Driver metadata
- Slug
- whisper
- Status
- active
- Protocol match
- proxyDriver measures a proxy of the cited metric. Validator C1 auto-downgrades not_reproduced to partial.
- Agent version
- v0.1.0-whisper-librispeech-3slice16
- Paper
- Robust Speech Recognition via Large-Scale Weak Supervision
- Model
- openai/whisper-tiny.en
Claim citation
This is the exact paper claim the driver compares against before the Verdict Validator allows any public WRONG path.
- Location
- Table 8 · whisper-tiny.en
- Metric
- LibriSpeech test-clean WER(wer)
- Reported value
- 0.056
- PDF page
- 25
- Quoted text
- “tiny.en 5.6”
Source
Every driver is a TypeScript orchestration script that routes the result through the Verdict Validator. Drivers with runnable public checkpoints also link the hermetic Modal job that loads the model and measures the paper claim.
- scripts/run-reproduction-whisper.ts ↗— TS orchestrator (claim citation, DB write, validator wiring)
- modal/reproduce_whisper.py ↗— Python Modal job (HF model load, evaluation harness, measurement)