{"previous": {"status": "measured", "value": 3.1964282989501953, "predicted_mos": 3.1964282989501953, "model": "UTMOSv2", "language": "ja", "style": "falsetto", "domain_validation": "未校正の探索的予測", "extraction_seed": 20261012, "repetitions": 1, "input_samples": 108438, "input_sample_rate": 24000, "input_seconds": 4.51825, "input_float32_sha256": "84f43b9942c00905fb27154315308069f5cf2a633d5da28c545d07f9d98f65fe", "internal_samples": 72292, "ssl_processed_samples": 48000, "spectrogram_processed_samples_per_frame": 22400, "ssl_tile_repetitions": 1, "spectrogram_tile_repetitions": 1, "reason": null, "seconds": 43.36582119995728}, "observed": {"status": "measured", "value": 3.196427822113037, "predicted_mos": 3.196427822113037, "model": "UTMOSv2", "language": "ja", "style": "falsetto", "domain_validation": "未校正の探索的予測", "extraction_seed": 20261012, "repetitions": 1, "input_samples": 108438, "input_sample_rate": 24000, "input_seconds": 4.51825, "input_float32_sha256": "84f43b9942c00905fb27154315308069f5cf2a633d5da28c545d07f9d98f65fe", "internal_samples": 72292, "ssl_processed_samples": 48000, "spectrogram_processed_samples_per_frame": 22400, "ssl_tile_repetitions": 1, "spectrogram_tile_repetitions": 1, "reason": null, "seconds": 23.87259740009904}, "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"}, "wall_seconds": 23.87409370020032, "waveform_sha256": "d057493783316e914981a3b3ab5d87910159e19ac7543c9b841d5883219a7a3d", "note": "CPU実行条件の観測。候補の選択には用いない。"}
