| { |
| "status": "selected_for_fresh_qualification", |
| "selected_batch": 2, |
| "selected_source_root": "work/mlx-native-q3-v1", |
| "weights_unchanged": true, |
| "selection_data": "Exposed AMI IS1000a only; reserved11 results not reused for selection", |
| "cases": [ |
| { |
| "batch": 4, |
| "report": "/Users/Ajab/conductor/workspaces/visualears/experiments/audio8-ternary/reports/mlx-native-q3-stream-v1/verification.json", |
| "report_sha256": "0d9e39bf29670fd54631b693e51a0cd8873105989a47239b588d0273a53255cc", |
| "generated_ids_sha256": "72f994120072e2b864ea8de08e44195efe481d10a74dedf2e78bd4a1a5eea4b3", |
| "word_errors": { |
| "hits": 303, |
| "substitutions": 39, |
| "deletions": 136, |
| "insertions": 4, |
| "reference_units": 478, |
| "hypothesis_units": 346, |
| "errors": 179, |
| "error_rate": 0.37447698744769875 |
| }, |
| "matched_word_timing": { |
| "measured": true, |
| "actual_arrival_paced": true, |
| "timing_basis": "Independent host receipt of flushed JSONL since paced20ms producer start; all source words retained", |
| "correctly_matched_units": 303, |
| "ambiguous_matched_units": 9, |
| "ambiguity_method": "all minimum unit-cost Levenshtein paths, without timing-based selection", |
| "unambiguous_matched_units": 294, |
| "unambiguous_maximum_seconds": 5.586898207997422, |
| "unmatched_reference_units": 175, |
| "matched_reference_fraction": 0.6338912133891214, |
| "min_seconds": -2.193736457971383, |
| "p50_seconds": 0.6946411670209045, |
| "p95_seconds": 0.8736333750304794, |
| "maximum_seconds": 5.586898207997422, |
| "matched_units_above_one_second": 4, |
| "all_matched_units_within_one_second": null, |
| "whole_release_pass": false, |
| "limitation": "Lag summaries use one deterministic text alignment and are descriptive when alternative optimal alignments exist. Only unambiguous matches can support per-word timing claims. Omissions/substitutions remain errors, never zero-latency successes. Published reference timing uncertainty and overlapping-speaker word order are not independently corrected." |
| }, |
| "causal_timing_audit": { |
| "count": 293, |
| "p95_seconds": 0.8712869580299412, |
| "max_seconds": 1.1078053750237586, |
| "above_one_second": 2 |
| }, |
| "cross_reset_matches": 1, |
| "mature_work_rtf": 0.7764502045640678 |
| }, |
| { |
| "batch": 2, |
| "report": "/Users/Ajab/conductor/workspaces/visualears/experiments/audio8-ternary/reports/mlx-native-q3-b2-exposed-v1/run.json", |
| "report_sha256": "789a6658c508b70aba1060c3cdd0412b87bd7b35e4793d47f8cb45e4c46c61e5", |
| "generated_ids_sha256": "3d3130b78aad7aa764dfbf1aeae3591847f84f89379f03fb6eab9a072cb6d18c", |
| "word_errors": { |
| "hits": 302, |
| "substitutions": 39, |
| "deletions": 137, |
| "insertions": 3, |
| "reference_units": 478, |
| "hypothesis_units": 344, |
| "errors": 179, |
| "error_rate": 0.37447698744769875 |
| }, |
| "matched_word_timing": { |
| "measured": true, |
| "actual_arrival_paced": true, |
| "timing_basis": "Independent host receipt of flushed JSONL since paced20ms producer start; all source words retained", |
| "correctly_matched_units": 302, |
| "ambiguous_matched_units": 8, |
| "ambiguity_method": "all minimum unit-cost Levenshtein paths, without timing-based selection", |
| "unambiguous_matched_units": 294, |
| "unambiguous_maximum_seconds": 5.489463750023845, |
| "unmatched_reference_units": 176, |
| "matched_reference_fraction": 0.6317991631799164, |
| "min_seconds": -2.2680539579666146, |
| "p50_seconds": 0.5616992500261517, |
| "p95_seconds": 0.7167164999945186, |
| "maximum_seconds": 5.489463750023845, |
| "matched_units_above_one_second": 1, |
| "all_matched_units_within_one_second": null, |
| "whole_release_pass": false, |
| "limitation": "Lag summaries use one deterministic text alignment and are descriptive when alternative optimal alignments exist. Only unambiguous matches can support per-word timing claims. Omissions/substitutions remain errors, never zero-latency successes. Published reference timing uncertainty and overlapping-speaker word order are not independently corrected." |
| }, |
| "causal_timing_audit": { |
| "count": 293, |
| "p95_seconds": 0.7042127919802397, |
| "max_seconds": 0.9364958749921044, |
| "above_one_second": 0 |
| }, |
| "cross_reset_matches": 1, |
| "mature_work_rtf": 0.8644507736555196 |
| }, |
| { |
| "batch": 3, |
| "report": "/Users/Ajab/conductor/workspaces/visualears/experiments/audio8-ternary/reports/mlx-native-q3-b3-exposed-v2/run.json", |
| "report_sha256": "ac20ccdcf5eb13b28227de3428465c6caac31f1e21032b2037b16308f43e5cd3", |
| "generated_ids_sha256": "3d3130b78aad7aa764dfbf1aeae3591847f84f89379f03fb6eab9a072cb6d18c", |
| "word_errors": { |
| "hits": 302, |
| "substitutions": 39, |
| "deletions": 137, |
| "insertions": 3, |
| "reference_units": 478, |
| "hypothesis_units": 344, |
| "errors": 179, |
| "error_rate": 0.37447698744769875 |
| }, |
| "matched_word_timing": { |
| "measured": true, |
| "actual_arrival_paced": true, |
| "timing_basis": "Independent host receipt of flushed JSONL since paced20ms producer start; all source words retained", |
| "correctly_matched_units": 302, |
| "ambiguous_matched_units": 8, |
| "ambiguity_method": "all minimum unit-cost Levenshtein paths, without timing-based selection", |
| "unambiguous_matched_units": 294, |
| "unambiguous_maximum_seconds": 5.515487207956596, |
| "unmatched_reference_units": 176, |
| "matched_reference_fraction": 0.6317991631799164, |
| "min_seconds": -2.1270811670180336, |
| "p50_seconds": 0.6239733329950923, |
| "p95_seconds": 0.7859231659909938, |
| "maximum_seconds": 5.515487207956596, |
| "matched_units_above_one_second": 4, |
| "all_matched_units_within_one_second": null, |
| "whole_release_pass": false, |
| "limitation": "Lag summaries use one deterministic text alignment and are descriptive when alternative optimal alignments exist. Only unambiguous matches can support per-word timing claims. Omissions/substitutions remain errors, never zero-latency successes. Published reference timing uncertainty and overlapping-speaker word order are not independently corrected." |
| }, |
| "causal_timing_audit": { |
| "count": 293, |
| "p95_seconds": 0.7846965410071789, |
| "max_seconds": 1.0496153329871731, |
| "above_one_second": 2 |
| }, |
| "cross_reset_matches": 1, |
| "mature_work_rtf": 0.8001077809603885 |
| } |
| ], |
| "reason": "Batch2 and batch3 have identical tokens. All three batches have179/478 errors, but batch2/3 have one more deletion and one fewer insertion than batch4. Batch2 has the lowest delay and no >1 second same-session unambiguous match in this exposed diagnostic while mature RTF remains below1. Cross-reset textual matches remain unresolved recognition evidence, not latency successes. No universal pass inferred.", |
| "new_inputs": [ |
| "data/continuous-v3/manifest.json", |
| "data/heldout-zh-v3/manifest.json" |
| ], |
| "selector_sha256": "7ba7cbed4125fd7834a57ba909ca464b4d970e39487b551b567e0b4b9265acd0", |
| "production_ready": false |
| } |
|
|
