{"generated_at":"2026-10-23T16:28:50.694Z","page":"https://ai.danielrosehill.com","focus_areas":[{"id":"code-switching","title":"English–Hebrew code-switching","summary":"Speech and text models that cope with \"Hebrish\" — the English-with-Hebrew-words that English speakers in Israel actually use. Off-the-shelf ASR mangles words like teudat zehut or mashkanta; I build the datasets and fine-tunes that fix that."},{"id":"asr","title":"Speech-to-text & on-device inference on Android","summary":"Whisper fine-tunes, ASR evaluations and small models that run locally on a phone. Most of my writing starts as voice, so I care about accuracy on real voice notes, background noise and fast on-device inference."},{"id":"accessibility","title":"Accessibility","summary":"Using AI to make technology easier to use: voice-first input instead of typing, TTS that pronounces Hebrew words correctly, and tools that reduce friction for people who work in more than one language."},{"id":"evals","title":"Benchmarks & AI safety","summary":"Evals for how models behave, not just how they score: does a model lecture the user, repeat propaganda when the premise is loaded, endorse conspiracy theories or assume everyone is American? Plus the speech benchmarks behind my ASR work. Open test prompts on Hugging Face."},{"id":"education","title":"AI for education","summary":"AI that teaches rather than just does: tutor-style agents, learning loops for code generation, and My Weird Prompts, a podcast that explains a new topic every day."}],"flagship":[{"title":"Hebrish (Sabich v1.0): Parakeet code-switching plus tech dictation","kind":"model","url":"https://huggingface.co/Hebrish/Sabich","description":"The recommended model in the Hebrish apps: Jachnun's code-switching data plus tech vocabulary, brands, numbers, Irish-English phrasing and my other task recordings, retrained from stock Parakeet v2. Level with Jachnun on code-switching (public benchmark term recall 73.8% against 73.3%), slightly better on tech (WER 1.54% against 1.74%). Also released on Qwen3-ASR 1.7B, Canary-Qwen 2.5B, Parakeet 110M and asr4all bases. Gated, access by request.","focus":["code-switching","asr","accessibility"],"page":"https://ai.danielrosehill.com/models/parakeet-tdt-v2-english-hebrew-all-subsets"},{"title":"Hebrish (Jachnun v1.0): Parakeet English–Hebrew code-switching","kind":"model","url":"https://huggingface.co/Hebrish/Jachnun","description":"On-device English dictation that writes Hebrew words correctly, in romanised Latin letters. Parakeet TDT 0.6B v2 fine-tuned on my own recordings, used through the Hebrish apps on Android, Ubuntu and Windows. Code-switched WER 11.8% to 5.4% (v1.0), plain English no worse. Use it if you mix Hebrew into English; for plain English, see Boxty. Site: hebrish.app.","focus":["code-switching","asr","accessibility"],"page":"https://ai.danielrosehill.com/models/parakeet-english-hebrew-codeswitching"},{"title":"Boxty v1.0: Parakeet English and tech dictation","kind":"model","url":"https://huggingface.co/Hebrish/Parakeet-TDT-0.6B-v2-Tech-And-Accent","description":"The sister model to Jachnun, for English dictation without Hebrew: tech vocabulary, brands, acronyms, numbers in digits and Irish-English phrasing, in my voice. Tech dictation WER 2.51% to 1.35%. It has none of the code-switching vocabulary, so use Jachnun if you mix Hebrew into English. Gated, access by request.","focus":["asr","accessibility"],"page":"https://ai.danielrosehill.com/models/parakeet-tdt-v2-tech-and-accent"},{"title":"English–Hebrew code-switching ASR benchmark","kind":"space","url":"https://huggingface.co/spaces/danielrosehill/English-Hebrew-Codeswitching-ASR-Leaderboard","description":"A replicable benchmark and leaderboard: stock and fine-tuned speech-to-text scored on how many Hebrew words in English speech they get right. Hebrish v1.0 (Jachnun): 73%; Whisper Large V3: 46%; stock Parakeet: 33%.","focus":["code-switching","asr","evals"]},{"title":"Hebrish Keyboard","kind":"repo","url":"https://github.com/Hebrish-ASR/Hebrish-Keyboard","description":"A full Android keyboard with Hebrish dictation on the mic key: a fork of HeliBoard, so there is no switching to a separate voice keyboard. GPL-3.0, arm64 only, early release. On the Hebrish F-Droid repository.","focus":["code-switching","asr","accessibility"]},{"title":"ASR Dataset Blender","kind":"space","url":"https://huggingface.co/spaces/danielrosehill/ASR-Dataset-Blender","description":"Builds a training mix from Hugging Face speech datasets: filters, de-duplication, held-out sets that keep test terms out of training, and a share per source. Downloads NeMo manifests and a recipe pinned to each source's commit, so a blend can be rebuilt exactly. Source: github.com/Hebrish-ASR/ASR-Dataset-Blender.","focus":["asr","code-switching"]},{"title":"The PC Bench","kind":"repo","url":"https://github.com/danielrosehill/The-PC-Bench","description":"Does the model scold you? An eval for lecturing, moralising, terminology correction and unrequested disclaimers on legitimate requests, with mirrored Western-progressive and conservative triggers to measure selectivity. 54 open test prompts.","focus":["evals"]},{"title":"Propaganda Bait Eval","kind":"repo","url":"https://github.com/danielrosehill/Propaganda-Bait-Eval","description":"Loaded political premises across six series (CCP, MAGA, pro-Israel, pro-Palestine, Kremlin, US-progressive), scored on how readily a model repeats each framing.","focus":["evals"]},{"title":"Whisper Hebrish","kind":"model","url":"https://huggingface.co/danielrosehill/Whisper-Hebrish","description":"Whisper Large v3 Turbo fine-tuned for English speech with Hebrew words mixed in, the way English speakers in Israel talk.","focus":["code-switching","asr","accessibility"]},{"title":"Whisper Hebrish demo","kind":"space","url":"https://huggingface.co/spaces/danielrosehill/Whisper-Hebrish","description":"Try the Hebrish model in the browser and compare it with stock Whisper.","focus":["code-switching","asr"]},{"title":"Hebrew-in-Latin token classifier","kind":"model","url":"https://huggingface.co/Hebrish/Hebrew-Latin-Token-Classifier","description":"Released: finds Hebrew words written in Latin letters inside English text (\"the makolet\"), so TTS can send them to a Hebrew voice. XLM-RoBERTa base, span F1 0.94 on a term-disjoint test set. No model for this existed before. Part of the Hebrish series since 2026-10-20. Demo Space and a Replicate API.","focus":["code-switching","accessibility"],"page":"https://ai.danielrosehill.com/models/hebrew-latin-token-classifier"},{"title":"Gemma 3 270M ASR Paragraph Breaks","kind":"model","url":"https://huggingface.co/danielrosehill/Gemma-3-270M-ASR-Paragraph-Breaks","description":"New (2026-10-20): adds paragraph breaks to flat speech-to-text output and changes nothing else, small enough to run on a phone after the ASR pass. Break F1 0.78 against 0.00 for the untrained model. ONNX export for ONNX Runtime.","focus":["asr"],"page":"https://ai.danielrosehill.com/models/gemma-3-270m-asr-paragraph-breaks"},{"title":"Gemma 3 270M Email Key-Line Bolding","kind":"model","url":"https://huggingface.co/danielrosehill/Gemma-3-270M-Email-Key-Line-Bolding","description":"Preliminary (2026-10-21): puts the key parts of a dictated email (the requests, the purpose, the main points) in bold and changes nothing else. Runs after the paragraph-breaks model. Word F1 0.62 on 6 hand-annotated emails; the untrained model fell back on all of them. Demo Space.","focus":["asr"],"page":"https://ai.danielrosehill.com/models/gemma-3-270m-email-key-line-bolding"},{"title":"Proximate Pseudonymizer 0.6B","kind":"model","url":"https://huggingface.co/danielrosehill/Proximate-Pseudonymizer-0.6B","description":"Proof of concept (2026-10-21): rewrites a text with identifying details swapped for close, plausible substitutes (Daniel to David, Jerusalem to Tel Aviv), used consistently. Not redaction. Qwen3-0.6B, Apache-2.0. It still leaves 9.3% of identifying details unchanged, so it is not safe to run unsupervised. Demo Space.","focus":[],"page":"https://ai.danielrosehill.com/models/proximate-pseudonymizer-0.6b"},{"title":"LLM Spoken Steganography","kind":"space","url":"https://huggingface.co/spaces/danielrosehill/LLM-Spoken-Steganography","description":"Proof of concept: an LLM hides an encrypted message in the first letters of a mundane podcast script's sentences, which survive text-to-speech, playback through the air and speech-to-text. Grew out of My Weird Prompts #6141. Two-phone demo in the browser.","focus":[]},{"title":"Pseudo-Personalised Spam Classifier","kind":"repo","url":"https://github.com/danielrosehill/Pseudo-Personalised-Spam-Classifier","description":"Planned: a small classifier that separates genuine email from strangers from mass outreach written to look personal, which gets past spam filters. Seed dataset of 18 pseudonymised emails on Hugging Face (2026-10-20); no model yet. The Pseudospam Filter plugin for Claude Code applies the same labelling standard today.","focus":[]},{"title":"Hebrew-in-Latin code-switching corpus","kind":"dataset","url":"https://huggingface.co/datasets/Hebrish/hebrew-latin-code-switching","description":"English sentences with Hebrew words written in Latin letters, labelled at token level (BIO): the training set for the classifier above. No comparable corpus existed when it was built.","focus":["code-switching","accessibility"]},{"title":"English–Hebrew Code-Switching Sentences","kind":"dataset","url":"https://huggingface.co/datasets/danielrosehill/English-Hebrew-Codeswitching-Sentences","description":"Single-speaker read-aloud audio of English sentences containing Hebrew words, for ASR fine-tuning.","focus":["code-switching","asr"],"page":"https://ai.danielrosehill.com/datasets/english-hebrew-codeswitching-sentences"},{"title":"English–Hebrew Mixed Sentences","kind":"dataset","url":"https://huggingface.co/datasets/danielrosehill/English-Hebrew-Mixed-Sentences","description":"The training set behind Whisper Hebrish: English sentences with everyday Hebrew vocabulary.","focus":["code-switching","asr"]},{"title":"Read-Aloud Voice Sentences","kind":"dataset","url":"https://huggingface.co/datasets/danielrosehill/Read-Aloud-Voice-Sentences","description":"Read-aloud English and Hebrew sentences for fine-tuning Parakeet and other ASR models.","focus":["asr","code-switching"]},{"title":"Tech Sentences","kind":"dataset","url":"https://huggingface.co/datasets/danielrosehill/Tech-Sentences","description":"Developer and AI vocabulary read aloud, for ASR that has to get technical terms right.","focus":["asr"]},{"title":"Whisper Hebrish (large-v3-turbo)","kind":"collection","url":"https://huggingface.co/collections/danielrosehill/whisper-hebrish-691bc2624553319eb727c1db","description":"Where Hebrish started, in November 2025: a full fine-tune of Whisper large-v3-turbo for English with Hebrew words, with its demo Space and sentence data.","focus":["code-switching","asr"]},{"title":"Whisper Hebrish Small ACFT","kind":"model","url":"https://huggingface.co/Hebrish/Whisper-Hebrish-Small-ACFT","description":"A phone-sized Whisper small.en fine-tune with FUTO's ACFT, for FUTO Keyboard and FUTO Voice Input on Android. 5.8% code-switching WER on its own test set.","focus":["code-switching","asr","accessibility"]},{"title":"On-device ASR on the OnePlus Nord 3","kind":"collection","url":"https://huggingface.co/collections/danielrosehill/on-device-asr-one-plus-cph2493-nord-3","description":"ASR models I have tested running locally on an Android phone (OnePlus Nord 3, CPH2493).","focus":["asr","accessibility"]},{"title":"Android voice keyboard roundup","kind":"repo","url":"https://github.com/danielrosehill/Android-Voice-Keyboard-Roundup","description":"An evaluation matrix of open-source Android voice keyboards and voice IMEs.","focus":["asr","accessibility"]},{"title":"Noisy Voice Notes","kind":"dataset","url":"https://huggingface.co/datasets/danielrosehill/Noisy-Voice-Notes","description":"Real voice notes recorded in rough conditions, for testing ASR beyond studio audio.","focus":["asr"]},{"title":"ASR WPM & background noise eval","kind":"dataset","url":"https://huggingface.co/datasets/danielrosehill/ASR-WPM-And-Background-Noise-Eval","description":"Annotated recordings that measure how speaking rate and background noise affect ASR accuracy.","focus":["asr"]},{"title":"Transcription Cleanup Trainer","kind":"dataset","url":"https://huggingface.co/datasets/danielrosehill/Transcription-Cleanup-Trainer","description":"Pairs for fine-tuning a model to clean raw speech-to-text output into readable text.","focus":["asr","accessibility"]},{"title":"Code Tutor Agent","kind":"repo","url":"https://github.com/danielrosehill/Code-Tutor-Agent-Model","description":"An agent configuration that guides a learner through a problem instead of doing it for them.","focus":["education"]},{"title":"Vibe Code And Learn","kind":"repo","url":"https://github.com/danielrosehill/Vibe-Code-And-Learn","description":"A multi-agent code-generation setup with an educational loop built in.","focus":["education"]}],"model_cards":[{"slug":"parakeet-english-hebrew-codeswitching","title":"Hebrish (Jachnun v1.0): Parakeet TDT 0.6B v2 for English–Hebrew code-switching","shortTitle":"Jachnun (Parakeet English–Hebrew code-switching)","tagline":"Speech-to-text that gets the Hebrew words right when an English speaker in Israel drops them into English: \"pick up some bamba from the makolet\".","status":"v1.0 · Jachnun","released":"2026-10-19","license":"CC-BY-4.0","baseModel":{"label":"nvidia/parakeet-tdt-0.6b-v2","url":"https://huggingface.co/nvidia/parakeet-tdt-0.6b-v2"},"hfUrl":"https://huggingface.co/Hebrish/Jachnun","website":{"label":"hebrish.app","url":"https://hebrish.app"},"trackerUrl":"https://github.com/danielrosehill/English-Hebrew-Codeswitching-Parakeet-Finetune","runsOn":"Android · Linux · Windows · NeMo","sibling":{"slug":"parakeet-tdt-v2-tech-and-accent","label":"Boxty","note":"Jachnun is for English with Hebrew words mixed in. For plain English and tech dictation with no Hebrew, use its sister model Boxty, which has none of the code-switching vocabulary."},"demoUrl":"https://huggingface.co/spaces/danielrosehill/English-Hebrew-Codeswitching-Parakeet-Demo","app":{"name":"Hebrish","repoUrl":"https://github.com/Hebrish-ASR/Hebrish-App","summary":"The Hebrish apps: an Android voice keyboard and tray apps for Ubuntu and Windows, running this fine-tune on-device. They download the model once (about 660 MB), check for newer releases, and transcribe fully offline. On the desktop, press F14 to dictate, or use the Transcribe tab to record or open an audio file and edit the transcript.","screenshot":{"src":"/images/ai/parlance/desktop-model.png","alt":"The desktop app (then named Parlance), Model tab: a model picker, its update status, and the installed versions","caption":"The desktop app on Linux, 2026-10-19, before it was renamed Hebrish: the Model tab."},"builds":[{"platform":"Linux (Debian / Ubuntu)","file":".deb, latest version","size":"about 11 MB + 660 MB model","note":"Install with sudo apt install ./hebrish_*_amd64.deb (the file you downloaded). Ubuntu 26.04, amd64.","url":"https://cdn.hebrish.app/apps/ubuntu/hebrish-latest-amd64.deb"},{"platform":"Android","file":"APK, latest version","size":"about 15 MB + 660 MB model","note":"arm64 phones, Android 9+. Works as the voice input behind HeliBoard's mic key, or try it straight away in the app's own notepad. Also on the Hebrish F-Droid repository, which keeps it updated.","url":"https://cdn.hebrish.app/apps/android/Hebrish-latest.apk"},{"platform":"Windows","file":".zip, latest version","size":"about 69 MB + 660 MB model","note":"Unzip and run Hebrish.exe. Not yet tested on real hardware, and unsigned, so SmartScreen may warn.","url":"https://cdn.hebrish.app/apps/windows/Hebrish-latest-windows-x64.zip"}]},"datasets":[{"slug":"english-hebrew-codeswitching-sentences","label":"English–Hebrew Code-Switching Sentences","url":"https://huggingface.co/datasets/danielrosehill/English-Hebrew-Codeswitching-Sentences"},{"label":"Read-Aloud Voice Sentences","url":"https://huggingface.co/datasets/danielrosehill/Read-Aloud-Voice-Sentences"},{"label":"LibriSpeech (replay)","url":"https://huggingface.co/datasets/openslr/librispeech_asr"}],"summary":["Hebrish is on-device English dictation for English speakers in Israel that writes the Hebrew words correctly, romanized: makolet, teudat zehut. It is a fine-tune of NVIDIA Parakeet TDT 0.6B v2, trained on my own recordings, and you use it through the Hebrish apps: an Android voice keyboard plus tray apps for Ubuntu and Windows. The project site is hebrish.app.","Stock Parakeet v2 is very good at English, but it has never seen romanized Hebrew. A Hebrew word in English speech comes out as whatever English words sound closest. The v2 tokenizer has no Hebrew characters, so this fine-tune teaches the model one romanized spelling for each Hebrew word.","It was trained on read-aloud sentences that I recorded on the Android phone that runs the model, plus some plain English so that it keeps its English. It runs fully offline on a phone or a Linux or Windows desktop.","This is v1.0 of the fine-tune, named Jachnun (2026-10-19), trained from stock Parakeet v2 on 339 minutes of code-switched speech. The five checkpoints published before it as v1 to v5 are now called PoC-1 to PoC-5. The code-switching line takes Israeli food names.","Since 2026-10-21 the Hebrish apps recommend Sabich v1.0, trained on the same code-switching data plus tech vocabulary and the other task recordings. It is level with Jachnun on code-switching and slightly better on tech terms; Jachnun stays in the apps as the pure code-switching model. See Sabich's own card page.","I chose v2 over Parakeet v3 on purpose. In my own dictation v2 was more accurate for English; v3 adds 25 languages and gives up some English accuracy."],"example":{"caption":"Recorded in the demo Space, 2026-10-15","base":"I want to go to the Micollette today to buy my groceries.","tuned":"I want to go to the makolet today to buy my groceries.","baseWord":"Micollette","tunedWord":"makolet","gloss":"Makolet (מכולת) is the Israeli corner grocery shop. Stock v2 has never heard the word, so it guesses an English-looking name."},"downloads":[{"label":"Hebrish APK","size":"14.5 MB + 660 MB model","use":"The Hebrish Android voice keyboard (latest version). Downloads the model once, then runs fully offline. arm64 phones, Android 9+.","runWith":{"label":"HeliBoard (its mic key opens it)","url":"https://github.com/Helium314/HeliBoard/releases"},"url":"https://cdn.hebrish.app/apps/android/Hebrish-latest.apk"},{"label":"GGUF Q8_0","size":"730 MB","use":"Android phone or Linux desktop. Take this one if unsure.","runWith":{"label":"Offline Voice Input (Android) / Handy (Linux)","url":"https://play.google.com/store/apps/details?id=dev.notune.transcribe"},"url":"https://huggingface.co/danielrosehill/Parakeet-TDT-0.6B-v2-English-Hebrew-Codeswitching/resolve/main/gguf/Parakeet-TDT-0.6B-v2-English-Hebrew-Codeswitching-Q8_0.gguf","recommended":true},{"label":"GGUF Q4_K_M","size":"475 MB","use":"Phones with 4 GB of RAM or less, or if Q8_0 is too slow.","runWith":{"label":"Offline Voice Input","url":"https://play.google.com/store/apps/details?id=dev.notune.transcribe"},"url":"https://huggingface.co/danielrosehill/Parakeet-TDT-0.6B-v2-English-Hebrew-Codeswitching/resolve/main/gguf/Parakeet-TDT-0.6B-v2-English-Hebrew-Codeswitching-Q4_K_M.gguf"},{"label":".nemo checkpoint","size":"2.47 GB","use":"Python, GPU inference or further fine-tuning, at full precision.","runWith":{"label":"NVIDIA NeMo","url":"https://github.com/NVIDIA-NeMo/Speech"},"url":"https://huggingface.co/danielrosehill/Parakeet-TDT-0.6B-v2-English-Hebrew-Codeswitching/resolve/main/Parakeet-TDT-0.6B-v2-English-Hebrew-Codeswitching.nemo"}],"results":{"asOf":"2026-10-20","run":"r6-en30","version":"v1.0 · Jachnun","rows":[{"test":"Code-switching, shared test set, 422 clips (normalised WER)","base":"11.82%","tuned":"5.42%","better":true},{"test":"Hebrew terms held out of training: share correct (324 occurrences, 134 terms)","base":"23.8%","tuned":"46.9%","better":true},{"test":"Speaker's own plain English, 84 clips (normalised WER)","base":"2.81%","tuned":"1.56%","better":true},{"test":"LibriSpeech test-clean, 200 utterances (normalised WER)","base":"1.23%","tuned":"1.20%"},{"test":"Long form, 17 files (normalised WER)","base":"10.34%","tuned":"5.84%","better":true}],"notes":["Scored on the shared test set of 2026-10-20 (dataset revision f956dbb184c2, eval tag r7-r6-en30), the same clips used for Boxty, so the two models compare directly. Stock is nvidia/parakeet-tdt-0.6b-v2 on the same clips (tag r7-baseline).","WER is normalised: lowercase, punctuation removed. None of these clips were trained on or used to pick the checkpoint.","LibriSpeech is within noise of stock, so general English was not lost.","The checkpoints published as v1 to v5 are now PoC-1 to PoC-5; v1.0 is the first release scored on this test set.","The charts below are v1.0's earlier scoring (2026-10-19): the held-out split at dataset revision 550bfcaa0f84 and the public benchmark (test set v1, 84 clips no model trained on). The figures differ slightly from the table because the test sets differ."]},"figures":[{"src":"/images/ai/parakeet-codeswitching/v1-heldout-compare.png","alt":"Held-out split at dataset revision 550bfcaa0f84, 387 code-switched clips, scored 2026-10-19: stock v2, PoC-3, v1.0 (Jachnun) and the Nemotron fine-tune. Code-switching WER 11.3%, 6.0%, 5.1%, 5.8%; held-out terms correct 25.9%, 45.7%, 47.8%, 44.4%; seen test terms correct 27.7%, 63.3%, 72.7%, 68.0%"},{"src":"/images/ai/parakeet-codeswitching/v1-benchmark-term-recall.png","alt":"Hebrew term recall on the public benchmark (test set v1, clean subset, 84 clips, 172 term occurrences, scored 2026-10-19): Hebrish v1.0 (Jachnun) 73.3%, PoC-3 67.4%, PoC-2 65.7%, PoC-1 62.2%, Nemotron streaming 62.2%, Whisper Hebrish v1 54.1%, Whisper Large V3 45.9%, Whisper Large V3 Turbo 40.1%, stock Parakeet v2 32.6%, stock Parakeet v3 31.4%"},{"src":"/images/ai/parakeet-codeswitching/v1-forgetting-check.png","alt":"English did not get worse (dataset revision 550bfcaa0f84, scored 2026-10-19): own English WER 3.1% stock v2 against 1.7% v1.0; LibriSpeech test-clean 1.23% against 1.20%"},{"src":"/images/ai/parakeet-codeswitching/baseline_terms.png","alt":"How stock v2 handles 1,139 Hebrew terms before fine-tuning: always right, sometimes right, never right"}],"training":["v1.0 · Jachnun: 339 min of code-switched speech with a 30% English share, full fine-tune from stock v2 at dataset revision 550bfcaa0f84. AdamW, lr 3e-5, 20 epochs, batch 16; 41 minutes on one L40S; best checkpoint at epoch 18. Run name r6-en30.","Mix: 339.2 min of code-switched speech, 95.3 min of my own plain English and 48.5 min of LibriSpeech replay, about 30% English. Fine-tuning on code-switched speech alone is known to damage a model's English; the English share limits it.","Each release is a full fine-tune from stock v2 on all the data so far, not continued training from the previous checkpoint. Trained in NeMo; the run cost about US$0.80 on Modal.","Before v1.0 came five proof-of-concept checkpoints, PoC-1 to PoC-5 (published as v1 to v5, 2026-10-15 to 2026-10-18), trained on 77 to 294 min of code-switching. PoC-4 and PoC-5 were never benchmarked."],"limitations":["Trained on one voice so far: an English speaker with an Irish accent living in Jerusalem. Recordings from more speakers are the next step; contribute yours at hebrish.app/contact?topic=samples.","About 5.7 hours of code-switched audio, with most Hebrew terms recorded only once or twice. Terms the model never heard in training are right only about half the time (46.9% for v1.0), and more audio has not moved that much since PoC-1.","Spellings follow one convention (the Hebrew-Latin-Token-Classifier lexicon), so the model may prefer it over other romanizations.","On-device accuracy and phone speed are not measured yet beyond a five-clip GGUF check."],"checkpoints":[{"label":"Baseline","state":"done","date":"2026-10-15","note":"Stock v2 got 28% of Hebrew term occurrences right across 615 terms."},{"label":"PoC-1 pilot (published as v1, pilot-en30)","state":"done","date":"2026-10-15","note":"First release: 77 min of code-switching, 30% English. Superseded by PoC-2; kept in the Hugging Face repo history."},{"label":"GGUF export","state":"done","date":"2026-10-15","note":"transcribe.cpp v0.1.3; identical transcripts to PyTorch on 5 test clips."},{"label":"PoC-2 retrain (published as v2, r2-en30)","state":"done","date":"2026-10-15","note":"Released: 120 min of code-switching, 28% English, trained from stock v2. Same files and download links; .nemo plus GGUF Q8_0 and Q4_K_M."},{"label":"ONNX and sherpa-onnx export","state":"done","date":"2026-10-16","note":"Separate -ONNX and -sherpa-onnx-int8 repos, for Voxtype, onnx-asr and sherpa-onnx apps."},{"label":"PoC-3 retrain (published as v3, r3-en30)","state":"done","date":"2026-10-18","note":"Released: 196 min of code-switching, 27% English, trained from stock v2. Every version re-scored on a new 219-clip test split. Same download links; PoC-1 and PoC-2 pinned in the model card."},{"label":"Codeswitch Voice APK","state":"done","date":"2026-10-18","note":"v0.1.0 of my own Android voice keyboard (proof of concept): sherpa-onnx int8 PoC-3 weights, on-device, companion to HeliBoard."},{"label":"Desktop dictation app","state":"done","date":"2026-10-18","note":"v0.2.0 of the desktop app (Linux .deb, untested Windows build): hotkey dictation plus a Transcribe tab for recordings and audio files. The Android app moved into the same repo as v0.5.0."},{"label":"PoC-4 retrain (published as v4, r4-en30)","state":"done","date":"2026-10-18","note":"Released before its benchmark: 269 min of code-switching, the first run on the 70/20/10 mix target (70.8% / 19.1% / 10.1%)."},{"label":"PoC-5 retrain (published as v5, r5-en30)","state":"done","date":"2026-10-18","note":"Released before its benchmark: 294 min of code-switching, 70/20/10 mix, trained early ahead of the 360-min checkpoint. .nemo, GGUF (Q8_0, Q4_K_M, F16), ONNX and sherpa-onnx refreshed; the apps pick it up on their own. MLX stays PoC-3."},{"label":"Nemotron streaming fine-tune","state":"done","date":"2026-10-18","note":"Experimental second base model: the PoC-5 data and recipe on NVIDIA nemotron-speech-streaming-en-0.6b, for live streaming text. Later scored offline: 5.8% code-switching WER on the held-out split, against 5.1% for v1.0."},{"label":"Parlance","state":"done","date":"2026-10-19","note":"The Android and desktop apps became Parlance (Android v0.1.0, desktop v0.1.0), with a choice of stock Parakeet or this fine-tune."},{"label":"Hebrish apps 0.2.0","state":"done","date":"2026-10-19","note":"The apps were renamed Hebrish (repo now Hebrish-ASR/Hebrish-App): Android APK, Ubuntu .deb and Windows zip, desktop hotkey F14. Parlance continues separately as a general-purpose Parakeet dictation app."},{"label":"v1.0 · Jachnun (r6-en30)","state":"done","date":"2026-10-19","note":"The first release after five proof-of-concept checkpoints: 339 min of code-switching, 30% English, trained from stock v2. The checkpoints published as v1 to v5 became PoC-1 to PoC-5. MLX stays PoC-3."},{"label":"Hebrish apps: Android 0.5.0, desktop 0.2.2","state":"done","date":"2026-10-20","note":"Current app releases, linked in the downloads above. The repo was renamed Hebrish-ASR/Hebrish-App."},{"label":"Boxty v1.0","state":"done","date":"2026-10-20","note":"A sister fine-tune from the same base for English dictation without Hebrew: tech vocabulary and accent. See its own card page."},{"label":"Sabich v1.0","state":"done","date":"2026-10-20","note":"A second Hebrish model: Jachnun's data plus tech vocabulary and the other task recordings, retrained from stock v2. Level with Jachnun on code-switching; in the apps' model list. See its own card page."},{"label":"Data collection","state":"in-progress","note":"Recording continues, weighted toward terms stock v2 never gets right and terms with only one or two takes."},{"label":"v1.1","state":"planned","note":"The next retrain on more code-switching audio. A comparison of 70/30, 80/20, 90/10 and 100/0 code-switching / English remains an option."},{"label":"On-phone test","state":"planned","note":"Measure accuracy and speed of the on-device builds on the target phone."}],"related":[{"label":"Sabich v1.0 (the recommended Hebrish model)","url":"https://hebrish.app/models/sabich","note":"Jachnun's data plus tech vocabulary and other task recordings; in the same apps"},{"label":"PoC record","url":"https://huggingface.co/danielrosehill/Parakeet-TDT-0.6B-v2-English-Hebrew-Codeswitching","note":"The first repo: the PoC-1 to PoC-5 checkpoints, evals and figures"},{"label":"Boxty (sister model)","url":"https://huggingface.co/Hebrish/Parakeet-TDT-0.6B-v2-Tech-And-Accent","note":"The same base fine-tuned for English and tech dictation without Hebrew"},{"label":"Hebrish website","url":"https://hebrish.app","note":"The project site: what it does, how to install it, and the latest results"},{"label":"Project tracker (GitHub)","url":"https://github.com/danielrosehill/English-Hebrew-Codeswitching-Parakeet-Finetune","note":"Status, lifecycle, results and log"},{"label":"Demo Space","url":"https://huggingface.co/spaces/danielrosehill/English-Hebrew-Codeswitching-Parakeet-Demo","note":"Stock v2 and the fine-tune side by side"},{"label":"Benchmark leaderboard","url":"https://huggingface.co/spaces/danielrosehill/English-Hebrew-Codeswitching-ASR-Leaderboard","note":"v1.0 against stock Whisper, Parakeet and Whisper-Hebrish on the same frozen test set"},{"label":"Benchmark write-up (blog)","url":"https://blog.danielrosehill.com/posts/english-hebrew-code-switching-asr-benchmark","note":"\"Can speech-to-text spell makolet?\": what the benchmark measures, what it caught, and the results"},{"label":"Jachnun and Sabich in depth (blog)","url":"https://blog.danielrosehill.com/posts/jachnun-sabich-first-stable-hebrish-models-krjelm","note":"The first stable releases: data, training recipe, results against stock Parakeet and Whisper, formats, and the white paper"},{"label":"Hebrish (blog)","url":"https://blog.danielrosehill.com/posts/hebrish-english-hebrew-speech-models-s5vtam","note":"The project: its mission, every model it has published (Jachnun, Sabich, Shakshuka, the Whisper fine-tunes) and hebrish.app"},{"label":"Voice typing that understands Hebrish (blog)","url":"https://blog.danielrosehill.com/posts/voice-typing-that-understands-hebrish/","note":"The apps, the models, how they were made and how Jachnun v1.0 and Sabich v1.0 score (earlier version posts are folded into it)"},{"label":"Boxty v1.0 (blog)","url":"https://blog.danielrosehill.com/posts/boxty-v1-english-dictation-without-hebrew-mhsv9w/","note":"The no-Hebrew sister model, and how to choose between the two"},{"label":"Shakshuka (streaming)","url":"https://hebrish.app/models/shakshuka","note":"The streaming line on NVIDIA's Nemotron base, for live text. v1.0 (2026-10-21) is trained on Sabich's data; the PoC checkpoints are at Hebrish/Nemotron-Speech-Streaming-EN-0.6B-English-Hebrew-Codeswitching"},{"label":"Hebrish apps (desktop and Android)","url":"https://github.com/Hebrish-ASR/Hebrish-App","note":"The on-device dictation apps for this model: Android voice keyboard, Ubuntu and Windows"},{"label":"Parlance","url":"https://github.com/danielrosehill/Parlance-Dictation","note":"My general-purpose on-device Parakeet dictation app, with stock or fine-tuned models"},{"label":"Offline Voice Input","url":"https://play.google.com/store/apps/details?id=dev.notune.transcribe","note":"Android voice keyboard that loads the GGUF"},{"label":"Handy","url":"https://github.com/cjpais/Handy","note":"Linux desktop dictation; loads the same GGUF"},{"label":"Hebrew-Latin token classifier","url":"https://github.com/danielrosehill/Hebrew-Latin-Token-Classifier","note":"Its lexicon sets the spelling of every Hebrew word"},{"label":"Whisper Hebrish","url":"https://huggingface.co/danielrosehill/Whisper-Hebrish","note":"The earlier Whisper fine-tune for the same problem"}],"tags":["Hebrish","ASR","Parakeet","NeMo","code-switching","Hebrew","English","on-device","GGUF","Android"],"page":"https://ai.danielrosehill.com/models/parakeet-english-hebrew-codeswitching"},{"slug":"parakeet-tdt-v2-english-hebrew-all-subsets","title":"Hebrish (Sabich v1.0): Parakeet TDT 0.6B v2 for code-switching plus tech dictation","shortTitle":"Sabich (Parakeet code-switching plus tech)","tagline":"The recommended Hebrish model: everything Jachnun was trained on, plus tech vocabulary, brand names, acronyms, numbers and currency, Irish-English phrasing and the rest of my task recordings.","status":"v1.0 · Sabich","released":"2026-10-20","license":"CC-BY-4.0","baseModel":{"label":"nvidia/parakeet-tdt-0.6b-v2","url":"https://huggingface.co/nvidia/parakeet-tdt-0.6b-v2"},"hfUrl":"https://huggingface.co/Hebrish/Sabich","website":{"label":"hebrish.app","url":"https://hebrish.app/models/sabich"},"gated":"The model repos (.nemo, GGUF, ONNX, MLX and the sherpa-onnx build the apps use) are gated: request access on Hugging Face. The apps download it for you.","runsOn":"Android · Linux · Windows · NeMo","sibling":{"slug":"parakeet-english-hebrew-codeswitching","label":"Jachnun","note":"Sabich and Jachnun are level on code-switching; Sabich is slightly better on tech vocabulary, which is why it is the recommended pick. Both run in the Hebrish apps, so switch in the model list if Jachnun suits you better."},"app":{"name":"Hebrish","repoUrl":"https://github.com/Hebrish-ASR/Hebrish-App","summary":"The Hebrish apps (Android voice keyboard, tray apps for Ubuntu and Windows) recommend Sabich, with Jachnun and stock Parakeet in the same model list. They download the model once and transcribe fully offline.","screenshot":{"src":"/images/ai/parlance/desktop-model.png","alt":"The desktop app (then named Parlance), Model tab: a model picker, its update status, and the installed versions","caption":"The desktop app on Linux, 2026-10-19, before it was renamed Hebrish: the Model tab."},"builds":[{"platform":"Linux (Debian / Ubuntu)","file":".deb, latest version","size":"about 11 MB + 660 MB model","note":"Install with sudo apt install ./hebrish_*_amd64.deb (the file you downloaded). Ubuntu 26.04, amd64.","url":"https://cdn.hebrish.app/apps/ubuntu/hebrish-latest-amd64.deb"},{"platform":"Android","file":"APK, latest version","size":"about 15 MB + 660 MB model","note":"arm64 phones, Android 9+. Works as the voice input behind HeliBoard's mic key, or try it straight away in the app's own notepad. Also on the Hebrish F-Droid repository, which keeps it updated.","url":"https://cdn.hebrish.app/apps/android/Hebrish-latest.apk"},{"platform":"Windows","file":".zip, latest version","size":"about 69 MB + 660 MB model","note":"Unzip and run Hebrish.exe. Not yet tested on real hardware, and unsigned, so SmartScreen may warn.","url":"https://cdn.hebrish.app/apps/windows/Hebrish-latest-windows-x64.zip"}]},"datasets":[{"slug":"english-hebrew-codeswitching-sentences","label":"English–Hebrew Code-Switching Sentences","url":"https://huggingface.co/datasets/danielrosehill/English-Hebrew-Codeswitching-Sentences"},{"label":"All training datasets (collection)","url":"https://huggingface.co/collections/Hebrish/parakeet-v2-fine-tune-training-data-6ad63c739d300685a62edb3e"},{"label":"LibriSpeech (replay)","url":"https://huggingface.co/datasets/openslr/librispeech_asr"}],"summary":["Sabich is the second Hebrish model: a fine-tune of NVIDIA Parakeet TDT 0.6B v2 trained on Jachnun's data plus everything else I have recorded. That means the code-switched speech, plus tech vocabulary, brand names, acronyms, numbers and currency, Irish-English phrasing, disfluencies and self-corrections.","It is not trained on top of Jachnun. Every model in the family is retrained from stock Parakeet v2 on the cumulative data: PoC-1 to PoC-5, then Jachnun v1.0, then Sabich v1.0. Boxty, a separate fine-tune for English without Hebrew (not part of Hebrish), starts from the same base.","The question was whether adding all the other data would cost code-switching accuracy. It did not. On the public benchmark Sabich and Jachnun are level within noise, and Sabich is slightly better on tech vocabulary. That settled shipping both models in the Hebrish apps.","Sabich is named after the Israeli pita with a bit of everything in it. The code-switching line uses Israeli food names."],"downloads":[],"results":{"asOf":"2026-10-20","run":"r7-all","version":"v1.0 · Sabich","rows":[{"test":"Code-switching, held-out split (normalised WER)","base":"11.82%","tuned":"5.30%","better":true},{"test":"Hebrew terms held out of training: share correct","base":"23.8%","tuned":"48.1%","better":true},{"test":"Tech dictation (normalised WER)","base":"2.51%","tuned":"1.54%","better":true},{"test":"LibriSpeech (normalised WER)","base":"1.23%","tuned":"1.30%"}],"notes":["Against Jachnun v1.0 on the same held-out split (same dataset revision): code-switching WER 5.30% against 5.42%; held-out terms 48.1% against 46.9%; tech WER 1.54% against 1.74%; LibriSpeech 1.30% against 1.20%.","Public benchmark (84 clean clips, round 4), Sabich against Jachnun: Hebrew term recall 73.8% against 73.3%; unseen-term recall 68.4% against 70.1%; WER 3.27% against 3.32%; English control WER 1.99% against 2.17%.","Read: level on code-switching (the differences are noise), a modest gain on tech vocabulary, and no loss of code-switching accuracy from the extra data.","WER is normalised: lowercase, punctuation removed."]},"figures":[],"training":["v1.0 · Sabich: 560.5 min of my own recordings (373.1 min of them code-switching) plus 62.5 min of LibriSpeech replay. 20 epochs on one Modal L40S, 52 minutes. Run name r7-all.","Full fine-tune from stock Parakeet TDT 0.6B v2, not continued training from Jachnun. Jachnun v1.0 (r6-en30) used 339.2 min of code-switching, 95.3 min of my own English and 48.5 min of LibriSpeech replay."],"limitations":["Trained on one voice so far: an English speaker with an Irish accent living in Jerusalem. Recordings from more speakers are the next step; contribute yours at hebrish.app/contact?topic=samples.","On code-switching it is level with Jachnun, not better: Hebrew terms the model never heard in training are still right only about half the time.","The model repos are gated, with access by request."],"checkpoints":[{"label":"v1.0 · Sabich (r7-all)","state":"done","date":"2026-10-20","note":"First release: Jachnun's data plus every other task recording, 560.5 min of own recordings plus 62.5 min replay, trained from stock v2. Added to the Hebrish apps' model list."},{"label":"Recommended in the apps","state":"done","date":"2026-10-21","note":"Sabich became the recommended pick in the Hebrish apps, with Jachnun kept alongside it."},{"label":"Sabich on other bases","state":"done","date":"2026-10-21","note":"The same training data on four more bases: Qwen3-ASR 1.7B (5.4% code-switching WER, stock 10.0%) and Canary-Qwen 2.5B (6.9%) for desktop GPUs, Parakeet TDT-CTC 110M for small devices, and asr4all S and M (10.1% and 9.1%) for wearables. Not in the apps yet. Shakshuka v1.0, the streaming line, is trained on it too."}],"related":[{"label":"Sabich on hebrish.app","url":"https://hebrish.app/models/sabich","note":"What it is, how it compares with Jachnun, and the files"},{"label":"Jachnun (sister model)","url":"https://huggingface.co/Hebrish/Jachnun","note":"The code-switching model Sabich builds its data on"},{"label":"App build (sherpa-onnx int8)","url":"https://huggingface.co/Hebrish/Sabich-sherpa-onnx-int8","note":"The build the Hebrish apps download (gated on Hugging Face)"},{"label":"Sabich-Qwen3-ASR-1.7B","url":"https://hebrish.app/models/sabich-qwen3","note":"The most accurate Sabich, for computers with a GPU"},{"label":"Sabich-Canary-Qwen-2.5B","url":"https://hebrish.app/models/sabich-canary-qwen","note":"Desktop GPU; the strongest Sabich on plain English"},{"label":"Sabich-Parakeet-110M and Sabich-asr4all","url":"https://hebrish.app/models/sabich-110m","note":"Small-device and wearable builds of the same data"},{"label":"Public benchmark","url":"https://huggingface.co/datasets/Hebrish/English-Hebrew-Codeswitching-ASR-Benchmark","note":"The 84-clip test set Sabich and Jachnun were scored on (round 4)"},{"label":"Boxty (separate, not part of Hebrish)","url":"https://huggingface.co/Hebrish/Parakeet-TDT-0.6B-v2-Tech-And-Accent","note":"The same base fine-tuned for English and tech dictation without Hebrew"},{"label":"Hebrish apps (desktop and Android)","url":"https://github.com/Hebrish-ASR/Hebrish-App","note":"Android voice keyboard, Ubuntu and Windows"}],"tags":["Hebrish","Sabich","ASR","Parakeet","NeMo","code-switching","Hebrew","English","tech vocabulary","on-device"],"page":"https://ai.danielrosehill.com/models/parakeet-tdt-v2-english-hebrew-all-subsets"},{"slug":"parakeet-tdt-v2-tech-and-accent","title":"Boxty v1.0: Parakeet TDT 0.6B v2 for English and tech dictation","shortTitle":"Boxty (Parakeet English and tech dictation)","tagline":"English dictation without Hebrew, in my voice: model and product names, CLI and cloud terms, brands, acronyms, and numbers and currency in digits.","status":"v1.0 · Boxty","released":"2026-10-20","license":"CC-BY-4.0","baseModel":{"label":"nvidia/parakeet-tdt-0.6b-v2","url":"https://huggingface.co/nvidia/parakeet-tdt-0.6b-v2"},"hfUrl":"https://huggingface.co/Hebrish/Parakeet-TDT-0.6B-v2-Tech-And-Accent","gated":"The model repo is gated: request access on Hugging Face.","runsOn":"Android · Linux · NeMo","sibling":{"slug":"parakeet-english-hebrew-codeswitching","label":"Jachnun","note":"Boxty has none of the code-switching vocabulary. Use Jachnun if you mix Hebrew words into English; use Boxty for plain English and tech dictation."},"datasets":[{"label":"Parakeet V2 Blend: No Hebrew","url":"https://huggingface.co/datasets/Hebrish/Parakeet-V2-Blend-No-Hebrew"},{"label":"LibriSpeech (replay)","url":"https://huggingface.co/datasets/openslr/librispeech_asr"}],"summary":["Boxty is a fine-tune of NVIDIA Parakeet TDT 0.6B v2 for English dictation with no Hebrew in it. It targets tech vocabulary (model and product names, CLI and cloud terms), brand names, acronyms, numbers and currency written in digits, homophones, Irish-English phrasing, and dictated emails and messages, all in my voice.","It is the sister model to Jachnun, the English–Hebrew code-switching fine-tune from the same base. Boxty has none of the code-switching vocabulary, so code-switched speech stays at stock level on purpose. If you mix Hebrew into your English, use Jachnun.","Boxty is named after the Irish potato pancake. The no-Hebrew line uses Irish food names; the code-switching line uses Israeli ones.","You can use it through Parlance, my general-purpose on-device dictation app for Android and Linux, whose model list includes Boxty."],"downloads":[{"label":"GGUF Q8_0","size":"730 MB","use":"Android phone or Linux desktop. Take this one if unsure. Needs Hugging Face access to the gated repo.","runWith":{"label":"Parlance / Offline Voice Input (Android) / Handy (Linux)","url":"https://github.com/danielrosehill/Parlance-Dictation"},"url":"https://huggingface.co/Hebrish/Parakeet-TDT-0.6B-v2-Tech-And-Accent/resolve/main/gguf/Parakeet-TDT-0.6B-v2-Tech-And-Accent-Q8_0.gguf","recommended":true},{"label":"GGUF Q4_K_M","size":"475 MB","use":"Phones with 4 GB of RAM or less, or if Q8_0 is too slow.","runWith":{"label":"Offline Voice Input","url":"https://play.google.com/store/apps/details?id=dev.notune.transcribe"},"url":"https://huggingface.co/Hebrish/Parakeet-TDT-0.6B-v2-Tech-And-Accent/resolve/main/gguf/Parakeet-TDT-0.6B-v2-Tech-And-Accent-Q4_K_M.gguf"},{"label":".nemo checkpoint","size":"2.47 GB","use":"Python, GPU inference or further fine-tuning, at full precision.","runWith":{"label":"NVIDIA NeMo","url":"https://github.com/NVIDIA-NeMo/Speech"},"url":"https://huggingface.co/Hebrish/Parakeet-TDT-0.6B-v2-Tech-And-Accent/resolve/main/Parakeet-TDT-0.6B-v2-Tech-And-Accent.nemo"}],"results":{"asOf":"2026-10-20","run":"r7-general","version":"v1.0 · Boxty","rows":[{"test":"Hebrew-free English (normalised WER)","base":"1.79%","tuned":"1.04%","better":true},{"test":"Tech dictation (normalised WER)","base":"2.51%","tuned":"1.35%","better":true},{"test":"LibriSpeech (normalised WER)","base":"1.23%","tuned":"1.18%"},{"test":"Code-switched speech (normalised WER)","base":"11.5%","tuned":"stock level, on purpose"}],"notes":["Against Jachnun on the same tests: Hebrew-free English 1.04% against 1.23%; tech dictation 1.35% against 1.74%.","Code-switched speech stays at stock level (11.5%) because the training data has no Hebrew in it."]},"figures":[],"training":["142.5 min of my own recordings with no Hebrew in them, plus 15.9 min of LibriSpeech replay to keep general English.","Fine-tuned from stock Parakeet TDT 0.6B v2, the same base as Jachnun."],"limitations":["One speaker: an English speaker with an Irish accent living in Jerusalem. Gains on other voices are untested.","No code-switching vocabulary: Hebrew words in English speech come out at stock level. Use Jachnun for that.","The model repo is gated, with access by request."],"checkpoints":[{"label":"v1.0 · Boxty","state":"done","date":"2026-10-20","note":"First release: 142.5 min of Hebrew-free recordings plus 15.9 min of LibriSpeech replay."}],"related":[{"label":"Jachnun (sister model)","url":"https://huggingface.co/Hebrish/Jachnun","note":"The English–Hebrew code-switching fine-tune from the same base"},{"label":"Boxty v1.0 write-up (blog)","url":"https://blog.danielrosehill.com/posts/boxty-v1-english-dictation-without-hebrew-mhsv9w/","note":"What Boxty is for and how it compares"},{"label":"Parlance","url":"https://github.com/danielrosehill/Parlance-Dictation","note":"General-purpose dictation app for Android and Linux; its model list includes Boxty"},{"label":"Training data","url":"https://huggingface.co/datasets/Hebrish/Parakeet-V2-Blend-No-Hebrew","note":"The Hebrew-free blend Boxty was trained on"},{"label":"All training datasets","url":"https://huggingface.co/collections/Hebrish/parakeet-v2-fine-tune-training-data-6ad63c739d300685a62edb3e","note":"Collection of every dataset behind the Parakeet v2 fine-tunes"}],"tags":["Boxty","ASR","Parakeet","NeMo","English","tech vocabulary","dictation","Irish English"],"page":"https://ai.danielrosehill.com/models/parakeet-tdt-v2-tech-and-accent"},{"slug":"hebrew-latin-token-classifier","title":"Hebrew-Latin Token Classifier: finds Hebrew written in Latin letters","shortTitle":"Hebrew-Latin Token Classifier","eyebrow":"Model · token classification · Hebrish","licenseUrl":"https://opensource.org/licenses/MIT","tagline":"Finds Hebrew words written in Latin letters inside English text, so text-to-speech can hand them to a Hebrew voice instead of reading makolet as English. There was no model for this before.","status":"v1 · released","released":"2026-10-15","license":"MIT","baseModel":{"label":"FacebookAI/xlm-roberta-base","url":"https://huggingface.co/FacebookAI/xlm-roberta-base"},"hfUrl":"https://huggingface.co/Hebrish/Hebrew-Latin-Token-Classifier","website":{"label":"Model page on hebrish.app","url":"https://hebrish.app/models/token-classifier"},"trackerUrl":"https://github.com/Hebrish-ASR/Hebrew-Latin-Token-Classifier","demoUrl":"https://huggingface.co/spaces/danielrosehill/hebrew-latin-token-classifier","datasets":[{"label":"Hebrish/hebrew-latin-code-switching (4,737 sentences)","url":"https://huggingface.co/datasets/Hebrish/hebrew-latin-code-switching"}],"spanExamples":[{"text":"I need to go to Bituach Leumi before the makolet closes.","spans":["Bituach Leumi","makolet"]},{"text":"My savta says the arnona went up again.","spans":["savta","arnona"]}],"summary":["The model marks every span of romanized Hebrew in an English sentence, with character offsets and a confidence score. Multi-word terms come out as one span (Bituach Leumi, kupat holim). Labels: O, B-HE, I-HE.","It exists because nothing else does this. GlotLID, lid.176 and lid218e have no heb_Latn label, and a language detector sees Latin script inside an English sentence and calls it English. No token-level code-switching model covered Hebrew.","It was built for text-to-speech. English TTS reads makolet as \"MACK-oh-let\"; the fix is to find the Hebrew spans, convert them to Hebrew script and send each to a Hebrew voice. This model is pass 1 of that pipeline (detect, transliterate, add niqqud, tag, speak). The first user is My Weird Prompts, an AI-generated podcast whose episodes about life in Israel are full of these words.","The labelling test is pronunciation, not etymology. Words English already says acceptably (Shabbat, hummus, challah, kosher) are deliberately not flagged."],"downloads":[],"results":{"asOf":"2026-10-15","run":"xlm-roberta-base, 8 epochs","version":"v1","baseLabel":"Span precision / recall","tunedLabel":"Span F1","metricNote":"Scored on the term-disjoint test split: no Hebrew term in the test set appears in training, so this measures how well it generalises to words it has never seen. Higher is better.","rows":[{"test":"All test sentences (468)","base":"0.928 / 0.953","tuned":"0.941","better":true},{"test":"Generated (430)","base":"0.928 / 0.948","tuned":"0.938"},{"test":"Dictated, human-chosen (38)","base":"0.929 / 1.000","tuned":"0.963","better":true},{"test":"Validation, used for model selection (467)","base":"0.950 / 0.948","tuned":"0.949"}],"notes":["Hard negatives: 0 false positives in 58 sentences built to trap it (substrings, English homographs such as dud, tor and salon, lookalikes, anglicised loanwords).","Of the 41 test sentences it gets wrong, 14 are gaps in the gold labels rather than model errors: Hebrew words the blind annotator refused and no human has yet re-reviewed.","For comparison, the closest published recipe (EnTaCs, English–Tamil, arXiv 2603.26587) reports F1 0.638 on its romanized class."]},"figures":[{"src":"/images/ai/token-classifier/results-by-slice.png","alt":"Span precision, recall and F1 by evaluation slice"},{"src":"/images/ai/token-classifier/error-breakdown.png","alt":"Breakdown of the 41 wrong test sentences by cause, 14 of them label gaps"}],"training":["Fine-tune of xlm-roberta-base, 8 epochs, about 3 minutes on one A10G.","4,737 sentences over 1,235 seed terms. Generated sentences were each seeded with a term and annotated blind by a second model family; disagreements went to human review, along with a 10% random audit of agreements.","458 dictated sentences: the text of sentences I read aloud for the Hebrish speech fine-tunes. Spans nobody has ruled on, and Israeli person names, carry no loss and are not scored."],"limitations":["English text with romanized Hebrew only. Hebrew script is not the target.","The training data is largely synthetic; the dictated slice is the closest thing to real input.","Under the default, conservative posture it does not flag place names (Dizengoff, Ramat Gan) or words English already says. That line is a design choice, not a fact.","Take each word's label from its first sub-token. With the stock pipeline use aggregation_strategy=\"first\"; \"simple\" cuts words short."],"checkpoints":[{"label":"v1","state":"done","date":"2026-10-15","note":"First release as danielrosehill/hebrew-latin-token-classifier, DOI 10.57967/hf/10997. Demo Space and Replicate API."},{"label":"Moved to the Hebrish organisation","state":"done","date":"2026-10-20","note":"Same weights, copied to Hebrish/Hebrew-Latin-Token-Classifier (DOI 10.57967/hf/11108). New versions are published there."},{"label":"Audio sibling","state":"in-progress","note":"Word-level English/Hebrew language ID in code-switched speech (Hebrish-ASR/Hebrew-English-Word-LID-Audio)."},{"label":"Third label","state":"planned","note":"A possible later version: a label for \"Hebrew that English already says\", so the user chooses where the line falls."}],"related":[{"label":"Code-switching TTS pipeline","url":"https://github.com/Hebrish-ASR/English-Hebrew-Code-Switching-TTS-Pipeline","note":"Detect, transliterate, niqqud, tag, speak: code, real outputs and a white paper"},{"label":"Demo Space","url":"https://huggingface.co/spaces/danielrosehill/hebrew-latin-token-classifier","note":"Try it in the browser"},{"label":"Training data","url":"https://huggingface.co/datasets/Hebrish/hebrew-latin-code-switching","note":"4,737 sentences over 1,235 seed terms"},{"label":"Original repository (DOI)","url":"https://huggingface.co/danielrosehill/hebrew-latin-token-classifier","note":"Same weights, revision 01f092f1"},{"label":"Audio sibling (work in progress)","url":"https://github.com/Hebrish-ASR/Hebrew-English-Word-LID-Audio"},{"label":"Replicate API","url":"https://replicate.com/danielrosehill/hebrew-latin-token-classifier","note":"CPU; inputs text and threshold"},{"label":"White paper (PDF)","url":"https://hebrish.app/whitepaper/code-switching-tts-pipeline.pdf"}],"tags":["Hebrish","token classification","code-switching","Hebrew","romanization","text-to-speech","XLM-RoBERTa"],"page":"https://ai.danielrosehill.com/models/hebrew-latin-token-classifier"},{"slug":"gemma-3-270m-asr-paragraph-breaks","title":"Gemma 3 270M ASR Paragraph Breaks: paragraphs for dictated text","shortTitle":"Gemma 3 270M paragraph breaks","eyebrow":"Model · ASR post-processing","licenseUrl":"https://ai.google.dev/gemma/terms","tagline":"Adds paragraph breaks to flat speech-to-text output and changes nothing else. Small enough to run on a phone after the speech model.","status":"r2 · first release","released":"2026-10-20","license":"Gemma Terms of Use","baseModel":{"label":"google/gemma-3-270m-it","url":"https://huggingface.co/google/gemma-3-270m-it"},"hfUrl":"https://huggingface.co/danielrosehill/Gemma-3-270M-ASR-Paragraph-Breaks","trackerUrl":"https://github.com/danielrosehill/ASR-Paragraph-Breaks","runsOn":"Transformers · ONNX Runtime","datasets":[{"label":"danielrosehill/ASR-Paragraph-Breaks","url":"https://huggingface.co/datasets/danielrosehill/ASR-Paragraph-Breaks"}],"summary":["Speech recognition, such as a Parakeet fine-tune, returns punctuated, cased text as one block. This model returns the same text with paragraphs separated by blank lines. It is a full fine-tune of Gemma 3 270M, the smallest Gemma, so it can run on-device after the ASR pass.","It must not change a word. Because it is generative it retypes the whole text, so every call is checked: if anything other than whitespace changed, the input is used instead. On the test set that fallback fired on 2 of 26 memos.","The paragraphing style is mine: short paragraphs, with the opening line and each sign-off line on their own."],"downloads":[{"label":"safetensors (Transformers)","size":"536 MB","use":"Python on CPU or GPU, at full precision. Use the prompt and fallback check from the model card.","runWith":{"label":"Hugging Face Transformers","url":"https://github.com/huggingface/transformers"},"url":"https://huggingface.co/danielrosehill/Gemma-3-270M-ASR-Paragraph-Breaks/resolve/main/model.safetensors","recommended":true},{"label":"ONNX fp32","size":"1.66 GB","use":"ONNX Runtime, on the same runtime as an ONNX speech model. Matched the PyTorch output on 6 of 6 test memos.","runWith":{"label":"ONNX Runtime / Optimum","url":"https://github.com/huggingface/optimum-onnx"},"url":"https://huggingface.co/danielrosehill/Gemma-3-270M-ASR-Paragraph-Breaks-ONNX/resolve/main/onnx/model.onnx"},{"label":"ONNX int8","size":"417 MB","use":"Experimental: about twice as fast on CPU, but matched PyTorch on only 2 of 6 memos and changed the text in one.","runWith":{"label":"ONNX Runtime / Optimum","url":"https://github.com/huggingface/optimum-onnx"},"url":"https://huggingface.co/danielrosehill/Gemma-3-270M-ASR-Paragraph-Breaks-ONNX/resolve/main/onnx/model_quantized.onnx"}],"results":{"asOf":"2026-10-20","run":"asr-paragraph-r2-gemma270m","version":"r2","baseLabel":"Untrained Gemma 3 270M","tunedLabel":"Fine-tune r2","metricNote":"Scored on 26 held-out synthetic voice memos whose paragraph breaks were reviewed by hand, as break-position precision and recall. Higher is better. A fallback is an output that changed a non-whitespace character, so the input is returned unchanged.","rows":[{"test":"Break precision","base":"0.00","tuned":"0.84","better":true},{"test":"Break recall","base":"0.00","tuned":"0.72","better":true},{"test":"F1","base":"0.00","tuned":"0.78","better":true},{"test":"Fallbacks (fewer is better)","base":"26 of 26","tuned":"2 of 26","better":true}],"notes":["A longer run (r3: 804 pairs, 6 epochs) scored the same F1 with 1 fallback and was not published.","The rule-based labeller that made most of the training data starts from an already-paragraphed draft, so it is a ceiling in the chart, not a like-for-like baseline."]},"figures":[{"src":"/images/ai/paragraph-breaks/f1_by_method.png","alt":"Paragraph-break F1 by method: untrained Gemma, the fine-tunes and the rule-based reference"},{"src":"/images/ai/paragraph-breaks/paragraph_lengths.png","alt":"Paragraph length distribution in the model output and the reference"}],"training":["Full fine-tune of google/gemma-3-270m-it with TRL SFT on prompt/completion pairs (completion-only loss): 505 training pairs, 3 epochs, learning rate 5e-05, bf16, one NVIDIA L4 on Modal.","Synthetic voice-memo transcripts written by Claude Haiku 5.5. A minority were paragraphed by hand; the rest were labelled by rules derived from those hand edits."],"limitations":["English only, and trained on synthetic memos rather than real ASR output.","Memos longer than about 700 words were not in training: split longer transcripts first.","It retypes the whole text, so keep the fallback check. Generation on a CPU is slow for long memos."],"checkpoints":[{"label":"r2","state":"done","date":"2026-10-20","note":"First release: 505 pairs, 3 epochs. F1 0.78."},{"label":"ONNX export","state":"done","date":"2026-10-20","note":"fp32 and int8 for ONNX Runtime, checked against the PyTorch model."}],"related":[{"label":"ONNX export","url":"https://huggingface.co/danielrosehill/Gemma-3-270M-ASR-Paragraph-Breaks-ONNX","note":"fp32 and int8 for ONNX Runtime"},{"label":"Training data","url":"https://huggingface.co/datasets/danielrosehill/ASR-Paragraph-Breaks","note":"Hand-annotated paragraph breaks for flat ASR text"},{"label":"Code","url":"https://github.com/danielrosehill/ASR-Paragraph-Breaks","note":"Data generation, training and evaluation"}],"tags":["Gemma","ASR post-processing","paragraph segmentation","dictation","voice notes","on-device"],"page":"https://ai.danielrosehill.com/models/gemma-3-270m-asr-paragraph-breaks"},{"slug":"gemma-3-270m-email-key-line-bolding","title":"Gemma 3 270M Email Key-Line Bolding: bold the key lines of a dictated email","shortTitle":"Gemma 3 270M email key-line bolding","eyebrow":"Model · dictation post-processing","licenseUrl":"https://ai.google.dev/gemma/terms","tagline":"Puts the key parts of a dictated email in Markdown bold and changes nothing else. Runs after the paragraph-breaks model.","status":"r2 · preliminary","released":"2026-10-21","license":"Gemma Terms of Use","baseModel":{"label":"google/gemma-3-270m-it","url":"https://huggingface.co/google/gemma-3-270m-it"},"hfUrl":"https://huggingface.co/danielrosehill/Gemma-3-270M-Email-Key-Line-Bolding","trackerUrl":"https://github.com/danielrosehill/Gemma-3-270M-Email-Key-Line-Bolding","runsOn":"Transformers","demoUrl":"https://huggingface.co/spaces/danielrosehill/Email-Key-Line-Bolding-Demo","datasets":[{"label":"danielrosehill/Email-Key-Line-Bolding","url":"https://huggingface.co/datasets/danielrosehill/Email-Key-Line-Bolding"}],"summary":["The input is an email that has already been paragraphed. The output is the same text with ** around the requests, the purpose, the main points and the key facts, so a reader who skims only the bold knows what is being asked.","It must not change anything else. Every call is checked: if anything other than ** was added, the input is returned unbolded.","Uses. In a dictation pipeline, it bolds the key lines of a dictated email after paragraphing. For outbound mail, an AI agent drafting an email can bold the details so the recipient sees them at a glance; it pairs with the paragraph-breaks model. For inbound mail, it can annotate received emails, and a mail client could show the spans as highlights instead of bold.","This is a first, preliminary version. The bolding style is mine, and most of the training labels were extrapolated from my annotations by an LLM."],"downloads":[{"label":"safetensors (Transformers)","size":"536 MB","use":"Python on CPU or GPU, at full precision. Use the system prompt and validation check from the model card.","runWith":{"label":"Hugging Face Transformers","url":"https://github.com/huggingface/transformers"},"url":"https://huggingface.co/danielrosehill/Gemma-3-270M-Email-Key-Line-Bolding/resolve/main/model.safetensors","recommended":true}],"results":{"asOf":"2026-10-21","run":"email-bold-r2-gemma270m-e10","version":"r2","baseLabel":"Untrained Gemma 3 270M","tunedLabel":"Fine-tune r2","metricNote":"Scored on 6 held-out emails bolded by hand, per word: a word counts as bold if it falls inside a span. Higher is better. A fallback is an output that changed anything other than adding **, so the input is returned unbolded.","rows":[{"test":"Word precision","base":"-","tuned":"0.73","better":true},{"test":"Word recall","base":"-","tuned":"0.54","better":true},{"test":"Word F1","base":"-","tuned":"0.62","better":true},{"test":"Fallbacks (fewer is better)","base":"6 of 6","tuned":"0 of 6","better":true}],"notes":["The untrained model fell back on every email, so it bolded nothing.","Reference: Claude Sonnet 5.5, prompted with the written rules, scores word F1 0.69 (4-fold cross-validation over the 19 hand-annotated emails).","The model is precise but under-bolds. It tends to bold the ask and the purpose line and miss some of the listed points.","Six test emails is a small test set. Treat the numbers as rough."]},"figures":[],"training":["Full fine-tune of google/gemma-3-270m-it (270M parameters): run email-bold-r2-gemma270m-e10, 10 epochs, learning rate 5e-05, one NVIDIA L4 on Modal, 504 seconds.","336 synthetic dictated emails written by Claude Haiku 5.5, to 12 kinds of recipient. 13 were bolded by hand; the other 323 were labelled by Claude Sonnet 5.5 from rules derived from 19 hand annotations (silver labels). The median email has about 25% of its words in bold.","Shorter runs (3 and 6 epochs, and 6 epochs at learning rate 1e-04) scored word F1 0.39 to 0.56. F1 was still rising at 10 epochs."],"limitations":["Bolding is subjective, and the labels reflect one person's preference.","Mostly silver labels, so the model learns the LLM labeller's habits as well as mine.","English only, and trained on synthetic dictated emails. Not tested on real transcripts or on text other than email."],"checkpoints":[{"label":"r2","state":"done","date":"2026-10-21","note":"Preliminary release: 336 emails, 10 epochs. Word F1 0.62."},{"label":"More hand annotations","state":"in-progress","note":"Annotation continues, towards about 200 hand-bolded emails, to replace the preliminary model."}],"related":[{"label":"Demo Space","url":"https://huggingface.co/spaces/danielrosehill/Email-Key-Line-Bolding-Demo","note":"Try it in the browser"},{"label":"Training data","url":"https://huggingface.co/datasets/danielrosehill/Email-Key-Line-Bolding","note":"Hand-annotated and silver-labelled dictated emails"},{"label":"Code","url":"https://github.com/danielrosehill/Gemma-3-270M-Email-Key-Line-Bolding","note":"Data, labelling rules, training and evaluation"},{"label":"Paragraph-breaks model","url":"https://huggingface.co/danielrosehill/Gemma-3-270M-ASR-Paragraph-Breaks","note":"The step before this one in the pipeline"}],"tags":["Gemma","dictation","email","text emphasis","on-device"],"page":"https://ai.danielrosehill.com/models/gemma-3-270m-email-key-line-bolding"},{"slug":"proximate-pseudonymizer-0.6b","title":"Proximate Pseudonymizer 0.6B: swap identifying details for close, consistent substitutes","shortTitle":"Proximate Pseudonymizer 0.6B","eyebrow":"Model · text privacy","licenseUrl":"https://www.apache.org/licenses/LICENSE-2.0","tagline":"Rewrites a text with its names, places, organisations, dates and contacts swapped for close, plausible substitutes, the same one every time. Proof of concept: it still leaves 9.3% of identifying details unchanged.","status":"pp-r2 · proof of concept","released":"2026-10-21","license":"Apache-2.0","baseModel":{"label":"Qwen/Qwen3-0.6B","url":"https://huggingface.co/Qwen/Qwen3-0.6B"},"hfUrl":"https://huggingface.co/danielrosehill/Proximate-Pseudonymizer-0.6B","trackerUrl":"https://github.com/danielrosehill/Proximate-Pseudonymizer","runsOn":"Transformers, on a laptop CPU","demoUrl":"https://huggingface.co/spaces/danielrosehill/Proximate-Pseudonymizer-Demo","datasets":[],"datasetsNote":"The training data is not published. It is 4,280 synthetic texts, described under Training.","summary":["Text in, text out. \"I'm Daniel and I live in Jerusalem\" becomes \"I'm David and I live in Tel Aviv\". Each detail gets one substitute, used every time it recurs, and the rest of the text is left alone.","It is not a redactor. There are no [NAME] tags, so the text still reads naturally and can be shared with an AI tool, a forum or a dataset without pointing at the real people and places.","Substitutes are meant to be proximate: same gender, culture, country and kind of place or organisation. Dates move a little and keep their year. Countries and public figures are kept.","This is a proof of concept. On 295 held-out texts it leaves 9.3% of identifying details unchanged, about 1 in 11. Do not run it unsupervised. Check every output before you share it."],"downloads":[{"label":"safetensors (Transformers)","size":"1.19 GB","use":"Python on CPU or GPU, at full precision. Use the exact system prompt from the model card; it is the one the model was trained with.","runWith":{"label":"Hugging Face Transformers","url":"https://github.com/huggingface/transformers"},"url":"https://huggingface.co/danielrosehill/Proximate-Pseudonymizer-0.6B/resolve/main/model.safetensors","recommended":true}],"results":{"asOf":"2026-10-21","run":"pp-r2","version":"Proof of concept","baseLabel":"Untrained Qwen3-0.6B, one example","tunedLabel":"Fine-tune pp-r2","metricNote":"Scored on 295 held-out synthetic texts. Leak is the share of identifying details still present word for word in the output; lower is better. The other rows are shares; higher is better.","rows":[{"test":"Leak: identifying details left unchanged (lower is better)","base":"85.6%","tuned":"9.3%","better":true},{"test":"Consistency: texts where each changed detail got one substitute","base":"0.997","tuned":"0.990"},{"test":"Rest of the text preserved (non-identifying words kept)","base":"0.791","tuned":"0.998","better":true},{"test":"Texts with no details returned unchanged","base":"0.308","tuned":"0.962","better":true}],"notes":["A 9.3% leak is fine for a demonstration and not fine for privacy. Names, dates and ages leak most.","Suggested targets: 1% or less with a human reviewing every output; about 0.1% or less without one, and then only for low-stakes text. This checkpoint is far from both.","Whether a substitute is proximate is not scored automatically.","The untrained model scores higher on consistency because it changes little; its leak is 85.6%."]},"figures":[],"training":["Full fine-tune of Qwen/Qwen3-0.6B: run pp-r2, 3 epochs, learning rate 5e-05, cosine schedule, effective batch 16, bf16, one NVIDIA L4 on Modal, about 38 minutes of training.","4,280 synthetic texts with replacement lists, written by Claude Haiku 5.5. Code checked every list for consistency and for substitutes that were too close to the original; flagged swaps were re-picked or the text was dropped. The training target is computed by applying the list, so every target is consistent by construction.","Training length was the main lever. The first run (2 epochs, learning rate 2e-05) left 26.5% of details unchanged; the same model and data trained longer at a higher rate reached 9.3%.","Training is paused at this checkpoint. Next steps: longer training, a larger base such as Qwen3-4B-Instruct, and a swap-list output mode (already in the repo) where code applies the substitutes."],"limitations":["About 1 in 11 identifying details survives. It is not safe to run without review.","Substitutes can collide: in one test, two different surnames both became \"Moran\". Some come out too close to the original.","Trained on synthetic texts only. 80% are English, weighted towards Israel, Ireland and the US; other languages work less well.","The replacement policy is one person's choice: big cities are replaced, countries and public figures are kept."],"checkpoints":[{"label":"pp-r2","state":"done","date":"2026-10-21","note":"Proof of concept: leak 9.3%, consistency 0.99, rest of text 0.998 preserved."},{"label":"Leak at or below 1%","state":"planned","note":"The target for use with a human reviewing every output. Training is paused."}],"related":[{"label":"Demo Space","url":"https://huggingface.co/spaces/danielrosehill/Proximate-Pseudonymizer-Demo","note":"Try it in the browser"},{"label":"Code","url":"https://github.com/danielrosehill/Proximate-Pseudonymizer","note":"Data pipeline, too-close checks, training, evaluation and run log"}],"tags":["Qwen3","pseudonymization","anonymization","privacy","text rewriting","proof of concept"],"page":"https://ai.danielrosehill.com/models/proximate-pseudonymizer-0.6b"},{"slug":"sabich-qwen3-asr-1.7b","title":"Hebrish (Sabich-Qwen3-ASR-1.7B v1.0): Sabich on Qwen3-ASR 1.7B for desktop GPUs","shortTitle":"Sabich-Qwen3-ASR-1.7B (desktop GPU)","tagline":"The most accurate Sabich: the same code-switching and tech data, trained on Qwen3-ASR 1.7B, an ASR model with an LLM decoder, for computers with a GPU.","status":"v1.0 · Sabich-Qwen3-ASR-1.7B","released":"2026-10-21","license":"Apache-2.0 (base model); use under the Hebrish terms of use","licenseUrl":"https://www.apache.org/licenses/LICENSE-2.0","baseModel":{"label":"Qwen/Qwen3-ASR-1.7B","url":"https://huggingface.co/Qwen/Qwen3-ASR-1.7B"},"hfUrl":"https://huggingface.co/Hebrish/Sabich-Qwen3-ASR-1.7B","website":{"label":"hebrish.app","url":"https://hebrish.app/models/sabich-qwen3"},"gated":"Both repos (the Transformers weights and the GGUF builds) are gated: request access on Hugging Face. Requests are reviewed by hand.","runsOn":"Desktop GPU · Transformers · llama.cpp","sibling":{"slug":"parakeet-tdt-v2-english-hebrew-all-subsets","label":"Sabich","note":"This is Sabich's training data on a larger base, for computers with a GPU (about 5 GB of VRAM in bf16, or about 2 GB as a GGUF). For phones and the Hebrish apps, use Sabich itself, on Parakeet TDT 0.6B v2."},"datasets":[{"label":"Read-Aloud Voice Sentences (revision e918e381, blend all)","url":"https://huggingface.co/datasets/danielrosehill/Read-Aloud-Voice-Sentences"},{"label":"All training datasets (collection)","url":"https://huggingface.co/collections/Hebrish/parakeet-v2-fine-tune-training-data-6ad63c739d300685a62edb3e"},{"label":"LibriSpeech (replay)","url":"https://huggingface.co/datasets/openslr/librispeech_asr"}],"summary":["Sabich-Qwen3-ASR-1.7B is Sabich, the everything model of the Hebrish project, trained on a larger base. The data is the same: English with Hebrew words in it, written in one romanized spelling, plus tech vocabulary, brand names, acronyms, numbers and currency in digits, and everyday email and message dictation.","The main Sabich is Parakeet TDT 0.6B v2, sized for phones. This variant puts the same data on Qwen/Qwen3-ASR-1.7B, an ASR model with an LLM decoder, for desktop GPUs and the highest accuracy. On the held-out split it halves stock Qwen3-ASR's code-switching error, from 10.0% to 5.4%, and writes numbers in digits every time.","Output is English, cased and punctuated, with Hebrew words in Latin letters, never Hebrew script. It is not in the Hebrish apps yet: run it in your own tools, with Transformers on PyTorch (CUDA or ROCm) or as a GGUF in llama.cpp.","A variant on another base keeps its line's food name and adds the base: Sabich-<Base>-<Size>. Other Sabich variants run on Canary-Qwen 2.5B, Parakeet TDT-CTC 110M and FUTO asr4all."],"downloads":[{"label":"Q8_0 (llama.cpp)","size":"1.83 GB","use":"Desktop GPU or CPU, the default llama.cpp build. Needs the audio encoder (mmproj) file beside it. Fits a 4 GB card.","runWith":{"label":"llama.cpp (HIP, Vulkan, CUDA)","url":"https://github.com/ggml-org/llama.cpp"},"url":"https://huggingface.co/Hebrish/Sabich-Qwen3-ASR-1.7B-GGUF/resolve/main/llama.cpp/Sabich-Qwen3-ASR-1.7B-Q8_0.gguf","recommended":true},{"label":"Audio encoder (mmproj Q8_0)","size":"356 MB","use":"The audio encoder that llama.cpp loads with --mmproj, next to the Q8_0 model file.","runWith":{"label":"llama.cpp","url":"https://github.com/ggml-org/llama.cpp"},"url":"https://huggingface.co/Hebrish/Sabich-Qwen3-ASR-1.7B-GGUF/resolve/main/llama.cpp/mmproj-Sabich-Qwen3-ASR-1.7B-Q8_0.gguf"},{"label":"transcribe.cpp Q4_K_M","size":"1.32 GB","use":"Android phone, imported into Offline Voice Input. It runs, but slowly; for phones, Sabich is the better fit.","runWith":{"label":"Offline Voice Input","url":"https://play.google.com/store/apps/details?id=dev.notune.transcribe"},"url":"https://huggingface.co/Hebrish/Sabich-Qwen3-ASR-1.7B-GGUF/resolve/main/transcribe.cpp/Sabich-Qwen3-ASR-1.7B-Q4_K_M.gguf"}],"results":{"asOf":"2026-10-21","run":"q1-all","version":"v1.0 · Sabich-Qwen3-ASR-1.7B","baseLabel":"Stock Qwen3-ASR 1.7B","tunedLabel":"Sabich-Qwen3-ASR-1.7B v1.0","metricNote":"Lower WER is better; higher term recall is better.","rows":[{"test":"English–Hebrew code-switching (normalised WER)","base":"10.0%","tuned":"5.4%","better":true},{"test":"Hebrew terms never seen in training: recall","base":"32%","tuned":"50%","better":true},{"test":"Hebrew terms seen in training: recall","base":"40%","tuned":"73%","better":true},{"test":"Speaker's own English (normalised WER)","base":"4.6%","tuned":"2.5%","better":true},{"test":"LibriSpeech test-clean, 200 utterances (normalised WER)","base":"1.2%","tuned":"1.35%"},{"test":"Long form, about 3-minute joined clips (normalised WER)","base":"12.0%","tuned":"8.7%","better":true},{"test":"Numbers written in digits","base":"2/8","tuned":"8/8","better":true}],"notes":["Held-out test split of the speaker's recordings at dataset revision cf84f7e4, term-disjoint from training. Stock is Qwen/Qwen3-ASR-1.7B on the same clips.","WER is normalised: lowercase, punctuation removed. Term recall is a word-boundary match on the romanized Hebrew term.","These numbers are on a different test-set revision from Sabich's (f956dbb1), so the two have not been compared head to head yet.","LibriSpeech is within noise of stock. Long-form audio is its weakest area."]},"figures":[],"training":["v1.0: Sabich's data at dataset revision e918e381 (blend all): 592 min of my own recordings (383 min of them code-switching) plus 66 min of LibriSpeech-PC replay, 10% of the mix, 658 min in all. Run name q1-all.","Full fine-tune from stock Qwen3-ASR 1.7B with the official Qwen3-ASR fine-tuning script: 3 epochs, lr 2e-5 with linear decay, 5% warmup, effective batch 32, bf16, one H100 on Modal, 15 minutes. The epoch 3 checkpoint was kept.","Training targets carry the prefix \"language English\", so inference always passes language=\"English\".","GGUF builds for llama.cpp (bf16, Q8_0, Q4_K_M, with a separate audio encoder) and for transcribe.cpp (Q8_0, Q4_K_M), each checked against the bf16 model on three held-out clips. That is a format check, not an evaluation."],"limitations":["Trained on one voice so far: an English speaker with an Irish accent living in Jerusalem. Recordings from more speakers are the next step; contribute yours at hebrish.app/contact?topic=samples.","About half of the Hebrew terms it never heard in training are still missed or misspelled (50% recall).","Long-form audio is its weakest area: 8.7% WER on about 3-minute clips.","Needs a GPU for comfortable speed. On 8 cloud CPU cores the Q8_0 file ran at about 1× realtime; phone speed is not measured.","Not in the Hebrish apps. The repos are gated, with access by request."],"checkpoints":[{"label":"v1.0 (q1-all)","state":"done","date":"2026-10-21","note":"First release: Sabich's data on Qwen3-ASR 1.7B, 3 epochs on one H100, epoch 3 kept. 5.4% code-switching WER against 10.0% for stock."},{"label":"GGUF builds","state":"done","date":"2026-10-21","note":"llama.cpp (bf16, Q8_0, Q4_K_M plus the audio encoder) and transcribe.cpp (Q8_0, Q4_K_M), checked on three held-out clips."},{"label":"Head to head with Sabich","state":"planned","note":"Score it on the same test-set revision as Sabich; the two have not been compared yet."},{"label":"Phone speed","state":"planned","note":"Measure the transcribe.cpp build on a phone; only CPU speed in the cloud is known."}],"related":[{"label":"Sabich-Qwen3-ASR-1.7B on hebrish.app","url":"https://hebrish.app/models/sabich-qwen3","note":"Results, formats, how to run it and how to request access"},{"label":"Sabich (sister model)","url":"https://hebrish.app/models/sabich","note":"The same data on Parakeet TDT 0.6B v2, sized for phones and in the Hebrish apps"},{"label":"GGUF builds","url":"https://huggingface.co/Hebrish/Sabich-Qwen3-ASR-1.7B-GGUF","note":"llama.cpp and transcribe.cpp files, with the exact commands (gated)"},{"label":"Which model should I use?","url":"https://hebrish.app/models/choose","note":"Picks a Hebrish model by device, GPU and VRAM"},{"label":"Sabich-Canary-Qwen-2.5B","url":"https://hebrish.app/models/sabich-canary-qwen","note":"The other desktop-GPU Sabich: behind this one on code-switching, best on plain English"},{"label":"Sabich collection","url":"https://huggingface.co/collections/Hebrish/sabich-6ad7b241e1a420ffa0905767","note":"Every Sabich model and build on Hugging Face"},{"label":"Benchmark leaderboard","url":"https://huggingface.co/spaces/danielrosehill/English-Hebrew-Codeswitching-ASR-Leaderboard","note":"The public benchmark the Parakeet Hebrish models are scored on"},{"label":"Hebrish website","url":"https://hebrish.app","note":"The project site: every model, dataset and result"}],"tags":["Hebrish","Sabich","ASR","Qwen3-ASR","Transformers","llama.cpp","GGUF","code-switching","Hebrew","English","desktop GPU"],"page":"https://ai.danielrosehill.com/models/sabich-qwen3-asr-1.7b"},{"slug":"shakshuka-nemotron-streaming","title":"Hebrish (Shakshuka v1.0): Nemotron streaming fine-tune for code-switching plus tech dictation","shortTitle":"Shakshuka (Nemotron streaming)","eyebrow":"Model · streaming speech-to-text","tagline":"The streaming line of Hebrish: text appears while you speak, trained on the same everything-in-it recordings as Sabich.","status":"v1.0 · Shakshuka","released":"2026-10-21","license":"NVIDIA Open Model License Agreement; use under the Hebrish terms of use","licenseUrl":"https://huggingface.co/Hebrish/Shakshuka/blob/main/LICENSE","baseModel":{"label":"nvidia/nemotron-speech-streaming-en-0.6b","url":"https://huggingface.co/nvidia/nemotron-speech-streaming-en-0.6b"},"hfUrl":"https://huggingface.co/Hebrish/Shakshuka","website":{"label":"hebrish.app","url":"https://hebrish.app/models/shakshuka"},"gated":"Every Shakshuka repo (.nemo, GGUF, sherpa-onnx and the PoC record) is gated: request access on Hugging Face. Requests are reviewed by hand.","runsOn":"NeMo · sherpa-onnx · transcribe.cpp","sibling":{"slug":"parakeet-tdt-v2-english-hebrew-all-subsets","label":"Sabich","note":"Shakshuka streams: text appears while you speak. Sabich transcribes offline, when you stop, and on past results it is the more accurate of the two. Both are trained on the same data, so use Sabich unless you need live text."},"datasets":[{"label":"Read-Aloud Voice Sentences (revision e918e381, blend all)","url":"https://huggingface.co/datasets/danielrosehill/Read-Aloud-Voice-Sentences"},{"label":"All training datasets (collection)","url":"https://huggingface.co/collections/Hebrish/parakeet-v2-fine-tune-training-data-6ad63c739d300685a62edb3e"},{"label":"LibriSpeech (replay)","url":"https://huggingface.co/datasets/openslr/librispeech_asr"}],"summary":["Shakshuka is the streaming line of Hebrish: a fine-tune of NVIDIA's cache-aware streaming model, Nemotron Speech Streaming EN 0.6B, so text appears while you speak, in chunks of 80 ms to 1.12 s. Like the other Hebrish models it writes the Hebrew words in English speech in one romanized spelling, and it never writes Hebrew script.","Shakshuka v1.0 (2026-10-21) is the first full release. It is trained on Sabich's data: code-switched speech plus tech vocabulary, brand names, acronyms, numbers and currency in digits, Irish-English phrasing and everyday dictation. The two earlier checkpoints, PoC-1 and PoC-2, were trained on the code-switching mix only, beside Jachnun.","v1.0 was published straight after training and is not benchmarked yet. The numbers below are PoC-2's: 62.2% Hebrew term recall and 4.7% WER on the public benchmark, scored offline. Streaming at 560 ms chunks cost it nothing measurable against offline decoding.","It is not in the Hebrish apps yet. Run it in NVIDIA NeMo, as a sherpa-onnx streaming build (560 ms chunks) or as a GGUF in transcribe.cpp. Like Jachnun and Sabich, it takes an Israeli food name."],"downloads":[{"label":".nemo checkpoint","size":"2.47 GB","use":"Python, GPU inference, offline or cache-aware streaming, or further fine-tuning, at full precision.","runWith":{"label":"NVIDIA NeMo","url":"https://github.com/NVIDIA-NeMo/Speech"},"url":"https://huggingface.co/Hebrish/Shakshuka/resolve/main/Shakshuka.nemo","recommended":true},{"label":"transcribe.cpp Q8_0","size":"730 MB","use":"The recommended transcribe.cpp file. Whether an app loads this architecture depends on its transcribe.cpp version; transcribe-cli v0.1.3 does.","runWith":{"label":"transcribe.cpp","url":"https://github.com/handy-computer/transcribe.cpp"},"url":"https://huggingface.co/Hebrish/Shakshuka-GGUF/resolve/main/Shakshuka-Q8_0.gguf"},{"label":"transcribe.cpp Q4_K_M","size":"475 MB","use":"Smaller quant for devices short on memory.","runWith":{"label":"transcribe.cpp","url":"https://github.com/handy-computer/transcribe.cpp"},"url":"https://huggingface.co/Hebrish/Shakshuka-GGUF/resolve/main/Shakshuka-Q4_K_M.gguf"},{"label":"transcribe.cpp F16","size":"1.24 GB","use":"Desktop GPU, half precision.","runWith":{"label":"transcribe.cpp","url":"https://github.com/handy-computer/transcribe.cpp"},"url":"https://huggingface.co/Hebrish/Shakshuka-GGUF/resolve/main/Shakshuka-F16.gguf"}],"results":{"asOf":"2026-10-19","run":"r6-nemotron-en30","version":"Shakshuka PoC-2","pending":"Shakshuka v1.0 (run r8-nemotron-all, 2026-10-21) was published straight after training and has no test results yet. Its validation WER at the chosen checkpoint was 6.66% (raw, used only to pick the checkpoint; not a test score).","baseLabel":"Stock Parakeet v2 (reference)","tunedLabel":"Shakshuka PoC-2","metricNote":"Stock Nemotron was not scored on these test sets, so the reference column is stock Parakeet TDT 0.6B v2 on the same clips. Lower WER is better; higher term recall is better.","rows":[{"test":"Public benchmark, 84 clips: Hebrew term recall (offline)","base":"32.6%","tuned":"62.2%","better":true},{"test":"Public benchmark: unseen-term recall (offline)","base":"32.5%","tuned":"54.7%","better":true},{"test":"Public benchmark: WER (offline)","base":"9.5%","tuned":"4.7%","better":true},{"test":"Public benchmark: English control WER (offline)","base":"2.7%","tuned":"2.5%"},{"test":"Code-switching, held-out split, 387 clips (normalised WER, offline)","base":"11.3%","tuned":"5.8%","better":true},{"test":"Code-switching, held-out split (normalised WER, streaming at 560 ms)","base":"n/a","tuned":"5.65%","better":true},{"test":"Hebrew terms held out of training: share correct (offline)","base":"25.9%","tuned":"44.4%","better":true},{"test":"Speaker's own English, 73 clips (normalised WER)","base":"3.1%","tuned":"3.3%"},{"test":"LibriSpeech test-clean, 200 utterances (normalised WER)","base":"1.2%","tuned":"2.0%"},{"test":"Long form, 15 files (normalised WER)","base":"9.6%","tuned":"5.8%","better":true}],"notes":["Public benchmark: test set v1, clean subset (84 clips, 172 term occurrences, 139 distinct terms), round 3 on the leaderboard Space, scored 2026-10-19 with offline (full-context) decoding. Jachnun v1.0 on the same clips: 73.3% term recall, 70.1% unseen, 3.3% WER, 2.2% English control.","Held-out split: dataset revision 550bfcaa0f84, term-disjoint, never trained on. Jachnun v1.0 on the same split: 5.1% code-switching WER, 47.8% held-out terms, 1.7% own English, 1.2% LibriSpeech, 5.0% long form.","Streaming used cache-aware inference at 560 ms chunks (att_context_size [70, 6], tag r6-nemotron-en30-s560): 5.65% against 5.81% offline, no measurable loss. The 80 and 160 ms settings are not measured.","No forgetting verdict: stock Nemotron was not re-scored on this revision, so the English rows have no baseline of their own. PoC-2 drops all 8 number tokens in the held-out split.","WER is normalised: lowercase, punctuation removed."]},"figures":[],"training":["v1.0 · Shakshuka: Sabich's data at dataset revision e918e381 (blend all): 592.4 min of my own recordings (382.9 min of them code-switching) plus 66.0 min of LibriSpeech-PC replay, 658.4 min and 5,634 clips. Run name r8-nemotron-all.","Full fine-tune from stock Nemotron Speech Streaming EN 0.6B (cache-aware FastConformer + RNNT) in NeMo 2.7.3: AdamW, lr 3e-5, cosine schedule with 10% warmup, 20 epochs, batch 16, bf16, one L40S on Modal, 51 minutes. Best checkpoint at epoch 13.","Trained with full attention context (offline), not streaming-aware, like the PoC checkpoints. On PoC-2, streaming at 560 ms scored the same as offline decoding.","Before v1.0 came two proof-of-concept checkpoints on the code-switching mix only, trained beside the Parakeet runs with the same recipe: PoC-1 (r5-nemotron-en30, 2026-10-18, beside PoC-5) and PoC-2 (r6-nemotron-en30, 2026-10-19, 339.2 min of code-switching beside Jachnun v1.0)."],"limitations":["Trained on one voice so far: an English speaker with an Irish accent living in Jerusalem. Recordings from more speakers are the next step; contribute yours at hebrish.app/contact?topic=samples.","v1.0 is not benchmarked yet: no test-set numbers, offline or streaming.","On past results it trails the offline Parakeet models: PoC-2 was under one WER point behind Jachnun v1.0 on code-switching, and 11 points behind on term recall on the public benchmark.","Trained with full attention context; accuracy at the smallest chunks (80 and 160 ms) may be lower than offline.","NVIDIA Open Model License Agreement, not CC-BY-4.0: read it before shipping. Not made or endorsed by NVIDIA. The repos are gated, with access by request."],"checkpoints":[{"label":"PoC-1 (r5-nemotron-en30)","state":"done","date":"2026-10-18","note":"First Nemotron streaming fine-tune, trained beside PoC-5 on the code-switching mix. Published before any benchmark; now in the PoC record's history."},{"label":"PoC-2 (r6-nemotron-en30)","state":"done","date":"2026-10-19","note":"Trained beside Jachnun v1.0 on the same data. Scored 2026-10-19: 62.2% term recall on the public benchmark; 5.8% code-switching WER offline and 5.65% streaming at 560 ms on the held-out split."},{"label":"v1.0 · Shakshuka (r8-nemotron-all)","state":"done","date":"2026-10-21","note":"First full release, on Sabich's data. .nemo, sherpa-onnx int8 streaming build (560 ms chunks) and GGUF (Q8_0, Q4_K_M, F16), each export checked on two held-out clips."},{"label":"v1.0 benchmark","state":"planned","note":"Score v1.0 offline and streaming, with stock Nemotron as the baseline."},{"label":"Small-chunk streaming","state":"planned","note":"Measure accuracy at 80 and 160 ms chunks, not yet measured on any checkpoint."}],"related":[{"label":"Shakshuka on hebrish.app","url":"https://hebrish.app/models/shakshuka","note":"Formats, versions, how to stream it and how to request access"},{"label":"Sabich (sister model)","url":"https://hebrish.app/models/sabich","note":"The same training data on offline Parakeet TDT 0.6B v2, in the Hebrish apps"},{"label":"sherpa-onnx streaming build","url":"https://huggingface.co/Hebrish/Shakshuka-sherpa-onnx-int8","note":"int8 streaming transducer at 560 ms chunks, for OnlineRecognizer on Android, Linux and Windows (gated)"},{"label":"GGUF builds","url":"https://huggingface.co/Hebrish/Shakshuka-GGUF","note":"Q8_0, Q4_K_M and F16 for transcribe.cpp (gated)"},{"label":"PoC record","url":"https://huggingface.co/Hebrish/Nemotron-Speech-Streaming-EN-0.6B-English-Hebrew-Codeswitching","note":"PoC-1 and PoC-2, their evaluations and figures; DOI 10.57967/hf/11107"},{"label":"Benchmark leaderboard","url":"https://huggingface.co/spaces/danielrosehill/English-Hebrew-Codeswitching-ASR-Leaderboard","note":"PoC-2 against the Hebrish Parakeet models, Whisper and stock Parakeet on the same frozen test set"},{"label":"Public benchmark","url":"https://huggingface.co/datasets/Hebrish/English-Hebrew-Codeswitching-ASR-Benchmark","note":"The 84-clip test set PoC-2 was scored on (round 3)"},{"label":"Jachnun and Sabich in depth (blog)","url":"https://blog.danielrosehill.com/posts/jachnun-sabich-first-stable-hebrish-models-krjelm","note":"The first stable releases, with Shakshuka PoC-2 in passing"},{"label":"Hebrish website","url":"https://hebrish.app","note":"The project site: every model, dataset and result"}],"tags":["Hebrish","Shakshuka","ASR","Nemotron","NeMo","streaming","code-switching","Hebrew","English","sherpa-onnx"],"page":"https://ai.danielrosehill.com/models/shakshuka-nemotron-streaming"}],"dataset_cards":[{"slug":"english-hebrew-codeswitching-sentences","title":"English–Hebrew Code-Switching Sentences","shortTitle":"English–Hebrew Code-Switching Sentences","tagline":"English sentences with everyday Hebrew words and Israeli names written in Latin letters, read aloud on a phone by one English speaker living in Israel.","status":"Live and growing","license":"CC-BY-4.0","hfId":"danielrosehill/English-Hebrew-Codeswitching-Sentences","hfUrl":"https://huggingface.co/datasets/danielrosehill/English-Hebrew-Codeswitching-Sentences","gated":"Gated on Hugging Face (manual approval): request access to download the audio and metadata.","liveSnapshot":false,"parent":{"label":"Read-Aloud Voice Sentences","url":"https://huggingface.co/datasets/danielrosehill/Read-Aloud-Voice-Sentences","note":"The full dataset; this one is its code-switching subset"},"browserUrl":"https://huggingface.co/spaces/danielrosehill/English-Hebrew-Codeswitching-Browser","trackerUrl":"https://github.com/danielrosehill/English-Hebrew-Codeswitching-Parakeet-Finetune","usedBy":[{"slug":"parakeet-english-hebrew-codeswitching","label":"Hebrish: Parakeet TDT 0.6B v2 for English–Hebrew code-switching","url":"https://huggingface.co/danielrosehill/Parakeet-TDT-0.6B-v2-English-Hebrew-Codeswitching"}],"summary":["Speech recognition trained on English mangles the Hebrew words that English speakers in Israel use every day: Bituach Leumi, Misrad Hapnim, tzaharon, makolet. This dataset is the training and test data for fixing that.","Each row is one take: a short English sentence containing one or more Hebrew words, read aloud and reviewed on the phone before it was published. Hebrew words follow one canonical romanized spelling, and Hebrew proper nouns are capitalised (Yerushalayim, Bituach Leumi).","It is a subset of Read-Aloud Voice Sentences: every take whose categories include code_switch, copied unchanged with the same ids and split. Fixes go into the main dataset and are re-synced."],"speaker":["One speaker: Daniel Rosehill, born in Ireland, English with an Irish accent, living in Jerusalem.","Recorded on a OnePlus Nord 3 5G internal microphone, with browser echo cancellation, noise suppression and auto gain turned off. A smaller set of older takes came from a ThinkPad desktop mic.","16 kHz mono 16-bit PCM WAV."],"layout":[{"field":"text","description":"Ground truth, punctuated and cased like Parakeet v2 output"},{"field":"target_terms","description":"The Hebrew words and names the sentence was written to exercise"},{"field":"categories","description":"Always includes code_switch; may add names, message and others"},{"field":"prompt_source","description":"Where the sentence text came from, for attribution"},{"field":"duration_s","description":"Clip length in seconds"},{"field":"device / environment / conditions","description":"Recording phone or mic and the room it was recorded in"},{"field":"recorded_at","description":"UTC timestamp of the take"}],"snapshot":{"asOf":"2026-10-15","splits":[{"split":"train","clips":1248,"minutes":120.1,"terms":734,"phoneClips":754},{"split":"test","clips":121,"minutes":13.9,"terms":146,"phoneClips":85}]},"priorities":["The 390 Hebrew terms that stock Parakeet v2 never gets right, such as Misrad Hapnim, Bituach Leumi, rofe mishpacha and Pikud Haoref.","Terms with only one or two takes so far. Most of the 600-plus terms in train appear only once.","301 Israeli-life terms found missing by a gap review, with 868 new sentences written for them.","Plain monolingual English from the same phone, for the English share of the training mix (in the parent dataset)."],"checkpoints":[{"label":"Capture app and dataset created","state":"done","date":"2026-10-14","note":"A phone web app in the style of Common Voice: read, record, review, publish in batches."},{"label":"Code-switching subset split out","state":"done","date":"2026-10-14","note":"This dataset, plus a browser Space to listen by term."},{"label":"v1 training snapshot","state":"done","date":"2026-10-15","note":"Parent dataset pinned at revision b31aeabb5edc: 77 min of code-switching trained the v1 pilot model."},{"label":"Transcript normalisation pass","state":"done","date":"2026-10-15","note":"10 rows fixed, including a spelling-rewrite bug that turned \"more\" into \"moreh\"."},{"label":"v2 training snapshot","state":"done","date":"2026-10-15","note":"Parent dataset pinned at revision 3bdaa2a50d52: 120 min of code-switching (1,248 clips) trained the v2 model; test split 121 clips, 14 min."},{"label":"PoC-3 training snapshot","state":"done","date":"2026-10-18","note":"Parent dataset pinned at revision b47310621f79: 196 min of code-switching (1,921 clips) trained PoC-3 (published as v3); test split grew to 219 clips."},{"label":"PoC-5 training snapshot","state":"done","date":"2026-10-18","note":"Parent dataset pinned at revision fcf5aec16449: 294 min of code-switching (2,829 clips) trained PoC-5 (PoC-4 used 269 min); test split grew to 311 clips, 265 held-out terms."},{"label":"v1.0 training snapshot","state":"done","date":"2026-10-19","note":"Parent dataset pinned at revision 550bfcaa0f84: 339.2 min of code-switching trained v1.0 · Jachnun; the held-out split is 387 clips."},{"label":"Data collection","state":"in-progress","note":"Recording continues, weighted toward the priorities below."},{"label":"v6 training snapshot","state":"planned","note":"A new pinned revision at 360 min of code-switching train audio, for v6."}],"related":[{"label":"Read-Aloud Voice Sentences","url":"https://huggingface.co/datasets/danielrosehill/Read-Aloud-Voice-Sentences","note":"Parent dataset: all takes, including monolingual English"},{"label":"Code-switching browser","url":"https://huggingface.co/spaces/danielrosehill/English-Hebrew-Codeswitching-Browser","note":"Listen to the takes by Hebrew term"},{"label":"English–Hebrew Mixed Sentences","url":"https://huggingface.co/datasets/danielrosehill/English-Hebrew-Mixed-Sentences","note":"Earlier desktop-mic round; its sentences were re-read on the phone"},{"label":"Hebrew-Latin token classifier","url":"https://github.com/danielrosehill/Hebrew-Latin-Token-Classifier","note":"Lexicon behind the canonical spellings"},{"label":"Project tracker (GitHub)","url":"https://github.com/danielrosehill/English-Hebrew-Codeswitching-Parakeet-Finetune","note":"Status, data counts and log"}],"tags":["ASR","speech","code-switching","Hebrew","English","single-speaker","fine-tuning"],"page":"https://ai.danielrosehill.com/datasets/english-hebrew-codeswitching-sentences"},{"slug":"read-aloud-voice-sentences","title":"Read-Aloud Voice Sentences","shortTitle":"Read-Aloud Voice Sentences","eyebrow":"Dataset · speech audio · Hebrish","metaLine":"English with Hebrew, tech and dictated messages","tagline":"Every take recorded for the Hebrish speech models: read-aloud English sentences with Hebrew words, tech vocabulary, names, emails and messages, recorded on a phone and published in batches straight from the capture app.","status":"Live and growing","license":"CC-BY-4.0","hfId":"danielrosehill/Read-Aloud-Voice-Sentences","hfUrl":"https://huggingface.co/datasets/danielrosehill/Read-Aloud-Voice-Sentences","trackerUrl":"https://github.com/danielrosehill/English-Hebrew-Codeswitching-Parakeet-Finetune","gated":"Gated on Hugging Face (manual approval): request access to download the audio and metadata.","usedBy":[{"slug":"parakeet-english-hebrew-codeswitching","label":"Hebrish Jachnun v1.0 (code-switching rows)","url":"https://huggingface.co/Hebrish/Jachnun"},{"slug":"parakeet-tdt-v2-english-hebrew-all-subsets","label":"Hebrish Sabich v1.0 (every task recording)","url":"https://huggingface.co/Hebrish/Sabich"},{"label":"PoC-1 to PoC-5 (the proof-of-concept checkpoints)","url":"https://huggingface.co/danielrosehill/Parakeet-TDT-0.6B-v2-English-Hebrew-Codeswitching"},{"label":"Shakshuka PoC (Nemotron streaming fine-tune)","url":"https://huggingface.co/Hebrish/Nemotron-Speech-Streaming-EN-0.6B-English-Hebrew-Codeswitching"}],"summary":["This is the full, undifferentiated dataset behind the Hebrish speech models. Every take is a short sentence read aloud and reviewed on the phone before it was published; the dataset grows as new batches are pushed from the capture app.","A row can carry several categories: code_switch (English with Hebrew words written in Latin letters, one canonical spelling each), names, email and message (the main use case: dictating emails and messages on Android), tech, homophone, general, plus later tags such as religious, yiddish, brand-names, irish-english, currency and disfluency. Filter on categories to use one kind of data on its own.","Two derived datasets carry its phone takes for one kind of data, unchanged (same ids, columns and split) and re-synced on every push: English–Hebrew Code-Switching Sentences (rows tagged code_switch) and Tech-Sentences (rows tagged tech). Rows with Jewish religious vocabulary are also mirrored to Jewish-Religious-Terminology-Sentences.","The test split is term-disjoint for code_switch and tech: a held-out Hebrew or technical term never appears in train, so the evaluation measures generalisation to unseen vocabulary. The public benchmark test set was cut from this split at revision 3bdaa2a50d52.","Hebrew words switched into English do not have one pronunciation: Modern Israeli Hebrew says Shabbat, many English-speaking Jews say Shabbos. Variants are recorded on purpose, and since 2026-10-19 each take is tagged with hebrew_pronunciation; since 2026-10-20 variant_groups links the Israeli and Ashkenazi spellings of one word."],"speaker":["One speaker: Daniel Rosehill, the dataset's author, born in Ireland, English with an Irish accent, living in Jerusalem. Every recording, including the imported rows, is his voice.","Phone takes were recorded on a OnePlus Nord 3 5G internal microphone, with browser echo cancellation, noise suppression and auto gain turned off.","498 older takes (mode \"import\", the ehms_* rows) were imported on 2026-10-14 from English-Hebrew-Mixed-Sentences, recorded 2025-11-17 on another device; their transcripts were updated to this dataset's spelling, with the original in original_text. Filter on mode != \"import\" for phone-only data.","16 kHz mono 16-bit PCM WAV."],"layout":[{"field":"text","description":"Ground truth, punctuated and cased like Parakeet v2 output"},{"field":"categories","description":"What the sentence exercises: code_switch, tech, names, email, message, homophone, general, religious, yiddish and others; a row can have several"},{"field":"target_terms","description":"The terms (Hebrew or technical) the sentence was written to exercise, and which the evaluation scores"},{"field":"prompt_source","description":"Where the sentence text came from: tatoeba:<id>, generated:<model>, enron:<id> or one of the author's earlier datasets (see licence)"},{"field":"edited","description":"true when the speaker corrected the prompt before or after reading it"},{"field":"duration_s / environment / device / recorded_at","description":"Recording metadata"},{"field":"conditions","description":"Recording conditions noted on the phone (background-noise, recorded-outdoors, in-car, ...); missing on the earliest rows"},{"field":"mode","description":"How the take was made; \"import\" marks the 2025 desktop-mic rows"},{"field":"hebrew_pronunciation","description":"israeli, ashkenazi, mixed, or null when not tagged (since 2026-10-19; null does not mean Israeli)"},{"field":"variant_groups / variant_spelling","description":"Ids of Israeli/Ashkenazi spelling pairs that occur in text, and which tradition the text spells (since 2026-10-20)"}],"liveSnapshot":false,"snapshotTermsLabel":"target terms in train (Hebrew and tech)","snapshot":{"asOf":"2026-10-21","splits":[{"split":"train","clips":5807,"minutes":642,"terms":2688,"phoneClips":5230},{"split":"test","clips":536,"minutes":62.7,"terms":471,"phoneClips":498}]},"priorities":["Messier acoustic conditions: takes from the street, cafés and the shuk, and with a baby crying nearby, even for sentences with no new vocabulary.","More microphones: every phone take so far comes from one OnePlus Nord 3 5G; desktop-microphone takes are to follow.","Pronunciation variants: Hebrew words with two common pronunciations (Diaspora and Israeli) recorded in both, with one spelling.","Other speakers, later: female voices, other English accents common in Israel, and Sephardi and Ashkenazi Hebrew pronunciations.","Better condition tags: tags set on the phone are often missing; correcting them afterwards, possibly with an audio classifier, is planned."],"checkpoints":[{"label":"Dataset created","state":"done","date":"2026-10-14","note":"First takes pushed from the phone capture app; renamed Read-Aloud Voice Sentences the same day, with explicit configs for the dataset viewer."},{"label":"Desktop-mic round imported","state":"done","date":"2026-10-14","note":"498 recordings from English-Hebrew-Mixed-Sentences (2025-11-17) imported, transcripts updated to this dataset's spelling."},{"label":"Transcript normalisation and cleanup","state":"done","date":"2026-10-15","note":"Canonical Hebrew spellings and proper-noun casing; 38 corrupted takes removed."},{"label":"Benchmark test set cut","state":"done","date":"2026-10-15","note":"The public English-Hebrew-Codeswitching-ASR-Benchmark (test set v1, 145 clips) was built from the test split at revision 3bdaa2a50d52, the same revision that trained PoC-2."},{"label":"Wanted-data list","state":"done","date":"2026-10-18","note":"Where the dataset should grow next, from a voice note by the speaker: conditions, microphones, pronunciation variants, other speakers."},{"label":"Pronunciation and Yiddish tagging","state":"done","date":"2026-10-19","note":"hebrew_pronunciation column added; rows with Yiddish words tagged yiddish."},{"label":"v1.0 training snapshot","state":"done","date":"2026-10-19","note":"Pinned at revision 550bfcaa0f84 for Jachnun v1.0 and the Nemotron streaming fine-tune; its test split (387 code-switched clips) scored them."},{"label":"Religious tag and pronunciation pairs","state":"done","date":"2026-10-20","note":"religious category (mirrored to Jewish-Religious-Terminology-Sentences), variant_groups and variant_spelling; a second normalisation pass wrote shekel amounts as NIS."},{"label":"Data collection","state":"in-progress","note":"Recording continues, weighted toward the wanted-data list below."}],"related":[{"label":"English–Hebrew Code-Switching Sentences","url":"https://huggingface.co/datasets/danielrosehill/English-Hebrew-Codeswitching-Sentences","note":"Derived subset: the rows tagged code_switch"},{"label":"Tech-Sentences","url":"https://huggingface.co/datasets/danielrosehill/Tech-Sentences","note":"Derived subset: the rows tagged tech"},{"label":"Jewish-Religious-Terminology-Sentences","url":"https://huggingface.co/datasets/Hebrish/Jewish-Religious-Terminology-Sentences","note":"Mirror of the rows tagged religious"},{"label":"English–Hebrew Mixed Sentences","url":"https://huggingface.co/datasets/Hebrish/English-Hebrew-Mixed-Sentences","note":"The 2025 desktop-mic round imported here"},{"label":"All training datasets (collection)","url":"https://huggingface.co/collections/Hebrish/parakeet-v2-fine-tune-training-data-6ad63c739d300685a62edb3e","note":"Every dataset behind the Parakeet v2 fine-tunes"},{"label":"Code-switching benchmark","url":"https://huggingface.co/datasets/Hebrish/English-Hebrew-Codeswitching-ASR-Benchmark","note":"The frozen test set cut from this dataset's test split"},{"label":"Hebrish website","url":"https://hebrish.app","note":"The apps, the results, and every model and dataset in one place"},{"label":"Project tracker (GitHub)","url":"https://github.com/danielrosehill/English-Hebrew-Codeswitching-Parakeet-Finetune","note":"Status, data counts and log"}],"tags":["Hebrish","ASR","speech","code-switching","Hebrew","English","dictation","fine-tuning"],"page":"https://ai.danielrosehill.com/datasets/read-aloud-voice-sentences"},{"slug":"english-hebrew-codeswitching-asr-benchmark","title":"English–Hebrew Code-Switching ASR Benchmark","shortTitle":"English–Hebrew Code-Switching ASR Benchmark","eyebrow":"Dataset · benchmark test set · Hebrish","metaLine":"English with Hebrew · frozen test set v1","tagline":"A frozen, versioned test set for scoring how well speech recognition writes the Hebrew words in English speech, \"pick up some bamba from the makolet\", with stock models and fine-tunes scored the same way on a public leaderboard.","status":"Test set v1 · early stage","license":"CC-BY-4.0","hfId":"Hebrish/English-Hebrew-Codeswitching-ASR-Benchmark","hfUrl":"https://huggingface.co/datasets/Hebrish/English-Hebrew-Codeswitching-ASR-Benchmark","parent":{"label":"Read-Aloud Voice Sentences","url":"https://huggingface.co/datasets/danielrosehill/Read-Aloud-Voice-Sentences","note":"Cut from its held-out test split at revision 3bdaa2a50d52"},"browserUrl":"https://huggingface.co/spaces/danielrosehill/English-Hebrew-Codeswitching-ASR-Leaderboard","browserLabel":"Leaderboard","usedByLabel":"Models scored on it","parentLabel":"Cut from","trackerUrl":"https://github.com/Hebrish-ASR/English-Hebrew-Codeswitching-ASR-Benchmark","gated":"The test set is gated on Hugging Face (manual approval): request access to download the audio. The runner, scorer, results and leaderboard are public.","usedBy":[{"slug":"parakeet-tdt-v2-english-hebrew-all-subsets","label":"Hebrish Sabich v1.0: 73.8% term recall, 3.3% WER (round 4)","url":"https://huggingface.co/Hebrish/Sabich"},{"slug":"parakeet-english-hebrew-codeswitching","label":"Hebrish Jachnun v1.0: 73.3% term recall, 3.3% WER (round 3)","url":"https://huggingface.co/Hebrish/Jachnun"},{"label":"PoC-1 to PoC-3: 62.2%, 65.7%, 67.4% term recall","url":"https://huggingface.co/danielrosehill/Parakeet-TDT-0.6B-v2-English-Hebrew-Codeswitching"},{"label":"Shakshuka PoC-2 (Nemotron streaming, scored offline): 62.2%","url":"https://huggingface.co/Hebrish/Nemotron-Speech-Streaming-EN-0.6B-English-Hebrew-Codeswitching"},{"label":"Whisper-Hebrish: 54.1% (known training-data overlap)","url":"https://huggingface.co/danielrosehill/Whisper-Hebrish"}],"summary":["English speakers in Israel drop Hebrew words into English, and English speech recognition writes the nearest English-sounding thing (Macaulay for makolet). This benchmark measures how often a model gets those words right, and whether it still transcribes plain English well. The references write each Hebrew word in one canonical romanized spelling; Hebrew script in the output is never a hit.","Test set v1 is 145 clips from the held-out test split of Read-Aloud Voice Sentences, each tagged with a subset. The headline is the clean subset: 84 phone recordings (11.8 min) of sentences that appear in no published training set, holding 172 Hebrew term occurrences of 139 distinct terms. english_control (24 clips) is the same speaker's English with no Hebrew, a regression check. The rest (contaminated_text, contaminated_audio, desktop_mic) are reported only in the all_code_switch view.","How it is scored. Term recall, the headline, is the share of Hebrew term occurrences whose canonical spelling appears in the transcript as a whole word, case-insensitive, with hyphen and space interchangeable; never a substring. Unseen-term recall counts only the 117 occurrences of terms absent from the source training split. Lenient recall also accepts the lexicon's variant spellings (tsaharon for tzaharon), which is fairer to stock models. WER is normalised (lowercase, punctuation removed), and English control WER catches a fine-tune that damages English.","Contamination is checked, not assumed. The 2025 desktop clips and the sentences Whisper-Hebrish was trained on were traced to its exact training files and split out of the headline subset (2026-10-15). Before a fine-tune is published, the benchmark's 145 clip ids and texts are checked against its training manifest; none were found for Jachnun v1.0 or the Nemotron fine-tune.","The test set never changes in place: a new test set is a new version, and every result records the commit it was scored against. Results are stored with every transcript, so any score can be recomputed with the public scorer. Stock models on the same clips: Whisper Large V3 45.9%, Whisper Large V3 Turbo 40.1%, Parakeet TDT 0.6B v2 32.6%, Parakeet v3 31.4%."],"speaker":["One speaker: Daniel Rosehill, the author, Irish-born, English-speaking, living in Jerusalem. Results rank models on this speaker's code-switching; they do not show how a model handles other accents or spontaneous speech.","Read speech. The clean and english_control clips were recorded on a OnePlus Nord 3 5G internal microphone (2026-10-14/15); the desktop_mic and contaminated_audio clips are desktop-mic recordings from 2025-11-17.","16 kHz mono WAV."],"stats":[{"value":"145","label":"clips in test set v1"},{"value":"84","label":"clean headline clips (11.8 min)"},{"value":"172","label":"Hebrew term occurrences (139 terms)"},{"value":"117","label":"occurrences of unseen terms"}],"statsNote":"Test set v1, frozen at dataset revision 3cb0b2b3b3e4 (2026-10-15). Later commits change only the card. 24 english_control clips (3.3 min) sit alongside the clean subset; 17.3 min of audio in all.","layoutIntro":"A single test split: data/test/audio/<id>.wav with one metadata.jsonl. The runner loads it at the pinned revision, so every result scores the same audio.","layout":[{"field":"file_name","description":"WAV, 16 kHz mono"},{"field":"id","description":"Same id as in Read-Aloud Voice Sentences"},{"field":"text","description":"Reference transcript: cased and punctuated, canonical romanized Hebrew"},{"field":"subset","description":"clean, english_control, contaminated_text, contaminated_audio or desktop_mic"},{"field":"hebrew_terms","description":"List of {term, variants, unseen}: the canonical spelling, accepted variants from the Hebrew-Latin token classifier lexicon, and whether the term is absent from the source train split"},{"field":"source_text","description":"Set on the 4 rows whose reference was corrected to the canonical term; the original transcript"},{"field":"duration_s / categories / recording / source_dataset","description":"Provenance"}],"priorities":["v1.1: relabel the 22 of 139 clean-subset terms that are English tech words or product names (JSON, grep, Proxmox, Tuya, ...), 26 of 172 occurrences, and rescore every model from its stored transcripts. Ranks are unaffected; term recall is 2 to 6 points high until then.","More speakers, better labels and more models: contributions are welcome as pull requests with a result file."],"checkpoints":[{"label":"Test set v1 and round 1","state":"done","date":"2026-10-15","note":"145 clips from Read-Aloud-Voice-Sentences@3bdaa2a50d52, contamination-tagged; 4 references canonicalised. Runner, scorer and leaderboard Space published with seven models: four stock and three fine-tunes."},{"label":"Known issue flagged","state":"done","date":"2026-10-15","note":"Marked early-stage: 22 of the 139 term labels are not Hebrew."},{"label":"Round 2","state":"done","date":"2026-10-18","note":"PoC-3 (published as v3): 67.4% term recall."},{"label":"Round 3","state":"done","date":"2026-10-19","note":"Jachnun v1.0 (73.3%, 3.3% WER) and the Nemotron streaming fine-tune (62.2%, 4.7%). No benchmark clip in either training manifest."},{"label":"Round 4","state":"done","date":"2026-10-20","note":"Sabich v1.0: 73.8% term recall, 3.3% WER; level with Jachnun within noise."},{"label":"Test set v1.1","state":"planned","note":"Relabel the non-Hebrew terms and rescore every stored result."}],"related":[{"label":"Leaderboard (Space)","url":"https://huggingface.co/spaces/danielrosehill/English-Hebrew-Codeswitching-ASR-Leaderboard","note":"Every model on test set v1, with the dated round archive and data dictionary"},{"label":"Runner and scorer (GitHub)","url":"https://github.com/Hebrish-ASR/English-Hebrew-Codeswitching-ASR-Benchmark","note":"ehcs-bench: run, score and rescore; methodology, contamination and limits"},{"label":"Benchmark write-up (blog)","url":"https://blog.danielrosehill.com/posts/english-hebrew-code-switching-asr-benchmark","note":"What the benchmark measures, what it caught, and the results"},{"label":"Benchmarks on hebrish.app","url":"https://hebrish.app/benchmarks","note":"This benchmark plus the held-out splits the Hebrish models are scored on"},{"label":"Read-Aloud Voice Sentences","url":"https://huggingface.co/datasets/danielrosehill/Read-Aloud-Voice-Sentences","note":"Source dataset; the test set is cut from its test split"},{"label":"Hebrew-Latin token classifier","url":"https://github.com/Hebrish-ASR/Hebrew-Latin-Token-Classifier","note":"Lexicon behind the canonical spellings and accepted variants"}],"tags":["Hebrish","ASR","benchmark","evaluation","code-switching","Hebrew","English","leaderboard"],"page":"https://ai.danielrosehill.com/datasets/english-hebrew-codeswitching-asr-benchmark"},{"slug":"hebrew-latin-code-switching","title":"Hebrew-in-Latin Code-Switching Corpus","shortTitle":"Hebrew-in-Latin Code-Switching Corpus","kind":"text","eyebrow":"Dataset · token classification · Hebrish","metaLine":"English text with romanized Hebrew · BIO spans","tagline":"English sentences with Hebrew words written in Latin letters, every Hebrew span marked token by token, for finding those words in text before text-to-speech mispronounces them. No corpus like it had been published.","status":"v1 · released","license":"MIT","hfId":"Hebrish/hebrew-latin-code-switching","hfUrl":"https://huggingface.co/datasets/Hebrish/hebrew-latin-code-switching","trackerUrl":"https://github.com/Hebrish-ASR/Hebrew-Latin-Token-Classifier","usedBy":[{"slug":"hebrew-latin-token-classifier","label":"Hebrew-Latin Token Classifier (span F1 0.941 on the test split)","url":"https://huggingface.co/Hebrish/Hebrew-Latin-Token-Classifier"}],"summary":["English text-to-speech reads romanized Hebrew as English: makolet comes out as \"MACK-oh-let\". The fix is to find the Hebrew spans, convert them to Hebrew script and send each to a Hebrew voice. This corpus trains the classifier that does the finding. There was no heb_Latn label in GlotLID v3, lid.176 or lid218e, no token-level code-switching model covering Hebrew, and no Hebrew-English code-switching corpus of any kind.","The labelling test is pronunciation, not etymology: would an English voice mispronounce this word, and would treating it as Hebrew fix that? So kosher, Shabbat and hummus are negatives, while makolet, mashkanta and arnona are positives. 574 sentences are hard negatives built to trap a classifier: English senses, substring traps, anglicised loanwords and lookalikes.","Generation was term-seeded: a model was given one Hebrew term and asked for sentences using it, and any sentence where the term did not match at a word boundary was rejected. A second model from a different family then annotated each sentence blind, without being told the seed term. Agreements (87.7%) were accepted, disagreements went to human review, and a random 10% of agreements was audited too: 757 human-reviewed tasks in all.","458 sentences are dictated: the text of sentences read aloud for the Hebrish speech fine-tunes, so a human chose to say every one. Their labels are recomputed against the term inventory, not copied. The rest of the corpus is synthetic; treat the audit as the quality estimate.","The splits are term-disjoint: no seed term appears in more than one split, so a model is scored on Hebrew words it has never seen."],"stats":[{"value":"4,737","label":"sentences"},{"value":"1,235","label":"seed terms"},{"value":"458","label":"dictated sentences"},{"value":"574","label":"hard negatives"}],"statsNote":"Counted from the dataset files at revision 2364d2f2654c (built 2026-10-15; unchanged since apart from the card). Splits: train 3,802 sentences / 987 seed terms, validation 467 / 124, test 468 / 124. 3,855 marked spans; 216 sentences carry masked spans.","layoutIntro":"One JSON Lines file per split, data/{train,validation,test}.jsonl, plus terms.csv: the term inventory with category, status and accepted spelling variants. load_dataset(\"Hebrish/hebrew-latin-code-switching\") works directly.","layout":[{"field":"id","description":"Row id (g… for generated)"},{"field":"text","description":"The English sentence"},{"field":"source","description":"generated or dictated"},{"field":"seed_term / category","description":"The term the sentence was generated around, and its topic (bureaucracy, food_shopping, … or negative:<kind> for hard negatives)"},{"field":"spans","description":"Hebrew spans as {start, end, surface, term}, character offsets into text"},{"field":"masked","description":"Character spans with no label: model disagreements no human has ruled on, and Israeli person names. Ignore their tokens when training or scoring (label -100)"},{"field":"tokens","description":"[token, tag] pairs, tags O / B-HE / I-HE; multi-word terms such as bituach leumi are B-HE I-HE"}],"licenseNote":"MIT. Built in Hebrish-ASR/Hebrew-Latin-Token-Classifier, which holds the generation, annotation and review code and the annotation policy.","checkpoints":[{"label":"Corpus built and published","state":"done","date":"2026-10-15","note":"4,737 sentences over 1,235 seed terms, term-disjoint train / validation / test splits, plus terms.csv."},{"label":"Token classifier v1 trained on it","state":"done","date":"2026-10-15","note":"xlm-roberta-base fine-tune: span F1 0.941 on the test split, 0 false positives on the 58 hard-negative test sentences."},{"label":"Moved to the Hebrish organisation","state":"done","date":"2026-10-19","note":"Now Hebrish/hebrew-latin-code-switching; data unchanged."}],"related":[{"label":"Hebrew-Latin Token Classifier","url":"https://huggingface.co/Hebrish/Hebrew-Latin-Token-Classifier","note":"The model trained on this corpus"},{"label":"Build code and annotation policy (GitHub)","url":"https://github.com/Hebrish-ASR/Hebrew-Latin-Token-Classifier","note":"Generation, blind annotation, review and split building"},{"label":"Code-switching TTS pipeline","url":"https://github.com/Hebrish-ASR/English-Hebrew-Code-Switching-TTS-Pipeline","note":"Detect, transliterate, niqqud, tag, speak: where the classifier is used"},{"label":"English–Hebrew Code-Switching Sentences","url":"https://huggingface.co/datasets/danielrosehill/English-Hebrew-Codeswitching-Sentences","note":"Source of the 458 dictated sentences (text only)"},{"label":"English–Hebrew Mixed Sentences","url":"https://huggingface.co/datasets/Hebrish/English-Hebrew-Mixed-Sentences","note":"Human-read audio, the predecessor"}],"tags":["Hebrish","token classification","code-switching","Hebrew","romanization","text-to-speech","BIO"],"page":"https://ai.danielrosehill.com/datasets/hebrew-latin-code-switching"}],"benchmarks":{"about":"Registry of Daniel Rosehill's public benchmarks and evals. Source of truth for the README table (scripts/build_table.py), the Hugging Face Space danielrosehill/My-Benchmarks and danielrosehill.com/ai/benchmarks. Edit this file, then run the script. Private repos are never listed.","updated":"2026-10-21","stages":{"results":{"label":"Results","meaning":"Run, with published results"},"built":{"label":"Built","meaning":"Repo, test material and harness exist; not yet fully run"},"prompts":{"label":"Prompts only","meaning":"Repo and published test prompts; no harness yet"},"planned":{"label":"Planned","meaning":"Design only: a plan file, or a public repo with no test prompts yet"}},"categories":{"behaviour":"Model behaviour & AI safety","speech":"Speech recognition & audio","image":"Image generation","hardware":"Local inference & hardware"},"benchmarks":[{"id":"english-hebrew-codeswitching-asr","name":"English–Hebrew Code-Switching ASR Benchmark","category":"speech","stage":"results","flagship":true,"summary":"Speech-to-text on English speech that contains Hebrew words: a frozen test set, runner, scorer and public leaderboard, comparing stock models with code-switching fine-tunes.","repo":"https://github.com/Hebrish-ASR/English-Hebrew-Codeswitching-ASR-Benchmark","dataset":"https://huggingface.co/datasets/Hebrish/English-Hebrew-Codeswitching-ASR-Benchmark","space":"https://huggingface.co/spaces/danielrosehill/English-Hebrew-Codeswitching-ASR-Leaderboard","date":"2026-10"},{"id":"the-pc-bench","name":"The PC Bench","category":"behaviour","stage":"prompts","flagship":true,"summary":"Does the model scold you? Measures lecturing, moralising, terminology correction and unrequested disclaimers on legitimate requests. 24 mirrored pairs (Western-progressive vs conservative/religious triggers) plus neutral controls.","repo":"https://github.com/danielrosehill/The-PC-Bench","dataset":"https://huggingface.co/datasets/danielrosehill/The-PC-Bench","prompts":54,"date":"2026-10-19"},{"id":"propaganda-bait-eval","name":"Propaganda Bait Eval","category":"behaviour","stage":"built","flagship":true,"summary":"Feeds models loaded political premises across six series (CCP, MAGA, pro-Israel, pro-Palestine, Kremlin, US-progressive) and scores how readily each framing is repeated, with per-series selectivity.","repo":"https://github.com/danielrosehill/Propaganda-Bait-Eval","dataset":"https://huggingface.co/datasets/danielrosehill/Propaganda-Bait-Eval","prompts":72,"date":"2026-10-19"},{"id":"conspiracy-endorsement-eval","name":"Conspiracy Endorsement Eval","category":"behaviour","stage":"results","summary":"Whether a model pushes back on, endorses or encourages 45 conspiracy theories, each asked in a curious and a believer framing. First hand-graded run (on-air-01): 7 models.","repo":"https://github.com/danielrosehill/Conspiracy-Endorsement-Eval","dataset":"https://huggingface.co/datasets/danielrosehill/Conspiracy-Endorsement-Eval","prompts":98,"date":"2026-10-19"},{"id":"israel-palestine-framing-eval","name":"Israel–Palestine Framing Eval","category":"behaviour","stage":"results","summary":"Whether a model treats mirrored pro-Israel and pro-Palestinian framings of the same events differently: loaded premises, framing pairs and open-neutral questions. First graded run (on-air-01): 6 models.","repo":"https://github.com/danielrosehill/Israel-Palestine-Framing-Eval","dataset":"https://huggingface.co/datasets/danielrosehill/Israel-Palestine-Framing-Eval","prompts":70,"date":"2026-10-20"},{"id":"american-centrism-eval","name":"American Centrism Eval","category":"behaviour","stage":"prompts","summary":"Whether a model assumes the user is in the US when nothing says so: geography, units, spelling, law and culture defaults, with controls.","repo":"https://github.com/danielrosehill/American-Centrism-Eval","dataset":"https://huggingface.co/datasets/danielrosehill/American-Centrism-Eval","prompts":48,"date":"2026-10-19"},{"id":"dan-eval","name":"Dan-Eval","category":"behaviour","stage":"built","summary":"A composite of the model-behaviour checks I run before adopting a model for an app, giving one scorecard per model. Each component lives in its own repo.","repo":"https://github.com/danielrosehill/Dan-Eval","date":"2026-10-19"},{"id":"audio-understanding-bitrate-eval","name":"Audio Understanding Bitrate Eval","category":"speech","stage":"results","summary":"How MP3 bitrate affects transcription accuracy across every audio-input LLM on OpenRouter.","repo":"https://github.com/danielrosehill/Audio-Understanding-Bitrate-Eval-0426","dataset":"https://huggingface.co/datasets/danielrosehill/Audio-Understanding-Bitrate-Eval-0426","date":"2026-04"},{"id":"whisper-fine-tune-accuracy-eval","name":"Whisper Fine-Tune Accuracy Eval","category":"speech","stage":"results","summary":"WER of personal Whisper fine-tunes vs stock Whisper at each model size. Fine-tuning helped the small models, hurt the large ones, and improved code-switching at every size.","repo":"https://github.com/danielrosehill/Whisper-Fine-Tune-Accuracy-Eval"},{"id":"whisper-wpm-background-noise-eval","name":"Whisper WPM & Background Noise Eval","category":"speech","stage":"results","summary":"How speaking pace and background noise affect ASR word error rate.","repo":"https://github.com/danielrosehill/Whisper-WPM-Background-Noise-Eval","dataset":"https://huggingface.co/datasets/danielrosehill/ASR-WPM-And-Background-Noise-Eval"},{"id":"long-form-audio-eval","name":"Long-Form Audio Eval","category":"speech","stage":"results","summary":"STT models on a 27-minute podcast against a professional reference transcript: WER, CER and punctuation accuracy.","repo":"https://github.com/danielrosehill/Long-Form-Audio-Eval"},{"id":"one-shot-transcription-microphone-eval","name":"One-Shot Transcription Microphone Eval","category":"speech","stage":"results","summary":"STT accuracy across a range of microphones recording the same material.","repo":"https://github.com/danielrosehill/One-Shot-Transcription-Microphone-Eval"},{"id":"transcription-cleanup-eval","name":"Transcription Cleanup Eval","category":"speech","stage":"results","summary":"Cloud audio-multimodal models transcribing and cleaning up a recording in one step. Subjective grading.","repo":"https://github.com/danielrosehill/Transcription-Cleanup-Eval-1225","date":"2025-12"},{"id":"local-asr-stt-benchmark","name":"Local ASR STT Benchmark","category":"speech","stage":"results","summary":"Whisper model sizes in Speech Note on Ubuntu with AMD/ROCm. A rough first-pass comparison.","repo":"https://github.com/danielrosehill/Local-ASR-STT-Benchmark"},{"id":"hebrew-image-generation-eval","name":"Hebrew Image Generation Eval","category":"image","stage":"results","summary":"Twelve text-to-image models scored pass/fail on rendering Hebrew text correctly.","repo":"https://github.com/danielrosehill/Hebrew-Image-Generation-Eval","dataset":"https://huggingface.co/datasets/danielrosehill/Hebrew-Image-Eval-111225","date":"2025-12"},{"id":"gemma-4-benchmarks","name":"Gemma 4 Benchmarks","category":"hardware","stage":"results","summary":"Gemma 4 E4B vs Qwen3-VL 8B via Ollama on a CPU-only i7-12700F with 64 GB RAM.","repo":"https://github.com/danielrosehill/Gemma-4-Benchmarks-0426","date":"2026-04"},{"id":"workstation-benchmarks","name":"Workstation Benchmarks","category":"hardware","stage":"results","summary":"Desktop workstation benchmarks (i7-12700F, RX 7700 XT, 64 GB DDR5, Ubuntu 25.10) with the full system inventory.","repo":"https://github.com/danielrosehill/Workstation-Benchmarks-0226","date":"2026-02"},{"id":"asian-model-state-censorship","name":"State Censorship in Asian Models","category":"behaviour","stage":"planned","summary":"State-aligned censorship and framing in Chinese and other Asian models.","plan":"https://github.com/danielrosehill/My-Benchmarks/blob/master/benchmark-plans/asian-model-state-censorship.md"},{"id":"us-partisan-framing-eval","name":"US Partisan Framing Eval","category":"behaviour","stage":"prompts","summary":"Asked a neutral question about a divisive US issue, does the model answer with the popular Democratic or Republican framing, and on foreign policy, hawkish or dovish? Blinded anchor-based grading, 14 topics plus factual controls.","repo":"https://github.com/danielrosehill/US-Partisan-Framing-Eval","dataset":"https://huggingface.co/datasets/danielrosehill/US-Partisan-Framing-Eval","prompts":54,"date":"2026-10-19"},{"id":"last-hour-news-recall","name":"Last-Hour News Recall","category":"behaviour","stage":"planned","summary":"Recall of the past hour's news by models that claim real-time retrieval.","plan":"https://github.com/danielrosehill/My-Benchmarks/blob/master/benchmark-plans/last-hour-news-recall.md"},{"id":"last-24-hours-news-recall","name":"Last-24-Hours News Recall","category":"behaviour","stage":"planned","summary":"Recall of the past day's news by models that claim real-time retrieval.","plan":"https://github.com/danielrosehill/My-Benchmarks/blob/master/benchmark-plans/last-24-hours-news-recall.md"}]},"collections":[{"title":"Code Switching (English and Hebrew)","url":"https://huggingface.co/collections/danielrosehill/code-switching-english-and-hebrew-all-resources-6acfdcb1e11911797af4c3b3"},{"title":"Hebrew ASR","url":"https://huggingface.co/collections/danielrosehill/hebrew-asr-and-audio-to-text-69cbb72271b4fe75732a465a"},{"title":"Whisper Hebrish","url":"https://huggingface.co/collections/danielrosehill/whisper-hebrish"},{"title":"Android STT favourites","url":"https://huggingface.co/collections/danielrosehill/android-stt-favorites"},{"title":"Small ASR models","url":"https://huggingface.co/collections/danielrosehill/small-asr-models"},{"title":"STT evaluations","url":"https://huggingface.co/collections/danielrosehill/stt-evaluations"},{"title":"STT cleanup (text post-processing)","url":"https://huggingface.co/collections/danielrosehill/stt-cleanup-text-postprocessing"},{"title":"My public audio datasets","url":"https://huggingface.co/collections/danielrosehill/my-public-audio-datasets"},{"title":"Hebrew TTS","url":"https://huggingface.co/collections/danielrosehill/hebrew-tts"}],"huggingface":[{"name":"English-Hebrew-Codeswitching-Sentences","kind":"dataset","url":"https://huggingface.co/datasets/danielrosehill/English-Hebrew-Codeswitching-Sentences","description":"English-Hebrew Code-Switching Sentences English sentences with everyday Hebrew words and Israeli names written in Latin letters, read aloud by one English speaker living in…","speech_or_language":true,"tags":["task_categories:automatic-speech-recognition","language:en","language:he","license:cc-by-4.0","size_categories:1k<n<10k","format:audiofolder","modality:audio","modality:text","library:datasets","library:mlcroissant","doi:10.57967/hf/10992","region:us","speech","asr","fine-tuning","single-speaker","code-switching","hebrew","english"],"updated_at":"2026-10-23T12:37:27.000Z"},{"name":"Tech-Sentences","kind":"dataset","url":"https://huggingface.co/datasets/danielrosehill/Tech-Sentences","description":"Tech Sentences Dictation-style English sentences dense in software, DevOps, cloud, self-hosting and AI vocabulary (product names, CLI tools, acronyms, model names), read aloud…","speech_or_language":true,"tags":["task_categories:automatic-speech-recognition","language:en","license:cc-by-4.0","size_categories:n<1k","region:us","speech","asr","fine-tuning","single-speaker","tech-vocabulary","developer","ai"],"updated_at":"2026-10-23T12:37:27.000Z"},{"name":"AI-Interactions-Sentences","kind":"dataset","url":"https://huggingface.co/datasets/danielrosehill/AI-Interactions-Sentences","description":"AI Interactions Sentences Instructions dictated to AI coding agents (Claude Code, Codex and others) and the git and GitHub workflow around them, read aloud by one speaker:…","speech_or_language":true,"tags":["task_categories:automatic-speech-recognition","language:en","license:cc-by-4.0","size_categories:n<1k","region:us","speech","asr","fine-tuning","single-speaker","tech-vocabulary","ai-agents","developer"],"updated_at":"2026-10-23T12:37:26.000Z"},{"name":"Read-Aloud-Voice-Sentences","kind":"dataset","url":"https://huggingface.co/datasets/danielrosehill/Read-Aloud-Voice-Sentences","description":"Read-Aloud Voice Sentences Read-aloud sentences from one speaker, recorded on a phone (OnePlus Nord 3 5G, internal microphone). Website: hebrish.app: get the Hebrish keyboard…","speech_or_language":true,"tags":["task_categories:automatic-speech-recognition","language:en","language:he","license:cc-by-4.0","size_categories:n<1k","doi:10.57967/hf/10996","region:us","speech","asr","fine-tuning","parakeet","code-switching","single-speaker"],"updated_at":"2026-10-23T12:37:24.000Z"},{"name":"General-English-Sentences","kind":"dataset","url":"https://huggingface.co/datasets/danielrosehill/General-English-Sentences","description":"General English Sentences Plain, everyday English sentences (emails, messages, general prose and homophone pairs) read aloud by one speaker, with no Hebrew or Yiddish words,…","speech_or_language":true,"tags":["task_categories:automatic-speech-recognition","language:en","license:cc-by-4.0","size_categories:n<1k","region:us","speech","asr","fine-tuning","single-speaker","accent-adaptation","english"],"updated_at":"2026-10-23T12:14:55.000Z"},{"name":"Linux-Admin-Sentences","kind":"dataset","url":"https://huggingface.co/datasets/danielrosehill/Linux-Admin-Sentences","description":"Linux Admin Sentences Dictation-style English sentences about Linux desktop and server administration (KDE Plasma, GNOME, systemd, storage, networking, homelab), read aloud by…","speech_or_language":true,"tags":["task_categories:automatic-speech-recognition","language:en","license:cc-by-4.0","size_categories:n<1k","region:us","speech","asr","fine-tuning","single-speaker","tech-vocabulary","linux","sysadmin"],"updated_at":"2026-10-23T12:14:55.000Z"},{"name":"Homophone-Context-Sentences","kind":"dataset","url":"https://huggingface.co/datasets/danielrosehill/Homophone-Context-Sentences","description":"Homophone Context Sentences English sentences whose context settles which spelling of a homophone is meant (their/there, principal/principle, and 58 more sets), read aloud by…","speech_or_language":true,"tags":["task_categories:automatic-speech-recognition","language:en","license:cc-by-4.0","size_categories:n<1k","region:us","speech","asr","fine-tuning","single-speaker","homophones","english"],"updated_at":"2026-10-23T12:14:55.000Z"},{"name":"Smart-Home-Sentences","kind":"dataset","url":"https://huggingface.co/datasets/danielrosehill/Smart-Home-Sentences","description":"Smart Home Sentences Dictation-style English sentences about home automation (Home Assistant, Zigbee, Z-Wave, Matter, Thread, ESPHome, MQTT and device brands), read aloud by…","speech_or_language":true,"tags":["task_categories:automatic-speech-recognition","language:en","license:cc-by-4.0","size_categories:n<1k","region:us","speech","asr","fine-tuning","single-speaker","tech-vocabulary","smart-home","home-assistant"],"updated_at":"2026-10-23T12:13:16.000Z"},{"name":"Personal-Vocabulary-Sentences","kind":"dataset","url":"https://huggingface.co/datasets/danielrosehill/Personal-Vocabulary-Sentences","description":"Personal Vocabulary Sentences English sentences built around names a speech model gets wrong for one speaker: his own name, his company and his project names (Daniel Rosehill,…","speech_or_language":true,"tags":["task_categories:automatic-speech-recognition","language:en","license:cc-by-4.0","size_categories:n<1k","region:us","speech","asr","fine-tuning","single-speaker","personal-vocabulary","english"],"updated_at":"2026-10-23T12:13:16.000Z"},{"name":"Israel-Election-2026-Forecast","kind":"space","url":"https://huggingface.co/spaces/danielrosehill/Israel-Election-2026-Forecast","description":"","speech_or_language":false,"tags":["static","region:us"],"updated_at":"2026-10-22T20:06:15.000Z"},{"name":"Israel-Election-2026-Polls","kind":"dataset","url":"https://huggingface.co/datasets/danielrosehill/Israel-Election-2026-Polls","description":"Israel 2026 Knesset Election: Opinion Polls This dataset contains every published seat-projection poll for the 27 October 2026 Knesset election, from 23 December 2022 to 22…","speech_or_language":false,"tags":["language:en","license:cc-by-sa-4.0","size_categories:10k<n<100k","format:csv","modality:tabular","modality:text","library:datasets","library:pandas","library:polars","library:mlcroissant","region:us","politics","israel","knesset","elections","polling","forecasting"],"updated_at":"2026-10-22T19:27:30.000Z"},{"name":"Israel-Election-2026-Candidate-Lists","kind":"dataset","url":"https://huggingface.co/datasets/danielrosehill/Israel-Election-2026-Candidate-Lists","description":"Israel 2026 Knesset Election: Candidate Lists This dataset contains every candidate list for the election to the 26th Knesset, held on 27 October 2026. Each candidate appears…","speech_or_language":false,"tags":["language:he","language:en","license:cc-by-sa-4.0","size_categories:1k<n<10k","format:csv","modality:tabular","modality:text","library:datasets","library:pandas","library:polars","library:mlcroissant","region:us","politics","israel","knesset","elections"],"updated_at":"2026-10-22T19:27:28.000Z"},{"name":"Israel-Knesset-25-Members","kind":"dataset","url":"https://huggingface.co/datasets/danielrosehill/Israel-Knesset-25-Members","description":"Israel 25th Knesset: Members and Factions This dataset covers the outgoing 25th Knesset, elected on 1 November 2022, at the end of its term ahead of the 27 October 2026…","speech_or_language":false,"tags":["language:en","language:he","license:cc-by-sa-4.0","size_categories:n<1k","format:csv","modality:tabular","modality:text","library:datasets","library:pandas","library:polars","library:mlcroissant","region:us","politics","israel","knesset","elections"],"updated_at":"2026-10-22T19:27:27.000Z"},{"name":"Me-And-My-Voice-Clone","kind":"space","url":"https://huggingface.co/spaces/danielrosehill/Me-And-My-Voice-Clone","description":"","speech_or_language":true,"tags":["static","text-to-speech","voice-cloning","qwen3-tts","hebrish","region:us"],"updated_at":"2026-10-22T17:56:48.000Z"},{"name":"paragraph-inference-postprocessing-v1-gemma-3-6ad7f15e8a82325523cc299a","kind":"collection","url":"https://huggingface.co/collections/danielrosehill/paragraph-inference-postprocessing-v1-gemma-3-6ad7f15e8a82325523cc299a","description":"Paragraph inference intended for use in ASR pipelines as a focused post-processing layer. Transformation: un-paragraphed text in, paragraphed text out","speech_or_language":true,"tags":[],"updated_at":"2026-10-21T00:43:56.148Z"},{"name":"demos-690130f7367b6d3a5c817fac","kind":"collection","url":"https://huggingface.co/collections/danielrosehill/demos-690130f7367b6d3a5c817fac","description":"Hugging Face collection: Demos","speech_or_language":false,"tags":[],"updated_at":"2026-10-21T00:43:43.854Z"},{"name":"Proximate-Pseudonymizer-Demo","kind":"space","url":"https://huggingface.co/spaces/danielrosehill/Proximate-Pseudonymizer-Demo","description":"","speech_or_language":false,"tags":["gradio","region:us"],"updated_at":"2026-10-21T00:39:40.000Z"},{"name":"ASR-Paragraph-Breaks-Demo","kind":"space","url":"https://huggingface.co/spaces/danielrosehill/ASR-Paragraph-Breaks-Demo","description":"","speech_or_language":true,"tags":["gradio","region:us"],"updated_at":"2026-10-21T00:39:09.000Z"},{"name":"pocs-6ad80732fb6699eecb953327","kind":"collection","url":"https://huggingface.co/collections/danielrosehill/pocs-6ad80732fb6699eecb953327","description":"Early stage proof of concept models that require more training data, training, or rearchitecting.","speech_or_language":false,"tags":[],"updated_at":"2026-10-21T00:30:31.353Z"},{"name":"Proximate-Pseudonymizer-0.6B","kind":"model","url":"https://huggingface.co/danielrosehill/Proximate-Pseudonymizer-0.6B","description":"","speech_or_language":false,"tags":["transformers","safetensors","qwen3","text-generation","pseudonymization","anonymization","privacy","pii","text-rewriting","proof-of-concept","conversational","en","multilingual","base_model:qwen/qwen3-0.6b","base_model:finetune:qwen/qwen3-0.6b","doi:10.57967/hf/11167","license:apache-2.0","text-generation-inference","endpoints_compatible","region:us"],"updated_at":"2026-10-21T00:23:43.000Z"},{"name":"Email-Key-Line-Bolding","kind":"dataset","url":"https://huggingface.co/datasets/danielrosehill/Email-Key-Line-Bolding","description":"Email Key-Line Bolding Hand annotations of which words in a dictated email should be bold: any run of whole words within a paragraph, from a phrase to several sentences. The…","speech_or_language":false,"tags":["task_categories:text-generation","language:en","license:cc-by-4.0","size_categories:n<1k","format:json","modality:text","library:datasets","library:dask","library:polars","library:mlcroissant","doi:10.57967/hf/11162","region:us","email","dictation","emphasis","text-formatting"],"updated_at":"2026-10-21T00:19:10.000Z"},{"name":"Email-Key-Line-Bolding-Demo","kind":"space","url":"https://huggingface.co/spaces/danielrosehill/Email-Key-Line-Bolding-Demo","description":"","speech_or_language":false,"tags":["gradio","region:us"],"updated_at":"2026-10-21T00:19:05.000Z"},{"name":"email-key-line-bolding-6ad802836b0ac72aa5d38e60","kind":"collection","url":"https://huggingface.co/collections/danielrosehill/email-key-line-bolding-6ad802836b0ac72aa5d38e60","description":"Hugging Face collection: Email Key Line Bolding","speech_or_language":false,"tags":[],"updated_at":"2026-10-21T00:08:35.334Z"},{"name":"Gemma-3-270M-Email-Key-Line-Bolding","kind":"model","url":"https://huggingface.co/danielrosehill/Gemma-3-270M-Email-Key-Line-Bolding","description":"","speech_or_language":false,"tags":["transformers","safetensors","gemma3_text","text-generation","dictation","email","text-emphasis","on-device","conversational","en","dataset:danielrosehill/email-key-line-bolding","base_model:google/gemma-3-270m-it","base_model:finetune:google/gemma-3-270m-it","doi:10.57967/hf/11163","license:gemma","text-generation-inference","endpoints_compatible","region:us"],"updated_at":"2026-10-20T23:55:47.000Z"},{"name":"ASR-Paragraph-Breaks","kind":"dataset","url":"https://huggingface.co/datasets/danielrosehill/ASR-Paragraph-Breaks","description":"ASR Paragraph Breaks Hand-annotated paragraph breaks for flat speech-to-text output. ASR engines emit punctuated, cased text as one block. This dataset pairs that block (text)…","speech_or_language":true,"tags":["task_categories:text-generation","language:en","license:cc-by-4.0","size_categories:1k<n<10k","format:json","modality:tabular","modality:text","library:datasets","library:dask","library:polars","library:mlcroissant","doi:10.57967/hf/11166","region:us","asr-post-processing","paragraph-segmentation","dictation","voice-notes","text-segmentation"],"updated_at":"2026-10-20T22:55:45.000Z"},{"name":"Gemma-3-270M-ASR-Paragraph-Breaks-ONNX","kind":"model","url":"https://huggingface.co/danielrosehill/Gemma-3-270M-ASR-Paragraph-Breaks-ONNX","description":"","speech_or_language":true,"tags":["onnxruntime","onnx","gemma3_text","asr-post-processing","paragraph-segmentation","on-device","text-generation","conversational","en","base_model:danielrosehill/gemma-3-270m-asr-paragraph-breaks","base_model:quantized:danielrosehill/gemma-3-270m-asr-paragraph-breaks","doi:10.57967/hf/11164","license:gemma","region:us"],"updated_at":"2026-10-20T22:50:37.000Z"},{"name":"Gemma-3-270M-ASR-Paragraph-Breaks","kind":"model","url":"https://huggingface.co/danielrosehill/Gemma-3-270M-ASR-Paragraph-Breaks","description":"","speech_or_language":true,"tags":["transformers","safetensors","gemma3_text","text-generation","asr-post-processing","paragraph-segmentation","dictation","voice-notes","on-device","conversational","en","dataset:danielrosehill/asr-paragraph-breaks","base_model:google/gemma-3-270m-it","base_model:finetune:google/gemma-3-270m-it","doi:10.57967/hf/11165","license:gemma","text-generation-inference","endpoints_compatible","region:us"],"updated_at":"2026-10-20T22:49:44.000Z"},{"name":"audio-datasets-68f25a2c4474685751e5279f","kind":"collection","url":"https://huggingface.co/collections/danielrosehill/audio-datasets-68f25a2c4474685751e5279f","description":"My training datasets for ASR fine tuning","speech_or_language":true,"tags":[],"updated_at":"2026-10-20T22:33:18.118Z"},{"name":"evaluations-68f25938b0a397bf973556b6","kind":"collection","url":"https://huggingface.co/collections/danielrosehill/evaluations-68f25938b0a397bf973556b6","description":"Hugging Face collection: Evaluations","speech_or_language":false,"tags":[],"updated_at":"2026-10-20T22:30:07.338Z"},{"name":"ASR-Paragraph-Annotator","kind":"space","url":"https://huggingface.co/spaces/danielrosehill/ASR-Paragraph-Annotator","description":"","speech_or_language":true,"tags":["gradio","region:us"],"updated_at":"2026-10-20T21:06:53.000Z"},{"name":"audio-training-datasets-sentences-6ad7aff9009411d3759e836f","kind":"collection","url":"https://huggingface.co/collections/danielrosehill/audio-training-datasets-sentences-6ad7aff9009411d3759e836f","description":"Datasets containing groups of sentences for ASR training (fine tuning). Note: datasets containing audio binaries for the actual training runs elsewher","speech_or_language":true,"tags":[],"updated_at":"2026-10-20T18:20:12.473Z"},{"name":"my-public-audio-datasets-6923462704ade4111eedf00d","kind":"collection","url":"https://huggingface.co/collections/danielrosehill/my-public-audio-datasets-6923462704ade4111eedf00d","description":"Open sourced audio datasets for STT/ASR. All recordings by me (Daniel Rosehill) unless otherwise accredited.","speech_or_language":true,"tags":[],"updated_at":"2026-10-20T14:00:55.254Z"},{"name":"ASR-Dataset-Blender","kind":"space","url":"https://huggingface.co/spaces/danielrosehill/ASR-Dataset-Blender","description":"","speech_or_language":true,"tags":["gradio","region:us"],"updated_at":"2026-10-20T13:39:26.000Z"},{"name":"pseudo-personalised-outreach","kind":"dataset","url":"https://huggingface.co/datasets/danielrosehill/pseudo-personalised-outreach","description":"Pseudo-personalised outreach Real unsolicited emails sent to one person, labelled genuine or pseudo_personalised, with every sender identity replaced by a plausible fake. Code,…","speech_or_language":false,"tags":["task_categories:text-classification","language:en","license:cc-by-4.0","size_categories:n<1k","format:json","modality:text","library:datasets","library:pandas","library:polars","library:mlcroissant","region:us","email","cold-outreach","spam","classification","synthetic-identities"],"updated_at":"2026-10-20T13:11:08.000Z"},{"name":"to-check-out-690be379a52bbafc7673bb31","kind":"collection","url":"https://huggingface.co/collections/danielrosehill/to-check-out-690be379a52bbafc7673bb31","description":"Hugging Face collection: To Check Out","speech_or_language":false,"tags":[],"updated_at":"2026-10-20T09:27:35.369Z"},{"name":"decision-making-support-6ab027e55d43860e7cf02961","kind":"collection","url":"https://huggingface.co/collections/danielrosehill/decision-making-support-6ab027e55d43860e7cf02961","description":"Hugging Face collection: Decision Making Support","speech_or_language":false,"tags":[],"updated_at":"2026-10-20T09:15:19.123Z"},{"name":"LLM-Spoken-Steganography","kind":"space","url":"https://huggingface.co/spaces/danielrosehill/LLM-Spoken-Steganography","description":"","speech_or_language":false,"tags":["static","region:us"],"updated_at":"2026-10-20T01:27:44.000Z"},{"name":"Israel-Palestine-Framing-Eval-Runs","kind":"dataset","url":"https://huggingface.co/datasets/danielrosehill/Israel-Palestine-Framing-Eval-Runs","description":"Israel-Palestine-Framing-Eval: runs Model outputs and grades from runs of Israel-Palestine-Framing-Eval. The prompts are a separate dataset,…","speech_or_language":false,"tags":["language:en","license:cc-by-4.0","size_categories:n<1k","format:json","modality:tabular","modality:text","library:datasets","library:pandas","library:polars","library:mlcroissant","region:us","evaluation","bias","llm-evaluation","israel","palestine","misinformation"],"updated_at":"2026-10-19T23:02:24.000Z"},{"name":"datasets-68f25a1c8d25bccc631337b3","kind":"collection","url":"https://huggingface.co/collections/danielrosehill/datasets-68f25a1c8d25bccc631337b3","description":"Hugging Face collection: Datasets","speech_or_language":false,"tags":[],"updated_at":"2026-10-19T22:12:06.167Z"},{"name":"hebrew-datasets-680428e2813deb0a8b8d1271","kind":"collection","url":"https://huggingface.co/collections/danielrosehill/hebrew-datasets-680428e2813deb0a8b8d1271","description":"Datasets in Hebrew","speech_or_language":true,"tags":[],"updated_at":"2026-10-19T22:06:09.577Z"},{"name":"hebrew-audio-datasets-6ad6923f2ea298b0df13adf5","kind":"collection","url":"https://huggingface.co/collections/danielrosehill/hebrew-audio-datasets-6ad6923f2ea298b0df13adf5","description":"Hugging Face collection: Hebrew audio datasets","speech_or_language":true,"tags":[],"updated_at":"2026-10-19T22:05:25.391Z"},{"name":"hebrew-6ad68f42575d0b5073d19e86","kind":"collection","url":"https://huggingface.co/collections/danielrosehill/hebrew-6ad68f42575d0b5073d19e86","description":"Hugging Face collection: Hebrew","speech_or_language":true,"tags":[],"updated_at":"2026-10-19T22:04:24.948Z"},{"name":"hebrew-llms-text-prediction-675f3e982858d6e9652e7bac","kind":"collection","url":"https://huggingface.co/collections/danielrosehill/hebrew-llms-text-prediction-675f3e982858d6e9652e7bac","description":"Language models supporting TTS (predominantly the text generation task). These are almost exclusively fine-tunes of Gemma, Mistral, or the Qwen models","speech_or_language":true,"tags":[],"updated_at":"2026-10-19T22:03:08.929Z"},{"name":"hebrew-asr-and-audio-to-text-69cbb72271b4fe75732a465a","kind":"collection","url":"https://huggingface.co/collections/danielrosehill/hebrew-asr-and-audio-to-text-69cbb72271b4fe75732a465a","description":"ASR and multimodal models supporting Hebrew to English through audio and text token processing.","speech_or_language":true,"tags":[],"updated_at":"2026-10-19T21:56:09.715Z"},{"name":"hebrew-tts-69cbb6f343675b83becaeae1","kind":"collection","url":"https://huggingface.co/collections/danielrosehill/hebrew-tts-69cbb6f343675b83becaeae1","description":"Hugging Face collection: Hebrew TTS","speech_or_language":true,"tags":[],"updated_at":"2026-10-19T21:54:51.039Z"},{"name":"Prompt-Separation","kind":"dataset","url":"https://huggingface.co/datasets/danielrosehill/Prompt-Separation","description":"Voice-typed podcast prompt transcripts decomposed into structured fields: discrete prompts (asks), a list of context chunks, and free-form host_notes. The dataset supports…","speech_or_language":true,"tags":["task_categories:text-classification","task_categories:text-generation","language:en","license:cc-by-4.0","size_categories:1k<n<10k","format:parquet","modality:tabular","modality:text","library:datasets","library:pandas","library:polars","library:mlcroissant","region:us","prompt-engineering","voice-typed","weak-supervision","prompt-separation"],"updated_at":"2026-10-19T20:14:13.000Z"},{"name":"open-data-packages-6ad6766e87ac469cd9db21ec","kind":"collection","url":"https://huggingface.co/collections/danielrosehill/open-data-packages-6ad6766e87ac469cd9db21ec","description":"Datasets packaging (third-party) open access datasets for AI/ML use","speech_or_language":false,"tags":[],"updated_at":"2026-10-19T20:09:06.651Z"},{"name":"image-task-training-6ad678d4493e164691c129ed","kind":"collection","url":"https://huggingface.co/collections/danielrosehill/image-task-training-6ad678d4493e164691c129ed","description":"Hugging Face collection: Image Task Training","speech_or_language":false,"tags":[],"updated_at":"2026-10-19T20:08:52.224Z"},{"name":"israel-context-datasets-6ad676a3773a64192165a96c","kind":"collection","url":"https://huggingface.co/collections/danielrosehill/israel-context-datasets-6ad676a3773a64192165a96c","description":"Datasets intended for (multimodal) training of models with understanding of real-life context of Israel - esp. settings and places","speech_or_language":false,"tags":[],"updated_at":"2026-10-19T20:03:10.053Z"},{"name":"israel-open-data-69fba702e180798fe3761073","kind":"collection","url":"https://huggingface.co/collections/danielrosehill/israel-open-data-69fba702e180798fe3761073","description":"Hugging Face collection: Israel Open Data","speech_or_language":false,"tags":[],"updated_at":"2026-10-19T20:02:33.160Z"},{"name":"jerusalem-68f258e850eb110423d0312f","kind":"collection","url":"https://huggingface.co/collections/danielrosehill/jerusalem-68f258e850eb110423d0312f","description":"Projects related to the city of Jerusalem","speech_or_language":false,"tags":[],"updated_at":"2026-10-19T20:02:06.610Z"},{"name":"evals-runs-datasets-6ad6761f2ea298b0df0f5d55","kind":"collection","url":"https://huggingface.co/collections/danielrosehill/evals-runs-datasets-6ad6761f2ea298b0df0f5d55","description":"Datasets of evaluation runs","speech_or_language":false,"tags":[],"updated_at":"2026-10-19T19:57:50.495Z"},{"name":"Conspiracy-Endorsement-Eval-Runs","kind":"dataset","url":"https://huggingface.co/datasets/danielrosehill/Conspiracy-Endorsement-Eval-Runs","description":"Conspiracy-Endorsement-Eval: runs Raw model outputs from runs of Conspiracy-Endorsement-Eval. The prompts themselves are a separate dataset:…","speech_or_language":false,"tags":["language:en","license:cc-by-4.0","size_categories:n<1k","format:json","modality:text","library:datasets","library:pandas","library:polars","library:mlcroissant","region:us","safety","evaluation","conspiracy-theories","llm-evaluation"],"updated_at":"2026-10-19T19:32:45.000Z"},{"name":"Propaganda-Bait-Eval-Runs","kind":"dataset","url":"https://huggingface.co/datasets/danielrosehill/Propaganda-Bait-Eval-Runs","description":"Propaganda-Bait-Eval: runs Raw model outputs from runs of Propaganda-Bait-Eval. The prompts themselves are a separate dataset: danielrosehill/Propaganda-Bait-Eval. This dataset…","speech_or_language":false,"tags":["language:en","language:zh","language:ru","license:cc-by-4.0","region:us","safety","evaluation","propaganda","censorship","llm-evaluation"],"updated_at":"2026-10-19T19:32:43.000Z"},{"name":"My-Benchmarks","kind":"space","url":"https://huggingface.co/spaces/danielrosehill/My-Benchmarks","description":"","speech_or_language":false,"tags":["static","region:us"],"updated_at":"2026-10-19T19:29:17.000Z"},{"name":"evaluation-datasets-6939bb47d4419e6cc8114ac2","kind":"collection","url":"https://huggingface.co/collections/danielrosehill/evaluation-datasets-6939bb47d4419e6cc8114ac2","description":"Hugging Face collection: Evaluation Datasets","speech_or_language":false,"tags":[],"updated_at":"2026-10-19T19:28:49.931Z"},{"name":"Israel-Palestine-Framing-Eval","kind":"dataset","url":"https://huggingface.co/datasets/danielrosehill/Israel-Palestine-Framing-Eval","description":"Israel-Palestine Framing Eval: test prompts Test prompts for an evaluation of whether a language model leans pro-Israel or pro-Palestinian. The test gives the model mirrored…","speech_or_language":false,"tags":["task_categories:text-generation","language:en","license:cc-by-4.0","size_categories:n<1k","format:json","modality:text","library:datasets","library:pandas","library:polars","library:mlcroissant","region:us","bias","political-bias","evaluation","israel","palestine"],"updated_at":"2026-10-19T19:26:05.000Z"},{"name":"US-Partisan-Framing-Eval","kind":"dataset","url":"https://huggingface.co/datasets/danielrosehill/US-Partisan-Framing-Eval","description":"US Partisan Framing Eval: test prompts Test prompts for an evaluation of whether a language model, asked a neutral question about a divisive US issue, answers with the popular…","speech_or_language":false,"tags":["task_categories:text-generation","language:en","license:cc-by-4.0","size_categories:n<1k","format:json","modality:text","library:datasets","library:pandas","library:polars","library:mlcroissant","region:us","evaluation","political-bias","us-politics","framing","foreign-policy"],"updated_at":"2026-10-19T19:22:27.000Z"},{"name":"The-PC-Bench","kind":"dataset","url":"https://huggingface.co/datasets/danielrosehill/The-PC-Bench","description":"The PC Bench: test prompts Test prompts for an evaluation of scolding: model responses that make the user feel judged for asking, by lecturing, moralising, correcting their…","speech_or_language":false,"tags":["task_categories:text-generation","language:en","license:cc-by-4.0","size_categories:n<1k","format:json","modality:text","library:datasets","library:pandas","library:polars","library:mlcroissant","region:us","evaluation","llm-evaluation","political-correctness","moralizing","over-refusal","political-bias"],"updated_at":"2026-10-19T19:22:06.000Z"},{"name":"Parakeet-TDT-0.6B-v2-English-Hebrew-Codeswitching-GGUF","kind":"model","url":"https://huggingface.co/danielrosehill/Parakeet-TDT-0.6B-v2-English-Hebrew-Codeswitching-GGUF","description":"","speech_or_language":true,"tags":["gguf","transcribe.cpp","parakeet","tdt","code-switching","hebrew","english","automatic-speech-recognition","en","he","base_model:danielrosehill/parakeet-tdt-0.6b-v2-english-hebrew-codeswitching","base_model:quantized:danielrosehill/parakeet-tdt-0.6b-v2-english-hebrew-codeswitching","license:cc-by-4.0","region:us"],"updated_at":"2026-10-19T19:15:43.000Z"},{"name":"Parakeet-TDT-0.6B-v2-English-Hebrew-Codeswitching-ONNX","kind":"model","url":"https://huggingface.co/danielrosehill/Parakeet-TDT-0.6B-v2-English-Hebrew-Codeswitching-ONNX","description":"","speech_or_language":true,"tags":["onnx","nemo-conformer-tdt","onnx-asr","voxtype","parakeet","tdt","code-switching","hebrew","english","automatic-speech-recognition","en","he","base_model:danielrosehill/parakeet-tdt-0.6b-v2-english-hebrew-codeswitching","base_model:quantized:danielrosehill/parakeet-tdt-0.6b-v2-english-hebrew-codeswitching","license:cc-by-4.0","region:us"],"updated_at":"2026-10-19T19:13:16.000Z"},{"name":"Conspiracy-Endorsement-Eval","kind":"dataset","url":"https://huggingface.co/datasets/danielrosehill/Conspiracy-Endorsement-Eval","description":"Conspiracy Endorsement Eval: test prompts Test prompts for an evaluation of whether a language model pushes back on, endorses, or encourages conspiracy theories that a user…","speech_or_language":false,"tags":["task_categories:text-generation","language:en","license:cc-by-4.0","size_categories:n<1k","format:json","modality:text","library:datasets","library:pandas","library:polars","library:mlcroissant","region:us","evaluation","safety","conspiracy-theories","misinformation","antisemitism","sycophancy"],"updated_at":"2026-10-19T19:06:23.000Z"},{"name":"Propaganda-Bait-Eval","kind":"dataset","url":"https://huggingface.co/datasets/danielrosehill/Propaganda-Bait-Eval","description":": test prompts Status: proof of concept (2026-10-19). Version 0.1. The prompts have been piloted on a small smoke run only. Expect wording to change. Pin a commit if you use…","speech_or_language":false,"tags":["task_categories:text-generation","language:en","license:cc-by-4.0","size_categories:n<1k","format:json","modality:text","library:datasets","library:pandas","library:polars","library:mlcroissant","region:us","bias","political-bias","propaganda","censorship","sycophancy","evaluation"],"updated_at":"2026-10-19T19:00:13.000Z"},{"name":"American-Centrism-Eval","kind":"dataset","url":"https://huggingface.co/datasets/danielrosehill/American-Centrism-Eval","description":"American Centrism Eval: test prompts Test prompts for an evaluation of whether a language model assumes the user is in the United States when nothing says so, and more broadly…","speech_or_language":false,"tags":["task_categories:text-generation","language:en","license:cc-by-4.0","size_categories:n<1k","format:json","modality:text","library:datasets","library:pandas","library:polars","library:mlcroissant","region:us","bias","cultural-bias","evaluation","localization"],"updated_at":"2026-10-19T13:34:59.000Z"},{"name":"parakeet-v2-english-hebrew-codeswitching-finetune-6ad26d79c70d28a49ed71c78","kind":"collection","url":"https://huggingface.co/collections/danielrosehill/parakeet-v2-english-hebrew-codeswitching-finetune-6ad26d79c70d28a49ed71c78","description":"Base fine tune and variants of Parakeet V2 English-Hebrew codeswitching finetune. Website: hebrish.app","speech_or_language":true,"tags":[],"updated_at":"2026-10-19T12:33:45.017Z"},{"name":"deepseek-v41-6ab00ff7592c120056c363a0","kind":"collection","url":"https://huggingface.co/collections/danielrosehill/deepseek-v41-6ab00ff7592c120056c363a0","description":"Hugging Face collection: Deepseek V4.1","speech_or_language":false,"tags":[],"updated_at":"2026-10-18T04:18:18.251Z"},{"name":"deepseek-41-flash-point-releases-6ad447df4a0566f79cd43a9c","kind":"collection","url":"https://huggingface.co/collections/danielrosehill/deepseek-41-flash-point-releases-6ad447df4a0566f79cd43a9c","description":"Hugging Face collection: Deepseek 4.1 Flash point releases","speech_or_language":false,"tags":[],"updated_at":"2026-10-18T04:15:27.370Z"},{"name":"code-switching-english-and-hebrew-all-resources-6acfdcb1e11911797af4c3b3","kind":"collection","url":"https://huggingface.co/collections/danielrosehill/code-switching-english-and-hebrew-all-resources-6acfdcb1e11911797af4c3b3","description":"Datasets and fine tunes (mine) focused on English (primary) with Hebrew code-switching. Targeted use-case: improved ASR accuracy among olim","speech_or_language":true,"tags":[],"updated_at":"2026-10-17T15:20:49.303Z"},{"name":"Annotated-Voice-Notes","kind":"dataset","url":"https://huggingface.co/datasets/danielrosehill/Annotated-Voice-Notes","description":"Annotated Voice Notes A small, hand-picked set of real voice notes. It is kept small on purpose. Each note has its audio, a raw AssemblyAI transcript, and a written annotation…","speech_or_language":true,"tags":["task_categories:automatic-speech-recognition","task_categories:text-classification","task_categories:token-classification","language:en","license:cc-by-4.0","size_categories:n<1k","format:audiofolder","modality:audio","modality:text","library:datasets","library:mlcroissant","region:us","voice-notes","task-extraction","disfluency","long-form","real-speech"],"updated_at":"2026-10-17T13:03:46.000Z"},{"name":"fine-tune-candidates-1026-6acf9c375edb1848d3baa0e6","kind":"collection","url":"https://huggingface.co/collections/danielrosehill/fine-tune-candidates-1026-6acf9c375edb1848d3baa0e6","description":"Candidates for a fine tuning idea for audio to text with deterministic post processing edits applied in one forward pass","speech_or_language":true,"tags":[],"updated_at":"2026-10-17T10:12:14.307Z"},{"name":"utilities-6954092ee990ed08f5c52627","kind":"collection","url":"https://huggingface.co/collections/danielrosehill/utilities-6954092ee990ed08f5c52627","description":"Hugging Face collection: Utilities","speech_or_language":false,"tags":[],"updated_at":"2026-10-17T09:59:14.477Z"},{"name":"to-read-6ad346553351cecf110cc3cc","kind":"collection","url":"https://huggingface.co/collections/danielrosehill/to-read-6ad346553351cecf110cc3cc","description":"Hugging Face collection: To Read","speech_or_language":false,"tags":[],"updated_at":"2026-10-17T09:57:52.932Z"},{"name":"benchmarks-67fbe5e1ade4f6ffafeb7f88","kind":"collection","url":"https://huggingface.co/collections/danielrosehill/benchmarks-67fbe5e1ade4f6ffafeb7f88","description":"Hugging Face collection: Benchmarks","speech_or_language":false,"tags":[],"updated_at":"2026-10-17T09:53:09.291Z"},{"name":"tools-6ad342862b4bd6a95101e9bf","kind":"collection","url":"https://huggingface.co/collections/danielrosehill/tools-6ad342862b4bd6a95101e9bf","description":"Hugging Face collection: Tools","speech_or_language":false,"tags":[],"updated_at":"2026-10-17T09:40:22.152Z"},{"name":"classifiers-6ab1917d0c0b293bdccb9deb","kind":"collection","url":"https://huggingface.co/collections/danielrosehill/classifiers-6ab1917d0c0b293bdccb9deb","description":"Hugging Face collection: Classifiers","speech_or_language":false,"tags":[],"updated_at":"2026-10-16T18:37:47.819Z"},{"name":"Parakeet-Dictation-Cleanup-Eval","kind":"dataset","url":"https://huggingface.co/datasets/danielrosehill/Parakeet-Dictation-Cleanup-Eval","description":"Parakeet Dictation Cleanup Eval Status (2026-10-16): placeholder. No clips recorded yet. Evaluation only. Never train on this set. Planned contents: 40 spontaneous dictations…","speech_or_language":true,"tags":["task_categories:text-generation","language:en","license:cc-by-4.0","size_categories:n<1k","region:us","asr","dictation","post-processing","evaluation","parakeet"],"updated_at":"2026-10-16T15:29:07.000Z"},{"name":"English-Hebrew-Codeswitching-ASR-Leaderboard","kind":"space","url":"https://huggingface.co/spaces/danielrosehill/English-Hebrew-Codeswitching-ASR-Leaderboard","description":"","speech_or_language":true,"tags":["gradio","leaderboard","asr","code-switching","hebrew","israel","codeswitching","speech-recognition","hebrew-nlp","english-hebrew","region:us"],"updated_at":"2026-10-15T14:08:40.000Z"},{"name":"english-hebrew-codeswitching-datasets-and-training-6ad0c4d11b32f13a90b466ea","kind":"collection","url":"https://huggingface.co/collections/danielrosehill/english-hebrew-codeswitching-datasets-and-training-6ad0c4d11b32f13a90b466ea","description":"Datasets and training materials for English-Hebrew codeswitching ASR and multimodal (audio+text to text) models","speech_or_language":true,"tags":[],"updated_at":"2026-10-15T12:22:46.191Z"},{"name":"english-hebrew-codeswitching-models-6ad0c4a69101a4c065dedc95","kind":"collection","url":"https://huggingface.co/collections/danielrosehill/english-hebrew-codeswitching-models-6ad0c4a69101a4c065dedc95","description":"Fine-tunes of ASR models on English-Hebrew codeswitching (English sentences with code-switching in Hebrew).","speech_or_language":true,"tags":[],"updated_at":"2026-10-15T12:20:47.743Z"},{"name":"inventory-accessories-6ad0c3ebf9e5ebfd04f0af1b","kind":"collection","url":"https://huggingface.co/collections/danielrosehill/inventory-accessories-6ad0c3ebf9e5ebfd04f0af1b","description":"Hugging Face collection: Inventory Accessories","speech_or_language":false,"tags":[],"updated_at":"2026-10-15T12:15:55.487Z"},{"name":"inventory-spec-lines","kind":"dataset","url":"https://huggingface.co/datasets/danielrosehill/inventory-spec-lines","description":"Inventory Spec Lines 14,426 product-specification lines from a real household inventory (cables, adapters, tools, electronics, household goods), each labelled with spans for a…","speech_or_language":false,"tags":["task_categories:token-classification","language:en","license:cc-by-4.0","size_categories:10k<n<100k","format:json","modality:text","library:datasets","library:pandas","library:polars","library:mlcroissant","region:us"],"updated_at":"2026-10-15T11:26:24.000Z"},{"name":"Voice-Notes","kind":"dataset","url":"https://huggingface.co/datasets/danielrosehill/Voice-Notes","description":"Voice Notes (screened) Real voice-note transcripts from one person's Voicenotes export (April 2026), filtered for privacy, for evaluating Voice-Note-Intent-Classifier. How…","speech_or_language":true,"tags":["license:other","size_categories:n<1k","format:json","modality:text","library:datasets","library:pandas","library:polars","library:mlcroissant","region:us","voice-notes","asr","intent-classification"],"updated_at":"2026-10-15T11:02:18.000Z"},{"name":"inventory-spec-extractor","kind":"model","url":"https://huggingface.co/danielrosehill/inventory-spec-extractor","description":"","speech_or_language":false,"tags":["onnx","inventory","product-specifications","information-extraction","token-classification","en","dataset:danielrosehill/inventory-spec-lines","base_model:baai/bge-small-en-v1.5","base_model:quantized:baai/bge-small-en-v1.5","doi:10.57967/hf/10998","license:apache-2.0","region:us"],"updated_at":"2026-10-15T09:14:31.000Z"},{"name":"English-Hebrew-Codeswitching-Parakeet-Demo","kind":"space","url":"https://huggingface.co/spaces/danielrosehill/English-Hebrew-Codeswitching-Parakeet-Demo","description":"","speech_or_language":true,"tags":["gradio","region:us"],"updated_at":"2026-10-15T08:49:09.000Z"},{"name":"Parakeet-TDT-0.6B-v2-English-Hebrew-Codeswitching","kind":"model","url":"https://huggingface.co/danielrosehill/Parakeet-TDT-0.6B-v2-English-Hebrew-Codeswitching","description":"","speech_or_language":true,"tags":["nemo","onnx","gguf","code-switching","hebrew","english","parakeet","tdt","speaker-adaptation","automatic-speech-recognition","en","he","dataset:danielrosehill/read-aloud-voice-sentences","dataset:danielrosehill/english-hebrew-codeswitching-sentences","dataset:hebrish/english-hebrew-mixed-sentences","dataset:openslr/librispeech_asr","base_model:nvidia/parakeet-tdt-0.6b-v2","base_model:finetune:nvidia/parakeet-tdt-0.6b-v2","doi:10.57967/hf/10995","license:cc-by-4.0","region:us"],"updated_at":"2026-10-15T08:41:37.000Z"},{"name":"hebrew-latin-token-classifier","kind":"space","url":"https://huggingface.co/spaces/danielrosehill/hebrew-latin-token-classifier","description":"","speech_or_language":true,"tags":["gradio","region:us"],"updated_at":"2026-10-15T08:30:25.000Z"},{"name":"my-models-classifiers-6ad08ac6093aba9665e48de4","kind":"collection","url":"https://huggingface.co/collections/danielrosehill/my-models-classifiers-6ad08ac6093aba9665e48de4","description":"Hugging Face collection: My Models - Classifiers","speech_or_language":false,"tags":[],"updated_at":"2026-10-15T08:12:04.690Z"},{"name":"Voice-Note-Intent-Classifier","kind":"model","url":"https://huggingface.co/danielrosehill/Voice-Note-Intent-Classifier","description":"","speech_or_language":true,"tags":["sentence-transformers","intent-classification","multi-label-classification","voice-notes","asr","routing","mcp","agents","text-classification","en","base_model:baai/bge-base-en-v1.5","base_model:finetune:baai/bge-base-en-v1.5","license:mit","region:us"],"updated_at":"2026-10-15T07:44:19.000Z"},{"name":"hebrew-latin-token-classifier","kind":"model","url":"https://huggingface.co/danielrosehill/hebrew-latin-token-classifier","description":"","speech_or_language":true,"tags":["transformers","safetensors","xlm-roberta","token-classification","code-switching","hebrew","transliteration","romanization","text-to-speech","en","he","dataset:hebrish/hebrew-latin-code-switching","base_model:facebookai/xlm-roberta-base","base_model:finetune:facebookai/xlm-roberta-base","doi:10.57967/hf/10997","license:mit","model-index","endpoints_compatible","region:us"],"updated_at":"2026-10-15T07:39:39.000Z"},{"name":"embedding-models-68f3e5ea9e4a9c5e39176af6","kind":"collection","url":"https://huggingface.co/collections/danielrosehill/embedding-models-68f3e5ea9e4a9c5e39176af6","description":"Hugging Face collection: Embedding Models","speech_or_language":false,"tags":[],"updated_at":"2026-10-13T09:29:56.435Z"},{"name":"baby-cry-detection-6acdf9ba3d47ac916de9d59c","kind":"collection","url":"https://huggingface.co/collections/danielrosehill/baby-cry-detection-6acdf9ba3d47ac916de9d59c","description":"Hugging Face collection: Baby Cry Detection","speech_or_language":false,"tags":[],"updated_at":"2026-10-13T09:28:26.816Z"},{"name":"asr-models-0926-6ab279a96bea27b574ad5e33","kind":"collection","url":"https://huggingface.co/collections/danielrosehill/asr-models-0926-6ab279a96bea27b574ad5e33","description":"A composite of models I've been trying and using across different devices in September 2026. Both ASR post-processing small models and multimodal.","speech_or_language":true,"tags":[],"updated_at":"2026-09-22T19:48:12.034Z"},{"name":"chatterbox-multilingual-6ab29a79189fd0c363d73502","kind":"collection","url":"https://huggingface.co/collections/danielrosehill/chatterbox-multilingual-6ab29a79189fd0c363d73502","description":"Hugging Face collection: Chatterbox Multilingual","speech_or_language":false,"tags":[],"updated_at":"2026-09-22T15:11:55.209Z"},{"name":"mwp-dataset-navigator","kind":"space","url":"https://huggingface.co/spaces/danielrosehill/mwp-dataset-navigator","description":"","speech_or_language":false,"tags":["static","corpus-linguistics","ai-generated","podcast","region:us"],"updated_at":"2026-09-22T14:53:56.000Z"},{"name":"token-classifiers-6ab27a123c7ea22f6bbc30c5","kind":"collection","url":"https://huggingface.co/collections/danielrosehill/token-classifiers-6ab27a123c7ea22f6bbc30c5","description":"Token classification models. Use-case: Linguistic identification for code switching in TTS script normalization.","speech_or_language":true,"tags":[],"updated_at":"2026-09-22T12:53:10.031Z"},{"name":"stt-cleanup-text-postprocessing-6ab00cf0422ffeb81648a0ce","kind":"collection","url":"https://huggingface.co/collections/danielrosehill/stt-cleanup-text-postprocessing-6ab00cf0422ffeb81648a0ce","description":"Models that apply a text-to-text transformation as a second pass on speech transcribed using ASR.","speech_or_language":true,"tags":[],"updated_at":"2026-09-22T04:47:26.431Z"},{"name":"on-device-asr-one-plus-cph2493-nord-3-6ab1247af50bd5d4bfdb5166","kind":"collection","url":"https://huggingface.co/collections/danielrosehill/on-device-asr-one-plus-cph2493-nord-3-6ab1247af50bd5d4bfdb5166","description":"Models used for on device speech to text with the OnePlus CPH2493.","speech_or_language":true,"tags":[],"updated_at":"2026-09-21T21:21:52.059Z"},{"name":"pidgeon-detection-nvr-6ab19cd724a4608509d6b729","kind":"collection","url":"https://huggingface.co/collections/danielrosehill/pidgeon-detection-nvr-6ab19cd724a4608509d6b729","description":"A few models to try in Frigate for presence detection of pigeons.","speech_or_language":false,"tags":[],"updated_at":"2026-09-21T21:09:48.959Z"},{"name":"small-asr-models-6ab19617a080f7833fa4e94e","kind":"collection","url":"https://huggingface.co/collections/danielrosehill/small-asr-models-6ab19617a080f7833fa4e94e","description":"that aren't Whisper (or derivs). On device inference for smartphones etc","speech_or_language":true,"tags":[],"updated_at":"2026-09-21T21:00:55.943Z"},{"name":"small-text-generation-6ab1923bd5cf3d44b6dc9985","kind":"collection","url":"https://huggingface.co/collections/danielrosehill/small-text-generation-6ab1923bd5cf3d44b6dc9985","description":"Hugging Face collection: Small Text Generation","speech_or_language":false,"tags":[],"updated_at":"2026-09-21T20:27:10.336Z"},{"name":"ivm-and-text-restoration-6ab17f81e542b0d5af03ace8","kind":"collection","url":"https://huggingface.co/collections/danielrosehill/ivm-and-text-restoration-6ab17f81e542b0d5af03ace8","description":"Hugging Face collection: IVM And Text Restoration","speech_or_language":false,"tags":[],"updated_at":"2026-09-21T19:21:51.547Z"},{"name":"punctuation-restoration-english-6ab17edb1784d8fb0439de9e","kind":"collection","url":"https://huggingface.co/collections/danielrosehill/punctuation-restoration-english-6ab17edb1784d8fb0439de9e","description":"Hugging Face collection: Punctuation Restoration- English","speech_or_language":false,"tags":[],"updated_at":"2026-09-21T19:02:04.750Z"},{"name":"encoders-6ab17e7611f3297412fd3a9c","kind":"collection","url":"https://huggingface.co/collections/danielrosehill/encoders-6ab17e7611f3297412fd3a9c","description":"Hugging Face collection: Encoders","speech_or_language":false,"tags":[],"updated_at":"2026-09-21T18:59:02.205Z"},{"name":"medical-68f3fa4fcb57e7f74aa2aee9","kind":"collection","url":"https://huggingface.co/collections/danielrosehill/medical-68f3fa4fcb57e7f74aa2aee9","description":"Hugging Face collection: Medical","speech_or_language":false,"tags":[],"updated_at":"2026-09-21T18:57:31.138Z"},{"name":"interesting-spaces-6ab17c6b22be5fc3c191396d","kind":"collection","url":"https://huggingface.co/collections/danielrosehill/interesting-spaces-6ab17c6b22be5fc3c191396d","description":"Hugging Face collection: Interesting Spaces","speech_or_language":false,"tags":[],"updated_at":"2026-09-21T18:50:19.363Z"},{"name":"android-stt-favorites-6ab00abc05f6ef971705ad01","kind":"collection","url":"https://huggingface.co/collections/danielrosehill/android-stt-favorites-6ab00abc05f6ef971705ad01","description":"Speech to text models that I've had the best experience with on my current hardware (One Plus Nord 3 ). Currently this is mostly ASR models.","speech_or_language":true,"tags":[],"updated_at":"2026-09-21T12:42:25.648Z"},{"name":"deepseek-v41-derivatives-6ab0e34756006c911e558a97","kind":"collection","url":"https://huggingface.co/collections/danielrosehill/deepseek-v41-derivatives-6ab0e34756006c911e558a97","description":"Hugging Face collection: DeepSeek V4.1 Derivatives","speech_or_language":false,"tags":[],"updated_at":"2026-09-21T07:57:53.628Z"},{"name":"on-device-models-6ab0e06a41fc046f99cc7c8d","kind":"collection","url":"https://huggingface.co/collections/danielrosehill/on-device-models-6ab0e06a41fc046f99cc7c8d","description":"Hugging Face collection: On Device Models","speech_or_language":false,"tags":[],"updated_at":"2026-09-21T07:44:42.496Z"},{"name":"use-case-diversity-6ab0289140bcb88107ba30a3","kind":"collection","url":"https://huggingface.co/collections/danielrosehill/use-case-diversity-6ab0289140bcb88107ba30a3","description":"A collection intended as a first entry store for small models, especially component models, which perform modular tightly bounded tasks.","speech_or_language":false,"tags":[],"updated_at":"2026-09-20T18:40:57.183Z"},{"name":"deepseek-v4-derivs-6ab010606122dd2c28756567","kind":"collection","url":"https://huggingface.co/collections/danielrosehill/deepseek-v4-derivs-6ab010606122dd2c28756567","description":"Community derivatives from Deepseek V4 and V4.1","speech_or_language":false,"tags":[],"updated_at":"2026-09-20T16:57:37.736Z"},{"name":"jerusalem-poster-detection","kind":"dataset","url":"https://huggingface.co/datasets/danielrosehill/jerusalem-poster-detection","description":"Jerusalem street poster sightings — detection dataset Geotagged street photographs of a messianic postering campaign in central Jerusalem, with bounding boxes around every…","speech_or_language":false,"tags":["task_categories:object-detection","annotations_creators:expert-generated","language:he","license:cc-by-4.0","size_categories:n<1k","format:imagefolder","modality:image","modality:text","library:datasets","library:mlcroissant","region:us","object-detection","street-photography","urban-studies","jerusalem"],"updated_at":"2026-09-01T10:09:58.000Z"},{"name":"jerusalem-poster-detector","kind":"model","url":"https://huggingface.co/danielrosehill/jerusalem-poster-detector","description":"","speech_or_language":false,"tags":["ultralytics","yolo11","object-detection","street-photography","dataset:danielrosehill/jerusalem-poster-detection","license:cc-by-4.0","region:us"],"updated_at":"2026-09-01T10:03:30.000Z"},{"name":"homegoods-taxonomy","kind":"dataset","url":"https://huggingface.co/datasets/danielrosehill/homegoods-taxonomy","description":"HomeGoods Taxonomy v1 A hierarchical classification of the physical goods in an ordinary home or small business, with a physical handling attribute layer attached to every…","speech_or_language":false,"tags":["task_categories:text-classification","task_categories:zero-shot-classification","language:en","license:mit","size_categories:n<1k","format:csv","modality:tabular","modality:text","library:datasets","library:pandas","library:polars","library:mlcroissant","region:us","taxonomy","classification","inventory","logistics","moving","household","dangerous-goods","hazmat"],"updated_at":"2026-08-03T19:28:44.000Z"},{"name":"storage-container-dimensions","kind":"dataset","url":"https://huggingface.co/datasets/danielrosehill/storage-container-dimensions","description":"Industrial Storage Container Dimensions Planning-grade volumetric reference for the plastic containers that European and North American warehouses actually run on: Euroboxes…","speech_or_language":false,"tags":["language:en","license:cc-by-4.0","size_categories:n<1k","format:csv","modality:document","modality:tabular","modality:text","library:datasets","library:pandas","library:polars","library:mlcroissant","region:us","logistics","supply-chain","warehousing","packaging","reference","tabular"],"updated_at":"2026-07-30T09:50:52.000Z"},{"name":"Israel-Open-Data-Catalogue","kind":"dataset","url":"https://huggingface.co/datasets/danielrosehill/Israel-Open-Data-Catalogue","description":"Israel Open Data Catalogue (data.gov.il) A bilingual (Hebrew/English) snapshot of the entire dataset catalogue published on data.gov.il, Israel's official open-data portal. The…","speech_or_language":true,"tags":["language:he","language:en","license:cc-by-4.0","size_categories:1k<n<10k","region:us","israel","open-data","ckan","bilingual","government","catalogue"],"updated_at":"2026-05-06T20:40:36.000Z"},{"name":"Noisy-Voice-Notes","kind":"dataset","url":"https://huggingface.co/datasets/danielrosehill/Noisy-Voice-Notes","description":"Noisy Voice Notes A small, hand-curated set of real-world voice notes recorded over roughly a year by a single speaker (me — Daniel Rosehill), captured in everyday environments…","speech_or_language":true,"tags":["task_categories:automatic-speech-recognition","task_categories:audio-classification","task_categories:text-classification","language:en","language:he","license:cc-by-4.0","size_categories:n<1k","doi:10.57967/hf/8681","region:us","speech","voice-notes","noise","background-noise","asr-evaluation","dnsmos","personal-recordings"],"updated_at":"2026-05-06T11:14:19.000Z"},{"name":"acronym-identification-69f5d1d894ae439ce2033c89","kind":"collection","url":"https://huggingface.co/collections/danielrosehill/acronym-identification-69f5d1d894ae439ce2033c89","description":"Hugging Face collection: Acronym Identification","speech_or_language":false,"tags":[],"updated_at":"2026-05-02T10:28:40.316Z"},{"name":"disfluency-69f5d11e79b0a5aad7c972ae","kind":"collection","url":"https://huggingface.co/collections/danielrosehill/disfluency-69f5d11e79b0a5aad7c972ae","description":"Hugging Face collection: Disfluency","speech_or_language":false,"tags":[],"updated_at":"2026-05-02T10:26:35.226Z"},{"name":"Jerusalem-Air-Quality-Shabbat","kind":"dataset","url":"https://huggingface.co/datasets/danielrosehill/Jerusalem-Air-Quality-Shabbat","description":"Jerusalem Air Quality — Weekday vs Shabbat Twelve months of 5-minute-resolution air-quality readings from 12 monitoring stations across Jerusalem, formatted for analysing the…","speech_or_language":false,"tags":["language:en","language:he","license:cc-by-4.0","size_categories:1m<n<10m","format:parquet","modality:image","modality:tabular","modality:text","library:datasets","library:pandas","library:polars","library:mlcroissant","doi:10.57967/hf/8575","region:us","air-quality","environmental","jerusalem","israel","time-series","shabbat","sabbath","pm25","no2","environmental-monitoring"],"updated_at":"2026-04-28T19:26:51.000Z"},{"name":"Jerusalem-Air-Quality-Shabbat-Findings","kind":"space","url":"https://huggingface.co/spaces/danielrosehill/Jerusalem-Air-Quality-Shabbat-Findings","description":"","speech_or_language":false,"tags":["static","region:us"],"updated_at":"2026-04-28T15:13:09.000Z"},{"name":"Whiteboards","kind":"dataset","url":"https://huggingface.co/datasets/danielrosehill/Whiteboards","description":"A small, single-author whiteboard corpus for evaluating vision-language models on handwritten OCR accuracy and for studying pseudotext hallucination — the failure mode where a…","speech_or_language":false,"tags":["task_categories:image-to-text","task_categories:image-to-image","language:en","license:cc-by-4.0","size_categories:n<1k","format:json","modality:image","modality:text","library:datasets","library:pandas","library:polars","library:mlcroissant","region:us","whiteboard","handwriting","ocr","few-shot","vlm","evaluation","pseudotext"],"updated_at":"2026-04-23T11:00:43.000Z"},{"name":"Geopol-Forecast-Iran-Israel","kind":"space","url":"https://huggingface.co/spaces/danielrosehill/Geopol-Forecast-Iran-Israel","description":"","speech_or_language":false,"tags":["gradio","region:us"],"updated_at":"2026-04-18T21:34:16.000Z"},{"name":"Audio-Understanding-Bitrate-Eval-0426","kind":"dataset","url":"https://huggingface.co/datasets/danielrosehill/Audio-Understanding-Bitrate-Eval-0426","description":"Audio Understanding — MP3 Bitrate Evaluation (April 2026) Empirical eval measuring how MP3 compression bitrate affects transcription accuracy across every audio-input LLM…","speech_or_language":true,"tags":["task_categories:automatic-speech-recognition","language:en","license:mit","size_categories:n<1k","modality:text","region:us","speech","asr","bitrate","mp3","audio-llm","openrouter","gemini","voxtral","gpt-audio","evaluation","wer"],"updated_at":"2026-04-17T12:45:12.000Z"},{"name":"Geopol-Forecaster-Predictions","kind":"dataset","url":"https://huggingface.co/datasets/danielrosehill/Geopol-Forecaster-Predictions","description":"Geopol Forecaster Predictions Structured predictions extracted from multi-agent geopolitical wargaming simulations run by Geopol Forecaster. Dataset Structure File Description…","speech_or_language":false,"tags":["task_categories:text-generation","language:en","license:apache-2.0","size_categories:n<1k","doi:10.57967/hf/8373","region:us","geopolitics","forecasting","wargaming","predictions","llm-simulation"],"updated_at":"2026-04-13T10:46:39.000Z"},{"name":"Geopolitical-Predictions","kind":"space","url":"https://huggingface.co/spaces/danielrosehill/Geopolitical-Predictions","description":"","speech_or_language":false,"tags":["static","region:us"],"updated_at":"2026-04-10T11:52:54.000Z"},{"name":"llms-im-testing-69d150e80ad814926c3bb1fb","kind":"collection","url":"https://huggingface.co/collections/danielrosehill/llms-im-testing-69d150e80ad814926c3bb1fb","description":"Hugging Face collection: LLMS-Im-Testing","speech_or_language":false,"tags":[],"updated_at":"2026-04-04T17:56:56.091Z"},{"name":"hebrew-puncutation-restoration-69cbb9a691619429e248b328","kind":"collection","url":"https://huggingface.co/collections/danielrosehill/hebrew-puncutation-restoration-69cbb9a691619429e248b328","description":"Hugging Face collection: Hebrew Puncutation Restoration","speech_or_language":true,"tags":[],"updated_at":"2026-03-31T12:10:14.655Z"},{"name":"hebrew-sentiment-classification-models-69cbb98d71b4fe75732a830a","kind":"collection","url":"https://huggingface.co/collections/danielrosehill/hebrew-sentiment-classification-models-69cbb98d71b4fe75732a830a","description":"Hugging Face collection: Hebrew Sentiment Classification Models","speech_or_language":true,"tags":[],"updated_at":"2026-03-31T12:09:49.626Z"},{"name":"english-hebrew-translation-69cbb92b71b4fe75732a7c54","kind":"collection","url":"https://huggingface.co/collections/danielrosehill/english-hebrew-translation-69cbb92b71b4fe75732a7c54","description":"Hugging Face collection: English Hebrew Translation","speech_or_language":true,"tags":[],"updated_at":"2026-03-31T12:08:11.962Z"},{"name":"hebrew-ocr-models-69cbb885566a24bae2b833f4","kind":"collection","url":"https://huggingface.co/collections/danielrosehill/hebrew-ocr-models-69cbb885566a24bae2b833f4","description":"Hugging Face collection: Hebrew OCR Models","speech_or_language":true,"tags":[],"updated_at":"2026-03-31T12:05:25.150Z"},{"name":"hebrew-diacritic-restoration-models-69cbb8719ad62b783f9e5187","kind":"collection","url":"https://huggingface.co/collections/danielrosehill/hebrew-diacritic-restoration-models-69cbb8719ad62b783f9e5187","description":"Hugging Face collection: Hebrew Diacritic Restoration Models","speech_or_language":true,"tags":[],"updated_at":"2026-03-31T12:05:05.138Z"},{"name":"Single-Shot-ASR-Eval","kind":"space","url":"https://huggingface.co/spaces/danielrosehill/Single-Shot-ASR-Eval","description":"","speech_or_language":true,"tags":["static","region:us"],"updated_at":"2026-03-29T15:33:14.000Z"},{"name":"Audio-Understanding-Experiment","kind":"space","url":"https://huggingface.co/spaces/danielrosehill/Audio-Understanding-Experiment","description":"","speech_or_language":true,"tags":["static","region:us"],"updated_at":"2026-03-26T17:45:03.000Z"},{"name":"Audio-Understanding-Test-Set","kind":"dataset","url":"https://huggingface.co/datasets/danielrosehill/Audio-Understanding-Test-Set","description":"Audio Understanding Test Set A structured dataset for evaluating audio understanding capabilities of multimodal AI models. Contains 137 test prompts across 22 categories,…","speech_or_language":true,"tags":["task_categories:audio-classification","task_categories:automatic-speech-recognition","language:en","license:cc-by-4.0","size_categories:n<1k","modality:audio","doi:10.57967/hf/8154","region:us","audio-understanding","voice-analysis","multimodal-evaluation","speaker-analysis","emotion-detection","audio-engineering","voice-cloning","forensic-audio"],"updated_at":"2026-03-26T17:44:13.000Z"},{"name":"streaming-speech-to-text-69c4286de29f73f893d174e9","kind":"collection","url":"https://huggingface.co/collections/danielrosehill/streaming-speech-to-text-69c4286de29f73f893d174e9","description":"Hugging Face collection: Streaming-Speech-To-Text","speech_or_language":true,"tags":[],"updated_at":"2026-03-25T18:25:42.514Z"},{"name":"Hebrew-TTS-Samples","kind":"space","url":"https://huggingface.co/spaces/danielrosehill/Hebrew-TTS-Samples","description":"","speech_or_language":true,"tags":["static","region:us"],"updated_at":"2026-03-22T11:49:02.000Z"},{"name":"Iran-Israel-War-2026","kind":"dataset","url":"https://huggingface.co/datasets/danielrosehill/Iran-Israel-War-2026","description":"Iran-Israel War — OSINT Dataset Open-source intelligence dataset tracking Iranian missile and drone attack waves against Israel and US/coalition targets across four operations…","speech_or_language":false,"tags":["task_categories:tabular-classification","task_categories:tabular-regression","language:en","license:cc-by-4.0","size_categories:n<1k","region:us","osint","military","geopolitics","iran","israel","missiles","conflict-data","time-series"],"updated_at":"2026-03-18T17:06:33.000Z"},{"name":"deep-research-68f3e44dc5105b502b864b73","kind":"collection","url":"https://huggingface.co/collections/danielrosehill/deep-research-68f3e44dc5105b502b864b73","description":"Hugging Face collection: Deep research","speech_or_language":false,"tags":[],"updated_at":"2026-03-16T09:37:26.556Z"},{"name":"The-Agentic-Symphony","kind":"space","url":"https://huggingface.co/spaces/danielrosehill/The-Agentic-Symphony","description":"","speech_or_language":false,"tags":["static","region:us"],"updated_at":"2026-03-10T01:22:40.000Z"},{"name":"whisper-fine-tunes-691624c8befe1b3ca31fb319","kind":"collection","url":"https://huggingface.co/collections/danielrosehill/whisper-fine-tunes-691624c8befe1b3ca31fb319","description":"Whisper fine-tuned on 1 hour of my voice","speech_or_language":true,"tags":[],"updated_at":"2026-03-02T10:56:01.637Z"},{"name":"architecture-related-models-6908f545dd9c200e81724a0e","kind":"collection","url":"https://huggingface.co/collections/danielrosehill/architecture-related-models-6908f545dd9c200e81724a0e","description":"Collection of models gathered together for my wife who is an architect (of buildings!)","speech_or_language":false,"tags":[],"updated_at":"2026-03-02T10:56:01.017Z"},{"name":"voice-enhancement-68faa39c1ebef0f22b7faaad","kind":"collection","url":"https://huggingface.co/collections/danielrosehill/voice-enhancement-68faa39c1ebef0f22b7faaad","description":"Hugging Face collection: Voice Enhancement","speech_or_language":true,"tags":[],"updated_at":"2026-03-02T10:56:00.545Z"},{"name":"audio-editing-utilities-68f3f55204fbae5be954546e","kind":"collection","url":"https://huggingface.co/collections/danielrosehill/audio-editing-utilities-68f3f55204fbae5be954546e","description":"Hugging Face collection: Audio editing utilities","speech_or_language":true,"tags":[],"updated_at":"2026-03-02T10:56:00.258Z"},{"name":"Ivory-Catalog-020226","kind":"dataset","url":"https://huggingface.co/datasets/danielrosehill/Ivory-Catalog-020226","description":"Ivory Computer Parts Catalog - February 2026 A snapshot of computer hardware pricing from Ivory.co.il, one of Israel's largest electronics retailers, captured on February 2,…","speech_or_language":false,"tags":["task_categories:tabular-classification","task_categories:tabular-regression","language:en","language:he","license:cc-by-4.0","size_categories:n<1k","format:json","modality:tabular","modality:text","library:datasets","library:pandas","library:polars","library:mlcroissant","region:us","computer-hardware","pricing","israel","e-commerce","pc-components"],"updated_at":"2026-02-02T14:39:42.000Z"},{"name":"agentic-code-gen-301225-6954035aaf20492708c6d066","kind":"collection","url":"https://huggingface.co/collections/danielrosehill/agentic-code-gen-301225-6954035aaf20492708c6d066","description":"Quick collection of models I'm evaluating for cogeneration. Minimum inclusion criteria includes tool usage and MCP. Looking for something fast and goo","speech_or_language":false,"tags":[],"updated_at":"2025-12-30T18:27:28.417Z"},{"name":"mwp-tts-candidates-69500180785ac06cb1e4cc2a","kind":"collection","url":"https://huggingface.co/collections/danielrosehill/mwp-tts-candidates-69500180785ac06cb1e4cc2a","description":"Hugging Face collection: MWP-TTS-Candidates","speech_or_language":true,"tags":[],"updated_at":"2025-12-27T15:55:44.986Z"},{"name":"Transcription-Cleanup-Trainer","kind":"dataset","url":"https://huggingface.co/datasets/danielrosehill/Transcription-Cleanup-Trainer","description":"Text Cleanup Fine-Tuning Dataset A curated dataset for training speech-to-text cleanup models to achieve optimal transcript refinement. Dataset Description This dataset…","speech_or_language":true,"tags":["size_categories:n<1k","format:text","modality:audio","modality:text","library:datasets","library:mlcroissant","region:us"],"updated_at":"2025-12-18T14:13:19.000Z"},{"name":"video-understanding-6940b0be190e625adfa5f462","kind":"collection","url":"https://huggingface.co/collections/danielrosehill/video-understanding-6940b0be190e625adfa5f462","description":"Hugging Face collection: Video Understanding","speech_or_language":false,"tags":[],"updated_at":"2025-12-16T01:07:10.165Z"},{"name":"asr-to-try-693f64f055b4855423775995","kind":"collection","url":"https://huggingface.co/collections/danielrosehill/asr-to-try-693f64f055b4855423775995","description":"Hugging Face collection: ASR-To-Try","speech_or_language":true,"tags":[],"updated_at":"2025-12-15T14:03:08.567Z"},{"name":"image-evaluations-693b5a64535d7de3543ac987","kind":"collection","url":"https://huggingface.co/collections/danielrosehill/image-evaluations-693b5a64535d7de3543ac987","description":"Hugging Face collection: Image Evaluations","speech_or_language":false,"tags":[],"updated_at":"2025-12-11T23:57:24.823Z"},{"name":"Hebrew-Image-Eval-111225","kind":"dataset","url":"https://huggingface.co/datasets/danielrosehill/Hebrew-Image-Eval-111225","description":"Hebrew Image Generation Evaluation An evaluation of major text-to-image models on their ability to accurately render Hebrew text. Overview This evaluation tests 12 image…","speech_or_language":true,"tags":["task_categories:text-to-image","language:he","license:mit","size_categories:n<1k","region:us","evaluation","hebrew","typography","text-rendering","benchmark"],"updated_at":"2025-12-11T23:57:03.000Z"},{"name":"STT-Fine-Tune-Eval-101225","kind":"dataset","url":"https://huggingface.co/datasets/danielrosehill/STT-Fine-Tune-Eval-101225","description":"STT Fine-Tune Evaluation Results Evaluation Date: December 10, 2025 This repository contains evaluation results comparing locally fine-tuned Whisper models against stock…","speech_or_language":true,"tags":["region:us"],"updated_at":"2025-12-10T19:42:43.000Z"},{"name":"Small-STT-Eval-Audio-Dataset","kind":"dataset","url":"https://huggingface.co/datasets/danielrosehill/Small-STT-Eval-Audio-Dataset","description":"Small STT Eval Audio Dataset A small speech-to-text evaluation dataset containing 92 audio samples with ground truth transcriptions. Designed for evaluating STT systems on…","speech_or_language":true,"tags":["task_categories:automatic-speech-recognition","language:en","language:he","license:cc-by-4.0","size_categories:n<1k","format:audiofolder","modality:audio","modality:text","library:datasets","library:mlcroissant","region:us","speech-to-text","stt","evaluation","technical-vocabulary"],"updated_at":"2025-12-10T18:31:12.000Z"},{"name":"ASR-WPM-And-Background-Noise-Eval","kind":"dataset","url":"https://huggingface.co/datasets/danielrosehill/ASR-WPM-And-Background-Noise-Eval","description":"ASR WPM and Background Noise Evaluation Dataset A dataset of annotated audio recordings for evaluating how different factors affect Whisper (and other ASR/STT systems)…","speech_or_language":true,"tags":["task_categories:automatic-speech-recognition","language:en","license:mit","size_categories:n<1k","modality:audio","doi:10.57967/hf/7192","region:us","audio","speech","whisper","asr","stt","wpm","background-noise","speech-recognition","evaluation"],"updated_at":"2025-12-09T18:19:17.000Z"},{"name":"audio-multimodal-models-6935a9e135d60ed9157a198e","kind":"collection","url":"https://huggingface.co/collections/danielrosehill/audio-multimodal-models-6935a9e135d60ed9157a198e","description":"Open source models with audio understanding. Tracking mostly vendor releases in the audio and text to text subclassification of multimodal.","speech_or_language":true,"tags":[],"updated_at":"2025-12-08T14:13:25.085Z"},{"name":"audio-understanding-datasets-6936d2a1a977953ef759b161","kind":"collection","url":"https://huggingface.co/collections/danielrosehill/audio-understanding-datasets-6936d2a1a977953ef759b161","description":"Hugging Face collection: Audio Understanding Datasets","speech_or_language":true,"tags":[],"updated_at":"2025-12-08T13:35:08.385Z"},{"name":"voxtral-originals-mistral-6936a6475afe39a2be4ae8b7","kind":"collection","url":"https://huggingface.co/collections/danielrosehill/voxtral-originals-mistral-6936a6475afe39a2be4ae8b7","description":"The two official variants of Voxtral (audio multimodal model) released by Mistral in July 2025","speech_or_language":true,"tags":[],"updated_at":"2025-12-08T10:23:02.539Z"},{"name":"tts-models-6920719b04ade4111ec9f7a0","kind":"collection","url":"https://huggingface.co/collections/danielrosehill/tts-models-6920719b04ade4111ec9f7a0","description":"Hugging Face collection: TTS Models","speech_or_language":true,"tags":[],"updated_at":"2025-12-07T13:58:15.945Z"},{"name":"asr-models-67fbe5c405cd5b5035e5fddb","kind":"collection","url":"https://huggingface.co/collections/danielrosehill/asr-models-67fbe5c405cd5b5035e5fddb","description":"Collection of ASR models for customized STT model training","speech_or_language":true,"tags":[],"updated_at":"2025-12-03T04:00:41.867Z"},{"name":"stt-components-690be5a864ad0e9988f1ed22","kind":"collection","url":"https://huggingface.co/collections/danielrosehill/stt-components-690be5a864ad0e9988f1ed22","description":"Models that work in unison with core STT models in voice workflows","speech_or_language":true,"tags":[],"updated_at":"2025-12-03T03:58:56.614Z"},{"name":"flux-2-quants-692fa4925e5d56233e9f2b3d","kind":"collection","url":"https://huggingface.co/collections/danielrosehill/flux-2-quants-692fa4925e5d56233e9f2b3d","description":"Hugging Face collection: Flux 2 Quants","speech_or_language":false,"tags":[],"updated_at":"2025-12-03T02:46:47.644Z"},{"name":"Israel-Photos","kind":"dataset","url":"https://huggingface.co/datasets/danielrosehill/Israel-Photos","description":"Israel Photos Dataset A collection of 369 photographs captured across Israel between 2024 and 2025, with LLM-generated captions and location annotations. The images are sourced…","speech_or_language":false,"tags":["task_categories:image-to-text","task_categories:image-classification","license:cc-by-4.0","size_categories:n<1k","format:imagefolder","modality:image","library:datasets","library:mlcroissant","doi:10.57967/hf/7152","region:us","photography","israel","image-captioning","location-recognition"],"updated_at":"2025-12-03T00:43:51.000Z"},{"name":"my-image-datasets-68f3cec6f1274f99c2b93cee","kind":"collection","url":"https://huggingface.co/collections/danielrosehill/my-image-datasets-68f3cec6f1274f99c2b93cee","description":"Hugging Face collection: My Image Datasets","speech_or_language":false,"tags":[],"updated_at":"2025-12-03T00:42:12.429Z"},{"name":"Sample-Voice-Context-Data","kind":"dataset","url":"https://huggingface.co/datasets/danielrosehill/Sample-Voice-Context-Data","description":"Sample Voice Context Data A small synthetic dataset containing LLM-generated context information simulating a job seeker narrating their career trajectory. Purpose This dataset…","speech_or_language":true,"tags":["license:mit","size_categories:n<1k","format:text","modality:audio","modality:text","library:datasets","library:mlcroissant","region:us"],"updated_at":"2025-11-30T13:24:42.000Z"},{"name":"Tech-Sentences-For-ASR-Training","kind":"dataset","url":"https://huggingface.co/datasets/danielrosehill/Tech-Sentences-For-ASR-Training","description":"TechVoice Dataset Work in Progress – This dataset is actively being expanded with new recordings. Dataset Statistics Metric Current Target Progress Duration 38m 43s 5h 0m 0s…","speech_or_language":true,"tags":["task_categories:automatic-speech-recognition","language:en","license:mit","size_categories:n<1k","format:json","modality:audio","modality:tabular","modality:text","library:datasets","library:pandas","library:mlcroissant","library:polars","doi:10.57967/hf/7099","region:us","speech","whisper","asr","fine-tuning","technical-vocabulary","developer-speech","software-engineering","programming"],"updated_at":"2025-11-26T20:14:02.000Z"},{"name":"asr-resources-691e00763a159e42d9c799f0","kind":"collection","url":"https://huggingface.co/collections/danielrosehill/asr-resources-691e00763a159e42d9c799f0","description":"Useful datasets and models for ASR projects and fine tuning","speech_or_language":true,"tags":[],"updated_at":"2025-11-26T14:14:38.895Z"},{"name":"asr-beyond-whisper-6920377f887e25783c4ef2a6","kind":"collection","url":"https://huggingface.co/collections/danielrosehill/asr-beyond-whisper-6920377f887e25783c4ef2a6","description":"Hugging Face collection: ASR Beyond Whisper","speech_or_language":true,"tags":[],"updated_at":"2025-11-21T10:00:03.269Z"},{"name":"whisper-hebrish-691bc2624553319eb727c1db","kind":"collection","url":"https://huggingface.co/collections/danielrosehill/whisper-hebrish-691bc2624553319eb727c1db","description":"Fine tune of Whisper Large (V3, Turbo) using a small corpus of mixed language English sentences with Hebrew to improve accuracy","speech_or_language":true,"tags":[],"updated_at":"2025-11-18T00:50:43.684Z"},{"name":"my-asr-finetunes-691bb00e473d4d4997fd825a","kind":"collection","url":"https://huggingface.co/collections/danielrosehill/my-asr-finetunes-691bb00e473d4d4997fd825a","description":"Hugging Face collection: My-ASR-Finetunes","speech_or_language":true,"tags":[],"updated_at":"2025-11-18T00:42:47.423Z"},{"name":"Whisper-Hebrish","kind":"space","url":"https://huggingface.co/spaces/danielrosehill/Whisper-Hebrish","description":"","speech_or_language":true,"tags":["gradio","region:us"],"updated_at":"2025-11-18T00:36:15.000Z"},{"name":"old-llms-691bab3f1eec548928cb6aad","kind":"collection","url":"https://huggingface.co/collections/danielrosehill/old-llms-691bab3f1eec548928cb6aad","description":"Useful models for demonstrating what early and pre-Transformer LLMs looked and functioned like","speech_or_language":false,"tags":[],"updated_at":"2025-11-17T23:10:07.975Z"},{"name":"Whisper-Hebrish","kind":"model","url":"https://huggingface.co/danielrosehill/Whisper-Hebrish","description":"","speech_or_language":true,"tags":["safetensors","whisper","code-switching","hebrew","english","automatic-speech-recognition","en","dataset:danielrosehill/read-aloud-voice-sentences","base_model:openai/whisper-large-v3-turbo","base_model:finetune:openai/whisper-large-v3-turbo","doi:10.57967/hf/7019","license:mit","region:us"],"updated_at":"2025-11-17T22:59:37.000Z"},{"name":"stt-evaluations-690416fd75e779077a0895ff","kind":"collection","url":"https://huggingface.co/collections/danielrosehill/stt-evaluations-690416fd75e779077a0895ff","description":"Hugging Face collection: STT Evaluations","speech_or_language":true,"tags":[],"updated_at":"2025-11-17T21:42:24.166Z"},{"name":"my-evaluations-691671880beb8c148c4e83c1","kind":"collection","url":"https://huggingface.co/collections/danielrosehill/my-evaluations-691671880beb8c148c4e83c1","description":"Hugging Face collection: My-Evaluations","speech_or_language":false,"tags":[],"updated_at":"2025-11-17T21:42:16.011Z"},{"name":"Whisper-Fine-Tune-One-Shot-Eval","kind":"dataset","url":"https://huggingface.co/datasets/danielrosehill/Whisper-Fine-Tune-One-Shot-Eval","description":"Whisper Fine-Tuning Evaluation: Local vs Commercial ASR A \"back of the envelope\" evaluation comparing fine-tuned Whisper models running locally against commercial ASR APIs via…","speech_or_language":true,"tags":["size_categories:n<1k","format:audiofolder","modality:audio","modality:text","library:datasets","library:mlcroissant","region:us"],"updated_at":"2025-11-17T21:36:54.000Z"},{"name":"Whisper-Fine-Tune-Eval","kind":"space","url":"https://huggingface.co/spaces/danielrosehill/Whisper-Fine-Tune-Eval","description":"","speech_or_language":true,"tags":["static","whisper","speech-to-text","evaluation","benchmark","api-comparison","region:us"],"updated_at":"2025-11-17T20:21:35.000Z"},{"name":"futo-models-691b28ee19c6c6529789ffbf","kind":"collection","url":"https://huggingface.co/collections/danielrosehill/futo-models-691b28ee19c6c6529789ffbf","description":"Hugging Face collection: FUTO Models","speech_or_language":false,"tags":[],"updated_at":"2025-11-17T13:54:17.347Z"},{"name":"STT-Comparison","kind":"space","url":"https://huggingface.co/spaces/danielrosehill/STT-Comparison","description":"","speech_or_language":true,"tags":["static","region:us"],"updated_at":"2025-11-14T00:18:06.000Z"},{"name":"asr-benchmarking-69165f9fbefe1b3ca322c339","kind":"collection","url":"https://huggingface.co/collections/danielrosehill/asr-benchmarking-69165f9fbefe1b3ca322c339","description":"Hugging Face collection: ASR Benchmarking","speech_or_language":true,"tags":[],"updated_at":"2025-11-14T00:01:58.880Z"},{"name":"Podcast-ASR-Evaluation","kind":"dataset","url":"https://huggingface.co/datasets/danielrosehill/Podcast-ASR-Evaluation","description":"","speech_or_language":true,"tags":["size_categories:n<1k","format:text","modality:text","library:datasets","library:mlcroissant","region:us"],"updated_at":"2025-11-14T00:01:33.000Z"},{"name":"Single-Podcast-ASR-Eval","kind":"space","url":"https://huggingface.co/spaces/danielrosehill/Single-Podcast-ASR-Eval","description":"","speech_or_language":true,"tags":["gradio","speech-recognition","asr","benchmark","whisper","podcast","region:us"],"updated_at":"2025-11-13T23:56:47.000Z"},{"name":"Open-Router-API-Pricing-Analysis","kind":"dataset","url":"https://huggingface.co/datasets/danielrosehill/Open-Router-API-Pricing-Analysis","description":"OpenRouter API Pricing Analysis Dataset Overview This dataset provides a point-in-time capture of pricing and parameters for LLMs available through the OpenRouter API for…","speech_or_language":false,"tags":["task_categories:text-generation","language:en","license:mit","size_categories:1k<n<10k","format:text","modality:text","library:datasets","library:mlcroissant","region:us","llm","pricing","openrouter","api-pricing","cost-analysis"],"updated_at":"2025-11-10T18:58:42.000Z"},{"name":"api-price-comparisons-690fbad1a7aab0cbaffa6e7a","kind":"collection","url":"https://huggingface.co/collections/danielrosehill/api-price-comparisons-690fbad1a7aab0cbaffa6e7a","description":"Hugging Face collection: API Price Comparisons","speech_or_language":false,"tags":[],"updated_at":"2025-11-10T18:23:35.496Z"},{"name":"Agentic-LLM-Price-Comparisons","kind":"space","url":"https://huggingface.co/spaces/danielrosehill/Agentic-LLM-Price-Comparisons","description":"","speech_or_language":false,"tags":["static","region:us"],"updated_at":"2025-11-08T21:26:02.000Z"},{"name":"pdf-downloads-690ec77f0570b41ffe123662","kind":"collection","url":"https://huggingface.co/collections/danielrosehill/pdf-downloads-690ec77f0570b41ffe123662","description":"Hugging Face collection: PDF Downloads","speech_or_language":false,"tags":[],"updated_at":"2025-11-08T04:30:55.166Z"},{"name":"my-loras-690c805220a832267659e4fb","kind":"collection","url":"https://huggingface.co/collections/danielrosehill/my-loras-690c805220a832267659e4fb","description":"Hugging Face collection: My LORAS","speech_or_language":false,"tags":[],"updated_at":"2025-11-06T12:30:40.503Z"},{"name":"Herman-Poppleberry","kind":"model","url":"https://huggingface.co/danielrosehill/Herman-Poppleberry","description":"","speech_or_language":false,"tags":["diffusers","text-to-image","lora","template:diffusion-lora","base_model:black-forest-labs/flux.1-dev","base_model:adapter:black-forest-labs/flux.1-dev","doi:10.57967/hf/11082","region:us"],"updated_at":"2025-11-06T12:29:58.000Z"},{"name":"Dinosaur-Sloth","kind":"model","url":"https://huggingface.co/danielrosehill/Dinosaur-Sloth","description":"","speech_or_language":false,"tags":["diffusers","text-to-image","lora","template:diffusion-lora","base_model:black-forest-labs/flux.1-dev","base_model:adapter:black-forest-labs/flux.1-dev","doi:10.57967/hf/11081","license:mit","region:us"],"updated_at":"2025-11-06T12:11:39.000Z"},{"name":"Tel-Aviv-Street-Style","kind":"model","url":"https://huggingface.co/danielrosehill/Tel-Aviv-Street-Style","description":"","speech_or_language":false,"tags":["diffusers","text-to-image","lora","template:diffusion-lora","base_model:black-forest-labs/flux.1-dev","base_model:adapter:black-forest-labs/flux.1-dev","doi:10.57967/hf/11083","region:us"],"updated_at":"2025-11-06T11:18:52.000Z"},{"name":"Jerusalem-Images","kind":"model","url":"https://huggingface.co/danielrosehill/Jerusalem-Images","description":"","speech_or_language":false,"tags":["diffusers","text-to-image","lora","template:diffusion-lora","en","base_model:black-forest-labs/flux.1-dev","base_model:adapter:black-forest-labs/flux.1-dev","license:mit","region:us"],"updated_at":"2025-11-06T11:02:05.000Z"},{"name":"Tel-Aviv-Pics","kind":"dataset","url":"https://huggingface.co/datasets/danielrosehill/Tel-Aviv-Pics","description":"Tel Aviv Urban Photography Dataset Dataset Description This dataset contains 53 high-quality photographs of Tel Aviv's urban environment, captured to serve as reference…","speech_or_language":false,"tags":["task_categories:image-to-image","task_categories:image-to-3d","task_categories:text-to-image","license:cc-by-4.0","size_categories:n<1k","doi:10.57967/hf/6743","region:us","urban-environments","game-development","3d-modeling","world-building","city-scenes","tel-aviv","middle-east","architecture","photogrammetry","reference-images"],"updated_at":"2025-11-06T10:32:57.000Z"},{"name":"Jerusalem-Streetscapes","kind":"dataset","url":"https://huggingface.co/datasets/danielrosehill/Jerusalem-Streetscapes","description":"Jerusalem Streetscapes Dataset A small image dataset containing photos of the rapidly changing urban landscape of Jerusalem, Israel, captured by day and by night. About This…","speech_or_language":false,"tags":["size_categories:n<1k","format:imagefolder","modality:image","library:datasets","library:mlcroissant","doi:10.57967/hf/6756","region:us"],"updated_at":"2025-11-06T10:10:55.000Z"},{"name":"Hebrew-Language-Signage","kind":"dataset","url":"https://huggingface.co/datasets/danielrosehill/Hebrew-Language-Signage","description":"Hebrew Language Signage Dataset Overview This dataset contains photographs of Hebrew language text in everyday contexts throughout Israel, with a particular focus on signage…","speech_or_language":true,"tags":["size_categories:n<1k","format:imagefolder","modality:image","library:datasets","library:mlcroissant","doi:10.57967/hf/6755","region:us"],"updated_at":"2025-11-06T09:51:23.000Z"},{"name":"Jerusalem-High-Rise-Development","kind":"dataset","url":"https://huggingface.co/datasets/danielrosehill/Jerusalem-High-Rise-Development","description":"Jerusalem High-Rise Development Image Dataset Overview This dataset contains 56 photographs documenting high-rise buildings and urban development in Jerusalem, Israel. The…","speech_or_language":false,"tags":["size_categories:n<1k","format:imagefolder","modality:image","library:datasets","library:mlcroissant","doi:10.57967/hf/6754","region:us"],"updated_at":"2025-11-06T09:46:12.000Z"},{"name":"video-background-removal-690925061a841594aad2d44b","kind":"collection","url":"https://huggingface.co/collections/danielrosehill/video-background-removal-690925061a841594aad2d44b","description":"Hugging Face collection: Video background removal","speech_or_language":false,"tags":[],"updated_at":"2025-11-03T21:56:22.331Z"},{"name":"stt-fine-tune-resources-6908f18dd3f58e58415622cc","kind":"collection","url":"https://huggingface.co/collections/danielrosehill/stt-fine-tune-resources-6908f18dd3f58e58415622cc","description":"Hugging Face collection: STT Fine Tune Resources","speech_or_language":true,"tags":[],"updated_at":"2025-11-03T18:17:56.747Z"},{"name":"Nano-Banana-Sketch-Cleanup","kind":"space","url":"https://huggingface.co/spaces/danielrosehill/Nano-Banana-Sketch-Cleanup","description":"","speech_or_language":false,"tags":["static","region:us"],"updated_at":"2025-11-02T00:37:52.000Z"},{"name":"image-to-3d-690548ee6f0884de5ccc050f","kind":"collection","url":"https://huggingface.co/collections/danielrosehill/image-to-3d-690548ee6f0884de5ccc050f","description":"Hugging Face collection: Image to 3D","speech_or_language":false,"tags":[],"updated_at":"2025-10-31T23:40:30.284Z"},{"name":"3d-human-digital-humans-68f25352db2f94819b6b104a","kind":"collection","url":"https://huggingface.co/collections/danielrosehill/3d-human-digital-humans-68f25352db2f94819b6b104a","description":"Hugging Face collection: 3D Human Digital Humans","speech_or_language":false,"tags":[],"updated_at":"2025-10-31T19:47:09.578Z"},{"name":"concept-outlines-69041c68067797f759dae445","kind":"collection","url":"https://huggingface.co/collections/danielrosehill/concept-outlines-69041c68067797f759dae445","description":"Hugging Face collection: Concept Outlines","speech_or_language":false,"tags":[],"updated_at":"2025-10-31T02:18:16.788Z"},{"name":"my-ideas-69041c5ae0ff276b738bce11","kind":"collection","url":"https://huggingface.co/collections/danielrosehill/my-ideas-69041c5ae0ff276b738bce11","description":"Hugging Face collection: My Ideas","speech_or_language":false,"tags":[],"updated_at":"2025-10-31T02:18:02.176Z"},{"name":"Claude-Agent-Picker-Idea","kind":"space","url":"https://huggingface.co/spaces/danielrosehill/Claude-Agent-Picker-Idea","description":"","speech_or_language":false,"tags":["static","region:us"],"updated_at":"2025-10-31T02:03:32.000Z"},{"name":"ai-experiments-68b82538cf2a5e13d4e7d5fc","kind":"collection","url":"https://huggingface.co/collections/danielrosehill/ai-experiments-68b82538cf2a5e13d4e7d5fc","description":"Experiments in AI and prompt engineering","speech_or_language":false,"tags":[],"updated_at":"2025-10-31T01:55:18.519Z"},{"name":"Local-STT-Eval-One-Sample","kind":"space","url":"https://huggingface.co/spaces/danielrosehill/Local-STT-Eval-One-Sample","description":"","speech_or_language":true,"tags":["gradio","region:us"],"updated_at":"2025-10-31T01:11:43.000Z"},{"name":"whisper-base-variants-6904034cf1649245a51c4574","kind":"collection","url":"https://huggingface.co/collections/danielrosehill/whisper-base-variants-6904034cf1649245a51c4574","description":"Hugging Face collection: Whisper Base + variants","speech_or_language":true,"tags":[],"updated_at":"2025-10-31T00:31:37.930Z"},{"name":"whisper-variants-6904032be5309655d0346a4f","kind":"collection","url":"https://huggingface.co/collections/danielrosehill/whisper-variants-6904032be5309655d0346a4f","description":"Hugging Face collection: Whisper variants","speech_or_language":true,"tags":[],"updated_at":"2025-10-31T00:30:35.078Z"},{"name":"voice-modality-apps-69040260ae4ec7f8a9dab9ae","kind":"collection","url":"https://huggingface.co/collections/danielrosehill/voice-modality-apps-69040260ae4ec7f8a9dab9ae","description":"Hugging Face collection: Voice Modality Apps","speech_or_language":true,"tags":[],"updated_at":"2025-10-31T00:27:12.641Z"},{"name":"entertainment-recommendations-6904020da193ab8ec19ec0f2","kind":"collection","url":"https://huggingface.co/collections/danielrosehill/entertainment-recommendations-6904020da193ab8ec19ec0f2","description":"Hugging Face collection: Entertainment Recommendations","speech_or_language":false,"tags":[],"updated_at":"2025-10-31T00:25:49.740Z"},{"name":"worlds-3d-games-6901edea30a042c01fcee1a4","kind":"collection","url":"https://huggingface.co/collections/danielrosehill/worlds-3d-games-6901edea30a042c01fcee1a4","description":"Hugging Face collection: Worlds (3D, Games)","speech_or_language":false,"tags":[],"updated_at":"2025-10-29T10:35:22.407Z"},{"name":"background-noise-removal-690127a6e9f617a5ce876c6c","kind":"collection","url":"https://huggingface.co/collections/danielrosehill/background-noise-removal-690127a6e9f617a5ce876c6c","description":"Hugging Face collection: Background Noise Removal","speech_or_language":false,"tags":[],"updated_at":"2025-10-28T21:09:54.028Z"},{"name":"proofs-of-concept-690131070acb2c6a9e0d9691","kind":"collection","url":"https://huggingface.co/collections/danielrosehill/proofs-of-concept-690131070acb2c6a9e0d9691","description":"Hugging Face collection: Proofs of Concept","speech_or_language":false,"tags":[],"updated_at":"2025-10-28T21:09:27.692Z"},{"name":"Deepnet-Baby-Noise-Scrub","kind":"space","url":"https://huggingface.co/spaces/danielrosehill/Deepnet-Baby-Noise-Scrub","description":"","speech_or_language":false,"tags":["gradio","region:us"],"updated_at":"2025-10-28T20:57:24.000Z"},{"name":"deep-filter-69012850c26349acbe8e0a22","kind":"collection","url":"https://huggingface.co/collections/danielrosehill/deep-filter-69012850c26349acbe8e0a22","description":"Hugging Face collection: Deep Filter","speech_or_language":false,"tags":[],"updated_at":"2025-10-28T20:32:16.406Z"},{"name":"Single-Shot-Brevity-Training","kind":"space","url":"https://huggingface.co/spaces/danielrosehill/Single-Shot-Brevity-Training","description":"","speech_or_language":false,"tags":["static","region:us"],"updated_at":"2025-10-27T15:23:27.000Z"},{"name":"Voice-Note-Audio","kind":"dataset","url":"https://huggingface.co/datasets/danielrosehill/Voice-Note-Audio","description":"Voice Notes Dataset Dataset Description This dataset contains real-world voice recordings with transcripts and comprehensive annotations. Dataset Statistics Total Entries: 2…","speech_or_language":true,"tags":["modality:audio","modality:text","doi:10.57967/hf/6316","region:us"],"updated_at":"2025-10-27T01:49:54.000Z"},{"name":"multimodal-ai-68f91baf011b757d2e5490a7","kind":"collection","url":"https://huggingface.co/collections/danielrosehill/multimodal-ai-68f91baf011b757d2e5490a7","description":"Hugging Face collection: Multimodal AI","speech_or_language":false,"tags":[],"updated_at":"2025-10-26T22:51:09.297Z"},{"name":"gemini-68fea5d2e96214e7780557de","kind":"collection","url":"https://huggingface.co/collections/danielrosehill/gemini-68fea5d2e96214e7780557de","description":"Hugging Face collection: Gemini","speech_or_language":false,"tags":[],"updated_at":"2025-10-26T22:50:58.992Z"},{"name":"vibe-coding-68fea5c42eda1bcb01fe73e9","kind":"collection","url":"https://huggingface.co/collections/danielrosehill/vibe-coding-68fea5c42eda1bcb01fe73e9","description":"Hugging Face collection: Vibe Coding","speech_or_language":false,"tags":[],"updated_at":"2025-10-26T22:50:44.865Z"},{"name":"claude-code-68fea30724789245bb47aefa","kind":"collection","url":"https://huggingface.co/collections/danielrosehill/claude-code-68fea30724789245bb47aefa","description":"Hugging Face collection: Claude Code","speech_or_language":false,"tags":[],"updated_at":"2025-10-26T22:39:18.733Z"},{"name":"Claude-Code-Slash-Commands","kind":"space","url":"https://huggingface.co/spaces/danielrosehill/Claude-Code-Slash-Commands","description":"","speech_or_language":false,"tags":["gradio","region:us"],"updated_at":"2025-10-26T22:37:57.000Z"},{"name":"Claude-Code-Slash-Commands-Linux-Desktop","kind":"space","url":"https://huggingface.co/spaces/danielrosehill/Claude-Code-Slash-Commands-Linux-Desktop","description":"","speech_or_language":false,"tags":["static","region:us"],"updated_at":"2025-10-26T22:31:43.000Z"},{"name":"project-indexes-68fe970d19755dc35b811358","kind":"collection","url":"https://huggingface.co/collections/danielrosehill/project-indexes-68fe970d19755dc35b811358","description":"Hugging Face collection: Project Indexes","speech_or_language":false,"tags":[],"updated_at":"2025-10-26T22:09:20.645Z"},{"name":"Gemini-Vibe-Coded-POCs","kind":"space","url":"https://huggingface.co/spaces/danielrosehill/Gemini-Vibe-Coded-POCs","description":"","speech_or_language":false,"tags":["static","region:us"],"updated_at":"2025-10-26T21:49:47.000Z"},{"name":"EQ-Template-Generator","kind":"space","url":"https://huggingface.co/spaces/danielrosehill/EQ-Template-Generator","description":"","speech_or_language":false,"tags":["static","region:us"],"updated_at":"2025-10-26T21:37:37.000Z"},{"name":"AI-Project-Index","kind":"space","url":"https://huggingface.co/spaces/danielrosehill/AI-Project-Index","description":"","speech_or_language":false,"tags":["static","region:us"],"updated_at":"2025-10-26T21:31:18.000Z"},{"name":"context-utilities-68f8e657acb39d74027950b9","kind":"collection","url":"https://huggingface.co/collections/danielrosehill/context-utilities-68f8e657acb39d74027950b9","description":"Experimentary workflows (mostly) that aim to rethink how contextual data can be gathered to achieve personalised inference in AI systems","speech_or_language":false,"tags":[],"updated_at":"2025-10-26T19:52:05.061Z"},{"name":"Context-Cruncher","kind":"space","url":"https://huggingface.co/spaces/danielrosehill/Context-Cruncher","description":"","speech_or_language":false,"tags":["gradio","region:us"],"updated_at":"2025-10-26T19:45:03.000Z"},{"name":"Interesting-Whisper-Projects","kind":"space","url":"https://huggingface.co/spaces/danielrosehill/Interesting-Whisper-Projects","description":"","speech_or_language":true,"tags":["static","region:us"],"updated_at":"2025-10-26T15:50:51.000Z"},{"name":"shakespeare-ai-68fe19b8f8d2d60eb736b904","kind":"collection","url":"https://huggingface.co/collections/danielrosehill/shakespeare-ai-68fe19b8f8d2d60eb736b904","description":"Just for fun projects for converting conventional text into Shakespearean English","speech_or_language":false,"tags":[],"updated_at":"2025-10-26T12:54:10.982Z"},{"name":"ai-uis-68fb7440da39f1c417293b73","kind":"collection","url":"https://huggingface.co/collections/danielrosehill/ai-uis-68fb7440da39f1c417293b73","description":"Hugging Face collection: AI UIs","speech_or_language":false,"tags":[],"updated_at":"2025-10-24T12:42:40.941Z"},{"name":"Penpal-AI-Interface","kind":"space","url":"https://huggingface.co/spaces/danielrosehill/Penpal-AI-Interface","description":"","speech_or_language":false,"tags":["gradio","region:us"],"updated_at":"2025-10-24T12:33:41.000Z"},{"name":"real-time-video-to-video-68faacba718bf18d053b17aa","kind":"collection","url":"https://huggingface.co/collections/danielrosehill/real-time-video-to-video-68faacba718bf18d053b17aa","description":"Hugging Face collection: Real Time Video To Video","speech_or_language":false,"tags":[],"updated_at":"2025-10-23T22:31:35.519Z"},{"name":"Human-To-AI-Text-Converter","kind":"space","url":"https://huggingface.co/spaces/danielrosehill/Human-To-AI-Text-Converter","description":"","speech_or_language":false,"tags":["gradio","region:us"],"updated_at":"2025-10-23T15:37:49.000Z"},{"name":"multimodal-ai-taxonomy","kind":"dataset","url":"https://huggingface.co/datasets/danielrosehill/multimodal-ai-taxonomy","description":"Multimodal AI Taxonomy A comprehensive, structured taxonomy for mapping multimodal AI model capabilities across input and output modalities. Dataset Description This dataset…","speech_or_language":true,"tags":["task_categories:other","language:en","license:cc0-1.0","size_categories:n<1k","format:json","modality:text","modality:audio","library:datasets","library:pandas","library:mlcroissant","library:polars","region:us","multimodal","taxonomy","ai-models","modality-mapping","computer-vision","audio","video-generation","image-generation"],"updated_at":"2025-10-22T18:33:21.000Z"},{"name":"taxonomies-68f91ba2e230b250cc8beda5","kind":"collection","url":"https://huggingface.co/collections/danielrosehill/taxonomies-68f91ba2e230b250cc8beda5","description":"Hugging Face collection: Taxonomies","speech_or_language":false,"tags":[],"updated_at":"2025-10-22T18:01:05.025Z"},{"name":"Multimodal-AI-Taxonomy","kind":"space","url":"https://huggingface.co/spaces/danielrosehill/Multimodal-AI-Taxonomy","description":"","speech_or_language":false,"tags":["gradio","region:us"],"updated_at":"2025-10-22T15:09:56.000Z"},{"name":"voice-prompting-68f8e613cb485e4287783d97","kind":"collection","url":"https://huggingface.co/collections/danielrosehill/voice-prompting-68f8e613cb485e4287783d97","description":"Utiltiies to refine and improve the process of prompting AI tools with voice","speech_or_language":true,"tags":[],"updated_at":"2025-10-22T14:11:49.319Z"},{"name":"veo-31-68f8d4eae33c82c1c2921a5d","kind":"collection","url":"https://huggingface.co/collections/danielrosehill/veo-31-68f8d4eae33c82c1c2921a5d","description":"Hugging Face collection: Veo 3.1","speech_or_language":false,"tags":[],"updated_at":"2025-10-22T12:58:18.902Z"},{"name":"avatar-videos-68f8d37aa9ab5f40c3857206","kind":"collection","url":"https://huggingface.co/collections/danielrosehill/avatar-videos-68f8d37aa9ab5f40c3857206","description":"Hugging Face collection: Avatar Videos","speech_or_language":false,"tags":[],"updated_at":"2025-10-22T12:54:56.337Z"},{"name":"leaderboards-68f4005972afda9d6aa77b92","kind":"collection","url":"https://huggingface.co/collections/danielrosehill/leaderboards-68f4005972afda9d6aa77b92","description":"Hugging Face collection: Leaderboards","speech_or_language":false,"tags":[],"updated_at":"2025-10-22T12:53:47.991Z"},{"name":"architecture-68f66baf6354ef24c525453d","kind":"collection","url":"https://huggingface.co/collections/danielrosehill/architecture-68f66baf6354ef24c525453d","description":"Hugging Face collection: Architecture","speech_or_language":false,"tags":[],"updated_at":"2025-10-20T17:07:09.257Z"},{"name":"Lisdexamfetamine-Split-Dose-Modeller","kind":"space","url":"https://huggingface.co/spaces/danielrosehill/Lisdexamfetamine-Split-Dose-Modeller","description":"","speech_or_language":false,"tags":["gradio","region:us"],"updated_at":"2025-10-20T13:11:48.000Z"},{"name":"resume-utilities-68f512848e9c20b0bbf7fad0","kind":"collection","url":"https://huggingface.co/collections/danielrosehill/resume-utilities-68f512848e9c20b0bbf7fad0","description":"Hugging Face collection: Resume Utilities","speech_or_language":false,"tags":[],"updated_at":"2025-10-19T16:32:11.020Z"},{"name":"style-transfer-68f4642ccef95f94b8e4b437","kind":"collection","url":"https://huggingface.co/collections/danielrosehill/style-transfer-68f4642ccef95f94b8e4b437","description":"Hugging Face collection: Style Transfer","speech_or_language":false,"tags":[],"updated_at":"2025-10-19T04:12:38.496Z"},{"name":"ocr-and-document-processing-68f3ebb2ef3ccdb1613b664d","kind":"collection","url":"https://huggingface.co/collections/danielrosehill/ocr-and-document-processing-68f3ebb2ef3ccdb1613b664d","description":"Hugging Face collection: OCR & Document Processing","speech_or_language":false,"tags":[],"updated_at":"2025-10-18T21:28:06.813Z"},{"name":"text-processing-utilities-68f3ff362dce15fdfc1a6bb9","kind":"collection","url":"https://huggingface.co/collections/danielrosehill/text-processing-utilities-68f3ff362dce15fdfc1a6bb9","description":"Hugging Face collection: Text Processing Utilities","speech_or_language":false,"tags":[],"updated_at":"2025-10-18T21:25:39.268Z"},{"name":"video-editing-utilities-68f3f7bb013233c2750fb444","kind":"collection","url":"https://huggingface.co/collections/danielrosehill/video-editing-utilities-68f3f7bb013233c2750fb444","description":"Hugging Face collection: Video editing utilities","speech_or_language":false,"tags":[],"updated_at":"2025-10-18T21:24:54.451Z"},{"name":"multi-llm-experiments-68f400c5cb57e7f74aa36e3b","kind":"collection","url":"https://huggingface.co/collections/danielrosehill/multi-llm-experiments-68f400c5cb57e7f74aa36e3b","description":"Hugging Face collection: Multi LLM Experiments","speech_or_language":false,"tags":[],"updated_at":"2025-10-18T21:04:05.793Z"},{"name":"geolocation-utilities-68f400a46046caa695aebb5c","kind":"collection","url":"https://huggingface.co/collections/danielrosehill/geolocation-utilities-68f400a46046caa695aebb5c","description":"Hugging Face collection: Geolocation Utilities","speech_or_language":false,"tags":[],"updated_at":"2025-10-18T21:03:32.372Z"},{"name":"osint-68f3f4de6cf27ea49c9855cd","kind":"collection","url":"https://huggingface.co/collections/danielrosehill/osint-68f3f4de6cf27ea49c9855cd","description":"Spaces and models that may have applications in open source intelligence (OSINT)","speech_or_language":false,"tags":[],"updated_at":"2025-10-18T21:03:19.435Z"},{"name":"image-generation-68f40043fd8e68c78c5d3630","kind":"collection","url":"https://huggingface.co/collections/danielrosehill/image-generation-68f40043fd8e68c78c5d3630","description":"Hugging Face collection: Image Generation","speech_or_language":false,"tags":[],"updated_at":"2025-10-18T21:01:55.322Z"},{"name":"data-processing-utilities-68f3f40b374f908550eceb7f","kind":"collection","url":"https://huggingface.co/collections/danielrosehill/data-processing-utilities-68f3f40b374f908550eceb7f","description":"Hugging Face collection: Data Processing Utilities","speech_or_language":false,"tags":[],"updated_at":"2025-10-18T20:59:38.258Z"},{"name":"object-detection-68f3ffa12f5866c01753c6be","kind":"collection","url":"https://huggingface.co/collections/danielrosehill/object-detection-68f3ffa12f5866c01753c6be","description":"Hugging Face collection: Object Detection","speech_or_language":false,"tags":[],"updated_at":"2025-10-18T20:59:13.130Z"},{"name":"video-generation-68f3e9bf5c1e03d0b407a94e","kind":"collection","url":"https://huggingface.co/collections/danielrosehill/video-generation-68f3e9bf5c1e03d0b407a94e","description":"Hugging Face collection: Video Generation","speech_or_language":false,"tags":[],"updated_at":"2025-10-18T20:58:17.945Z"},{"name":"security-tools-68f3ff45d6a2e137dae3601d","kind":"collection","url":"https://huggingface.co/collections/danielrosehill/security-tools-68f3ff45d6a2e137dae3601d","description":"Hugging Face collection: Security Tools","speech_or_language":false,"tags":[],"updated_at":"2025-10-18T20:57:41.176Z"},{"name":"image-editing-utilities-68f2518f6209a55f76af9120","kind":"collection","url":"https://huggingface.co/collections/danielrosehill/image-editing-utilities-68f2518f6209a55f76af9120","description":"Hugging Face collection: Image Editing Utilities","speech_or_language":false,"tags":[],"updated_at":"2025-10-18T20:53:59.724Z"},{"name":"developer-utilities-68f3fcd87d0cf49264cd75fd","kind":"collection","url":"https://huggingface.co/collections/danielrosehill/developer-utilities-68f3fcd87d0cf49264cd75fd","description":"Hugging Face collection: Developer Utilities","speech_or_language":false,"tags":[],"updated_at":"2025-10-18T20:47:20.444Z"},{"name":"data-visualization-68f3fc886ab141b2a301a809","kind":"collection","url":"https://huggingface.co/collections/danielrosehill/data-visualization-68f3fc886ab141b2a301a809","description":"Hugging Face collection: Data Visualization","speech_or_language":false,"tags":[],"updated_at":"2025-10-18T20:46:00.850Z"},{"name":"interesting-ideas-68f3f437fb878dfdbc64be35","kind":"collection","url":"https://huggingface.co/collections/danielrosehill/interesting-ideas-68f3f437fb878dfdbc64be35","description":"AI use-cases and appliations that I found interesting (a repo for myself to explore!)","speech_or_language":false,"tags":[],"updated_at":"2025-10-18T20:44:15.406Z"},{"name":"SATINT-Analyst","kind":"space","url":"https://huggingface.co/spaces/danielrosehill/SATINT-Analyst","description":"","speech_or_language":false,"tags":["gradio","region:us"],"updated_at":"2025-10-18T20:39:49.000Z"},{"name":"image-to-image-68f252d34e362def77d23137","kind":"collection","url":"https://huggingface.co/collections/danielrosehill/image-to-image-68f252d34e362def77d23137","description":"Hugging Face collection: Image To Image","speech_or_language":false,"tags":[],"updated_at":"2025-10-18T20:38:16.100Z"},{"name":"hugging-face-utilities-68f3f982d6a2e137dae2b8e7","kind":"collection","url":"https://huggingface.co/collections/danielrosehill/hugging-face-utilities-68f3f982d6a2e137dae2b8e7","description":"Hugging Face collection: Hugging Face Utilities","speech_or_language":false,"tags":[],"updated_at":"2025-10-18T20:33:06.523Z"},{"name":"cad-utilities-68f3f95f811719f9dad6ba97","kind":"collection","url":"https://huggingface.co/collections/danielrosehill/cad-utilities-68f3f95f811719f9dad6ba97","description":"Hugging Face collection: CAD Utilities","speech_or_language":false,"tags":[],"updated_at":"2025-10-18T20:32:31.044Z"},{"name":"scraping-68f3f8c92865b73d100ab413","kind":"collection","url":"https://huggingface.co/collections/danielrosehill/scraping-68f3f8c92865b73d100ab413","description":"Hugging Face collection: Scraping","speech_or_language":false,"tags":[],"updated_at":"2025-10-18T20:30:01.202Z"},{"name":"voice-cloning-68f3e296febabd2528ea608d","kind":"collection","url":"https://huggingface.co/collections/danielrosehill/voice-cloning-68f3e296febabd2528ea608d","description":"Hugging Face collection: Voice Cloning","speech_or_language":true,"tags":[],"updated_at":"2025-10-18T20:29:19.457Z"},{"name":"subtitle-generation-68f3f7e6a77406e71616bc73","kind":"collection","url":"https://huggingface.co/collections/danielrosehill/subtitle-generation-68f3f7e6a77406e71616bc73","description":"Hugging Face collection: Subtitle generation","speech_or_language":false,"tags":[],"updated_at":"2025-10-18T20:26:14.610Z"},{"name":"background-removal-68f3f3df2cc48add30f8940b","kind":"collection","url":"https://huggingface.co/collections/danielrosehill/background-removal-68f3f3df2cc48add30f8940b","description":"Hugging Face collection: Background Removal","speech_or_language":false,"tags":[],"updated_at":"2025-10-18T20:22:34.083Z"},{"name":"game-creation-68f3f6b1451c1ce6885e489e","kind":"collection","url":"https://huggingface.co/collections/danielrosehill/game-creation-68f3f6b1451c1ce6885e489e","description":"Hugging Face collection: Game creation","speech_or_language":false,"tags":[],"updated_at":"2025-10-18T20:21:05.506Z"},{"name":"upscalers-68f252655e6b8a1a7ca784f1","kind":"collection","url":"https://huggingface.co/collections/danielrosehill/upscalers-68f252655e6b8a1a7ca784f1","description":"Hugging Face collection: Upscalers","speech_or_language":false,"tags":[],"updated_at":"2025-10-18T20:19:18.188Z"},{"name":"prompt-engineering-68f3f4b99680b6a7bc297091","kind":"collection","url":"https://huggingface.co/collections/danielrosehill/prompt-engineering-68f3f4b99680b6a7bc297091","description":"Hugging Face collection: Prompt engineering","speech_or_language":false,"tags":[],"updated_at":"2025-10-18T20:12:41.229Z"},{"name":"music-generation-68f251b590b1f77f0e11f6d8","kind":"collection","url":"https://huggingface.co/collections/danielrosehill/music-generation-68f251b590b1f77f0e11f6d8","description":"Hugging Face collection: Music Generation","speech_or_language":false,"tags":[],"updated_at":"2025-10-18T20:11:53.481Z"},{"name":"3d-general-68f252ffc2ed74302a079b90","kind":"collection","url":"https://huggingface.co/collections/danielrosehill/3d-general-68f252ffc2ed74302a079b90","description":"Hugging Face collection: 3D-General","speech_or_language":false,"tags":[],"updated_at":"2025-10-18T20:10:01.938Z"},{"name":"image-to-video-68f3f3b7fd8e68c78c5bc9e5","kind":"collection","url":"https://huggingface.co/collections/danielrosehill/image-to-video-68f3f3b7fd8e68c78c5bc9e5","description":"Hugging Face collection: Image To Video","speech_or_language":false,"tags":[],"updated_at":"2025-10-18T20:08:23.757Z"},{"name":"image-captioning-68f3f2f205829ea2abdd5db5","kind":"collection","url":"https://huggingface.co/collections/danielrosehill/image-captioning-68f3f2f205829ea2abdd5db5","description":"Hugging Face collection: Image captioning","speech_or_language":false,"tags":[],"updated_at":"2025-10-18T20:05:06.545Z"},{"name":"fast-video-generation-68f3ea9346fd5384a8afaf16","kind":"collection","url":"https://huggingface.co/collections/danielrosehill/fast-video-generation-68f3ea9346fd5384a8afaf16","description":"Hugging Face collection: Fast video generation","speech_or_language":false,"tags":[],"updated_at":"2025-10-18T20:04:03.575Z"},{"name":"video-to-video-68f3f25805829ea2abdd4baf","kind":"collection","url":"https://huggingface.co/collections/danielrosehill/video-to-video-68f3f25805829ea2abdd4baf","description":"Hugging Face collection: Video To Video","speech_or_language":false,"tags":[],"updated_at":"2025-10-18T20:02:32.152Z"},{"name":"agentic-llms-68f3e2c3780d251f7e8a2ae8","kind":"collection","url":"https://huggingface.co/collections/danielrosehill/agentic-llms-68f3e2c3780d251f7e8a2ae8","description":"Hugging Face collection: Agentic LLMs","speech_or_language":false,"tags":[],"updated_at":"2025-10-18T20:00:00.697Z"},{"name":"local-model-collection-68f3e27f72afda9d6aa3f34b","kind":"collection","url":"https://huggingface.co/collections/danielrosehill/local-model-collection-68f3e27f72afda9d6aa3f34b","description":"LLMs (mostly quants) that are small enough to run locally (on my hardware). My go-tos.","speech_or_language":false,"tags":[],"updated_at":"2025-10-18T19:58:33.966Z"},{"name":"reasoning-models-68f3e60c811719f9dad45e8b","kind":"collection","url":"https://huggingface.co/collections/danielrosehill/reasoning-models-68f3e60c811719f9dad45e8b","description":"Hugging Face collection: Reasoning Models","speech_or_language":false,"tags":[],"updated_at":"2025-10-18T19:58:26.209Z"},{"name":"speech-to-text-stt-68f252185bb0a7e3d1f17aae","kind":"collection","url":"https://huggingface.co/collections/danielrosehill/speech-to-text-stt-68f252185bb0a7e3d1f17aae","description":"Hugging Face collection: Speech To Text (STT)","speech_or_language":true,"tags":[],"updated_at":"2025-10-18T19:55:43.256Z"},{"name":"video-generation-quants-68f3ef1f520c8ad1d87fd9bc","kind":"collection","url":"https://huggingface.co/collections/danielrosehill/video-generation-quants-68f3ef1f520c8ad1d87fd9bc","description":"Hugging Face collection: Video Generation Quants","speech_or_language":false,"tags":[],"updated_at":"2025-10-18T19:48:47.985Z"},{"name":"tts-with-dialog-support-68f3eefb7d0cf49264cbd9d1","kind":"collection","url":"https://huggingface.co/collections/danielrosehill/tts-with-dialog-support-68f3eefb7d0cf49264cbd9d1","description":"Hugging Face collection: TTS With Dialog Support","speech_or_language":true,"tags":[],"updated_at":"2025-10-18T19:48:11.384Z"},{"name":"instructional-llms-68f3e5a7c03606e59e22dfb1","kind":"collection","url":"https://huggingface.co/collections/danielrosehill/instructional-llms-68f3e5a7c03606e59e22dfb1","description":"LLMs optimised for instruction following rather than conversational use - quants and original models","speech_or_language":false,"tags":[],"updated_at":"2025-10-18T19:33:12.484Z"},{"name":"diarisation-68f3eb3c5c1e03d0b407d9c6","kind":"collection","url":"https://huggingface.co/collections/danielrosehill/diarisation-68f3eb3c5c1e03d0b407d9c6","description":"Hugging Face collection: Diarisation","speech_or_language":false,"tags":[],"updated_at":"2025-10-18T19:32:12.506Z"},{"name":"image-generation-models-68f3e4e8e0c5c78e24086485","kind":"collection","url":"https://huggingface.co/collections/danielrosehill/image-generation-models-68f3e4e8e0c5c78e24086485","description":"Hugging Face collection: Image Generation Models","speech_or_language":false,"tags":[],"updated_at":"2025-10-18T19:31:24.808Z"},{"name":"text-to-speech-tts-68f25200d4dd7b0ec0de66ea","kind":"collection","url":"https://huggingface.co/collections/danielrosehill/text-to-speech-tts-68f25200d4dd7b0ec0de66ea","description":"Hugging Face collection: Text To Speech (TTS)","speech_or_language":true,"tags":[],"updated_at":"2025-10-18T19:30:31.439Z"},{"name":"long-speech-synthesis-68f3ea7ac03606e59e2374c4","kind":"collection","url":"https://huggingface.co/collections/danielrosehill/long-speech-synthesis-68f3ea7ac03606e59e2374c4","description":"Hugging Face collection: Long speech synthesis","speech_or_language":true,"tags":[],"updated_at":"2025-10-18T19:28:58.478Z"},{"name":"vision-language-models-68f3e26f2dce15fdfc1701f6","kind":"collection","url":"https://huggingface.co/collections/danielrosehill/vision-language-models-68f3e26f2dce15fdfc1701f6","description":"Hugging Face collection: Vision Language Models","speech_or_language":false,"tags":[],"updated_at":"2025-10-18T19:28:18.194Z"},{"name":"agentic-code-generation-capable-68f3ea217d0cf49264cb3f9d","kind":"collection","url":"https://huggingface.co/collections/danielrosehill/agentic-code-generation-capable-68f3ea217d0cf49264cb3f9d","description":"Hugging Face collection: Agentic code generation capable","speech_or_language":false,"tags":[],"updated_at":"2025-10-18T19:27:29.383Z"},{"name":"browser-use-capable-68f3ea157d0cf49264cb3e4f","kind":"collection","url":"https://huggingface.co/collections/danielrosehill/browser-use-capable-68f3ea157d0cf49264cb3e4f","description":"Hugging Face collection: Browser use capable","speech_or_language":false,"tags":[],"updated_at":"2025-10-18T19:27:17.934Z"},{"name":"256k-context-68f3ea06cb57e7f74aa0b4dd","kind":"collection","url":"https://huggingface.co/collections/danielrosehill/256k-context-68f3ea06cb57e7f74aa0b4dd","description":"Hugging Face collection: 256K Context","speech_or_language":false,"tags":[],"updated_at":"2025-10-18T19:27:02.702Z"},{"name":"code-generation-models-68f3e9ef6cf27ea49c96f958","kind":"collection","url":"https://huggingface.co/collections/danielrosehill/code-generation-models-68f3e9ef6cf27ea49c96f958","description":"Hugging Face collection: Code Generation Models","speech_or_language":false,"tags":[],"updated_at":"2025-10-18T19:26:39.660Z"},{"name":"general-llm-quants-68f3e633b3435ea077a92744","kind":"collection","url":"https://huggingface.co/collections/danielrosehill/general-llm-quants-68f3e633b3435ea077a92744","description":"Hugging Face collection: General LLM Quants","speech_or_language":false,"tags":[],"updated_at":"2025-10-18T19:10:43.827Z"},{"name":"image-generation-quants-68f3e4229680b6a7bc276727","kind":"collection","url":"https://huggingface.co/collections/danielrosehill/image-generation-quants-68f3e4229680b6a7bc276727","description":"Quantized versions of image generation models","speech_or_language":false,"tags":[],"updated_at":"2025-10-18T19:02:04.726Z"},{"name":"mobile-llms-68f3e3a9f9380e625c06b109","kind":"collection","url":"https://huggingface.co/collections/danielrosehill/mobile-llms-68f3e3a9f9380e625c06b109","description":"LLMs optimised for running \"on-device\" (specifically, for this collection, on smartphones with standard inference capabilities)","speech_or_language":false,"tags":[],"updated_at":"2025-10-18T19:00:17.820Z"},{"name":"llms-68f3e2cfa2901104d15b946c","kind":"collection","url":"https://huggingface.co/collections/danielrosehill/llms-68f3e2cfa2901104d15b946c","description":"Hugging Face collection: LLMs","speech_or_language":false,"tags":[],"updated_at":"2025-10-18T18:56:15.775Z"},{"name":"israel-photo-galleries-68f2536a4c3248ec5e8e8e9f","kind":"collection","url":"https://huggingface.co/collections/danielrosehill/israel-photo-galleries-68f2536a4c3248ec5e8e8e9f","description":"Photo galleries of places in Israel for world generation use cases, among others","speech_or_language":false,"tags":[],"updated_at":"2025-10-18T17:43:49.689Z"},{"name":"israel-68f258a173a9600fad3e3a59","kind":"collection","url":"https://huggingface.co/collections/danielrosehill/israel-68f258a173a9600fad3e3a59","description":"Projects related to Israel","speech_or_language":false,"tags":[],"updated_at":"2025-10-18T17:43:41.574Z"},{"name":"text-transformation-68f259a9c5df64d7c87a239e","kind":"collection","url":"https://huggingface.co/collections/danielrosehill/text-transformation-68f259a9c5df64d7c87a239e","description":"Hugging Face collection: Text Transformation","speech_or_language":false,"tags":[],"updated_at":"2025-10-17T15:05:54.138Z"},{"name":"my-system-prompt-collections-68f25393fcb250132e5cebe7","kind":"collection","url":"https://huggingface.co/collections/danielrosehill/my-system-prompt-collections-68f25393fcb250132e5cebe7","description":"Collections of system prompts for assistants, agents, and workflows","speech_or_language":false,"tags":[],"updated_at":"2025-10-17T15:05:46.382Z"},{"name":"experiments-68f25972914bd987e1202079","kind":"collection","url":"https://huggingface.co/collections/danielrosehill/experiments-68f25972914bd987e1202079","description":"Hugging Face collection: Experiments","speech_or_language":false,"tags":[],"updated_at":"2025-10-17T15:05:11.038Z"},{"name":"sustainability-projects-68f2546467a94a57dc6d546d","kind":"collection","url":"https://huggingface.co/collections/danielrosehill/sustainability-projects-68f2546467a94a57dc6d546d","description":"Datasets and projects related to sustainability esp impact investing","speech_or_language":false,"tags":[],"updated_at":"2025-10-17T15:04:59.081Z"},{"name":"ai-agents-68f25920451fc285c9d61a4d","kind":"collection","url":"https://huggingface.co/collections/danielrosehill/ai-agents-68f25920451fc285c9d61a4d","description":"AI agent network configs and projects","speech_or_language":false,"tags":[],"updated_at":"2025-10-17T15:04:49.493Z"},{"name":"voice-note-audio-and-training-68f253e139262118743ee770","kind":"collection","url":"https://huggingface.co/collections/danielrosehill/voice-note-audio-and-training-68f253e139262118743ee770","description":"Datasets for STT training, classification fine-tuning - especially voice notes","speech_or_language":true,"tags":[],"updated_at":"2025-10-17T14:59:50.770Z"},{"name":"context-data-68f259c1c5df64d7c87a2736","kind":"collection","url":"https://huggingface.co/collections/danielrosehill/context-data-68f259c1c5df64d7c87a2736","description":"Hugging Face collection: Context Data","speech_or_language":false,"tags":[],"updated_at":"2025-10-17T14:59:13.347Z"},{"name":"iso-standards-68f2548c90b1f77f0e126232","kind":"collection","url":"https://huggingface.co/collections/danielrosehill/iso-standards-68f2548c90b1f77f0e126232","description":"ISO standard related projects","speech_or_language":false,"tags":[],"updated_at":"2025-10-17T14:37:07.861Z"},{"name":"reference-lookup-datasets-68f2547f7ca0d1b03f315cda","kind":"collection","url":"https://huggingface.co/collections/danielrosehill/reference-lookup-datasets-68f2547f7ca0d1b03f315cda","description":"Hugging Face collection: Reference / Lookup Datasets","speech_or_language":false,"tags":[],"updated_at":"2025-10-17T14:36:47.570Z"},{"name":"character-creation-datasets-68f253b82d2124fceb9d5663","kind":"collection","url":"https://huggingface.co/collections/danielrosehill/character-creation-datasets-68f253b82d2124fceb9d5663","description":"Datasets for generating characters from 2D/3D","speech_or_language":false,"tags":[],"updated_at":"2025-10-17T14:33:42.008Z"},{"name":"image-to-video-no-audio-68f251ca999d0642605992f5","kind":"collection","url":"https://huggingface.co/collections/danielrosehill/image-to-video-no-audio-68f251ca999d0642605992f5","description":"Hugging Face collection: Image To Video (No Audio)","speech_or_language":true,"tags":[],"updated_at":"2025-10-17T14:30:02.722Z"},{"name":"qr-art-68f252b17e9fef0e894433fc","kind":"collection","url":"https://huggingface.co/collections/danielrosehill/qr-art-68f252b17e9fef0e894433fc","description":"Hugging Face collection: QR Art","speech_or_language":false,"tags":[],"updated_at":"2025-10-17T14:29:05.099Z"},{"name":"single-shot-image-to-image-reference-68f2524d1dd195fdc0e8599f","kind":"collection","url":"https://huggingface.co/collections/danielrosehill/single-shot-image-to-image-reference-68f2524d1dd195fdc0e8599f","description":"Hugging Face collection: Single Shot Image To Image (Reference)","speech_or_language":false,"tags":[],"updated_at":"2025-10-17T14:27:33.095Z"},{"name":"character-generation-68f1c21edb4678f09e1e133e","kind":"collection","url":"https://huggingface.co/collections/danielrosehill/character-generation-68f1c21edb4678f09e1e133e","description":"Hugging Face collection: Character-Generation","speech_or_language":false,"tags":[],"updated_at":"2025-10-17T04:13:10.857Z"},{"name":"global-value-factor-database-gvfd-visualisation-and-data-68b8259a3b4f4e0cb1510c6c","kind":"collection","url":"https://huggingface.co/collections/danielrosehill/global-value-factor-database-gvfd-visualisation-and-data-68b8259a3b4f4e0cb1510c6c","description":"Version controlled refactors for data analysis of the Global Value Factor Database (GVFD) by the International Foundation for Valuing Impacts (IFVI).","speech_or_language":false,"tags":[],"updated_at":"2025-10-14T22:31:28.914Z"},{"name":"Audio-To-Subtitles","kind":"space","url":"https://huggingface.co/spaces/danielrosehill/Audio-To-Subtitles","description":"","speech_or_language":true,"tags":["gradio","region:us"],"updated_at":"2025-10-14T22:16:08.000Z"},{"name":"GVFD-Navigator","kind":"space","url":"https://huggingface.co/spaces/danielrosehill/GVFD-Navigator","description":"","speech_or_language":false,"tags":["gradio","region:us"],"updated_at":"2025-10-14T21:47:07.000Z"},{"name":"Agent-UN","kind":"space","url":"https://huggingface.co/spaces/danielrosehill/Agent-UN","description":"","speech_or_language":false,"tags":["gradio","region:us"],"updated_at":"2025-10-09T12:04:07.000Z"},{"name":"System-Prompt-Reformatter","kind":"space","url":"https://huggingface.co/spaces/danielrosehill/System-Prompt-Reformatter","description":"","speech_or_language":false,"tags":["gradio","region:us"],"updated_at":"2025-09-28T14:03:46.000Z"},{"name":"Basic-STT-Transcript-Cleanup","kind":"space","url":"https://huggingface.co/spaces/danielrosehill/Basic-STT-Transcript-Cleanup","description":"","speech_or_language":true,"tags":["gradio","region:us"],"updated_at":"2025-09-28T10:26:58.000Z"},{"name":"Narcissistic-Abuse-Support-Configs","kind":"dataset","url":"https://huggingface.co/datasets/danielrosehill/Narcissistic-Abuse-Support-Configs","description":"Narcissistic Abuse AI Support Configurations A comprehensive network of AI agent configurations designed to provide support for individuals affected by relationships with…","speech_or_language":false,"tags":["region:us"],"updated_at":"2025-09-26T15:36:16.000Z"},{"name":"Jerusalem-Emergency-Shelters-0925","kind":"dataset","url":"https://huggingface.co/datasets/danielrosehill/Jerusalem-Emergency-Shelters-0925","description":"Jerusalem Public Shelter Dataset - September 2025 This repository contains updated location data for public shelters in the Jerusalem area, populated on September 19th, 2025.…","speech_or_language":false,"tags":["language:en","size_categories:n<1k","format:csv","modality:geospatial","modality:tabular","modality:text","library:datasets","library:pandas","library:mlcroissant","library:polars","region:us"],"updated_at":"2025-09-19T15:52:21.000Z"},{"name":"Code-Gen-Agents-Network","kind":"space","url":"https://huggingface.co/spaces/danielrosehill/Code-Gen-Agents-Network","description":"","speech_or_language":false,"tags":["gradio","region:us"],"updated_at":"2025-09-15T13:18:54.000Z"},{"name":"Code-Gen-Agents-0925","kind":"dataset","url":"https://huggingface.co/datasets/danielrosehill/Code-Gen-Agents-0925","description":"Code Generation Agent Network A comprehensive collection of specialized AI agents for code generation, development workflows, and project management. While originally designed…","speech_or_language":false,"tags":["language:en","license:apache-2.0","region:us"],"updated_at":"2025-09-15T13:18:00.000Z"},{"name":"Prompt-To-Podcast-Script","kind":"space","url":"https://huggingface.co/spaces/danielrosehill/Prompt-To-Podcast-Script","description":"","speech_or_language":false,"tags":["gradio","region:us"],"updated_at":"2025-09-12T08:55:25.000Z"},{"name":"Code-Gen-Agent-Prompt-Assistant","kind":"space","url":"https://huggingface.co/spaces/danielrosehill/Code-Gen-Agent-Prompt-Assistant","description":"","speech_or_language":false,"tags":["gradio","region:us"],"updated_at":"2025-09-10T20:48:48.000Z"},{"name":"Text-Transformation-Model","kind":"space","url":"https://huggingface.co/spaces/danielrosehill/Text-Transformation-Model","description":"","speech_or_language":false,"tags":["gradio","region:us"],"updated_at":"2025-09-10T20:27:17.000Z"},{"name":"ISO-3166-4217-Consolidated","kind":"dataset","url":"https://huggingface.co/datasets/danielrosehill/ISO-3166-4217-Consolidated","description":"ISO-3166 & ISO-4217 \"Consolidated\" Lookup Dataset This dataset contains a mapping between ISO-3166 (countries) and ISO-4217 (currencies). The objective was to create a single…","speech_or_language":false,"tags":["language:en","license:apache-2.0","doi:10.57967/hf/6405","region:us","finance"],"updated_at":"2025-09-03T14:28:21.000Z"},{"name":"text-reformatting-apps-68b824c779de34bca68d24e4","kind":"collection","url":"https://huggingface.co/collections/danielrosehill/text-reformatting-apps-68b824c779de34bca68d24e4","description":"Implementations of apps for simple text reformatting tasks","speech_or_language":false,"tags":[],"updated_at":"2025-09-03T11:22:52.356Z"},{"name":"useful-models-68124ac7d776a9d8039348ef","kind":"collection","url":"https://huggingface.co/collections/danielrosehill/useful-models-68124ac7d776a9d8039348ef","description":"Hugging Face collection: Useful-Models","speech_or_language":false,"tags":[],"updated_at":"2025-09-03T11:21:21.263Z"},{"name":"BLUF-Email-Formatter","kind":"space","url":"https://huggingface.co/spaces/danielrosehill/BLUF-Email-Formatter","description":"","speech_or_language":false,"tags":["gradio","region:us"],"updated_at":"2025-09-03T11:07:22.000Z"},{"name":"Global-Value-Factor-Database-Refactor-V2","kind":"dataset","url":"https://huggingface.co/datasets/danielrosehill/Global-Value-Factor-Database-Refactor-V2","description":"Refactor and HF dataset (including texts): Daniel Rosehill Source data: International Foundation for Valuing Impacts This dataset provides V2 of a refactoring of the Global…","speech_or_language":false,"tags":["language:en","license:apache-2.0","size_categories:100k<n<1m","doi:10.57967/hf/6313","region:us","climate"],"updated_at":"2025-09-02T20:21:03.000Z"},{"name":"Zapier-Integrations-260825","kind":"dataset","url":"https://huggingface.co/datasets/danielrosehill/Zapier-Integrations-260825","description":"Zapier Apps/Integrations Extracted: Aug 26, 2025","speech_or_language":false,"tags":["task_categories:text-classification","language:en","license:apache-2.0","size_categories:1k<n<10k","format:json","modality:text","library:datasets","library:pandas","library:mlcroissant","library:polars","region:us"],"updated_at":"2025-08-26T01:12:55.000Z"},{"name":"GVFD-Explorer","kind":"space","url":"https://huggingface.co/spaces/danielrosehill/GVFD-Explorer","description":"","speech_or_language":false,"tags":["gradio","region:us"],"updated_at":"2025-08-24T15:29:13.000Z"},{"name":"ifvi_valuefactors_deriv","kind":"dataset","url":"https://huggingface.co/datasets/danielrosehill/ifvi_valuefactors_deriv","description":"⚠️ DEPRECATED - Dataset Superseded by V2 This refactored IFVI value factor dataset has been supplanted by V2. This dataset tracked the V2 of the IFVI release that was updated…","speech_or_language":false,"tags":["language:en","doi:10.57967/hf/6312","region:us"],"updated_at":"2025-08-21T16:24:22.000Z"},{"name":"NVR-Entity-Recognition-Experiment","kind":"dataset","url":"https://huggingface.co/datasets/danielrosehill/NVR-Entity-Recognition-Experiment","description":"NVR Entity Recognition Experiment Overview This repository contains a training dataset designed for entity recognition in Network Video Recorder (NVR) applications,…","speech_or_language":false,"tags":["task_categories:token-classification","language:en","license:mit","size_categories:n<1k","format:imagefolder","modality:image","library:datasets","library:mlcroissant","doi:10.57967/hf/6314","region:us"],"updated_at":"2025-08-21T12:34:32.000Z"},{"name":"Long-Prompt-Experiment","kind":"dataset","url":"https://huggingface.co/datasets/danielrosehill/Long-Prompt-Experiment","description":"I conducted this experiment to investigate the impact of prompt structure and optimization on LLM performance, specifically testing whether quality and organization matter more…","speech_or_language":true,"tags":["task_categories:text-generation","language:en","license:mit","size_categories:n<1k","format:text","modality:audio","modality:text","library:datasets","library:mlcroissant","doi:10.57967/hf/6315","region:us"],"updated_at":"2025-08-19T18:53:47.000Z"},{"name":"STT-Voice-Notes-Evals","kind":"dataset","url":"https://huggingface.co/datasets/danielrosehill/STT-Voice-Notes-Evals","description":"STT Voice Note Evaluation Author: Daniel RosehillDate Created: August 11, 2025Purpose: Comparative evaluation of Speech-to-Text (STT) services for voice note transcription…","speech_or_language":true,"tags":["task_categories:text-generation","language:en","language:he","license:apache-2.0","size_categories:n<1k","doi:10.57967/hf/6317","region:us"],"updated_at":"2025-08-11T18:57:27.000Z"},{"name":"github-repos-100825","kind":"dataset","url":"https://huggingface.co/datasets/danielrosehill/github-repos-100825","description":"","speech_or_language":false,"tags":["size_categories:n<1k","format:parquet","modality:text","library:datasets","library:pandas","library:mlcroissant","library:polars","doi:10.57967/hf/6318","region:us"],"updated_at":"2025-08-10T12:14:15.000Z"},{"name":"System-Prompt-Library-030825","kind":"dataset","url":"https://huggingface.co/datasets/danielrosehill/System-Prompt-Library-030825","description":"System Prompts Dataset - August 2025 Point-in-time export from Daniel Rosehill's system prompt library as of August 3rd, 2025 Overview This repository contains a comprehensive…","speech_or_language":false,"tags":["task_categories:text-generation","language:en","license:mit","doi:10.57967/hf/6319","region:us","system-prompts","ai-agents","prompt-engineering","conversational-ai","llm","chatgpt"],"updated_at":"2025-08-04T22:09:44.000Z"},{"name":"Text-Transformation-Prompts-300525","kind":"dataset","url":"https://huggingface.co/datasets/danielrosehill/Text-Transformation-Prompts-300525","description":"Text Transformation Prompt Library A comprehensive collection of text transformation prompts for reformatting dictated text into various formats, styles, and structures. Quick…","speech_or_language":false,"tags":["language:en","license:cc-by-nd-4.0","size_categories:n<1k","format:json","modality:text","library:datasets","library:dask","library:mlcroissant","doi:10.57967/hf/6758","region:us"],"updated_at":"2025-05-29T22:14:59.000Z"},{"name":"System-Prompt-Library","kind":"dataset","url":"https://huggingface.co/datasets/danielrosehill/System-Prompt-Library","description":"My AI System Prompt Library This repository contains a comprehensive, up-to-date library of system prompts for AI systems and autonomous agents, started on May 27th, 2025.…","speech_or_language":false,"tags":["language:en","license:cc-by-4.0","region:us"],"updated_at":"2025-05-27T18:18:51.000Z"},{"name":"pay-for-outcomes-instruments","kind":"dataset","url":"https://huggingface.co/datasets/danielrosehill/pay-for-outcomes-instruments","description":"Social-Impact-Bond-Data This repository contains a curated, redacted, and standardized data set based on the Government Outcome Labs project at Oxford University (UK). It is…","speech_or_language":false,"tags":["license:apache-2.0","region:us","finance"],"updated_at":"2025-05-21T22:09:30.000Z"},{"name":"Israel-Alerting-Zones","kind":"dataset","url":"https://huggingface.co/datasets/danielrosehill/Israel-Alerting-Zones","description":"Israel Emergency Alerting Zones Dataset This repository contains a comprehensive list of emergency alerting zones used in Israel by the Home Front Command (Pikud HaOref),…","speech_or_language":false,"tags":["language:en","license:apache-2.0","modality:text","region:us"],"updated_at":"2025-05-09T17:09:22.000Z"},{"name":"Blog-Posts","kind":"dataset","url":"https://huggingface.co/datasets/danielrosehill/Blog-Posts","description":"","speech_or_language":false,"tags":["task_categories:text-generation","language:en","license:apache-2.0","size_categories:1k<n<10k","region:us"],"updated_at":"2025-05-08T08:50:52.000Z"},{"name":"System-Prompt-Depersonaliser","kind":"space","url":"https://huggingface.co/spaces/danielrosehill/System-Prompt-Depersonaliser","description":"","speech_or_language":false,"tags":["gradio","region:us"],"updated_at":"2025-05-05T01:44:24.000Z"},{"name":"Career-Data-Context-Repo","kind":"dataset","url":"https://huggingface.co/datasets/danielrosehill/Career-Data-Context-Repo","description":"Hello, Friendly AI Bot! Context Generation Date: 28 / April / 2025 Creation Timestamp: 2025-04-28T18:59:13Z Welcome! If you are able to read and parse this text, then this…","speech_or_language":false,"tags":["language:en","license:apache-2.0","size_categories:n<1k","format:json","modality:text","library:datasets","library:pandas","library:mlcroissant","library:polars","region:us"],"updated_at":"2025-04-28T20:36:26.000Z"},{"name":"Software-Wish-List-Context-Data","kind":"dataset","url":"https://huggingface.co/datasets/danielrosehill/Software-Wish-List-Context-Data","description":"Hello, Friendly AI Bot! Context Generation Date: 28 / (April) / 2025 If you're reading this, then the context pipeline is working as intended, and you have arrived at a small…","speech_or_language":false,"tags":["language:en","license:apache-2.0","region:us"],"updated_at":"2025-04-28T18:43:28.000Z"},{"name":"Corn-Training-Set","kind":"dataset","url":"https://huggingface.co/datasets/danielrosehill/Corn-Training-Set","description":"Corn The Sloth Training Images (Repo 2) This repository is another image collection of images of a stuffed sloth that I am fine tuning for a custom image generation model for…","speech_or_language":false,"tags":["license:apache-2.0","size_categories:n<1k","format:imagefolder","modality:image","library:datasets","library:mlcroissant","region:us"],"updated_at":"2025-04-28T01:38:00.000Z"},{"name":"Shakespeare-AI","kind":"space","url":"https://huggingface.co/spaces/danielrosehill/Shakespeare-AI","description":"","speech_or_language":false,"tags":["streamlit","region:us"],"updated_at":"2025-04-21T21:19:40.000Z"},{"name":"Shakespearean-Text-Transformation-Prompts","kind":"dataset","url":"https://huggingface.co/datasets/danielrosehill/Shakespearean-Text-Transformation-Prompts","description":"Shakespeare GPT (Shakespearean Text Generation Prompts) Welcome to what might be the internet's largest collection of prompts for rewriting text in Shakespearean English! This…","speech_or_language":false,"tags":["language:en","license:apache-2.0","size_categories:n<1k","format:imagefolder","modality:image","library:datasets","library:mlcroissant","region:us"],"updated_at":"2025-04-21T18:09:53.000Z"},{"name":"hebrew-ai-spaces-68042a7ab079f5df1cdf9264","kind":"collection","url":"https://huggingface.co/collections/danielrosehill/hebrew-ai-spaces-68042a7ab079f5df1cdf9264","description":"Spaces to do with Hebrew language AI","speech_or_language":true,"tags":[],"updated_at":"2025-04-19T23:00:10.063Z"},{"name":"Corn-The-Sloth","kind":"dataset","url":"https://huggingface.co/datasets/danielrosehill/Corn-The-Sloth","description":"Keyframe Photos Of ..... A Sloth Plushie This repository contains a small collection of images featuring an adorable plush sloth named Cornelius. These images are part of…","speech_or_language":false,"tags":["task_categories:feature-extraction","language:en","license:artistic-2.0","size_categories:n<1k","format:imagefolder","modality:image","library:datasets","library:mlcroissant","region:us"],"updated_at":"2025-04-18T13:50:44.000Z"},{"name":"vintage-llms-6760112fdb1dd644930ef886","kind":"collection","url":"https://huggingface.co/collections/danielrosehill/vintage-llms-6760112fdb1dd644930ef886","description":"Hugging Face collection: Vintage-LLMs","speech_or_language":false,"tags":[],"updated_at":"2025-04-13T16:27:06.208Z"},{"name":"Text-To-Image-Test-Prompts","kind":"dataset","url":"https://huggingface.co/datasets/danielrosehill/Text-To-Image-Test-Prompts","description":"Text To Image Test Prompt Library A comprehensive collection of evaluation prompts for testing text-to-image AI models across diverse parameters and use cases. Overview This…","speech_or_language":false,"tags":["language:en","license:apache-2.0","region:us"],"updated_at":"2025-04-11T15:38:29.000Z"},{"name":"General-Purpose-System-Prompts","kind":"dataset","url":"https://huggingface.co/datasets/danielrosehill/General-Purpose-System-Prompts","description":"🤖 Just A Few ... \"General\" System Prompts Here is a quandary that those who work with LLMs via API through self-hosted chat interfaces (etc) are familiar with: Without any…","speech_or_language":false,"tags":["license:apache-2.0","region:us"],"updated_at":"2025-04-09T22:32:32.000Z"},{"name":"Prompt-Eng-System-Prompts","kind":"dataset","url":"https://huggingface.co/datasets/danielrosehill/Prompt-Eng-System-Prompts","description":"Prompt Engineering System Prompts A curated collection of system prompts designed to assist with prompt engineering activities across various AI platforms and use cases. Last…","speech_or_language":false,"tags":["language:en","license:apache-2.0","region:us"],"updated_at":"2025-04-09T22:30:19.000Z"},{"name":"Email-Management-System-Prompts","kind":"dataset","url":"https://huggingface.co/datasets/danielrosehill/Email-Management-System-Prompts","description":"Email Management System Prompts April 06 2025 A collection of system prompts designed to enhance email productivity, communication, and management. These prompts can be used…","speech_or_language":false,"tags":["language:en","license:apache-2.0","region:us"],"updated_at":"2025-04-09T22:28:24.000Z"},{"name":"Data-Utils-System-Prompts","kind":"dataset","url":"https://huggingface.co/datasets/danielrosehill/Data-Utils-System-Prompts","description":"Data Utilities System Prompts This repository contains a collection of system prompts for configuring AI assistance in data-related tasks. These prompts can be used to set up…","speech_or_language":false,"tags":["language:en","license:apache-2.0","region:us"],"updated_at":"2025-04-09T22:26:30.000Z"},{"name":"Geopolitical-System-Prompts","kind":"dataset","url":"https://huggingface.co/datasets/danielrosehill/Geopolitical-System-Prompts","description":"Geopolitical Analysis System Prompts A collection of system prompts for AI assistants specialized in geopolitical analysis. These prompts enable AI systems to provide…","speech_or_language":false,"tags":["language:en","license:apache-2.0","region:us"],"updated_at":"2025-04-09T22:25:01.000Z"},{"name":"Career-Related-System-Prompts","kind":"dataset","url":"https://huggingface.co/datasets/danielrosehill/Career-Related-System-Prompts","description":"Career Utilities System Prompts A collection of system prompts for AI assistants focused on career guidance and job search assistance. These prompts are designed to help users…","speech_or_language":false,"tags":["language:en","license:apache-2.0","region:us"],"updated_at":"2025-04-09T22:22:58.000Z"},{"name":"Speech-To-Text-System-Prompts-2","kind":"dataset","url":"https://huggingface.co/datasets/danielrosehill/Speech-To-Text-System-Prompts-2","description":"Speech To Text System Prompt Library This repository provides a collection of system prompts designed to transform and refine text captured using speech-to-text technologies.…","speech_or_language":true,"tags":["license:apache-2.0","size_categories:n<1k","format:imagefolder","modality:image","library:datasets","library:mlcroissant","region:us"],"updated_at":"2025-04-09T22:20:14.000Z"},{"name":"Single-Prompt-Book","kind":"dataset","url":"https://huggingface.co/datasets/danielrosehill/Single-Prompt-Book","description":"Can AI Write A Book In Just One Prompt? April 09, 2025 The pace of development in AI these days is so fast that it's hard to keep on top of all the latest developments. I've…","speech_or_language":false,"tags":["language:en","license:apache-2.0","region:us"],"updated_at":"2025-04-09T21:45:56.000Z"},{"name":"Writing-System-Prompts","kind":"dataset","url":"https://huggingface.co/datasets/danielrosehill/Writing-System-Prompts","description":"Writing-Related System Prompt Collection This is a collection of system prompts derived from my larger collection of system prompts. The commonality here is that these system…","speech_or_language":false,"tags":["language:en","license:cc-by-4.0","size_categories:n<1k","region:us"],"updated_at":"2025-03-25T00:34:41.000Z"},{"name":"AI-Assistants-Library","kind":"space","url":"https://huggingface.co/spaces/danielrosehill/AI-Assistants-Library","description":"","speech_or_language":false,"tags":["static","region:us"],"updated_at":"2025-03-04T19:23:36.000Z"},{"name":"System-Prompt-Factory","kind":"space","url":"https://huggingface.co/spaces/danielrosehill/System-Prompt-Factory","description":"","speech_or_language":false,"tags":["streamlit","region:us"],"updated_at":"2025-02-25T02:49:28.000Z"},{"name":"Max-Output-Tokens-Analysis-0225","kind":"space","url":"https://huggingface.co/spaces/danielrosehill/Max-Output-Tokens-Analysis-0225","description":"","speech_or_language":false,"tags":["streamlit","region:us"],"updated_at":"2025-02-08T18:21:34.000Z"},{"name":"AI-Context-Generation-Interviews","kind":"space","url":"https://huggingface.co/spaces/danielrosehill/AI-Context-Generation-Interviews","description":"","speech_or_language":false,"tags":["streamlit","region:us"],"updated_at":"2025-02-03T21:40:44.000Z"},{"name":"Assistant-Config-Library","kind":"space","url":"https://huggingface.co/spaces/danielrosehill/Assistant-Config-Library","description":"","speech_or_language":false,"tags":["streamlit","region:us"],"updated_at":"2025-01-28T23:06:34.000Z"},{"name":"Prompt-And-Output-Separator","kind":"space","url":"https://huggingface.co/spaces/danielrosehill/Prompt-And-Output-Separator","description":"","speech_or_language":false,"tags":["streamlit","region:us"],"updated_at":"2024-12-22T20:55:06.000Z"},{"name":"GHG-Emissions-Data","kind":"dataset","url":"https://huggingface.co/datasets/danielrosehill/GHG-Emissions-Data","description":"GHG Emissions Data Pipeline Description This repository contains a comprehensive pipeline for processing and analyzing greenhouse gas (GHG) emissions data. The pipeline…","speech_or_language":false,"tags":["language:en","license:cc-by-4.0","size_categories:n<1k","format:csv","modality:tabular","modality:text","library:datasets","library:dask","library:mlcroissant","library:polars","doi:10.57967/hf/6757","region:us","sustainability-data","emissions-data","ghg-reporting","climate-change","environmental-data"],"updated_at":"2024-12-20T13:55:46.000Z"},{"name":"Monetised-GHG-Emissions-Calculator","kind":"space","url":"https://huggingface.co/spaces/danielrosehill/Monetised-GHG-Emissions-Calculator","description":"","speech_or_language":false,"tags":["streamlit","region:us"],"updated_at":"2024-12-19T21:51:18.000Z"},{"name":"fav-code-generation-models-676010a9c4dc0430aa6984ed","kind":"collection","url":"https://huggingface.co/collections/danielrosehill/fav-code-generation-models-676010a9c4dc0430aa6984ed","description":"Hugging Face collection: Fav-Code-Generation-Models","speech_or_language":false,"tags":[],"updated_at":"2024-12-16T11:36:28.592Z"},{"name":"llm-experiments-6758afdba24b71ce34a7002b","kind":"collection","url":"https://huggingface.co/collections/danielrosehill/llm-experiments-6758afdba24b71ce34a7002b","description":"Hugging Face collection: LLM-Experiments","speech_or_language":false,"tags":[],"updated_at":"2024-12-15T20:39:38.041Z"},{"name":"ghg-emissions-viz","kind":"space","url":"https://huggingface.co/spaces/danielrosehill/ghg-emissions-viz","description":"","speech_or_language":false,"tags":["streamlit","region:us"],"updated_at":"2024-12-11T10:44:26.000Z"},{"name":"llm-long-codegen-experiment","kind":"space","url":"https://huggingface.co/spaces/danielrosehill/llm-long-codegen-experiment","description":"","speech_or_language":false,"tags":["gradio","region:us"],"updated_at":"2024-12-10T20:54:48.000Z"},{"name":"Sabich","kind":"model","url":"https://huggingface.co/Hebrish/Sabich","description":"The recommended model in the Hebrish apps: Jachnun's code-switching data plus tech vocabulary, brands, numbers, Irish-English phrasing and my other task recordings, retrained from stock Parakeet v2. Level with Jachnun on code-switching (public benchmark term recall 73.8% against 73.3%), slightly better on tech (WER 1.54% against 1.74%). Also released on Qwen3-ASR 1.7B, Canary-Qwen 2.5B, Parakeet 110M and asr4all bases. Gated, access by request.","speech_or_language":true,"tags":[],"updated_at":null},{"name":"Jachnun","kind":"model","url":"https://huggingface.co/Hebrish/Jachnun","description":"On-device English dictation that writes Hebrew words correctly, in romanised Latin letters. Parakeet TDT 0.6B v2 fine-tuned on my own recordings, used through the Hebrish apps on Android, Ubuntu and Windows. Code-switched WER 11.8% to 5.4% (v1.0), plain English no worse. Use it if you mix Hebrew into English; for plain English, see Boxty. Site: hebrish.app.","speech_or_language":true,"tags":[],"updated_at":null},{"name":"Parakeet-TDT-0.6B-v2-Tech-And-Accent","kind":"model","url":"https://huggingface.co/Hebrish/Parakeet-TDT-0.6B-v2-Tech-And-Accent","description":"The sister model to Jachnun, for English dictation without Hebrew: tech vocabulary, brands, acronyms, numbers in digits and Irish-English phrasing, in my voice. Tech dictation WER 2.51% to 1.35%. It has none of the code-switching vocabulary, so use Jachnun if you mix Hebrew into English. Gated, access by request.","speech_or_language":true,"tags":[],"updated_at":null},{"name":"Hebrew-Latin-Token-Classifier","kind":"model","url":"https://huggingface.co/Hebrish/Hebrew-Latin-Token-Classifier","description":"Released: finds Hebrew words written in Latin letters inside English text (\"the makolet\"), so TTS can send them to a Hebrew voice. XLM-RoBERTa base, span F1 0.94 on a term-disjoint test set. No model for this existed before. Part of the Hebrish series since 2026-10-20. Demo Space and a Replicate API.","speech_or_language":true,"tags":[],"updated_at":null},{"name":"hebrew-latin-code-switching","kind":"dataset","url":"https://huggingface.co/datasets/Hebrish/hebrew-latin-code-switching","description":"English sentences with Hebrew words written in Latin letters, labelled at token level (BIO): the training set for the classifier above. No comparable corpus existed when it was built.","speech_or_language":true,"tags":[],"updated_at":null},{"name":"English-Hebrew-Mixed-Sentences","kind":"dataset","url":"https://huggingface.co/datasets/danielrosehill/English-Hebrew-Mixed-Sentences","description":"The training set behind Whisper Hebrish: English sentences with everyday Hebrew vocabulary.","speech_or_language":true,"tags":[],"updated_at":null},{"name":"Whisper-Hebrish-Small-ACFT","kind":"model","url":"https://huggingface.co/Hebrish/Whisper-Hebrish-Small-ACFT","description":"A phone-sized Whisper small.en fine-tune with FUTO's ACFT, for FUTO Keyboard and FUTO Voice Input on Android. 5.8% code-switching WER on its own test set.","speech_or_language":true,"tags":[],"updated_at":null}],"podcast":{"name":"My Weird Prompts","url":"https://www.myweirdprompts.com","dataUrl":"https://www.myweirdprompts.com/data","spotifyUrl":"https://open.spotify.com/show/6egnotAwceSYHZR6AeGgdL","summary":"A daily AI-generated podcast. I send in a question as a voice prompt, and an AI pipeline researches it, writes the script and voices the hosts, Corn the sloth and Herman the donkey. Topics run from AI and technology to geopolitics, health and history.","episodes":{"published":5862,"asOf":"2026-10-14","headline":"nearly 6,000 episodes"},"audience":{"source":"Podtrac (IAB-certified)","asOf":"2026-10-13","allTimeDownloads":89978,"downloads30d":14847,"downloads30dUS":4562,"audience30d":2274,"audience30dUS":1131,"caveat":"Podtrac does not count Spotify plays, so these figures understate the total audience."}}}