{
  "mos": {
    "origins": {
      "reused_previous_same_float32_audio": 360,
      "new_inference": 960
    },
    "new_inference_recorded_seconds": 2481.5735438320125,
    "reused_rows": 360,
    "reused_historical_seconds_excluded": true,
    "initialization": [
      {
        "source": "runs/added-metrics-20261014/mos/batch-00/initialization.json",
        "seconds": 12.347712126999795,
        "provenance": {
          "name": "UTMOSv2",
          "quantity": "predicted_mos_not_human_rating",
          "source_repository": "https://github.com/sarulab-speech/UTMOSv2",
          "source_commit": "cc2700db57bb83ee13dc31ebe1b868c254e15d09",
          "checkpoint_revision": "506474f2b33dc77c234d668cc419be1861899cad",
          "checkpoint_sha256": "c8149d988e4bbf3f347e6966b5d769de347a5f8c59ffca1dc4bd4bf5b8585e57",
          "fold": 0,
          "device": "cpu",
          "configuration": "fusion_stage3",
          "predict_dataset": "bvcc",
          "repetitions": 1,
          "extraction_seed": 20261012,
          "cpu_threads": 4,
          "remove_silent_section": false,
          "nnpack_enabled": false,
          "internal_sample_rate": 16000,
          "ssl_crop_seconds": 3.0,
          "spectrogram_crop_seconds": 1.4,
          "spectrogram_frames_per_repetition": 2,
          "spectrogram_inner_mixup": true,
          "short_input_extension": "tile",
          "code_license": "MIT",
          "weights_license": "MIT",
          "independent_human_rating_calibration": "not_performed_for_this_project",
          "scope": "JVS/LibriSpeechの再合成への適用は未校正。言語・発声別に集計する。",
          "paper_protocol_difference": "論文の5fold×5抽出平均ではなく、fold0と上記反復数。",
          "initialization": "official_architecture_then_complete_strict_state_dict",
          "packages": {
            "torch": "2.8.0+cpu",
            "torchaudio": "2.8.0+cpu",
            "torchvision": "0.23.0+cpu",
            "transformers": "4.51.3",
            "timm": "1.0.15",
            "librosa": "0.11.0",
            "numpy": "2.3.5",
            "utmosv2": "1.3.1.dev0"
          },
          "checkpoint": {
            "url": "https://huggingface.co/sarulab-speech/UTMOSv2/resolve/506474f2b33dc77c234d668cc419be1861899cad/fold0_s42_best_model.pth",
            "sha256": "c8149d988e4bbf3f347e6966b5d769de347a5f8c59ffca1dc4bd4bf5b8585e57",
            "bytes": 818531314
          },
          "feature_configuration": {
            "config.json": {
              "url": "https://huggingface.co/facebook/wav2vec2-base/resolve/0b5b8e868dd84f03fd87d01f9c4ff0f080fecfe8/config.json",
              "sha256": "4937977e24d12d1bba70cdce8709c3c04807a8e4ae8ddac4229c48c436ae99ae",
              "bytes": 1842
            },
            "preprocessor_config.json": {
              "url": "https://huggingface.co/facebook/wav2vec2-base/resolve/0b5b8e868dd84f03fd87d01f9c4ff0f080fecfe8/preprocessor_config.json",
              "sha256": "b225d617c025463b9e157e06afea8b90dc7078fc70b013c533328423e0486b4a",
              "bytes": 159
            }
          },
          "initialization_status": "ready"
        },
        "environment": {
          "python": "3.12.14",
          "platform": "Linux-6.18.44-x86_64-with-glibc2.41",
          "cpu": "",
          "logical_cpu_count": 9,
          "packages": {
            "numpy": "2.3.5",
            "scipy": "1.17.0",
            "torch": "2.8.0+cpu",
            "torchaudio": "2.8.0+cpu",
            "onnxruntime": "1.22.0",
            "speechbrain": "1.0.3",
            "utmosv2": "1.3.1.dev0",
            "transformers": "4.51.3",
            "timm": "1.0.15",
            "librosa": "0.11.0"
          }
        }
      },
      {
        "source": "runs/added-metrics-20261014/mos/batch-01/initialization.json",
        "seconds": 7.311520341000687,
        "provenance": {
          "name": "UTMOSv2",
          "quantity": "predicted_mos_not_human_rating",
          "source_repository": "https://github.com/sarulab-speech/UTMOSv2",
          "source_commit": "cc2700db57bb83ee13dc31ebe1b868c254e15d09",
          "checkpoint_revision": "506474f2b33dc77c234d668cc419be1861899cad",
          "checkpoint_sha256": "c8149d988e4bbf3f347e6966b5d769de347a5f8c59ffca1dc4bd4bf5b8585e57",
          "fold": 0,
          "device": "cpu",
          "configuration": "fusion_stage3",
          "predict_dataset": "bvcc",
          "repetitions": 1,
          "extraction_seed": 20261012,
          "cpu_threads": 4,
          "remove_silent_section": false,
          "nnpack_enabled": false,
          "internal_sample_rate": 16000,
          "ssl_crop_seconds": 3.0,
          "spectrogram_crop_seconds": 1.4,
          "spectrogram_frames_per_repetition": 2,
          "spectrogram_inner_mixup": true,
          "short_input_extension": "tile",
          "code_license": "MIT",
          "weights_license": "MIT",
          "independent_human_rating_calibration": "not_performed_for_this_project",
          "scope": "JVS/LibriSpeechの再合成への適用は未校正。言語・発声別に集計する。",
          "paper_protocol_difference": "論文の5fold×5抽出平均ではなく、fold0と上記反復数。",
          "initialization": "official_architecture_then_complete_strict_state_dict",
          "packages": {
            "torch": "2.8.0+cpu",
            "torchaudio": "2.8.0+cpu",
            "torchvision": "0.23.0+cpu",
            "transformers": "4.51.3",
            "timm": "1.0.15",
            "librosa": "0.11.0",
            "numpy": "2.3.5",
            "utmosv2": "1.3.1.dev0"
          },
          "checkpoint": {
            "url": "https://huggingface.co/sarulab-speech/UTMOSv2/resolve/506474f2b33dc77c234d668cc419be1861899cad/fold0_s42_best_model.pth",
            "sha256": "c8149d988e4bbf3f347e6966b5d769de347a5f8c59ffca1dc4bd4bf5b8585e57",
            "bytes": 818531314
          },
          "feature_configuration": {
            "config.json": {
              "url": "https://huggingface.co/facebook/wav2vec2-base/resolve/0b5b8e868dd84f03fd87d01f9c4ff0f080fecfe8/config.json",
              "sha256": "4937977e24d12d1bba70cdce8709c3c04807a8e4ae8ddac4229c48c436ae99ae",
              "bytes": 1842
            },
            "preprocessor_config.json": {
              "url": "https://huggingface.co/facebook/wav2vec2-base/resolve/0b5b8e868dd84f03fd87d01f9c4ff0f080fecfe8/preprocessor_config.json",
              "sha256": "b225d617c025463b9e157e06afea8b90dc7078fc70b013c533328423e0486b4a",
              "bytes": 159
            }
          },
          "initialization_status": "ready"
        },
        "environment": {
          "python": "3.12.14",
          "platform": "Linux-6.18.44-x86_64-with-glibc2.41",
          "cpu": "",
          "logical_cpu_count": 9,
          "packages": {
            "numpy": "2.3.5",
            "scipy": "1.17.0",
            "torch": "2.8.0+cpu",
            "torchaudio": "2.8.0+cpu",
            "onnxruntime": "1.22.0",
            "speechbrain": "1.0.3",
            "utmosv2": "1.3.1.dev0",
            "transformers": "4.51.3",
            "timm": "1.0.15",
            "librosa": "0.11.0"
          }
        }
      },
      {
        "source": "runs/added-metrics-20261014/mos/batch-02/initialization.json",
        "seconds": 6.75253375200009,
        "provenance": {
          "name": "UTMOSv2",
          "quantity": "predicted_mos_not_human_rating",
          "source_repository": "https://github.com/sarulab-speech/UTMOSv2",
          "source_commit": "cc2700db57bb83ee13dc31ebe1b868c254e15d09",
          "checkpoint_revision": "506474f2b33dc77c234d668cc419be1861899cad",
          "checkpoint_sha256": "c8149d988e4bbf3f347e6966b5d769de347a5f8c59ffca1dc4bd4bf5b8585e57",
          "fold": 0,
          "device": "cpu",
          "configuration": "fusion_stage3",
          "predict_dataset": "bvcc",
          "repetitions": 1,
          "extraction_seed": 20261012,
          "cpu_threads": 4,
          "remove_silent_section": false,
          "nnpack_enabled": false,
          "internal_sample_rate": 16000,
          "ssl_crop_seconds": 3.0,
          "spectrogram_crop_seconds": 1.4,
          "spectrogram_frames_per_repetition": 2,
          "spectrogram_inner_mixup": true,
          "short_input_extension": "tile",
          "code_license": "MIT",
          "weights_license": "MIT",
          "independent_human_rating_calibration": "not_performed_for_this_project",
          "scope": "JVS/LibriSpeechの再合成への適用は未校正。言語・発声別に集計する。",
          "paper_protocol_difference": "論文の5fold×5抽出平均ではなく、fold0と上記反復数。",
          "initialization": "official_architecture_then_complete_strict_state_dict",
          "packages": {
            "torch": "2.8.0+cpu",
            "torchaudio": "2.8.0+cpu",
            "torchvision": "0.23.0+cpu",
            "transformers": "4.51.3",
            "timm": "1.0.15",
            "librosa": "0.11.0",
            "numpy": "2.3.5",
            "utmosv2": "1.3.1.dev0"
          },
          "checkpoint": {
            "url": "https://huggingface.co/sarulab-speech/UTMOSv2/resolve/506474f2b33dc77c234d668cc419be1861899cad/fold0_s42_best_model.pth",
            "sha256": "c8149d988e4bbf3f347e6966b5d769de347a5f8c59ffca1dc4bd4bf5b8585e57",
            "bytes": 818531314
          },
          "feature_configuration": {
            "config.json": {
              "url": "https://huggingface.co/facebook/wav2vec2-base/resolve/0b5b8e868dd84f03fd87d01f9c4ff0f080fecfe8/config.json",
              "sha256": "4937977e24d12d1bba70cdce8709c3c04807a8e4ae8ddac4229c48c436ae99ae",
              "bytes": 1842
            },
            "preprocessor_config.json": {
              "url": "https://huggingface.co/facebook/wav2vec2-base/resolve/0b5b8e868dd84f03fd87d01f9c4ff0f080fecfe8/preprocessor_config.json",
              "sha256": "b225d617c025463b9e157e06afea8b90dc7078fc70b013c533328423e0486b4a",
              "bytes": 159
            }
          },
          "initialization_status": "ready"
        },
        "environment": {
          "python": "3.12.14",
          "platform": "Linux-6.18.44-x86_64-with-glibc2.41",
          "cpu": "",
          "logical_cpu_count": 9,
          "packages": {
            "numpy": "2.3.5",
            "scipy": "1.17.0",
            "torch": "2.8.0+cpu",
            "torchaudio": "2.8.0+cpu",
            "onnxruntime": "1.22.0",
            "speechbrain": "1.0.3",
            "utmosv2": "1.3.1.dev0",
            "transformers": "4.51.3",
            "timm": "1.0.15",
            "librosa": "0.11.0"
          }
        }
      },
      {
        "source": "runs/added-metrics-20261014/mos/batch-03/initialization.json",
        "seconds": 6.619538762000047,
        "provenance": {
          "name": "UTMOSv2",
          "quantity": "predicted_mos_not_human_rating",
          "source_repository": "https://github.com/sarulab-speech/UTMOSv2",
          "source_commit": "cc2700db57bb83ee13dc31ebe1b868c254e15d09",
          "checkpoint_revision": "506474f2b33dc77c234d668cc419be1861899cad",
          "checkpoint_sha256": "c8149d988e4bbf3f347e6966b5d769de347a5f8c59ffca1dc4bd4bf5b8585e57",
          "fold": 0,
          "device": "cpu",
          "configuration": "fusion_stage3",
          "predict_dataset": "bvcc",
          "repetitions": 1,
          "extraction_seed": 20261012,
          "cpu_threads": 4,
          "remove_silent_section": false,
          "nnpack_enabled": false,
          "internal_sample_rate": 16000,
          "ssl_crop_seconds": 3.0,
          "spectrogram_crop_seconds": 1.4,
          "spectrogram_frames_per_repetition": 2,
          "spectrogram_inner_mixup": true,
          "short_input_extension": "tile",
          "code_license": "MIT",
          "weights_license": "MIT",
          "independent_human_rating_calibration": "not_performed_for_this_project",
          "scope": "JVS/LibriSpeechの再合成への適用は未校正。言語・発声別に集計する。",
          "paper_protocol_difference": "論文の5fold×5抽出平均ではなく、fold0と上記反復数。",
          "initialization": "official_architecture_then_complete_strict_state_dict",
          "packages": {
            "torch": "2.8.0+cpu",
            "torchaudio": "2.8.0+cpu",
            "torchvision": "0.23.0+cpu",
            "transformers": "4.51.3",
            "timm": "1.0.15",
            "librosa": "0.11.0",
            "numpy": "2.3.5",
            "utmosv2": "1.3.1.dev0"
          },
          "checkpoint": {
            "url": "https://huggingface.co/sarulab-speech/UTMOSv2/resolve/506474f2b33dc77c234d668cc419be1861899cad/fold0_s42_best_model.pth",
            "sha256": "c8149d988e4bbf3f347e6966b5d769de347a5f8c59ffca1dc4bd4bf5b8585e57",
            "bytes": 818531314
          },
          "feature_configuration": {
            "config.json": {
              "url": "https://huggingface.co/facebook/wav2vec2-base/resolve/0b5b8e868dd84f03fd87d01f9c4ff0f080fecfe8/config.json",
              "sha256": "4937977e24d12d1bba70cdce8709c3c04807a8e4ae8ddac4229c48c436ae99ae",
              "bytes": 1842
            },
            "preprocessor_config.json": {
              "url": "https://huggingface.co/facebook/wav2vec2-base/resolve/0b5b8e868dd84f03fd87d01f9c4ff0f080fecfe8/preprocessor_config.json",
              "sha256": "b225d617c025463b9e157e06afea8b90dc7078fc70b013c533328423e0486b4a",
              "bytes": 159
            }
          },
          "initialization_status": "ready"
        },
        "environment": {
          "python": "3.12.14",
          "platform": "Linux-6.18.44-x86_64-with-glibc2.41",
          "cpu": "",
          "logical_cpu_count": 9,
          "packages": {
            "numpy": "2.3.5",
            "scipy": "1.17.0",
            "torch": "2.8.0+cpu",
            "torchaudio": "2.8.0+cpu",
            "onnxruntime": "1.22.0",
            "speechbrain": "1.0.3",
            "utmosv2": "1.3.1.dev0",
            "transformers": "4.51.3",
            "timm": "1.0.15",
            "librosa": "0.11.0"
          }
        }
      },
      {
        "source": "runs/added-metrics-20261014/mos/batch-04/initialization.json",
        "seconds": 6.423138477000066,
        "provenance": {
          "name": "UTMOSv2",
          "quantity": "predicted_mos_not_human_rating",
          "source_repository": "https://github.com/sarulab-speech/UTMOSv2",
          "source_commit": "cc2700db57bb83ee13dc31ebe1b868c254e15d09",
          "checkpoint_revision": "506474f2b33dc77c234d668cc419be1861899cad",
          "checkpoint_sha256": "c8149d988e4bbf3f347e6966b5d769de347a5f8c59ffca1dc4bd4bf5b8585e57",
          "fold": 0,
          "device": "cpu",
          "configuration": "fusion_stage3",
          "predict_dataset": "bvcc",
          "repetitions": 1,
          "extraction_seed": 20261012,
          "cpu_threads": 4,
          "remove_silent_section": false,
          "nnpack_enabled": false,
          "internal_sample_rate": 16000,
          "ssl_crop_seconds": 3.0,
          "spectrogram_crop_seconds": 1.4,
          "spectrogram_frames_per_repetition": 2,
          "spectrogram_inner_mixup": true,
          "short_input_extension": "tile",
          "code_license": "MIT",
          "weights_license": "MIT",
          "independent_human_rating_calibration": "not_performed_for_this_project",
          "scope": "JVS/LibriSpeechの再合成への適用は未校正。言語・発声別に集計する。",
          "paper_protocol_difference": "論文の5fold×5抽出平均ではなく、fold0と上記反復数。",
          "initialization": "official_architecture_then_complete_strict_state_dict",
          "packages": {
            "torch": "2.8.0+cpu",
            "torchaudio": "2.8.0+cpu",
            "torchvision": "0.23.0+cpu",
            "transformers": "4.51.3",
            "timm": "1.0.15",
            "librosa": "0.11.0",
            "numpy": "2.3.5",
            "utmosv2": "1.3.1.dev0"
          },
          "checkpoint": {
            "url": "https://huggingface.co/sarulab-speech/UTMOSv2/resolve/506474f2b33dc77c234d668cc419be1861899cad/fold0_s42_best_model.pth",
            "sha256": "c8149d988e4bbf3f347e6966b5d769de347a5f8c59ffca1dc4bd4bf5b8585e57",
            "bytes": 818531314
          },
          "feature_configuration": {
            "config.json": {
              "url": "https://huggingface.co/facebook/wav2vec2-base/resolve/0b5b8e868dd84f03fd87d01f9c4ff0f080fecfe8/config.json",
              "sha256": "4937977e24d12d1bba70cdce8709c3c04807a8e4ae8ddac4229c48c436ae99ae",
              "bytes": 1842
            },
            "preprocessor_config.json": {
              "url": "https://huggingface.co/facebook/wav2vec2-base/resolve/0b5b8e868dd84f03fd87d01f9c4ff0f080fecfe8/preprocessor_config.json",
              "sha256": "b225d617c025463b9e157e06afea8b90dc7078fc70b013c533328423e0486b4a",
              "bytes": 159
            }
          },
          "initialization_status": "ready"
        },
        "environment": {
          "python": "3.12.14",
          "platform": "Linux-6.18.44-x86_64-with-glibc2.41",
          "cpu": "",
          "logical_cpu_count": 9,
          "packages": {
            "numpy": "2.3.5",
            "scipy": "1.17.0",
            "torch": "2.8.0+cpu",
            "torchaudio": "2.8.0+cpu",
            "onnxruntime": "1.22.0",
            "speechbrain": "1.0.3",
            "utmosv2": "1.3.1.dev0",
            "transformers": "4.51.3",
            "timm": "1.0.15",
            "librosa": "0.11.0"
          }
        }
      },
      {
        "source": "runs/added-metrics-20261014/mos/batch-05/initialization.json",
        "seconds": 6.687216815999818,
        "provenance": {
          "name": "UTMOSv2",
          "quantity": "predicted_mos_not_human_rating",
          "source_repository": "https://github.com/sarulab-speech/UTMOSv2",
          "source_commit": "cc2700db57bb83ee13dc31ebe1b868c254e15d09",
          "checkpoint_revision": "506474f2b33dc77c234d668cc419be1861899cad",
          "checkpoint_sha256": "c8149d988e4bbf3f347e6966b5d769de347a5f8c59ffca1dc4bd4bf5b8585e57",
          "fold": 0,
          "device": "cpu",
          "configuration": "fusion_stage3",
          "predict_dataset": "bvcc",
          "repetitions": 1,
          "extraction_seed": 20261012,
          "cpu_threads": 4,
          "remove_silent_section": false,
          "nnpack_enabled": false,
          "internal_sample_rate": 16000,
          "ssl_crop_seconds": 3.0,
          "spectrogram_crop_seconds": 1.4,
          "spectrogram_frames_per_repetition": 2,
          "spectrogram_inner_mixup": true,
          "short_input_extension": "tile",
          "code_license": "MIT",
          "weights_license": "MIT",
          "independent_human_rating_calibration": "not_performed_for_this_project",
          "scope": "JVS/LibriSpeechの再合成への適用は未校正。言語・発声別に集計する。",
          "paper_protocol_difference": "論文の5fold×5抽出平均ではなく、fold0と上記反復数。",
          "initialization": "official_architecture_then_complete_strict_state_dict",
          "packages": {
            "torch": "2.8.0+cpu",
            "torchaudio": "2.8.0+cpu",
            "torchvision": "0.23.0+cpu",
            "transformers": "4.51.3",
            "timm": "1.0.15",
            "librosa": "0.11.0",
            "numpy": "2.3.5",
            "utmosv2": "1.3.1.dev0"
          },
          "checkpoint": {
            "url": "https://huggingface.co/sarulab-speech/UTMOSv2/resolve/506474f2b33dc77c234d668cc419be1861899cad/fold0_s42_best_model.pth",
            "sha256": "c8149d988e4bbf3f347e6966b5d769de347a5f8c59ffca1dc4bd4bf5b8585e57",
            "bytes": 818531314
          },
          "feature_configuration": {
            "config.json": {
              "url": "https://huggingface.co/facebook/wav2vec2-base/resolve/0b5b8e868dd84f03fd87d01f9c4ff0f080fecfe8/config.json",
              "sha256": "4937977e24d12d1bba70cdce8709c3c04807a8e4ae8ddac4229c48c436ae99ae",
              "bytes": 1842
            },
            "preprocessor_config.json": {
              "url": "https://huggingface.co/facebook/wav2vec2-base/resolve/0b5b8e868dd84f03fd87d01f9c4ff0f080fecfe8/preprocessor_config.json",
              "sha256": "b225d617c025463b9e157e06afea8b90dc7078fc70b013c533328423e0486b4a",
              "bytes": 159
            }
          },
          "initialization_status": "ready"
        },
        "environment": {
          "python": "3.12.14",
          "platform": "Linux-6.18.44-x86_64-with-glibc2.41",
          "cpu": "",
          "logical_cpu_count": 9,
          "packages": {
            "numpy": "2.3.5",
            "scipy": "1.17.0",
            "torch": "2.8.0+cpu",
            "torchaudio": "2.8.0+cpu",
            "onnxruntime": "1.22.0",
            "speechbrain": "1.0.3",
            "utmosv2": "1.3.1.dev0",
            "transformers": "4.51.3",
            "timm": "1.0.15",
            "librosa": "0.11.0"
          }
        }
      },
      {
        "source": "runs/added-metrics-20261014/mos/batch-06/initialization.json",
        "seconds": 6.370313589999569,
        "provenance": {
          "name": "UTMOSv2",
          "quantity": "predicted_mos_not_human_rating",
          "source_repository": "https://github.com/sarulab-speech/UTMOSv2",
          "source_commit": "cc2700db57bb83ee13dc31ebe1b868c254e15d09",
          "checkpoint_revision": "506474f2b33dc77c234d668cc419be1861899cad",
          "checkpoint_sha256": "c8149d988e4bbf3f347e6966b5d769de347a5f8c59ffca1dc4bd4bf5b8585e57",
          "fold": 0,
          "device": "cpu",
          "configuration": "fusion_stage3",
          "predict_dataset": "bvcc",
          "repetitions": 1,
          "extraction_seed": 20261012,
          "cpu_threads": 4,
          "remove_silent_section": false,
          "nnpack_enabled": false,
          "internal_sample_rate": 16000,
          "ssl_crop_seconds": 3.0,
          "spectrogram_crop_seconds": 1.4,
          "spectrogram_frames_per_repetition": 2,
          "spectrogram_inner_mixup": true,
          "short_input_extension": "tile",
          "code_license": "MIT",
          "weights_license": "MIT",
          "independent_human_rating_calibration": "not_performed_for_this_project",
          "scope": "JVS/LibriSpeechの再合成への適用は未校正。言語・発声別に集計する。",
          "paper_protocol_difference": "論文の5fold×5抽出平均ではなく、fold0と上記反復数。",
          "initialization": "official_architecture_then_complete_strict_state_dict",
          "packages": {
            "torch": "2.8.0+cpu",
            "torchaudio": "2.8.0+cpu",
            "torchvision": "0.23.0+cpu",
            "transformers": "4.51.3",
            "timm": "1.0.15",
            "librosa": "0.11.0",
            "numpy": "2.3.5",
            "utmosv2": "1.3.1.dev0"
          },
          "checkpoint": {
            "url": "https://huggingface.co/sarulab-speech/UTMOSv2/resolve/506474f2b33dc77c234d668cc419be1861899cad/fold0_s42_best_model.pth",
            "sha256": "c8149d988e4bbf3f347e6966b5d769de347a5f8c59ffca1dc4bd4bf5b8585e57",
            "bytes": 818531314
          },
          "feature_configuration": {
            "config.json": {
              "url": "https://huggingface.co/facebook/wav2vec2-base/resolve/0b5b8e868dd84f03fd87d01f9c4ff0f080fecfe8/config.json",
              "sha256": "4937977e24d12d1bba70cdce8709c3c04807a8e4ae8ddac4229c48c436ae99ae",
              "bytes": 1842
            },
            "preprocessor_config.json": {
              "url": "https://huggingface.co/facebook/wav2vec2-base/resolve/0b5b8e868dd84f03fd87d01f9c4ff0f080fecfe8/preprocessor_config.json",
              "sha256": "b225d617c025463b9e157e06afea8b90dc7078fc70b013c533328423e0486b4a",
              "bytes": 159
            }
          },
          "initialization_status": "ready"
        },
        "environment": {
          "python": "3.12.14",
          "platform": "Linux-6.18.44-x86_64-with-glibc2.41",
          "cpu": "",
          "logical_cpu_count": 9,
          "packages": {
            "numpy": "2.3.5",
            "scipy": "1.17.0",
            "torch": "2.8.0+cpu",
            "torchaudio": "2.8.0+cpu",
            "onnxruntime": "1.22.0",
            "speechbrain": "1.0.3",
            "utmosv2": "1.3.1.dev0",
            "transformers": "4.51.3",
            "timm": "1.0.15",
            "librosa": "0.11.0"
          }
        }
      },
      {
        "source": "runs/added-metrics-20261014/mos/batch-07/initialization.json",
        "seconds": 6.414642003000154,
        "provenance": {
          "name": "UTMOSv2",
          "quantity": "predicted_mos_not_human_rating",
          "source_repository": "https://github.com/sarulab-speech/UTMOSv2",
          "source_commit": "cc2700db57bb83ee13dc31ebe1b868c254e15d09",
          "checkpoint_revision": "506474f2b33dc77c234d668cc419be1861899cad",
          "checkpoint_sha256": "c8149d988e4bbf3f347e6966b5d769de347a5f8c59ffca1dc4bd4bf5b8585e57",
          "fold": 0,
          "device": "cpu",
          "configuration": "fusion_stage3",
          "predict_dataset": "bvcc",
          "repetitions": 1,
          "extraction_seed": 20261012,
          "cpu_threads": 4,
          "remove_silent_section": false,
          "nnpack_enabled": false,
          "internal_sample_rate": 16000,
          "ssl_crop_seconds": 3.0,
          "spectrogram_crop_seconds": 1.4,
          "spectrogram_frames_per_repetition": 2,
          "spectrogram_inner_mixup": true,
          "short_input_extension": "tile",
          "code_license": "MIT",
          "weights_license": "MIT",
          "independent_human_rating_calibration": "not_performed_for_this_project",
          "scope": "JVS/LibriSpeechの再合成への適用は未校正。言語・発声別に集計する。",
          "paper_protocol_difference": "論文の5fold×5抽出平均ではなく、fold0と上記反復数。",
          "initialization": "official_architecture_then_complete_strict_state_dict",
          "packages": {
            "torch": "2.8.0+cpu",
            "torchaudio": "2.8.0+cpu",
            "torchvision": "0.23.0+cpu",
            "transformers": "4.51.3",
            "timm": "1.0.15",
            "librosa": "0.11.0",
            "numpy": "2.3.5",
            "utmosv2": "1.3.1.dev0"
          },
          "checkpoint": {
            "url": "https://huggingface.co/sarulab-speech/UTMOSv2/resolve/506474f2b33dc77c234d668cc419be1861899cad/fold0_s42_best_model.pth",
            "sha256": "c8149d988e4bbf3f347e6966b5d769de347a5f8c59ffca1dc4bd4bf5b8585e57",
            "bytes": 818531314
          },
          "feature_configuration": {
            "config.json": {
              "url": "https://huggingface.co/facebook/wav2vec2-base/resolve/0b5b8e868dd84f03fd87d01f9c4ff0f080fecfe8/config.json",
              "sha256": "4937977e24d12d1bba70cdce8709c3c04807a8e4ae8ddac4229c48c436ae99ae",
              "bytes": 1842
            },
            "preprocessor_config.json": {
              "url": "https://huggingface.co/facebook/wav2vec2-base/resolve/0b5b8e868dd84f03fd87d01f9c4ff0f080fecfe8/preprocessor_config.json",
              "sha256": "b225d617c025463b9e157e06afea8b90dc7078fc70b013c533328423e0486b4a",
              "bytes": 159
            }
          },
          "initialization_status": "ready"
        },
        "environment": {
          "python": "3.12.14",
          "platform": "Linux-6.18.44-x86_64-with-glibc2.41",
          "cpu": "",
          "logical_cpu_count": 9,
          "packages": {
            "numpy": "2.3.5",
            "scipy": "1.17.0",
            "torch": "2.8.0+cpu",
            "torchaudio": "2.8.0+cpu",
            "onnxruntime": "1.22.0",
            "speechbrain": "1.0.3",
            "utmosv2": "1.3.1.dev0",
            "transformers": "4.51.3",
            "timm": "1.0.15",
            "librosa": "0.11.0"
          }
        }
      },
      {
        "source": "runs/added-metrics-20261014/mos/batch-08/initialization.json",
        "seconds": 6.136912787000256,
        "provenance": {
          "name": "UTMOSv2",
          "quantity": "predicted_mos_not_human_rating",
          "source_repository": "https://github.com/sarulab-speech/UTMOSv2",
          "source_commit": "cc2700db57bb83ee13dc31ebe1b868c254e15d09",
          "checkpoint_revision": "506474f2b33dc77c234d668cc419be1861899cad",
          "checkpoint_sha256": "c8149d988e4bbf3f347e6966b5d769de347a5f8c59ffca1dc4bd4bf5b8585e57",
          "fold": 0,
          "device": "cpu",
          "configuration": "fusion_stage3",
          "predict_dataset": "bvcc",
          "repetitions": 1,
          "extraction_seed": 20261012,
          "cpu_threads": 4,
          "remove_silent_section": false,
          "nnpack_enabled": false,
          "internal_sample_rate": 16000,
          "ssl_crop_seconds": 3.0,
          "spectrogram_crop_seconds": 1.4,
          "spectrogram_frames_per_repetition": 2,
          "spectrogram_inner_mixup": true,
          "short_input_extension": "tile",
          "code_license": "MIT",
          "weights_license": "MIT",
          "independent_human_rating_calibration": "not_performed_for_this_project",
          "scope": "JVS/LibriSpeechの再合成への適用は未校正。言語・発声別に集計する。",
          "paper_protocol_difference": "論文の5fold×5抽出平均ではなく、fold0と上記反復数。",
          "initialization": "official_architecture_then_complete_strict_state_dict",
          "packages": {
            "torch": "2.8.0+cpu",
            "torchaudio": "2.8.0+cpu",
            "torchvision": "0.23.0+cpu",
            "transformers": "4.51.3",
            "timm": "1.0.15",
            "librosa": "0.11.0",
            "numpy": "2.3.5",
            "utmosv2": "1.3.1.dev0"
          },
          "checkpoint": {
            "url": "https://huggingface.co/sarulab-speech/UTMOSv2/resolve/506474f2b33dc77c234d668cc419be1861899cad/fold0_s42_best_model.pth",
            "sha256": "c8149d988e4bbf3f347e6966b5d769de347a5f8c59ffca1dc4bd4bf5b8585e57",
            "bytes": 818531314
          },
          "feature_configuration": {
            "config.json": {
              "url": "https://huggingface.co/facebook/wav2vec2-base/resolve/0b5b8e868dd84f03fd87d01f9c4ff0f080fecfe8/config.json",
              "sha256": "4937977e24d12d1bba70cdce8709c3c04807a8e4ae8ddac4229c48c436ae99ae",
              "bytes": 1842
            },
            "preprocessor_config.json": {
              "url": "https://huggingface.co/facebook/wav2vec2-base/resolve/0b5b8e868dd84f03fd87d01f9c4ff0f080fecfe8/preprocessor_config.json",
              "sha256": "b225d617c025463b9e157e06afea8b90dc7078fc70b013c533328423e0486b4a",
              "bytes": 159
            }
          },
          "initialization_status": "ready"
        },
        "environment": {
          "python": "3.12.14",
          "platform": "Linux-6.18.44-x86_64-with-glibc2.41",
          "cpu": "",
          "logical_cpu_count": 9,
          "packages": {
            "numpy": "2.3.5",
            "scipy": "1.17.0",
            "torch": "2.8.0+cpu",
            "torchaudio": "2.8.0+cpu",
            "onnxruntime": "1.22.0",
            "speechbrain": "1.0.3",
            "utmosv2": "1.3.1.dev0",
            "transformers": "4.51.3",
            "timm": "1.0.15",
            "librosa": "0.11.0"
          }
        }
      },
      {
        "source": "runs/added-metrics-20261014/mos/batch-09/initialization.json",
        "seconds": 6.77110944099968,
        "provenance": {
          "name": "UTMOSv2",
          "quantity": "predicted_mos_not_human_rating",
          "source_repository": "https://github.com/sarulab-speech/UTMOSv2",
          "source_commit": "cc2700db57bb83ee13dc31ebe1b868c254e15d09",
          "checkpoint_revision": "506474f2b33dc77c234d668cc419be1861899cad",
          "checkpoint_sha256": "c8149d988e4bbf3f347e6966b5d769de347a5f8c59ffca1dc4bd4bf5b8585e57",
          "fold": 0,
          "device": "cpu",
          "configuration": "fusion_stage3",
          "predict_dataset": "bvcc",
          "repetitions": 1,
          "extraction_seed": 20261012,
          "cpu_threads": 4,
          "remove_silent_section": false,
          "nnpack_enabled": false,
          "internal_sample_rate": 16000,
          "ssl_crop_seconds": 3.0,
          "spectrogram_crop_seconds": 1.4,
          "spectrogram_frames_per_repetition": 2,
          "spectrogram_inner_mixup": true,
          "short_input_extension": "tile",
          "code_license": "MIT",
          "weights_license": "MIT",
          "independent_human_rating_calibration": "not_performed_for_this_project",
          "scope": "JVS/LibriSpeechの再合成への適用は未校正。言語・発声別に集計する。",
          "paper_protocol_difference": "論文の5fold×5抽出平均ではなく、fold0と上記反復数。",
          "initialization": "official_architecture_then_complete_strict_state_dict",
          "packages": {
            "torch": "2.8.0+cpu",
            "torchaudio": "2.8.0+cpu",
            "torchvision": "0.23.0+cpu",
            "transformers": "4.51.3",
            "timm": "1.0.15",
            "librosa": "0.11.0",
            "numpy": "2.3.5",
            "utmosv2": "1.3.1.dev0"
          },
          "checkpoint": {
            "url": "https://huggingface.co/sarulab-speech/UTMOSv2/resolve/506474f2b33dc77c234d668cc419be1861899cad/fold0_s42_best_model.pth",
            "sha256": "c8149d988e4bbf3f347e6966b5d769de347a5f8c59ffca1dc4bd4bf5b8585e57",
            "bytes": 818531314
          },
          "feature_configuration": {
            "config.json": {
              "url": "https://huggingface.co/facebook/wav2vec2-base/resolve/0b5b8e868dd84f03fd87d01f9c4ff0f080fecfe8/config.json",
              "sha256": "4937977e24d12d1bba70cdce8709c3c04807a8e4ae8ddac4229c48c436ae99ae",
              "bytes": 1842
            },
            "preprocessor_config.json": {
              "url": "https://huggingface.co/facebook/wav2vec2-base/resolve/0b5b8e868dd84f03fd87d01f9c4ff0f080fecfe8/preprocessor_config.json",
              "sha256": "b225d617c025463b9e157e06afea8b90dc7078fc70b013c533328423e0486b4a",
              "bytes": 159
            }
          },
          "initialization_status": "ready"
        },
        "environment": {
          "python": "3.12.14",
          "platform": "Linux-6.18.44-x86_64-with-glibc2.41",
          "cpu": "",
          "logical_cpu_count": 9,
          "packages": {
            "numpy": "2.3.5",
            "scipy": "1.17.0",
            "torch": "2.8.0+cpu",
            "torchaudio": "2.8.0+cpu",
            "onnxruntime": "1.22.0",
            "speechbrain": "1.0.3",
            "utmosv2": "1.3.1.dev0",
            "transformers": "4.51.3",
            "timm": "1.0.15",
            "librosa": "0.11.0"
          }
        }
      },
      {
        "source": "runs/added-metrics-20261014/mos/batch-10/initialization.json",
        "seconds": 6.443110997999611,
        "provenance": {
          "name": "UTMOSv2",
          "quantity": "predicted_mos_not_human_rating",
          "source_repository": "https://github.com/sarulab-speech/UTMOSv2",
          "source_commit": "cc2700db57bb83ee13dc31ebe1b868c254e15d09",
          "checkpoint_revision": "506474f2b33dc77c234d668cc419be1861899cad",
          "checkpoint_sha256": "c8149d988e4bbf3f347e6966b5d769de347a5f8c59ffca1dc4bd4bf5b8585e57",
          "fold": 0,
          "device": "cpu",
          "configuration": "fusion_stage3",
          "predict_dataset": "bvcc",
          "repetitions": 1,
          "extraction_seed": 20261012,
          "cpu_threads": 4,
          "remove_silent_section": false,
          "nnpack_enabled": false,
          "internal_sample_rate": 16000,
          "ssl_crop_seconds": 3.0,
          "spectrogram_crop_seconds": 1.4,
          "spectrogram_frames_per_repetition": 2,
          "spectrogram_inner_mixup": true,
          "short_input_extension": "tile",
          "code_license": "MIT",
          "weights_license": "MIT",
          "independent_human_rating_calibration": "not_performed_for_this_project",
          "scope": "JVS/LibriSpeechの再合成への適用は未校正。言語・発声別に集計する。",
          "paper_protocol_difference": "論文の5fold×5抽出平均ではなく、fold0と上記反復数。",
          "initialization": "official_architecture_then_complete_strict_state_dict",
          "packages": {
            "torch": "2.8.0+cpu",
            "torchaudio": "2.8.0+cpu",
            "torchvision": "0.23.0+cpu",
            "transformers": "4.51.3",
            "timm": "1.0.15",
            "librosa": "0.11.0",
            "numpy": "2.3.5",
            "utmosv2": "1.3.1.dev0"
          },
          "checkpoint": {
            "url": "https://huggingface.co/sarulab-speech/UTMOSv2/resolve/506474f2b33dc77c234d668cc419be1861899cad/fold0_s42_best_model.pth",
            "sha256": "c8149d988e4bbf3f347e6966b5d769de347a5f8c59ffca1dc4bd4bf5b8585e57",
            "bytes": 818531314
          },
          "feature_configuration": {
            "config.json": {
              "url": "https://huggingface.co/facebook/wav2vec2-base/resolve/0b5b8e868dd84f03fd87d01f9c4ff0f080fecfe8/config.json",
              "sha256": "4937977e24d12d1bba70cdce8709c3c04807a8e4ae8ddac4229c48c436ae99ae",
              "bytes": 1842
            },
            "preprocessor_config.json": {
              "url": "https://huggingface.co/facebook/wav2vec2-base/resolve/0b5b8e868dd84f03fd87d01f9c4ff0f080fecfe8/preprocessor_config.json",
              "sha256": "b225d617c025463b9e157e06afea8b90dc7078fc70b013c533328423e0486b4a",
              "bytes": 159
            }
          },
          "initialization_status": "ready"
        },
        "environment": {
          "python": "3.12.14",
          "platform": "Linux-6.18.44-x86_64-with-glibc2.41",
          "cpu": "",
          "logical_cpu_count": 9,
          "packages": {
            "numpy": "2.3.5",
            "scipy": "1.17.0",
            "torch": "2.8.0+cpu",
            "torchaudio": "2.8.0+cpu",
            "onnxruntime": "1.22.0",
            "speechbrain": "1.0.3",
            "utmosv2": "1.3.1.dev0",
            "transformers": "4.51.3",
            "timm": "1.0.15",
            "librosa": "0.11.0"
          }
        }
      },
      {
        "source": "runs/added-metrics-20261014/mos/batch-11/initialization.json",
        "seconds": 6.268164124999203,
        "provenance": {
          "name": "UTMOSv2",
          "quantity": "predicted_mos_not_human_rating",
          "source_repository": "https://github.com/sarulab-speech/UTMOSv2",
          "source_commit": "cc2700db57bb83ee13dc31ebe1b868c254e15d09",
          "checkpoint_revision": "506474f2b33dc77c234d668cc419be1861899cad",
          "checkpoint_sha256": "c8149d988e4bbf3f347e6966b5d769de347a5f8c59ffca1dc4bd4bf5b8585e57",
          "fold": 0,
          "device": "cpu",
          "configuration": "fusion_stage3",
          "predict_dataset": "bvcc",
          "repetitions": 1,
          "extraction_seed": 20261012,
          "cpu_threads": 4,
          "remove_silent_section": false,
          "nnpack_enabled": false,
          "internal_sample_rate": 16000,
          "ssl_crop_seconds": 3.0,
          "spectrogram_crop_seconds": 1.4,
          "spectrogram_frames_per_repetition": 2,
          "spectrogram_inner_mixup": true,
          "short_input_extension": "tile",
          "code_license": "MIT",
          "weights_license": "MIT",
          "independent_human_rating_calibration": "not_performed_for_this_project",
          "scope": "JVS/LibriSpeechの再合成への適用は未校正。言語・発声別に集計する。",
          "paper_protocol_difference": "論文の5fold×5抽出平均ではなく、fold0と上記反復数。",
          "initialization": "official_architecture_then_complete_strict_state_dict",
          "packages": {
            "torch": "2.8.0+cpu",
            "torchaudio": "2.8.0+cpu",
            "torchvision": "0.23.0+cpu",
            "transformers": "4.51.3",
            "timm": "1.0.15",
            "librosa": "0.11.0",
            "numpy": "2.3.5",
            "utmosv2": "1.3.1.dev0"
          },
          "checkpoint": {
            "url": "https://huggingface.co/sarulab-speech/UTMOSv2/resolve/506474f2b33dc77c234d668cc419be1861899cad/fold0_s42_best_model.pth",
            "sha256": "c8149d988e4bbf3f347e6966b5d769de347a5f8c59ffca1dc4bd4bf5b8585e57",
            "bytes": 818531314
          },
          "feature_configuration": {
            "config.json": {
              "url": "https://huggingface.co/facebook/wav2vec2-base/resolve/0b5b8e868dd84f03fd87d01f9c4ff0f080fecfe8/config.json",
              "sha256": "4937977e24d12d1bba70cdce8709c3c04807a8e4ae8ddac4229c48c436ae99ae",
              "bytes": 1842
            },
            "preprocessor_config.json": {
              "url": "https://huggingface.co/facebook/wav2vec2-base/resolve/0b5b8e868dd84f03fd87d01f9c4ff0f080fecfe8/preprocessor_config.json",
              "sha256": "b225d617c025463b9e157e06afea8b90dc7078fc70b013c533328423e0486b4a",
              "bytes": 159
            }
          },
          "initialization_status": "ready"
        },
        "environment": {
          "python": "3.12.14",
          "platform": "Linux-6.18.44-x86_64-with-glibc2.41",
          "cpu": "",
          "logical_cpu_count": 9,
          "packages": {
            "numpy": "2.3.5",
            "scipy": "1.17.0",
            "torch": "2.8.0+cpu",
            "torchaudio": "2.8.0+cpu",
            "onnxruntime": "1.22.0",
            "speechbrain": "1.0.3",
            "utmosv2": "1.3.1.dev0",
            "transformers": "4.51.3",
            "timm": "1.0.15",
            "librosa": "0.11.0"
          }
        }
      }
    ],
    "initialization_recorded_seconds": 84.54591321899898
  },
  "quality": {
    "first_wall_seconds": 597.5332919320181,
    "repeat_wall_seconds": 574.638993208996,
    "initialization_seconds": 48.3202147710017
  },
  "synthesis": {
    "analysis_seconds": 265.47181602100045,
    "decode_wall_seconds": 305.5550459199976,
    "execution": {
      "started_at_utc": "2026-10-13T16:12:55.407846+00:00",
      "finished_at_utc": "2026-10-13T16:23:35.259724+00:00",
      "saved_outputs": 1200,
      "planned_outputs": 1200
    },
    "elapsed_wall_seconds": 639.851878
  },
  "asr": {
    "original": {
      "count": 55,
      "audio_seconds": 225.935,
      "inference_wall_seconds": 47.293222766003055,
      "inference_process_cpu_seconds": 66.60619294799994,
      "inference_real_time_factor": 0.209322250939443,
      "per_waveform_wall_seconds": {
        "count": 55,
        "mean": 0.8598767775636919,
        "median": 0.7258015839997825,
        "min": 0.2892741890000252,
        "max": 2.5560034919999453
      }
    },
    "hnm": {
      "count": 55,
      "audio_seconds": 225.935,
      "inference_wall_seconds": 50.58416748799846,
      "inference_process_cpu_seconds": 67.55157825599996,
      "inference_real_time_factor": 0.22388814255426762,
      "per_waveform_wall_seconds": {
        "count": 55,
        "mean": 0.9197121361454265,
        "median": 0.6685842580000099,
        "min": 0.3060445009996329,
        "max": 3.3895693220001704
      }
    },
    "world_harvest": {
      "count": 55,
      "audio_seconds": 225.935,
      "inference_wall_seconds": 44.27901779199783,
      "inference_process_cpu_seconds": 63.85978081600007,
      "inference_real_time_factor": 0.19598122376788824,
      "per_waveform_wall_seconds": {
        "count": 55,
        "mean": 0.8050730507635969,
        "median": 0.653488531000221,
        "min": 0.2686074430002918,
        "max": 2.4574500849994365
      }
    },
    "assessment_first_wall_seconds": 0.30167030199663714,
    "assessment_repeat_wall_seconds": 0.27746703100274317
  },
  "assessment_status_counts": {
    "ok": 1920
  },
  "note": "同時実行・分割初期化を含む記録。各工程の和は全体経過時間ではない。初回＋1反復で速度安定性は主張しない。",
  "assessment": {
    "initialization_scope": "assessment本体は学習モデルを初期化しない。品質点は保存済み証跡。",
    "first_wall_seconds": 118.79901729399262,
    "repeat_wall_seconds": 114.91574849200333,
    "repetitions": 1
  }
}