{
  "format": 1,
  "language": "en",
  "source": {
    "repo": "jonatasgrosman/wav2vec2-large-xlsr-53-english",
    "revision": "569a6236e92bd5f7652a0420bfe9bb94c5664080",
    "license": "apache-2.0",
    "base_model": "facebook/wav2vec2-large-xlsr-53",
    "datasets": [
      "common_voice"
    ],
    "files": {
      "model.safetensors": {
        "sha256": "6144f8464c6aaa220dd57c5a2ad4039b5710dcf8ee6e67057675f76597c19875",
        "size": 1261942732
      },
      "README.md": {
        "sha256": "5b50a1c21a9d7361efb0b336922756057a964b4415b2475fdd0a118257fb929d",
        "size": 5327
      },
      "config.json": {
        "sha256": "ffcc5c417fe11433447975d5053b2279fbeafd6bca03dd2753082e72ad2d36b7",
        "size": 1531
      },
      "preprocessor_config.json": {
        "sha256": "ca5999a45e98bb76ea87a461ba28a23ad32a5bb9f733b8e0f6546ff38b6c612d",
        "size": 262
      },
      "special_tokens_map.json": {
        "sha256": "bb7068de1150661a10b55f9e4b12a0e77af8bf91f5e45e1b58afaf1d0e17f675",
        "size": 85
      },
      "vocab.json": {
        "sha256": "1ec5eeced764da0e4a4d2e95c1c3470c3235222db0182088beac3e7c67eda8ff",
        "size": 300
      }
    }
  },
  "model": {
    "file": "model.int8.onnx",
    "precision": "int8-dynamic",
    "input": "input_values",
    "output": "logits",
    "sample_rate": 16000,
    "frame_samples": 320,
    "min_samples": 400,
    "normalize": "zero-mean-unit-variance",
    "blank_id": 0,
    "vocab_size": 33,
    "word_delimiter": "|",
    "text_case": "preserve",
    "unit": "char"
  },
  "text": {
    "spaces_separate_words": true,
    "unknown_characters": "wildcard"
  },
  "verify": {
    "tools": {
      "onnxruntime": [
        "1.30.0",
        "1.23.2"
      ]
    },
    "logit_maxabs_fp32_vs_torch": 7.295608520507812e-05,
    "frame_argmax_agreement_int8": 1.0,
    "fp32_size": 1262324439,
    "int8_size": 356284613,
    "frames": 49,
    "runtime": {
      "1.30.0": {
        "provider": "CPUExecutionProvider",
        "runs": {
          "0.5": {
            "samples": 8000,
            "shape": [
              1,
              24,
              33
            ],
            "seconds": 0.069899
          },
          "60.0": {
            "samples": 960000,
            "shape": [
              1,
              2999,
              33
            ],
            "seconds": 11.763146
          }
        }
      },
      "1.23.2": {
        "provider": "CPUExecutionProvider",
        "runs": {
          "0.5": {
            "samples": 8000,
            "shape": [
              1,
              24,
              33
            ],
            "seconds": 0.048931
          },
          "60.0": {
            "samples": 960000,
            "shape": [
              1,
              2999,
              33
            ],
            "seconds": 8.14218
          }
        }
      }
    },
    "word_start_delta_ms_int8_vs_fp32": {
      "median": 0.0,
      "p95": 0.0,
      "max": 0.0,
      "spans": 2,
      "test_audio": "espeak-ng generated sentence"
    },
    "sanity": {
      "monotonic_sentence_spans": 2,
      "unknown_wildcard_spans": 2
    },
    "cost": {
      "audio_seconds": 30,
      "alignment_seconds": 5.709291,
      "peak_rss_bytes": 2924044288
    },
    "synthetic_sentence": {
      "voice": "en-us",
      "caption": "hello world",
      "samples": 16828,
      "note": "pipeline/monotonicity check only; not an accuracy claim"
    },
    "ground_truth_generated_boundaries": {
      "words": [
        "hello",
        "world"
      ],
      "silence_ms": 500,
      "segment_start_errors_ms": [
        20.063762626262626,
        24.829545454545496
      ],
      "median": 22.446654040404063,
      "p95": 24.591256313131353,
      "max": 24.829545454545496,
      "note": "espeak-ng segments with known 500 ms join; segment boundaries, not human speech labels"
    }
  },
  "files": {
    "model.int8.onnx": {
      "sha256": "731ccae6a5d1cddce647cd7c4a2faef9fb31d4a12eab64714e9ff7b6048e7d7b",
      "size": 356284613
    }
  },
  "elapsed_seconds": 33.019
}