English Hebrew Codeswitching ASR Benchmark

English-Hebrew Code-Switching ASR Benchmark (test set v1) Status: early-stage, work in progress. Known issue: 22 of the 139 clean-subset hebrew_terms labels are English tech words or product names (JSON, grep, Proxmox, Tuya, ...), 26 of 172 occurrences. v1.1 will relabel them; until then, term recall is 2 to 6 points higher than a Hebrew-only count. Details: methodology, limits. A frozen, versioned test set for measuring how well speech recognition handles English speech with… See the full description on the dataset page: https://huggingface.co/datasets/danielrosehill/English-Hebrew-Codeswitching-ASR-Benchmark.