{
  "format": 1,
  "language": "zh",
  "source": {
    "repo": "ydshieh/wav2vec2-large-xlsr-53-chinese-zh-cn-gpt",
    "revision": "cf511f7e00de089e67e80fbedd5fbfb8e76ea067",
    "license": "apache-2.0",
    "base_model": "facebook/wav2vec2-large-xlsr-53",
    "datasets": [
      "common_voice"
    ],
    "files": {
      "README.md": {
        "sha256": "a285e286a678275dcfe176524160cca3ab637ad26acfe9fe35e8406f05c65a04",
        "size": 15169
      },
      "config.json": {
        "sha256": "ad7e01a16ff4d6b1848b696bd7db4c4214ec1532bf00071b4c806258b36d3b8e",
        "size": 1605
      },
      "preprocessor_config.json": {
        "sha256": "c403ce09975b90dff0dd8302c42d422e9de1f166cd7772df23490069893cb0cf",
        "size": 158
      },
      "pytorch_model.bin": {
        "sha256": "4a91f92daaf8e7be42192a05e41ba37e80a26dcc9729d22dd922c2e703b53736",
        "size": 1348558615
      },
      "special_tokens_map.json": {
        "sha256": "50eb73d51191696209d30d42d6ede50e57e7a542ca1db12df714b2c0aa3da8e2",
        "size": 85
      },
      "tokenizer_config.json": {
        "sha256": "3160c256a4d10e1fc5133a2d318e63406cc382b957563198931c8a90f9b9242d",
        "size": 138
      },
      "vocab.json": {
        "sha256": "e2fa9e1f77e7b4c26d31d60dca783175cb82da4227c6695e76fcf616829a41a8",
        "size": 309714
      }
    }
  },
  "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": 21128,
    "word_delimiter": "|",
    "text_case": "preserve",
    "unit": "char"
  },
  "text": {
    "spaces_separate_words": false,
    "unknown_characters": "wildcard"
  },
  "verify": {
    "tools": {
      "onnxruntime": [
        "1.30.0",
        "1.23.2"
      ]
    },
    "logit_maxabs_fp32_vs_torch": 0.000202178955078125,
    "frame_argmax_agreement_fp32_vs_torch": 1.0,
    "frame_argmax_agreement_int8": 1.0,
    "fp32_size": 1348813949,
    "int8_size": 378078776,
    "frames": 49,
    "runtime": {
      "1.30.0": {
        "provider": "CPUExecutionProvider",
        "runs": {
          "0.5": {
            "samples": 8000,
            "shape": [
              1,
              24,
              21128
            ],
            "seconds": 0.178673
          },
          "60.0": {
            "samples": 960000,
            "shape": [
              1,
              2999,
              21128
            ],
            "seconds": 11.422376
          }
        }
      },
      "1.23.2": {
        "provider": "CPUExecutionProvider",
        "runs": {
          "0.5": {
            "samples": 8000,
            "shape": [
              1,
              24,
              21128
            ],
            "seconds": 0.126522
          },
          "60.0": {
            "samples": 960000,
            "shape": [
              1,
              2999,
              21128
            ],
            "seconds": 10.217091
          }
        }
      }
    },
    "word_start_delta_ms_int8_vs_fp32": {
      "median": 0.0,
      "p95": 0.0,
      "max": 0.0,
      "spans": 4,
      "test_audio": "espeak-ng generated sentence"
    },
    "sanity": {
      "monotonic_sentence_spans": 4,
      "unknown_wildcard_spans": 1
    },
    "cost": {
      "audio_seconds": 30,
      "alignment_seconds": 5.006784,
      "peak_rss_bytes": 3979509760
    },
    "synthetic_sentence": {
      "voice": "cmn",
      "caption": "\u4f60\u597d\u4e16\u754c",
      "samples": 20216,
      "note": "pipeline/monotonicity check only; not an accuracy claim"
    }
  },
  "files": {
    "model.int8.onnx": {
      "sha256": "a2ed0362fec4989a81854b0485b61b551ef3dd9d3761a489960848a07acae985",
      "size": 378078776
    }
  },
  "elapsed_seconds": 40.785
}