{ "schemaVersion": 1, "engineFamily": "ml-engine-asr", "published": { "sanitizedAt": "2026-09-20T16:26:27.059Z", "sourceFileName": "ml-engine-mac-71e4cdd00-2026-09-20_14-34-47.json", "note": "Paths, transcripts, and local machine identifiers removed for public hosting." }, "environment": {}, "runConfig": { "provider": "cpu", "numThreads": 8, "languageFromCase": true, "notes": "Default matches App file-transcription ml-engine path (CPU). Override numThreads only via CLI if needed." }, "manifestPath": "manifest.json", "modelsPath": "models.json", "modelRuns": [ { "modelId": "en-zipformer-onnx", "modelArch": "transducer", "batch": { "engine": "ml-engine-asr", "results": [ { "engine": "ml-engine-asr", "model": "transducer", "caseId": "audio-benchmark", "segmentCount": 15, "errorMetricKind": "wer", "wer": 0.043, "cer": null, "insertionRate": 0.0108, "deletionRate": 0, "substitutionRate": 0.0323, "duplicateTokenRate": 0, "audioDurationMs": 34536, "preprocessElapsedMs": 0, "transcribeElapsedMs": 1785, "totalElapsedMs": 1785, "realTimeFactor": 0.0517, "sceneHint": "general" }, { "engine": "ml-engine-asr", "model": "transducer", "caseId": "en-long-sentence-pause", "segmentCount": 33, "errorMetricKind": "wer", "wer": 0.1133, "cer": null, "insertionRate": 0.0099, "deletionRate": 0, "substitutionRate": 0.1034, "duplicateTokenRate": 0, "audioDurationMs": 70992, "preprocessElapsedMs": 0, "transcribeElapsedMs": 2587, "totalElapsedMs": 2587, "realTimeFactor": 0.0364, "sceneHint": "speech" }, { "engine": "ml-engine-asr", "model": "transducer", "caseId": "en-oral-correction", "segmentCount": 35, "errorMetricKind": "wer", "wer": 0.0941, "cer": null, "insertionRate": 0.0248, "deletionRate": 0.0198, "substitutionRate": 0.0495, "duplicateTokenRate": 0.0148, "audioDurationMs": 68448, "preprocessElapsedMs": 0, "transcribeElapsedMs": 2461, "totalElapsedMs": 2461, "realTimeFactor": 0.036, "sceneHint": "dialogue" }, { "engine": "ml-engine-asr", "model": "transducer", "caseId": "en-tech-talk-podcast", "segmentCount": 26, "errorMetricKind": "wer", "wer": 0.1018, "cer": null, "insertionRate": 0.024, "deletionRate": 0.012, "substitutionRate": 0.0659, "duplicateTokenRate": 0.0059, "audioDurationMs": 79680, "preprocessElapsedMs": 0, "transcribeElapsedMs": 2093, "totalElapsedMs": 2093, "realTimeFactor": 0.0263, "sceneHint": "course" } ], "failures": [], "summary": { "caseCount": 4, "successCount": 4, "crashCount": 0, "crashRate": 0, "averageWer": 0.0881, "averageCer": null, "averageInsertionRate": 0.0174, "averageDeletionRate": 0.008, "averageSubstitutionRate": 0.0628, "averageDuplicateTokenRate": 0.0052, "averagePreprocessElapsedMs": 0, "averageTotalElapsedMs": 2231.5, "averageRealTimeFactor": 0.0376, "byScene": { "general": { "caseCount": 1, "successCount": 1, "crashCount": 0, "crashRate": 0, "averageWer": 0.043, "averageCer": null, "averageInsertionRate": 0.0108, "averageDeletionRate": 0, "averageSubstitutionRate": 0.0323, "averageDuplicateTokenRate": 0, "averagePreprocessElapsedMs": 0, "averageTotalElapsedMs": 1785, "averageRealTimeFactor": 0.0517 }, "speech": { "caseCount": 1, "successCount": 1, "crashCount": 0, "crashRate": 0, "averageWer": 0.1133, "averageCer": null, "averageInsertionRate": 0.0099, "averageDeletionRate": 0, "averageSubstitutionRate": 0.1034, "averageDuplicateTokenRate": 0, "averagePreprocessElapsedMs": 0, "averageTotalElapsedMs": 2587, "averageRealTimeFactor": 0.0364 }, "dialogue": { "caseCount": 1, "successCount": 1, "crashCount": 0, "crashRate": 0, "averageWer": 0.0941, "averageCer": null, "averageInsertionRate": 0.0248, "averageDeletionRate": 0.0198, "averageSubstitutionRate": 0.0495, "averageDuplicateTokenRate": 0.0148, "averagePreprocessElapsedMs": 0, "averageTotalElapsedMs": 2461, "averageRealTimeFactor": 0.036 }, "course": { "caseCount": 1, "successCount": 1, "crashCount": 0, "crashRate": 0, "averageWer": 0.1018, "averageCer": null, "averageInsertionRate": 0.024, "averageDeletionRate": 0.012, "averageSubstitutionRate": 0.0659, "averageDuplicateTokenRate": 0.0059, "averagePreprocessElapsedMs": 0, "averageTotalElapsedMs": 2093, "averageRealTimeFactor": 0.0263 } } } } }, { "modelId": "zh-en-zipformer-onnx", "modelArch": "transducer", "batch": { "engine": "ml-engine-asr", "results": [ { "engine": "ml-engine-asr", "model": "transducer", "caseId": "audio-benchmark", "segmentCount": 15, "errorMetricKind": "wer", "wer": 0.0645, "cer": null, "insertionRate": 0.0215, "deletionRate": 0, "substitutionRate": 0.043, "duplicateTokenRate": 0, "audioDurationMs": 34536, "preprocessElapsedMs": 0, "transcribeElapsedMs": 1215, "totalElapsedMs": 1215, "realTimeFactor": 0.0352, "sceneHint": "general" }, { "engine": "ml-engine-asr", "model": "transducer", "caseId": "en-long-sentence-pause", "segmentCount": 33, "errorMetricKind": "wer", "wer": 0.0788, "cer": null, "insertionRate": 0.0099, "deletionRate": 0.0049, "substitutionRate": 0.064, "duplicateTokenRate": 0, "audioDurationMs": 70992, "preprocessElapsedMs": 0, "transcribeElapsedMs": 2199, "totalElapsedMs": 2199, "realTimeFactor": 0.031, "sceneHint": "speech" }, { "engine": "ml-engine-asr", "model": "transducer", "caseId": "en-oral-correction", "segmentCount": 35, "errorMetricKind": "wer", "wer": 0.1188, "cer": null, "insertionRate": 0.0248, "deletionRate": 0.0248, "substitutionRate": 0.0693, "duplicateTokenRate": 0.0099, "audioDurationMs": 68448, "preprocessElapsedMs": 0, "transcribeElapsedMs": 1860, "totalElapsedMs": 1860, "realTimeFactor": 0.0272, "sceneHint": "dialogue" }, { "engine": "ml-engine-asr", "model": "transducer", "caseId": "en-tech-talk-podcast", "segmentCount": 26, "errorMetricKind": "wer", "wer": 0.1317, "cer": null, "insertionRate": 0.024, "deletionRate": 0.018, "substitutionRate": 0.0898, "duplicateTokenRate": 0.006, "audioDurationMs": 79680, "preprocessElapsedMs": 0, "transcribeElapsedMs": 2205, "totalElapsedMs": 2205, "realTimeFactor": 0.0277, "sceneHint": "course" }, { "engine": "ml-engine-asr", "model": "transducer", "caseId": "zh-tech-podcast", "segmentCount": 29, "errorMetricKind": "cer", "wer": null, "cer": 0.0226, "insertionRate": 0, "deletionRate": 0, "substitutionRate": 0.0226, "duplicateTokenRate": 0.125, "audioDurationMs": 74664, "preprocessElapsedMs": 0, "transcribeElapsedMs": 1829, "totalElapsedMs": 1829, "realTimeFactor": 0.0245, "sceneHint": "course" } ], "failures": [], "summary": { "caseCount": 5, "successCount": 5, "crashCount": 0, "crashRate": 0, "averageWer": 0.0985, "averageCer": 0.0226, "averageInsertionRate": 0.016, "averageDeletionRate": 0.0095, "averageSubstitutionRate": 0.0577, "averageDuplicateTokenRate": 0.0282, "averagePreprocessElapsedMs": 0, "averageTotalElapsedMs": 1861.6, "averageRealTimeFactor": 0.0291, "byScene": { "general": { "caseCount": 1, "successCount": 1, "crashCount": 0, "crashRate": 0, "averageWer": 0.0645, "averageCer": null, "averageInsertionRate": 0.0215, "averageDeletionRate": 0, "averageSubstitutionRate": 0.043, "averageDuplicateTokenRate": 0, "averagePreprocessElapsedMs": 0, "averageTotalElapsedMs": 1215, "averageRealTimeFactor": 0.0352 }, "speech": { "caseCount": 1, "successCount": 1, "crashCount": 0, "crashRate": 0, "averageWer": 0.0788, "averageCer": null, "averageInsertionRate": 0.0099, "averageDeletionRate": 0.0049, "averageSubstitutionRate": 0.064, "averageDuplicateTokenRate": 0, "averagePreprocessElapsedMs": 0, "averageTotalElapsedMs": 2199, "averageRealTimeFactor": 0.031 }, "dialogue": { "caseCount": 1, "successCount": 1, "crashCount": 0, "crashRate": 0, "averageWer": 0.1188, "averageCer": null, "averageInsertionRate": 0.0248, "averageDeletionRate": 0.0248, "averageSubstitutionRate": 0.0693, "averageDuplicateTokenRate": 0.0099, "averagePreprocessElapsedMs": 0, "averageTotalElapsedMs": 1860, "averageRealTimeFactor": 0.0272 }, "course": { "caseCount": 2, "successCount": 2, "crashCount": 0, "crashRate": 0, "averageWer": 0.1317, "averageCer": 0.0226, "averageInsertionRate": 0.012, "averageDeletionRate": 0.009, "averageSubstitutionRate": 0.0562, "averageDuplicateTokenRate": 0.0655, "averagePreprocessElapsedMs": 0, "averageTotalElapsedMs": 2017, "averageRealTimeFactor": 0.0261 } } } } }, { "modelId": "qwen3-asr", "modelArch": "qwen3-asr", "batch": { "engine": "ml-engine-asr", "results": [ { "engine": "ml-engine-asr", "model": "qwen3-asr", "caseId": "audio-benchmark", "segmentCount": 15, "errorMetricKind": "wer", "wer": 0, "cer": null, "insertionRate": 0, "deletionRate": 0, "substitutionRate": 0, "duplicateTokenRate": 0, "audioDurationMs": 34536, "preprocessElapsedMs": 0, "transcribeElapsedMs": 10357, "totalElapsedMs": 10357, "realTimeFactor": 0.2999, "sceneHint": "general" }, { "engine": "ml-engine-asr", "model": "qwen3-asr", "caseId": "en-long-sentence-pause", "segmentCount": 33, "errorMetricKind": "wer", "wer": 0, "cer": null, "insertionRate": 0, "deletionRate": 0, "substitutionRate": 0, "duplicateTokenRate": 0, "audioDurationMs": 70992, "preprocessElapsedMs": 0, "transcribeElapsedMs": 16294, "totalElapsedMs": 16294, "realTimeFactor": 0.2295, "sceneHint": "speech" }, { "engine": "ml-engine-asr", "model": "qwen3-asr", "caseId": "en-oral-correction", "segmentCount": 35, "errorMetricKind": "wer", "wer": 0.0297, "cer": null, "insertionRate": 0, "deletionRate": 0.0248, "substitutionRate": 0.005, "duplicateTokenRate": 0.0254, "audioDurationMs": 68448, "preprocessElapsedMs": 0, "transcribeElapsedMs": 16743, "totalElapsedMs": 16743, "realTimeFactor": 0.2446, "sceneHint": "dialogue" }, { "engine": "ml-engine-asr", "model": "qwen3-asr", "caseId": "en-tech-talk-podcast", "segmentCount": 26, "errorMetricKind": "wer", "wer": 0.018, "cer": null, "insertionRate": 0, "deletionRate": 0, "substitutionRate": 0.018, "duplicateTokenRate": 0.006, "audioDurationMs": 79680, "preprocessElapsedMs": 0, "transcribeElapsedMs": 13494, "totalElapsedMs": 13494, "realTimeFactor": 0.1694, "sceneHint": "course" }, { "engine": "ml-engine-asr", "model": "qwen3-asr", "caseId": "zh-tech-podcast", "segmentCount": 29, "errorMetricKind": "cer", "wer": null, "cer": 0.0065, "insertionRate": 0, "deletionRate": 0, "substitutionRate": 0.0065, "duplicateTokenRate": 0.8333, "audioDurationMs": 74664, "preprocessElapsedMs": 0, "transcribeElapsedMs": 14132, "totalElapsedMs": 14132, "realTimeFactor": 0.1893, "sceneHint": "course" } ], "failures": [], "summary": { "caseCount": 5, "successCount": 5, "crashCount": 0, "crashRate": 0, "averageWer": 0.0119, "averageCer": 0.0065, "averageInsertionRate": 0, "averageDeletionRate": 0.005, "averageSubstitutionRate": 0.0059, "averageDuplicateTokenRate": 0.1729, "averagePreprocessElapsedMs": 0, "averageTotalElapsedMs": 14204, "averageRealTimeFactor": 0.2265, "byScene": { "general": { "caseCount": 1, "successCount": 1, "crashCount": 0, "crashRate": 0, "averageWer": 0, "averageCer": null, "averageInsertionRate": 0, "averageDeletionRate": 0, "averageSubstitutionRate": 0, "averageDuplicateTokenRate": 0, "averagePreprocessElapsedMs": 0, "averageTotalElapsedMs": 10357, "averageRealTimeFactor": 0.2999 }, "speech": { "caseCount": 1, "successCount": 1, "crashCount": 0, "crashRate": 0, "averageWer": 0, "averageCer": null, "averageInsertionRate": 0, "averageDeletionRate": 0, "averageSubstitutionRate": 0, "averageDuplicateTokenRate": 0, "averagePreprocessElapsedMs": 0, "averageTotalElapsedMs": 16294, "averageRealTimeFactor": 0.2295 }, "dialogue": { "caseCount": 1, "successCount": 1, "crashCount": 0, "crashRate": 0, "averageWer": 0.0297, "averageCer": null, "averageInsertionRate": 0, "averageDeletionRate": 0.0248, "averageSubstitutionRate": 0.005, "averageDuplicateTokenRate": 0.0254, "averagePreprocessElapsedMs": 0, "averageTotalElapsedMs": 16743, "averageRealTimeFactor": 0.2446 }, "course": { "caseCount": 2, "successCount": 2, "crashCount": 0, "crashRate": 0, "averageWer": 0.018, "averageCer": 0.0065, "averageInsertionRate": 0, "averageDeletionRate": 0, "averageSubstitutionRate": 0.0122, "averageDuplicateTokenRate": 0.4197, "averagePreprocessElapsedMs": 0, "averageTotalElapsedMs": 13813, "averageRealTimeFactor": 0.1794 } } } } }, { "modelId": "parakeet", "modelArch": "parakeet", "batch": { "engine": "ml-engine-asr", "results": [ { "engine": "ml-engine-asr", "model": "parakeet", "caseId": "audio-benchmark", "segmentCount": 15, "errorMetricKind": "wer", "wer": 0, "cer": null, "insertionRate": 0, "deletionRate": 0, "substitutionRate": 0, "duplicateTokenRate": 0, "audioDurationMs": 34536, "preprocessElapsedMs": 0, "transcribeElapsedMs": 4139, "totalElapsedMs": 4139, "realTimeFactor": 0.1198, "sceneHint": "general" }, { "engine": "ml-engine-asr", "model": "parakeet", "caseId": "en-long-sentence-pause", "segmentCount": 33, "errorMetricKind": "wer", "wer": 0, "cer": null, "insertionRate": 0, "deletionRate": 0, "substitutionRate": 0, "duplicateTokenRate": 0, "audioDurationMs": 70992, "preprocessElapsedMs": 0, "transcribeElapsedMs": 7577, "totalElapsedMs": 7577, "realTimeFactor": 0.1067, "sceneHint": "speech" }, { "engine": "ml-engine-asr", "model": "parakeet", "caseId": "en-oral-correction", "segmentCount": 35, "errorMetricKind": "wer", "wer": 0.0446, "cer": null, "insertionRate": 0.005, "deletionRate": 0.0248, "substitutionRate": 0.0149, "duplicateTokenRate": 0.0253, "audioDurationMs": 68448, "preprocessElapsedMs": 0, "transcribeElapsedMs": 7109, "totalElapsedMs": 7109, "realTimeFactor": 0.1039, "sceneHint": "dialogue" }, { "engine": "ml-engine-asr", "model": "parakeet", "caseId": "en-tech-talk-podcast", "segmentCount": 26, "errorMetricKind": "wer", "wer": 0.024, "cer": null, "insertionRate": 0, "deletionRate": 0.018, "substitutionRate": 0.006, "duplicateTokenRate": 0.0061, "audioDurationMs": 79680, "preprocessElapsedMs": 0, "transcribeElapsedMs": 6551, "totalElapsedMs": 6551, "realTimeFactor": 0.0822, "sceneHint": "course" } ], "failures": [], "summary": { "caseCount": 4, "successCount": 4, "crashCount": 0, "crashRate": 0, "averageWer": 0.0171, "averageCer": null, "averageInsertionRate": 0.0013, "averageDeletionRate": 0.0107, "averageSubstitutionRate": 0.0052, "averageDuplicateTokenRate": 0.0078, "averagePreprocessElapsedMs": 0, "averageTotalElapsedMs": 6344, "averageRealTimeFactor": 0.1032, "byScene": { "general": { "caseCount": 1, "successCount": 1, "crashCount": 0, "crashRate": 0, "averageWer": 0, "averageCer": null, "averageInsertionRate": 0, "averageDeletionRate": 0, "averageSubstitutionRate": 0, "averageDuplicateTokenRate": 0, "averagePreprocessElapsedMs": 0, "averageTotalElapsedMs": 4139, "averageRealTimeFactor": 0.1198 }, "speech": { "caseCount": 1, "successCount": 1, "crashCount": 0, "crashRate": 0, "averageWer": 0, "averageCer": null, "averageInsertionRate": 0, "averageDeletionRate": 0, "averageSubstitutionRate": 0, "averageDuplicateTokenRate": 0, "averagePreprocessElapsedMs": 0, "averageTotalElapsedMs": 7577, "averageRealTimeFactor": 0.1067 }, "dialogue": { "caseCount": 1, "successCount": 1, "crashCount": 0, "crashRate": 0, "averageWer": 0.0446, "averageCer": null, "averageInsertionRate": 0.005, "averageDeletionRate": 0.0248, "averageSubstitutionRate": 0.0149, "averageDuplicateTokenRate": 0.0253, "averagePreprocessElapsedMs": 0, "averageTotalElapsedMs": 7109, "averageRealTimeFactor": 0.1039 }, "course": { "caseCount": 1, "successCount": 1, "crashCount": 0, "crashRate": 0, "averageWer": 0.024, "averageCer": null, "averageInsertionRate": 0, "averageDeletionRate": 0.018, "averageSubstitutionRate": 0.006, "averageDuplicateTokenRate": 0.0061, "averagePreprocessElapsedMs": 0, "averageTotalElapsedMs": 6551, "averageRealTimeFactor": 0.0822 } } } } }, { "modelId": "sense-voice", "modelArch": "sense-voice", "batch": { "engine": "ml-engine-asr", "results": [ { "engine": "ml-engine-asr", "model": "sense-voice", "caseId": "audio-benchmark", "segmentCount": 15, "errorMetricKind": "wer", "wer": 0, "cer": null, "insertionRate": 0, "deletionRate": 0, "substitutionRate": 0, "duplicateTokenRate": 0, "audioDurationMs": 34536, "preprocessElapsedMs": 0, "transcribeElapsedMs": 1833, "totalElapsedMs": 1833, "realTimeFactor": 0.0531, "sceneHint": "general" }, { "engine": "ml-engine-asr", "model": "sense-voice", "caseId": "en-long-sentence-pause", "segmentCount": 33, "errorMetricKind": "wer", "wer": 0.0148, "cer": null, "insertionRate": 0, "deletionRate": 0, "substitutionRate": 0.0148, "duplicateTokenRate": 0, "audioDurationMs": 70992, "preprocessElapsedMs": 0, "transcribeElapsedMs": 3065, "totalElapsedMs": 3065, "realTimeFactor": 0.0432, "sceneHint": "speech" }, { "engine": "ml-engine-asr", "model": "sense-voice", "caseId": "en-oral-correction", "segmentCount": 35, "errorMetricKind": "wer", "wer": 0.0743, "cer": null, "insertionRate": 0.005, "deletionRate": 0.0297, "substitutionRate": 0.0396, "duplicateTokenRate": 0.0254, "audioDurationMs": 68448, "preprocessElapsedMs": 0, "transcribeElapsedMs": 2885, "totalElapsedMs": 2885, "realTimeFactor": 0.0421, "sceneHint": "dialogue" }, { "engine": "ml-engine-asr", "model": "sense-voice", "caseId": "en-tech-talk-podcast", "segmentCount": 26, "errorMetricKind": "wer", "wer": 0.0359, "cer": null, "insertionRate": 0, "deletionRate": 0.006, "substitutionRate": 0.0299, "duplicateTokenRate": 0.006, "audioDurationMs": 79680, "preprocessElapsedMs": 0, "transcribeElapsedMs": 2832, "totalElapsedMs": 2832, "realTimeFactor": 0.0355, "sceneHint": "course" }, { "engine": "ml-engine-asr", "model": "sense-voice", "caseId": "zh-tech-podcast", "segmentCount": 29, "errorMetricKind": "cer", "wer": null, "cer": 0.0161, "insertionRate": 0.0065, "deletionRate": 0, "substitutionRate": 0.0097, "duplicateTokenRate": 0.2857, "audioDurationMs": 74664, "preprocessElapsedMs": 0, "transcribeElapsedMs": 3366, "totalElapsedMs": 3366, "realTimeFactor": 0.0451, "sceneHint": "course" } ], "failures": [], "summary": { "caseCount": 5, "successCount": 5, "crashCount": 0, "crashRate": 0, "averageWer": 0.0313, "averageCer": 0.0161, "averageInsertionRate": 0.0023, "averageDeletionRate": 0.0071, "averageSubstitutionRate": 0.0188, "averageDuplicateTokenRate": 0.0634, "averagePreprocessElapsedMs": 0, "averageTotalElapsedMs": 2796.2, "averageRealTimeFactor": 0.0438, "byScene": { "general": { "caseCount": 1, "successCount": 1, "crashCount": 0, "crashRate": 0, "averageWer": 0, "averageCer": null, "averageInsertionRate": 0, "averageDeletionRate": 0, "averageSubstitutionRate": 0, "averageDuplicateTokenRate": 0, "averagePreprocessElapsedMs": 0, "averageTotalElapsedMs": 1833, "averageRealTimeFactor": 0.0531 }, "speech": { "caseCount": 1, "successCount": 1, "crashCount": 0, "crashRate": 0, "averageWer": 0.0148, "averageCer": null, "averageInsertionRate": 0, "averageDeletionRate": 0, "averageSubstitutionRate": 0.0148, "averageDuplicateTokenRate": 0, "averagePreprocessElapsedMs": 0, "averageTotalElapsedMs": 3065, "averageRealTimeFactor": 0.0432 }, "dialogue": { "caseCount": 1, "successCount": 1, "crashCount": 0, "crashRate": 0, "averageWer": 0.0743, "averageCer": null, "averageInsertionRate": 0.005, "averageDeletionRate": 0.0297, "averageSubstitutionRate": 0.0396, "averageDuplicateTokenRate": 0.0254, "averagePreprocessElapsedMs": 0, "averageTotalElapsedMs": 2885, "averageRealTimeFactor": 0.0421 }, "course": { "caseCount": 2, "successCount": 2, "crashCount": 0, "crashRate": 0, "averageWer": 0.0359, "averageCer": 0.0161, "averageInsertionRate": 0.0032, "averageDeletionRate": 0.003, "averageSubstitutionRate": 0.0198, "averageDuplicateTokenRate": 0.1459, "averagePreprocessElapsedMs": 0, "averageTotalElapsedMs": 3099, "averageRealTimeFactor": 0.0403 } } } } }, { "modelId": "fire-red-ctc", "modelArch": "fire-red-ctc", "batch": { "engine": "ml-engine-asr", "results": [ { "engine": "ml-engine-asr", "model": "fire-red-ctc", "caseId": "audio-benchmark", "segmentCount": 15, "errorMetricKind": "wer", "wer": 0.0753, "cer": null, "insertionRate": 0.0108, "deletionRate": 0, "substitutionRate": 0.0645, "duplicateTokenRate": 0, "audioDurationMs": 34536, "preprocessElapsedMs": 0, "transcribeElapsedMs": 8113, "totalElapsedMs": 8113, "realTimeFactor": 0.2349, "sceneHint": "general" }, { "engine": "ml-engine-asr", "model": "fire-red-ctc", "caseId": "en-long-sentence-pause", "segmentCount": 33, "errorMetricKind": "wer", "wer": 0.1527, "cer": null, "insertionRate": 0.0099, "deletionRate": 0.0049, "substitutionRate": 0.1379, "duplicateTokenRate": 0, "audioDurationMs": 70992, "preprocessElapsedMs": 0, "transcribeElapsedMs": 15368, "totalElapsedMs": 15368, "realTimeFactor": 0.2165, "sceneHint": "speech" }, { "engine": "ml-engine-asr", "model": "fire-red-ctc", "caseId": "en-oral-correction", "segmentCount": 35, "errorMetricKind": "wer", "wer": 0.1089, "cer": null, "insertionRate": 0.0099, "deletionRate": 0.0198, "substitutionRate": 0.0792, "duplicateTokenRate": 0.025, "audioDurationMs": 68448, "preprocessElapsedMs": 0, "transcribeElapsedMs": 16222, "totalElapsedMs": 16222, "realTimeFactor": 0.237, "sceneHint": "dialogue" }, { "engine": "ml-engine-asr", "model": "fire-red-ctc", "caseId": "en-tech-talk-podcast", "segmentCount": 26, "errorMetricKind": "wer", "wer": 0.1198, "cer": null, "insertionRate": 0.012, "deletionRate": 0.012, "substitutionRate": 0.0958, "duplicateTokenRate": 0.006, "audioDurationMs": 79680, "preprocessElapsedMs": 0, "transcribeElapsedMs": 14834, "totalElapsedMs": 14834, "realTimeFactor": 0.1862, "sceneHint": "course" }, { "engine": "ml-engine-asr", "model": "fire-red-ctc", "caseId": "zh-tech-podcast", "segmentCount": 29, "errorMetricKind": "cer", "wer": null, "cer": 0.0258, "insertionRate": 0.0032, "deletionRate": 0.0065, "substitutionRate": 0.0161, "duplicateTokenRate": 0.2857, "audioDurationMs": 74664, "preprocessElapsedMs": 0, "transcribeElapsedMs": 15444, "totalElapsedMs": 15444, "realTimeFactor": 0.2068, "sceneHint": "course" } ], "failures": [], "summary": { "caseCount": 5, "successCount": 5, "crashCount": 0, "crashRate": 0, "averageWer": 0.1142, "averageCer": 0.0258, "averageInsertionRate": 0.0092, "averageDeletionRate": 0.0086, "averageSubstitutionRate": 0.0787, "averageDuplicateTokenRate": 0.0633, "averagePreprocessElapsedMs": 0, "averageTotalElapsedMs": 13996.2, "averageRealTimeFactor": 0.2163, "byScene": { "general": { "caseCount": 1, "successCount": 1, "crashCount": 0, "crashRate": 0, "averageWer": 0.0753, "averageCer": null, "averageInsertionRate": 0.0108, "averageDeletionRate": 0, "averageSubstitutionRate": 0.0645, "averageDuplicateTokenRate": 0, "averagePreprocessElapsedMs": 0, "averageTotalElapsedMs": 8113, "averageRealTimeFactor": 0.2349 }, "speech": { "caseCount": 1, "successCount": 1, "crashCount": 0, "crashRate": 0, "averageWer": 0.1527, "averageCer": null, "averageInsertionRate": 0.0099, "averageDeletionRate": 0.0049, "averageSubstitutionRate": 0.1379, "averageDuplicateTokenRate": 0, "averagePreprocessElapsedMs": 0, "averageTotalElapsedMs": 15368, "averageRealTimeFactor": 0.2165 }, "dialogue": { "caseCount": 1, "successCount": 1, "crashCount": 0, "crashRate": 0, "averageWer": 0.1089, "averageCer": null, "averageInsertionRate": 0.0099, "averageDeletionRate": 0.0198, "averageSubstitutionRate": 0.0792, "averageDuplicateTokenRate": 0.025, "averagePreprocessElapsedMs": 0, "averageTotalElapsedMs": 16222, "averageRealTimeFactor": 0.237 }, "course": { "caseCount": 2, "successCount": 2, "crashCount": 0, "crashRate": 0, "averageWer": 0.1198, "averageCer": 0.0258, "averageInsertionRate": 0.0076, "averageDeletionRate": 0.0092, "averageSubstitutionRate": 0.0559, "averageDuplicateTokenRate": 0.1459, "averagePreprocessElapsedMs": 0, "averageTotalElapsedMs": 15139, "averageRealTimeFactor": 0.1965 } } } } }, { "modelId": "moonshine", "modelArch": "moonshine", "batch": { "engine": "ml-engine-asr", "results": [ { "engine": "ml-engine-asr", "model": "moonshine", "caseId": "audio-benchmark", "segmentCount": 15, "errorMetricKind": "wer", "wer": 0, "cer": null, "insertionRate": 0, "deletionRate": 0, "substitutionRate": 0, "duplicateTokenRate": 0, "audioDurationMs": 34536, "preprocessElapsedMs": 0, "transcribeElapsedMs": 667, "totalElapsedMs": 667, "realTimeFactor": 0.0193, "sceneHint": "general" }, { "engine": "ml-engine-asr", "model": "moonshine", "caseId": "en-long-sentence-pause", "segmentCount": 33, "errorMetricKind": "wer", "wer": 0.0542, "cer": null, "insertionRate": 0, "deletionRate": 0, "substitutionRate": 0.0542, "duplicateTokenRate": 0, "audioDurationMs": 70992, "preprocessElapsedMs": 0, "transcribeElapsedMs": 1558, "totalElapsedMs": 1558, "realTimeFactor": 0.022, "sceneHint": "speech" }, { "engine": "ml-engine-asr", "model": "moonshine", "caseId": "en-oral-correction", "segmentCount": 35, "errorMetricKind": "wer", "wer": 0.0594, "cer": null, "insertionRate": 0.005, "deletionRate": 0.0149, "substitutionRate": 0.0396, "duplicateTokenRate": 0.025, "audioDurationMs": 68448, "preprocessElapsedMs": 0, "transcribeElapsedMs": 1310, "totalElapsedMs": 1310, "realTimeFactor": 0.0191, "sceneHint": "dialogue" }, { "engine": "ml-engine-asr", "model": "moonshine", "caseId": "en-tech-talk-podcast", "segmentCount": 26, "errorMetricKind": "wer", "wer": 0.0479, "cer": null, "insertionRate": 0.024, "deletionRate": 0.006, "substitutionRate": 0.018, "duplicateTokenRate": 0.0059, "audioDurationMs": 79680, "preprocessElapsedMs": 0, "transcribeElapsedMs": 1131, "totalElapsedMs": 1131, "realTimeFactor": 0.0142, "sceneHint": "course" } ], "failures": [], "summary": { "caseCount": 4, "successCount": 4, "crashCount": 0, "crashRate": 0, "averageWer": 0.0404, "averageCer": null, "averageInsertionRate": 0.0073, "averageDeletionRate": 0.0052, "averageSubstitutionRate": 0.0279, "averageDuplicateTokenRate": 0.0077, "averagePreprocessElapsedMs": 0, "averageTotalElapsedMs": 1166.5, "averageRealTimeFactor": 0.0186, "byScene": { "general": { "caseCount": 1, "successCount": 1, "crashCount": 0, "crashRate": 0, "averageWer": 0, "averageCer": null, "averageInsertionRate": 0, "averageDeletionRate": 0, "averageSubstitutionRate": 0, "averageDuplicateTokenRate": 0, "averagePreprocessElapsedMs": 0, "averageTotalElapsedMs": 667, "averageRealTimeFactor": 0.0193 }, "speech": { "caseCount": 1, "successCount": 1, "crashCount": 0, "crashRate": 0, "averageWer": 0.0542, "averageCer": null, "averageInsertionRate": 0, "averageDeletionRate": 0, "averageSubstitutionRate": 0.0542, "averageDuplicateTokenRate": 0, "averagePreprocessElapsedMs": 0, "averageTotalElapsedMs": 1558, "averageRealTimeFactor": 0.022 }, "dialogue": { "caseCount": 1, "successCount": 1, "crashCount": 0, "crashRate": 0, "averageWer": 0.0594, "averageCer": null, "averageInsertionRate": 0.005, "averageDeletionRate": 0.0149, "averageSubstitutionRate": 0.0396, "averageDuplicateTokenRate": 0.025, "averagePreprocessElapsedMs": 0, "averageTotalElapsedMs": 1310, "averageRealTimeFactor": 0.0191 }, "course": { "caseCount": 1, "successCount": 1, "crashCount": 0, "crashRate": 0, "averageWer": 0.0479, "averageCer": null, "averageInsertionRate": 0.024, "averageDeletionRate": 0.006, "averageSubstitutionRate": 0.018, "averageDuplicateTokenRate": 0.0059, "averagePreprocessElapsedMs": 0, "averageTotalElapsedMs": 1131, "averageRealTimeFactor": 0.0142 } } } } }, { "modelId": "cohere-transcribe-multilingual", "modelArch": "cohere", "batch": { "engine": "ml-engine-asr", "results": [ { "engine": "ml-engine-asr", "model": "cohere", "caseId": "audio-benchmark", "segmentCount": 15, "errorMetricKind": "wer", "wer": 0.0108, "cer": null, "insertionRate": 0, "deletionRate": 0.0108, "substitutionRate": 0, "duplicateTokenRate": 0, "audioDurationMs": 34536, "preprocessElapsedMs": 0, "transcribeElapsedMs": 13095, "totalElapsedMs": 13095, "realTimeFactor": 0.3792, "sceneHint": "general" }, { "engine": "ml-engine-asr", "model": "cohere", "caseId": "en-long-sentence-pause", "segmentCount": 33, "errorMetricKind": "wer", "wer": 0.0049, "cer": null, "insertionRate": 0, "deletionRate": 0, "substitutionRate": 0.0049, "duplicateTokenRate": 0, "audioDurationMs": 70992, "preprocessElapsedMs": 0, "transcribeElapsedMs": 27822, "totalElapsedMs": 27822, "realTimeFactor": 0.3919, "sceneHint": "speech" }, { "engine": "ml-engine-asr", "model": "cohere", "caseId": "en-oral-correction", "segmentCount": 35, "errorMetricKind": "wer", "wer": 0.0495, "cer": null, "insertionRate": 0, "deletionRate": 0.0149, "substitutionRate": 0.0347, "duplicateTokenRate": 0.0251, "audioDurationMs": 68448, "preprocessElapsedMs": 0, "transcribeElapsedMs": 28106, "totalElapsedMs": 28106, "realTimeFactor": 0.4106, "sceneHint": "dialogue" }, { "engine": "ml-engine-asr", "model": "cohere", "caseId": "en-tech-talk-podcast", "segmentCount": 26, "errorMetricKind": "wer", "wer": 0.006, "cer": null, "insertionRate": 0, "deletionRate": 0, "substitutionRate": 0.006, "duplicateTokenRate": 0.006, "audioDurationMs": 79680, "preprocessElapsedMs": 0, "transcribeElapsedMs": 18184, "totalElapsedMs": 18184, "realTimeFactor": 0.2282, "sceneHint": "course" }, { "engine": "ml-engine-asr", "model": "cohere", "caseId": "zh-tech-podcast", "segmentCount": 29, "errorMetricKind": "cer", "wer": null, "cer": 0.1032, "insertionRate": 0, "deletionRate": 0.0903, "substitutionRate": 0.0129, "duplicateTokenRate": 0.5, "audioDurationMs": 74664, "preprocessElapsedMs": 0, "transcribeElapsedMs": 23821, "totalElapsedMs": 23821, "realTimeFactor": 0.319, "sceneHint": "course" } ], "failures": [], "summary": { "caseCount": 5, "successCount": 5, "crashCount": 0, "crashRate": 0, "averageWer": 0.0178, "averageCer": 0.1032, "averageInsertionRate": 0, "averageDeletionRate": 0.0232, "averageSubstitutionRate": 0.0117, "averageDuplicateTokenRate": 0.1062, "averagePreprocessElapsedMs": 0, "averageTotalElapsedMs": 22205.6, "averageRealTimeFactor": 0.3458, "byScene": { "general": { "caseCount": 1, "successCount": 1, "crashCount": 0, "crashRate": 0, "averageWer": 0.0108, "averageCer": null, "averageInsertionRate": 0, "averageDeletionRate": 0.0108, "averageSubstitutionRate": 0, "averageDuplicateTokenRate": 0, "averagePreprocessElapsedMs": 0, "averageTotalElapsedMs": 13095, "averageRealTimeFactor": 0.3792 }, "speech": { "caseCount": 1, "successCount": 1, "crashCount": 0, "crashRate": 0, "averageWer": 0.0049, "averageCer": null, "averageInsertionRate": 0, "averageDeletionRate": 0, "averageSubstitutionRate": 0.0049, "averageDuplicateTokenRate": 0, "averagePreprocessElapsedMs": 0, "averageTotalElapsedMs": 27822, "averageRealTimeFactor": 0.3919 }, "dialogue": { "caseCount": 1, "successCount": 1, "crashCount": 0, "crashRate": 0, "averageWer": 0.0495, "averageCer": null, "averageInsertionRate": 0, "averageDeletionRate": 0.0149, "averageSubstitutionRate": 0.0347, "averageDuplicateTokenRate": 0.0251, "averagePreprocessElapsedMs": 0, "averageTotalElapsedMs": 28106, "averageRealTimeFactor": 0.4106 }, "course": { "caseCount": 2, "successCount": 2, "crashCount": 0, "crashRate": 0, "averageWer": 0.006, "averageCer": 0.1032, "averageInsertionRate": 0, "averageDeletionRate": 0.0452, "averageSubstitutionRate": 0.0095, "averageDuplicateTokenRate": 0.253, "averagePreprocessElapsedMs": 0, "averageTotalElapsedMs": 21002.5, "averageRealTimeFactor": 0.2736 } } } } } ], "skipped": [ { "modelId": "en-zipformer-onnx", "caseId": "zh-tech-podcast", "reason": "language_not_supported" }, { "modelId": "parakeet", "caseId": "zh-tech-podcast", "reason": "language_not_supported" }, { "modelId": "moonshine", "caseId": "zh-tech-podcast", "reason": "language_not_supported" } ] }