{
  "schema": "transcribe-golden-manifest-v1",
  "family": "whisper",
  "variant": "whisper-small.en",
  "source_model": {
    "hf_repo": "openai/whisper-small.en",
    "hf_revision": "e8727524f962ee844a7319d92be39ac1bd25655a"
  },
  "reference": {
    "kind": "transformers",
    "source": "https://github.com/huggingface/transformers",
    "revision": "v5.6.1",
    "entrypoint": "scripts/dump_reference_whisper_transformers.py"
  },
  "expected_dtype": "float32",
  "dtype_source": "config",
  "transcript_compare": "normalized",
  "frontend": {
    "sample_rate": 16000,
    "n_mels": 80,
    "hop_length": 160,
    "fft_size": 400,
    "win_length": 400,
    "window": "hann_periodic",
    "normalization": "global",
    "preemphasis": null,
    "dither": null
  },
  "tokenizer_summary": {
    "type": "bpe",
    "vocab_size": 51864,
    "special_tokens": {
      "bos": 50256,
      "eos": 50256,
      "pad": 50256,
      "decoder_start": 50257,
      "transcribe": 50358,
      "translate": 50357,
      "no_timestamps": 50362,
      "prev_sot": 50360
    }
  },
  "capabilities": {
    "languages": [
      "en"
    ],
    "language_detection": false,
    "translation": false,
    "timestamps": [
      "segment",
      "word"
    ],
    "streaming": false,
    "speaker_diarization": false
  },
  "tolerance_file": "tests/tolerances/whisper-small.en.json",
  "cases": [
    {
      "audio": "jfk"
    }
  ]
}