diff --git a/analysis/current/analysis-report-hybrid.json b/analysis/current/analysis-report-hybrid.json index 909546d7a8c70d9b8bffb61405ed0b6f78316f66..50161d03e511b3d348b487662470a38a6906ecbb 100644 --- a/analysis/current/analysis-report-hybrid.json +++ b/analysis/current/analysis-report-hybrid.json @@ -1,8 +1,8 @@ { "schema_version": "1.0", "repo": "huggingface/transformers", - "snapshot_id": "20260718T060036Z", - "generated_at": "2026-07-18T06:12:22Z", + "snapshot_id": "20260718T120032Z", + "generated_at": "2026-07-18T12:14:56Z", "evidence_quality": "full", "llm_enrichment": true, "meta_bugs": [ @@ -1050,111 +1050,23 @@ ] }, { - "cluster_id": "cluster-43451-5", - "summary": "Cluster of 1 issues and 4 PRs centered on issue #47039.", - "status": "open", - "confidence": 0.6, - "canonical_issue_number": 47039, - "canonical_pr_number": 47291, - "issue_numbers": [ - 47039 - ], - "pr_numbers": [ - 43451, - 47163, - 47175, - 47291 - ], - "evidence_types": [ - "closing_reference", - "soft_similarity" - ], - "pr_comparisons": [ - { - "left_pr_number": 43451, - "right_pr_number": 47163, - "code_similarity": 0.004, - "size_similarity": 0.004, - "file_overlap": 0.0, - "area_overlap": 0.0, - "patch_similarity": 0.02, - "shared_filenames": [], - "shared_file_areas": [] - }, - { - "left_pr_number": 43451, - "right_pr_number": 47175, - "code_similarity": 0.002, - "size_similarity": 0.002, - "file_overlap": 0.0, - "area_overlap": 0.0, - "patch_similarity": 0.008, - "shared_filenames": [], - "shared_file_areas": [] - }, - { - "left_pr_number": 43451, - "right_pr_number": 47291, - "code_similarity": 0.002, - "size_similarity": 0.005, - "file_overlap": 0.0, - "area_overlap": 0.0, - "patch_similarity": 0.006, - "shared_filenames": [], - "shared_file_areas": [] - }, - { - "left_pr_number": 47163, - "right_pr_number": 47175, - "code_similarity": 0.112, - "size_similarity": 0.536, - "file_overlap": 0.0, - "area_overlap": 0.0, - "patch_similarity": 0.03, - "shared_filenames": [], - "shared_file_areas": [] - }, - { - "left_pr_number": 47163, - "right_pr_number": 47291, - "code_similarity": 0.176, - "size_similarity": 0.875, - "file_overlap": 0.0, - "area_overlap": 0.0, - "patch_similarity": 0.007, - "shared_filenames": [], - "shared_file_areas": [] - }, - { - "left_pr_number": 47175, - "right_pr_number": 47291, - "code_similarity": 0.096, - "size_similarity": 0.469, - "file_overlap": 0.0, - "area_overlap": 0.0, - "patch_similarity": 0.012, - "shared_filenames": [], - "shared_file_areas": [] - } - ] - }, - { - "cluster_id": "cluster-46894-7", - "summary": "Cluster of 3 issues and 4 PRs centered on issue #46894.", + "cluster_id": "cluster-46894-8", + "summary": "Cluster of 3 issues and 5 PRs centered on issue #46894.", "status": "open", "confidence": 0.8, "canonical_issue_number": 46894, "canonical_pr_number": 46903, "issue_numbers": [ 46894, - 46897, - 46901 + 46901, + 47394 ], "pr_numbers": [ 46895, 46903, 46958, - 47090 + 47090, + 47395 ], "evidence_types": [ "closing_reference", @@ -1349,6 +1261,17 @@ } ] }, + { + "left_pr_number": 46895, + "right_pr_number": 47395, + "code_similarity": 0.187, + "size_similarity": 0.683, + "file_overlap": 0.0, + "area_overlap": 0.0, + "patch_similarity": 0.333, + "shared_filenames": [], + "shared_file_areas": [] + }, { "left_pr_number": 46903, "right_pr_number": 46958, @@ -1479,6 +1402,17 @@ } ] }, + { + "left_pr_number": 46903, + "right_pr_number": 47395, + "code_similarity": 0.083, + "size_similarity": 0.276, + "file_overlap": 0.0, + "area_overlap": 0.0, + "patch_similarity": 0.186, + "shared_filenames": [], + "shared_file_areas": [] + }, { "left_pr_number": 46958, "right_pr_number": 47090, @@ -1551,6 +1485,168 @@ ] } ] + }, + { + "left_pr_number": 46958, + "right_pr_number": 47395, + "code_similarity": 0.201, + "size_similarity": 0.747, + "file_overlap": 0.0, + "area_overlap": 0.0, + "patch_similarity": 0.342, + "shared_filenames": [], + "shared_file_areas": [] + }, + { + "left_pr_number": 47090, + "right_pr_number": 47395, + "code_similarity": 0.078, + "size_similarity": 0.192, + "file_overlap": 0.0, + "area_overlap": 0.0, + "patch_similarity": 0.262, + "shared_filenames": [], + "shared_file_areas": [] + } + ] + }, + { + "cluster_id": "cluster-46509-6", + "summary": "Cluster of 2 issues and 4 PRs centered on issue #47333.", + "status": "open", + "confidence": 0.6, + "canonical_issue_number": 47333, + "canonical_pr_number": 47336, + "issue_numbers": [ + 47332, + 47333 + ], + "pr_numbers": [ + 46509, + 47143, + 47301, + 47336 + ], + "evidence_types": [ + "closing_reference", + "soft_similarity" + ], + "pr_comparisons": [ + { + "left_pr_number": 46509, + "right_pr_number": 47143, + "code_similarity": 0.254, + "size_similarity": 0.793, + "file_overlap": 0.071, + "area_overlap": 0.143, + "patch_similarity": 0.161, + "shared_filenames": [ + "src/transformers/models/auto/auto_mappings.py", + "src/transformers/models/auto/modeling_auto.py" + ], + "shared_file_areas": [ + { + "filename": "src/transformers/models/auto/auto_mappings.py", + "left_ranges": [ + [ + 412, + 418 + ], + [ + 1077, + 1083 + ], + [ + 1206, + 1212 + ] + ], + "right_ranges": [ + [ + 412, + 418 + ], + [ + 975, + 981 + ] + ] + }, + { + "filename": "src/transformers/models/auto/modeling_auto.py", + "left_ranges": [ + [ + 1097, + 1103 + ] + ], + "right_ranges": [ + [ + 348, + 354 + ], + [ + 1960, + 1966 + ] + ] + } + ] + }, + { + "left_pr_number": 46509, + "right_pr_number": 47301, + "code_similarity": 0.01, + "size_similarity": 0.013, + "file_overlap": 0.0, + "area_overlap": 0.0, + "patch_similarity": 0.047, + "shared_filenames": [], + "shared_file_areas": [] + }, + { + "left_pr_number": 46509, + "right_pr_number": 47336, + "code_similarity": 0.003, + "size_similarity": 0.008, + "file_overlap": 0.0, + "area_overlap": 0.0, + "patch_similarity": 0.007, + "shared_filenames": [], + "shared_file_areas": [] + }, + { + "left_pr_number": 47143, + "right_pr_number": 47301, + "code_similarity": 0.006, + "size_similarity": 0.01, + "file_overlap": 0.0, + "area_overlap": 0.0, + "patch_similarity": 0.026, + "shared_filenames": [], + "shared_file_areas": [] + }, + { + "left_pr_number": 47143, + "right_pr_number": 47336, + "code_similarity": 0.002, + "size_similarity": 0.006, + "file_overlap": 0.0, + "area_overlap": 0.0, + "patch_similarity": 0.004, + "shared_filenames": [], + "shared_file_areas": [] + }, + { + "left_pr_number": 47301, + "right_pr_number": 47336, + "code_similarity": 0.12, + "size_similarity": 0.6, + "file_overlap": 0.0, + "area_overlap": 0.0, + "patch_similarity": 0.0, + "shared_filenames": [], + "shared_file_areas": [] } ] }, @@ -1753,32 +1849,52 @@ ] }, { - "cluster_id": "cluster-46658-3", - "summary": "Cluster of 1 issues and 2 PRs centered on issue #47321.", + "cluster_id": "cluster-47175-3", + "summary": "Cluster of 3 related pull requests linked by soft_similarity.", "status": "open", - "confidence": 0.6, - "canonical_issue_number": 47321, - "canonical_pr_number": 47322, - "issue_numbers": [ - 47321 - ], + "confidence": 0.5, + "canonical_issue_number": null, + "canonical_pr_number": 47298, + "issue_numbers": [], "pr_numbers": [ - 46658, - 47322 + 47175, + 47294, + 47298 ], "evidence_types": [ - "closing_reference", "soft_similarity" ], "pr_comparisons": [ { - "left_pr_number": 46658, - "right_pr_number": 47322, - "code_similarity": 0.001, - "size_similarity": 0.002, + "left_pr_number": 47175, + "right_pr_number": 47294, + "code_similarity": 0.008, + "size_similarity": 0.032, + "file_overlap": 0.0, + "area_overlap": 0.0, + "patch_similarity": 0.011, + "shared_filenames": [], + "shared_file_areas": [] + }, + { + "left_pr_number": 47175, + "right_pr_number": 47298, + "code_similarity": 0.019, + "size_similarity": 0.043, + "file_overlap": 0.0, + "area_overlap": 0.0, + "patch_similarity": 0.067, + "shared_filenames": [], + "shared_file_areas": [] + }, + { + "left_pr_number": 47294, + "right_pr_number": 47298, + "code_similarity": 0.153, + "size_similarity": 0.734, "file_overlap": 0.0, "area_overlap": 0.0, - "patch_similarity": 0.002, + "patch_similarity": 0.044, "shared_filenames": [], "shared_file_areas": [] } @@ -2534,6 +2650,81 @@ } ] }, + { + "cluster_id": "cluster-43656-2", + "summary": "Cluster of 2 related pull requests linked by soft_similarity.", + "status": "open", + "confidence": 0.5, + "canonical_issue_number": null, + "canonical_pr_number": 43656, + "issue_numbers": [], + "pr_numbers": [ + 43656, + 43842 + ], + "evidence_types": [ + "soft_similarity" + ], + "pr_comparisons": [ + { + "left_pr_number": 43656, + "right_pr_number": 43842, + "code_similarity": 0.405, + "size_similarity": 0.5, + "file_overlap": 1.0, + "area_overlap": 0.0, + "patch_similarity": 0.036, + "shared_filenames": [ + "src/transformers/cli/serve.py" + ], + "shared_file_areas": [ + { + "filename": "src/transformers/cli/serve.py", + "left_ranges": [ + [ + 11, + 18 + ], + [ + 30, + 36 + ], + [ + 315, + 323 + ], + [ + 665, + 671 + ], + [ + 931, + 937 + ], + [ + 1843, + 1849 + ], + [ + 1868, + 1874 + ] + ], + "right_ranges": [ + [ + 54, + 61 + ], + [ + 587, + 593 + ] + ] + } + ] + } + ] + }, { "cluster_id": "cluster-47123-2", "summary": "Cluster of 2 related pull requests linked by soft_similarity.", @@ -2593,6 +2784,35 @@ } ] }, + { + "cluster_id": "cluster-45157-2", + "summary": "Cluster of 2 related pull requests linked by soft_similarity.", + "status": "open", + "confidence": 0.5, + "canonical_issue_number": null, + "canonical_pr_number": 47387, + "issue_numbers": [], + "pr_numbers": [ + 45157, + 47387 + ], + "evidence_types": [ + "soft_similarity" + ], + "pr_comparisons": [ + { + "left_pr_number": 45157, + "right_pr_number": 47387, + "code_similarity": 0.176, + "size_similarity": 0.842, + "file_overlap": 0.0, + "area_overlap": 0.0, + "patch_similarity": 0.053, + "shared_filenames": [], + "shared_file_areas": [] + } + ] + }, { "cluster_id": "cluster-47282-2", "summary": "Cluster of 2 related pull requests linked by soft_similarity.", @@ -2665,290 +2885,52 @@ ] }, { - "cluster_id": "cluster-44018-2", + "cluster_id": "cluster-47163-2", "summary": "Cluster of 2 related pull requests linked by soft_similarity.", "status": "open", "confidence": 0.5, "canonical_issue_number": null, - "canonical_pr_number": 44068, + "canonical_pr_number": 47163, "issue_numbers": [], "pr_numbers": [ - 44018, - 44068 + 47163, + 47210 ], "evidence_types": [ "soft_similarity" ], "pr_comparisons": [ { - "left_pr_number": 44018, - "right_pr_number": 44068, - "code_similarity": 0.766, - "size_similarity": 0.939, - "file_overlap": 1.0, - "area_overlap": 0.425, - "patch_similarity": 0.866, - "shared_filenames": [ - "src/transformers/models/gpt_neo/modeling_gpt_neo.py", - "tests/models/gpt_neo/test_modeling_gpt_neo.py" - ], - "shared_file_areas": [ - { - "filename": "src/transformers/models/gpt_neo/modeling_gpt_neo.py", - "left_ranges": [ - [ - 26, - 46 - ], - [ - 138, - 143 - ], - [ - 182, - 187 - ], - [ - 283, - 295 - ], - [ - 327, - 341 - ], - [ - 348, - 354 - ], - [ - 360, - 369 - ], - [ - 399, - 406 - ], - [ - 411, - 419 - ], - [ - 428, - 433 - ], - [ - 472, - 492 - ], - [ - 507, - 513 - ], - [ - 519, - 528 - ], - [ - 541, - 559 - ], - [ - 562, - 567 - ], - [ - 595, - 601 - ], - [ - 607, - 614 - ], - [ - 627, - 643 - ], - [ - 685, - 690 - ], - [ - 708, - 714 - ], - [ - 720, - 727 - ], - [ - 740, - 757 - ], - [ - 761, - 766 - ], - [ - 780, - 786 - ], - [ - 791, - 798 - ], - [ - 807, - 822 - ], - [ - 840, - 845 - ] - ], - "right_ranges": [ - [ - 26, - 31 - ], - [ - 34, - 44 - ], - [ - 136, - 141 - ], - [ - 180, - 185 - ], - [ - 281, - 293 - ], - [ - 325, - 339 - ], - [ - 346, - 352 - ], - [ - 358, - 367 - ], - [ - 397, - 404 - ], - [ - 409, - 417 - ], - [ - 426, - 434 - ], - [ - 463, - 483 - ], - [ - 498, - 504 - ], - [ - 510, - 519 - ], - [ - 532, - 550 - ], - [ - 553, - 564 - ], - [ - 586, - 592 - ], - [ - 598, - 605 - ], - [ - 618, - 634 - ], - [ - 676, - 688 - ], - [ - 699, - 705 - ], - [ - 711, - 718 - ], - [ - 731, - 748 - ], - [ - 752, - 762 - ], - [ - 771, - 777 - ], - [ - 782, - 789 - ], - [ - 798, - 813 - ], - [ - 831, - 836 - ] - ] - }, - { - "filename": "tests/models/gpt_neo/test_modeling_gpt_neo.py", - "left_ranges": [ - [ - 458, - 464 - ] - ], - "right_ranges": [ - [ - 458, - 464 - ] - ] - } - ] + "left_pr_number": 47163, + "right_pr_number": 47210, + "code_similarity": 0.021, + "size_similarity": 0.076, + "file_overlap": 0.0, + "area_overlap": 0.0, + "patch_similarity": 0.036, + "shared_filenames": [], + "shared_file_areas": [] } ] } ], "duplicate_issues": [ { - "cluster_id": "cluster-46894-7", + "cluster_id": "cluster-46509-6", + "canonical_issue_number": 47333, + "duplicate_issue_numbers": [ + 47332 + ], + "reason": "Issues in cluster-46509-6 are treated as duplicates because they share closing_reference, soft_similarity evidence." + }, + { + "cluster_id": "cluster-46894-8", "canonical_issue_number": 46894, "duplicate_issue_numbers": [ - 46897, - 46901 + 46901, + 47394 ], - "reason": "Issues in cluster-46894-7 are treated as duplicates because they share closing_reference, shared_issue_target, soft_similarity evidence." + "reason": "Issues in cluster-46894-8 are treated as duplicates because they share closing_reference, shared_issue_target, soft_similarity evidence." }, { "cluster_id": "cluster-47138-4", @@ -2979,15 +2961,13 @@ "reason": "PRs in cluster-43240-3 are treated as duplicates because they converge on issue #43240 with closing_reference, shared_issue_target evidence." }, { - "cluster_id": "cluster-43451-5", - "canonical_pr_number": 47291, + "cluster_id": "cluster-43656-2", + "canonical_pr_number": 43656, "duplicate_pr_numbers": [ - 43451, - 47163, - 47175 + 43842 ], - "target_issue_number": 47039, - "reason": "PRs in cluster-43451-5 are treated as duplicates because they converge on issue #47039 with closing_reference, soft_similarity evidence." + "target_issue_number": null, + "reason": "PRs in cluster-43656-2 are treated as duplicates because they share soft_similarity evidence." }, { "cluster_id": "cluster-43698-3", @@ -3015,15 +2995,6 @@ "target_issue_number": 43979, "reason": "PRs in cluster-43979-11 are treated as duplicates because they converge on issue #43979 with closing_reference, shared_issue_target, soft_similarity evidence." }, - { - "cluster_id": "cluster-44018-2", - "canonical_pr_number": 44068, - "duplicate_pr_numbers": [ - 44018 - ], - "target_issue_number": null, - "reason": "PRs in cluster-44018-2 are treated as duplicates because they share soft_similarity evidence." - }, { "cluster_id": "cluster-45081-3", "canonical_pr_number": 45317, @@ -3033,6 +3004,15 @@ "target_issue_number": 45081, "reason": "PRs in cluster-45081-3 are treated as duplicates because they converge on issue #45081 with closing_reference, shared_issue_target, soft_similarity evidence." }, + { + "cluster_id": "cluster-45157-2", + "canonical_pr_number": 47387, + "duplicate_pr_numbers": [ + 45157 + ], + "target_issue_number": null, + "reason": "PRs in cluster-45157-2 are treated as duplicates because they share soft_similarity evidence." + }, { "cluster_id": "cluster-45561-3", "canonical_pr_number": 45645, @@ -3070,13 +3050,15 @@ "reason": "PRs in cluster-46305-3 are treated as duplicates because they converge on issue #46305 with closing_reference, shared_issue_target, soft_similarity evidence." }, { - "cluster_id": "cluster-46658-3", - "canonical_pr_number": 47322, + "cluster_id": "cluster-46509-6", + "canonical_pr_number": 47336, "duplicate_pr_numbers": [ - 46658 + 46509, + 47143, + 47301 ], - "target_issue_number": 47321, - "reason": "PRs in cluster-46658-3 are treated as duplicates because they converge on issue #47321 with closing_reference, soft_similarity evidence." + "target_issue_number": 47333, + "reason": "PRs in cluster-46509-6 are treated as duplicates because they converge on issue #47333 with closing_reference, soft_similarity evidence." }, { "cluster_id": "cluster-46869-3", @@ -3088,15 +3070,16 @@ "reason": "PRs in cluster-46869-3 are treated as duplicates because they converge on issue #46869 with closing_reference, shared_issue_target evidence." }, { - "cluster_id": "cluster-46894-7", + "cluster_id": "cluster-46894-8", "canonical_pr_number": 46903, "duplicate_pr_numbers": [ 46895, 46958, - 47090 + 47090, + 47395 ], "target_issue_number": 46894, - "reason": "PRs in cluster-46894-7 are treated as duplicates because they converge on issue #46894 with closing_reference, shared_issue_target, soft_similarity evidence." + "reason": "PRs in cluster-46894-8 are treated as duplicates because they converge on issue #46894 with closing_reference, shared_issue_target, soft_similarity evidence." }, { "cluster_id": "cluster-47123-2", @@ -3116,6 +3099,25 @@ "target_issue_number": 47138, "reason": "PRs in cluster-47138-4 are treated as duplicates because they converge on issue #47138 with closing_reference, soft_similarity evidence." }, + { + "cluster_id": "cluster-47163-2", + "canonical_pr_number": 47163, + "duplicate_pr_numbers": [ + 47210 + ], + "target_issue_number": null, + "reason": "PRs in cluster-47163-2 are treated as duplicates because they share soft_similarity evidence." + }, + { + "cluster_id": "cluster-47175-3", + "canonical_pr_number": 47298, + "duplicate_pr_numbers": [ + 47175, + 47294 + ], + "target_issue_number": null, + "reason": "PRs in cluster-47175-3 are treated as duplicates because they share soft_similarity evidence." + }, { "cluster_id": "cluster-47282-2", "canonical_pr_number": 47380, @@ -3127,10 +3129,10 @@ } ], "best_issue": { - "cluster_id": "cluster-46894-7", + "cluster_id": "cluster-46894-8", "issue_number": 46894, - "reason": "Issue #46894 is the strongest global issue candidate because it is open, belongs to a cluster with 7 artifacts, and carries 1 discussion comments plus 10 inbound references.", - "score": 182.0 + "reason": "Issue #46894 is the strongest global issue candidate because it is open, belongs to a cluster with 8 artifacts, and carries 1 discussion comments plus 10 inbound references.", + "score": 192.0 }, "best_pr": { "cluster_id": "cluster-43979-11", diff --git a/analysis/current/analysis-report-hybrid.llm-reviews.json b/analysis/current/analysis-report-hybrid.llm-reviews.json index 21cf3f7d1370dbb277a6dbb29be5cd183d1b5f2f..e65d32944397fadfed36c438c2455f8a8a7977af 100644 --- a/analysis/current/analysis-report-hybrid.llm-reviews.json +++ b/analysis/current/analysis-report-hybrid.llm-reviews.json @@ -1,8 +1,8 @@ { "schema_version": "1.0", "repo": "huggingface/transformers", - "snapshot_id": "20260718T060036Z", - "generated_at": "2026-07-18T06:12:22Z", + "snapshot_id": "20260718T120032Z", + "generated_at": "2026-07-18T12:14:56Z", "model": "gpt-5.4-mini?service_tier=flex", "reviews": [ { @@ -16,7 +16,7 @@ "source": "llm", "cache_hit": false, "model": "gpt-5.4-mini?service_tier=flex", - "cluster_id": "cluster-10000-529", + "cluster_id": "cluster-10000-530", "nodes": [ "issue:24643", "issue:30277", @@ -125,20 +125,20 @@ "issue:43039|issue:44918", "issue:41762|issue:43638" ], - "prepared_review_unit_hash": "aa29dc0725a6f91bd032d2842f6cfa880b9c21ed2cb1fceeeb15f227e453c40d", + "prepared_review_unit_hash": "9a400136cc34d904348faa448fd0abd472fb13ee6274a452a4b0d81aad08433e", "component_budget": { - "node_count": 529, - "item_count": 529, - "soft_pair_count": 1953, - "serialized_chars": 1012074, - "estimated_input_tokens": 253019, - "estimated_eval_tokens": 506294 + "node_count": 530, + "item_count": 530, + "soft_pair_count": 1954, + "serialized_chars": 1013004, + "estimated_input_tokens": 253251, + "estimated_eval_tokens": 506758 }, "budget": { "node_count": 48, "item_count": 48, "soft_pair_count": 56, - "serialized_chars": 44520, + "serialized_chars": 44518, "estimated_input_tokens": 11130, "estimated_eval_tokens": 22516 }, @@ -147,321 +147,261 @@ "aggressively_trimmed": false, "split": true, "analyst_result": { - "summary": "This cluster is mostly heterogeneous: it mixes unrelated model-specific runtime errors, docs issues, backend/device failures, and a few duplicate-looking bug reports. Only a couple of pairs look like true duplicates.", - "confidence": 0.78, - "canonical_issue_reason": "Issue 43638 is the clearest representative of the only strong duplicate-like subcluster here: the DeepSpeed ZeRO-3/BERT index-0-out-of-bounds failure also reported in 43596.", + "summary": "Mostly heterogeneous cluster. The only clear duplicate subcluster is the DeepSpeed ZeRO-3 BERT IndexError pair (43596/43638); the other issues span different models, hardware paths, docs, and export/inference bugs.", + "confidence": 0.94, + "canonical_issue_reason": "issue:43596 is the best canonical issue because it is one of the only true duplicates in the set and gives a concrete, reproducible DeepSpeed ZeRO-3/BERT IndexError that matches 43638 almost exactly.", "canonical_pr_reason": null, - "best_issue_reason": "Issue 43638 is the best issue to anchor on, but only for the BERT/ZeRO-3 subcluster; it is not a good representative of the whole heterogeneous set.", + "best_issue_reason": "issue:43596 is the best global issue choice because it represents the only confirmed duplicate pair and is the clearest, most self-contained report for that bug.", "best_pr_reason": null, "soft_edge_verdicts": [ { "left": "issue:44322", "right": "issue:45464", "accept": false, - "reason": "Different problems: Qwen3.5 config missing num_attention_heads vs streaming chat/completions inference failure." + "reason": "Different failures: Qwen3_5Config missing num_attention_heads versus a Qwen3.5-0.8B streaming chat/completions error." }, { "left": "issue:45237", "right": "issue:45685", "accept": false, - "reason": "Different hardware/backend bugs: AMD GPU device-map failure vs MPS histogram on Int." + "reason": "Different model/platform bugs: GPT-OSS on AMD GPUs versus an MPS histogram kernel limitation for Int." }, { "left": "issue:43716", "right": "issue:45237", "accept": false, - "reason": "Different models and failure modes: Mistral-3 dtype mismatch vs GPT-OSS on AMD GPUs." + "reason": "Different root causes and subsystems: image preprocessor/model dtype mismatch versus AMD GPU device issues." }, { "left": "issue:43828", "right": "issue:45237", "accept": false, - "reason": "Autocast dtype mismatch in Phi-tiny-MoE is unrelated to GPT-OSS AMD-GPU failure." + "reason": "Autocast dtype mismatch in Phi-tiny-MoE is unrelated to GPT-OSS failing on AMD GPUs." }, { "left": "issue:45237", "right": "issue:46468", "accept": false, - "reason": "Different model-specific runtime failures; no shared code path is evident." + "reason": "Different models and code paths: GPT-OSS AMD GPU failure versus a broken Nemotron example." }, { "left": "issue:45237", "right": "issue:45507", "accept": false, - "reason": "Different model/back-end issues: AMD GPU failure vs GraniteMoEHybrid invalid method." + "reason": "Different bugs: AMD GPU compatibility failure versus GraniteMoEHybrid calling an invalid method." }, { "left": "issue:30277", "right": "issue:35545", "accept": false, - "reason": "DeepSpeed ZeRO-3 collective mismatch is unrelated to ModernBERT ONNX export errors." + "reason": "DeepSpeed collective mismatch and ModernBERT ONNX export are different failures." }, { "left": "issue:35545", "right": "issue:42915", "accept": false, - "reason": "Export failure vs FP8 quantization failure; different code paths." + "reason": "ONNX export error and Qwen3Moe FineGrainedFP8 failure are unrelated." }, { "left": "issue:35545", "right": "issue:41720", "accept": false, - "reason": "ModernBERT ONNX export and Qwen3 auto device mapping are unrelated." + "reason": "Different code paths: ONNX export failure versus a CUDA assert during auto device mapping." }, { "left": "issue:45901", "right": "issue:46015", "accept": false, - "reason": "Task crash vs documentation cleanup; not the same bug." + "reason": "Task crash and outdated RAG docs are not the same underlying issue." }, { "left": "issue:35545", "right": "issue:36010", "accept": false, - "reason": "ONNX export issue vs GenerationMixin import regression." + "reason": "ModernBERT ONNX export is unrelated to the GenerationMixin import error." }, { "left": "issue:44509", "right": "issue:44512", "accept": false, - "reason": "Both are docs regressions, but they mention different removed content and are not the same fix." + "reason": "Both are v5 docs issues, but they cover different removed documentation topics and are not the same change." }, { "left": "issue:44918", "right": "issue:46165", "accept": false, - "reason": "Different issues: TRL SFT embedding unpacking vs resume-training with trust_remote_code." + "reason": "TRL SFT embedding unpacking and resume-training with trust_remote_code are different bugs." }, { "left": "issue:46165", "right": "issue:46869", "accept": false, - "reason": "Different training/device-map failures; no shared bug evidence." + "reason": "Resume-training/trust_remote_code bug is unrelated to YOLO multi-GPU device_map auto failure." }, { "left": "issue:44805", "right": "issue:45823", "accept": false, - "reason": "Both are forward-time tensor issues, but one is a mask shape mismatch and the other is a Gemma4 device mismatch." + "reason": "Mask/tensor shape mismatch is a different root cause from Gemma4 PLE device mismatch." }, { "left": "issue:43638", "right": "issue:44805", "accept": false, - "reason": "BERT ZeRO-3 init index error vs a generic mask/indexing shape mismatch; different underlying bugs." - }, - { - "left": "issue:43638", - "right": "issue:46165", - "accept": false, - "reason": "BERT ZeRO-3 init error is unrelated to model resume/trust_remote_code." + "reason": "Both involve IndexError, but ZeRO-3/BERT init and mask-shape indexing are different bugs." }, { "left": "issue:43596", "right": "issue:44805", "accept": false, - "reason": "Both mention shape/index errors, but the BERT ZeRO-3 initialization bug is not the same as the mask indexing failure." + "reason": "Different error shapes and code paths; not the same underlying bug." }, { "left": "issue:44805", "right": "issue:46294", "accept": false, - "reason": "Different modalities and code paths: mask indexing vs video token/feature mismatch." - }, - { - "left": "issue:46165", - "right": "issue:46468", - "accept": false, - "reason": "Unrelated training-resume bug vs official example breakage." + "reason": "Tensor mask mismatch is unrelated to Qwen3.5 video token/feature mismatch." }, { "left": "issue:44805", "right": "issue:45161", "accept": false, - "reason": "Generic tensor indexing error vs GPT-OSS tensor-parallel issue; unrelated." + "reason": "Different failures: mask/tensor indexing versus GPT-OSS tensor-parallel behavior." }, { "left": "issue:44560", "right": "issue:46165", "accept": false, - "reason": "Video StopIteration in Qwen3-vl-embedding is unrelated to resume-training issues." + "reason": "Qwen3-vl video StopIteration is unrelated to resume-training with local models and trust_remote_code." }, { "left": "issue:34634", "right": "issue:35545", "accept": false, - "reason": "BarkProcessor voice_preset and ModernBERT export are unrelated." + "reason": "BarkProcessor voice_preset bug is unrelated to ModernBERT ONNX export." }, { "left": "issue:30277", "right": "issue:43638", "accept": false, - "reason": "Both involve DeepSpeed, but one is a collective mismatch and the other is a ZeRO-3 BERT init index error." + "reason": "Both involve DeepSpeed, but the concrete failures differ: collective mismatch versus zero-sized tensor IndexError." }, { "left": "issue:43596", "right": "issue:43638", "accept": true, - "reason": "Same underlying bug: BERT/DeepSpeed ZeRO-3 initialization hits index-0-out-of-bounds on empty/sharded tensors." + "reason": "Near-identical DeepSpeed ZeRO-3 BERT reports with the same out-of-bounds index-0 failure during initialization." }, { "left": "issue:43039", "right": "issue:45161", "accept": false, - "reason": "Liger kernel cross-entropy call vs GPT-OSS tensor-parallel failure are unrelated." + "reason": "Liger cross_entropy being called is a different issue from GPT-OSS tensor-parallel failure." }, { "left": "issue:43975", "right": "issue:44457", "accept": false, - "reason": "DeepSeek detokenization bug is unrelated to LoRA merge/save mismatch." + "reason": "Incorrect detokenization and LoRA merge/reload output mismatch are different bugs." }, { "left": "issue:44292", "right": "issue:44457", "accept": false, - "reason": "Different models and failures: Qwen NVFP4 runtime error vs LoRA reload mismatch." + "reason": "Qwen-3 NVFP4 runtime error is unrelated to LoRA save/load mismatch." }, { "left": "issue:34928", "right": "issue:35308", "accept": false, - "reason": "Both are distributed-training crashes, but the causes differ: activation checkpointing/FSDP mismatch vs iterable-dataset length handling." + "reason": "Activation-checkpoint/FSDP recomputation mismatch and IterableDataset length mismatch are different distributed-training bugs." }, { "left": "issue:44841", "right": "issue:46661", "accept": false, - "reason": "Processor failure for Voxtral is unrelated to remote-code config_class errors." + "reason": "Processor failure for Voxtral is unrelated to the remote-code config_class AttributeError." }, { "left": "issue:24643", "right": "issue:30277", "accept": false, - "reason": "Both are DeepSpeed training bugs, but the concrete failures are different: weight dimensionality vs collective mismatch." + "reason": "Different DeepSpeed failures: 2-D weight runtime error versus collective mismatch." }, { "left": "issue:43716", "right": "issue:46869", "accept": false, - "reason": "Different models and failure modes: Mistral-3 dtype mismatch vs YOLO multi-GPU device-map failure." - }, - { - "left": "issue:34928", - "right": "issue:44805", - "accept": false, - "reason": "Different tensor issues: recomputed-size mismatch under checkpointing/FSDP vs mask shape mismatch." + "reason": "Different models and causes: image preprocessing dtype mismatch versus YOLO multi-GPU device-map failure." }, { "left": "issue:45685", "right": "issue:46869", "accept": false, - "reason": "MPS histogram unsupported for Int is unrelated to YOLO device-map failure." - }, - { - "left": "issue:34928", - "right": "issue:36331", - "accept": false, - "reason": "Distributed tensor-size mismatch is unrelated to CustomTrainer.compute_loss signature regression." - }, - { - "left": "issue:45859", - "right": "issue:45902", - "accept": false, - "reason": "Same model component, but different bugs: CPU-offload incompatibility vs uninitialized inv_freq after meta-to-empty materialization." - }, - { - "left": "issue:39290", - "right": "issue:45081", - "accept": false, - "reason": "Gemma3 config attribute regression vs Mistral regex-patch crash; unrelated." - }, - { - "left": "issue:45800", - "right": "issue:46515", - "accept": true, - "reason": "Same underlying issue: both report incompatibility with torch 2.4.1 / transformers 5.8.0." - }, - { - "left": "issue:45823", - "right": "issue:45874", - "accept": false, - "reason": "Gemma4 device mismatch and expensive reverse-embedding lookup are different bugs." - }, - { - "left": "issue:34928", - "right": "issue:41720", - "accept": false, - "reason": "FSDP/checkpointing size mismatch is unrelated to Qwen3 auto device-map cuda asserts." + "reason": "MPS histogram kernel limitation is unrelated to YOLO device-map auto failure." }, { "left": "issue:43425", "right": "issue:44292", "accept": false, - "reason": "Torch 2.10 incompatibility and Qwen3 NVFP4 runtime error are not the same bug." + "reason": "Torch version incompatibility is not the same as a Qwen NVFP4 model runtime error." }, { "left": "issue:43828", "right": "issue:45161", "accept": false, - "reason": "Phi-tiny-MoE autocast dtype mismatch is unrelated to GPT-OSS tensor-parallel failure." + "reason": "Autocast dtype mismatch in Phi-tiny-MoE is a different bug from GPT-OSS tensor-parallel failure." }, { "left": "issue:43541", "right": "issue:43716", "accept": false, - "reason": "Mixtral torch-dynamo grouped_mm tracing error is unrelated to Mistral-3 image/text dtype mismatch." + "reason": "Torch dynamo grouped_mm tracing error is unrelated to Mistral-3 image dtype mismatch." }, { "left": "issue:43828", "right": "issue:44292", "accept": false, - "reason": "Different models and error classes; no shared code path." + "reason": "Different model families and failure modes." }, { "left": "issue:43716", "right": "issue:45685", "accept": false, - "reason": "Mistral-3 dtype mismatch is unrelated to MPS histogram unsupported for Int." + "reason": "Different subsystem and hardware-specific failures." }, { "left": "issue:45507", "right": "issue:45685", "accept": false, - "reason": "GraniteMoEHybrid invalid method is unrelated to the MPS Int histogram limitation." + "reason": "GraniteMoEHybrid invalid method is unrelated to the MPS histogram Int limitation." }, { "left": "issue:44292", "right": "issue:44918", "accept": false, - "reason": "Qwen NVFP4 runtime error and Qwen3.5/TRL embedding unpacking are different issues." + "reason": "Different Qwen-related failures with different triggers and stack behavior." }, { "left": "issue:43716", "right": "issue:44292", "accept": false, - "reason": "Different models and failures: Mistral-3 dtype mismatch vs Qwen NVFP4 runtime error." - }, - { - "left": "issue:45874", - "right": "issue:46294", - "accept": false, - "reason": "Gemma4 reverse-embedding lookup is unrelated to Qwen3.5 video feature/token mismatch." + "reason": "Different model and runtime issues." }, { - "left": "issue:41720", + "left": "issue:43596", "right": "issue:43638", - "accept": false, - "reason": "Qwen3 auto device-map cuda assert is unrelated to the BERT ZeRO-3 index-0-out-of-bounds bug." + "accept": true, + "reason": "Same underlying ZeRO-3 BERT initialization bug and essentially the same error report." } ] }, "evaluator_result": { - "accept": true, - "feedback": "Summary is grounded in the packet and stays conservative. The main duplicate-like subcluster around 43596/43638 is well supported by the titles, and the other soft-edge calls do not overstate similarity. No PR-mergeability issue applies here since these are issues, not PR pairs." + "accept": false, + "feedback": "Mostly conservative, but I\u2019d reject for precision: one soft-edge reason misstates the packet (issue:44322 is titled about Qwen3_5Config lacking num_attention_heads, not Qwen3.5_5Moe), which makes the evidence handling less exact than required. The summary is otherwise plausible, but the review needs to stay tighter to the provided titles." }, "evaluator_used": true, - "retried": false, + "retried": true, "accepted_nontrivial_soft_edge": true, "error_kind": null, "error_message": null @@ -477,7 +417,7 @@ "source": "llm", "cache_hit": false, "model": "gpt-5.4-mini?service_tier=flex", - "cluster_id": "cluster-10000-529", + "cluster_id": "cluster-10000-530", "nodes": [ "issue:30277", "issue:34928", @@ -493,7 +433,6 @@ "issue:43296", "issue:43425", "issue:43531", - "issue:43541", "issue:43596", "issue:43638", "issue:43792", @@ -525,7 +464,8 @@ "issue:46304", "issue:46402", "issue:46772", - "issue:47228" + "issue:46901", + "issue:47394" ], "soft_pairs": [ "issue:43827|issue:44512", @@ -537,8 +477,8 @@ "issue:43039|issue:44805", "issue:45800|issue:46304", "issue:43296|issue:46304", - "issue:43866|issue:46402", "issue:39290|issue:41720", + "issue:43866|issue:46402", "issue:41762|issue:43596", "issue:45237|issue:45901", "issue:44589|issue:46304", @@ -552,6 +492,7 @@ "issue:44863|issue:46015", "issue:35308|issue:42617", "issue:45812|issue:45920", + "issue:46901|issue:47394", "issue:44292|issue:46015", "issue:43872|issue:46304", "issue:39290|issue:43531", @@ -559,64 +500,60 @@ "issue:41093|issue:43638", "issue:43792|issue:45901", "issue:43854|issue:43866", - "issue:43425|issue:45070", - "issue:43541|issue:47228", - "issue:43975|issue:45901", - "issue:43596|issue:45800", - "issue:43792|issue:45070" + "issue:43425|issue:45070" ], - "prepared_review_unit_hash": "00953a4c5351dac17a8ca2346d9429f172d72a69f910384cd7732a1cccd27992", + "prepared_review_unit_hash": "ca06b3e7403c76a2f7235fb71c4a34eee0e0e1b8ce5f6b6c75b229347550b0ee", "component_budget": { - "node_count": 529, - "item_count": 529, - "soft_pair_count": 1953, - "serialized_chars": 1012074, - "estimated_input_tokens": 253019, - "estimated_eval_tokens": 506294 + "node_count": 530, + "item_count": 530, + "soft_pair_count": 1954, + "serialized_chars": 1013004, + "estimated_input_tokens": 253251, + "estimated_eval_tokens": 506758 }, "budget": { "node_count": 47, "item_count": 47, - "soft_pair_count": 36, - "serialized_chars": 37817, - "estimated_input_tokens": 9455, - "estimated_eval_tokens": 19166 + "soft_pair_count": 33, + "serialized_chars": 36962, + "estimated_input_tokens": 9241, + "estimated_eval_tokens": 18738 }, "overflow_policy": "truncate_then_skip", "trimmed": false, "aggressively_trimmed": false, "split": true, "analyst_result": { - "summary": "This cluster is mostly heterogeneous Transformers regressions and docs issues. Only two soft edges look like true duplicates: the AutoTokenizer wrong-token-ID regression reports, and the DeepSpeed ZeRO-3 index-out-of-bounds loading reports.", - "confidence": 0.72, - "canonical_issue_reason": "Issue 45920 is the best canonical issue for the main duplicate group: it describes the broadest version of the same AutoTokenizer wrong-token-ID regression across multiple model families, not just a single model.", + "summary": "Most pairs are unrelated despite surface-level keyword similarity. I only accept a few where the titles point to the same concrete regression/path: Zero-3 init index-out-of-bounds, AutoTokenizer wrong IDs, and double-shifted training loss.", + "confidence": 0.78, + "canonical_issue_reason": "issue:45920 is the strongest canonical issue candidate: it describes the broad v5 AutoTokenizer regression across multiple models, and it cleanly subsumes the narrower Granite-only report (#45812).", "canonical_pr_reason": null, - "best_issue_reason": "45920 is the most representative issue in the cluster because it generalizes the tokenizer regression across several affected models and reads like the umbrella report for that bug.", + "best_issue_reason": "issue:45920 is the best overall issue to keep as the canonical duplicate target because it is broad, specific about the regression, and already aggregates community evidence.", "best_pr_reason": null, "soft_edge_verdicts": [ { "left": "issue:43827", "right": "issue:44512", "accept": false, - "reason": "Both are docs-related, but they concern different removals and different pages; not the same underlying issue." + "reason": "Both are docs-related, but they refer to different removed surfaces (pipeline() vs run command) and do not describe the same bug." }, { "left": "issue:43828", "right": "issue:45685", "accept": false, - "reason": "Different failure modes and backends: autocast dtype mismatch vs MPS histogram not implemented for Int." + "reason": "Different failures and code paths: autocast dtype mismatch vs MPS histogram op not implemented for Int." }, { "left": "issue:43039", "right": "issue:43828", "accept": false, - "reason": "Liger Kernel cross_entropy call and Phi-tiny-MoE autocast dtype mismatch are unrelated bugs." + "reason": "One is a Liger Kernel cross-entropy complaint; the other is an autocast dtype mismatch. Too different." }, { "left": "issue:30277", "right": "issue:36331", "accept": false, - "reason": "Different subsystems and errors: distributed collective mismatch vs CustomTrainer signature change." + "reason": "Unrelated training/distributed errors: collective mismatch with ZeRO3 vs compute_loss API breakage." }, { "left": "issue:34928", @@ -634,181 +571,169 @@ "left": "issue:43039", "right": "issue:44805", "accept": false, - "reason": "Different bugs: Liger Kernel cross_entropy usage vs mask/tensor shape mismatch." + "reason": "Both mention shape/mask problems, but the underlying reports are different and not the same code-path." }, { "left": "issue:45800", "right": "issue:46304", "accept": false, - "reason": "Both are version-compatibility complaints, but one is torch 2.4.1 vs transformers 5.8.0 and the other is a distributed-build regression in v5.9.0; not the same bug." + "reason": "These are separate compatibility regressions: torch/transformers versioning vs USE_DISTRIBUTED=0 support." }, { "left": "issue:43296", "right": "issue:46304", "accept": false, - "reason": "PaddleOCR-VL loading in vLLM and USE_DISTRIBUTED=0 support are different code paths and different breakages." + "reason": "Different model-loading failure in vLLM vs distributed-build incompatibility." }, { - "left": "issue:43866", - "right": "issue:46402", + "left": "issue:39290", + "right": "issue:41720", "accept": false, - "reason": "Checkpoint corruption and missing text-encoder weights are different model-loading problems." + "reason": "Different model families and different failures: missing sliding_window_pattern vs cudaErrorAssert in auto device mapping." }, { - "left": "issue:39290", - "right": "issue:41720", + "left": "issue:43866", + "right": "issue:46402", "accept": false, - "reason": "Gemma3 sliding_window_pattern in vLLM and Qwen3 auto-device-map cudaErrorAssert are unrelated." + "reason": "Checkpoint corruption/loader failure for Ovis2 is not the same issue as Sam3Model weights not loading." }, { "left": "issue:41762", "right": "issue:43596", "accept": true, - "reason": "Both report the same ZeRO-3 loading path failing with index 0 out of bounds from a zero-sized tensor; model class differs, but the underlying bug looks the same." + "reason": "Both report the same underlying DeepSpeed ZeRO-3 index-out-of-bounds on loading a model, with nearly identical failure mode and code path." }, { "left": "issue:45237", "right": "issue:45901", "accept": false, - "reason": "GPU/model compatibility failure and table-question-answering crash are unrelated." + "reason": "Different tasks and failure modes; no evidence they share the same underlying bug." }, { "left": "issue:44589", "right": "issue:46304", "accept": false, - "reason": "Float8 storage lookup and distributed-build support regression are different issues." + "reason": "Float8 storage lookup error is unrelated to the distributed build compatibility regression." }, { "left": "issue:43901", "right": "issue:44512", "accept": false, - "reason": "TextClassificationPipeline docs mention return_all_scores, while the other issue is about a removed run command; separate docs problems." + "reason": "Both are docs issues, but they concern different documented behaviors and removals." }, { "left": "issue:43039", "right": "issue:44560", "accept": false, - "reason": "Liger Kernel cross_entropy and Qwen3.5 video chat-template mismatch are unrelated." + "reason": "Cross-entropy being called by Liger Kernel is unrelated to Qwen3-VL video StopIteration." }, { "left": "issue:44841", "right": "issue:45823", "accept": false, - "reason": "Processor failure for Voxtral and Gemma4 PLE device mismatch are different model-specific bugs." + "reason": "Different models and failures: processor handling vs device-mapped forward mismatch." }, { "left": "issue:45823", "right": "issue:46294", "accept": false, - "reason": "Device placement bug in Gemma4 and video token/feature mismatch in Qwen3.5-9B are unrelated." + "reason": "Different subsystems and symptoms: device_map forward mismatch vs chat-template video token mismatch." }, { "left": "issue:43975", "right": "issue:46015", "accept": false, - "reason": "Incorrect detokenization and outdated RAG examples are unrelated." + "reason": "Detokenization regression is unrelated to outdated RAG examples." }, { "left": "issue:43866", "right": "issue:46772", "accept": false, - "reason": "Corrupted Ovis2 checkpoint and NVFP4 compatibility for DiffusionGemma are different issues." + "reason": "Checkpoint corruption vs ModelOpt/NVFP4 compatibility are different problems." }, { "left": "issue:45020", "right": "issue:46304", "accept": false, - "reason": "Remote-code breakage and USE_DISTRIBUTED=0 support regression are different compatibility problems." + "reason": "Remote-code regressions and distributed-build incompatibility are not the same bug." }, { "left": "issue:44863", "right": "issue:46015", "accept": false, - "reason": "NemotronH checkpoint loading and outdated RAG examples are unrelated." + "reason": "NemotronH checkpoint loading and RAG docs are unrelated." }, { "left": "issue:35308", "right": "issue:42617", "accept": false, - "reason": "IterableDataset length mismatch and a generic 3d_parallel.py failure do not look like the same bug." + "reason": "Both involve multi-GPU contexts, but the actual crashes are different and not clearly the same code path." }, { "left": "issue:45812", "right": "issue:45920", "accept": true, - "reason": "Both describe the same silent AutoTokenizer v5 regression producing wrong token IDs; 45920 is just the broader multi-model report." + "reason": "Both describe the same v5 AutoTokenizer regression producing wrong token IDs; #45920 is the broader report and #45812 is a narrower model-specific instance." + }, + { + "left": "issue:46901", + "right": "issue:47394", + "accept": true, + "reason": "Both point to the same training-loss double-shift regression in causal LM code; the GitForCausalLM report adds an extra symptom, but the core bug matches." }, { "left": "issue:44292", "right": "issue:46015", "accept": false, - "reason": "Qwen-3-8B-NVFP4 runtime failure and RAG docs issues are unrelated." + "reason": "Qwen-3-8B-NVFP4 runtime failure is unrelated to outdated RAG docs." }, { "left": "issue:43872", "right": "issue:46304", "accept": false, - "reason": "bitsandbytes Int8Params incompatibility is unrelated to distributed-build support changes." + "reason": "bitsandbytes Int8Params incompatibility is unrelated to the distributed build issue." }, { "left": "issue:39290", "right": "issue:43531", "accept": false, - "reason": "Different sliding-window bugs on different model families and different contexts; not the same underlying issue." + "reason": "Both mention sliding_window, but they affect different model/config paths and are not clearly the same bug." }, { "left": "issue:43866", "right": "issue:44863", "accept": false, - "reason": "Corrupted checkpoint vs model implementation unable to load checkpoints are not the same bug." + "reason": "Different model-loading failures; no evidence of a shared root cause." }, { "left": "issue:41093", "right": "issue:43638", "accept": false, - "reason": "Mask/tensor shape mismatch is a different failure from the ZeRO-3 BertModel index error." + "reason": "Similar-looking indexing errors, but one is a mask shape mismatch and the other is a ZeRO-3 zero-size index error." }, { "left": "issue:43792", "right": "issue:45901", "accept": false, - "reason": "Whisper loading failure and table-question-answering crash are unrelated." + "reason": "Whisper loading/runtime failure is unrelated to table-question-answering crashes." }, { "left": "issue:43854", "right": "issue:43866", "accept": false, - "reason": "GLM-4.7-Flash unit-test load failure and Ovis2 checkpoint corruption are different problems." + "reason": "Unit-test loading failure for GLM-4.7-Flash is not the same as a corrupted Ovis2 checkpoint." }, { "left": "issue:43425", "right": "issue:45070", "accept": false, - "reason": "Torch 2.10 incompatibility and v5.4.0 pydantic PretrainedConfig breakage are unrelated." - }, - { - "left": "issue:43541", - "right": "issue:47228", - "accept": false, - "reason": "Mixtral grouped_mm tracing failure and Sam3 bf16 compilation regression are different model/compiler issues." - }, - { - "left": "issue:43596", - "right": "issue:45800", - "accept": false, - "reason": "ZeRO-3 index-out-of-bounds during model init and torch/transformers version incompatibility are not the same bug." - }, - { - "left": "issue:43792", - "right": "issue:45070", - "accept": false, - "reason": "Whisper loading failure and pydantic PretrainedConfig regression are unrelated." + "reason": "Torch 2.10 compatibility and a pydantic PretrainedConfig regression are unrelated." } ] }, "evaluator_result": { "accept": true, - "feedback": "The summary is grounded in the packet: 45920 is a reasonable canonical issue for the AutoTokenizer v5 regression cluster, and the two accepted soft edges are the only ones that look well-supported as duplicates from the titles alone. The reject calls are conservative and do not overclaim shared causality." + "feedback": "The summary is conservative and grounded in the packet: it only accepts a few pairs that clearly share the same concrete failure mode (ZeRO-3 index-out-of-bounds, AutoTokenizer wrong IDs, and double-shifted training loss), and the canonical-issue choice of issue:45920 over issue:45812 is well supported by the broader title and the duplicate relationship. I did not see evidence of overstatement in the remaining rejects." }, "evaluator_used": true, "retried": false, @@ -827,7 +752,7 @@ "source": "llm", "cache_hit": false, "model": "gpt-5.4-mini?service_tier=flex", - "cluster_id": "cluster-10000-529", + "cluster_id": "cluster-10000-530", "nodes": [ "issue:36010", "issue:39186", @@ -842,11 +767,12 @@ "issue:43541", "issue:43572", "issue:43596", - "issue:43645", "issue:43716", + "issue:43792", "issue:43828", "issue:43866", "issue:43872", + "issue:43975", "issue:43994", "issue:44291", "issue:44292", @@ -878,6 +804,10 @@ "issue:47228" ], "soft_pairs": [ + "issue:43541|issue:47228", + "issue:43975|issue:45901", + "issue:43596|issue:45800", + "issue:43792|issue:45070", "issue:45362|issue:45464", "issue:43541|issue:46869", "issue:44292|issue:45901", @@ -889,8 +819,8 @@ "issue:44841|issue:45084", "issue:45507|issue:46869", "issue:43404|issue:43866", - "issue:45399|issue:46661", "issue:41093|issue:41762", + "issue:45399|issue:46661", "issue:43572|issue:45070", "issue:45800|issue:45823", "issue:43872|issue:47228", @@ -914,181 +844,204 @@ "issue:45759|issue:45902", "issue:46640|issue:46661", "issue:39186|issue:44918", - "issue:43716|issue:44918", - "issue:43645|issue:45399" + "issue:43716|issue:44918" ], - "prepared_review_unit_hash": "2fb91fc3dcdba710c64c417a128321bea99f27e55934e2597b45894484b18b5c", + "prepared_review_unit_hash": "fdb555bdfbfc411514e38acdf9b4f09251f604e6cb009d2a960a42661a9f5ce0", "component_budget": { - "node_count": 529, - "item_count": 529, - "soft_pair_count": 1953, - "serialized_chars": 1012074, - "estimated_input_tokens": 253019, - "estimated_eval_tokens": 506294 + "node_count": 530, + "item_count": 530, + "soft_pair_count": 1954, + "serialized_chars": 1013004, + "estimated_input_tokens": 253251, + "estimated_eval_tokens": 506758 }, "budget": { - "node_count": 47, - "item_count": 47, - "soft_pair_count": 38, - "serialized_chars": 38436, - "estimated_input_tokens": 9609, - "estimated_eval_tokens": 19474 + "node_count": 48, + "item_count": 48, + "soft_pair_count": 41, + "serialized_chars": 39837, + "estimated_input_tokens": 9960, + "estimated_eval_tokens": 20176 }, "overflow_policy": "truncate_then_skip", "trimmed": false, "aggressively_trimmed": false, "split": true, "analyst_result": { - "summary": "Mostly heterogeneous issue cluster with one clearly supported duplicate family: Qwen video-input/template regressions. The SigLIP2 pair is too weak to merge confidently and is rejected.", - "confidence": 0.83, - "canonical_issue_reason": "No single canonical issue for the whole set; the only well-supported duplicate family is the Qwen video-input/template regression, anchored by 44479 with 46294 as a narrower manifestation.", + "summary": "Mostly unrelated issue set with one strong soft-duplicate pair around Qwen3.5 video-input regressions. I\u2019d use the broad remote-code regression report as the best issue representative.", + "confidence": 0.79, + "canonical_issue_reason": "issue:45020 is the broadest, most representative regression report in the set; it captures a general recent-version compatibility break affecting remote_code models, while most other candidates are narrower, model-specific symptoms.", "canonical_pr_reason": null, - "best_issue_reason": "issue:44479 is the best anchor because it describes the broader Qwen video-input regression across multiple Qwen3.5-family variants, and it subsumes the narrower 46294 report.", + "best_issue_reason": "issue:45020 is the best standalone representative because it is broad, cross-model, and framed as a general transformers regression rather than a single downstream symptom.", "best_pr_reason": null, "soft_edge_verdicts": [ + { + "left": "issue:43541", + "right": "issue:47228", + "accept": false, + "reason": "Different failures: Mixtral grouped_mm tracing error vs Sam3 bf16 compilation regression." + }, + { + "left": "issue:43975", + "right": "issue:45901", + "accept": false, + "reason": "Different subsystems and symptoms: detokenization bug vs table-question-answering crash." + }, + { + "left": "issue:43596", + "right": "issue:45800", + "accept": false, + "reason": "Zero-3/BertModel loading failure is unrelated to torch/transformers version incompatibility." + }, + { + "left": "issue:43792", + "right": "issue:45070", + "accept": false, + "reason": "Whisper runtime failure and a PretrainedConfig pydantic-field regression are not the same bug." + }, { "left": "issue:45362", "right": "issue:45464", "accept": false, - "reason": "Different Qwen3.5 failures: chat/model crash vs streaming video-input API failure." + "reason": "Both involve Qwen3.5, but one is a chat crash and the other is a streaming API failure; not clearly one code path." }, { "left": "issue:43541", "right": "issue:46869", "accept": false, - "reason": "Different models and code paths: Mixtral grouped_mm tracing vs YOLO multi-GPU device-map issues." + "reason": "Mixtral tracing/grouped_mm issue is unrelated to YOLO multi-GPU device-map behavior." }, { "left": "issue:44292", "right": "issue:45901", "accept": false, - "reason": "Different failures and subsystems: Qwen-3-8B-NVFP4 loading vs table-question-answering crashes." + "reason": "NVFP4 model load error is unrelated to table-question-answering crashes." }, { "left": "issue:39290", "right": "issue:45823", "accept": false, - "reason": "Both are Gemma/Gemma4-related, but one is a missing config attribute in vLLM and the other is a device mismatch in forward." + "reason": "Gemma3 sliding_window_pattern/vLLM issue is different from Gemma4 PLE device mismatch." }, { "left": "issue:45464", "right": "issue:45685", "accept": false, - "reason": "Streaming inference on Qwen3.5 is unrelated to the MPS histogram unsupported-int error." + "reason": "Qwen3.5 streaming inference failure is unrelated to the MPS histogram op error." }, { "left": "issue:36010", "right": "issue:39290", "accept": false, - "reason": "GenerationMixin import failure is unrelated to the Gemma3TextConfig missing-attribute bug." + "reason": "ImportError for GenerationMixin is unrelated to Gemma3 sliding_window_pattern behavior." }, { "left": "issue:43541", "right": "issue:45685", "accept": false, - "reason": "Different model families and failures: Mixtral grouped_mm runtime vs MPS histogram on int tensors." + "reason": "Different models and code paths: Mixtral tracing vs MPS MoE histogram failure." }, { "left": "issue:43541", "right": "issue:43828", "accept": false, - "reason": "Both mention MoE/dtype-related problems, but they affect different models and concrete code paths." + "reason": "Both are MoE-related, but grouped_mm tracing and autocast dtype mismatch are distinct failures." }, { "left": "issue:44841", "right": "issue:45084", "accept": false, - "reason": "Processor failure for Voxtral is not the same bug as the template-node compilation TypeError." + "reason": "Voxtral processor failure is unrelated to the non-template node compilation error." }, { "left": "issue:45507", "right": "issue:46869", "accept": false, - "reason": "GraniteMoEHybrid invalid-method call and YOLO device-map auto are different issues." + "reason": "GraniteMoEHybrid invalid method call is not the same bug as YOLO device-map auto failure." }, { "left": "issue:43404", "right": "issue:43866", "accept": false, - "reason": "A tied-weights bug in Mistral3 is unrelated to a corrupted Ovis2 checkpoint." + "reason": "Tied lm_head weights and a corrupted Ovis2 checkpoint are different problems." }, { - "left": "issue:45399", - "right": "issue:46661", + "left": "issue:41093", + "right": "issue:41762", "accept": false, - "reason": "Flash-Attn fallback logic and remote-code config_class failures are different code paths." + "reason": "Both are index/shape errors, but the mask-shape mismatch and ZeRO-3 empty-tensor index-out-of-bounds are distinct." }, { - "left": "issue:41093", - "right": "issue:41762", + "left": "issue:45399", + "right": "issue:46661", "accept": false, - "reason": "Similar IndexError wording, but different models and failure contexts; not enough evidence of one bug." + "reason": "Flash-Attn fallback gating and remote-code config_class errors do not share the same code path." }, { "left": "issue:43572", "right": "issue:45070", "accept": false, - "reason": "StableLm pad_token_idx regression is not the same as the PretrainedConfig pydantic-field breakage." + "reason": "Missing StableLmConfig pad_token_idx and a pydantic PretrainedConfig field regression are different config issues." }, { "left": "issue:45800", "right": "issue:45823", "accept": false, - "reason": "A version incompatibility report is not the same as a Gemma4 device-mismatch bug." + "reason": "Torch/transformers version incompatibility is unrelated to Gemma4 device placement mismatch." }, { "left": "issue:43872", "right": "issue:47228", "accept": false, - "reason": "bitsandbytes init-arg incompatibility and Sam3 bf16 compilation regression are unrelated." + "reason": "bitsandbytes Int8Params API incompatibility is unrelated to Sam3 compilation regression." }, { "left": "issue:44805", "right": "issue:44918", "accept": false, - "reason": "Both mention tensor-shape problems, but one is a generic mask mismatch and the other is a TRL embedding-unpacking failure." + "reason": "Both mention tensor shape problems, but one is a generic mask-shape error and the other is a Qwen3.5/TRl embedding unpacking issue." }, { "left": "issue:43828", "right": "issue:45464", "accept": false, - "reason": "Autocast dtype mismatch on Phi-tiny-MoE is unrelated to Qwen3.5 streaming inference." + "reason": "Autocast dtype mismatch in Phi-tiny-MoE is unrelated to Qwen3.5 streaming API failure." }, { "left": "issue:43716", "right": "issue:47228", "accept": false, - "reason": "Mistral-3 image-preprocessor dtype mismatch and Sam3 bf16 compilation regression are different bugs." + "reason": "Mistral-3 image preprocessor dtype mismatch and Sam3 bf16 compilation regression are unrelated." }, { "left": "issue:43296", "right": "issue:43366", "accept": false, - "reason": "PaddleOCR-VL load failure is unrelated to GGUF support for gpt-oss." + "reason": "PaddleOCR-VL loading in vLLM and GGUF gpt-oss support are different feature areas." }, { "left": "issue:43366", "right": "issue:45464", "accept": false, - "reason": "GGUF architecture support and Qwen3.5 streaming inference are unrelated." + "reason": "GGUF architecture support and Qwen3.5 streaming inference failure are unrelated." }, { "left": "issue:44291", "right": "issue:45070", "accept": false, - "reason": "Unexpected _is_hf_initialized in init_empty_weights is a different regression from the pydantic PretrainedConfig issue." + "reason": "init_empty_weights/_is_hf_initialized TypeError is a different failure from the pydantic PretrainedConfig field regression." }, { "left": "issue:43531", "right": "issue:45070", "accept": false, - "reason": "Qwen3-MoE sliding_window handling and pydantic PretrainedConfig breakage are not the same bug." + "reason": "Qwen3-MoE sliding_window behavior is unrelated to the PretrainedConfig pydantic-model bug." }, { "left": "issue:44387", "right": "issue:45005", "accept": false, - "reason": "int4 OOM under Transformers 5.x and tied-weights translation-model issues are different failures." + "reason": "int4 CUDA memory/OOM regression and tied-weights translation-model issues are different bugs." }, { "left": "issue:46869", @@ -1100,88 +1053,82 @@ "left": "issue:45020", "right": "issue:45800", "accept": false, - "reason": "Remote-code model breakage and torch/transformers incompatibility are distinct issue classes." + "reason": "Both are broad compatibility reports, but one is about remote_code models and the other is torch-version incompatibility." }, { "left": "issue:45507", "right": "issue:47228", "accept": false, - "reason": "GraniteMoEHybrid invalid method does not share a concrete code path with the Sam3 regression." + "reason": "GraniteMoEHybrid invalid method call is unrelated to Sam3 bf16 compilation." }, { "left": "issue:43716", "right": "issue:46294", "accept": false, - "reason": "Different multimodal bugs: Mistral-3 image dtype mismatch vs Qwen3.5 video token/feature mismatch." + "reason": "Different multimodal bugs: image preprocessor dtype mismatch vs video feature/token mismatch." }, { "left": "issue:41093", "right": "issue:43596", "accept": false, - "reason": "Both are IndexErrors, but one is a mask-shape mismatch and the other is a zero3 init out-of-bounds error." + "reason": "Mask-length mismatch and zero-size index-out-of-bounds during load are not the same issue." }, { "left": "issue:44479", "right": "issue:46294", "accept": true, - "reason": "Same Qwen video-input/template bug family: both describe mismatched video features/tokens in Qwen3.5-family processing." + "reason": "Both describe the same Qwen3.5-family video input regression, with 46294 looking like a specific manifestation of the broader bug in 44479." }, { "left": "issue:43054", "right": "issue:43994", "accept": false, - "reason": "Both are SigLIP2-related quality regressions, but the evidence only shows separate user-visible output problems, not one concrete bug." + "reason": "Both involve SigLIP2, but the reported symptoms differ enough that they are not clearly the same concrete bug." }, { "left": "issue:46468", "right": "issue:46869", "accept": false, - "reason": "Official Nemotron example breakage and YOLO device-map auto are unrelated." + "reason": "Nemotron example breakage is unrelated to YOLO multi-GPU device-map behavior." }, { "left": "issue:43596", "right": "issue:46165", "accept": false, - "reason": "Zero3 init index error and remote-code resume failure are different issues." + "reason": "ZeRO-3/BertModel loading failure and remote-code resume-training issue are different problems." }, { "left": "issue:45759", "right": "issue:45902", "accept": false, - "reason": "Both involve Qwen3.5 multimodal internals, but one is config unwrapping and the other is rotary-embedding materialization; different bugs." + "reason": "Composite-config unwrapping and uninitialized rotary-embedding memory are separate bugs." }, { "left": "issue:46640", "right": "issue:46661", "accept": false, - "reason": "Flash Attention 2 issue and remote-code config_class failure are unrelated." + "reason": "Flash Attention 2 issue and remote-code model_class/config_class error are unrelated." }, { "left": "issue:39186", "right": "issue:44918", "accept": false, - "reason": "FSDP weight-shape error is unrelated to the TRL SFT embedding-unpacking failure." + "reason": "FSDP 2-D weight error is unrelated to Qwen3.5/TRl embedding unpacking." }, { "left": "issue:43716", "right": "issue:44918", "accept": false, - "reason": "Mistral-3 image dtype mismatch and Qwen3.5 TRL embedding unpacking are different models and code paths." - }, - { - "left": "issue:43645", - "right": "issue:45399", - "accept": false, - "reason": "Jupyter custom-model initialization breakage is unrelated to the Flash-Attn fallback check ordering bug." + "reason": "Image-preprocessor dtype mismatch is unrelated to Qwen3.5 embedding unpacking under TRL." } ] }, "evaluator_result": { - "accept": false, - "feedback": "The summary is mostly grounded, but one soft-edge acceptance is too strong: issue:43054 \u2194 issue:43994. The packet only shows both are SigLIP2-related and user-visible quality/output problems, which is suggestive but not enough to confidently call them the same bug family. The Qwen video-input pair (44479 \u2194 46294) is conservative and well supported, but the SigLIP2 pair overstates the evidence, so I can\u2019t accept the cluster as reviewed." + "accept": true, + "feedback": "Grounded overall. The representative issue choice is conservative and supported by the packet, and the soft-edge verdicts generally avoid overstating duplicates. The one accepted pair (44479/46294) looks plausibly like the same Qwen3.5 video-input regression at different specificity levels." }, "evaluator_used": true, - "retried": true, + "retried": false, "accepted_nontrivial_soft_edge": true, "error_kind": null, "error_message": null @@ -1197,7 +1144,7 @@ "source": "llm", "cache_hit": false, "model": "gpt-5.4-mini?service_tier=flex", - "cluster_id": "cluster-10000-529", + "cluster_id": "cluster-10000-530", "nodes": [ "issue:33666", "issue:35308", @@ -1209,8 +1156,8 @@ "issue:42915", "issue:43054", "issue:43404", - "issue:43493", "issue:43644", + "issue:43645", "issue:43828", "issue:43872", "issue:43873", @@ -1249,6 +1196,7 @@ "issue:47333" ], "soft_pairs": [ + "issue:43645|issue:45399", "issue:40990|issue:44960", "issue:42915|issue:45399", "issue:45632|issue:45698", @@ -1268,8 +1216,8 @@ "issue:43950|issue:44534", "issue:36331|issue:41762", "issue:43644|issue:43950", - "issue:46032|issue:47246", "issue:33666|issue:35308", + "issue:46032|issue:47246", "issue:43872|issue:44589", "issue:46899|issue:46932", "issue:43054|issue:45701", @@ -1278,232 +1226,231 @@ "issue:47332|issue:47333", "issue:43828|issue:44928", "issue:45987|issue:46858", - "issue:43950|issue:45902", - "issue:43054|issue:43493" + "issue:43950|issue:45902" ], - "prepared_review_unit_hash": "1723bf3f91eac32eeeb15598e50253a544af1fa3a1182cc359b84ab5cc239d54", + "prepared_review_unit_hash": "8fd0cd03b24e8a44694e7bfc8ee18d0bb7a21d7f6d527c14e5f8780ed824a9a6", "component_budget": { - "node_count": 529, - "item_count": 529, - "soft_pair_count": 1953, - "serialized_chars": 1012074, - "estimated_input_tokens": 253019, - "estimated_eval_tokens": 506294 + "node_count": 530, + "item_count": 530, + "soft_pair_count": 1954, + "serialized_chars": 1013004, + "estimated_input_tokens": 253251, + "estimated_eval_tokens": 506758 }, "budget": { "node_count": 48, "item_count": 48, "soft_pair_count": 31, - "serialized_chars": 37112, - "estimated_input_tokens": 9278, - "estimated_eval_tokens": 18812 + "serialized_chars": 37131, + "estimated_input_tokens": 9283, + "estimated_eval_tokens": 18822 }, "overflow_policy": "truncate_then_skip", "trimmed": false, "aggressively_trimmed": false, "split": true, "analyst_result": { - "summary": "The cluster is mostly a set of unrelated reports with a few true duplicate groups: the v5 non-persistent-buffer corruption regression (#43644/#43950/#44534), the cache.get_seq_length contract bug (#45987/#46858), Gemma4 quantized embedder dtype-casting (#46899/#46932), and the disk-offloaded save failure (#47332/#47333). #43644 is the earliest report in the non-persistent-buffer group, while #43950 is the clearest explicit regression/duplicate target.", - "confidence": 0.88, - "canonical_issue_reason": "#43950 is the clearest canonical representative for the non-persistent-buffer regression because it states the from_pretrained corruption bug explicitly and is directly matched by #44534; however, #43644 is the earlier report in the same cluster.", + "summary": "This cluster is mostly heterogeneous, but it contains a few clear duplicate sets: the v5 non-persistent-buffer regression, the Gemma4 quantized embedder dtype-cast bug, the Qwen3 offloading save failure, and the StaticCache get_seq_length contract issue. A couple of other pairs are thematically related but not the same concrete bug.", + "confidence": 0.71, + "canonical_issue_reason": "Issue #43950 is the clearest and broadest statement of the shared v5 non-persistent-buffer regression, and it best covers the closely related follow-ups.", "canonical_pr_reason": null, - "best_issue_reason": "#43950 is the best overall issue to anchor this cluster because it most explicitly describes the concrete v5 corruption regression and sits at the center of the duplicate set, even though #43644 predates it.", + "best_issue_reason": "#43950 is the strongest hub issue for the cluster\u2019s duplicate set around non-persistent buffer corruption and has the cleanest root-cause framing.", "best_pr_reason": null, "soft_edge_verdicts": [ + { + "left": "issue:43645", + "right": "issue:45399", + "accept": false, + "reason": "Unrelated problems: custom model initialization in notebooks vs flash-attn fallback gating." + }, { "left": "issue:40990", "right": "issue:44960", "accept": false, - "reason": "Different models and failure modes; one is GPT-OSS perplexity, the other is a GLM5-specific issue." + "reason": "Different bugs and models: high perplexity on gpt-oss vs GLM5 model issue." }, { "left": "issue:42915", "right": "issue:45399", "accept": false, - "reason": "Both touch quantization/fallback broadly, but they describe different concrete code-path failures." + "reason": "FP8 quantization failure on Qwen3Moe is unrelated to flash-attn fallback checks." }, { "left": "issue:45632", "right": "issue:45698", "accept": false, - "reason": "Both involve remote-code loading, but one is a leaf-directory cache collision and the other loads the wrong custom module after save/load." + "reason": "Both involve custom-module loading, but the symptoms and failure modes differ enough to avoid treating as the same bug." }, { "left": "issue:43404", "right": "issue:45005", "accept": false, - "reason": "Both mention tied weights, but the actual bugs are different: a specific Mistral3 weight-tying issue versus a broader translation-model regression." + "reason": "Both mention tied weights, but they target different model families and code paths." }, { "left": "issue:43873", "right": "issue:45005", "accept": false, - "reason": "Offloading/quantization behavior is unrelated to the tied-weights regression." + "reason": "Quantization/offloading behavior is unrelated to tied-weights regressions." }, { "left": "issue:44360", "right": "issue:44512", "accept": false, - "reason": "A DSA indexer ReLU discussion is unrelated to a docs typo about a removed run command." + "reason": "DSA indexer discussion is unrelated to a docs typo about the removed run command." }, { "left": "issue:35308", "right": "issue:36331", "accept": false, - "reason": "Multi-GPU IterableDataset crashes are unrelated to the custom trainer compute_loss signature breakage." + "reason": "Different failure classes: iterable-dataset multi-GPU crash vs compute_loss signature change." }, { "left": "issue:44492", "right": "issue:44512", "accept": false, - "reason": "A cache-strategy typo is unrelated to outdated docs mentioning a removed command." + "reason": "A cache-strategy typo is not the same as outdated docs about the removed run command." }, { "left": "issue:44918", "right": "issue:45507", "accept": false, - "reason": "Different models and different failures; no shared underlying code-path is evident." + "reason": "Different model-specific runtime failures; no shared concrete bug." }, { "left": "issue:44530", "right": "issue:45005", "accept": false, - "reason": "PagedAttentionCache linear_attention failure is unrelated to tied-weight handling." + "reason": "PagedAttentionCache linear_attention crash is unrelated to tied-weight handling." }, { "left": "issue:43975", "right": "issue:44292", "accept": false, - "reason": "Tokenizer detokenization regression and Qwen-3 NVFP4 runtime error are different problems." + "reason": "Tokenizer detokenization issues for different models are not the same bug." }, { "left": "issue:45183", "right": "issue:46463", "accept": false, - "reason": "Related validation area, but the T5 decoder-input error is not clearly the same bug as the broad XOR validation regression." + "reason": "Both are validation/error-message issues, but they affect different validation paths and are not clearly the same underlying defect." }, { "left": "issue:44863", "right": "issue:45005", "accept": false, - "reason": "Checkpoint loading failure for NemotronH is unrelated to tied-weights on translation models." + "reason": "Checkpoint loading for NemotronH is unrelated to tied-weight regressions." }, { "left": "issue:35308", "right": "issue:42915", "accept": false, - "reason": "Different bugs: iterable-dataset multi-GPU crash versus FP8/quantization failure." + "reason": "Completely different topics: dataloader/multi-GPU crash vs FP8 quantization failure." }, { "left": "issue:39692", "right": "issue:43054", "accept": false, - "reason": "Both are SigLIP2-related, but one is a docs example error and the other is a model quality discrepancy." + "reason": "SigLIP2 docs errors and SigLIP2 embedding quality are related to the same model family but not the same concrete bug." }, { "left": "issue:46489", "right": "issue:46491", "accept": false, - "reason": "Both are tokenizer regressions, but they affect different models and different symptoms." + "reason": "Both are tokenizer regressions, but they hit different models and different observed behavior." }, { "left": "issue:43950", "right": "issue:44534", "accept": true, - "reason": "Near-verbatim duplicate reports of the same v5 non-persistent-buffer corruption regression." + "reason": "Same non-persistent-buffer regression in v5; 44534 is a near-duplicate phrasing of the corruption/junk-buffer issue." }, { "left": "issue:36331", "right": "issue:41762", "accept": false, - "reason": "Different bugs: custom-trainer kwarg mismatch versus DeepSpeed ZeRO-3 loading crash." + "reason": "Unexpected compute_loss kwarg and ZeRO-3 load-time index error are unrelated." }, { "left": "issue:43644", "right": "issue:43950", "accept": true, - "reason": "Same underlying v5 regression corrupting non-persistent buffers; #43644 is the earlier report and #43950 is the more explicit duplicate target." + "reason": "Same v5 non-persistent-buffer corruption bug, with 43644 describing the same junk-filled buffer symptom." }, { - "left": "issue:46032", - "right": "issue:47246", + "left": "issue:33666", + "right": "issue:35308", "accept": false, - "reason": "Both mention Mamba2-style recurrence, but one is a generic use_cache correctness bug and the other is a NemotronH slow-path propagation issue." + "reason": "Both mention multi-GPU training, but the model and failure mechanism are different." }, { - "left": "issue:33666", - "right": "issue:35308", + "left": "issue:46032", + "right": "issue:47246", "accept": false, - "reason": "Both mention multi-GPU training, but they are different failure scenarios and code paths." + "reason": "Both concern Mamba2-family recurrence, but one is a generic use_cache correctness bug and the other is a NemotronH slow-path propagation issue." }, { "left": "issue:43872", "right": "issue:44589", "accept": false, - "reason": "Different bitsandbytes/storage compatibility failures with no shared concrete bug." + "reason": "Different quantization/storage errors with different root causes." }, { "left": "issue:46899", "right": "issue:46932", "accept": true, - "reason": "Same Gemma4 quantized vision/audio embedder dtype-casting bug; #46932 explicitly says it is the same issue." + "reason": "Same Gemma4 quantized vision/audio embedder dtype-cast bug; 46932 is the follow-up fix issue." }, { "left": "issue:43054", "right": "issue:45701", "accept": false, - "reason": "SigLIP2 embedding quality complaints are unrelated to tokenization changing across transformers versions." + "reason": "SigLIP2 embedding quality and tokenization-version changes are different bugs." }, { "left": "issue:45920", "right": "issue:46489", "accept": false, - "reason": "Both are tokenizer regressions, but they involve different model families and different underlying fixes." + "reason": "Both are tokenizer regressions, but the affected models and failure modes differ." }, { "left": "issue:38175", "right": "issue:43054", "accept": false, - "reason": "Both are SigLIP2-related, but one is zero probabilities and the other is degraded text embeddings." + "reason": "Unexpected zero probabilities and worse text embeddings are not the same concrete SigLIP2 issue." }, { "left": "issue:47332", "right": "issue:47333", "accept": true, - "reason": "Same disk-offloaded save failure; only the model name differs." + "reason": "Same offloading save failure for disk-offloaded Qwen3 models; only the model name differs." }, { "left": "issue:43828", "right": "issue:44928", "accept": false, - "reason": "Autocast dtype mismatch and Qwen3.5 gradient explosion are unrelated failures." + "reason": "Autocast dtype mismatch and RLHF NaN explosion from 3D position_ids are different bugs." }, { "left": "issue:45987", "right": "issue:46858", "accept": true, - "reason": "Same cache.get_seq_length contract bug: it returns a tensor instead of an int-like length, which breaks static-cache arithmetic." + "reason": "Same StaticCache.get_seq_length contract bug: it returns a tensor instead of a plain int, breaking downstream arithmetic." }, { "left": "issue:43950", "right": "issue:45902", - "accept": false, - "reason": "Non-persistent-buffer corruption is unrelated to Qwen3.5 MoE rotary embedding uninitialized-memory reads." - }, - { - "left": "issue:43054", - "right": "issue:43493", - "accept": false, - "reason": "Different SigLIP2 problems: relative text-embedding quality versus HF/JAX implementation discrepancy." + "accept": true, + "reason": "Same non-persistent-buffer regression family; 45902 is a specific manifestation via uninitialized `inv_freq` after materialization." } ] }, "evaluator_result": { - "accept": false, - "feedback": "The cluster summary is mostly grounded, but one key statement is not: issue #43950 is not the earliest report in the non-persistent-buffer group, since #43644 was created earlier. That makes the \"likely earliest report\" rationale overstated. The rest of the duplicate judgments look conservative enough, but this precision error means I can't accept the output as written." + "accept": true, + "feedback": "Grounded overall: the accepted duplicate pairs are supported by the titles and the cluster summary stays conservative. The rejected soft edges are mostly appropriately cautious, with no clear overstatement of evidence." }, "evaluator_used": true, - "retried": true, + "retried": false, "accepted_nontrivial_soft_edge": true, "error_kind": null, "error_message": null @@ -1519,7 +1466,7 @@ "source": "llm", "cache_hit": false, "model": "gpt-5.4-mini?service_tier=flex", - "cluster_id": "cluster-10000-529", + "cluster_id": "cluster-10000-530", "nodes": [ "issue:24643", "issue:28530", @@ -1528,7 +1475,9 @@ "issue:42491", "issue:42915", "issue:42947", + "issue:43054", "issue:43404", + "issue:43493", "issue:43643", "issue:43782", "issue:43824", @@ -1566,15 +1515,14 @@ "issue:46315", "issue:46489", "issue:46661", - "issue:46869", - "issue:46894", - "issue:46897" + "issue:46869" ], "soft_pairs": [ + "issue:43054|issue:43493", "issue:44910|issue:45859", "issue:45663|issue:45902", - "issue:42947|issue:44387", "issue:44403|issue:44451", + "issue:42947|issue:44387", "issue:24643|issue:28530", "issue:43975|issue:45237", "issue:44403|issue:44488", @@ -1600,233 +1548,229 @@ "issue:45488|issue:46489", "issue:43404|issue:43950", "issue:43827|issue:44509", - "issue:43881|issue:43883", - "issue:46894|issue:46897" + "issue:43881|issue:43883" ], - "prepared_review_unit_hash": "d353e95ce8280280e3e320c38b3c5e0643687cb3fc1d929abc00249012ac2b5a", + "prepared_review_unit_hash": "47656424ac07705379ab1645ef65bc607edbabd3c8fa8245c715d5190d698238", "component_budget": { - "node_count": 529, - "item_count": 529, - "soft_pair_count": 1953, - "serialized_chars": 1012074, - "estimated_input_tokens": 253019, - "estimated_eval_tokens": 506294 + "node_count": 530, + "item_count": 530, + "soft_pair_count": 1954, + "serialized_chars": 1013004, + "estimated_input_tokens": 253251, + "estimated_eval_tokens": 506758 }, "budget": { "node_count": 48, "item_count": 48, "soft_pair_count": 31, - "serialized_chars": 37495, - "estimated_input_tokens": 9374, - "estimated_eval_tokens": 19004 + "serialized_chars": 37489, + "estimated_input_tokens": 9373, + "estimated_eval_tokens": 19002 }, "overflow_policy": "truncate_then_skip", "trimmed": false, "aggressively_trimmed": false, "split": true, "analyst_result": { - "summary": "This set is mostly heterogeneous. I found three likely duplicate pairs: stale v5 docs about removed pipeline tasks, remote-code model loading regressions, and the Qwen/Florence/Cohere training-loss double-shift pattern. Most other pairs look like different bugs that only share a broad subsystem or model family.", - "confidence": 0.81, - "canonical_issue_reason": "Issue 45020 is the best umbrella issue in this set: it describes the broad remote-code loading regression class, which cleanly subsumes the more specific AutoConfig/AutoModel remote-code failures.", + "summary": "The cluster is semantically related but not made of true duplicates: most items are separate regressions around model loading, tokenizers, or device/memory behavior. I would use issue 45020 as the best umbrella representative for the broad remote-code loading regression theme, but I would not merge any of the soft-edge pairs.", + "confidence": 0.69, + "canonical_issue_reason": "issue:45020 is the broadest and most central representative here: it describes a general class of recent-version breakages for remote-code models, which matches several of the loading-failure reports better than the narrower, model-specific issues.", "canonical_pr_reason": null, - "best_issue_reason": "45020 is the most representative and reusable canonical issue because it names the broad underlying regression rather than a single model-specific symptom.", + "best_issue_reason": "issue:45020 is the best single issue to summarize the dominant cluster theme because it captures the widest shared symptom set without being tied to one specific model or subsystem.", "best_pr_reason": null, "soft_edge_verdicts": [ + { + "left": "issue:43054", + "right": "issue:43493", + "accept": false, + "reason": "Both concern SigLIP2 text-embedding quality/implementation, but one is a performance complaint and the other is a HF-vs-JAX discrepancy report; not enough evidence they are the same concrete bug." + }, { "left": "issue:44910", "right": "issue:45859", "accept": false, - "reason": "Different failures: one is a Flash Attention crash on Qwen3.5 with 3D position_ids, the other is a CPU-offload incompatibility in Qwen3_5MoeTextRotaryEmbedding." + "reason": "Both mention Qwen3.5/MoE and runtime behavior, but one is a FlashAttention crash from 3D position_ids handling while the other is CPU-offload incompatibility in rotary embeddings; different failure paths." }, { "left": "issue:45663", "right": "issue:45902", "accept": false, - "reason": "Both mention MoE/Gemma/Qwen materialization issues, but they are different models and different bugs (FSDP2 KeyError vs uninitialized inv_freq after to_empty)." + "reason": "Both involve Gemma4/Qwen-style training edge cases, but the KeyError in FSDP2 attention and the uninitialized-memory bug after meta-to-empty materialization are distinct defects." }, { - "left": "issue:42947", - "right": "issue:44387", + "left": "issue:44403", + "right": "issue:44451", "accept": false, - "reason": "One is about gradient checkpointing being ineffective with LoRA, the other is an int4 quantization memory regression leading to OOM." + "reason": "Generic loading noise vs inability to load a specific model; the titles do not point to the same underlying bug." }, { - "left": "issue:44403", - "right": "issue:44451", + "left": "issue:42947", + "right": "issue:44387", "accept": false, - "reason": "Generic load-time noise vs a specific model loading failure for ScandiBERT; not the same bug." + "reason": "Gradient checkpointing with LoRA and increased CUDA reserved memory under int4 quantization are unrelated training regressions." }, { "left": "issue:24643", "right": "issue:28530", "accept": false, - "reason": "Unrelated training/deepspeed tensor shape error vs early-stopping metric lookup warning." + "reason": "DeepSpeed weight-dimension error vs early-stopping metric configuration warning; completely different subsystems." }, { "left": "issue:43975", "right": "issue:45237", "accept": false, - "reason": "Tokenizer detokenization regression is unrelated to GPT-OSS AMD GPU execution failure." + "reason": "Tokenizer detokenization regression and AMD GPU model-loading failure are unrelated." }, { "left": "issue:44403", "right": "issue:44488", "accept": false, - "reason": "Both are model loading reports, but one is about noisy logging and the other is a concrete tokenizer/model load failure." + "reason": "These are both model-loading complaints, but they target different models and likely different root causes." }, { "left": "issue:41762", "right": "issue:42915", "accept": false, - "reason": "Different load-time failures in different model families (Gemma3 ZeRO-3 vs Qwen3Moe FP8Config)." + "reason": "Gemma3 ZeRO-3 loading crash and Qwen3Moe FineGrainedFP8 failure are separate model-specific issues." }, { "left": "issue:43643", "right": "issue:45020", - "accept": true, - "reason": "Same remote-code loading regression class: both report pretrained loading/config objects missing expected fields or behavior when trust_remote_code is used." + "accept": false, + "reason": "Missing fields from AutoConfig with trust_remote_code=True is related to remote-code loading, but it is a narrower configuration bug, not the same concrete issue as the broader remote-code regression report." }, { "left": "issue:44991", "right": "issue:45399", "accept": false, - "reason": "Tokenizer loading failure and flash-attn fallback gating are unrelated." + "reason": "Tokenizer loading failure for one model and flash-attn fallback blocking for another are different code paths." }, { "left": "issue:42947", "right": "issue:43856", "accept": false, - "reason": "Different issues: checkpointing behavior with LoRA vs memory usage during Qwen3 MoE training." + "reason": "LoRA checkpointing behavior and Qwen3 MoE memory usage are different problems despite both involving training efficiency." }, { "left": "issue:45081", "right": "issue:45399", "accept": false, - "reason": "Tokenizer regex crash vs flash-attn fallback blocked by checks; different code paths." + "reason": "Mistral regex patch crash in tokenizer loading and flash-attn fallback gating are unrelated." }, { "left": "issue:43824", "right": "issue:43872", "accept": false, - "reason": "Importing a missing model class is unrelated to the bitsandbytes Int8Params constructor incompatibility." + "reason": "ImportError for a missing VL class and bitsandbytes Int8Params incompatibility are distinct API/regression issues." }, { "left": "issue:45685", "right": "issue:45823", "accept": false, - "reason": "MPS histogram kernel limitation is unrelated to Gemma4 device-map mismatch." + "reason": "MPS histogram support in MoE and Gemma4 PLE device mismatch are different backend/device bugs." }, { "left": "issue:41553", "right": "issue:43883", "accept": false, - "reason": "Bad AutoTokenizer error messaging and missing all_tied_weights_keys are different bugs." + "reason": "Bad AutoTokenizer error message for Voxtral and missing all_tied_weights_keys on Molmo are unrelated model-specific issues." }, { "left": "issue:45020", "right": "issue:45823", "accept": false, - "reason": "Remote-code loading regressions are unrelated to Gemma4 PLE forward device mismatch." + "reason": "Remote-code loading regressions and Gemma4 device-map forward mismatch are different problem classes." }, { "left": "issue:43782", "right": "issue:44960", "accept": false, - "reason": "Qwen3VL weight-only loading failure and GLM5 issue are not the same underlying problem." + "reason": "Qwen3VL weight_only=True load error and the generic GLM5 issue do not describe the same failure." }, { "left": "issue:44661", "right": "issue:45020", "accept": false, - "reason": "Tokenizer mapping-registration failure is a different area from remote-code model loading regressions." + "reason": "Failure in add-new-model-like for TOKENIZER_MAPPING_NAMES is a registry-management bug, not the same as the broader remote-code loading regression." }, { "left": "issue:43828", "right": "issue:45823", "accept": false, - "reason": "Autocast dtype mismatch and device-map mismatch are different runtime errors." + "reason": "Autocast dtype mismatch in Phi-tiny-MoE-instruct and Gemma4 PLE device mismatch are separate runtime issues." }, { "left": "issue:45020", "right": "issue:45356", "accept": false, - "reason": "Remote-code loading regressions do not match the Kimi-K2.5 tokenizer codec/mistral-regex regression." + "reason": "General remote-code breakage and the Kimi-K2.5 tokenizer regression have different symptoms and likely different fixes." }, { "left": "issue:41553", "right": "issue:43881", "accept": false, - "reason": "Tokenizer import/error-message issue is unrelated to glm-4v-9b load failure." + "reason": "Bad AutoTokenizer error messaging for Voxtral and glm-4v-9b load failure are not the same bug." }, { "left": "issue:42491", "right": "issue:44368", "accept": false, - "reason": "LoRA compatibility across HF versions is unrelated to a warning about tie_word_embeddings." + "reason": "LoRA compatibility across versions and a tie_word_embeddings warning during Qwen3.5 fine-tuning are not duplicates." }, { "left": "issue:43872", "right": "issue:46869", "accept": false, - "reason": "bitsandbytes constructor incompatibility is unrelated to a YOLO multi-GPU device-map issue." + "reason": "bitsandbytes Int8Params incompatibility and YOLO multi-GPU device-map failure are unrelated." }, { "left": "issue:45759", "right": "issue:46661", "accept": false, - "reason": "Composite Qwen multimodal config handling is a different bug from remote-code model_class/config_class loading failures." + "reason": "Composite Qwen config unwrapping and remote-code model_class.config_class AttributeError are both AutoModel-related, but they are different code paths and not clearly the same bug." }, { "left": "issue:45715", "right": "issue:46315", "accept": false, - "reason": "Both are tokenizer performance regressions, but they affect opposite conversion directions and are not the same code path." + "reason": "Slow-tokenizer conversion inefficiency in two different methods is performance-related, but the titles point to separate hot paths and not one shared defect." }, { "left": "issue:45020", "right": "issue:46661", - "accept": true, - "reason": "Same remote-code loading regression: both describe failures when loading remote-code models, with missing config_class/fields as the concrete symptom." + "accept": false, + "reason": "Remote-code breakage is broad, but this is a specific AutoModel.from_pretrained AttributeError with remote-code models, not the same concrete issue as the umbrella report." }, { "left": "issue:45488", "right": "issue:46489", "accept": false, - "reason": "Both concern tokenizer correctness in DeepSeek-family models, but they describe different model families and different tokenizer bugs." + "reason": "Both are DeepSeek tokenizer regressions, but one is a ByteLevel-vs-Metaspace override issue and the other is an incorrect detokenization gap from a prior fix; too different to merge." }, { "left": "issue:43404", "right": "issue:43950", "accept": false, - "reason": "Un-tied lm_head weights in one model are unrelated to non-persistent buffer corruption in from_pretrained." + "reason": "lm_head tying bug in Mistral3 and non-persistent buffer corruption during from_pretrained are different state-loading problems." }, { "left": "issue:43827", "right": "issue:44509", - "accept": true, - "reason": "Same documentation regression: both note stale docs still referring to pipeline tasks removed in v5." + "accept": false, + "reason": "These are both docs cleanups after pipeline removals, but the pair is not explicitly the same artifact set and the titles suggest separate documentation mentions." }, { "left": "issue:43881", "right": "issue:43883", "accept": false, - "reason": "Different model-specific loading errors with no shared concrete failure mode." - }, - { - "left": "issue:46894", - "right": "issue:46897", - "accept": true, - "reason": "Same training-loss double-shift bug pattern; these look like model-specific reports of the same underlying loss-computation error." + "reason": "glm-4v-9b loading failure and Molmo missing all_tied_weights_keys are different model-specific loading errors." } ] }, - "evaluator_result": { - "accept": true, - "feedback": "The clustering is grounded in the packet and the accepted soft pairs are stated conservatively. The canonical issue choice (45020) is supported by its broad remote-code loading title, and the rejected pairs mostly avoid over-merging unrelated model- or subsystem-specific bugs. One minor imprecision: the summary says three duplicate pairs, but the accepted evidence effectively covers four pairwise links (two under the remote-code umbrella)." - }, - "evaluator_used": true, + "evaluator_result": null, + "evaluator_used": false, "retried": false, - "accepted_nontrivial_soft_edge": true, + "accepted_nontrivial_soft_edge": false, "error_kind": null, "error_message": null }, @@ -1841,7 +1785,7 @@ "source": "llm", "cache_hit": false, "model": "gpt-5.4-mini?service_tier=flex", - "cluster_id": "cluster-10000-529", + "cluster_id": "cluster-10000-530", "nodes": [ "issue:24643", "issue:30064", @@ -1856,8 +1800,6 @@ "issue:43531", "issue:43596", "issue:43638", - "issue:43645", - "issue:43646", "issue:43716", "issue:43792", "issue:43824", @@ -1890,12 +1832,15 @@ "issue:45923", "issue:46402", "issue:46468", - "issue:46515" + "issue:46515", + "issue:46894", + "issue:46897" ], "soft_pairs": [ "issue:43792|issue:44220", "issue:43257|issue:45305", "issue:36331|issue:41093", + "issue:46894|issue:46897", "issue:42947|issue:45923", "issue:43940|issue:45313", "issue:43854|issue:44863", @@ -1921,227 +1866,226 @@ "issue:42617|issue:43366", "issue:43950|issue:45834", "issue:44560|issue:45685", - "issue:43645|issue:43646", "issue:41553|issue:44488" ], - "prepared_review_unit_hash": "9e32dfb8e6366e83d5b4351e4b886c384f3b0f5a22fea5f666e744681e5696b5", + "prepared_review_unit_hash": "324d6ffc40127287feb70477bf0e00c696fcaa086f775b6ef038349d0fad6a57", "component_budget": { - "node_count": 529, - "item_count": 529, - "soft_pair_count": 1953, - "serialized_chars": 1012074, - "estimated_input_tokens": 253019, - "estimated_eval_tokens": 506294 + "node_count": 530, + "item_count": 530, + "soft_pair_count": 1954, + "serialized_chars": 1013004, + "estimated_input_tokens": 253251, + "estimated_eval_tokens": 506758 }, "budget": { "node_count": 48, "item_count": 48, "soft_pair_count": 30, - "serialized_chars": 36765, - "estimated_input_tokens": 9192, - "estimated_eval_tokens": 18640 + "serialized_chars": 36758, + "estimated_input_tokens": 9190, + "estimated_eval_tokens": 18636 }, "overflow_policy": "truncate_then_skip", "trimmed": false, "aggressively_trimmed": false, "split": true, "analyst_result": { - "summary": "The cluster is mostly heterogeneous: many pairs share a subsystem or model family, but only a few appear to be the same underlying bug. I accept the near-exact duplicates for the mask-shape IndexError, the Qwen3-VL video regression, and the Transformers 5.0 custom-model initialization breakage.", - "confidence": 0.78, - "canonical_issue_reason": "Issue 44479 is the best overall issue representative among the true duplicate-like reports because it captures a broad, concrete regression across multiple Qwen3/VL variants; the other accepted duplicate pairs are either narrower or more exact-error follow-ups.", + "summary": "The cluster is mostly title-level similarity rather than confirmed duplicates. I would not accept any soft-edge merge on the packet alone; issue:44805 is still the best anchor for the mask/tensor shape-mismatch family, but its relationship to issue:41093 remains tentative.", + "confidence": 0.64, + "canonical_issue_reason": "issue:44805 is the best representative for the mask/tensor shape-mismatch family because it has the clearest error signature and reproduction context; issue:41093 is only loosely similar by title.", "canonical_pr_reason": null, - "best_issue_reason": "44479 is the strongest canonical issue candidate: it describes a concrete, reproducible regression with a broader blast radius than the other accepted pairs, so it is the best anchor for deduping this set.", + "best_issue_reason": "issue:44805 is the strongest triage anchor in this set: the report is detailed, the failure signature is explicit, and it is the easiest case to compare against later shape-mismatch reports.", "best_pr_reason": null, "soft_edge_verdicts": [ { "left": "issue:43792", "right": "issue:44220", "accept": false, - "reason": "Both are audio/model runtime problems, but they describe different failures and no shared concrete code-path is evident." + "reason": "Both are audio/Whisper-related, but one is a broad run/load failure and the other is a specific fbank helper issue; different code paths." }, { "left": "issue:43257", "right": "issue:45305", "accept": false, - "reason": "Both mention DeepSpeed, but one is about Qwen3 MoE weight conversion and the other about gradient averaging with loss kwargs; different bugs." + "reason": "Both mention DeepSpeed, but the failures differ: weight conversion vs. gradient averaging with GAS." }, { "left": "issue:36331", "right": "issue:41093", "accept": false, - "reason": "Unexpected kwarg in CustomTrainer is unrelated to the mask/tensor shape IndexError." + "reason": "Unexpected compute_loss kwargs and a mask/tensor shape mismatch are different error classes and likely unrelated bugs." + }, + { + "left": "issue:46894", + "right": "issue:46897", + "accept": false, + "reason": "They share a bug pattern, but the packet does not show the same concrete model/path or a shared fix." }, { "left": "issue:42947", "right": "issue:45923", "accept": false, - "reason": "Both touch gradient checkpointing, but one is LoRA checkpointing ineffectiveness and the other is a missing capability flag on a trust-remote-code model." + "reason": "Both touch gradient checkpointing, but one is PEFT LoRA behavior and the other is a missing model flag; different issues." }, { "left": "issue:43940", "right": "issue:45313", "accept": false, - "reason": "Both are ZeRO-3 weight-loading issues in Qwen models, but they affect different model families and the failure descriptions are not the same bug." + "reason": "The titles are similar in the Qwen3/ZeRO-3 area, but there is no explicit linkage or remediation evidence to confirm the same bug." }, { "left": "issue:43854", "right": "issue:44863", "accept": false, - "reason": "Different models and different loading failures; only the broad theme of checkpoint loading overlaps." + "reason": "These are model-loading failures for different models with different symptoms; not the same concrete bug." }, { "left": "issue:24643", "right": "issue:30064", "accept": false, - "reason": "Training with DeepSpeed and void segmentation map processing are unrelated problems." + "reason": "DeepSpeed training and void segmentation-map preprocessing are unrelated failure modes." }, { "left": "issue:44479", "right": "issue:44560", - "accept": true, - "reason": "Both report the same Qwen3/VL video-input regression in transformers 5.3.0, with matching StopIteration-style symptoms." + "accept": false, + "reason": "Both are video-related, but one is a broad regression across several models and the other is a specific StopIteration report; not enough to merge." }, { "left": "issue:41093", "right": "issue:44805", - "accept": true, - "reason": "These describe the same mask-vs-indexed-tensor shape mismatch IndexError; 44805 looks like a later, more detailed report of the same bug." + "accept": false, + "reason": "Same general shape-mismatch theme, but the packet only gives title-level similarity and no shared repro or fix evidence." }, { "left": "issue:44928", "right": "issue:45588", "accept": false, - "reason": "A NaN gradient explosion in RLHF is not the same bug as the flash-attention crash on s_aux=None." + "reason": "A gradient-explosion report from 3D position_ids and a flash-attention AttributeError are different bugs." }, { "left": "issue:43366", "right": "issue:43828", "accept": false, - "reason": "GGUF/gpt-oss support is unrelated to the torch.autocast dtype mismatch in Phi-tiny-MoE." + "reason": "GGUF gpt-oss support and a Phi-tiny-MoE autocast dtype mismatch are unrelated." }, { "left": "issue:43716", "right": "issue:46468", "accept": false, - "reason": "An image preprocessor dtype mismatch is not the same as the broken Nemotron example." + "reason": "Different model families and different failure modes; no shared code path is evident." }, { "left": "issue:44530", "right": "issue:44863", "accept": false, - "reason": "PagedAttentionCache linear_attention failure on Qwen3.5 is unrelated to NemotronH checkpoint loading." + "reason": "PagedAttentionCache linear_attention crash and NemotronH checkpoint loading are unrelated issues." }, { "left": "issue:44863", "right": "issue:46402", "accept": false, - "reason": "Different models and different symptoms: NemotronH checkpoint loading versus Sam3 text encoder weights not loading." + "reason": "Both are model-loading problems, but they affect different model families and different loading paths." }, { "left": "issue:43366", "right": "issue:43531", "accept": false, - "reason": "GGUF gpt-oss support and Qwen3-MoE sliding_window behavior are unrelated." + "reason": "GGUF gpt-oss support and a Qwen3-MoE sliding_window issue do not describe the same bug." }, { "left": "issue:43596", "right": "issue:46515", "accept": false, - "reason": "DeepSpeed ZeRO-3 zero-sized tensor indexing and a torch 2.4.1 incompatibility are not the same issue." + "reason": "A ZeRO-3 Bert index error and a generic torch incompatibility complaint are not the same concrete failure." }, { "left": "issue:41553", "right": "issue:44991", "accept": false, - "reason": "Both are tokenizer-loading complaints, but they target different models and different root causes." + "reason": "Both are tokenizer-loading reports, but they target different models and error conditions." }, { "left": "issue:45325", "right": "issue:45902", "accept": false, - "reason": "One is a Qwen2.5-VL rope-index scaling bug, the other is uninitialized memory in a Qwen3.5-MoE rotary embedding path." + "reason": "Qwen2.5-VL rope-index scaling and a Qwen3_5Moe rotary-buffer materialization bug are different code paths." }, { "left": "issue:43366", "right": "issue:44589", "accept": false, - "reason": "GGUF support for gpt-oss and missing Float8 storage support are unrelated." + "reason": "GGUF support and a Float8 storage lookup failure are unrelated." }, { "left": "issue:43638", "right": "issue:46515", "accept": false, - "reason": "The BertModel ZeRO-3 index error is unrelated to the torch version incompatibility report." + "reason": "The Bert ZeRO-3 index error and a torch version complaint are not the same bug." }, { "left": "issue:45663", "right": "issue:45823", "accept": false, - "reason": "Both are Gemma4 issues, but one is an FSDP2 attention-state bug and the other is a device-map forward mismatch." + "reason": "Same model family, but FSDP2 KeyError and device-map forward mismatch are different problems." }, { "left": "issue:43716", "right": "issue:44560", "accept": false, - "reason": "Different model families and different error paths: image preprocessor dtype mismatch versus video input StopIteration." + "reason": "Mistral-3 dtype mismatch and Qwen3-VL StopIteration are unrelated failures." }, { "left": "issue:42915", "right": "issue:43957", "accept": false, - "reason": "FineGrainedFP8Config failure in Qwen3MoE is not the same bug as meta-device model loading breakage." + "reason": "Qwen3-MoE FP8 failure and meta-device loading regression are different issues." }, { "left": "issue:44155", "right": "issue:45584", "accept": false, - "reason": "Both are audio-related, but one is a batched inference token leak and the other is empty-transcription generation failure." + "reason": "AudioFlamingo3 batched inference token leakage and Whisper empty-transcription generation failure are unrelated." }, { "left": "issue:43824", "right": "issue:43957", "accept": false, - "reason": "Missing Qwen2.5-VL import and meta-device loading regressions are different problems." + "reason": "A missing import and a meta-device loading problem are different failure modes." }, { "left": "issue:42617", "right": "issue:43366", "accept": false, - "reason": "Running 3d_parallel.py and GGUF model architecture support are unrelated." + "reason": "3d_parallel.py execution failure and GGUF support are unrelated." }, { "left": "issue:43950", "right": "issue:45834", "accept": false, - "reason": "Non-persistent buffer corruption and long OCR index errors are different bugs." + "reason": "Non-persistent buffer corruption and a Kosmos2.5 OCR index error do not share the same code path." }, { "left": "issue:44560", "right": "issue:45685", "accept": false, - "reason": "Video-input StopIteration on Qwen3-VL is unrelated to the MPS histogram implementation error in MoE." - }, - { - "left": "issue:43645", - "right": "issue:43646", - "accept": true, - "reason": "These are the same Transformers 5.0 custom-model initialization regression, with 43645 being the notebook-specific report and 43646 the broader one." + "reason": "Qwen3-VL video StopIteration and an MPS histogram limitation in MoE are unrelated." }, { "left": "issue:41553", "right": "issue:44488", "accept": false, - "reason": "Different tokenizer-loading failures on different models; no evidence they share the same underlying bug." + "reason": "Both mention tokenizer/model loading, but the models and symptoms differ and there is no shared fix evidence." } ] }, "evaluator_result": { - "accept": true, - "feedback": "The summary is grounded in the packet and stays conservative overall. The accepted pairs are well-supported by near-identical titles/symptoms, and the rejected pairs generally avoid overclaiming beyond the available evidence." + "accept": false, + "feedback": "The writeup is mostly careful, but it still overstates evidence on a few soft-edge pairs. In particular, the three accepted pairs (43940\u219445313, 44479\u219444560, 41093\u219444805) are plausible matches by title, but the packet provides no explicit linkage or shared remediation evidence, so calling them the same underlying bug / true duplicates is stronger than warranted. The summary also states 'true duplicates' more broadly than the packet supports. Conservative rephrasing or keeping those as tentative would be safer." }, "evaluator_used": true, - "retried": false, - "accepted_nontrivial_soft_edge": true, + "retried": true, + "accepted_nontrivial_soft_edge": false, "error_kind": null, "error_message": null }, @@ -2156,7 +2100,7 @@ "source": "llm", "cache_hit": false, "model": "gpt-5.4-mini?service_tier=flex", - "cluster_id": "cluster-10000-529", + "cluster_id": "cluster-10000-530", "nodes": [ "issue:36010", "issue:38175", @@ -2172,6 +2116,7 @@ "issue:43638", "issue:43644", "issue:43645", + "issue:43646", "issue:43720", "issue:43994", "issue:44451", @@ -2197,10 +2142,8 @@ "issue:45812", "issue:45823", "issue:45834", - "issue:46459", "issue:46468", "issue:46489", - "issue:46512", "issue:46515", "issue:46846", "issue:46858", @@ -2208,6 +2151,7 @@ "issue:47254" ], "soft_pairs": [ + "issue:43645|issue:43646", "issue:44560|issue:44918", "issue:44683|issue:46515", "issue:36010|issue:42915", @@ -2236,200 +2180,204 @@ "issue:44451|issue:44488", "issue:45313|issue:45357", "issue:44863|issue:46846", - "issue:45507|issue:46468", - "issue:46459|issue:46512", - "issue:43644|issue:43645" + "issue:45507|issue:46468" ], - "prepared_review_unit_hash": "0a02fcb93af3415f5989c8651709c62af17fcfed639351ef9604a62ad79445ae", + "prepared_review_unit_hash": "3bd944c90e1a26f579667044dc859dee165294f9c310c7f8c7efd3dba1ddce3a", "component_budget": { - "node_count": 529, - "item_count": 529, - "soft_pair_count": 1953, - "serialized_chars": 1012074, - "estimated_input_tokens": 253019, - "estimated_eval_tokens": 506294 + "node_count": 530, + "item_count": 530, + "soft_pair_count": 1954, + "serialized_chars": 1013004, + "estimated_input_tokens": 253251, + "estimated_eval_tokens": 506758 }, "budget": { - "node_count": 48, - "item_count": 48, - "soft_pair_count": 31, - "serialized_chars": 36943, - "estimated_input_tokens": 9236, - "estimated_eval_tokens": 18728 - }, + "node_count": 47, + "item_count": 47, + "soft_pair_count": 30, + "serialized_chars": 35975, + "estimated_input_tokens": 8994, + "estimated_eval_tokens": 18244 + }, "overflow_policy": "truncate_then_skip", "trimmed": false, "aggressively_trimmed": false, "split": true, "analyst_result": { - "summary": "This cluster is mostly a mix of unrelated issues, with only two strong duplicate pairs: the Transformers v5 non-persistent-buffer regression and the Gemma4 4-bit multimodal dtype-cast bug. The SigLIP2 pair is too weakly supported to treat as a duplicate, so I rejected it.", - "confidence": 0.88, - "canonical_issue_reason": "issue:44534 is the clearest exact-duplicate anchor: it matches issue:43644 on the same Transformers v5 non-persistent-buffer junk regression.", + "summary": "This packet contains one clearly exact duplicate pair, a couple of plausible same-bug pairs, and several lookalikes that are too weak to merge. The strongest anchor is the non-persistent-buffer regression (43644/44534). SigLIP2 output regressions (38175/43994) and Gemma4 4-bit multimodal dtype issues (45672/46899) look like real duplicates, while the custom-model-init reports (43645/43646) are related but not evidenced well enough to treat as the same root cause.", + "confidence": 0.84, + "canonical_issue_reason": "issue:43644 is the best canonical issue because it is the earlier, broader report of the exact same 'Transformers v5 fills non-persistent buffers with junk' regression later duplicated by 44534, with broader discussion and an inbound reference.", "canonical_pr_reason": null, - "best_issue_reason": "issue:44534 is the strongest standalone issue because it is a precise, reproducible regression and has an exact duplicate counterpart.", + "best_issue_reason": "issue:43644 is the best global representative: it is an exact duplicate target, predates the follow-up report, and its title captures the underlying regression without narrowing it to one reproduction path.", "best_pr_reason": null, "soft_edge_verdicts": [ + { + "left": "issue:43645", + "right": "issue:43646", + "accept": false, + "reason": "Closely related custom-model-init failures, but the packet does not show a shared root cause or code path; treat as related reports, not duplicates." + }, { "left": "issue:44560", "right": "issue:44918", "accept": false, - "reason": "Same broad Qwen3 area, but the failures are different code paths: video StopIteration versus SFT embedding unpacking." + "reason": "Different model families and different failure modes: video StopIteration versus TRL SFT unpacking." }, { "left": "issue:44683", "right": "issue:46515", "accept": false, - "reason": "Both mention torch, but one is flex_attention on torch>=2.9 and the other is a generic torch 2.4.1 incompatibility; too broad." + "reason": "Both mention torch compatibility, but they describe different versions and different concrete failures." }, { "left": "issue:36010", "right": "issue:42915", "accept": false, - "reason": "ImportError for GenerationMixin is unrelated to Qwen3Moe fine-grained FP8 loading/config behavior." + "reason": "Unrelated issues: GenerationMixin import error versus Qwen3 MoE LoRA hf4.x/hf5.x incompatibility." }, { "left": "issue:44910", "right": "issue:46858", "accept": false, - "reason": "Both are cache/attention-adjacent, but the root causes differ: 3D position_ids misinterpretation versus static-cache slice arithmetic." + "reason": "Both are cache/attention-related, but they hit different mechanisms and backends; not enough evidence of one bug." }, { "left": "issue:43550", "right": "issue:44910", "accept": false, - "reason": "Different models and mechanisms: Bamba torch.compile+SDPA versus Qwen3.5 flash-attention position_ids handling." + "reason": "Different model and problem class: Bamba torch.compile+SDPA versus Qwen3.5 flash-attention position_ids handling." }, { "left": "issue:42947", "right": "issue:47254", "accept": false, - "reason": "Both involve PEFT and gradient checkpointing, but one is ineffective checkpointing and the other is a ZeRO-3 CheckpointError." + "reason": "Related training stack, but one is ineffective gradient checkpointing and the other is a ZeRO-3 checkpoint error." }, { "left": "issue:45823", "right": "issue:46515", "accept": false, - "reason": "A Gemma4 device-mapping forward mismatch is not the same as a generic torch version incompatibility report." + "reason": "Gemma4 device-mismatch during forward is not the same as a generic torch version incompatibility report." }, { "left": "issue:45812", "right": "issue:46489", "accept": false, - "reason": "Both are tokenizer-related regressions, but they affect different model families and likely different code paths." + "reason": "Both are tokenizer regressions, but for different model families and different concrete defects." }, { "left": "issue:44661", "right": "issue:45823", "accept": false, - "reason": "Tokenizer-mapping edge case versus Gemma4 device-mapping forward bug; no concrete duplicate-level overlap." + "reason": "Tokenizer-mapping failure and Gemma4 device mismatch are unrelated code paths." }, { "left": "issue:44560", "right": "issue:44805", "accept": false, - "reason": "One is Qwen3-vl video StopIteration; the other is a generic mask-shape IndexError." + "reason": "Different symptoms and different model paths; no shared underlying defect is shown." }, { "left": "issue:44877", "right": "issue:45030", "accept": false, - "reason": "Both mention config validation, but the affected models and failure modes differ too much to be the same bug." + "reason": "Both involve stricter config validation, but the affected configs and fixes appear different." }, { "left": "issue:38175", "right": "issue:43994", - "accept": false, - "reason": "The shared SigLIP2 model name and vague bad-output symptom are not enough to establish the same root cause or a duplicate-level match." + "accept": true, + "reason": "Both report the same SigLIP2 model regression producing wrong/zero outputs; the later issue looks like the same underlying bug re-reported." }, { "left": "issue:43638", "right": "issue:44661", "accept": false, - "reason": "A DeepSpeed ZeRO-3 Bert index error is unrelated to tokenizer mapping-name handling." + "reason": "DeepSpeed ZeRO-3 Bert indexing error is unrelated to add-new-model-like/tokenizer mapping logic." }, { "left": "issue:43596", "right": "issue:44661", "accept": false, - "reason": "Different areas entirely: ZeRO-3/Bert initialization versus tokenizer mapping logic." + "reason": "Different bug classes: Bert zero3 init IndexError versus tokenizer-mapping failure." }, { "left": "issue:44661", "right": "issue:44805", "accept": false, - "reason": "Tokenizer helper failure versus a tensor mask-shape IndexError; not the same bug." + "reason": "Only a broad runtime/load-failure resemblance; no evidence of the same code-path problem." }, { "left": "issue:42491", "right": "issue:43720", "accept": false, - "reason": "Both touch loading/weights, but one is a Qwen3 MoE LoRA compatibility break and the other is BitNet packed-weight unpacking." + "reason": "Different models and failure modes; not the same underlying bug." }, { "left": "issue:45672", "right": "issue:46899", "accept": true, - "reason": "Both describe the same Gemma4 multimodal 4-bit quantization problem, centered on unsafe dtype handling in audio/vision embedders." + "reason": "Both are Gemma4 multimodal 4-bit quantization failures involving unsafe dtype handling in audio/vision embedders." }, { "left": "issue:43632", "right": "issue:45834", "accept": false, - "reason": "Transformers v5 initialization flag breakage is unrelated to a Kosmos2.5 OCR index error." + "reason": "_is_hf_initialized flag regression is unrelated to the Kosmos2.5 OCR indexing issue." }, { "left": "issue:43317", "right": "issue:44756", "accept": false, - "reason": "GPU/CPU offload loading failure and Strix Halo mmap OOM are different load-time problems." + "reason": "Different resource/loading problems: dequantized offload failure versus mmap OOM workaround." }, { "left": "issue:41720", "right": "issue:42915", "accept": false, - "reason": "Both are Qwen3-related, but one is auto device mapping cudaErrorAssert and the other is FineGrainedFP8Config loading." + "reason": "Qwen3 auto device-mapping cuda assert and Qwen3 MoE LoRA hf4.x/hf5.x incompatibility are different bugs." }, { "left": "issue:38175", "right": "issue:39692", "accept": false, - "reason": "Same model family, but one is a runtime output anomaly and the other is a docs-example model/processor mismatch plus quantization issue." + "reason": "Same model family, but one is incorrect outputs and the other is a docs-example mismatch plus quantization failure." }, { "left": "issue:42915", "right": "issue:44756", "accept": false, - "reason": "No real overlap beyond being issues in the same repository." + "reason": "No shared underlying bug; one is a model compatibility issue, the other a platform-specific mmap/OOM workaround." }, { "left": "issue:43644", "right": "issue:44534", "accept": true, - "reason": "Exact duplicate regression: both report non-persistent buffers being filled with junk in Transformers v5." + "reason": "Exact duplicate regression: both report Transformers v5 filling non-persistent buffers with junk." }, { "left": "issue:42491", "right": "issue:45276", "accept": false, - "reason": "Qwen3 MoE LoRA compatibility and Gemma4 resize_token_embeddings behavior are unrelated." + "reason": "LoRA hf4.x/hf5.x incompatibility is unrelated to Gemma4 token-embedding resize behavior." }, { "left": "issue:44960", "right": "issue:44977", "accept": false, - "reason": "Different models and failures: GLM5 versus Qwen3.5 flash-attention generation." + "reason": "Unrelated GLM5 and Qwen3.5 flash-attention reports." }, { "left": "issue:44451", "right": "issue:44488", "accept": false, - "reason": "Both are model-loading complaints, but the titles do not show the same underlying code-path or fix." + "reason": "Both are BERT load failures, but the packet does not show enough evidence that they share the same root cause." }, { "left": "issue:45313", "right": "issue:45357", "accept": false, - "reason": "One is ZeRO-3 loading weights for language_model, the other is save_pretrained writing incorrect visual encoder keys; separate bugs." + "reason": "Distinct code paths: ZeRO-3 weight loading versus incorrect save_pretrained visual-encoder keys." }, { "left": "issue:44863", @@ -2441,25 +2389,13 @@ "left": "issue:45507", "right": "issue:46468", "accept": false, - "reason": "GraniteMoEHybrid invalid method calls and a broken Nemotron example do not share the same bug." - }, - { - "left": "issue:46459", - "right": "issue:46512", - "accept": false, - "reason": "Both involve dtype, but the load-time silent ignore and the stale config.dtype issue affect different code paths." - }, - { - "left": "issue:43644", - "right": "issue:43645", - "accept": false, - "reason": "The buffer-junk regression and Jupyter custom-model initialization breakage are unrelated." + "reason": "Different models and failures; no duplicate evidence." } ] }, "evaluator_result": { "accept": false, - "feedback": "The summary is mostly grounded, but one soft-edge acceptance is too strong for the evidence provided. In particular, issue:38175 vs issue:43994 shares the same SigLIP2 model name and a vague bad-output symptom, but the packet does not establish the same root cause or a clear duplicate-level match. The other accepted pairs (44534/43644 and 45672/46899) look solid, but this borderline acceptance makes the overall verdict insufficiently conservative." + "feedback": "The packet supports several plausible duplicate clusters, but the response is not conservative enough. In particular, issue:43645 vs issue:43646 is overstated as the same underlying regression; the packet only shows closely related custom-model-init failures, not proof of a shared root cause. The summary also overstates that custom model initialization is a clear duplicate subcluster. The 43644 vs 44534 pair looks well grounded, and 38175/43994 and 45672/46899 are plausible, but the overall precision standard is not met." }, "evaluator_used": true, "retried": true, @@ -2478,17 +2414,18 @@ "source": "llm", "cache_hit": false, "model": "gpt-5.4-mini?service_tier=flex", - "cluster_id": "cluster-10000-529", + "cluster_id": "cluster-10000-530", "nodes": [ "issue:39692", "issue:41720", "issue:42491", "issue:42947", "issue:43257", - "issue:43335", "issue:43404", "issue:43425", "issue:43454", + "issue:43644", + "issue:43645", "issue:43708", "issue:43854", "issue:43856", @@ -2497,7 +2434,6 @@ "issue:43957", "issue:43994", "issue:44075", - "issue:44079", "issue:44155", "issue:44164", "issue:44291", @@ -2524,17 +2460,20 @@ "issue:45823", "issue:45859", "issue:46050", - "issue:46829", + "issue:46459", + "issue:46512", "issue:46897", "issue:46901" ], "soft_pairs": [ + "issue:46459|issue:46512", "issue:44488|issue:44991", - "issue:39692|issue:43994", + "issue:43644|issue:43645", "issue:43257|issue:43854", + "issue:39692|issue:43994", "issue:43856|issue:43873", - "issue:43854|issue:43931", "issue:42491|issue:45335", + "issue:43854|issue:43931", "issue:42491|issue:45292", "issue:43856|issue:44387", "issue:43708|issue:44075", @@ -2545,236 +2484,234 @@ "issue:45362|issue:45406", "issue:43425|issue:43957", "issue:44291|issue:44756", - "issue:46897|issue:46901", "issue:41720|issue:44155", "issue:44402|issue:44488", "issue:45663|issue:45759", "issue:39692|issue:44960", "issue:44910|issue:45588", "issue:43404|issue:43454", + "issue:46897|issue:46901", "issue:43873|issue:44387", "issue:42947|issue:45704", "issue:45663|issue:45859", "issue:44292|issue:44912", "issue:44155|issue:45823", - "issue:44079|issue:46829", - "issue:43957|issue:44756", - "issue:43335|issue:45663" + "issue:43957|issue:44756" ], - "prepared_review_unit_hash": "da6e9ece9cac4796ce9fefc589f373f22300b661810f442afa787195894d59fb", + "prepared_review_unit_hash": "fa46344100f9b5bf8c1da13868d5a1537ab28ad95fc5ea6baae2604c0e637264", "component_budget": { - "node_count": 529, - "item_count": 529, - "soft_pair_count": 1953, - "serialized_chars": 1012074, - "estimated_input_tokens": 253019, - "estimated_eval_tokens": 506294 + "node_count": 530, + "item_count": 530, + "soft_pair_count": 1954, + "serialized_chars": 1013004, + "estimated_input_tokens": 253251, + "estimated_eval_tokens": 506758 }, "budget": { - "node_count": 47, - "item_count": 47, + "node_count": 48, + "item_count": 48, "soft_pair_count": 31, - "serialized_chars": 36625, - "estimated_input_tokens": 9157, - "estimated_eval_tokens": 18570 + "serialized_chars": 37206, + "estimated_input_tokens": 9302, + "estimated_eval_tokens": 18860 }, "overflow_policy": "truncate_then_skip", "trimmed": false, "aggressively_trimmed": false, "split": true, "analyst_result": { - "summary": "These items are mostly unrelated model-loading, tokenizer, quantization, and training bugs. I found no soft pair that clearly represents the same underlying issue, so all soft edges are rejected.", + "summary": "This cluster is highly heterogeneous; most pairs share only a broad theme like loading, tokenization, quantization, or model-init regressions, but not the same concrete bug. No soft pair looks safe to merge as duplicates.", "confidence": 0.82, - "canonical_issue_reason": "issue:46901 is the clearest standalone report in this set: narrow scope, specific failure mode, and an open issue that could anchor future duplicates. The soft pairs do not justify forming a duplicate cluster around any other item.", + "canonical_issue_reason": "issue:45812 is the strongest representative because it describes a broad, concrete, user-visible regression (all Granite token IDs are wrong) with clear impact and a well-scoped fix target.", "canonical_pr_reason": null, - "best_issue_reason": "issue:46901 is the best representative issue here because it is precise, self-contained, and actionable; it is more suitable as a canonical reference than the broader or more ambiguous reports.", + "best_issue_reason": "issue:45812 is the best anchor for the cluster: it is specific, high-impact, and reads like a clear regression with a single underlying code path.", "best_pr_reason": null, "soft_edge_verdicts": [ + { + "left": "issue:46459", + "right": "issue:46512", + "accept": false, + "reason": "Both mention dtype handling, but one is about loading composite checkpoints and the other about from_config initialization/config staleness; different bug surfaces." + }, { "left": "issue:44488", "right": "issue:44991", "accept": false, - "reason": "Both are tokenizer/model loading complaints, but they target different models and failure modes; not the same bug." + "reason": "Both are tokenizer-loading failures, but they concern different models and likely different tokenizer files/code paths." }, { - "left": "issue:39692", - "right": "issue:43994", + "left": "issue:43644", + "right": "issue:43645", "accept": false, - "reason": "Both involve SigLIP2, but one is a documentation/example mismatch plus quantization failure and the other is nonsensical inference output; not the same concrete issue." + "reason": "One is about junk-filled non-persistent buffers; the other is a notebook/custom-model initialization break. Similar release theme, not the same bug." }, { "left": "issue:43257", "right": "issue:43854", "accept": false, - "reason": "Different models and different loading problems; no evidence of a shared underlying bug." + "reason": "Different models and failure modes: Qwen3 MoE weight conversion under accelerate+deepspeed vs GLM-4.7-Flash loading in tests." }, { - "left": "issue:43856", - "right": "issue:43873", + "left": "issue:39692", + "right": "issue:43994", "accept": false, - "reason": "Both mention memory, but one is Qwen3 MoE training inefficiency and the other is quantization/offloading behavior; distinct code paths." + "reason": "Both involve SigLIP2, but one is a docs example with model/processor mismatch and quantization trouble; the other is nonsensical outputs with AutoModel/pipeline." }, { - "left": "issue:43854", - "right": "issue:43931", + "left": "issue:43856", + "right": "issue:43873", "accept": false, - "reason": "Separate model families with different load-time errors; not duplicates." + "reason": "Both are memory-related, but training-memory inefficiency and quantization offload behavior are distinct issues." }, { "left": "issue:42491", "right": "issue:45335", "accept": false, - "reason": "LoRA hf4.x/hf5.x incompatibility is unrelated to the T5Gemma resize_token_embeddings propagation bug." + "reason": "LoRA incompatibility for Qwen3 MoE is unrelated to resize_token_embeddings not updating decoder.embed_tokens." + }, + { + "left": "issue:43854", + "right": "issue:43931", + "accept": false, + "reason": "Different models and symptoms: GLM-4.7-Flash test loading vs Qwen3-VL weight-shape mismatch." }, { "left": "issue:42491", "right": "issue:45292", "accept": false, - "reason": "Different model-specific and generic resize-token issues; no shared concrete failure." + "reason": "The LoRA/qwen3_moe compatibility regression is not the same as a generic output_embeddings resize propagation bug." }, { "left": "issue:43856", "right": "issue:44387", "accept": false, - "reason": "Both are memory-related, but the root causes differ: training memory usage vs int4 quantization reserved-memory regression." + "reason": "Both mention memory, but one is Qwen3 MoE training efficiency and the other is int4 quantization causing CUDA reserved-memory OOM." }, { "left": "issue:43708", "right": "issue:44075", "accept": false, - "reason": "Trainer checkpoint step \u0917\u0923ulation and unused SGD args are unrelated bugs." + "reason": "Trainer resume/max_steps calculation is unrelated to SGD argument handling." }, { "left": "issue:44164", "right": "issue:44291", "accept": false, - "reason": "One is save/from_pretrained extra_state handling; the other is init_empty_weights / _is_hf_initialized argument mismatch." + "reason": "Both involve model save/load flows, but extra_state handling and init_empty_weights/_is_hf_initialized are different failure points." }, { "left": "issue:45488", "right": "issue:45812", "accept": false, - "reason": "Both are tokenizer regressions, but they affect different tokenizer implementations and models." + "reason": "Both are tokenizer regressions, but one changes LlamaTokenizer pre-tokenization for DeepSeek-style models; the other produces wrong token IDs for Granite." }, { "left": "issue:44960", "right": "issue:46050", "accept": false, - "reason": "Generic GLM5 mention vs a sinq save/load quantization break; not the same issue." + "reason": "GLM5 loading and sinq quantization save/load breakage are unrelated bugs." }, { "left": "issue:45276", "right": "issue:45335", "accept": false, - "reason": "Both are resize_token_embeddings propagation bugs, but they are model-specific and not clearly the same code-path problem." + "reason": "Both are resize_token_embeddings propagation bugs, but they hit different architectures and output paths; not one concrete bug." }, { "left": "issue:45362", "right": "issue:45406", "accept": false, - "reason": "A Qwen3.5 chat crash and a Gemma4Processor attribute error are unrelated." + "reason": "Qwen3.5 chat crashes and Gemma4Processor missing _tokenizer are different code paths." }, { "left": "issue:43425", "right": "issue:43957", "accept": false, - "reason": "Torch version incompatibility and meta-device loading failures are different compatibility issues." + "reason": "Torch 2.10 compatibility is a dependency/version issue; meta-device loading breakage is a separate model-loading bug." }, { "left": "issue:44291", "right": "issue:44756", "accept": false, - "reason": "init_empty_weights argument handling and mmap OOM mitigation are unrelated." - }, - { - "left": "issue:46897", - "right": "issue:46901", - "accept": false, - "reason": "Both describe training-loss double-shift symptoms, but they are different model families with no evidence of the same underlying implementation bug." + "reason": "Unexpected init arg in init_empty_weights is unrelated to Strix Halo mmap OOM behavior." }, { "left": "issue:41720", "right": "issue:44155", "accept": false, - "reason": "CUDA assert on device mapping and token/embedding leak between tracks are unrelated failures." + "reason": "CUDA assert on Qwen3 auto device mapping is not the same as AudioFlamingo3 batched inference token/embedding leakage." }, { "left": "issue:44402", "right": "issue:44488", "accept": false, - "reason": "One is tokenizer vocab-size mismatch, the other is a model loading failure for a different model." + "reason": "Both are tokenizer/model-loading problems, but they concern different models and different underlying defects." }, { "left": "issue:45663", "right": "issue:45759", "accept": false, - "reason": "Gemma4 FSDP2 attention-state bug and Qwen multimodal config unwrapping are different code paths." + "reason": "Gemma4 FSDP2 shared_kv_states bug and Qwen multimodal text_config unwrapping are separate composite-model issues." }, { "left": "issue:39692", "right": "issue:44960", "accept": false, - "reason": "SigLIP2 documentation/quantization issue and GLM5 loading issue are unrelated." + "reason": "SigLIP2 docs/quantization errors and GLM5 loading are unrelated." }, { "left": "issue:44910", "right": "issue:45588", "accept": false, - "reason": "Both are flash-attention crashes, but one is about 3D position_ids being misread and the other is an s_aux=None attribute error." + "reason": "Both are flash-attention crashes, but one is a packed-sequence/3D position_ids bug and the other is an s_aux=None attribute error." }, { "left": "issue:43404", "right": "issue:43454", "accept": false, - "reason": "Both concern lm_head tying, but they are separate model implementations and not clearly the same defect." + "reason": "Both are lm_head tying problems, but they affect different model families and likely different implementations." + }, + { + "left": "issue:46897", + "right": "issue:46901", + "accept": false, + "reason": "Both describe a double-shift training-loss symptom, but they are different models without evidence of one shared code path." }, { "left": "issue:43873", "right": "issue:44387", "accept": false, - "reason": "Both involve memory/OOM, but the mechanisms and affected paths differ." + "reason": "Quantization/offload behavior and int4 reserved-memory OOM are related broadly, but not the same bug." }, { "left": "issue:42947", "right": "issue:45704", "accept": false, - "reason": "Gradient checkpointing with PEFT LoRA is unrelated to the apex.FusedRMSNorm memory leak." + "reason": "Gradient checkpointing with PEFT LoRA is unrelated to T5 using apex.FusedRMSNorm memory leak." }, { "left": "issue:45663", "right": "issue:45859", "accept": false, - "reason": "Gemma4 FSDP2 KeyError and Qwen3_5Moe CPU-offload incompatibility are different issues." + "reason": "Gemma4 FSDP2 attention-state rebuild and Qwen3_5MoeTextRotaryEmbedding CPU-offload incompatibility are different bugs." }, { "left": "issue:44292", "right": "issue:44912", "accept": false, - "reason": "Different models, different quantization formats, and different outcomes; not duplicates." + "reason": "Qwen-3-8B-NVFP4 runtime error and git-oss-20b MXFP4 fallback are separate quantization issues." }, { "left": "issue:44155", "right": "issue:45823", "accept": false, - "reason": "A batched inference leak and a device-mismatch forward crash are unrelated." - }, - { - "left": "issue:44079", - "right": "issue:46829", - "accept": false, - "reason": "Generic ModelOutput key assignment logic and a VideoPrism output-field bug are not the same underlying bug." + "reason": "AudioFlamingo3 batched inference corruption and Gemma4 PLE device mismatch are unrelated." }, { "left": "issue:43957", "right": "issue:44756", "accept": false, - "reason": "Meta-device loading regressions and mmap OOM avoidance are unrelated." - }, - { - "left": "issue:43335", - "right": "issue:45663", - "accept": false, - "reason": "Sparse-layer creation in SwitchTransformers and Gemma4 FSDP2 attention state rebuilding are unrelated." + "reason": "Meta-device loading regressions and mmap OOM on Strix Halo are different failure modes." } ] }, @@ -2796,14 +2733,14 @@ "source": "llm", "cache_hit": false, "model": "gpt-5.4-mini?service_tier=flex", - "cluster_id": "cluster-10000-529", + "cluster_id": "cluster-10000-530", "nodes": [ "issue:36010", "issue:39404", "issue:42796", + "issue:43335", "issue:43381", "issue:43425", - "issue:43611", "issue:43645", "issue:43646", "issue:43708", @@ -2811,6 +2748,7 @@ "issue:43828", "issue:43854", "issue:43957", + "issue:44079", "issue:44164", "issue:44387", "issue:44402", @@ -2820,7 +2758,6 @@ "issue:44589", "issue:44683", "issue:44756", - "issue:44805", "issue:44863", "issue:44877", "issue:44928", @@ -2840,7 +2777,7 @@ "issue:46326", "issue:46468", "issue:46491", - "issue:46665", + "issue:46829", "issue:46869", "issue:46894", "issue:46901", @@ -2848,6 +2785,8 @@ "issue:47332" ], "soft_pairs": [ + "issue:44079|issue:46829", + "issue:43335|issue:45663", "issue:36010|issue:39404", "issue:44863|issue:46468", "issue:44479|issue:44877", @@ -2859,247 +2798,231 @@ "issue:42796|issue:46326", "issue:45588|issue:45663", "issue:43645|issue:44164", - "issue:43425|issue:47228", "issue:43854|issue:45216", - "issue:45684|issue:45698", + "issue:43425|issue:47228", "issue:45292|issue:45335", + "issue:45684|issue:45698", "issue:44589|issue:44756", "issue:43425|issue:43828", - "issue:43646|issue:43708", "issue:44402|issue:44451", - "issue:46894|issue:46901", + "issue:43646|issue:43708", "issue:43425|issue:46869", + "issue:46894|issue:46901", + "issue:44530|issue:45084", "issue:46315|issue:46491", - "issue:45812|issue:46315", "issue:44683|issue:45823", - "issue:44530|issue:45084", + "issue:45812|issue:46315", "issue:43425|issue:43716", - "issue:39404|issue:45701", - "issue:43611|issue:43645", - "issue:44479|issue:44805", - "issue:44589|issue:45084", - "issue:39404|issue:46665" + "issue:39404|issue:45701" ], - "prepared_review_unit_hash": "460fa9768dfc0d5c5a3a0dac5ba6601bd899027ccb627560f2f73ff07a1d281f", + "prepared_review_unit_hash": "486d32fb6ae4261496044dabd70e127209272e06ffa6fc9a5ed3aa8bd3c3fc8b", "component_budget": { - "node_count": 529, - "item_count": 529, - "soft_pair_count": 1953, - "serialized_chars": 1012074, - "estimated_input_tokens": 253019, - "estimated_eval_tokens": 506294 + "node_count": 530, + "item_count": 530, + "soft_pair_count": 1954, + "serialized_chars": 1013004, + "estimated_input_tokens": 253251, + "estimated_eval_tokens": 506758 }, "budget": { "node_count": 48, "item_count": 48, - "soft_pair_count": 31, - "serialized_chars": 36927, - "estimated_input_tokens": 9232, - "estimated_eval_tokens": 18720 + "soft_pair_count": 29, + "serialized_chars": 36354, + "estimated_input_tokens": 9089, + "estimated_eval_tokens": 18434 }, "overflow_policy": "truncate_then_skip", "trimmed": false, "aggressively_trimmed": false, "split": true, "analyst_result": { - "summary": "Most pairs are not true duplicates. The only clear duplicate families here are the resize_token_embeddings regression trio and the training-loss double-shift pair; the rest are similar by subsystem or symptom only.", - "confidence": 0.81, - "canonical_issue_reason": "issue:45292 is the best canonical issue because it states the underlying resize_token_embeddings regression in the most general form, and it anchors the related model-specific variants.", + "summary": "The set is mostly heterogeneous. I see two real duplicate pairs: the resize_token_embeddings/output_embeddings reports, and the training-loss double-shift reports. The rest are broad thematic overlaps, not the same underlying bug.", + "confidence": 0.67, + "canonical_issue_reason": "46901 is the strongest canonical anchor because it states the bug at implementation level (training against labels[..., 1:] instead of labels), while 46894 is the more model-specific duplicate. Most other items are unrelated enough that they should not influence the canonical choice.", "canonical_pr_reason": null, - "best_issue_reason": "issue:45292 is the most representative and reusable issue for the duplicate family: concise, model-agnostic, and directly describes the shared bug.", + "best_issue_reason": "46901 is the clearest overall issue to keep: it is precise, concrete, and directly actionable, whereas the remaining items are mostly broader or model-specific variants of separate problems.", "best_pr_reason": null, "soft_edge_verdicts": [ + { + "left": "issue:44079", + "right": "issue:46829", + "accept": false, + "reason": "Both are output-structure bugs, but one is generic ModelOutput key bookkeeping and the other is a VideoPrism classifier field mix-up; not the same concrete failure." + }, + { + "left": "issue:43335", + "right": "issue:45663", + "accept": false, + "reason": "Different bugs in different models: sparse-layer creation in SwitchTransformers vs shared_kv_states rebuild under FSDP2 in Gemma4." + }, { "left": "issue:36010", "right": "issue:39404", "accept": false, - "reason": "Different failures: an import error for GenerationMixin vs a Whisper pipeline return_language regression." + "reason": "Unrelated failures: an import error for GenerationMixin vs Whisper pipeline return_language behavior." }, { "left": "issue:44863", "right": "issue:46468", "accept": false, - "reason": "Both mention Nemotron, but one is a checkpoint-loading bug and the other is an example failure; not the same concrete issue." + "reason": "Same model family, but one is a checkpoint-loading failure and the other is a broken example; no clear same code path." }, { "left": "issue:44479", "right": "issue:44877", "accept": false, - "reason": "Unrelated: video-input regression for Qwen VL models vs strict config blocking granite_speech loading." + "reason": "Both touch multimodal/model loading, but one is a video-input regression and the other is strict config parsing for granite_speech; too broad to merge." }, { "left": "issue:44387", "right": "issue:44756", "accept": false, - "reason": "Both involve memory/OOM, but the causes are different (int4 reserved memory vs mmap on Strix Halo)." + "reason": "Both mention OOM, but the causes differ: int4 quantization memory growth vs mmap on Strix Halo." }, { "left": "issue:45276", "right": "issue:45292", "accept": true, - "reason": "Same resize_token_embeddings regression; 45276 is a model-specific manifestation, 45292 is the generic report." + "reason": "Same resize_token_embeddings propagation bug; 45276 is the Gemma4-specific variant while 45292 is the broader output_embeddings report." }, { "left": "issue:43381", "right": "issue:44928", "accept": false, - "reason": "Different bugs: gradient checkpointing in eval mode vs RLHF NaN explosion from 3D position_ids / SDPA fallback." + "reason": "Gradient checkpointing eval-mode restriction is unrelated to the RLHF NaN/SDPA math-fallback issue." }, { "left": "issue:44164", "right": "issue:44756", "accept": false, - "reason": "Save/from_pretrained extra_state handling is unrelated to the Strix Halo mmap OOM report." + "reason": "save/from_pretrained extra_state handling is unrelated to mmap-related OOM." }, { "left": "issue:43957", "right": "issue:47332", "accept": false, - "reason": "Both concern loading/offloading, but one is meta-device model loading and the other is saving a disk-offloaded model." + "reason": "Both involve loading/offloading, but one is meta-device model loading and the other is saving a disk-offloaded model; different problems." }, { "left": "issue:42796", "right": "issue:46326", "accept": false, - "reason": "Both are tokenizer-related, but BPE marker handling and continue_final_message on mistral_common are different code paths." + "reason": "Both are tokenizer-related, but one is BPE marker handling and the other is continue_final_message behavior for mistral_common tokenizers." }, { "left": "issue:45588", "right": "issue:45663", "accept": false, - "reason": "Different failures in different modules: flash_attention sink-less crash vs Gemma-4 FSDP2 shared_kv_states error." + "reason": "Different failures in different subsystems: flash_attention sinkless crash vs Gemma4 FSDP attention state rebuild." }, { "left": "issue:43645", "right": "issue:44164", "accept": false, - "reason": "Custom model initialization in notebooks is not the same bug as extra_state save/from_pretrained handling." + "reason": "Custom model initialization in notebooks is not the same bug as save/from_pretrained extra_state handling." + }, + { + "left": "issue:43854", + "right": "issue:45216", + "accept": false, + "reason": "Different save/load regressions affecting different models and APIs; no shared concrete failure mode." }, { "left": "issue:43425", "right": "issue:47228", "accept": false, - "reason": "Different torch-version issues: a version incompatibility claim vs a bf16 compilation regression in Sam3." + "reason": "Both are torch-version compatibility regressions, but one is a generic Torch 2.10 incompatibility and the other is a Sam3 bf16 compilation regression." }, { - "left": "issue:43854", - "right": "issue:45216", + "left": "issue:45292", + "right": "issue:45335", "accept": false, - "reason": "Different model-loading regressions affecting different models and APIs." + "reason": "Both are resize_token_embeddings regressions, but they target different model internals (output_embeddings vs decoder.embed_tokens) and are not clearly the same code path." }, { "left": "issue:45684", "right": "issue:45698", "accept": false, - "reason": "Both involve save_pretrained/from_pretrained for custom models, but one is about file permissions and the other about loading the wrong module." - }, - { - "left": "issue:45292", - "right": "issue:45335", - "accept": true, - "reason": "Same underlying resize_token_embeddings bug: embeddings are not propagated to other tied outputs/decoder embeddings." + "reason": "One is about permission propagation during save_pretrained; the other is about loading the wrong custom module after save_pretrained." }, { "left": "issue:44589", "right": "issue:44756", "accept": false, - "reason": "Storage lookup failure for Float8 tensors is unrelated to the mmap OOM issue." + "reason": "Float8 storage lookup failure is unrelated to the Strix Halo mmap OOM issue." }, { "left": "issue:43425", "right": "issue:43828", "accept": false, - "reason": "Torch version incompatibility is not the same as the Phi-tiny-MoE autocast dtype mismatch." + "reason": "Torch compatibility and autocast dtype mismatch are separate problems." + }, + { + "left": "issue:44402", + "right": "issue:44451", + "accept": false, + "reason": "One is tokenizer vocab size / special token placement; the other is a model-loading failure for ScandiBERT." }, { "left": "issue:43646", "right": "issue:43708", "accept": false, - "reason": "Custom model init breaks in Transformers 5.0, while the other is a Trainer resume_from_checkpoint max_steps miscalculation." + "reason": "Custom model initialization breakage is unrelated to Trainer resume_from_checkpoint max_steps miscalculation." }, { - "left": "issue:44402", - "right": "issue:44451", + "left": "issue:43425", + "right": "issue:46869", "accept": false, - "reason": "Possible related tokenizer loading problems, but the evidence is too broad to call them the same underlying bug." + "reason": "Generic Torch 2.10 incompatibility is not the same as YOLO multi-GPU device_map auto failure." }, { "left": "issue:46894", "right": "issue:46901", "accept": true, - "reason": "Same training-loss double-shift bug: both indicate labels are shifted incorrectly during loss computation." + "reason": "Same double-shift training-loss bug; both say the model is training against labels[..., 1:] instead of the intended labels." }, { - "left": "issue:43425", - "right": "issue:46869", + "left": "issue:44530", + "right": "issue:45084", "accept": false, - "reason": "Torch compatibility and YOLO multi-GPU device_map failures are different issues." + "reason": "PagedAttentionCache group-type crash and non-template-node compile TypeError are different failures." }, { "left": "issue:46315", "right": "issue:46491", "accept": false, - "reason": "Tokenizer performance regression and CodeLlama leading-space round-trip bug are distinct." - }, - { - "left": "issue:45812", - "right": "issue:46315", - "accept": false, - "reason": "Wrong token IDs for Granite models is unrelated to slow convert_tokens_to_ids behavior." + "reason": "Performance regression in convert_tokens_to_ids is unrelated to CodeLlama leading-space round-trip corruption." }, { "left": "issue:44683", "right": "issue:45823", "accept": false, - "reason": "Compiled flex_attention on torch>=2.9 and Gemma4 device mismatch are different code paths." + "reason": "Compiled flex_attention on torch >=2.9 and Gemma4 PLE device mismatch are different bugs." }, { - "left": "issue:44530", - "right": "issue:45084", + "left": "issue:45812", + "right": "issue:46315", "accept": false, - "reason": "PagedAttentionCache linear_attention crash is not the same as a compile non-template-nodes TypeError." + "reason": "Wrong Granite token IDs and slow convert_tokens_to_ids are separate tokenizer issues." }, { "left": "issue:43425", "right": "issue:43716", "accept": false, - "reason": "Torch compatibility issues are different from the Mistral-3 image preprocessor/model dtype mismatch." + "reason": "Torch 2.10 incompatibility is unrelated to Mistral-3 image preprocessor/model dtype mismatch." }, { "left": "issue:39404", "right": "issue:45701", "accept": false, - "reason": "Whisper pipeline return_language regression is unrelated to a general tokenization behavior change." - }, - { - "left": "issue:43611", - "right": "issue:43645", - "accept": false, - "reason": "Both are Transformers 5.0 breakages, but they affect different loading/init paths and are not the same bug." - }, - { - "left": "issue:44479", - "right": "issue:44805", - "accept": false, - "reason": "Video input regression and mask/tensor shape mismatch are unrelated failures." - }, - { - "left": "issue:44589", - "right": "issue:45084", - "accept": false, - "reason": "Float8 storage lookup failure is unrelated to the non-template-nodes compile TypeError." - }, - { - "left": "issue:39404", - "right": "issue:46665", - "accept": false, - "reason": "Whisper pipeline behavior is unrelated to from_pretrained inside Python classes." + "reason": "Whisper return_language pipeline regression is unrelated to the tokenization behavior change reported in 45701." } ] }, "evaluator_result": { "accept": true, - "feedback": "The summary is conservative and grounded in the packet. The accepted duplicate calls are well supported by the titles: the resize_token_embeddings cluster and the training-loss double-shift pair. The rejected pairs are framed narrowly and do not overstate identity. No PR-pair mergeability concerns apply here." + "feedback": "The analysis is grounded in the packet and stays mostly conservative. The two accepted soft pairs are supported by the issue titles, and the rejected pairs are framed as broad thematic similarities rather than the same concrete bug. The canonical and best-issue choices are also justified by the more precise, implementation-level wording of issue:46901." }, "evaluator_used": true, "retried": false, @@ -3118,7 +3041,7 @@ "source": "llm", "cache_hit": false, "model": "gpt-5.4-mini?service_tier=flex", - "cluster_id": "cluster-10000-529", + "cluster_id": "cluster-10000-530", "nodes": [ "issue:36032", "issue:39404", @@ -3132,6 +3055,7 @@ "issue:43611", "issue:43632", "issue:43644", + "issue:43645", "issue:43716", "issue:43746", "issue:43856", @@ -3142,40 +3066,43 @@ "issue:44079", "issue:44164", "issue:44265", - "issue:44462", + "issue:44479", "issue:44488", "issue:44509", "issue:44530", + "issue:44589", "issue:44671", "issue:44683", "issue:44756", + "issue:44805", "issue:44898", "issue:44912", "issue:44987", "issue:44991", "issue:45081", "issue:45084", - "issue:45479", "issue:45759", "issue:45823", "issue:46050", "issue:46326", - "issue:46489", + "issue:46665", "issue:46821", - "issue:46894", - "issue:46897", "issue:46899", "issue:47050", "issue:47331", "issue:47333" ], "soft_pairs": [ - "issue:40990|issue:47331", + "issue:43611|issue:43645", + "issue:44479|issue:44805", + "issue:44589|issue:45084", + "issue:39404|issue:46665", "issue:44164|issue:45759", + "issue:40990|issue:47331", "issue:36032|issue:43452", "issue:43994|issue:44079", - "issue:44265|issue:44898", "issue:43901|issue:44509", + "issue:44265|issue:44898", "issue:44991|issue:45081", "issue:46050|issue:46326", "issue:43906|issue:46326", @@ -3183,225 +3110,228 @@ "issue:43906|issue:44530", "issue:39404|issue:44912", "issue:45823|issue:46899", - "issue:44164|issue:46821", "issue:43596|issue:44683", "issue:44756|issue:47333", + "issue:44164|issue:46821", "issue:43856|issue:44756", "issue:44488|issue:44987", "issue:44488|issue:45081", - "issue:43632|issue:47050", "issue:36032|issue:42915", - "issue:45479|issue:46894", - "issue:43746|issue:47331", + "issue:43632|issue:47050", "issue:43716|issue:45084", "issue:43299|issue:43931", + "issue:43746|issue:47331", "issue:43611|issue:43644", - "issue:45479|issue:46897", - "issue:42796|issue:43404", - "issue:44462|issue:46489" + "issue:42796|issue:43404" ], - "prepared_review_unit_hash": "0e2e0500194053a39d0c2f3b1cc29c07760cfdbb2afa831b1509a9e65630c21c", + "prepared_review_unit_hash": "d548a79f40a64e471e2c645aca3bfc84d6c09548221ae2463697912d08bda8a6", "component_budget": { - "node_count": 529, - "item_count": 529, - "soft_pair_count": 1953, - "serialized_chars": 1012074, - "estimated_input_tokens": 253019, - "estimated_eval_tokens": 506294 + "node_count": 530, + "item_count": 530, + "soft_pair_count": 1954, + "serialized_chars": 1013004, + "estimated_input_tokens": 253251, + "estimated_eval_tokens": 506758 }, "budget": { "node_count": 48, "item_count": 48, - "soft_pair_count": 29, - "serialized_chars": 36654, - "estimated_input_tokens": 9164, - "estimated_eval_tokens": 18584 + "soft_pair_count": 30, + "serialized_chars": 36942, + "estimated_input_tokens": 9236, + "estimated_eval_tokens": 18728 }, "overflow_policy": "truncate_then_skip", "trimmed": false, "aggressively_trimmed": false, "split": true, "analyst_result": { - "summary": "Heterogeneous set of mostly unrelated bug reports. I did not find any soft pair that looks like the same underlying issue, so all proposed merges are rejected.", - "confidence": 0.94, - "canonical_issue_reason": "No single canonical issue: the items span distinct regressions in tokenizers, model loading, quantization, export, docs, and training behavior.", + "summary": "This cluster is a broad mix of mostly unrelated Transformers v5 regression reports, heavily centered on loading/tokenizer/model-init issues. I do not see any soft edge that clearly represents the same concrete bug, so all candidate duplicate links should stay unmerged.", + "confidence": 0.42, + "canonical_issue_reason": "issue:43611 is the best anchor because it is a clear, concrete v5 loading regression and broadly representative of the common \"v5 breaks loading\" theme across the cluster.", "canonical_pr_reason": null, - "best_issue_reason": "No global best issue for this set; 43611 is the broadest v5-loading regression, but it does not represent the rest of the cluster well enough to serve as an umbrella.", + "best_issue_reason": "issue:43611 is the strongest representative issue: it is specific enough to describe a real regression while still being broad compared with the more model-specific reports.", "best_pr_reason": null, "soft_edge_verdicts": [ { - "left": "issue:40990", - "right": "issue:47331", + "left": "issue:43611", + "right": "issue:43645", + "accept": false, + "reason": "Different failure surfaces: base_model_prefix loading vs. defining/initializing custom models in Jupyter notebooks." + }, + { + "left": "issue:44479", + "right": "issue:44805", + "accept": false, + "reason": "Both are multimodal regressions, but one is a video-input regression across Qwen VL variants and the other is a mask/tensor shape IndexError." + }, + { + "left": "issue:44589", + "right": "issue:45084", + "accept": false, + "reason": "Unrelated bugs: Float8 storage lookup failure vs. a template-compilation error." + }, + { + "left": "issue:39404", + "right": "issue:46665", "accept": false, - "reason": "Different failures: gpt-oss perplexity on WikiText-2 vs MetalConfig batched quantized generate corruption." + "reason": "Whisper pipeline return_language regression is a different API/path from from_pretrained behavior inside Python classes." }, { "left": "issue:44164", "right": "issue:45759", "accept": false, - "reason": "Different code paths: extra_state save/load handling vs unwrapping text_config in composite Qwen multimodal configs." + "reason": "Different code paths: extra_state save/load handling vs. unwrapping text_config for composite Qwen multimodal configs." + }, + { + "left": "issue:40990", + "right": "issue:47331", + "accept": false, + "reason": "Perplexity quality issue on gpt-oss-20b is not the same as Metal quantization batch corruption in generation." }, { "left": "issue:36032", "right": "issue:43452", "accept": false, - "reason": "Tokenizer init conflict with add_special_tokens is unrelated to gguf_file loading in AutoTokenizer/AutoModel." + "reason": "Tokenizer special-token method conflict and gguf_file from_pretrained handling are separate loading bugs." }, { "left": "issue:43994", "right": "issue:44079", "accept": false, - "reason": "Siglip2 nonsensical outputs and ModelOutput key assignment bugs are different problems." + "reason": "SigLIP2 nonsensical outputs and ModelOutput key assignment with None are different layers of the stack." }, { - "left": "issue:44265", - "right": "issue:44898", + "left": "issue:43901", + "right": "issue:44509", "accept": false, - "reason": "Different export/classification paths: torch_compilable_check export failure vs Perceiver interpolate_pos_encoding image classification failure." + "reason": "Both touch docs, but they cover different stale documentation problems and different removed/changed API surface." }, { - "left": "issue:43901", - "right": "issue:44509", + "left": "issue:44265", + "right": "issue:44898", "accept": false, - "reason": "Both are docs issues, but they cover different text and different pipeline-task deprecations." + "reason": "torch.export failures with torch_compilable_check are not the same as Perceiver non-default-resolution interpolation failures." }, { "left": "issue:44991", "right": "issue:45081", "accept": false, - "reason": "Different tokenizer bugs: est-roberta loading vs Mistral regex patch crashing on backend_tokenizer access." + "reason": "Different tokenizer/loading bugs: EMBEDDIA/est-roberta vs. Mistral regex patching backend_tokenizer." }, { "left": "issue:46050", "right": "issue:46326", "accept": false, - "reason": "Quantized save/load breakage is not the same as continue_final_message behavior for mistral_common tokenizers." + "reason": "Quantized save/load corruption is unrelated to continue_final_message behavior for mistral_common tokenizers." }, { "left": "issue:43906", "right": "issue:46326", "accept": false, - "reason": "One is a reproduction of issue #38071; the other is a tokenizer continue_final_message bug." + "reason": "An isolated reproduction of an older issue does not match the specific continue_final_message tokenizer bug." }, { "left": "issue:44671", "right": "issue:44991", "accept": false, - "reason": "CamemBERT masked-LM prediction regression is unrelated to est-roberta tokenizer loading." + "reason": "CamemBERT masked LM prediction regression is unrelated to tokenizer loading for EMBEDDIA/est-roberta." }, { "left": "issue:43906", "right": "issue:44530", "accept": false, - "reason": "Isolated reproduction of #38071 does not match the PagedAttentionCache linear_attention crash." + "reason": "No shared concrete bug: one is a reproduction issue, the other is a Qwen3.5 cache crash on linear_attention group type." }, { "left": "issue:39404", "right": "issue:44912", "accept": false, - "reason": "Whisper pipeline return_language regression is unrelated to gpt-oss MXFP4 quantization falling back to bf16." + "reason": "Whisper return_language and MXFP4 quantization fallback are separate regressions." }, { "left": "issue:45823", "right": "issue:46899", "accept": false, - "reason": "Same model family, but different root causes: device_map auto placement vs unsafe dtype cast in 4-bit multimodal embedders." - }, - { - "left": "issue:44164", - "right": "issue:46821", - "accept": false, - "reason": "Extra_state save/load failure is unrelated to the TimesFM window_size AttributeError." + "reason": "Both involve Gemma4 multimodal paths, but one is device-map placement and the other is an unsafe dtype cast in 4-bit embedders." }, { "left": "issue:43596", "right": "issue:44683", "accept": false, - "reason": "DeepSpeed zero3/BertModel index error is a different failure from compiled flex_attention on torch >= 2.9." + "reason": "ZeRO3/BertModel init IndexError and flex_attention compile failure on torch>=2.9 are unrelated failures." }, { "left": "issue:44756", "right": "issue:47333", "accept": false, - "reason": "Strix Halo mmap OOM mitigation and disk-offloaded Qwen3 save failure are different memory/offload bugs." + "reason": "mmap OOM on Strix Halo and disk-offloaded Qwen3 save failure are different memory/offload issues." + }, + { + "left": "issue:44164", + "right": "issue:46821", + "accept": false, + "reason": "extra_state save/load handling is unrelated to the TimesFM window_size AttributeError." }, { "left": "issue:43856", "right": "issue:44756", "accept": false, - "reason": "Both mention memory, but one is MoE training memory use and the other is hardware-specific mmap OOM avoidance." + "reason": "Qwen3 MoE training memory inefficiency is not the same as disabling mmap to avoid OOM on Strix Halo." }, { "left": "issue:44488", "right": "issue:44987", "accept": false, - "reason": "Different model-loading failures for different repositories and different breakages." + "reason": "Both are load failures, but they affect different models and the titles point to different root causes." }, { "left": "issue:44488", "right": "issue:45081", "accept": false, - "reason": "sleng-bert loading failure is unrelated to the Mistral tokenizer regex crash." - }, - { - "left": "issue:43632", - "right": "issue:47050", - "accept": false, - "reason": "HF initialization flag regression is unrelated to CanineModel fp16 ONNX attention dtype mismatch." + "reason": "Different model-loading problems: cjvt/sleng-bert vs. the Mistral regex/backend_tokenizer crash." }, { "left": "issue:36032", "right": "issue:42915", "accept": false, - "reason": "Tokenizer add_special_tokens conflict is unrelated to Qwen3Moe FineGrainedFP8Config loading failure." + "reason": "Tokenizer add_special_tokens conflict and Qwen3Moe FineGrainedFP8Config failure are unrelated." }, { - "left": "issue:45479", - "right": "issue:46894", + "left": "issue:43632", + "right": "issue:47050", "accept": false, - "reason": "Single-label zero-loss classification bug is not the same as CohereASR's double-shift training-loss bug." - }, - { - "left": "issue:43746", - "right": "issue:47331", - "accept": false, - "reason": "GraniteSpeech PEFT checkpoint loading and MetalConfig batched quantization corruption are unrelated." + "reason": "_is_hf_initialized flag regression and ONNX attention dtype mismatch are separate bugs." }, { "left": "issue:43716", "right": "issue:45084", "accept": false, - "reason": "Mistral-3 image preprocessor dtype mismatch is unrelated to template compilation TypeError." + "reason": "Mistral-3 dtype mismatch in preprocessing/modeling is unrelated to the template-node compilation error." }, { "left": "issue:43299", "right": "issue:43931", "accept": false, - "reason": "Both involve Qwen3-VL, but one is a v5 loading regression and the other is a weight-shape mismatch on a specific checkpoint." + "reason": "Same model family, but different concrete problems: Qwen3VL Moe loading break vs. weight-shape mismatch on a specific checkpoint." }, { - "left": "issue:43611", - "right": "issue:43644", + "left": "issue:43746", + "right": "issue:47331", "accept": false, - "reason": "Both are v5 regressions, but one concerns base_model_prefix loading and the other non-persistent buffer initialization; distinct bugs." + "reason": "GraniteSpeech PEFT local checkpoint loading and MetalConfig batched quantization corruption are unrelated." }, { - "left": "issue:45479", - "right": "issue:46897", + "left": "issue:43611", + "right": "issue:43644", "accept": false, - "reason": "Different model families and different implementations of the loss bug; not the same concrete issue." + "reason": "base_model_prefix loading regression and non-persistent buffers being filled with junk are different code paths." }, { "left": "issue:42796", "right": "issue:43404", "accept": false, - "reason": "Ministral tokenizer BPE-marker handling is unrelated to Mistral3 lm_head weight tying." - }, - { - "left": "issue:44462", - "right": "issue:46489", - "accept": false, - "reason": "AutoTokenizer ignoring tokenizer.json is a different bug from DeepSeek-Coder tokenizer output regression on v5." + "reason": "Tokenizer BPE-marker handling and lm_head weight tying in Mistral3 are different underlying issues." } ] }, @@ -3423,7 +3353,7 @@ "source": "llm", "cache_hit": false, "model": "gpt-5.4-mini?service_tier=flex", - "cluster_id": "cluster-10000-529", + "cluster_id": "cluster-10000-530", "nodes": [ "issue:34567", "issue:36032", @@ -3432,9 +3362,7 @@ "issue:40990", "issue:42175", "issue:42915", - "issue:43208", "issue:43262", - "issue:43317", "issue:43421", "issue:43452", "issue:43482", @@ -3451,6 +3379,7 @@ "issue:44220", "issue:44403", "issue:44451", + "issue:44462", "issue:44530", "issue:44661", "issue:44671", @@ -3461,6 +3390,7 @@ "issue:44991", "issue:45081", "issue:45084", + "issue:45479", "issue:45507", "issue:45584", "issue:45859", @@ -3468,254 +3398,252 @@ "issue:46285", "issue:46326", "issue:46402", - "issue:46661", - "issue:47013", + "issue:46489", + "issue:46894", + "issue:46897", "issue:47312", "issue:47331" ], "soft_pairs": [ + "issue:45479|issue:46894", + "issue:44462|issue:46489", "issue:44661|issue:44683", "issue:46285|issue:46402", "issue:43541|issue:45084", + "issue:45479|issue:46897", "issue:44451|issue:44987", "issue:42915|issue:44530", "issue:43873|issue:47331", - "issue:40990|issue:43421", "issue:43643|issue:44530", + "issue:40990|issue:43421", "issue:36032|issue:43645", "issue:40444|issue:46326", + "issue:39404|issue:44671", "issue:34567|issue:40444", "issue:44737|issue:45859", - "issue:39404|issue:44671", "issue:43644|issue:43646", "issue:42175|issue:46178", "issue:44987|issue:44991", - "issue:43582|issue:43632", "issue:44206|issue:44220", "issue:45084|issue:45507", + "issue:43582|issue:43632", "issue:43452|issue:43482", "issue:43482|issue:47312", "issue:43582|issue:47331", - "issue:44811|issue:45584", "issue:43746|issue:46326", - "issue:43262|issue:47331", - "issue:43208|issue:47013", + "issue:44811|issue:45584", "issue:44403|issue:45081", - "issue:43317|issue:43482", - "issue:43645|issue:46661" + "issue:43262|issue:47331" ], - "prepared_review_unit_hash": "8da3bdda165ef06a3536193520633072e304ebee39bf3296390ecef180a524f1", + "prepared_review_unit_hash": "3cc426d2001703671d4aa90f7a34d07d42a68edbf8f86986ae81cbf9670dee5d", "component_budget": { - "node_count": 529, - "item_count": 529, - "soft_pair_count": 1953, - "serialized_chars": 1012074, - "estimated_input_tokens": 253019, - "estimated_eval_tokens": 506294 + "node_count": 530, + "item_count": 530, + "soft_pair_count": 1954, + "serialized_chars": 1013004, + "estimated_input_tokens": 253251, + "estimated_eval_tokens": 506758 }, "budget": { - "node_count": 47, - "item_count": 47, + "node_count": 48, + "item_count": 48, "soft_pair_count": 29, - "serialized_chars": 35907, - "estimated_input_tokens": 8977, - "estimated_eval_tokens": 18210 + "serialized_chars": 36482, + "estimated_input_tokens": 9121, + "estimated_eval_tokens": 18498 }, "overflow_policy": "truncate_then_skip", "trimmed": false, "aggressively_trimmed": false, "split": true, "analyst_result": { - "summary": "This set is mostly heterogeneous v5/regression reports, not a single tight duplicate cluster. I only see two plausible duplicate pairs: GGUF loading regressions and an audio feature-extraction crash path.", - "confidence": 0.63, - "canonical_issue_reason": "issue:43873 is the broadest, most actively discussed report here and has the strongest evidence of being a central standalone bug rather than a narrow one-off.", + "summary": "This cluster is mostly a noisy collection of unrelated Transformers issues that happen to share broad v5-regression vocabulary; none of the soft pairs look like true duplicates of the same bug/change.", + "confidence": 0.87, + "canonical_issue_reason": "If a single issue must represent the cluster, issue 43646 is the broadest and most central v5 regression report here (custom model initialization). That said, the set is heterogeneous, so it is only a loose representative, not a true duplicate umbrella.", "canonical_pr_reason": null, - "best_issue_reason": "issue:43873 is the best overall representative issue in this set because it is broad, high-signal, and has the most inbound attention.", + "best_issue_reason": "Issue 43646 is the best single anchor because it is concise, clearly states a concrete regression, and is broad enough to be a representative reference for nearby v5 initialization complaints.", "best_pr_reason": null, "soft_edge_verdicts": [ + { + "left": "issue:45479", + "right": "issue:46894", + "accept": false, + "reason": "Different model families and different manifestations of the training-loss double-shift bug; related pattern, not the same underlying issue." + }, + { + "left": "issue:44462", + "right": "issue:46489", + "accept": false, + "reason": "Both concern tokenizer loading, but one is about ignoring tokenizer.json and the other is a DeepSeek-Coder-specific tokenizer regression from a prior fix gap." + }, { "left": "issue:44661", "right": "issue:44683", "accept": false, - "reason": "Different subsystems: tokenizer-mapping bug vs compiled flex_attention regression." + "reason": "One is a tokenizer mapping/add-new-model failure, the other is a torch flex_attention compilation issue; separate code paths." }, { "left": "issue:46285", "right": "issue:46402", "accept": false, - "reason": "Different model families and failures: CLIP structure inconsistency vs Sam3 text-encoder loading." + "reason": "Different bugs: class structure inconsistency vs text encoder weights not being loaded." }, { "left": "issue:43541", "right": "issue:45084", "accept": false, - "reason": "Both are runtime/compile errors, but the concrete code paths and symptoms are unrelated." + "reason": "Unrelated failures in torch tracing/grouped_mm versus Jinja template compilation." + }, + { + "left": "issue:45479", + "right": "issue:46897", + "accept": false, + "reason": "Same general double-shift pattern, but different affected model and separate issue reports." }, { "left": "issue:44451", "right": "issue:44987", "accept": false, - "reason": "Both are load failures, but the affected models and likely root causes differ." + "reason": "Both are load failures, but on different models with different breakages and no clear shared bug." }, { "left": "issue:42915", "right": "issue:44530", "accept": false, - "reason": "Both mention Qwen, but one is FineGrainedFP8Config and the other is a PagedAttentionCache/linear_attention crash." + "reason": "Qwen3Moe FP8 config failure and Qwen3.5 PagedAttentionCache invalid group type are distinct runtime problems." }, { "left": "issue:43873", "right": "issue:47331", "accept": false, - "reason": "Both involve quantization/offload, but one is generic offloading behavior and the other is a specific Metal batched-generate corruption bug." + "reason": "Quantization/offloading difficulty versus a specific MetalConfig batched-generate corruption bug; not the same code path." }, { - "left": "issue:40990", - "right": "issue:43421", + "left": "issue:43643", + "right": "issue:44530", "accept": false, - "reason": "Perplexity/debugging issue vs tokenizer post-processor update behavior; no same underlying bug evident." + "reason": "AutoConfig trust_remote_code missing fields is unrelated to Qwen3.5 cache/group-type crashes." }, { - "left": "issue:43643", - "right": "issue:44530", + "left": "issue:40990", + "right": "issue:43421", "accept": false, - "reason": "Remote-code config loading vs Qwen3.5 PagedAttentionCache crash are unrelated." + "reason": "Perplexity regression on gpt-oss is not the same as runtime post-processor updates for tokenizer special tokens." }, { "left": "issue:36032", "right": "issue:43645", "accept": false, - "reason": "Tokenizer method-name conflict vs custom model initialization in notebooks are different failures." + "reason": "Tokenizer method conflict during load is unrelated to custom model initialization in notebooks." }, { "left": "issue:40444", "right": "issue:46326", "accept": false, - "reason": "Different multimodal tokenizer/training paths with no shared concrete bug shown." + "reason": "Multi-image IterableDataset fine-tuning and mistral_common continue_final_message handling are different bugs." + }, + { + "left": "issue:39404", + "right": "issue:44671", + "accept": false, + "reason": "Whisper pipeline return_language regression and CamemBERT masked-LM predictions are unrelated." }, { "left": "issue:34567", "right": "issue:40444", "accept": false, - "reason": "Trainer token counting vs Qwen2.5-VL dataset handling are unrelated." + "reason": "TrainerState token counter not updating is unrelated to Qwen2.5-VL IterableDataset multi-image failures." }, { "left": "issue:44737", "right": "issue:45859", "accept": false, - "reason": "Both are device/offload related, but XLNet positional encoding and Qwen3_5Moe rotary embedding are distinct code paths." - }, - { - "left": "issue:39404", - "right": "issue:44671", - "accept": false, - "reason": "Whisper pipeline behavior vs CamemBERT masked LM predictions are unrelated." + "reason": "CPU arange device bug in XLNet and CPU-offload incompatibility in Qwen3_5Moe are different implementation issues." }, { "left": "issue:43644", "right": "issue:43646", "accept": false, - "reason": "Both are v5 regressions, but one is about junk-filled buffers and the other about custom model initialization; not the same concrete bug." + "reason": "Both are v5 regressions, but one is about non-persistent buffers and the other about custom model initialization; not the same bug." }, { "left": "issue:42175", "right": "issue:46178", "accept": false, - "reason": "Packaging/backend-mapping issues, but one is missing tensorflow in an extra and the other is an invalid tensorflow_text backend entry." + "reason": "Package/backend dependency inclusion and backend mapping validation are distinct installation/backends issues." }, { "left": "issue:44987", "right": "issue:44991", "accept": false, - "reason": "Both are tokenizer/model loading failures in v5, but they appear model-specific rather than the same bug." - }, - { - "left": "issue:43582", - "right": "issue:43632", - "accept": false, - "reason": "AppleSilicon allocator TypeError vs _is_hf_initialized regression are unrelated." + "reason": "Different model-loading failures with different symptoms; no evidence they share one underlying tokenizer bug." }, { "left": "issue:44206", "right": "issue:44220", - "accept": true, - "reason": "Both point to the same audio feature-extraction path failing around _torch_extract_fbank_features / unsupported center handling." + "accept": false, + "reason": "Unsupported center arg in feature extractor versus _torch_extract_fbank_features() behavior; related audio area but not the same failure." }, { "left": "issue:45084", "right": "issue:45507", "accept": false, - "reason": "Template compilation error vs GraniteMoEHybrid invalid-method call are different failure modes." + "reason": "Template compilation error and GraniteMoEHybrid invalid method calls are separate problems." + }, + { + "left": "issue:43582", + "right": "issue:43632", + "accept": false, + "reason": "Apple Silicon allocator warmup TypeError is unrelated to the _is_hf_initialized flag regression." }, { "left": "issue:43452", "right": "issue:43482", - "accept": true, - "reason": "Both describe the same GGUF loading regression in transformers v5; one is generic from_pretrained/gguf_file breakage and the other is a model-specific manifestation." + "accept": false, + "reason": "GGUF loading breakage and Qwen2.5-GGUF loading failure are different model/loading issues without a shared concrete fix." }, { "left": "issue:43482", "right": "issue:47312", "accept": false, - "reason": "GGUF loading regression vs kernel compatibility issue; different bugs." + "reason": "Qwen2.5-GGUF v5 loading failure is unrelated to kernels 0.16 incompatibility." }, { "left": "issue:43582", "right": "issue:47331", "accept": false, - "reason": "AppleSilicon caching allocator error vs Metal quantization corruption are different code paths." - }, - { - "left": "issue:44811", - "right": "issue:45584", - "accept": false, - "reason": "Both are Whisper-related, but one is batch_decode skip_special_tokens behavior and the other is empty transcription after align_special_tokens." + "reason": "Allocator warmup TypeError and MetalConfig batched-generate corruption are different bugs and code paths." }, { "left": "issue:43746", "right": "issue:46326", "accept": false, - "reason": "PEFT checkpoint loading vs continue_final_message tokenizer behavior are unrelated." + "reason": "PEFT-adapter local checkpoint loading for GraniteSpeech is unrelated to mistral_common continue_final_message behavior." }, { - "left": "issue:43262", - "right": "issue:47331", - "accept": false, - "reason": "Audio processor sampling-rate defaulting vs Metal quantization batch corruption are unrelated." - }, - { - "left": "issue:43208", - "right": "issue:47013", + "left": "issue:44811", + "right": "issue:45584", "accept": false, - "reason": "Both are xLSTM-related, but one is a trio of training bugs and the other is a specific return_last_states=False crash." + "reason": "Whisper batch_decode skip_special_tokens handling and empty-transcription generation after align_special_tokens are different issues." }, { "left": "issue:44403", "right": "issue:45081", "accept": false, - "reason": "Generic loading noise vs Mistral regex patch crash are not the same underlying issue." - }, - { - "left": "issue:43317", - "right": "issue:43482", - "accept": false, - "reason": "Dequantized model offload failure vs GGUF loading failure are different problems." + "reason": "Loading noise vs Mistral regex patch backend_tokenizer crash are not the same underlying defect." }, { - "left": "issue:43645", - "right": "issue:46661", + "left": "issue:43262", + "right": "issue:47331", "accept": false, - "reason": "Custom model initialization in notebooks vs remote-code config_class error are distinct initialization paths." + "reason": "Audio processor chat-template sampling-rate default and Metal batched-generate quantization corruption are unrelated." } ] }, - "evaluator_result": { - "accept": true, - "feedback": "The summary is grounded in the packet: issue:43873 is a reasonable canonical/best representative based on its generic title, high discussion, and inbound references. The soft-edge verdicts are mostly conservative and the two accepted pairs are plausibly the same bug family based on the titles alone. I do not see evidence of overstatement." - }, - "evaluator_used": true, + "evaluator_result": null, + "evaluator_used": false, "retried": false, - "accepted_nontrivial_soft_edge": true, + "accepted_nontrivial_soft_edge": false, "error_kind": null, "error_message": null }, @@ -3730,33 +3658,32 @@ "source": "llm", "cache_hit": false, "model": "gpt-5.4-mini?service_tier=flex", - "cluster_id": "cluster-10000-529", + "cluster_id": "cluster-10000-530", "nodes": [ "issue:28530", "issue:34567", "issue:36032", - "issue:40444", "issue:41720", "issue:42915", - "issue:43089", + "issue:43208", "issue:43262", - "issue:43296", + "issue:43317", "issue:43322", "issue:43377", "issue:43404", "issue:43425", + "issue:43482", "issue:43531", "issue:43540", "issue:43632", + "issue:43645", "issue:43646", "issue:43742", "issue:43792", "issue:43819", - "issue:43872", "issue:43994", "issue:44008", "issue:44062", - "issue:44220", "issue:44265", "issue:44291", "issue:44403", @@ -3767,7 +3694,6 @@ "issue:44534", "issue:44589", "issue:44610", - "issue:44683", "issue:44849", "issue:44877", "issue:44987", @@ -3776,251 +3702,239 @@ "issue:45325", "issue:45923", "issue:45987", + "issue:46661", "issue:46821", "issue:46869", "issue:46899", + "issue:47013", "issue:47254" ], "soft_pairs": [ + "issue:43317|issue:43482", + "issue:43208|issue:47013", + "issue:43645|issue:46661", "issue:43792|issue:46869", "issue:42915|issue:45084", "issue:43646|issue:44534", "issue:44987|issue:45081", - "issue:43262|issue:43377", "issue:42915|issue:44291", + "issue:43262|issue:43377", "issue:36032|issue:43742", "issue:28530|issue:34567", "issue:43377|issue:43819", "issue:44062|issue:44987", - "issue:43404|issue:44466", "issue:44451|issue:45081", + "issue:43404|issue:44466", "issue:43994|issue:44610", - "issue:44849|issue:45325", "issue:42915|issue:44589", + "issue:44849|issue:45325", "issue:36032|issue:41720", "issue:43531|issue:44877", "issue:44265|issue:44610", "issue:44008|issue:46899", - "issue:45923|issue:47254", "issue:44479|issue:45325", - "issue:44464|issue:45987", "issue:43322|issue:43540", - "issue:43632|issue:46821", - "issue:43089|issue:45987", + "issue:45923|issue:47254", + "issue:44464|issue:45987", "issue:43425|issue:44403", - "issue:43296|issue:43531", - "issue:40444|issue:45325", - "issue:44220|issue:44683", - "issue:43872|issue:44291" + "issue:43632|issue:46821" ], - "prepared_review_unit_hash": "65bb2145487be7243b16f8b2c61f4d04f89e95fdc4efa40c6c070abf83ca81cc", + "prepared_review_unit_hash": "f363cab83bac1eb1b8153ad2829f08ef154aaa05c3d4be217d30f2ab65ee6607", "component_budget": { - "node_count": 529, - "item_count": 529, - "soft_pair_count": 1953, - "serialized_chars": 1012074, - "estimated_input_tokens": 253019, - "estimated_eval_tokens": 506294 + "node_count": 530, + "item_count": 530, + "soft_pair_count": 1954, + "serialized_chars": 1013004, + "estimated_input_tokens": 253251, + "estimated_eval_tokens": 506758 }, "budget": { "node_count": 48, "item_count": 48, - "soft_pair_count": 30, - "serialized_chars": 36779, - "estimated_input_tokens": 9195, - "estimated_eval_tokens": 18646 + "soft_pair_count": 28, + "serialized_chars": 36256, + "estimated_input_tokens": 9064, + "estimated_eval_tokens": 18384 }, "overflow_policy": "truncate_then_skip", "trimmed": false, "aggressively_trimmed": false, "split": true, "analyst_result": { - "summary": "The set is heterogeneous, but the strongest repeated theme is Transformers v5 initialization-compatibility regressions around `_is_hf_initialized` / `init_empty_weights`; no soft pair looks like a true duplicate, and there are several unrelated issue subclusters.", - "confidence": 0.79, - "canonical_issue_reason": "Issue 44291 is the clearest, most general representative of the `_is_hf_initialized`/model-init compatibility break that also appears in nearby issues like 43872 and 43632.", + "summary": "This is not a duplicate cluster: the items are a broad mix of unrelated Transformers bugs across training, tokenization, multimodal processing, device mapping, quantization, and serialization. All soft pairs look like loose topical similarity rather than the same underlying issue.", + "confidence": 0.97, + "canonical_issue_reason": "No canonical issue stands out because the set is heterogeneous and does not share one concrete code-path failure.", "canonical_pr_reason": null, - "best_issue_reason": "44291 has the broadest root-cause framing and is easiest to use as the cluster representative for the shared v5 initialization regression theme.", + "best_issue_reason": "No issue is globally suitable as a canonical representative; choosing any one would be arbitrary given the lack of a common bug theme.", "best_pr_reason": null, "soft_edge_verdicts": [ + { + "left": "issue:43317", + "right": "issue:43482", + "accept": false, + "reason": "Both involve loading/dequantization behavior, but one is device_map offload and the other is GGUF/v5 loading; different failure modes and code paths." + }, + { + "left": "issue:43208", + "right": "issue:47013", + "accept": false, + "reason": "Both mention xLSTM, but one is a trio of training bugs for smaller-than-7B models and the other is a specific return_last_states unpacking crash." + }, + { + "left": "issue:43645", + "right": "issue:46661", + "accept": false, + "reason": "Both concern custom/remote-code model initialization under v5, but one is a Jupyter-specific notebook issue and the other is an AutoModel config_class error for remote-code models." + }, { "left": "issue:43792", "right": "issue:46869", "accept": false, - "reason": "Different models and failure modes: Whisper cannot run vs YOLO multi-GPU device-map auto failure." + "reason": "Both mention model loading/runtime failure, but the first is Whisper not running and the second is YOLO failing with multi-GPU device_map auto; unrelated." }, { "left": "issue:42915", "right": "issue:45084", "accept": false, - "reason": "Unrelated bugs: Qwen3Moe FP8 config failure vs template compilation error." + "reason": "Qwen3MoE FP8 failure and a chat-template compilation TypeError are unrelated subsystems." }, { "left": "issue:43646", "right": "issue:44534", "accept": false, - "reason": "Both are v5-related, but one is custom model initialization breakage and the other is non-persistent buffer corruption; not the same bug." + "reason": "One is custom model initialization breaking in v5, the other is non-persistent buffer serialization corruption; different bugs." }, { "left": "issue:44987", "right": "issue:45081", "accept": false, - "reason": "Different code paths: loading physical-intelligence/fast vs Mistral tokenizer regex patching." + "reason": "Both are loading failures, but one is a model-specific transformers>=5.1 regression and the other is a Mistral tokenizer regex patch crash." }, { - "left": "issue:43262", - "right": "issue:43377", + "left": "issue:42915", + "right": "issue:44291", "accept": false, - "reason": "Audio chat-template sampling-rate default bug vs MIMI encoder batching/padding-mask mismatch." + "reason": "Both reference v5-era model loading, but one is Qwen3Moe FP8 config failure and the other is init_empty_weights/_is_hf_initialized argument handling." }, { - "left": "issue:42915", - "right": "issue:44291", + "left": "issue:43262", + "right": "issue:43377", "accept": false, - "reason": "Both mention loading/config issues, but the concrete problems differ: FP8 config failure vs `_is_hf_initialized` init-argument incompatibility." + "reason": "Audio processor sampling-rate defaulting and MIMI encoder padding-mask behavior are distinct multimodal bugs." }, { "left": "issue:36032", "right": "issue:43742", "accept": false, - "reason": "Tokenizer method-name conflict is unrelated to a key error loading a specific model." + "reason": "Tokenizer special-token conflict and MobileLLM key error are unrelated model-loading problems." }, { "left": "issue:28530", "right": "issue:34567", "accept": false, - "reason": "Training early-stopping metric handling vs token-count tracking; unrelated trainer issues." + "reason": "Early stopping metric selection and num_input_tokens_seen tracking are separate Trainer features." }, { "left": "issue:43377", "right": "issue:43819", "accept": false, - "reason": "Different multimodal/audio bugs: batching parity in MIMI encoder vs DAC latent decoding mismatch." + "reason": "MIMI encoder batching/padding bug and DAC from_latents/forward mismatch are different audio-model correctness issues." }, { "left": "issue:44062", "right": "issue:44987", "accept": false, - "reason": "Tokenizer AddedToken keyword collision vs model loading failure for a specific checkpoint." + "reason": "AddedToken keyword collision and a specific model-loading regression are unrelated tokenizer vs model-loading failures." }, { - "left": "issue:43404", - "right": "issue:44466", + "left": "issue:44451", + "right": "issue:45081", "accept": false, - "reason": "Weight tying bug in a model class vs inconsistent lm_head serialization depending on device." + "reason": "ScandiBERT loading failure and Mistral regex patch crash are different tokenizer/model issues." }, { - "left": "issue:44451", - "right": "issue:45081", + "left": "issue:43404", + "right": "issue:44466", "accept": false, - "reason": "Generic model loading failure vs tokenizer regex patch crash; no shared underlying defect." + "reason": "lm_head weight tying in Mistral3 and device-dependent lm_head serialization are related only broadly by tied weights, but they are not the same concrete bug." }, { "left": "issue:43994", "right": "issue:44610", "accept": false, - "reason": "Different vision model problems: nonsensical outputs vs processor/model input-size mismatch." + "reason": "SigLIP2 nonsensical outputs and OmDet-Turbo processor size mismatch are separate vision-model bugs." }, { - "left": "issue:44849", - "right": "issue:45325", + "left": "issue:42915", + "right": "issue:44589", "accept": false, - "reason": "Output-hidden-states bug in Qwen3.5 vs rope-index scaling bug in Qwen2.5-VL." + "reason": "Qwen3Moe FP8 config failure and Float8 storage lookup failure are different quantization/runtime issues." }, { - "left": "issue:42915", - "right": "issue:44589", + "left": "issue:44849", + "right": "issue:45325", "accept": false, - "reason": "FP8 config failure is unrelated to float8 storage lookup error." + "reason": "Qwen3.5 output_hidden_states behavior and Qwen2.5-VL rope-index scaling are different model-specific logic bugs." }, { "left": "issue:36032", "right": "issue:41720", "accept": false, - "reason": "Tokenizer loading error vs Qwen3 auto device-mapping cuda assert; different failures." + "reason": "Tokenizer method-name conflict and Qwen3 auto-device-map cuda assert are unrelated." }, { "left": "issue:43531", "right": "issue:44877", "accept": false, - "reason": "Qwen3-MoE sliding-window behavior vs strict-config loading rejection; not the same bug." + "reason": "Qwen3-MoE sliding_window behavior and strict config loading for granite_speech are different configuration bugs." }, { "left": "issue:44265", "right": "issue:44610", "accept": false, - "reason": "torch.export export failure via torch_compilable_check is unrelated to OmDet-Turbo processor size mismatch." + "reason": "torch.export/torch_compilable_check failure and OmDet-Turbo image size mismatch are unrelated export vs preprocessing issues." }, { "left": "issue:44008", "right": "issue:46899", "accept": false, - "reason": "Different multimodal models and causes: Gemma3n variable collision vs Gemma4 unsafe dtype cast under 4-bit quantization." - }, - { - "left": "issue:45923", - "right": "issue:47254", - "accept": false, - "reason": "Missing supports_gradient_checkpointing flag is unrelated to a PEFT+DeepSpeed checkpointing error." + "reason": "Gemma 3n variable-name collision and Gemma4 unsafe dtype cast are different multimodal model bugs." }, { "left": "issue:44479", "right": "issue:45325", "accept": false, - "reason": "Video input regression across Qwen variants vs still-image rope-index scaling bug; separate code paths." - }, - { - "left": "issue:44464", - "right": "issue:45987", - "accept": false, - "reason": "Chunked generation inconsistency under compiled forward is not the same as StaticCache get_seq_length returning the wrong type." + "reason": "Both touch Qwen VL models, but one is a video-input regression and the other is still-image rope indexing; not the same bug." }, { "left": "issue:43322", "right": "issue:43540", "accept": false, - "reason": "Llava Next loading segfault vs Qwen3OmniMoe video-input ValueError; distinct multimodal load/processing failures." + "reason": "Llava Next segfault and Qwen3OmniMoe video-input ValueError are distinct multimodal loading/processing failures." }, { - "left": "issue:43632", - "right": "issue:46821", + "left": "issue:45923", + "right": "issue:47254", "accept": false, - "reason": "Transformers v5 `_is_hf_initialized` breakage is unrelated to TimesFM window_size AttributeError." + "reason": "Missing supports_gradient_checkpointing on a remote-code model is unrelated to the PEFT/DeepSpeed checkpoint error." }, { - "left": "issue:43089", + "left": "issue:44464", "right": "issue:45987", "accept": false, - "reason": "Generation performance overhead and GPU syncs are unrelated to StaticCache length return-type bug." + "reason": "Chunked generation inconsistency with compiled forward and StaticCache.get_seq_length() returning the wrong type are separate generation/cache bugs." }, { "left": "issue:43425", "right": "issue:44403", "accept": false, - "reason": "Torch 2.10 incompatibility is not the same as noisy model-loading logs." - }, - { - "left": "issue:43296", - "right": "issue:43531", - "accept": false, - "reason": "PaddleOCR-VL loading failure vs Qwen3-MoE sliding-window issue; different models and root causes." - }, - { - "left": "issue:40444", - "right": "issue:45325", - "accept": false, - "reason": "IterableDataset multi-image finetuning failure is unrelated to Qwen2.5-VL rope-index scaling." - }, - { - "left": "issue:44220", - "right": "issue:44683", - "accept": false, - "reason": "_torch_extract_fbank_features issue is unrelated to compiled flex_attention failing on newer torch." + "reason": "Torch 2.10 incompatibility and loading noise are not the same underlying problem." }, { - "left": "issue:43872", - "right": "issue:44291", + "left": "issue:43632", + "right": "issue:46821", "accept": false, - "reason": "These are related in theme, but not the same concrete bug: bitsandbytes constructor incompatibility vs broader init_empty_weights/_is_hf_initialized loading failure." + "reason": "_is_hf_initialized flag breakage and TimesFM window_size AttributeError are unrelated v5 compatibility issues." } ] }, @@ -4042,7 +3956,7 @@ "source": "llm", "cache_hit": false, "model": "gpt-5.4-mini?service_tier=flex", - "cluster_id": "cluster-10000-529", + "cluster_id": "cluster-10000-530", "nodes": [ "issue:33666", "issue:35707", @@ -4051,8 +3965,8 @@ "issue:41720", "issue:42175", "issue:42796", + "issue:43089", "issue:43296", - "issue:43317", "issue:43381", "issue:43421", "issue:43482", @@ -4060,14 +3974,13 @@ "issue:43531", "issue:43611", "issue:43632", - "issue:43646", "issue:43716", "issue:43746", "issue:43792", - "issue:43856", + "issue:43872", "issue:43883", "issue:44079", - "issue:44265", + "issue:44220", "issue:44291", "issue:44403", "issue:44462", @@ -4075,18 +3988,19 @@ "issue:44534", "issue:44568", "issue:44589", + "issue:44683", "issue:44756", "issue:44811", "issue:44991", - "issue:45072", "issue:45081", "issue:45216", + "issue:45325", "issue:45357", - "issue:45538", "issue:45542", "issue:45663", "issue:45812", "issue:45923", + "issue:45987", "issue:46315", "issue:46491", "issue:46710", @@ -4094,15 +4008,20 @@ "issue:46899" ], "soft_pairs": [ + "issue:43089|issue:45987", + "issue:43296|issue:43531", + "issue:40444|issue:45325", + "issue:44220|issue:44683", + "issue:43872|issue:44291", "issue:43883|issue:44488", "issue:43611|issue:44534", "issue:44991|issue:46315", - "issue:43716|issue:46899", "issue:45216|issue:45357", + "issue:43716|issue:46899", "issue:44403|issue:44991", "issue:42796|issue:43746", - "issue:45812|issue:46491", "issue:44291|issue:44589", + "issue:45812|issue:46491", "issue:43381|issue:45923", "issue:43792|issue:44589", "issue:33666|issue:40444", @@ -4119,233 +4038,228 @@ "issue:43296|issue:46821", "issue:42175|issue:43296", "issue:42175|issue:45542", - "issue:44462|issue:45812", - "issue:44079|issue:44534", - "issue:43317|issue:43856", - "issue:43482|issue:45538", - "issue:43611|issue:43646", - "issue:44265|issue:45072" + "issue:44462|issue:45812" ], - "prepared_review_unit_hash": "12f771bc243cda319a5677c78428d09b526ff91c2fc9946151d92248e46c647f", + "prepared_review_unit_hash": "9eabcd0cfefb81fefa10e7a47cebd42e258a5f082d6a7ded1550a14bd5005ab9", "component_budget": { - "node_count": 529, - "item_count": 529, - "soft_pair_count": 1953, - "serialized_chars": 1012074, - "estimated_input_tokens": 253019, - "estimated_eval_tokens": 506294 + "node_count": 530, + "item_count": 530, + "soft_pair_count": 1954, + "serialized_chars": 1013004, + "estimated_input_tokens": 253251, + "estimated_eval_tokens": 506758 }, "budget": { "node_count": 48, "item_count": 48, "soft_pair_count": 31, - "serialized_chars": 37075, - "estimated_input_tokens": 9269, - "estimated_eval_tokens": 18794 + "serialized_chars": 37128, + "estimated_input_tokens": 9282, + "estimated_eval_tokens": 18820 }, "overflow_policy": "truncate_then_skip", "trimmed": false, "aggressively_trimmed": false, "split": true, "analyst_result": { - "summary": "This cluster is mostly heterogeneous: tokenizer-loading regressions, multimodal training issues, model init/load bugs, and backend/install problems. Only the Qwen3.5 save_pretrained pair looks like a clear duplicate family.", - "confidence": 0.87, - "canonical_issue_reason": "Issue 45216 is the clearest duplicate anchor because 45357 is a follow-up of the same Qwen3.5 save_pretrained regression; the rest of the set is unrelated.", + "summary": "Most items are unrelated issue reports from different models/code paths. I found only two soft-duplicate pairs that look like the same underlying bug: the transformers v5 `_is_hf_initialized`/bitsandbytes loading regression, and the Qwen3.5 `save_pretrained` checkpoint regression.", + "confidence": 0.84, + "canonical_issue_reason": "issue:43872 is the strongest representative: it has the clearest, broadly impactful regression title, high discussion/activity, and inbound references, and it matches a concrete v5 compatibility failure that also appears in a narrower follow-up issue.", "canonical_pr_reason": null, - "best_issue_reason": "Issue 45812 is the strongest standalone representative overall: it describes a broad silent v4\u2192v5 regression affecting all Granite models and has multiple inbound references.", + "best_issue_reason": "issue:43872 is the best overall issue to keep as the canonical anchor because it is concrete, high-signal, and central to one of the accepted duplicate pairs.", "best_pr_reason": null, "soft_edge_verdicts": [ { - "left": "issue:43883", - "right": "issue:44488", + "left": "issue:43089", + "right": "issue:45987", + "accept": false, + "reason": "Different problems: GPU sync/dispatch overhead vs a StaticCache return-type contract bug." + }, + { + "left": "issue:43296", + "right": "issue:43531", + "accept": false, + "reason": "Different models and failures: PaddleOCR-VL load failure vs Qwen3-MoE sliding_window behavior." + }, + { + "left": "issue:40444", + "right": "issue:45325", + "accept": false, + "reason": "Both are Qwen2.5-VL related, but one is multi-image finetuning input handling and the other is ROPE/position_id scaling." + }, + { + "left": "issue:44220", + "right": "issue:44683", + "accept": false, + "reason": "Unrelated subsystems: fbank feature extraction vs compiled flex_attention on newer torch." + }, + { + "left": "issue:43872", + "right": "issue:44291", + "accept": true, + "reason": "Same underlying transformers v5 loading regression around `_is_hf_initialized`, producing the same bitsandbytes TypeError path." + }, + { + "left": "issue:43883", + "right": "issue:44488", "accept": false, - "reason": "Different failures: missing all_tied_weights_keys vs model load failure; no clear same code path." + "reason": "Different model-loading failures with different missing/ignored tokenizer/model assets." }, { "left": "issue:43611", "right": "issue:44534", "accept": false, - "reason": "One is a model-loading regression, the other is non-persistent buffer initialization; unrelated bugs." + "reason": "Both are v5 regressions, but they affect different loader internals and different symptoms." }, { "left": "issue:44991", "right": "issue:46315", "accept": false, - "reason": "Tokenizer load failure vs convert_tokens_to_ids performance regression; different symptoms and likely different fixes." - }, - { - "left": "issue:43716", - "right": "issue:46899", - "accept": false, - "reason": "Both are dtype-related multimodal issues, but on different models and with different root causes." + "reason": "Tokenizer loading failure vs a performance regression in `convert_tokens_to_ids`; different bug classes." }, { "left": "issue:45216", "right": "issue:45357", "accept": true, - "reason": "Same Qwen3.5 save_pretrained regression; the second issue is a residual manifestation involving visual encoder keys." + "reason": "Same Qwen3.5 `save_pretrained` regression; 45357 is a narrower follow-up about incorrect visual encoder keys." + }, + { + "left": "issue:43716", + "right": "issue:46899", + "accept": false, + "reason": "Both are dtype-related multimodal issues, but they affect different models, modalities, and code paths." }, { "left": "issue:44403", "right": "issue:44991", "accept": false, - "reason": "Noise on loading vs tokenizer loading failure; not the same underlying bug." + "reason": "Noise/logging during load is not the same as tokenizer repository resolution failure." }, { "left": "issue:42796", "right": "issue:43746", "accept": false, - "reason": "Tokenizer BPE marker handling and PEFT/local-checkpoint loading are unrelated." + "reason": "Tokenizer BPE marker handling is unrelated to PEFT checkpoint loading." }, { - "left": "issue:45812", - "right": "issue:46491", + "left": "issue:44291", + "right": "issue:44589", "accept": false, - "reason": "Granite AutoTokenizer regression and CodeLlama leading-space round-trip bug are different tokenizer issues." + "reason": "Different regressions: unexpected `_is_hf_initialized` kwarg vs missing Float8 storage type." }, { - "left": "issue:44291", - "right": "issue:44589", + "left": "issue:45812", + "right": "issue:46491", "accept": false, - "reason": "init_empty_weights/_is_hf_initialized vs Float8 storage lookup; different breakages." + "reason": "Different tokenizer regressions affecting different models and different behaviors." }, { "left": "issue:43381", "right": "issue:45923", "accept": false, - "reason": "Eval-mode gradient checkpointing constraint vs missing supports_gradient_checkpointing flag; separate issues." + "reason": "Eval-mode gradient checkpointing restriction vs a missing `supports_gradient_checkpointing` flag on a trust_remote_code model." }, { "left": "issue:43792", "right": "issue:44589", "accept": false, - "reason": "Whisper run failure and Float8 storage lookup error are unrelated." + "reason": "Whisper runtime failure is not the same as a Float8 storage lookup error." }, { "left": "issue:33666", "right": "issue:40444", "accept": false, - "reason": "Both are Qwen2-VL training-related, but one is multi-GPU training and the other is iterable-dataset/multi-image finetuning; not the same bug." + "reason": "Both involve training, but they are distinct Qwen-VL issues with different failure modes." }, { "left": "issue:40990", "right": "issue:44811", "accept": false, - "reason": "Perplexity regression on gpt-oss vs Whisper batch_decode special-token handling; unrelated." + "reason": "Perplexity on a language model dataset is unrelated to Whisper `batch_decode` special-token handling." }, { "left": "issue:43632", "right": "issue:44756", "accept": false, - "reason": "_is_hf_initialized flag regression vs mmap/OOM workaround; no shared fix path." + "reason": "`_is_hf_initialized` flag regression and Strix Halo mmap OOM are unrelated." }, { "left": "issue:42796", "right": "issue:46710", "accept": false, - "reason": "Tokenizer BPE marker bug and DeepSeek accuracy regression are unrelated." + "reason": "Tokenizer marker handling vs accuracy regression after version upgrade; different symptoms and likely different causes." }, { "left": "issue:43531", "right": "issue:44291", "accept": false, - "reason": "Qwen3-MoE sliding_window issue vs init_empty_weights argument mismatch; different code paths." + "reason": "Qwen3-MoE sliding_window behavior is unrelated to the `_is_hf_initialized` loading error." }, { "left": "issue:35707", "right": "issue:40444", "accept": false, - "reason": "Progressive generation with inputs_embeds/past_key_values is unrelated to Qwen2.5-VL multi-image finetuning." + "reason": "Progressive generation with cache inputs is unrelated to Qwen2.5-VL multi-image finetuning failure." }, { "left": "issue:44079", "right": "issue:45663", "accept": false, - "reason": "ModelOutput key assignment bug vs Gemma4 FSDP2 shared_kv_states issue; unrelated." + "reason": "ModelOutput key assignment bug vs Gemma4 FSDP2 shared_kv_states issue." }, { "left": "issue:43883", "right": "issue:44991", "accept": false, - "reason": "Model attribute error vs tokenizer loading regression; not the same bug." + "reason": "Missing model attribute error is unrelated to tokenizer repository loading." }, { "left": "issue:41720", "right": "issue:45081", "accept": false, - "reason": "Qwen3 auto device map cuda assert and Mistral regex patch crash are different failures." + "reason": "Auto device mapping CUDA assert is unrelated to the Mistral regex patch crash." }, { "left": "issue:43482", "right": "issue:43525", "accept": false, - "reason": "GGUF loading failure and missing pad_token_id on Llama4Config are separate model-load problems." + "reason": "Qwen2.5-GGUF loading and Llama4Config missing `pad_token_id` are different model/config issues." }, { "left": "issue:43421", "right": "issue:44568", "accept": false, - "reason": "Runtime special-token post-processor update bug vs add_special_tokens not adding BOS/EOS for one tokenizer; related area but not clearly the same underlying issue." + "reason": "Runtime special-token post-processing update bug is different from BOS/EOS not being added on encode." }, { "left": "issue:43296", "right": "issue:46821", "accept": false, - "reason": "PaddleOCR-VL load failure and TimesFM window_size AttributeError are unrelated." + "reason": "PaddleOCR-VL load failure vs `window_size` AttributeError in TimesFM 2.5 are unrelated." }, { "left": "issue:42175", "right": "issue:43296", "accept": false, - "reason": "TensorFlow backend packaging issue vs model load failure; different problems." + "reason": "Packaging/backend dependency issue is unrelated to model loading failure in vLLM." }, { "left": "issue:42175", "right": "issue:45542", "accept": false, - "reason": "Install/backend packaging issue vs undefined tf backend error; not the same concrete bug." + "reason": "Both mention TensorFlow backend, but one is an install/extra dependency issue and the other is a runtime backend detection error." }, { "left": "issue:44462", "right": "issue:45812", "accept": false, - "reason": "Tokenizer.json lookup bug and Granite wrong-token-IDs regression may be related, but the titles do not show the same concrete failure path strongly enough." - }, - { - "left": "issue:44079", - "right": "issue:44534", - "accept": false, - "reason": "ModelOutput key assignment vs non-persistent buffer junk; unrelated." - }, - { - "left": "issue:43317", - "right": "issue:43856", - "accept": false, - "reason": "device_map/offload loading bug vs Qwen3 MoE training memory inefficiency; different issues." - }, - { - "left": "issue:43482", - "right": "issue:45538", - "accept": false, - "reason": "GGUF loading failure vs CLIPTokenizer model_max_length default; unrelated." - }, - { - "left": "issue:43611", - "right": "issue:43646", - "accept": false, - "reason": "Both are Transformer v5 regressions, but one is base_model_prefix loading and the other is custom model initialization; not enough evidence they are the same bug." - }, - { - "left": "issue:44265", - "right": "issue:45072", - "accept": false, - "reason": "torch.export/exportability bug vs bfloat16 dtype mismatch CI issue; different code paths and fixes." + "reason": "Repository tokenizer.json being ignored is a different tokenizer loading bug than Granite token ID regression." } ] }, "evaluator_result": { "accept": true, - "feedback": "The summary is grounded in the packet and stays conservative. The chosen representatives are well supported by titles and inbound references, and the only accepted soft-edge pair (45216/45357) looks like a plausible same-bug follow-up. The other pair rejections are cautious and do not overclaim equivalence." + "feedback": "The summary is grounded in the packet and stays conservative. The two accepted soft-duplicate pairs are plausibly the same underlying bug based on the titles, and the rest of the verdicts read as careful non-duplicates rather than overclaims. Canonical issue choice is also reasonable from the available metadata." }, "evaluator_used": true, "retried": false, @@ -4364,30 +4278,29 @@ "source": "llm", "cache_hit": false, "model": "gpt-5.4-mini?service_tier=flex", - "cluster_id": "cluster-10000-529", + "cluster_id": "cluster-10000-530", "nodes": [ "issue:34634", "issue:35707", "issue:40444", "issue:40990", "issue:41720", - "issue:42617", "issue:42796", - "issue:42915", - "issue:42947", - "issue:43257", "issue:43262", + "issue:43317", "issue:43377", "issue:43454", + "issue:43482", "issue:43493", - "issue:43550", "issue:43606", + "issue:43611", "issue:43644", + "issue:43646", "issue:43701", "issue:43746", "issue:43818", - "issue:43819", "issue:43828", + "issue:43856", "issue:43873", "issue:43994", "issue:44079", @@ -4397,36 +4310,42 @@ "issue:44466", "issue:44479", "issue:44509", + "issue:44534", "issue:44829", "issue:44877", + "issue:45072", "issue:45325", "issue:45357", "issue:45381", "issue:45479", "issue:45491", + "issue:45538", "issue:45584", "issue:45701", "issue:45823", "issue:45834", - "issue:45923", - "issue:46209", "issue:46402", "issue:46821", "issue:46909", "issue:47217" ], "soft_pairs": [ + "issue:43317|issue:43856", + "issue:44079|issue:44534", + "issue:43482|issue:45538", + "issue:43611|issue:43646", "issue:44509|issue:46909", + "issue:44265|issue:45072", "issue:45325|issue:45834", "issue:43377|issue:45491", - "issue:43644|issue:44079", "issue:40444|issue:45381", + "issue:43644|issue:44079", "issue:44291|issue:44877", "issue:45325|issue:45381", "issue:43493|issue:43994", "issue:43818|issue:46402", - "issue:43454|issue:44466", "issue:40444|issue:47217", + "issue:43454|issue:44466", "issue:43818|issue:45357", "issue:43873|issue:45701", "issue:43701|issue:45325", @@ -4437,248 +4356,226 @@ "issue:43828|issue:46821", "issue:40990|issue:45584", "issue:43262|issue:43746", - "issue:35707|issue:41720", "issue:34634|issue:35707", + "issue:35707|issue:41720", "issue:41720|issue:45823", - "issue:44186|issue:44265", - "issue:43606|issue:43819", - "issue:43454|issue:43550", - "issue:42617|issue:42915", - "issue:45823|issue:46821", - "issue:43257|issue:45923", - "issue:40990|issue:42947", - "issue:42796|issue:46209" + "issue:44186|issue:44265" ], - "prepared_review_unit_hash": "9fc24cbb4fa861a59553cc5094cfbab3e5696ae365ac93a9945c3e98a5c8d878", + "prepared_review_unit_hash": "7ded0d42bc09e44c2b70be4028d3e37df1103dbf1837f7942ebdcc57c66bc826", "component_budget": { - "node_count": 529, - "item_count": 529, - "soft_pair_count": 1953, - "serialized_chars": 1012074, - "estimated_input_tokens": 253019, - "estimated_eval_tokens": 506294 + "node_count": 530, + "item_count": 530, + "soft_pair_count": 1954, + "serialized_chars": 1013004, + "estimated_input_tokens": 253251, + "estimated_eval_tokens": 506758 }, "budget": { "node_count": 48, "item_count": 48, - "soft_pair_count": 32, - "serialized_chars": 37507, - "estimated_input_tokens": 9377, - "estimated_eval_tokens": 19010 + "soft_pair_count": 30, + "serialized_chars": 36885, + "estimated_input_tokens": 9222, + "estimated_eval_tokens": 18700 }, "overflow_policy": "truncate_then_skip", "trimmed": false, "aggressively_trimmed": false, "split": true, "analyst_result": { - "summary": "This cluster is mostly a grab-bag of unrelated bugs across different models and subsystems. The only high-confidence duplicate pair is the SigLIP2 regression; the rest are distinct issues and should stay separate.", - "confidence": 0.79, - "canonical_issue_reason": "issue:43493 is the clearest root-cause report in the only strong duplicate pair: it directly compares Hugging Face SigLIP2 behavior against the original JAX implementation and frames the bug precisely.", + "summary": "This set is mostly a loose collection of Transformers regressions and model-specific bugs; none of the soft pairs is close enough to merge from titles alone. The most representative active issue is the broad offloading/quantization tracker.", + "confidence": 0.61, + "canonical_issue_reason": "issue:43873 is the broadest active issue here and best matches the cluster\u2019s recurring offloading/device-map symptoms; it has the most tracker-like shape and inbound references.", "canonical_pr_reason": null, - "best_issue_reason": "issue:43493 is the best representative issue overall because it is the most specific and technically diagnostic report among the accepted duplicates.", + "best_issue_reason": "issue:43873 is the best global representative because it is open, actively discussed, and broad enough to serve as the landing issue for related offloading/quantization reports.", "best_pr_reason": null, "soft_edge_verdicts": [ + { + "left": "issue:43317", + "right": "issue:43856", + "accept": false, + "reason": "Different failures: CPU/GPU offload loading error versus MoE training memory inefficiency. Same general area, not the same bug." + }, + { + "left": "issue:44079", + "right": "issue:44534", + "accept": false, + "reason": "ModelOutput key assignment and non-persistent buffer initialization are different code paths and symptoms." + }, + { + "left": "issue:43482", + "right": "issue:45538", + "accept": false, + "reason": "GGUF loading failure with v5 is unrelated to CLIPTokenizer max-length handling." + }, + { + "left": "issue:43611", + "right": "issue:43646", + "accept": false, + "reason": "Both are v5 regressions, but one is about base_model_prefix loading and the other about custom model initialization; not the same underlying change." + }, { "left": "issue:44509", "right": "issue:46909", "accept": false, - "reason": "Both are documentation-related, but they target different docs problems: removed pipeline tasks vs a broken DIA example." + "reason": "Both are docs issues, but they concern different documentation problems and examples." + }, + { + "left": "issue:44265", + "right": "issue:45072", + "accept": false, + "reason": "torch.export failure with torch_compilable_check is a different bug from dtype mismatches in bfloat16 inference." }, { "left": "issue:45325", "right": "issue:45834", "accept": false, - "reason": "Different models and failures: Qwen2.5-VL rope/index handling vs Kosmos2.5 OCR long-input index errors." + "reason": "Qwen2.5-VL rope-index scaling and Kosmos2.5 long-OCR index errors are different model-specific failures." }, { "left": "issue:43377", "right": "issue:45491", "accept": false, - "reason": "Different models and code paths: MIMI padding-mask batching bug vs Gemma3 mixed-length sliding-window NaNs." + "reason": "Missing padding-mask support in MIMI is unrelated to Gemma3 sliding-window NaN embeddings." }, { - "left": "issue:43644", - "right": "issue:44079", + "left": "issue:40444", + "right": "issue:45381", "accept": false, - "reason": "Unrelated bugs: junk-filled non-persistent buffers vs ModelOutput key assignment when a key was previously None." + "reason": "Both involve Qwen VL inputs, but one is multiple images in an iterable dataset and the other is video vision_position_ids; not the same defect." }, { - "left": "issue:40444", - "right": "issue:45381", + "left": "issue:43644", + "right": "issue:44079", "accept": false, - "reason": "Same model family, but different symptoms and input modes: multiple images per prompt vs wrong video position ids." + "reason": "Non-persistent buffer serialization junk and ModelOutput key assignment are separate internal issues." }, { "left": "issue:44291", "right": "issue:44877", "accept": false, - "reason": "Different loading/config issues: unexpected _is_hf_initialized in init_empty_weights vs strict config blocking granite_speech." + "reason": "init_empty_weights argument handling and strict config loading are different regressions." }, { "left": "issue:45325", "right": "issue:45381", "accept": false, - "reason": "Both mention Qwen2.5-VL position ids, but one is still-image temporal scaling and the other is video input position ids; not enough to treat as the same bug." + "reason": "Both are Qwen2.5-VL position-id bugs, but one is still-image temporal scaling and the other is video input positions." }, { "left": "issue:43493", "right": "issue:43994", - "accept": true, - "reason": "Both point to the same SigLIP2 implementation correctness problem: HF outputs are wrong compared with expected behavior and produce nonsensical inference results." + "accept": false, + "reason": "Same model family, but the issues describe different symptoms and no shared concrete code-path evidence." }, { "left": "issue:43818", "right": "issue:46402", "accept": false, - "reason": "Different models and failures: Video-LLaVA temporal-attention/weight-sharing issue vs Sam3 text-encoder weights not loading." + "reason": "Video-LLaVA tower/weight structure issue is unrelated to Sam3Model text encoder loading." }, { - "left": "issue:43454", - "right": "issue:44466", + "left": "issue:40444", + "right": "issue:47217", "accept": false, - "reason": "Both involve lm_head/tied weights, but one is a generation bug and the other is a serialization/device-dependent save bug." + "reason": "Both are VL input-handling problems, but one is multi-image finetuning and the other is placeholder placement in user text." }, { - "left": "issue:40444", - "right": "issue:47217", + "left": "issue:43454", + "right": "issue:44466", "accept": false, - "reason": "Both are VL-input issues, but the failures are different: multi-image prompt handling vs image placeholder placement in user text." + "reason": "AyaVision weight-tying failure and inconsistent lm_head serialization are different tie/serialization bugs." }, { "left": "issue:43818", "right": "issue:45357", "accept": false, - "reason": "Different bugs: missing temporal attention / shared weights in Video-LLaVA vs incorrect visual encoder keys on save_pretrained." + "reason": "Video-LLaVA architecture mismatch and Qwen3.5 visual-encoder key saving are unrelated." }, { "left": "issue:43873", "right": "issue:45701", "accept": false, - "reason": "Unrelated: quantization/offloading behavior vs tokenization changes across versions." + "reason": "Quantized offloading behavior and tokenization-version changes are not the same bug." }, { "left": "issue:43701", "right": "issue:45325", "accept": false, - "reason": "Checkpoint resume key mismatch is unrelated to Qwen2.5-VL position-id computation." + "reason": "Checkpoint key mismatch and Qwen2.5-VL rope-index scaling are unrelated." }, { "left": "issue:44829", "right": "issue:45479", "accept": false, - "reason": "Both are training degeneration reports, but one is flash_attention_3-related loss collapse and the other is the num_labels=1/single_label_classification zero-loss issue." + "reason": "FlashAttention-3 training degeneration and single-label classification zero-loss are different training-path bugs." }, { "left": "issue:43606", "right": "issue:44265", "accept": false, - "reason": "Different failure modes: CPU offload device mismatch vs torch.export failing on torch_compilable_check." + "reason": "CPU offload device mismatch and torch.export failures are different failure modes." }, { "left": "issue:42796", "right": "issue:45381", "accept": false, - "reason": "Tokenizer BPE-marker handling is unrelated to Qwen video vision-position-id bugs." + "reason": "Tokenizer BPE marker handling is unrelated to Qwen video vision_position_ids." }, { "left": "issue:44479", "right": "issue:45823", "accept": false, - "reason": "Different models and problems: Qwen video-input regression vs Gemma4 PLE device mismatch." + "reason": "Qwen video-input regression and Gemma4 PLE device mismatch are distinct model-specific issues." }, { "left": "issue:43828", "right": "issue:46821", "accept": false, - "reason": "Different runtime errors: autocast dtype mismatch vs TimesFM window_size AttributeError." + "reason": "Autocast dtype mismatch and TimesFM window_size AttributeError do not share a concrete bug." }, { "left": "issue:40990", "right": "issue:45584", "accept": false, - "reason": "Different symptoms and models: high perplexity on gpt-oss-20b vs Whisper empty-transcription generation failure." + "reason": "High perplexity on gpt-oss-20b and Whisper empty-transcription failure are unrelated." }, { "left": "issue:43262", "right": "issue:43746", "accept": false, - "reason": "Audio chat-template sampling-rate default vs GraniteSpeech PEFT checkpoint loading are separate issues." + "reason": "Audio chat-template sampling-rate defaulting and PEFT checkpoint loading are different concerns." }, { - "left": "issue:35707", - "right": "issue:41720", + "left": "issue:34634", + "right": "issue:35707", "accept": false, - "reason": "Progressive generation with inputs_embeds/past_key_values is unrelated to Qwen3 auto device-map CUDA asserts." + "reason": "Bark voice_preset handling and progressive generation with inputs_embeds/past_key_values are unrelated." }, { - "left": "issue:34634", - "right": "issue:35707", + "left": "issue:35707", + "right": "issue:41720", "accept": false, - "reason": "BarkProcessor voice_preset and progressive generation are unrelated bugs." + "reason": "Different code paths: generation with inputs_embeds/past_key_values versus auto device mapping cuda asserts." }, { "left": "issue:41720", "right": "issue:45823", "accept": false, - "reason": "Different models and failures: Qwen3 CUDA assert under auto device map vs Gemma4 PLE forward device mismatch." + "reason": "Both mention auto device mapping, but the model-specific device failures and symptoms are different." }, { "left": "issue:44186", "right": "issue:44265", "accept": false, - "reason": "Tokenizer crashes are unrelated to torch.export failures in compiler-check code paths." - }, - { - "left": "issue:43606", - "right": "issue:43819", - "accept": false, - "reason": "CPU offload device mismatch and DAC.from_latents/STE mismatch are different bugs." - }, - { - "left": "issue:43454", - "right": "issue:43550", - "accept": false, - "reason": "Different model failures: tied-weight generation/serialization issue vs torch.compile SDPA failure." - }, - { - "left": "issue:42617", - "right": "issue:42915", - "accept": false, - "reason": "3d_parallel.py execution failure is unrelated to Qwen3MoE FP8Config loading/execution." - }, - { - "left": "issue:45823", - "right": "issue:46821", - "accept": false, - "reason": "Gemma4 device mismatch and TimesFM window_size AttributeError are unrelated." - }, - { - "left": "issue:43257", - "right": "issue:45923", - "accept": false, - "reason": "Qwen3 MoE weight conversion under accelerate+deepspeed is unrelated to missing supports_gradient_checkpointing on Nemotron remote-code variant." - }, - { - "left": "issue:40990", - "right": "issue:42947", - "accept": false, - "reason": "High perplexity on gpt-oss-20b is unrelated to gradient checkpointing being ineffective with PEFT LoRA." - }, - { - "left": "issue:42796", - "right": "issue:46209", - "accept": false, - "reason": "Tokenizer BPE-marker handling is unrelated to FP8 inference failure on B200." + "reason": "Tokenizer crashes on NER/padding and torch.export failures with torch_compilable_check are different bugs." } ] }, - "evaluator_result": { - "accept": true, - "feedback": "The summary is well grounded in the packet and stays conservative. The only accepted duplicate pair, issue:43493 with issue:43994, is supported by the titles and the stated SigLIP2 discrepancy. The remaining soft pairs are rejected with cautious, model- or bug-specific reasoning rather than overclaiming duplication. No mergeability concerns apply here because this is an issue-clustering judgment, not a PR-pair merge decision." - }, - "evaluator_used": true, + "evaluator_result": null, + "evaluator_used": false, "retried": false, - "accepted_nontrivial_soft_edge": true, + "accepted_nontrivial_soft_edge": false, "error_kind": null, "error_message": null }, @@ -4693,7 +4590,7 @@ "source": "llm", "cache_hit": false, "model": "gpt-5.4-mini?service_tier=flex", - "cluster_id": "cluster-10000-529", + "cluster_id": "cluster-10000-530", "nodes": [ "issue:26413", "issue:28530", @@ -4702,22 +4599,24 @@ "issue:41553", "issue:41720", "issue:42617", + "issue:42796", "issue:42915", - "issue:43065", - "issue:43295", + "issue:42947", + "issue:43257", "issue:43317", "issue:43322", "issue:43377", "issue:43450", "issue:43454", + "issue:43550", "issue:43572", "issue:43582", + "issue:43606", "issue:43720", "issue:43749", + "issue:43819", "issue:43873", - "issue:44186", "issue:44387", - "issue:44451", "issue:44466", "issue:44534", "issue:44568", @@ -4733,249 +4632,262 @@ "issue:45701", "issue:45704", "issue:45759", + "issue:45823", "issue:45920", + "issue:45923", "issue:46050", + "issue:46209", "issue:46305", "issue:46315", - "issue:46402", - "issue:46821", - "issue:46897", - "issue:47020", - "issue:47317" + "issue:46821" ], "soft_pairs": [ + "issue:43606|issue:43819", + "issue:43454|issue:43550", + "issue:42617|issue:42915", + "issue:45823|issue:46821", + "issue:43257|issue:45923", + "issue:42796|issue:46209", + "issue:40990|issue:42947", "issue:43454|issue:45072", "issue:44568|issue:45920", "issue:43377|issue:44610", "issue:36331|issue:42617", "issue:41553|issue:43582", - "issue:43317|issue:43873", - "issue:43454|issue:43720", "issue:41720|issue:43322", + "issue:43317|issue:43873", "issue:43749|issue:44534", "issue:36331|issue:42915", + "issue:43454|issue:43720", "issue:43572|issue:45759", "issue:44877|issue:44991", "issue:43377|issue:44898", "issue:46305|issue:46315", "issue:41720|issue:44977", - "issue:44387|issue:45704", "issue:41720|issue:42617", - "issue:40990|issue:45701", "issue:43322|issue:46050", + "issue:44387|issue:45704", + "issue:40990|issue:45701", "issue:43450|issue:45701", "issue:43873|issue:44756", "issue:44466|issue:45357", "issue:45685|issue:46821", - "issue:26413|issue:28530", - "issue:44451|issue:47020", - "issue:44186|issue:45072", - "issue:43295|issue:45701", - "issue:46897|issue:47317", - "issue:43065|issue:46402" + "issue:26413|issue:28530" ], - "prepared_review_unit_hash": "0b93f21e3eeda7af69a0f2b9a31f86f208fa4aefd78de95e973852e695cf2440", + "prepared_review_unit_hash": "f1b7766010b9d430701405e7ff173f81b2ee6ea7806992c7ad54ce503ccb2e78", "component_budget": { - "node_count": 529, - "item_count": 529, - "soft_pair_count": 1953, - "serialized_chars": 1012074, - "estimated_input_tokens": 253019, - "estimated_eval_tokens": 506294 + "node_count": 530, + "item_count": 530, + "soft_pair_count": 1954, + "serialized_chars": 1013004, + "estimated_input_tokens": 253251, + "estimated_eval_tokens": 506758 }, "budget": { - "node_count": 47, - "item_count": 47, - "soft_pair_count": 29, - "serialized_chars": 35756, - "estimated_input_tokens": 8939, - "estimated_eval_tokens": 18134 + "node_count": 48, + "item_count": 48, + "soft_pair_count": 31, + "serialized_chars": 36996, + "estimated_input_tokens": 9249, + "estimated_eval_tokens": 18754 }, "overflow_policy": "truncate_then_skip", "trimmed": false, "aggressively_trimmed": false, "split": true, "analyst_result": { - "summary": "This cluster is heterogeneous: the soft-similarity pairs share broad themes like loading, tokenization, or quantization, but they do not describe the same underlying bug/change. No PRs are present.", - "confidence": 0.97, - "canonical_issue_reason": "No true canonical issue: the items span unrelated training, loading, tokenizer, multimodal, and quantization bugs.", + "summary": "This set looks like a noisy grab-bag of largely unrelated Transformers issues across tokenization, quantization/offloading, configs, model-specific regressions, and training utilities. I don't see any pair that is clearly the same underlying bug or change, so I reject all soft edges. If forced to pick one representative issue, #43873 is the broadest umbrella around offloading/quantization failures.", + "confidence": 0.94, + "canonical_issue_reason": "No true duplicate cluster emerges from the titles alone; #43873 is the broadest issue and the closest umbrella for nearby offloading/quantization failures.", "canonical_pr_reason": null, - "best_issue_reason": "issue:45920 is the broadest single representative item here (multi-model tokenizer regression), but it still does not unify the cluster into one duplicate set.", + "best_issue_reason": "#43873 is the most general representative of the set, but only weakly so; the rest are mostly distinct model- or subsystem-specific bugs.", "best_pr_reason": null, "soft_edge_verdicts": [ { - "left": "issue:43454", - "right": "issue:45072", + "left": "issue:43606", + "right": "issue:43819", "accept": false, - "reason": "Different bugs: lm_head weight tying vs bfloat16 dtype mismatch in unrelated models/code paths." + "reason": "Different bugs: CPU offload device mismatch for Bark vs a DAC latent/STE mismatch. No shared code-path is evident." }, { - "left": "issue:44568", - "right": "issue:45920", + "left": "issue:43454", + "right": "issue:43550", "accept": false, - "reason": "Both are tokenizer-related, but one is missing add_special_tokens behavior while the other is wrong token IDs across multiple models." + "reason": "AyaVision lm_head tying bug and Bamba torch.compile/SDPA failure are unrelated." }, { - "left": "issue:43377", - "right": "issue:44610", + "left": "issue:42617", + "right": "issue:42915", "accept": false, - "reason": "Different domains and failures: MIMI batched padding-mask behavior vs OmDet-Turbo image input shape mismatch." + "reason": "3d_parallel.py runtime failure is unrelated to Qwen3Moe fine-grained FP8 loading." }, { - "left": "issue:36331", - "right": "issue:42617", + "left": "issue:45823", + "right": "issue:46821", "accept": false, - "reason": "Unrelated issues: Trainer compute_loss API break vs failure to run a 3D parallel script." + "reason": "Gemma4 device-mapping mismatch and TimesFM window_size AttributeError are different subsystems." }, { - "left": "issue:41553", - "right": "issue:43582", + "left": "issue:43257", + "right": "issue:45923", "accept": false, - "reason": "Different subsystems: AutoTokenizer error messaging for Voxtral vs Apple Silicon allocator TypeError." + "reason": "Qwen3 MoE weight conversion under accelerate+deepspeed is unrelated to a missing gradient-checkpointing flag." }, { - "left": "issue:43317", - "right": "issue:43873", + "left": "issue:42796", + "right": "issue:46209", + "accept": false, + "reason": "Tokenizer BPE-marker handling and FP8 inference on B200 are not the same bug." + }, + { + "left": "issue:40990", + "right": "issue:42947", "accept": false, - "reason": "Both involve offloading/quantization, but they are distinct failures with different trigger conditions and code paths." + "reason": "High perplexity on a model/dataset pair is unrelated to LoRA gradient checkpointing ineffectiveness." }, { "left": "issue:43454", - "right": "issue:43720", + "right": "issue:45072", + "accept": false, + "reason": "Both mention model failures, but one is weight tying and the other is a dtype mismatch; different code paths." + }, + { + "left": "issue:44568", + "right": "issue:45920", + "accept": false, + "reason": "Both are tokenizer regressions, but one is missing special tokens and the other is wrong token IDs across several models." + }, + { + "left": "issue:43377", + "right": "issue:44610", + "accept": false, + "reason": "Batching-related issues, but one is MIMI padding-mask behavior and the other is a video processor shape bug." + }, + { + "left": "issue:36331", + "right": "issue:42617", + "accept": false, + "reason": "CustomTrainer API change and 3d_parallel.py failure are unrelated." + }, + { + "left": "issue:41553", + "right": "issue:43582", "accept": false, - "reason": "Different model-loading bugs: weight tying in AyaVision vs unpacking packed BitNet weights during accelerate loading." + "reason": "Tokenizer loading error for Voxtral and AppleSilicon allocator TypeError share no concrete bug." }, { "left": "issue:41720", "right": "issue:43322", "accept": false, - "reason": "Both are load-time GPU failures, but they affect different models and manifest as different errors (cudaErrorAssert vs segfault)." + "reason": "Auto device-map assert on Qwen3 and Llava Next segfault are different failures with no common concrete cause shown." + }, + { + "left": "issue:43317", + "right": "issue:43873", + "accept": false, + "reason": "These are both offload/quantization-adjacent, but one is a specific dequantized-model loading failure and the other is a broad offloading complaint; not enough evidence they are the same bug." }, { "left": "issue:43749", "right": "issue:44534", "accept": false, - "reason": "Both touch loading/state handling, but one is broken FSDP CPU RAM efficient loading and the other is a non-persistent buffer serialization regression." + "reason": "FSDP CPU RAM efficient loading and non-persistent buffer junk are distinct regressions." }, { "left": "issue:36331", "right": "issue:42915", "accept": false, - "reason": "Unrelated: custom Trainer loss signature break vs Qwen3Moe FineGrainedFP8Config failure." + "reason": "Trainer signature change and Qwen3Moe FP8 loading failure are unrelated." + }, + { + "left": "issue:43454", + "right": "issue:43720", + "accept": false, + "reason": "AyaVision weight tying and BitNet packed-weight unpacking are different model-specific issues." }, { "left": "issue:43572", "right": "issue:45759", "accept": false, - "reason": "Different config-loading problems: missing pad_token_idx vs failure to unwrap text_config in composite multimodal configs." + "reason": "Missing pad_token_idx in StableLmConfig and text_config unwrapping for composite Qwen configs are separate config-loading problems." }, { "left": "issue:44877", "right": "issue:44991", "accept": false, - "reason": "Both are load failures, but one is strict config rejection and the other is tokenizer-loading failure for a different model family." + "reason": "Strict config validation for granite_speech and tokenizer loading failure for est-roberta are unrelated." }, { "left": "issue:43377", "right": "issue:44898", "accept": false, - "reason": "Different issues in different model classes: MIMI batching/padding-mask bug vs Perceiver non-default resolution handling." + "reason": "Different models and symptoms: MIMI batching/padding vs Perceiver positional interpolation." }, { "left": "issue:46305", "right": "issue:46315", "accept": false, - "reason": "One is a streaming mask off-by-one correctness bug; the other is a tokenizer performance regression." + "reason": "MoonshineStreaming sliding-window off-by-one and tokenizer conversion slowdown are unrelated." }, { "left": "issue:41720", "right": "issue:44977", "accept": false, - "reason": "Same model family, but not the same bug: device-mapping load assert vs flash-attention generation failure." + "reason": "Both involve Qwen3-family models, but one is device-map CUDA assert and the other is flash-attention generation quality; not the same concrete bug." }, { - "left": "issue:44387", - "right": "issue:45704", + "left": "issue:41720", + "right": "issue:42617", "accept": false, - "reason": "Different memory issues in different code paths: int4 CUDA reserved memory/OOM vs T5 apex RMSNorm leak." + "reason": "Qwen3 auto device mapping and 3d_parallel.py runtime failure do not share a concrete code-path." }, { - "left": "issue:41720", - "right": "issue:42617", + "left": "issue:43322", + "right": "issue:46050", "accept": false, - "reason": "Unrelated: Qwen3 auto device-map failure vs a 3d_parallel.py runtime problem." + "reason": "Llava Next segfault and sinq save/load breakage are different serialization/loading failures." }, { - "left": "issue:40990", - "right": "issue:45701", + "left": "issue:44387", + "right": "issue:45704", "accept": false, - "reason": "Different symptoms and likely causes: high perplexity on a specific eval vs tokenizer behavior changing across versions." + "reason": "Int4 quantization OOM and T5 apex memory leak are different memory issues in different implementations." }, { - "left": "issue:43322", - "right": "issue:46050", + "left": "issue:40990", + "right": "issue:45701", "accept": false, - "reason": "Different bugs: Llava Next load-time segfault vs sinq quantization save/load corruption." + "reason": "Perplexity regression on one model and tokenization changes across versions are not the same bug." }, { "left": "issue:43450", "right": "issue:45701", "accept": false, - "reason": "Video processor batching/shape bug is unrelated to tokenizer-version behavior changes." + "reason": "Video processor batch-shape bug and version-dependent tokenization changes are unrelated." }, { "left": "issue:43873", "right": "issue:44756", "accept": false, - "reason": "Both concern memory/offload behavior, but one is quantization offloading and the other is mmap OOM on a specific platform." + "reason": "Quantization/offloading issue and mmap OOM on Strix Halo are different resource problems." }, { "left": "issue:44466", "right": "issue:45357", "accept": false, - "reason": "Both involve save_pretrained serialization, but one is lm_head weight serialization and the other is visual encoder key saving in a different model." + "reason": "Both concern save_pretrained regressions, but one is lm_head serialization and the other is visual encoder key naming in Qwen3.5." }, { "left": "issue:45685", "right": "issue:46821", "accept": false, - "reason": "Unrelated: MPS histogram support issue vs TimesFM window_size attribute error." + "reason": "MPS histogram error for Int and TimesFM window_size AttributeError are unrelated." }, { "left": "issue:26413", "right": "issue:28530", "accept": false, - "reason": "Different training-loop issues: checkpoint resumption epoch_iterator failure vs early stopping metric selection." - }, - { - "left": "issue:44451", - "right": "issue:47020", - "accept": false, - "reason": "Both are tokenizer-loading issues, but they affect different tokenizer implementations and model families." - }, - { - "left": "issue:44186", - "right": "issue:45072", - "accept": false, - "reason": "Tokenizer crashes on NER/batched padding are unrelated to a bfloat16 inference dtype mismatch." - }, - { - "left": "issue:43295", - "right": "issue:45701", - "accept": false, - "reason": "A multimodal processor/tokenizer API regression is not the same underlying bug as generic tokenization changes across versions." - }, - { - "left": "issue:46897", - "right": "issue:47317", - "accept": false, - "reason": "Both are training-loss bugs, but they involve different models and no evidence of the same code path or fix." - }, - { - "left": "issue:43065", - "right": "issue:46402", - "accept": false, - "reason": "Same Sam3 family, but different problems: dummy Conv2d in the pixel decoder vs missing text encoder weights loading." + "reason": "resume_from_checkpoint epoch_iterator failure and early-stopping metric lookup failure are distinct trainer features." } ] }, @@ -4997,20 +4909,18 @@ "source": "llm", "cache_hit": false, "model": "gpt-5.4-mini?service_tier=flex", - "cluster_id": "cluster-10000-529", + "cluster_id": "cluster-10000-530", "nodes": [ "issue:28530", - "issue:29159", "issue:30064", "issue:33666", - "issue:36032", "issue:41720", "issue:42915", - "issue:43316", + "issue:43065", + "issue:43295", "issue:43317", "issue:43404", "issue:43441", - "issue:43452", "issue:43531", "issue:43550", "issue:43572", @@ -5025,255 +4935,256 @@ "issue:44186", "issue:44361", "issue:44423", + "issue:44451", "issue:44464", "issue:44488", "issue:44610", - "issue:44671", "issue:44734", - "issue:44743", "issue:44898", - "issue:44913", + "issue:45072", "issue:45161", "issue:45357", "issue:45406", - "issue:45479", + "issue:45701", "issue:45823", "issue:46291", + "issue:46402", "issue:46612", "issue:46665", "issue:46821", "issue:46858", "issue:46869", + "issue:46897", "issue:47020", "issue:47317", "issue:47337" ], "soft_pairs": [ - "issue:46612|issue:46858", + "issue:44186|issue:45072", + "issue:43295|issue:45701", + "issue:44451|issue:47020", + "issue:43065|issue:46402", "issue:43550|issue:43720", "issue:43646|issue:46665", - "issue:43577|issue:43708", + "issue:46612|issue:46858", "issue:44186|issue:44610", "issue:43550|issue:43606", + "issue:43577|issue:43708", + "issue:46897|issue:47317", "issue:46821|issue:46869", "issue:42915|issue:43317", "issue:41720|issue:45161", "issue:43404|issue:45357", - "issue:44610|issue:44898", "issue:44423|issue:45406", - "issue:44008|issue:45823", + "issue:44610|issue:44898", "issue:28530|issue:33666", + "issue:44008|issue:45823", "issue:43881|issue:44488", "issue:42915|issue:46291", "issue:43531|issue:43572", "issue:43441|issue:47337", - "issue:44734|issue:46858", "issue:44186|issue:44361", - "issue:43716|issue:46821", + "issue:44734|issue:46858", "issue:28530|issue:30064", - "issue:44464|issue:46858", - "issue:44671|issue:44743", - "issue:45479|issue:47317", - "issue:29159|issue:36032", - "issue:43316|issue:44913", - "issue:43452|issue:47020" + "issue:43716|issue:46821", + "issue:44464|issue:46858" ], - "prepared_review_unit_hash": "f9046e3382ee847a7601cc8606aee9c7d92cbb22c33d84c40f353d020aa4c05a", + "prepared_review_unit_hash": "3955e1e0d2c12f2c04856365199ab49c8efa5133779f3cf92f1c474b70bfab7e", "component_budget": { - "node_count": 529, - "item_count": 529, - "soft_pair_count": 1953, - "serialized_chars": 1012074, - "estimated_input_tokens": 253019, - "estimated_eval_tokens": 506294 + "node_count": 530, + "item_count": 530, + "soft_pair_count": 1954, + "serialized_chars": 1013004, + "estimated_input_tokens": 253251, + "estimated_eval_tokens": 506758 }, "budget": { - "node_count": 48, - "item_count": 48, + "node_count": 47, + "item_count": 47, "soft_pair_count": 28, - "serialized_chars": 36250, - "estimated_input_tokens": 9063, - "estimated_eval_tokens": 18382 + "serialized_chars": 35509, + "estimated_input_tokens": 8878, + "estimated_eval_tokens": 18012 }, "overflow_policy": "truncate_then_skip", "trimmed": false, "aggressively_trimmed": false, "split": true, "analyst_result": { - "summary": "This cluster is mostly a bag of unrelated Transformers issues that only share broad subsystem vocabulary (loading, tokenizers, cache, device maps, multimodal serving). None of the soft pairs look like the same concrete bug, so I would not merge any of them.", - "confidence": 0.92, - "canonical_issue_reason": "issue:43646 is the best broad representative: it describes a general Transformers 5.0 initialization regression, which is closer to the cluster\u2019s repeated loading/initialization failure themes than the model-specific reports.", + "summary": "This is a heterogeneous cluster of mostly unrelated bug reports spanning tokenization, loading, device-map/offload, dtype, and model-specific regressions. None of the soft pairs look like the same concrete underlying bug, so all should be rejected. If a representative issue is needed, #45701 is the broadest and most referenced report in the set, but the cluster itself is not cohesive.", + "confidence": 0.39, + "canonical_issue_reason": "#45701 is the broadest, most heavily referenced report here and the closest thing to an umbrella representative, but there is no truly canonical issue because the cluster is not a single bug family.", "canonical_pr_reason": null, - "best_issue_reason": "Among these mixed issues, issue:43646 is the most general and reusable as a representative umbrella for the cluster, because it targets core model initialization rather than a single model or edge-case path.", + "best_issue_reason": "#45701 is the best standalone representative because it is broad and well-linked, though it does not actually unify the other issues.", "best_pr_reason": null, "soft_edge_verdicts": [ { - "left": "issue:46612", - "right": "issue:46858", + "left": "issue:44186", + "right": "issue:45072", + "accept": false, + "reason": "Both are multimodal/vision-related failures, but one is tokenizer padding/truncation and the other is dtype mismatch in bfloat16 inference; not the same bug." + }, + { + "left": "issue:43295", + "right": "issue:45701", + "accept": false, + "reason": "Both mention regressions around custom code/tokenization, but one is a processor/tokenizer API break and the other is a broad version-caused tokenization change." + }, + { + "left": "issue:44451", + "right": "issue:47020", + "accept": false, + "reason": "Both are tokenizer loading issues, but they involve different models and different failure modes; no shared concrete code path." + }, + { + "left": "issue:43065", + "right": "issue:46402", "accept": false, - "reason": "Both involve generation/cache code, but one is beam-search cache reorder skipping and the other is a static-cache seq-length tensor arithmetic bug; different code paths and fixes." + "reason": "Same model family (Sam3), but one is a dummy Conv2d placeholder and the other is missing text encoder weight loading; separate defects." }, { "left": "issue:43550", "right": "issue:43720", "accept": false, - "reason": "Different failures: torch.compile with SDPA in Bamba vs. packed BitNet weights not unpacked during accelerate loading." + "reason": "Both involve loading/model execution with accelerate-like paths, but one is torch.compile+SDPA on Bamba and the other is packed-weight unpacking for BitNet; different bugs." }, { "left": "issue:43646", "right": "issue:46665", "accept": false, - "reason": "Both are loading/init regressions, but one is a v5 custom model initialization breakage and the other is inconsistent from_pretrained behavior inside Python classes; not the same bug." + "reason": "Both concern custom model/loading behavior, but one is custom model initialization in Transformers 5.0 and the other is from_pretrained inside Python classes; not the same issue." }, { - "left": "issue:43577", - "right": "issue:43708", + "left": "issue:46612", + "right": "issue:46858", "accept": false, - "reason": "One is a dtype propagation bug when loading Blip2, the other is a Trainer checkpoint resume max_steps calculation bug." + "reason": "Both touch generation cache behavior, but one is beam-search cache reorder skipped for specific model classes and the other is a static-cache seq-length tensor issue under torch.compile." }, { "left": "issue:44186", "right": "issue:44610", "accept": false, - "reason": "Both are preprocessing bugs, but they affect different components and different root causes: tokenizer crashes vs. processor image-size mismatch." + "reason": "Both are vision processor issues, but one is LayoutLMv2 tokenizer/NER padding and the other is OmDet-Turbo image size mismatch; unrelated." }, { "left": "issue:43550", "right": "issue:43606", "accept": false, - "reason": "Both mention device/offload style failures, but one is SDPA/torch.compile on a specific model and the other is a CPU-offload device mismatch in a different model." + "reason": "Both are device-map/offload-related, but one is a torch.compile SDPA failure and the other is a CPU-offload device mismatch in bark-small; different code paths." + }, + { + "left": "issue:43577", + "right": "issue:43708", + "accept": false, + "reason": "One is a dtype propagation bug in Blip2 loading and the other is Trainer checkpoint resumption max_steps calculation; unrelated." + }, + { + "left": "issue:46897", + "right": "issue:47317", + "accept": false, + "reason": "Both are training-loss bugs, but they affect different models and likely different loss computation paths; not the same change." }, { "left": "issue:46821", "right": "issue:46869", "accept": false, - "reason": "Unrelated: a TimesFM window_size attribute error versus a YOLO multi-GPU device_map auto failure." + "reason": "Both mention multi-GPU/device-map style failures, but one is a TimesFM window_size AttributeError and the other is YOLO auto device-map loading; distinct bugs." }, { "left": "issue:42915", "right": "issue:43317", "accept": false, - "reason": "Both touch loading with advanced config/offload, but one is FineGrainedFP8Config on Qwen3Moe and the other is dequantized model loading with gpu+cpu offload." + "reason": "Both are model loading failures, but one is Qwen3Moe with FineGrainedFP8Config and the other is dequantized model loading with gpu+cpu offload; different failure modes." }, { "left": "issue:41720", "right": "issue:45161", "accept": false, - "reason": "Different model families and different distributed/runtime failures: cudaErrorAssert on A800 vs. tensor-parallel support not working." + "reason": "Both concern model-parallel/device placement, but one is Qwen3 auto device mapping on A800 and the other is GPT-OSS MoE tensor-parallel behavior; not the same bug." }, { "left": "issue:43404", "right": "issue:45357", "accept": false, - "reason": "One is a tied-weights bug in Mistral3 generation, the other is incorrect visual encoder keys saved in Qwen3.5." - }, - { - "left": "issue:44610", - "right": "issue:44898", - "accept": false, - "reason": "Both are vision preprocessing issues, but the concrete bugs differ: wrong output resolution vs. non-default-resolution interpolation failure." + "reason": "Both involve multimodal generation models and weights, but one is tied lm_head in Mistral3 and the other is incorrect visual encoder keys in Qwen3.5 save_pretrained; separate defects." }, { "left": "issue:44423", "right": "issue:45406", "accept": false, - "reason": "Both are serve-related multimodal crashes, but the failures are different: one is a string/.to issue in continuous batching, the other is a missing _tokenizer attribute on Gemma4Processor." + "reason": "Both are serve/runtime crashes for multimodal processors, but one is a 'str' object .to() failure in continuous batching and the other is missing _tokenizer on Gemma4Processor; different root causes." }, { - "left": "issue:44008", - "right": "issue:45823", + "left": "issue:44610", + "right": "issue:44898", "accept": false, - "reason": "Different Gemma variants and different root causes: a variable-name collision in forward vs. a device mismatch in PLE forwarding." + "reason": "Both are image-size/preprocessing mismatches, but OmDet-Turbo has 640px vs 224px processor/model mismatch while Perceiver fails on non-default resolution despite interpolate_pos_encoding; not the same bug." }, { "left": "issue:28530", "right": "issue:33666", "accept": false, - "reason": "Early stopping/metric selection is unrelated to Qwen2-VL multi-GPU training." + "reason": "Early stopping metric lookup and Qwen2-VL multi-GPU training are unrelated." + }, + { + "left": "issue:44008", + "right": "issue:45823", + "accept": false, + "reason": "Both are Gemma/Mistral-family multimodal device issues, but one is a variable-name collision causing an AttributeError and the other is a device mismatch under auto mapping in forward." }, { "left": "issue:43881", "right": "issue:44488", "accept": false, - "reason": "Both are model loading failures, but one is a generic glm-4v-9b load issue and the other is a specific tokenizer/model loading problem for cjvt/sleng-bert." + "reason": "Both are model loading failures, but they involve different models and likely different tokenizer/config issues; no clear shared bug." }, { "left": "issue:42915", "right": "issue:46291", "accept": false, - "reason": "One is a Qwen3Moe FP8 loading/config issue; the other is an import-time kernels dependency/version problem." + "reason": "One is a FineGrainedFP8Config load failure and the other is a pipelines import/version constraint on kernels; unrelated." }, { "left": "issue:43531", "right": "issue:43572", "accept": false, - "reason": "Different config bugs: Qwen3-MoE sliding_window behavior vs. StableLmConfig missing pad_token_idx after v5 update." + "reason": "Both are config-related regressions in newer versions, but one is sliding_window behavior for Qwen3-MoE and the other is missing pad_token_idx in StableLmConfig; separate issues." }, { "left": "issue:43441", "right": "issue:47337", "accept": false, - "reason": "Completely unrelated: FlashAttention on Ministral-3 vs. a flaky bbox-range test." - }, - { - "left": "issue:44734", - "right": "issue:46858", - "accept": false, - "reason": "Both involve KV cache behavior, but one is serve continuation indexing and the other is static-cache seq-length arithmetic on non-CUDA backends." + "reason": "FlashAttention failure in Ministral-3 and a flaky bbox coordinate test are unrelated." }, { "left": "issue:44186", "right": "issue:44361", "accept": false, - "reason": "Both are tokenizer/task crashes, but they affect different tokenizers and different failure modes." + "reason": "Both are tokenizer/attribute errors on task inputs, but they affect different tokenizers and different task flows; no shared code path." }, { - "left": "issue:43716", - "right": "issue:46821", + "left": "issue:44734", + "right": "issue:46858", "accept": false, - "reason": "Different models and bugs: a dtype mismatch between image preprocessor and model vs. a TimesFM window_size attribute error." + "reason": "Both involve serve/cache behavior, but one is KV cache continuation in /v1/responses and the other is static cache seq-length arithmetic under torch.compile; different bugs." }, { "left": "issue:28530", "right": "issue:30064", "accept": false, - "reason": "Early stopping metric selection is unrelated to image processor void segmentation map handling." - }, - { - "left": "issue:44464", - "right": "issue:46858", - "accept": false, - "reason": "Both mention compiled/cache paths, but one is inconsistent chunked generation output and the other is a static-cache slice arithmetic failure." - }, - { - "left": "issue:44671", - "right": "issue:44743", - "accept": false, - "reason": "Different model bugs: CamemBERT masked-LM prediction regression vs. Qwen3.5 recurrent states resetting with cache." - }, - { - "left": "issue:45479", - "right": "issue:47317", - "accept": false, - "reason": "Generic training-loss wording is not enough; one is a zero-loss classification bug across models, the other is a PPFormulaNet-specific training-loss bug." - }, - { - "left": "issue:29159", - "right": "issue:36032", - "accept": false, - "reason": "Both are tokenizer issues, but one is slow/fast inconsistency and the other is a T5Tokenizer method-name conflict during load." + "reason": "Early stopping metric selection and void segmentation map processing are unrelated." }, { - "left": "issue:43316", - "right": "issue:44913", + "left": "issue:43716", + "right": "issue:46821", "accept": false, - "reason": "Both are config persistence/API discrepancies, but one concerns Gemma3TextConfig and the other rotary_pct silently resetting on reload." + "reason": "Both mention shape/dtype-ish model input issues, but one is Mistral image-preprocessor dtype mismatch and the other is TimesFM window_size AttributeError; not the same bug." }, { - "left": "issue:43452", - "right": "issue:47020", + "left": "issue:44464", + "right": "issue:46858", "accept": false, - "reason": "Both affect from_pretrained/tokenizer loading, but one is a gguf_file regression across APIs and the other is an XLMRobertaTokenizer constructor type mismatch." + "reason": "Both concern compiled generation behavior, but one is inconsistent chunked generation outputs and the other is cache length tensor arithmetic on non-CUDA backends; different failures." } ] }, @@ -5295,20 +5206,22 @@ "source": "llm", "cache_hit": false, "model": "gpt-5.4-mini?service_tier=flex", - "cluster_id": "cluster-10000-529", + "cluster_id": "cluster-10000-530", "nodes": [ + "issue:29159", "issue:30064", "issue:33357", - "issue:34567", "issue:34634", "issue:35707", + "issue:36032", "issue:41720", "issue:42490", "issue:42915", "issue:43257", + "issue:43316", "issue:43322", "issue:43329", - "issue:43388", + "issue:43452", "issue:43526", "issue:43531", "issue:43540", @@ -5321,264 +5234,240 @@ "issue:43873", "issue:43881", "issue:43883", - "issue:44077", - "issue:44451", "issue:44492", "issue:44509", "issue:44534", "issue:44589", - "issue:44625", + "issue:44671", + "issue:44743", + "issue:44913", "issue:44991", - "issue:45059", - "issue:45406", - "issue:45464", + "issue:45479", "issue:45632", "issue:45684", "issue:45701", "issue:45704", "issue:45759", - "issue:45859", "issue:46665", "issue:46821", - "issue:46858", "issue:46894", + "issue:47020", "issue:47145", "issue:47254", "issue:47317" ], "soft_pairs": [ - "issue:45632|issue:45684", + "issue:44671|issue:44743", + "issue:29159|issue:36032", + "issue:43316|issue:44913", "issue:30064|issue:34634", + "issue:45632|issue:45684", + "issue:43452|issue:47020", + "issue:45479|issue:47317", "issue:45701|issue:47145", "issue:42915|issue:43322", "issue:43873|issue:45704", "issue:33357|issue:35707", "issue:43540|issue:43782", - "issue:46894|issue:47317", "issue:43531|issue:45759", - "issue:41720|issue:43746", "issue:43881|issue:44991", + "issue:41720|issue:43746", "issue:43550|issue:43653", "issue:44534|issue:46665", + "issue:46894|issue:47317", "issue:43257|issue:47254", "issue:43526|issue:43883", "issue:42490|issue:43329", "issue:43606|issue:43720", "issue:43531|issue:46821", - "issue:34634|issue:41720", "issue:44492|issue:44509", - "issue:43531|issue:44589", - "issue:34567|issue:35707", - "issue:45406|issue:45464", - "issue:43606|issue:43653", - "issue:45759|issue:45859", - "issue:44451|issue:44991", - "issue:43540|issue:46858", - "issue:43388|issue:45059", - "issue:44077|issue:44625" + "issue:34634|issue:41720", + "issue:43531|issue:44589" ], - "prepared_review_unit_hash": "0aa694607cec554133aae7f60b4b1210e21b1c56599ad4588fc387ec05f640b2", + "prepared_review_unit_hash": "7fd7cb8141e7c1850c4520de4244f1d18bc3074726f107c13c3a0c17b4142be5", "component_budget": { - "node_count": 529, - "item_count": 529, - "soft_pair_count": 1953, - "serialized_chars": 1012074, - "estimated_input_tokens": 253019, - "estimated_eval_tokens": 506294 + "node_count": 530, + "item_count": 530, + "soft_pair_count": 1954, + "serialized_chars": 1013004, + "estimated_input_tokens": 253251, + "estimated_eval_tokens": 506758 }, "budget": { - "node_count": 48, - "item_count": 48, - "soft_pair_count": 29, - "serialized_chars": 36318, - "estimated_input_tokens": 9080, - "estimated_eval_tokens": 18416 + "node_count": 47, + "item_count": 47, + "soft_pair_count": 26, + "serialized_chars": 34870, + "estimated_input_tokens": 8718, + "estimated_eval_tokens": 17692 }, "overflow_policy": "truncate_then_skip", "trimmed": false, "aggressively_trimmed": false, "split": true, "analyst_result": { - "summary": "This cluster is mostly a grab bag of unrelated Transformers issues; the soft pairs share vocabulary or a broad subsystem, but not the same concrete bug. No pull requests are present. Among the issues, #43873 is the best representative because it is a broad, high-engagement open regression report with multiple references.", - "confidence": 0.68, - "canonical_issue_reason": "#43873 is the strongest cluster representative: broad offloading/quantization bug, open, and one of the highest-engagement reports in the set.", + "summary": "No soft edge looks like a true duplicate or mergeable same-bug pair; the cluster is mostly a loose similarity set spanning tokenizer, config, loading, quantization, and model-specific runtime bugs. If one issue must represent the cluster, the Qwen3-MoE sliding-window bug is the closest anchor, but it still does not subsume the others.", + "confidence": 0.77, + "canonical_issue_reason": "issue:43531 is the most central item in the sample, appearing in multiple soft candidates and being a concrete code-path bug rather than a broad meta report; it is the closest representative of the cluster, though not a duplicate of the linked issues.", "canonical_pr_reason": null, - "best_issue_reason": "#43873 is the best standalone issue to anchor the cluster because it is concrete, widely discussed, and closest to a reusable triage target.", + "best_issue_reason": "issue:43531 is the best available issue representative because it is specific, technical, and the most connected of the sampled issues; however, the cluster is too heterogeneous for any issue to be a strong canonical summary.", "best_pr_reason": null, "soft_edge_verdicts": [ { - "left": "issue:45632", - "right": "issue:45684", + "left": "issue:44671", + "right": "issue:44743", + "accept": false, + "reason": "Both are model-specific behavior bugs, but one is CamemBERT masked-LM output and the other is Qwen modular cache state reset; different models and code paths." + }, + { + "left": "issue:29159", + "right": "issue:36032", + "accept": false, + "reason": "Both mention tokenizers, but one is slow-vs-fast inconsistency and the other is a T5Tokenizer load-time attribute conflict; not the same bug." + }, + { + "left": "issue:43316", + "right": "issue:44913", "accept": false, - "reason": "Different bugs: cache-path collision for local trust_remote_code loading vs permission propagation in save_pretrained." + "reason": "One is a Gemma3TextConfig API discrepancy, the other is GPTNeoX rotary_pct persistence on reload; different configuration fields and failure modes." }, { "left": "issue:30064", "right": "issue:34634", "accept": false, - "reason": "Unrelated processor failures: void segmentation-map handling vs Bark voice_preset parsing." + "reason": "Image processor void segmentation maps and BarkProcessor voice_preset are unrelated processor-specific issues." + }, + { + "left": "issue:45632", + "right": "issue:45684", + "accept": false, + "reason": "Cache-path collision for trust_remote_code and read-only permission propagation in save_pretrained are different file-system concerns with different fixes." + }, + { + "left": "issue:43452", + "right": "issue:47020", + "accept": false, + "reason": "Both are tokenizer/loading-related, but gguf_file breaks from_pretrained broadly while the XLMRobertaTokenizer issue is a specific Unigram vocab type mismatch." + }, + { + "left": "issue:45479", + "right": "issue:47317", + "accept": false, + "reason": "Both involve training loss, but they affect different models and likely different loss computations; too broad to be the same bug." }, { "left": "issue:45701", "right": "issue:47145", "accept": false, - "reason": "Tokenization/version behavior change is unrelated to FP8 frozen-checkpoint weight_scale loading." + "reason": "Tokenization changes across versions and FP8 weight_scale loading corruption are unrelated." }, { "left": "issue:42915", "right": "issue:43322", "accept": false, - "reason": "Different model paths and failure modes: Qwen3Moe FineGrainedFP8Config vs Llava Next segfault." + "reason": "FineGrainedFP8Config failure in Qwen3Moe and Llava Next segmentation fault are different model families and distinct failure classes." }, { "left": "issue:43873", "right": "issue:45704", "accept": false, - "reason": "Both mention loading/runtime behavior, but one is offloading+quantization and the other is a T5/apex RMSNorm memory leak." + "reason": "Quantization offloading behavior and T5/apex RMSNorm memory leak are not the same underlying issue." }, { "left": "issue:33357", "right": "issue:35707", "accept": false, - "reason": "MacOS bus error with CLIP is unrelated to progressive generation with inputs_embeds/past_key_values." + "reason": "MacOS bus error in a community CLIP model and progressive generation with inputs_embeds/past_key_values are unrelated." }, { "left": "issue:43540", "right": "issue:43782", "accept": false, - "reason": "Same model family, but different concrete failures in different APIs/code paths (video input processing vs from_pretrained weight_only)." - }, - { - "left": "issue:46894", - "right": "issue:47317", - "accept": false, - "reason": "Both are training-loss bugs, but they affect different models and likely different loss pipelines." + "reason": "Both are Qwen multimodal-related, but one is a video-input processing error and the other is a from_pretrained weight_only load failure." }, { "left": "issue:43531", "right": "issue:45759", "accept": false, - "reason": "Sliding-window handling in Qwen3-MoE is not the same as config unwrapping for composite Qwen multimodal models." + "reason": "Both involve Qwen3, but one is a sliding_window recurrent-state bug and the other is config unwrapping for composite multimodal models; different code paths." }, { - "left": "issue:41720", - "right": "issue:43746", + "left": "issue:43881", + "right": "issue:44991", "accept": false, - "reason": "CUDA assert during auto device mapping is a different failure than PEFT adapter local-checkpoint loading." + "reason": "GLM-4V loading failure and tokenizer loading for EMBEDDIA/est-roberta are unrelated." }, { - "left": "issue:43881", - "right": "issue:44991", + "left": "issue:41720", + "right": "issue:43746", "accept": false, - "reason": "Different loading problems: one is a model-load failure for glm-4v-9b, the other is tokenizer loading for a specific RoBERTa model." + "reason": "Auto device mapping CUDA assert on Qwen3 and GraniteSpeech PEFT checkpoint loading are different problems." }, { "left": "issue:43550", "right": "issue:43653", "accept": false, - "reason": "Torch.compile/SDPA failure in Bamba is unrelated to BigBirdTokenizer special-token registration." + "reason": "torch.compile/SDPA on Bamba-9B-v2 and BigBirdTokenizer special-token registration are unrelated." }, { "left": "issue:44534", "right": "issue:46665", "accept": false, - "reason": "Non-persistent buffer corruption in v5 is unrelated to from_pretrained behavior inside Python classes." + "reason": "Non-persistent buffer junk in v5 and from_pretrained behavior inside Python classes do not describe the same failure." + }, + { + "left": "issue:46894", + "right": "issue:47317", + "accept": false, + "reason": "Both are training-loss bugs, but they concern different models and likely different shifting logic; not enough to treat as one issue." }, { "left": "issue:43257", "right": "issue:47254", "accept": false, - "reason": "Both involve distributed loading/training, but the concrete problems differ: MoE weight conversion vs PEFT+ZeRO-3 checkpointing." + "reason": "Qwen3 MoE weight conversion under accelerate+deepspeed and PEFT+ZeRO-3 gradient checkpointing are distinct deep-learning stack interactions." }, { "left": "issue:43526", "right": "issue:43883", "accept": false, - "reason": "Beit label reduction is unrelated to Molmo's all_tied_weights_keys attribute error." + "reason": "BEiTImageProcessorFast reduce_labels and Molmo all_tied_weights_keys are unrelated attribute/processing bugs." }, { "left": "issue:42490", "right": "issue:43329", "accept": false, - "reason": "A trailing-slash import warning is unrelated to an undefined-variable bug in multimodal token counting." + "reason": "Local-path trailing-slash import warning and undefined video-branch variables in multimodal token counting are unrelated." }, { "left": "issue:43606", "right": "issue:43720", "accept": false, - "reason": "Both touch loading/offload, but the actual bugs are different: CPU offload device mismatch vs unpacking packed weights." + "reason": "CPU offload device mismatch and BitNet packed-weight unpacking during accelerate loading are both loading-related but not the same concrete bug." }, { "left": "issue:43531", "right": "issue:46821", "accept": false, - "reason": "Qwen3-MoE sliding_window behavior is not the same as TimesFM window_size attribute handling." - }, - { - "left": "issue:34634", - "right": "issue:41720", - "accept": false, - "reason": "BarkProcessor voice_preset and Qwen3 auto device-mapping failures are unrelated." + "reason": "Qwen3-MoE sliding-window behavior and TimesFM window_size AttributeError are different model implementations and different parameter handling." }, { "left": "issue:44492", "right": "issue:44509", "accept": false, - "reason": "A typo in cache strategies is unrelated to stale docs for removed pipeline tasks." - }, - { - "left": "issue:43531", - "right": "issue:44589", - "accept": false, - "reason": "Sliding-window model logic is unrelated to missing Float8 storage during checkpoint loading." - }, - { - "left": "issue:34567", - "right": "issue:35707", - "accept": false, - "reason": "TrainerState token-count tracking is unrelated to progressive generation with embeddings and cache state." - }, - { - "left": "issue:45406", - "right": "issue:45464", - "accept": false, - "reason": "Gemma4Processor server crash and Qwen3.5 streaming inference failure are different code paths and symptoms." - }, - { - "left": "issue:43606", - "right": "issue:43653", - "accept": false, - "reason": "CPU offload/device mismatch is unrelated to tokenizer mask-token registration." - }, - { - "left": "issue:45759", - "right": "issue:45859", - "accept": false, - "reason": "Config unwrapping for multimodal Qwen models is unrelated to CPU-offload incompatibility in rotary embedding." - }, - { - "left": "issue:44451", - "right": "issue:44991", - "accept": false, - "reason": "Loading ScandiBERT and loading EMBEDDIA/est-roberta are different tokenizer/model issues." - }, - { - "left": "issue:43540", - "right": "issue:46858", - "accept": false, - "reason": "Video-input processing in Qwen3OmniMoe is unrelated to cache length arithmetic under static cache and torch.compile." + "reason": "A typo in cache strategies and stale docs for removed pipeline tasks are unrelated documentation/API issues." }, { - "left": "issue:43388", - "right": "issue:45059", + "left": "issue:34634", + "right": "issue:41720", "accept": false, - "reason": "Label gathering in metrics is unrelated to SAM3 text/bounding-box behavior." + "reason": "BarkProcessor voice_preset and Qwen3 auto device mapping CUDA assert are different models and failure modes." }, { - "left": "issue:44077", - "right": "issue:44625", + "left": "issue:43531", + "right": "issue:44589", "accept": false, - "reason": "PatchTSMixer post_init policy and Qwen3.5 num_labels propagation are separate configuration issues." + "reason": "Qwen3-MoE sliding-window logic and Float8 storage lookup failure are unrelated runtime bugs." } ] }, @@ -5600,46 +5489,48 @@ "source": "llm", "cache_hit": false, "model": "gpt-5.4-mini?service_tier=flex", - "cluster_id": "cluster-10000-529", + "cluster_id": "cluster-10000-530", "nodes": [ "issue:30064", + "issue:34567", + "issue:35707", "issue:36010", "issue:39692", "issue:41720", "issue:42886", "issue:43381", + "issue:43388", "issue:43421", - "issue:43493", + "issue:43540", + "issue:43606", "issue:43644", "issue:43653", - "issue:43673", "issue:43746", - "issue:43844", "issue:43873", "issue:43927", - "issue:43994", "issue:44062", - "issue:44368", + "issue:44077", "issue:44451", - "issue:44464", "issue:44514", "issue:44545", "issue:44610", "issue:44625", - "issue:44671", "issue:44843", "issue:44898", "issue:44918", "issue:44936", "issue:44987", - "issue:44995", + "issue:44991", "issue:45042", + "issue:45059", "issue:45072", - "issue:45081", - "issue:45356", "issue:45357", + "issue:45406", + "issue:45464", "issue:45479", "issue:45704", + "issue:45759", + "issue:45859", "issue:45923", "issue:46050", "issue:46402", @@ -5647,235 +5538,226 @@ "issue:46665", "issue:46693", "issue:46858", - "issue:47050", - "issue:47333", - "issue:47337" + "issue:47333" ], "soft_pairs": [ + "issue:45406|issue:45464", + "issue:34567|issue:35707", + "issue:43606|issue:43653", + "issue:45759|issue:45859", + "issue:44451|issue:44991", + "issue:43540|issue:46858", + "issue:43388|issue:45059", + "issue:44077|issue:44625", "issue:43644|issue:46665", "issue:44987|issue:45042", - "issue:45357|issue:46402", "issue:45923|issue:47333", + "issue:45357|issue:46402", "issue:44898|issue:45072", "issue:44062|issue:44451", "issue:41720|issue:45704", "issue:44918|issue:44936", "issue:39692|issue:44610", "issue:43746|issue:46050", - "issue:44625|issue:45479", "issue:43381|issue:45704", - "issue:44514|issue:44545", + "issue:44625|issue:45479", "issue:42886|issue:44843", "issue:30064|issue:36010", - "issue:43421|issue:43927", + "issue:44514|issue:44545", "issue:43746|issue:43873", - "issue:43381|issue:43844", - "issue:46655|issue:46693", - "issue:43421|issue:43653", - "issue:39692|issue:43493", - "issue:44464|issue:47050", - "issue:44995|issue:46858", - "issue:45081|issue:45356", - "issue:43673|issue:47337", - "issue:44368|issue:45357", - "issue:43994|issue:44671" + "issue:43421|issue:43927", + "issue:46655|issue:46693" ], - "prepared_review_unit_hash": "23b065ca5a0908c3781f5842f10eae8f811fe8a3b268a45a6c9bda14c1d0250d", + "prepared_review_unit_hash": "b0e8fd75f592ce1dc6850b91f57006920fd9c7dbe4341af763916e2c30af9d67", "component_budget": { - "node_count": 529, - "item_count": 529, - "soft_pair_count": 1953, - "serialized_chars": 1012074, - "estimated_input_tokens": 253019, - "estimated_eval_tokens": 506294 + "node_count": 530, + "item_count": 530, + "soft_pair_count": 1954, + "serialized_chars": 1013004, + "estimated_input_tokens": 253251, + "estimated_eval_tokens": 506758 }, "budget": { "node_count": 48, "item_count": 48, - "soft_pair_count": 27, - "serialized_chars": 36221, - "estimated_input_tokens": 9056, - "estimated_eval_tokens": 18368 + "soft_pair_count": 26, + "serialized_chars": 35719, + "estimated_input_tokens": 8930, + "estimated_eval_tokens": 18116 }, "overflow_policy": "truncate_then_skip", "trimmed": false, "aggressively_trimmed": false, "split": true, "analyst_result": { - "summary": "Most pairs are false positives across unrelated bug reports. The only clear duplicate here is the Qwen2_5_VLProcessor batched padding=False crash pair (44514/44545). No PRs are present.", - "confidence": 0.88, - "canonical_issue_reason": "No single canonical issue spans the whole set; the only true duplicate pair is 44514/44545, and 44514 is the earlier, cleaner report to anchor on.", + "summary": "These items are mostly unrelated issues from different code paths; only the Qwen2_5_VLProcessor batched-padding reports look like a true duplicate pair. The rest share broad themes like loading/offloading/tokenization but describe distinct failures.", + "confidence": 0.86, + "canonical_issue_reason": "issue:43873 is the broadest and most actively discussed issue here, with the most inbound references; if a single issue must represent the set, it is the best umbrella candidate, though this is not a true duplicate cluster.", "canonical_pr_reason": null, - "best_issue_reason": "Issue 44514 is the best issue candidate because it is the earliest report of the exact same batched-input crash and has the clearest title.", + "best_issue_reason": "issue:43873 has the strongest signal as a representative issue because it is high-activity, open, and broadly scoped around offloading/quantization behavior.", "best_pr_reason": null, "soft_edge_verdicts": [ + { + "left": "issue:45406", + "right": "issue:45464", + "accept": false, + "reason": "Different products and failures: Gemma4Processor missing _tokenizer vs Qwen3.5-0.8B streaming chat/completions crash. Same broad area, not the same bug." + }, + { + "left": "issue:34567", + "right": "issue:35707", + "accept": false, + "reason": "Both involve trainer-related state/generation behavior, but one is num_input_tokens_seen not updating and the other is progressive generation with inputs_embeds/past_key_values. Not the same underlying bug." + }, + { + "left": "issue:43606", + "right": "issue:43653", + "accept": false, + "reason": "Different tokenizer/model problems: CPU offload device mismatch for bark-small vs BigBirdTokenizer mask token registration and empty decode output." + }, + { + "left": "issue:45759", + "right": "issue:45859", + "accept": false, + "reason": "Related to Qwen 3.5/3.6 multimodal config handling, but one is AutoModelForCausalLM.from_config text_config unwrapping and the other is a rotary embedding CPU-offload compatibility issue." + }, + { + "left": "issue:44451", + "right": "issue:44991", + "accept": false, + "reason": "Both are tokenizer loading regressions, but they affect different models and likely different root causes; not enough evidence of one concrete bug." + }, + { + "left": "issue:43540", + "right": "issue:46858", + "accept": false, + "reason": "Video-input ValueError in Qwen3OmniMoe is unrelated to cache.get_seq_length() returning a device tensor in static cache/torch.compile prefill." + }, + { + "left": "issue:43388", + "right": "issue:45059", + "accept": false, + "reason": "Different issues: gather_for_metrics dropping tuple labels in the last batch versus SAM3 PCS odd behavior with text and bounding boxes." + }, + { + "left": "issue:44077", + "right": "issue:44625", + "accept": false, + "reason": "patchtsmixer post_init allowance and Qwen3.5 num_labels propagation are unrelated model/config bugs." + }, { "left": "issue:43644", "right": "issue:46665", "accept": false, - "reason": "Both mention loading/behavior regressions, but they describe different problems: non-persistent buffers vs from_pretrained behavior inside Python classes." + "reason": "Non-persistent buffers filled with junk is a serialization/init bug; from_pretrained inside Python classes is a separate behavioral inconsistency." }, { "left": "issue:44987", "right": "issue:45042", "accept": false, - "reason": "Different failure modes and subsystems: model loading failure for a specific repo vs PIL image processor requiring torchvision." + "reason": "One is a loading regression for a specific model; the other is PIL image processors incorrectly requiring torchvision. Same general dependency/loading surface, different bugs." }, { - "left": "issue:45356", - "right": "issue:46402", + "left": "issue:45923", + "right": "issue:47333", "accept": false, - "reason": "Unrelated topics: tokenizer regex/codec regression vs Sam3Model text encoder weights not loading." + "reason": "Both concern offloading on Qwen models, but one is a missing supports_gradient_checkpointing flag and the other is inability to save a disk-offloaded model. Not the same code-path problem." }, { - "left": "issue:45923", - "right": "issue:47333", + "left": "issue:45357", + "right": "issue:46402", "accept": false, - "reason": "Different offloading/checkpoint concerns: missing supports_gradient_checkpointing flag vs inability to save a disk-offloaded model." + "reason": "Both involve Qwen/SAM model loading, but one is incorrect visual encoder keys on save_pretrained and the other is text encoder weights not loaded. Distinct defects." }, { "left": "issue:44898", "right": "issue:45072", "accept": false, - "reason": "Both involve dtype/vision issues, but they are different concrete bugs in different models and code paths." + "reason": "Perceiver image classification resolution handling and bfloat16 dtype mismatches in SwitchTransformers/TimmWrapperModel are unrelated." }, { "left": "issue:44062", "right": "issue:44451", "accept": false, - "reason": "Tokenizer AddedToken argument duplication is unrelated to failing to load a specific model checkpoint." + "reason": "AddedToken special-argument duplication and inability to load a specific tokenizer/model are different tokenizer-layer issues, not the same bug." }, { "left": "issue:41720", "right": "issue:45704", "accept": false, - "reason": "Completely different bugs: cudaErrorAssert during auto device mapping vs apex FusedRMSNorm memory leak." + "reason": "Qwen3 auto device mapping CUDA assert and T5 apex.FusedRMSNorm memory leak are unrelated inference/runtime problems." }, { "left": "issue:44918", "right": "issue:44936", "accept": false, - "reason": "Trainer embedding unpacking failure is not the same as evaluate() failing after train()." + "reason": "TRL SFT input-embedding unpacking and trainer.evaluate() after trainer.train() are different trainer behaviors." }, { "left": "issue:39692", "right": "issue:44610", "accept": false, - "reason": "Both are model/processor mismatch flavored, but they affect different models and distinct bugs." + "reason": "SigLIP2 doc example errors are documentation/example problems; OmDet-Turbo processor size mismatch is a runtime processor/model config bug. Not the same issue." }, { "left": "issue:43746", "right": "issue:46050", "accept": false, - "reason": "Different save/load regressions: PEFT local checkpoint loading vs quantized model round-trip after save_pretrained." - }, - { - "left": "issue:44625", - "right": "issue:45479", - "accept": false, - "reason": "Both mention labels, but one is config propagation for Qwen3.5 and the other is a zero-loss classification bug across models." + "reason": "PEFT adapters failing to load from local checkpoints and sinq-quantized models breaking after save/load are different persistence/quantization bugs." }, { "left": "issue:43381", "right": "issue:45704", "accept": false, - "reason": "Gradient checkpointing eval-mode restriction is unrelated to the T5 apex RMSNorm memory leak." + "reason": "Gradient checkpointing in eval mode and T5 apex RMSNorm memory leak are unrelated." }, { - "left": "issue:44514", - "right": "issue:44545", - "accept": true, - "reason": "These are the same crash report with only minor title/formatting differences: Qwen2_5_VLProcessor.apply_chat_template fails on batched input when padding=False." + "left": "issue:44625", + "right": "issue:45479", + "accept": false, + "reason": "Qwen3.5 num_labels propagation and single-label classification with num_labels=1 causing zero loss are different configuration/classification bugs." }, { "left": "issue:42886", "right": "issue:44843", "accept": false, - "reason": "Both involve offline/cache behavior, but one is tokenizer cache loading and the other is an unconditional model_info() call in a regex patch." + "reason": "Both involve offline/loading behavior, but one is tokenizer cache loading with HF_HUB_OFFLINE and the other is an unconditional model_info() call in _patch_mistral_regex. Different root causes." }, { "left": "issue:30064", "right": "issue:36010", "accept": false, - "reason": "Segmentation-map processing failure and GenerationMixin import error are unrelated." + "reason": "Void segmentation map processing and GenerationMixin import failure are unrelated subsystems." }, { - "left": "issue:43421", - "right": "issue:43927", - "accept": false, - "reason": "Special-token post-processor updates and DiaConfig custom token IDs being lost on save/load are different defects." + "left": "issue:44514", + "right": "issue:44545", + "accept": true, + "reason": "These titles describe the same exact Qwen2_5_VLProcessor.apply_chat_template batched-input crash with padding=False; they look like duplicate reports of the same bug." }, { "left": "issue:43746", "right": "issue:43873", "accept": false, - "reason": "Local checkpoint loading with PEFT adapters is unrelated to quantization/offloading behavior." + "reason": "Both mention loading/offloading/quantization, but one is PEFT local checkpoint loading and the other is general offloading with quantization. Not the same bug." }, { - "left": "issue:43381", - "right": "issue:43844", + "left": "issue:43421", + "right": "issue:43927", "accept": false, - "reason": "Eval-mode gradient checkpointing restriction is not the same as the ZeRO-3 gradient increase regression." + "reason": "Runtime post-processor update for special tokens and DiaConfig losing custom token IDs after save/load are related to token metadata but are distinct failures." }, { "left": "issue:46655", "right": "issue:46693", "accept": false, - "reason": "Video processor padding calculation and FA2 long-context performance regression are different issues." - }, - { - "left": "issue:43421", - "right": "issue:43653", - "accept": false, - "reason": "Runtime post-processor updates for special tokens is not the same as BigBirdTokenizer mask token registration/decoding." - }, - { - "left": "issue:39692", - "right": "issue:43493", - "accept": false, - "reason": "Both are SigLIP2-related, but one is a doc example problem and the other is an implementation discrepancy with the original JAX model." - }, - { - "left": "issue:44464", - "right": "issue:47050", - "accept": false, - "reason": "Chunked generation inconsistency with compiled forward is unrelated to ONNX export dtype mismatch in CanineModel." - }, - { - "left": "issue:44995", - "right": "issue:46858", - "accept": false, - "reason": "A stale indexer cache on second forward pass is not the same as cache.get_seq_length() returning a device tensor in prefill." - }, - { - "left": "issue:45081", - "right": "issue:45356", - "accept": false, - "reason": "Both touch fix_mistral_regex, but one is a crash on tokenizer patching and the other is a Kimi-K2.5 codec/warning regression; not the same concrete bug." - }, - { - "left": "issue:43673", - "right": "issue:47337", - "accept": false, - "reason": "Generation cache missing during chunked_prefill and a flaky bbox coordinate test are unrelated." - }, - { - "left": "issue:44368", - "right": "issue:45357", - "accept": false, - "reason": "A tied-word-embeddings warning during LoRA fine-tuning is unrelated to incorrect visual encoder keys being saved." - }, - { - "left": "issue:43994", - "right": "issue:44671", - "accept": false, - "reason": "Different models and failures: SigLIP2 nonsensical outputs vs CamemBERT masked LM predictions." + "reason": "SmolVLM video padding calculation and FA2 long-context generation performance regression are separate issues." } ] }, "evaluator_result": { "accept": true, - "feedback": "Grounded overall: the only accepted soft pair is the exact-title duplicate 44514/44545, and the summary stays conservative about the rest. No PRs are present, so the null PR fields are consistent with the packet." + "feedback": "The analysis stays grounded in the packet and the soft-edge calls are conservative. The only clearly supported duplicate pair is issue:44514 / issue:44545, and the choice of issue:43873 as an umbrella representative is reasonable based on activity and breadth. No obvious overstatement." }, "evaluator_used": true, "retried": false, @@ -5894,62 +5776,71 @@ "source": "llm", "cache_hit": false, "model": "gpt-5.4-mini?service_tier=flex", - "cluster_id": "cluster-10000-529", + "cluster_id": "cluster-10000-530", "nodes": [ "issue:30064", - "issue:34634", - "issue:36010", "issue:36331", "issue:39692", "issue:41720", "issue:43295", + "issue:43381", "issue:43421", "issue:43526", "issue:43531", "issue:43540", "issue:43550", "issue:43596", - "issue:43611", "issue:43643", "issue:43653", + "issue:43673", "issue:43720", "issue:43782", "issue:43819", + "issue:43844", "issue:43873", "issue:43881", "issue:43927", - "issue:44008", + "issue:43994", "issue:44062", "issue:44186", "issue:44220", + "issue:44368", + "issue:44464", "issue:44488", - "issue:44661", - "issue:44821", + "issue:44671", "issue:44857", - "issue:44898", "issue:44991", + "issue:44995", "issue:45072", - "issue:45084", + "issue:45081", "issue:45356", + "issue:45357", "issue:45381", "issue:45834", "issue:45907", "issue:46032", "issue:46294", "issue:46463", - "issue:46665", "issue:46693", - "issue:46829", - "issue:46832", + "issue:46858", "issue:47020", - "issue:47050" + "issue:47050", + "issue:47337" ], "soft_pairs": [ - "issue:43421|issue:44186", + "issue:43381|issue:43844", + "issue:43421|issue:43653", + "issue:39692|issue:43873", + "issue:44464|issue:47050", + "issue:45081|issue:45356", + "issue:44995|issue:46858", + "issue:44368|issue:45357", + "issue:43994|issue:44671", + "issue:43673|issue:47337", "issue:43782|issue:45381", + "issue:43421|issue:44186", "issue:46032|issue:46463", "issue:44857|issue:47050", - "issue:39692|issue:43873", "issue:45356|issue:45907", "issue:39692|issue:47050", "issue:43596|issue:44220", @@ -5966,231 +5857,216 @@ "issue:45072|issue:47050", "issue:41720|issue:44991", "issue:30064|issue:36331", - "issue:45381|issue:46693", - "issue:44220|issue:44661", - "issue:44821|issue:45084", - "issue:43611|issue:46665", - "issue:46829|issue:46832", - "issue:44186|issue:44898", - "issue:43540|issue:44008", - "issue:43653|issue:43819", - "issue:44991|issue:47020", - "issue:34634|issue:36010" + "issue:45381|issue:46693" ], - "prepared_review_unit_hash": "4895872fc91f8e266fc05b69854b03379f2d5c4fdcbbcb4297a6174bcbebe2e9", + "prepared_review_unit_hash": "6b6daa3bdd3de88d7b5bfa972d7609886809d4c0cc81f82347a9c8cbb7ebe37e", "component_budget": { - "node_count": 529, - "item_count": 529, - "soft_pair_count": 1953, - "serialized_chars": 1012074, - "estimated_input_tokens": 253019, - "estimated_eval_tokens": 506294 + "node_count": 530, + "item_count": 530, + "soft_pair_count": 1954, + "serialized_chars": 1013004, + "estimated_input_tokens": 253251, + "estimated_eval_tokens": 506758 }, "budget": { - "node_count": 47, - "item_count": 47, - "soft_pair_count": 31, - "serialized_chars": 36553, - "estimated_input_tokens": 9139, - "estimated_eval_tokens": 18534 + "node_count": 48, + "item_count": 48, + "soft_pair_count": 30, + "serialized_chars": 37015, + "estimated_input_tokens": 9254, + "estimated_eval_tokens": 18764 }, "overflow_policy": "truncate_then_skip", "trimmed": false, "aggressively_trimmed": false, "split": true, "analyst_result": { - "summary": "This cluster is a grab-bag of unrelated Transformers bugs across tokenizers, vision models, loading, compilation, and dtype/runtime paths. The soft edges are mostly superficial similarity rather than the same underlying defect, so none should be merged.", - "confidence": 0.82, - "canonical_issue_reason": "issue:43873 is the best representative anchor: it is the most active item, has broad user impact, and is the only open issue among many narrow runtime bugs.", + "summary": "This is a heterogeneous cluster of tokenizer, model-loading, generation, vision, and training bugs. I do not see any soft pair that is clearly the same underlying defect, so none should be merged as duplicates.", + "confidence": 0.94, + "canonical_issue_reason": "If forced to pick one anchor, issue 43421 is the broadest reusable tokenizer-runtime regression in the set; however the cluster overall is too mixed to treat as a single duplicate group.", "canonical_pr_reason": null, - "best_issue_reason": "issue:43873 is the strongest single issue to anchor the cluster because it has the highest discussion/inbound activity and describes a broad runtime regression rather than a niche model-specific edge case.", + "best_issue_reason": "43421 is the best representative only in the weak sense that it is a general tokenizer/backend behavior bug rather than a model-specific crash; it still does not summarize the whole cluster well.", "best_pr_reason": null, "soft_edge_verdicts": [ + { + "left": "issue:43381", + "right": "issue:43844", + "accept": false, + "reason": "Both involve training/runtime behavior, but one is gradient checkpointing in eval mode and the other is a DeepSpeed ZeRO-3 gradient anomaly; different failure modes and code paths." + }, { "left": "issue:43421", - "right": "issue:44186", + "right": "issue:43653", + "accept": false, + "reason": "Both touch tokenizers and special tokens, but 43421 is about runtime post-processor updates while 43653 is a BigBirdTokenizer mask-token registration/empty decode bug." + }, + { + "left": "issue:39692", + "right": "issue:43873", + "accept": false, + "reason": "One is a SigLIP2 doc/example problem; the other is an offloading-with-quantization runtime bug. Same broad area, but not the same defect." + }, + { + "left": "issue:44464", + "right": "issue:47050", + "accept": false, + "reason": "Both mention inference/compiled execution, but 44464 is chunked generation inconsistency and 47050 is an ONNX fp16 attention dtype mismatch." + }, + { + "left": "issue:45081", + "right": "issue:45356", + "accept": false, + "reason": "Both involve tokenizer regex handling, but one crashes on missing backend_tokenizer and the other reports a version regression with incorrect codec handling; not the same concrete bug." + }, + { + "left": "issue:44995", + "right": "issue:46858", + "accept": false, + "reason": "Both mention cache/indexing, but one is a stale indexer cache on repeated forwards and the other is a device-tensor seq-length arithmetic issue in static cache/torch.compile." + }, + { + "left": "issue:44368", + "right": "issue:45357", + "accept": false, + "reason": "One is a warning about tie_word_embeddings in fine-tuning, the other is incorrect keys saved for Qwen3.5 visual encoders; unrelated defects." + }, + { + "left": "issue:43994", + "right": "issue:44671", "accept": false, - "reason": "Different tokenizer problems: runtime special-token/post-processor sync vs LayoutLMv2 NER/padding crashes." + "reason": "Both are SigLIP/CamemBERT model behavior issues, but one is generic wrong outputs with AutoModel/pipeline and the other is masked-LM prediction correctness; different bugs." + }, + { + "left": "issue:43673", + "right": "issue:47337", + "accept": false, + "reason": "One is a generation cache missing in chunked prefill, the other is a flaky bbox range assertion in tests; not the same issue." }, { "left": "issue:43782", "right": "issue:45381", "accept": false, - "reason": "Different model families and failures: Qwen3VL loading/weight_only vs Qwen2.5-VL video position ids." + "reason": "Both are multimodal Qwen video/model issues, but one is a from_pretrained weight_only load error and the other is video vision_position_ids mismatch; distinct code paths." + }, + { + "left": "issue:43421", + "right": "issue:44186", + "accept": false, + "reason": "Both are tokenizer-related, but 43421 is about special-token post-processing updates and 44186 is LayoutLMv2 batching/padding/truncation crashes." }, { "left": "issue:46032", "right": "issue:46463", "accept": false, - "reason": "Unrelated bugs: Mamba2Mixer cache correctness vs T5 decoder-input validation message." + "reason": "One is a cache correctness bug in Mamba2Mixer, the other is an error-message validation issue for T5 decoder inputs; not the same problem." }, { "left": "issue:44857", "right": "issue:47050", "accept": false, - "reason": "Both are dtype-related, but they hit different models and code paths: LwDetr loss AMP vs Canine ONNX export attention." - }, - { - "left": "issue:39692", - "right": "issue:43873", - "accept": false, - "reason": "Doc/example mismatch and quantization failure is not the same as a runtime offloading-with-quantization bug." + "reason": "Both are dtype/AMP-adjacent runtime issues, but one is LwDetrImageLoss under float16 AMP and the other is CanineModel ONNX export attention dtype mismatch." }, { "left": "issue:45356", "right": "issue:45907", "accept": false, - "reason": "Tokenizer regression in Kimi-K2.5 is unrelated to list_repo_files failing to forward revision." + "reason": "One is a tokenizer regression, the other is a hub API forwarding issue (`list_repo_files` revision); unrelated." }, { "left": "issue:39692", "right": "issue:47050", "accept": false, - "reason": "A broken documentation example is not the same bug as fp16 ONNX export dtype mismatch." + "reason": "One is a SigLIP2 documentation/example mismatch, the other is an fp16 ONNX export dtype bug in CanineModel; different artifacts and failures." }, { "left": "issue:43596", "right": "issue:44220", "accept": false, - "reason": "Different subsystems: deepspeed zero3/BertModel init vs audio fbank feature extraction." + "reason": "Both are runtime errors, but one is DeepSpeed ZeRO-3 init with BertModel and the other is `_torch_extract_fbank_features()`; no concrete overlap." }, { "left": "issue:43720", "right": "issue:43819", "accept": false, - "reason": "Different mechanisms and models: packed-weight loading in BitNet vs missing STE behavior in DAC.from_latents." + "reason": "One is about unpacking packed weights during accelerate loading, the other is a DAC latent/STE mismatch; unrelated model internals." }, { "left": "issue:43295", "right": "issue:45834", "accept": false, - "reason": "Both mention processor/tokenizer behavior, but one is a regression in custom code and the other is an OCR length/index bug in Kosmos2.5." + "reason": "Both are regression-style correctness issues, but one is processor/tokenizer access in custom code and the other is OCR bbox indexing on long inputs." }, { "left": "issue:43540", "right": "issue:46294", "accept": false, - "reason": "Both are video-related Qwen issues, but one is a Qwen3OmniMoe processing error and the other is a chat-template token/feature mismatch in Qwen3.5-9B." + "reason": "Both involve Qwen video inputs, but one is a ValueError in Qwen3OmniMoe video processing and the other is a chat-template token/feature count mismatch." }, { "left": "issue:43526", "right": "issue:43881", "accept": false, - "reason": "BEiT reduce_labels bug is unrelated to glm-4v-9b loading failure." + "reason": "One is a BeitImageProcessorFast label-reduction bug, the other is a glm-4v-9b load failure; different components and symptoms." }, { "left": "issue:43550", "right": "issue:43927", "accept": false, - "reason": "torch.compile+SDPA failure in Bamba is unrelated to DiaConfig losing custom token IDs after save/load." + "reason": "One is Bamba-9B-v2 torch.compile + SDPA failure, the other is DiaConfig save/load losing custom token IDs; unrelated." }, { "left": "issue:43550", "right": "issue:43819", "accept": false, - "reason": "These are different model-path failures: compile/SDPA vs DAC latent reconstruction." + "reason": "Different models and failure modes: compile/SDPA attention bug versus DAC latent consistency bug." }, { "left": "issue:44488", "right": "issue:47020", "accept": false, - "reason": "Both are loading/tokenizer issues, but they affect different classes and failure modes: model loading vs Unigram init with dict vocab." + "reason": "Both are tokenizer/loading issues, but one is a model failing to load and the other is XLMRobertaTokenizer passing the wrong type to Unigram." }, { "left": "issue:44062", "right": "issue:44488", "accept": false, - "reason": "AddedToken special-argument duplication is not the same as the cjvt/sleng-bert loading failure." + "reason": "One is an AddedToken argument duplication error, the other is a tokenizer loading failure for a specific model; not the same defect." }, { "left": "issue:43531", "right": "issue:43643", "accept": false, - "reason": "Qwen3-MoE sliding-window behavior and trust_remote_code config field loss are unrelated." + "reason": "Sliding-window behavior in Qwen3-MoE is unrelated to AutoConfig missing fields with trust_remote_code=True." }, { "left": "issue:43653", "right": "issue:43720", "accept": false, - "reason": "BigBird special-token registration and BitNet packed-weight loading are different tokenizer/model-loading bugs." + "reason": "Tokenizer special-token registration and packed-weight unloading are entirely different subsystems." }, { "left": "issue:45072", "right": "issue:47050", "accept": false, - "reason": "Both are dtype mismatches, but they occur in different models and inference/export paths." + "reason": "Both mention dtype mismatches, but one is bfloat16 inference in SwitchTransformers/TimmWrapperModel and the other is fp16 ONNX export attention mismatch." }, { "left": "issue:41720", "right": "issue:44991", "accept": false, - "reason": "Auto device mapping CUDA assert on Qwen3 is unrelated to tokenizer loading for EMBEDDIA/est-roberta." + "reason": "One is a CUDA assert during auto device mapping for Qwen3; the other is a tokenizer loading failure for EMBEDDIA/est-roberta." }, { "left": "issue:30064", "right": "issue:36331", "accept": false, - "reason": "Image segmentation-map processing and Trainer.compute_loss signature mismatch are different issues." + "reason": "An image-processor void segmentation-map bug is unrelated to a custom trainer compute_loss signature regression." }, { "left": "issue:45381", "right": "issue:46693", "accept": false, - "reason": "Video position-id bug and FA2 long-context performance regression are distinct generation-path problems." - }, - { - "left": "issue:44220", - "right": "issue:44661", - "accept": false, - "reason": "Audio fbank feature extraction is unrelated to add-new-model-like failing inside TOKENIZER_MAPPING_NAMES." - }, - { - "left": "issue:44821", - "right": "issue:45084", - "accept": false, - "reason": "AutoImageProcessor URL loading and template-node compilation are unrelated." - }, - { - "left": "issue:43611", - "right": "issue:46665", - "accept": false, - "reason": "Both involve loading behavior, but base_model_prefix breakage is a different regression than from_pretrained inside Python classes." - }, - { - "left": "issue:46829", - "right": "issue:46832", - "accept": false, - "reason": "Different model-specific forward bugs: VideoPrism output fields vs RT-DETR feature-level mismatch." - }, - { - "left": "issue:44186", - "right": "issue:44898", - "accept": false, - "reason": "LayoutLMv2 tokenizer crashes and Perceiver resolution/interpolation failure are unrelated vision/tokenizer issues." - }, - { - "left": "issue:43540", - "right": "issue:44008", - "accept": false, - "reason": "Different models and bugs: Qwen3OmniMoe video input handling vs Gemma3n variable-name collision in forward()." - }, - { - "left": "issue:43653", - "right": "issue:43819", - "accept": false, - "reason": "BigBird special-token registration and DAC latent mismatch do not share the same code path." - }, - { - "left": "issue:44991", - "right": "issue:47020", - "accept": false, - "reason": "Tokenizer loading for a specific model is not the same as Unigram expecting a sequence during XLMRoberta init." - }, - { - "left": "issue:34634", - "right": "issue:36010", - "accept": false, - "reason": "Bark voice_preset behavior and GenerationMixin import failure are unrelated." + "reason": "One is a video position-id bug for qwen2.5-vl, the other is a long-context FA2 performance regression; different code paths and outcomes." } ] }, @@ -6212,43 +6088,43 @@ "source": "llm", "cache_hit": false, "model": "gpt-5.4-mini?service_tier=flex", - "cluster_id": "cluster-10000-529", + "cluster_id": "cluster-10000-530", "nodes": [ + "issue:34634", + "issue:36010", "issue:39692", "issue:43322", "issue:43335", "issue:43421", - "issue:43550", - "issue:43577", + "issue:43540", "issue:43606", + "issue:43611", "issue:43618", "issue:43643", - "issue:43644", + "issue:43653", "issue:43697", - "issue:43716", - "issue:43749", "issue:43756", - "issue:43782", - "issue:43818", + "issue:43819", "issue:44008", "issue:44186", + "issue:44220", "issue:44442", "issue:44448", - "issue:44479", "issue:44561", + "issue:44661", "issue:44671", "issue:44805", "issue:44811", + "issue:44821", "issue:44843", "issue:44877", + "issue:44898", "issue:44936", - "issue:44977", "issue:44991", "issue:45020", "issue:45072", "issue:45081", - "issue:45183", - "issue:45479", + "issue:45084", "issue:45584", "issue:45685", "issue:45834", @@ -6256,18 +6132,28 @@ "issue:46291", "issue:46294", "issue:46463", - "issue:46752", + "issue:46665", "issue:46829", + "issue:46832", "issue:46858", "issue:46932", - "issue:47337" + "issue:47020" ], "soft_pairs": [ + "issue:44220|issue:44661", + "issue:44821|issue:45084", + "issue:43611|issue:46665", + "issue:46829|issue:46832", + "issue:44186|issue:44898", + "issue:43653|issue:43819", + "issue:43540|issue:44008", + "issue:44991|issue:47020", + "issue:34634|issue:36010", "issue:44843|issue:45081", - "issue:45987|issue:46829", - "issue:44008|issue:46858", "issue:43643|issue:44877", + "issue:45987|issue:46829", "issue:43335|issue:43756", + "issue:44008|issue:46858", "issue:39692|issue:45834", "issue:44561|issue:45020", "issue:44186|issue:44442", @@ -6278,202 +6164,200 @@ "issue:44991|issue:46291", "issue:43322|issue:45685", "issue:45584|issue:45834", - "issue:43421|issue:44811", "issue:43606|issue:45072", - "issue:46294|issue:46463", - "issue:43550|issue:45072", - "issue:46752|issue:47337", - "issue:43716|issue:44008", - "issue:44479|issue:45834", - "issue:43782|issue:44977", - "issue:43644|issue:43749", - "issue:45183|issue:45479", - "issue:43577|issue:43818" + "issue:43421|issue:44811", + "issue:46294|issue:46463" ], - "prepared_review_unit_hash": "8dfbfcef45bb8883965f50511ce99e66b3ac37927fd6b6f904943f9e2a23fde0", + "prepared_review_unit_hash": "dffc3460cdb5c149203c5b27f085b37385d0bc7ee5afe599af16697ea5fb9dda", "component_budget": { - "node_count": 529, - "item_count": 529, - "soft_pair_count": 1953, - "serialized_chars": 1012074, - "estimated_input_tokens": 253019, - "estimated_eval_tokens": 506294 + "node_count": 530, + "item_count": 530, + "soft_pair_count": 1954, + "serialized_chars": 1013004, + "estimated_input_tokens": 253251, + "estimated_eval_tokens": 506758 }, "budget": { - "node_count": 47, - "item_count": 47, - "soft_pair_count": 26, - "serialized_chars": 35369, - "estimated_input_tokens": 8843, - "estimated_eval_tokens": 17942 + "node_count": 48, + "item_count": 48, + "soft_pair_count": 27, + "serialized_chars": 36051, + "estimated_input_tokens": 9013, + "estimated_eval_tokens": 18282 }, "overflow_policy": "truncate_then_skip", "trimmed": false, "aggressively_trimmed": false, "split": true, "analyst_result": { - "summary": "This cluster is noisy: the items span many unrelated model-, tokenizer-, cache-, and runtime-specific regressions. None of the soft pairs look like the same underlying bug/change, so they should stay separate.", - "confidence": 0.71, - "canonical_issue_reason": "Issue 45020 is the broadest umbrella item here, covering a general class of recent `remote_code` breakages rather than a narrow model-specific symptom.", + "summary": "This cluster is noisy and mostly heterogeneous: the paired items share broad HF/Transformers vocabulary, but each points to a different concrete bug path. I did not find any soft pair that looks like the same underlying issue.", + "confidence": 0.84, + "canonical_issue_reason": "issue:45020 is the broadest umbrella-style report in the set and best represents the general v5/remote_code regression theme, even though the cluster overall is not a true duplicate group.", "canonical_pr_reason": null, - "best_issue_reason": "45020 is the best representative anchor for triage because it is the most general and closest to the few other `remote_code`-related reports in the set.", + "best_issue_reason": "issue:45020 has the widest scope and highest chance of serving as a central tracker for adjacent loading/regression problems.", "best_pr_reason": null, "soft_edge_verdicts": [ { - "left": "issue:44843", - "right": "issue:45081", + "left": "issue:44220", + "right": "issue:44661", "accept": false, - "reason": "Both involve Mistral regex handling, but one is an offline `model_info()` call in `AutoTokenizer`, while the other is a `backend_tokenizer` attribute crash in `_patch_mistral_regex`; different failure points and fixes." + "reason": "Different subsystems and failure modes: audio feature extraction vs tokenizer-mapping/model-addition logic." }, { - "left": "issue:45987", - "right": "issue:46829", + "left": "issue:44821", + "right": "issue:45084", "accept": false, - "reason": "`StaticCache.get_seq_length()` returning a tensor is unrelated to VideoPrism misreporting `hidden_states` in `ImageClassifierOutput`." + "reason": "URL image processor loading and template-node compilation are unrelated code paths." }, { - "left": "issue:44008", - "right": "issue:46858", + "left": "issue:43611", + "right": "issue:46665", "accept": false, - "reason": "Gemma3n\u2019s `audio_mel_mask` variable collision is a model-forward bug; `cache.get_seq_length()` returning a device tensor is a cache API/backend issue." + "reason": "Both involve loading, but one is a base_model_prefix v5 breakage and the other is a broader from_pretrained-in-classes inconsistency." }, { - "left": "issue:43643", - "right": "issue:44877", + "left": "issue:46829", + "right": "issue:46832", "accept": false, - "reason": "`trust_remote_code=True` missing fields and strict config rejection of `granite_speech` are different config-loading problems with different root causes." + "reason": "Different model families and bugs: output field mismatch vs RT-DETR forward shape/type error." }, { - "left": "issue:43335", - "right": "issue:43756", + "left": "issue:44186", + "right": "issue:44898", "accept": false, - "reason": "One is sparse-layer construction when `num_sparse_encoder_layers=0`; the other is a RoPE-layer count mismatch in Smollm3." + "reason": "Tokenizer padding/NER crash is unrelated to Perceiver image-classification interpolation behavior." }, { - "left": "issue:39692", - "right": "issue:45834", + "left": "issue:43653", + "right": "issue:43819", "accept": false, - "reason": "SigLIP2 doc/example errors and Kosmos2.5 long-OCR index errors are unrelated artifacts and code paths." + "reason": "Tokenizer special-token registration and DAC latent conversion are separate components with no shared code path." }, { - "left": "issue:44561", - "right": "issue:45020", + "left": "issue:43540", + "right": "issue:44008", "accept": false, - "reason": "Both mention `remote_code`, but #44561 is specifically the removal of `is_torch_fx_available`, while #45020 is a broader recent-regression report across models." + "reason": "Distinct multimodal model bugs: video input validation vs Gemma3n forward variable collision." }, { - "left": "issue:44186", - "right": "issue:44442", + "left": "issue:44991", + "right": "issue:47020", "accept": false, - "reason": "LayoutLMv2 tokenizer crashes on NER/padding are not the same as AutoTokenizer failing to load FastSpeech2ConformerTokenizer." + "reason": "Both are tokenizer-loading problems, but they affect different tokenizer classes and likely different constructor paths." }, { - "left": "issue:44448", - "right": "issue:44671", + "left": "issue:34634", + "right": "issue:36010", "accept": false, - "reason": "Both are v4-v5 output differences, but they affect different model families and different behaviors; not the same code-path bug." + "reason": "BarkProcessor voice preset behavior is unrelated to the GenerationMixin import error." }, { - "left": "issue:45987", - "right": "issue:46932", + "left": "issue:44843", + "right": "issue:45081", "accept": false, - "reason": "`get_seq_length()` return-type issues are unrelated to Gemma4 quantized vision/audio embedder dtype casting." + "reason": "Same helper name, but different concrete failures: offline hub access vs backend_tokenizer attribute crash." }, { - "left": "issue:43618", - "right": "issue:43697", + "left": "issue:43643", + "right": "issue:44877", "accept": false, - "reason": "Missing `CLIPOutput.attentions` and RTDetrV2 output drift are separate output-shape/return-value regressions." + "reason": "Different config-loading failures: missing fields from trust_remote_code vs strict config rejecting granite_speech." }, { - "left": "issue:44805", - "right": "issue:44936", + "left": "issue:45987", + "right": "issue:46829", "accept": false, - "reason": "A mask-shape `IndexError` during model execution is unrelated to `trainer.evaluate()` failing after `trainer.train()`." + "reason": "StaticCache sequence-length contract bug is unrelated to VideoPrism output tensor assignment." }, { - "left": "issue:44991", - "right": "issue:46291", + "left": "issue:43335", + "right": "issue:43756", "accept": false, - "reason": "Tokenizer loading failure for a specific model and `transformers.pipelines` import failing due to `kernels` versioning are different dependency issues." + "reason": "One is sparse-layer creation from zero sparse layers; the other is RoPE layer dropping in Smollm3." }, { - "left": "issue:43322", - "right": "issue:45685", + "left": "issue:44008", + "right": "issue:46858", "accept": false, - "reason": "Llava Next segfault/loading issue and MPS `histogram_mps` integer error in MOE are unrelated backend/model problems." + "reason": "Both mention tensor/device issues, but they arise in different models and different forward/cache logic." }, { - "left": "issue:45584", + "left": "issue:39692", "right": "issue:45834", "accept": false, - "reason": "Whisper empty-transcription generation and Kosmos2.5 long OCR input indexing are different input-handling bugs." + "reason": "Documentation example errors and long-ocr index error are different problems in different models." }, { - "left": "issue:43421", - "right": "issue:44811", + "left": "issue:44561", + "right": "issue:45020", "accept": false, - "reason": "Runtime special-token/post-processor update behavior is distinct from Whisper `batch_decode()` ignoring `skip_special_tokens`." + "reason": "Both are v5 regressions, but one is specifically is_torch_fx_available removal while the other is a broader remote_code breakage report." }, { - "left": "issue:43606", - "right": "issue:45072", + "left": "issue:44186", + "right": "issue:44442", "accept": false, - "reason": "CPU offload device mismatch in suno/bark-small and bfloat16 dtype mismatches in SwitchTransformers/TimmWrapperModel are not the same defect." + "reason": "Different tokenizer/model classes and distinct crash conditions." }, { - "left": "issue:46294", - "right": "issue:46463", + "left": "issue:44448", + "right": "issue:44671", "accept": false, - "reason": "Qwen3.5-9B video chat-template feature mismatch and T5 decoder-input validation are unrelated bugs." + "reason": "Version-v4/v5 output drift in different models; no shared code path is evident." }, { - "left": "issue:43550", - "right": "issue:45072", + "left": "issue:45987", + "right": "issue:46932", "accept": false, - "reason": "`torch.compile`/SDPA failure in Bamba-9B-v2 and dtype-mismatch inference issues are different concrete runtime paths." + "reason": "StaticCache length return type and Gemma4 dtype casting are unrelated issues." }, { - "left": "issue:46752", - "right": "issue:47337", + "left": "issue:43618", + "right": "issue:43697", "accept": false, - "reason": "Empty-list `apply_chat_template` crash and flaky bbox-range `IndexError` are unrelated." + "reason": "CLIP attentions field assignment and RTDetrV2 output divergence are separate model-output bugs." }, { - "left": "issue:43716", - "right": "issue:44008", + "left": "issue:44805", + "right": "issue:44936", "accept": false, - "reason": "Mistral-3 image preprocessor/model dtype mismatch and Gemma3n variable collision in `forward()` are different model bugs." + "reason": "Mask shape IndexError and trainer evaluate-after-train failure are unrelated training/runtime problems." }, { - "left": "issue:44479", - "right": "issue:45834", + "left": "issue:44991", + "right": "issue:46291", "accept": false, - "reason": "Qwen video-input regression and Kosmos2.5 OCR index error do not share the same underlying change." + "reason": "Tokenizer loading failure and pipelines import failure are different entrypoints and failure modes." }, { - "left": "issue:43782", - "right": "issue:44977", + "left": "issue:43322", + "right": "issue:45685", "accept": false, - "reason": "Qwen3VL `weight_only=True` loading error and Qwen3.5 flash-attention generation failure are separate issues in different models/code paths." + "reason": "Llava Next segfault and MPS histogram/int dtype error are not the same bug." }, { - "left": "issue:43644", - "right": "issue:43749", + "left": "issue:45584", + "right": "issue:45834", "accept": false, - "reason": "Non-persistent buffer junk in Transformers v5 and broken FSDP CPU RAM efficient loading are unrelated initialization/loading regressions." + "reason": "Both are multimodal generation issues, but empty-transcription Whisper behavior and Kosmos2.5 long OCR indexing are different code paths." }, { - "left": "issue:45183", - "right": "issue:45479", + "left": "issue:43606", + "right": "issue:45072", "accept": false, - "reason": "Input validation error-message issues are not the same as the zero-loss bug from `problem_type=\"single_label_classification\"` with `num_labels=1`." + "reason": "CPU offload device mismatch and bfloat16 dtype mismatch are separate backend-specific bugs." }, { - "left": "issue:43577", - "right": "issue:43818", + "left": "issue:43421", + "right": "issue:44811", + "accept": false, + "reason": "Runtime special-token/post-processor updates and Whisper batch_decode skip_special_tokens behavior are distinct tokenizer/processor issues." + }, + { + "left": "issue:46294", + "right": "issue:46463", "accept": false, - "reason": "Blip2 dtype fields staying float32 and Video-LLaVA temporal-attention/weight-sharing issues are different multimodal-model defects." + "reason": "Qwen video chat-template mismatch and T5 decoder-input validation are unrelated models and validations." } ] }, @@ -6495,51 +6379,50 @@ "source": "llm", "cache_hit": false, "model": "gpt-5.4-mini?service_tier=flex", - "cluster_id": "cluster-10000-529", + "cluster_id": "cluster-10000-530", "nodes": [ "issue:28218", "issue:29159", "issue:34634", - "issue:36010", "issue:36331", "issue:42175", "issue:42915", - "issue:43262", "issue:43388", "issue:43540", + "issue:43550", "issue:43577", "issue:43582", + "issue:43644", "issue:43645", - "issue:43653", + "issue:43716", "issue:43720", "issue:43742", "issue:43749", "issue:43782", - "issue:43819", + "issue:43818", "issue:43825", "issue:43931", "issue:43994", + "issue:44008", "issue:44291", - "issue:44589", + "issue:44479", "issue:44610", "issue:44671", - "issue:44912", - "issue:44964", + "issue:44977", "issue:45020", - "issue:45042", "issue:45072", "issue:45081", - "issue:45325", + "issue:45183", "issue:45381", + "issue:45479", "issue:45563", "issue:45701", + "issue:45834", "issue:46032", "issue:46512", "issue:46665", - "issue:46697", - "issue:46728", + "issue:46752", "issue:46846", - "issue:47217", "issue:47254", "issue:47276", "issue:47332", @@ -6547,12 +6430,20 @@ "issue:47337" ], "soft_pairs": [ + "issue:43550|issue:45072", + "issue:46752|issue:47337", + "issue:44479|issue:45834", + "issue:43716|issue:44008", + "issue:43782|issue:44977", + "issue:43644|issue:43749", + "issue:43577|issue:43818", "issue:28218|issue:29159", - "issue:45563|issue:47276", + "issue:45183|issue:45479", "issue:43582|issue:46512", - "issue:43825|issue:43994", + "issue:45563|issue:47276", "issue:43931|issue:46846", "issue:44610|issue:45072", + "issue:43825|issue:43994", "issue:43540|issue:45381", "issue:43749|issue:44291", "issue:34634|issue:36331", @@ -6560,228 +6451,217 @@ "issue:43720|issue:45072", "issue:47276|issue:47337", "issue:47254|issue:47333", - "issue:43388|issue:46032", "issue:42175|issue:45081", + "issue:43388|issue:46032", "issue:44671|issue:45701", "issue:43742|issue:45020", "issue:47254|issue:47332", "issue:43577|issue:43582", - "issue:42915|issue:47333", - "issue:46697|issue:46728", - "issue:44589|issue:45042", - "issue:43782|issue:45325", - "issue:43262|issue:47217", - "issue:36010|issue:43645", - "issue:43653|issue:44912", - "issue:43262|issue:43819", - "issue:44964|issue:47337" + "issue:42915|issue:47333" ], - "prepared_review_unit_hash": "f78a36a892515fc6b22735e481557d24256891840dce75c309ec88cceb28cc00", + "prepared_review_unit_hash": "8f595d4afb45226d037ef6bdb7af21ad21c45628071c7db0a349d4b4cb6ef2fc", "component_budget": { - "node_count": 529, - "item_count": 529, - "soft_pair_count": 1953, - "serialized_chars": 1012074, - "estimated_input_tokens": 253019, - "estimated_eval_tokens": 506294 + "node_count": 530, + "item_count": 530, + "soft_pair_count": 1954, + "serialized_chars": 1013004, + "estimated_input_tokens": 253251, + "estimated_eval_tokens": 506758 }, "budget": { - "node_count": 48, - "item_count": 48, + "node_count": 47, + "item_count": 47, "soft_pair_count": 28, - "serialized_chars": 36297, - "estimated_input_tokens": 9075, - "estimated_eval_tokens": 18406 + "serialized_chars": 35750, + "estimated_input_tokens": 8938, + "estimated_eval_tokens": 18132 }, "overflow_policy": "truncate_then_skip", "trimmed": false, "aggressively_trimmed": false, "split": true, "analyst_result": { - "summary": "Revised conservatively: 28218\u219429159 is the only clear duplicate pair here. The 44671\u219445701 link is too speculative from the titles alone and should be rejected.", - "confidence": 0.82, - "canonical_issue_reason": "Issue 45701 is the broadest tokenizer-regression umbrella in the packet, while 28218\u219429159 is the clearest concrete duplicate pair underneath it.", + "summary": "The cluster is heterogeneous: the soft pairs are mostly thematically similar only at a broad subsystem level, but they do not describe the same concrete bug or change. I would not merge any of the candidate edges.", + "confidence": 0.92, + "canonical_issue_reason": "No single canonical issue stands out. The set spans unrelated failures across tokenization, model loading, dtype handling, video inputs, training, offloading, and warnings, so the soft matches are too weak for duplicate triage.", "canonical_pr_reason": null, - "best_issue_reason": "45701 is the best global issue umbrella for this small tokenizer cluster because it is the most general report and can absorb closely related tokenizer-regression duplicates.", + "best_issue_reason": "There is no clear global best issue for deduplication because the cluster is not a true duplicate set. If forced to pick one representative, the most general/central wording is still not enough to subsume the rest.", "best_pr_reason": null, "soft_edge_verdicts": [ + { + "left": "issue:43550", + "right": "issue:45072", + "accept": false, + "reason": "Different bugs: torch.compile/SDPA failure vs dtype mismatch during inference." + }, + { + "left": "issue:46752", + "right": "issue:47337", + "accept": false, + "reason": "Empty chat template input crash is unrelated to the bbox range flaky test." + }, + { + "left": "issue:44479", + "right": "issue:45834", + "accept": false, + "reason": "Video-input regression across Qwen VL models is unrelated to Kosmos2.5 long OCR indexing." + }, + { + "left": "issue:43716", + "right": "issue:44008", + "accept": false, + "reason": "Different root causes: image-preprocessor dtype mismatch vs variable name collision in Gemma3nModel.forward()." + }, + { + "left": "issue:43782", + "right": "issue:44977", + "accept": false, + "reason": "Qwen3VL weight_only load error is not the same as Qwen3.5 flash-attention generation failure." + }, + { + "left": "issue:43644", + "right": "issue:43749", + "accept": false, + "reason": "Both concern loading, but one is junk-filled buffers and the other is FSDP CPU RAM efficient loading." + }, + { + "left": "issue:43577", + "right": "issue:43818", + "accept": false, + "reason": "Blip2 dtype propagation bug is unrelated to missing temporal attention/shared weights in Video-LLaVA." + }, { "left": "issue:28218", "right": "issue:29159", - "accept": true, - "reason": "Both describe the same tokenizer-behavior regression around added-token spacing and slow/fast tokenizer inconsistency." + "accept": false, + "reason": "Both are tokenizer-related, but one is an added-token spacing bug and the other is slow/fast inconsistency more broadly." }, { - "left": "issue:45563", - "right": "issue:47276", + "left": "issue:45183", + "right": "issue:45479", "accept": false, - "reason": "Both are warning-related, but they are different warnings in different generation paths." + "reason": "Input validation error messaging is unrelated to the single-label classification zero-loss bug." }, { "left": "issue:43582", "right": "issue:46512", "accept": false, - "reason": "Both mention dtype-ish behavior, but one is an Apple Silicon warmup TypeError and the other is stale config dtype after construction." + "reason": "Apple Silicon warmup TypeError is unrelated to stale model.config.dtype after from_config(dtype=...)." }, { - "left": "issue:43825", - "right": "issue:43994", + "left": "issue:45563", + "right": "issue:47276", "accept": false, - "reason": "One is an error-message/pipeline-support issue; the other is a model output quality/loading issue." + "reason": "Both are warning-message issues, but they refer to different warnings in different generation paths." }, { "left": "issue:43931", "right": "issue:46846", "accept": false, - "reason": "Both involve shape problems, but they affect different models and different root causes." + "reason": "Shape-mismatch loading failure is not the same as missing base_model_tp_plan entries causing TP OOM/shape errors." }, { "left": "issue:44610", "right": "issue:45072", "accept": false, - "reason": "Different subsystems: image preprocessing size mismatch vs dtype mismatch during inference." + "reason": "Processor image-size mismatch is unrelated to bfloat16 dtype mismatches in inference wrappers." + }, + { + "left": "issue:43825", + "right": "issue:43994", + "accept": false, + "reason": "Unsupported translation-task messaging is unrelated to SigLIP2 nonsense outputs." }, { "left": "issue:43540", "right": "issue:45381", "accept": false, - "reason": "Both are Qwen video-input issues, but the failure modes and model variants differ." + "reason": "Both touch video inputs, but the concrete failures and affected models differ." }, { "left": "issue:43749", "right": "issue:44291", "accept": false, - "reason": "Different loading failures with different triggers and code paths." + "reason": "FSDP CPU RAM efficient loading failure is unrelated to init_empty_weights receiving an unexpected argument." }, { "left": "issue:34634", "right": "issue:36331", "accept": false, - "reason": "Unrelated bugs: BarkProcessor voice_preset vs Trainer.compute_loss signature change." + "reason": "BarkProcessor voice preset handling and CustomTrainer.compute_loss signature are unrelated." }, { "left": "issue:43645", "right": "issue:46665", "accept": false, - "reason": "Both concern custom-model loading, but the notebook/Jupyter problem is distinct from from_pretrained-in-classes behavior." + "reason": "Notebook custom-model initialization breakage is not the same as from_pretrained behavior inside Python classes." }, { "left": "issue:43720", "right": "issue:45072", "accept": false, - "reason": "Packed-weight loading failure and dtype mismatch are different bugs." + "reason": "Packed-weight unpacking during accelerate loading is a different failure mode from dtype mismatch inference bugs." }, { "left": "issue:47276", "right": "issue:47337", "accept": false, - "reason": "A stale warning and a flaky bbox-range test are not the same underlying issue." + "reason": "A stale synced_gpus warning is unrelated to the bbox coordinate flaky test." }, { "left": "issue:47254", "right": "issue:47333", "accept": false, - "reason": "Separate offloading/checkpointing failures with different symptoms." + "reason": "Checkpointing/ZeRO-3 failure and disk-offloaded Qwen3 save failure are different code paths and symptoms." }, { - "left": "issue:43388", - "right": "issue:46032", + "left": "issue:42175", + "right": "issue:45081", "accept": false, - "reason": "Metric gathering on tuple labels is unrelated to Mamba2 use_cache correctness." + "reason": "Backend packaging issue is unrelated to the Mistral tokenizer regex patch crash." }, { - "left": "issue:42175", - "right": "issue:45081", + "left": "issue:43388", + "right": "issue:46032", "accept": false, - "reason": "Packaging/backend dependency issue vs tokenizer patch crash; no shared bug." + "reason": "Metric-gathering label truncation is unrelated to Mamba2Mixer use_cache correctness." }, { "left": "issue:44671", "right": "issue:45701", "accept": false, - "reason": "The CamemBERT masked-LM prediction issue is too specific and not clearly the same tokenizer regression as the generic version-changes-tokenization report." + "reason": "CamemBERT masked-LM regression is not the same as version-dependent tokenization changes." }, { "left": "issue:43742", "right": "issue:45020", "accept": false, - "reason": "A specific MobileLLM loading key error is not the same as the broader remote_code regression." + "reason": "MobileLLM key error and remote_code breakage are different loading failures." }, { "left": "issue:47254", "right": "issue:47332", "accept": false, - "reason": "Separate offloading/checkpointing save/load problems." + "reason": "Gradient-checkpointing checkpoint errors are unrelated to saving disk-offloaded Qwen3-VL models." }, { "left": "issue:43577", "right": "issue:43582", "accept": false, - "reason": "Both mention dtype, but they are different failures in different subsystems." + "reason": "Blip2 dtype issues and Apple Silicon warmup TypeError are unrelated." }, { "left": "issue:42915", "right": "issue:47333", "accept": false, - "reason": "Different Qwen model failures with different causes." - }, - { - "left": "issue:46697", - "right": "issue:46728", - "accept": false, - "reason": "Both are image-size-related, but the exact padding/counting bugs differ." - }, - { - "left": "issue:44589", - "right": "issue:45042", - "accept": false, - "reason": "Float8 storage error and torchvision dependency requirement are unrelated." - }, - { - "left": "issue:43782", - "right": "issue:45325", - "accept": false, - "reason": "One is a weight_only load error; the other is a rope-index/position-id scaling bug." - }, - { - "left": "issue:43262", - "right": "issue:47217", - "accept": false, - "reason": "Audio chat-template sampling-rate handling and VL image-placeholder handling are different problems." - }, - { - "left": "issue:36010", - "right": "issue:43645", - "accept": false, - "reason": "GenerationMixin import failure is unrelated to notebook custom-model initialization issues." - }, - { - "left": "issue:43653", - "right": "issue:44912", - "accept": false, - "reason": "Tokenizer special-token registration and MXFP4 quantization fallback are unrelated." - }, - { - "left": "issue:43262", - "right": "issue:43819", - "accept": false, - "reason": "Audio processor sampling-rate handling and DAC latent reconstruction mismatch are different code paths." - }, - { - "left": "issue:44964", - "right": "issue:47337", - "accept": false, - "reason": "A model-loading regression and a flaky bbox test are unrelated." + "reason": "Qwen3Moe FP8 config failure is unrelated to disk-offloaded Qwen3 save errors." } ] }, - "evaluator_result": { - "accept": false, - "feedback": "The overall summary is reasonable, and the 28218\u219429159 duplicate call is well grounded. However, the accepted 44671\u219445701 edge is too assertive for the packet: 44671 is about incorrect masked LM predictions, and the title alone does not clearly support a tokenizer-regression duplicate claim. Because the soft-edge verdicts must stay conservative, I can\u2019t accept this as written." - }, - "evaluator_used": true, - "retried": true, - "accepted_nontrivial_soft_edge": true, + "evaluator_result": null, + "evaluator_used": false, + "retried": false, + "accepted_nontrivial_soft_edge": false, "error_kind": null, "error_message": null }, @@ -6796,13 +6676,13 @@ "source": "llm", "cache_hit": false, "model": "gpt-5.4-mini?service_tier=flex", - "cluster_id": "cluster-10000-529", + "cluster_id": "cluster-10000-530", "nodes": [ "issue:36010", "issue:36331", "issue:42175", - "issue:42915", "issue:43257", + "issue:43262", "issue:43329", "issue:43452", "issue:43526", @@ -6810,10 +6690,10 @@ "issue:43550", "issue:43582", "issue:43638", - "issue:43643", - "issue:43696", + "issue:43645", + "issue:43653", "issue:43697", - "issue:43746", + "issue:43782", "issue:43819", "issue:43825", "issue:43827", @@ -6822,228 +6702,242 @@ "issue:43927", "issue:44060", "issue:44368", - "issue:44811", - "issue:44821", + "issue:44589", "issue:44857", "issue:44912", "issue:44936", + "issue:44964", "issue:44977", "issue:44991", "issue:45003", "issue:45042", "issue:45081", - "issue:45084", + "issue:45325", "issue:45372", "issue:45406", - "issue:45538", - "issue:45759", - "issue:45859", "issue:46175", "issue:46209", "issue:46285", "issue:46291", "issue:46326", + "issue:46697", + "issue:46728", "issue:46932", - "issue:47050" + "issue:47217", + "issue:47337" ], "soft_pairs": [ + "issue:46697|issue:46728", + "issue:44589|issue:45042", + "issue:43262|issue:47217", + "issue:43782|issue:45325", + "issue:36010|issue:43645", + "issue:43653|issue:44912", + "issue:43262|issue:43819", "issue:46175|issue:46932", - "issue:43582|issue:44857", + "issue:44964|issue:47337", "issue:45003|issue:46285", + "issue:43582|issue:44857", "issue:36010|issue:36331", "issue:43825|issue:43827", "issue:43531|issue:45042", "issue:43638|issue:44936", - "issue:45372|issue:46326", "issue:43550|issue:44912", + "issue:45372|issue:46326", "issue:42175|issue:46291", "issue:43452|issue:44991", "issue:43257|issue:43329", "issue:45081|issue:45406", "issue:43819|issue:43927", - "issue:43526|issue:43697", "issue:43906|issue:46209", + "issue:43526|issue:43697", "issue:44060|issue:44368", - "issue:43854|issue:44977", - "issue:43746|issue:44811", - "issue:42915|issue:44821", - "issue:43257|issue:43696", - "issue:45084|issue:45538", - "issue:43643|issue:45759", - "issue:45859|issue:47050" + "issue:43854|issue:44977" ], - "prepared_review_unit_hash": "7075b69adedf75b3923deee7d2f0ac26af95ad6b97cb62d8f52ed6ffb85c09a4", + "prepared_review_unit_hash": "7dd8c8d3d2ad771da3e9e30641da6b27f88cc8e3078047e351e81d69442430a9", "component_budget": { - "node_count": 529, - "item_count": 529, - "soft_pair_count": 1953, - "serialized_chars": 1012074, - "estimated_input_tokens": 253019, - "estimated_eval_tokens": 506294 + "node_count": 530, + "item_count": 530, + "soft_pair_count": 1954, + "serialized_chars": 1013004, + "estimated_input_tokens": 253251, + "estimated_eval_tokens": 506758 }, "budget": { "node_count": 47, "item_count": 47, - "soft_pair_count": 24, - "serialized_chars": 34750, - "estimated_input_tokens": 8688, - "estimated_eval_tokens": 17632 + "soft_pair_count": 26, + "serialized_chars": 35500, + "estimated_input_tokens": 8875, + "estimated_eval_tokens": 18006 }, "overflow_policy": "truncate_then_skip", "trimmed": false, "aggressively_trimmed": false, "split": true, "analyst_result": { - "summary": "These items are mostly unrelated standalone issues; the soft pairs only show superficial topical overlap, not the same underlying bug. No PRs were provided.", - "confidence": 0.91, - "canonical_issue_reason": "issue:43825 is the strongest representative: a concrete, user-facing regression with clear scope and inbound references, whereas the other issues are narrower or more incidental.", + "summary": "The cluster is broadly heterogeneous: the soft-similarity pairs mostly share a subsystem name or API surface, but they do not look like the same underlying bug. No PRs are present.", + "confidence": 0.93, + "canonical_issue_reason": "No clear canonical issue: the issues span unrelated failure modes across tokenization, vision, training, quantization, and docs, so there is no single shared root cause.", "canonical_pr_reason": null, - "best_issue_reason": "issue:43825 is the best overall issue candidate for the cluster because it is well-scoped, clearly described, and has the most evidence of being a reference point for related reports.", + "best_issue_reason": "None of the issues is a strong cluster anchor; even the more discussed items target different code paths and should not be treated as duplicates.", "best_pr_reason": null, "soft_edge_verdicts": [ + { + "left": "issue:46697", + "right": "issue:46728", + "accept": false, + "reason": "Both are image-processor bugs, but one is over-padding in Swin2SR and the other is a wrong patch-count calculation in Aria; different code paths and symptoms." + }, + { + "left": "issue:44589", + "right": "issue:45042", + "accept": false, + "reason": "One is an FP8 storage/type error, the other is an image-processor torchvision dependency issue; unrelated failures." + }, + { + "left": "issue:43262", + "right": "issue:47217", + "accept": false, + "reason": "Audio sampling-rate defaults and VL image-placeholder handling are different subsystems with different bug mechanisms." + }, + { + "left": "issue:43782", + "right": "issue:45325", + "accept": false, + "reason": "Different Qwen2.5-VL/Qwen3VL model issues: one is a from_pretrained weight_only error, the other is a rope index scaling bug." + }, + { + "left": "issue:36010", + "right": "issue:43645", + "accept": false, + "reason": "Both involve Transformers v5 breakage, but one is an import-path regression and the other is notebook custom-model initialization; not the same bug." + }, + { + "left": "issue:43653", + "right": "issue:44912", + "accept": false, + "reason": "Tokenizer special-token registration is unrelated to MXFP4 quantization fallback." + }, + { + "left": "issue:43262", + "right": "issue:43819", + "accept": false, + "reason": "Audio chat-template sampling rate and DAC latent reconstruction are unrelated features." + }, { "left": "issue:46175", "right": "issue:46932", "accept": false, - "reason": "Both involve Gemma4, but one is about a huge one-hot position embedding and the other is about dtype casting in quantized embedders; different bugs." + "reason": "Both are Gemma4-related, but one is a huge one-hot position-embedding implementation and the other is a dtype-casting fix for quantized embedders." }, { - "left": "issue:43582", - "right": "issue:44857", + "left": "issue:44964", + "right": "issue:47337", "accept": false, - "reason": "Completely different failures: AppleSilicon allocator warmup TypeError vs CUDA AMP loss crash in LwDetr." + "reason": "A multimodal model load failure and a flaky bbox-range test are unrelated." }, { "left": "issue:45003", "right": "issue:46285", "accept": false, - "reason": "sys.modules access safety vs CLIP text model structure inconsistency are unrelated code paths." + "reason": "Unsafe sys.modules access and CLIP text-model structure changes are different issues." + }, + { + "left": "issue:43582", + "right": "issue:44857", + "accept": false, + "reason": "Apple Silicon allocator warmup TypeError and CUDA AMP loss crash are unrelated runtime problems." }, { "left": "issue:36010", "right": "issue:36331", "accept": false, - "reason": "Importing GenerationMixin and CustomTrainer.compute_loss are unrelated issues." + "reason": "ImportError for GenerationMixin and CustomTrainer.compute_loss signature mismatch are separate regressions." }, { "left": "issue:43825", "right": "issue:43827", "accept": false, - "reason": "Both mention v5 pipeline removals, but one is an error message bug and the other is stale docs; related topic, not the same defect." + "reason": "One is a bad runtime error message, the other is stale docs; related topic but not the same concrete bug." }, { "left": "issue:43531", "right": "issue:45042", "accept": false, - "reason": "Qwen3-MoE sliding-window handling and PIL/torchvision backend requirements are unrelated." + "reason": "Qwen3-MoE sliding-window behavior and PIL backend torchvision dependency are unrelated." }, { "left": "issue:43638", "right": "issue:44936", "accept": false, - "reason": "Both touch trainer behavior, but zero3/non-pretrained Bert indexing and post-train evaluate failure are different problems." + "reason": "DeepSpeed ZeRO-3 initialization failure and evaluate-after-train failure are different trainer bugs." }, { - "left": "issue:45372", - "right": "issue:46326", + "left": "issue:43550", + "right": "issue:44912", "accept": false, - "reason": "Gemma4/mistral_common import failure and continue_final_message behavior are distinct tokenizer/processor issues." + "reason": "torch.compile with SDPA and MXFP4 quantization fallback are different model/runtime paths." }, { - "left": "issue:43550", - "right": "issue:44912", + "left": "issue:45372", + "right": "issue:46326", "accept": false, - "reason": "torch.compile + SDPA on Bamba is unrelated to MXFP4 quantization fallback on git-oss-20b." + "reason": "Gemma4 processor loading import failure and mistral_common continue_final_message behavior are unrelated." }, { "left": "issue:42175", "right": "issue:46291", "accept": false, - "reason": "Packaging missing TensorFlow in the torch extra is unrelated to pipelines import breaking on kernels v0.15." + "reason": "Packaging/backend dependency omission and pipelines import failure with kernels v0.15 are different dependency issues." }, { "left": "issue:43452", "right": "issue:44991", "accept": false, - "reason": "Both are loading/tokenizer reports, but one is gguf/from_pretrained handling and the other is a specific tokenizer load regression for a model repo." + "reason": "gguf_file handling and tokenizer loading for a specific model are separate from different loaders." }, { "left": "issue:43257", "right": "issue:43329", "accept": false, - "reason": "Qwen3 MOE weight conversion with accelerate/deepspeed is unrelated to multimodal token counting in the video branch." + "reason": "Qwen3 MoE weight conversion and undefined video-token helper usage are unrelated bugs." }, { "left": "issue:45081", "right": "issue:45406", "accept": false, - "reason": "Mistral regex patch crash and Gemma4 processor missing _tokenizer are different loading bugs." + "reason": "Both hit tokenizer/processor loading, but one is a Mistral regex patch crash and the other is a Gemma4Processor missing tokenizer attribute." }, { "left": "issue:43819", "right": "issue:43927", "accept": false, - "reason": "DAC latent decoding mismatch and DiaConfig losing custom token IDs are unrelated model/config issues." + "reason": "DAC latent reconstruction mismatch and DiaConfig token-id persistence loss are unrelated model/config bugs." }, { - "left": "issue:43526", - "right": "issue:43697", + "left": "issue:43906", + "right": "issue:46209", "accept": false, - "reason": "BEiT fast label reduction and RTDetrV2 output differences in v5 do not describe the same bug." + "reason": "A reproduction of an older issue and an FP8 inference failure on B200 are not the same root cause." }, { - "left": "issue:43906", - "right": "issue:46209", + "left": "issue:43526", + "right": "issue:43697", "accept": false, - "reason": "An isolated reproduction of an old issue is not the same as the FP8/B200 inference failure." + "reason": "A BEiT image-processor label reduction bug and an RTDetrV2 output regression are different code paths." }, { "left": "issue:44060", "right": "issue:44368", "accept": false, - "reason": "Both mention tied-weights warnings in Qwen models, but the titles point to different warning logic and different triggers; not enough to treat as the same defect." + "reason": "Both mention tied-weights warnings, but they concern different Qwen model families and different warning causes." }, { "left": "issue:43854", "right": "issue:44977", "accept": false, - "reason": "GLM-4.7-Flash unit-test loading and Qwen3.5 flash-attention generation are unrelated." - }, - { - "left": "issue:43746", - "right": "issue:44811", - "accept": false, - "reason": "PEFT/local checkpoint loading and Whisper batch_decode skip_special_tokens behavior are separate bugs." - }, - { - "left": "issue:42915", - "right": "issue:44821", - "accept": false, - "reason": "Qwen3Moe FP8 backend failure and AutoImageProcessor URL loading are unrelated." - }, - { - "left": "issue:43257", - "right": "issue:43696", - "accept": false, - "reason": "MOE weight conversion with accelerate/deepspeed is unrelated to a plain OOM report." - }, - { - "left": "issue:45084", - "right": "issue:45538", - "accept": false, - "reason": "Template compilation error and CLIPTokenizer model_max_length are unrelated." - }, - { - "left": "issue:43643", - "right": "issue:45759", - "accept": false, - "reason": "Missing fields from trust_remote_code config loading and composite Qwen multimodal config unwrapping are different config-path bugs." - }, - { - "left": "issue:45859", - "right": "issue:47050", - "accept": false, - "reason": "CPU offload incompatibility in Qwen3_5Moe rotary embedding and Canine ONNX dtype mismatch are different issues." + "reason": "GLM-4.7-Flash test loading and Qwen3.5 flash-attention generation issues are unrelated model-specific failures." } ] }, @@ -7065,12 +6959,11 @@ "source": "llm", "cache_hit": false, "model": "gpt-5.4-mini?service_tier=flex", - "cluster_id": "cluster-10000-529", + "cluster_id": "cluster-10000-530", "nodes": [ "issue:28282", "issue:29159", "issue:29870", - "issue:33357", "issue:34567", "issue:34689", "issue:38175", @@ -7083,15 +6976,12 @@ "issue:43329", "issue:43425", "issue:43450", - "issue:43454", "issue:43525", "issue:43526", - "issue:43540", - "issue:43577", - "issue:43650", + "issue:43643", + "issue:43696", "issue:43742", - "issue:43749", - "issue:43906", + "issue:43746", "issue:43931", "issue:44291", "issue:44387", @@ -7100,248 +6990,258 @@ "issue:44589", "issue:44610", "issue:44617", - "issue:44756", "issue:44811", + "issue:44821", "issue:44912", "issue:44987", "issue:45042", "issue:45081", + "issue:45084", "issue:45459", "issue:45538", "issue:45600", + "issue:45759", "issue:45803", + "issue:45859", "issue:46291", - "issue:46701", "issue:47020", + "issue:47050", "issue:47145", "issue:47333" ], "soft_pairs": [ + "issue:42915|issue:44821", + "issue:43746|issue:44811", + "issue:43257|issue:43696", + "issue:45084|issue:45538", + "issue:45859|issue:47050", + "issue:43643|issue:45759", "issue:38175|issue:44610", "issue:42490|issue:43257", "issue:43262|issue:44811", - "issue:43329|issue:45600", "issue:42915|issue:43742", + "issue:43329|issue:45600", "issue:43329|issue:44479", "issue:44291|issue:45042", "issue:43526|issue:44554", "issue:43329|issue:43450", "issue:39692|issue:44912", "issue:34567|issue:40990", - "issue:44387|issue:44617", "issue:29159|issue:34567", + "issue:44387|issue:44617", "issue:44589|issue:47333", "issue:43257|issue:43931", "issue:45081|issue:47020", "issue:44987|issue:45538", "issue:43425|issue:46291", "issue:44912|issue:47145", - "issue:45459|issue:45803", "issue:43525|issue:46291", "issue:28282|issue:29870", - "issue:28282|issue:34689", - "issue:43540|issue:46701", - "issue:43749|issue:44756", - "issue:33357|issue:34567", - "issue:43650|issue:43906", - "issue:43454|issue:43577" + "issue:45459|issue:45803", + "issue:28282|issue:34689" ], - "prepared_review_unit_hash": "8584446607a7c5fcdb563e5cb289312eb6868ed8d4715315270f46c1d5d583ea", + "prepared_review_unit_hash": "6881fd00cbc7e32cc0e7e3a2047c5eb884b303ea908f27c85b8d982a862162ac", "component_budget": { - "node_count": 529, - "item_count": 529, - "soft_pair_count": 1953, - "serialized_chars": 1012074, - "estimated_input_tokens": 253019, - "estimated_eval_tokens": 506294 + "node_count": 530, + "item_count": 530, + "soft_pair_count": 1954, + "serialized_chars": 1013004, + "estimated_input_tokens": 253251, + "estimated_eval_tokens": 506758 }, "budget": { - "node_count": 48, - "item_count": 48, - "soft_pair_count": 28, - "serialized_chars": 36235, - "estimated_input_tokens": 9059, - "estimated_eval_tokens": 18374 + "node_count": 47, + "item_count": 47, + "soft_pair_count": 29, + "serialized_chars": 36095, + "estimated_input_tokens": 9024, + "estimated_eval_tokens": 18304 }, "overflow_policy": "truncate_then_skip", "trimmed": false, "aggressively_trimmed": false, "split": true, "analyst_result": { - "summary": "This cluster is very heterogeneous: most items are standalone loading, tokenizer, processor, or quantization bugs rather than true duplicates. I don\u2019t see a strong single canonical bug; at most, there\u2019s a loose video-input regression theme among a few multimodal issues.", - "confidence": 0.39, - "canonical_issue_reason": "No exact duplicate center stands out, so I\u2019d use issue 42490 as the least model-specific representative of the loading-related subset in this mixed cluster.", + "summary": "All proposed soft edges look like false positives: the items span unrelated bugs across model loading, quantization, tokenizers, processors, video/audio paths, and trainer state. No single canonical issue stands out as a safe duplicate target.", + "confidence": 0.96, + "canonical_issue_reason": null, "canonical_pr_reason": null, - "best_issue_reason": "Issue 42490 is the most broadly applicable, user-facing loading-path bug among these mostly unrelated reports, so it is the best weak representative if one must be chosen.", + "best_issue_reason": null, "best_pr_reason": null, "soft_edge_verdicts": [ + { + "left": "issue:42915", + "right": "issue:44821", + "accept": false, + "reason": "Qwen3Moe FineGrainedFP8Config failure vs AutoImageProcessor URL loading; different subsystems and symptoms." + }, + { + "left": "issue:43746", + "right": "issue:44811", + "accept": false, + "reason": "PEFT local-checkpoint loading bug vs Whisper batch_decode skip_special_tokens handling; unrelated paths." + }, + { + "left": "issue:43257", + "right": "issue:43696", + "accept": false, + "reason": "MoE weight conversion under accelerate+deepspeed vs CUDA OOM on GPT-oss-20b; different root causes." + }, + { + "left": "issue:45084", + "right": "issue:45538", + "accept": false, + "reason": "Template compilation TypeError vs CLIPTokenizer model_max_length bug; no shared code path." + }, + { + "left": "issue:45859", + "right": "issue:47050", + "accept": false, + "reason": "CPU offload rotary embedding issue vs fp16 ONNX dtype mismatch; unrelated model/runtime problems." + }, + { + "left": "issue:43643", + "right": "issue:45759", + "accept": false, + "reason": "AutoConfig missing fields with trust_remote_code vs AutoModelForCausalLM text_config unwrapping; distinct config bugs." + }, { "left": "issue:38175", "right": "issue:44610", "accept": false, - "reason": "Both are vision-model/processor problems, but one is SigLIP2 zero-probability behavior and the other is an OmDet input-size mismatch; different bugs." + "reason": "Zero probabilities in SigLIP2 vs OmDet-Turbo processor/model input-size mismatch; unrelated model families." }, { "left": "issue:42490", "right": "issue:43257", "accept": false, - "reason": "Both involve model loading, but one is a local-path import warning and the other is Qwen3 MoE weight conversion under accelerate/deepspeed; not the same defect." + "reason": "Trailing-slash local-path warning vs Qwen3 MoE weight conversion; different loading issues." }, { "left": "issue:43262", "right": "issue:44811", "accept": false, - "reason": "Both are audio-related, but one is a sampling-rate default bug and the other is Whisper batch_decode skipping special tokens; unrelated code paths." + "reason": "Audio chat-template sampling-rate default vs Whisper batch_decode special-token handling; separate processor behavior." }, { - "left": "issue:43329", - "right": "issue:45600", + "left": "issue:42915", + "right": "issue:43742", "accept": false, - "reason": "Both touch multimodal internals, but one is a video-token counting bug and the other is stale auto-mapping references; not the same underlying issue." + "reason": "Qwen3Moe FP8 config failure vs MobileLLM key error on load; different models and failures." }, { - "left": "issue:42915", - "right": "issue:43742", + "left": "issue:43329", + "right": "issue:45600", "accept": false, - "reason": "Different loading failures: FineGrainedFP8Config vs a key error for MobileLLM-125M. Similar surface area, not the same bug." + "reason": "Undefined video token helper vars vs stale auto_mappings config references; unrelated maintenance bugs." }, { "left": "issue:43329", "right": "issue:44479", "accept": false, - "reason": "Both mention video-input regressions in VLMs, but the titles point to different failure modes and model sets; too broad to treat as one duplicate." + "reason": "Video token counting helper crash vs model-specific video regression; both video-related but not the same bug." }, { "left": "issue:44291", "right": "issue:45042", "accept": false, - "reason": "One is an init_empty_weights argument incompatibility, the other is an image-processor torchvision requirement regression; unrelated." + "reason": "init_empty_weights argument mismatch vs PIL backend requiring torchvision; unrelated compatibility issues." }, { "left": "issue:43526", "right": "issue:44554", "accept": false, - "reason": "BeitImageProcessorFast label reduction and an MPS attention correctness issue are unrelated bugs." + "reason": "BeitImageProcessorFast label reduction bug vs MPS attention correctness issue; different areas entirely." }, { "left": "issue:43329", "right": "issue:43450", "accept": false, - "reason": "Both involve video processing, but one is a missing helper in multimodal token counting and the other is a batched video processor shape bug; not the same defect." + "reason": "Undefined multimodal video token helper vs batched video processor output shape bug; not the same code path." }, { "left": "issue:39692", "right": "issue:44912", "accept": false, - "reason": "Documentation/example errors in SigLIP2 and MXFP4 loading fallback for git-oss-20b are separate problems." + "reason": "SigLIP2 doc example errors vs gpt-oss MXFP4 quantization fallback; documentation/model-loading mismatch." }, { "left": "issue:34567", "right": "issue:40990", "accept": false, - "reason": "Trainer token accounting and high perplexity on gpt-oss-20b are unrelated." + "reason": "TrainerState token counter not updating vs high perplexity on gpt-oss-20b; unrelated training/eval issues." }, { - "left": "issue:44387", - "right": "issue:44617", + "left": "issue:29159", + "right": "issue:34567", "accept": false, - "reason": "Both can lead to OOM, but one is a quantization memory regression and the other is a Sam3Video runtime memory issue; different causes." + "reason": "Slow/fast tokenizer inconsistency vs TrainerState token accounting; different components." }, { - "left": "issue:29159", - "right": "issue:34567", + "left": "issue:44387", + "right": "issue:44617", "accept": false, - "reason": "Tokenizer consistency and trainer token counting are different subsystems and bugs." + "reason": "Reserved-memory regression under int4 quantization vs Sam3Video OOM; both memory-related but distinct failures." }, { "left": "issue:44589", "right": "issue:47333", "accept": false, - "reason": "Float8 storage lookup failure and disk-offloaded Qwen3 save failure are different loading/offloading problems." + "reason": "Missing Float8 storage class vs disk-offloaded Qwen3 save failure; unrelated serialization/offload bugs." }, { "left": "issue:43257", "right": "issue:43931", "accept": false, - "reason": "Both are Qwen3-family load failures, but one is MoE weight conversion under accelerate/deepspeed and the other is a shape-mismatch on a VL model; not clearly the same bug." + "reason": "Qwen3 MoE weight conversion under distributed loading vs Qwen3-VL weight-shape mismatch; different load-time problems." }, { "left": "issue:45081", "right": "issue:47020", "accept": false, - "reason": "Both are tokenizer construction bugs, but they involve different tokenizers and different failure mechanisms." + "reason": "Mistral regex patch backend_tokenizer crash vs XLMRobertaTokenizer Unigram init bug; distinct tokenizer implementations." }, { "left": "issue:44987", "right": "issue:45538", "accept": false, - "reason": "A model-loading failure for physical-intelligence/fast and CLIPTokenizer max-length misconfiguration are unrelated." + "reason": "Loading failure for physical-intelligence/fast under 5.1.0 vs CLIPTokenizer max-length constant bug; unrelated." }, { "left": "issue:43425", "right": "issue:46291", "accept": false, - "reason": "Torch 2.10 incompatibility and kernels import requiring version/revision are different dependency issues." + "reason": "Torch version incompatibility vs kernels package requiring version/revision; different dependency issues." }, { "left": "issue:44912", "right": "issue:47145", "accept": false, - "reason": "Both are quantization-loading bugs, but MXFP4 fallback to bf16 and FP8 frozen-checkpoint weight_scale omission are distinct defects." - }, - { - "left": "issue:45459", - "right": "issue:45803", - "accept": false, - "reason": "Both concern exception handling, but they are different functions and different masked-error scenarios." + "reason": "MXFP4 quantization fallback to bf16 vs FP8 compressed-tensors missing weight_scale; both quantization-related but different formats and bugs." }, { "left": "issue:43525", "right": "issue:46291", "accept": false, - "reason": "Llama4Config missing pad_token_id and kernels import requirements are unrelated." + "reason": "Llama4Config missing pad_token_id vs pipelines import failing on kernels dependency; unrelated." }, { "left": "issue:28282", "right": "issue:29870", "accept": false, - "reason": "Missing PyTorch for AutoModel and Bart generate decoder_attention_mask assignment error are unrelated." - }, - { - "left": "issue:28282", - "right": "issue:34689", - "accept": false, - "reason": "A missing-dependency import error and a Llama 3.2 Vision model-loading regression are not the same bug." - }, - { - "left": "issue:43540", - "right": "issue:46701", - "accept": false, - "reason": "Both are multimodal pipeline failures, but one is Qwen3OmniMoe video processing and the other is DiffusionGemma image-text-to-text pipeline behavior; too different." - }, - { - "left": "issue:43749", - "right": "issue:44756", - "accept": false, - "reason": "Both involve memory/offloading, but one is an FSDP CPU RAM-efficient loading regression and the other is mmap on Strix Halo OOM avoidance; different issues." + "reason": "PyTorch missing ImportError vs Bart generate decoder_attention_mask UnboundLocalError; unrelated load vs generation bugs." }, { - "left": "issue:33357", - "right": "issue:34567", - "accept": false, - "reason": "A MacOS CLIP bus error and trainer token counting are unrelated." - }, - { - "left": "issue:43650", - "right": "issue:43906", + "left": "issue:45459", + "right": "issue:45803", "accept": false, - "reason": "ADD THE DATA is not a concrete bug report, and the isolated reproduction of #38071 is a separate issue." + "reason": "Tokenizer decode error handling with protobuf missing vs bare except swallowing interrupts; same exception theme but different functions and bugs." }, { - "left": "issue:43454", - "right": "issue:43577", + "left": "issue:28282", + "right": "issue:34689", "accept": false, - "reason": "AyaVision weight tying and BLIP2 dtype propagation are different model bugs." + "reason": "Missing PyTorch ImportError vs Llama 3.2 Vision model-loading regression; different underlying causes." } ] }, @@ -7363,36 +7263,36 @@ "source": "llm", "cache_hit": false, "model": "gpt-5.4-mini?service_tier=flex", - "cluster_id": "cluster-10000-529", + "cluster_id": "cluster-10000-530", "nodes": [ "issue:24643", - "issue:27957", "issue:28282", "issue:29127", "issue:29159", + "issue:33357", "issue:34567", - "issue:36331", "issue:39379", - "issue:41720", "issue:42175", "issue:42915", "issue:43404", - "issue:43452", + "issue:43454", "issue:43531", "issue:43540", "issue:43577", "issue:43582", + "issue:43650", "issue:43697", "issue:43742", "issue:43749", "issue:43824", + "issue:43906", "issue:44062", "issue:44291", "issue:44479", "issue:44554", - "issue:44560", "issue:44568", "issue:44589", + "issue:44756", "issue:44821", "issue:44857", "issue:44936", @@ -7406,8 +7306,8 @@ "issue:45538", "issue:46421", "issue:46424", - "issue:46512", "issue:46693", + "issue:46701", "issue:46821", "issue:46832", "issue:46869", @@ -7415,245 +7315,224 @@ "issue:47332" ], "soft_pairs": [ - "issue:46421|issue:46424", - "issue:44568|issue:45356", + "issue:43650|issue:43906", + "issue:33357|issue:34567", + "issue:43540|issue:46701", + "issue:43749|issue:44756", + "issue:43454|issue:43577", "issue:43742|issue:44291", + "issue:44568|issue:45356", "issue:44291|issue:44821", + "issue:46421|issue:46424", "issue:44936|issue:46821", "issue:42175|issue:43824", "issue:28282|issue:29127", "issue:44589|issue:47332", "issue:43404|issue:45127", - "issue:44479|issue:46693", "issue:24643|issue:29159", "issue:45137|issue:45405", + "issue:44479|issue:46693", "issue:42915|issue:45538", "issue:43697|issue:45357", "issue:43582|issue:47254", - "issue:44857|issue:46832", "issue:44062|issue:44991", "issue:43531|issue:44821", - "issue:39379|issue:44554", + "issue:44857|issue:46832", "issue:44568|issue:45488", + "issue:39379|issue:44554", "issue:45538|issue:46869", - "issue:42915|issue:43749", - "issue:27957|issue:28282", - "issue:41720|issue:43540", - "issue:43540|issue:44560", - "issue:43452|issue:43582", - "issue:43540|issue:44479", - "issue:34567|issue:36331", - "issue:42915|issue:47332", - "issue:43577|issue:46512" + "issue:42915|issue:43749" ], - "prepared_review_unit_hash": "eb8d937d3fd29fd0945e66a7efb77adc2a8aba7462a11dab2ff1a5d9c717555d", + "prepared_review_unit_hash": "a99a5c582806dc8ac989947efe0e76ac643205ea9b8d771737abf96dfe3982ca", "component_budget": { - "node_count": 529, - "item_count": 529, - "soft_pair_count": 1953, - "serialized_chars": 1012074, - "estimated_input_tokens": 253019, - "estimated_eval_tokens": 506294 + "node_count": 530, + "item_count": 530, + "soft_pair_count": 1954, + "serialized_chars": 1013004, + "estimated_input_tokens": 253251, + "estimated_eval_tokens": 506758 }, "budget": { "node_count": 48, "item_count": 48, - "soft_pair_count": 30, - "serialized_chars": 36695, - "estimated_input_tokens": 9174, - "estimated_eval_tokens": 18604 + "soft_pair_count": 27, + "serialized_chars": 35705, + "estimated_input_tokens": 8927, + "estimated_eval_tokens": 18110 }, "overflow_policy": "truncate_then_skip", "trimmed": false, "aggressively_trimmed": false, "split": true, "analyst_result": { - "summary": "All suggested soft edges are false positives: the pairs share broad subsystems or keywords, but not the same underlying bug/change. No PRs are present. I chose a high-signal tokenizer regression issue as the best canonical issue candidate.", - "confidence": 0.92, - "canonical_issue_reason": "issue:45488 is the strongest representative issue here: it describes a concrete, user-visible regression in a core tokenizer path, has multiple inbound references, and is scoped clearly enough to serve as a canonical example.", + "summary": "This cluster is mostly a grab bag of unrelated issues across tokenizers, model loading, cache behavior, DeepSpeed/FSDP, and multimodal regressions. The soft pairs share at most broad subsystem similarity, not the same concrete bug, so none should be merged as duplicates.", + "confidence": 0.69, + "canonical_issue_reason": "issue:45488 is the strongest representative: it is a concrete, well-described regression with clear user impact, multiple inbound references, and a specific code-path cause in tokenizer handling.", "canonical_pr_reason": null, - "best_issue_reason": "issue:45488 is also the best issue overall because it is a high-impact regression with a clear failure mode and strong external impact across multiple model families.", + "best_issue_reason": "issue:45488 is the best standalone issue in the set because it is specific, reproducible, and broad enough to represent a meaningful bug in the cluster.", "best_pr_reason": null, "soft_edge_verdicts": [ { - "left": "issue:46421", - "right": "issue:46424", + "left": "issue:43650", + "right": "issue:43906", "accept": false, - "reason": "Both are cache-related, but one is about lazy cache init and recompilation while the other is about max_cache_length only applying to quantized cache; different bugs." + "reason": "One is a vague data request, the other is an isolated reproduction of a separate issue; no shared bug." }, { - "left": "issue:44568", - "right": "issue:45356", + "left": "issue:33357", + "right": "issue:34567", + "accept": false, + "reason": "MacOS bus error in CLIP loading vs TrainerState token counting; unrelated code paths." + }, + { + "left": "issue:43540", + "right": "issue:46701", + "accept": false, + "reason": "Both are multimodal failures, but they affect different models and different pipelines, so not the same bug." + }, + { + "left": "issue:43749", + "right": "issue:44756", "accept": false, - "reason": "Both are tokenizer regressions, but they affect different models and different behaviors; no shared concrete code-path bug is evident." + "reason": "FSDP CPU RAM efficient loading vs mmap/OOM on Strix Halo are different memory-management problems." + }, + { + "left": "issue:43454", + "right": "issue:43577", + "accept": false, + "reason": "One is a weight-tying bug in AyaVision, the other a dtype-loading issue in BLIP2; different failures." }, { "left": "issue:43742", "right": "issue:44291", "accept": false, - "reason": "One is a model-loading KeyError for MobileLLM, the other is an init_empty_weights argument mismatch; unrelated failures." + "reason": "Model loading KeyError vs init_empty_weights/_is_hf_initialized TypeError; distinct failure mechanisms." + }, + { + "left": "issue:44568", + "right": "issue:45356", + "accept": false, + "reason": "Both tokenizer regressions, but for different models and different root causes; not mergeable as one fix." }, { "left": "issue:44291", "right": "issue:44821", "accept": false, - "reason": "Different loading paths and different error classes; not the same bug." + "reason": "init_empty_weights argument handling is unrelated to loading AutoImageProcessor from URL." + }, + { + "left": "issue:46421", + "right": "issue:46424", + "accept": false, + "reason": "Both involve cache behavior, but one is lazy-init recompilation and the other is max_cache_length handling." }, { "left": "issue:44936", "right": "issue:46821", "accept": false, - "reason": "Trainer lifecycle failure vs a TimesFM window_size AttributeError; different code paths." + "reason": "Trainer evaluate-after-train bug is unrelated to TimesFM window_size attribute errors." }, { "left": "issue:42175", "right": "issue:43824", "accept": false, - "reason": "Packaging/backend inclusion issue vs missing model import; unrelated." + "reason": "Packaging/backend dependency issue vs missing Qwen2.5-VL import; unrelated." }, { "left": "issue:28282", "right": "issue:29127", "accept": false, - "reason": "AutoModel import dependency failure vs a LayoutLMv3 error-message/validation issue; not the same bug." + "reason": "Missing PyTorch import dependency vs unclear LayoutLMv3 error message; different bugs." }, { "left": "issue:44589", "right": "issue:47332", "accept": false, - "reason": "Float8 storage lookup failure vs disk-offload save failure; different mechanisms." + "reason": "Float8 storage lookup failure vs saving a disk-offloaded model; not the same underlying defect." }, { "left": "issue:43404", "right": "issue:45127", "accept": false, - "reason": "Both involve tied weights, but one is a Mistral3 lm_head tying bug and the other is LoRA merge with extended vocab causing collapse; not the same concrete issue." - }, - { - "left": "issue:44479", - "right": "issue:46693", - "accept": false, - "reason": "One is a multimodal video-input regression; the other is a performance regression with FA2 and long-context generation." + "reason": "Both mention tied weights, but one is a missing tie in a model, the other is LoRA merge collapse after vocab extension." }, { "left": "issue:24643", "right": "issue:29159", "accept": false, - "reason": "DeepSpeed training runtime error vs tokenizer slow/fast inconsistency; unrelated." + "reason": "DeepSpeed training shape error vs slow/fast tokenizer inconsistency; no overlap beyond being bugs." }, { "left": "issue:45137", "right": "issue:45405", "accept": false, - "reason": "ZeRO-3 deque error vs a version-bump/package-release complaint; different topics entirely." + "reason": "ZeRO-3 deque underflow vs a dependency version bump; entirely separate." + }, + { + "left": "issue:44479", + "right": "issue:46693", + "accept": false, + "reason": "Video-input regression and FA2 long-context performance regression are different code paths and symptoms." }, { "left": "issue:42915", "right": "issue:45538", "accept": false, - "reason": "Qwen3-MoE FP8 failure and CLIPTokenizer max_length behavior are unrelated." + "reason": "FineGrainedFP8Config failure in Qwen3Moe is unrelated to CLIPTokenizer model_max_length." }, { "left": "issue:43697", "right": "issue:45357", "accept": false, - "reason": "RT-DETR output regression vs Qwen3.5 save_pretrained key serialization bug; different code paths." + "reason": "RT-DETR output drift in v5 vs Qwen3.5 save_pretrained key mismatch; unrelated model families and bugs." }, { "left": "issue:43582", "right": "issue:47254", "accept": false, - "reason": "Apple Silicon allocator TypeError vs PEFT/ZeRO-3 checkpointing error; not the same bug." - }, - { - "left": "issue:44857", - "right": "issue:46832", - "accept": false, - "reason": "Both are RT-DETR family issues, but one is an AMP/CUDA crash and the other is a feature-level/backbone mismatch; distinct failure modes." + "reason": "Apple Silicon allocator warmup TypeError vs PEFT/ZeRO-3 checkpointing error; different failures." }, { "left": "issue:44062", "right": "issue:44991", "accept": false, - "reason": "Tokenizers.AddedToken argument duplication vs tokenizer loading failure for a specific checkpoint; different root causes." + "reason": "AddedToken keyword conflict vs tokenizer loading failure for est-roberta; not the same root cause." }, { "left": "issue:43531", "right": "issue:44821", "accept": false, - "reason": "Qwen3-MoE sliding_window behavior vs AutoImageProcessor loading from URL; unrelated." + "reason": "Qwen3-MoE sliding_window behavior vs AutoImageProcessor URL loading; unrelated." }, { - "left": "issue:39379", - "right": "issue:44554", + "left": "issue:44857", + "right": "issue:46832", "accept": false, - "reason": "mllama mask handling bug vs MPS attention correctness issue; no shared concrete bug." + "reason": "Both are RT-DETR-related, but one is an AMP crash in loss and the other is a forward shape/type issue in a different condition." }, { "left": "issue:44568", "right": "issue:45488", "accept": false, - "reason": "Both are tokenizer regressions, but they concern different tokenizer classes and different breakage; too broad to merge." - }, - { - "left": "issue:45538", - "right": "issue:46869", - "accept": false, - "reason": "Tokenizer max_length default vs YOLO multi-GPU device map failure; unrelated." - }, - { - "left": "issue:42915", - "right": "issue:43749", - "accept": false, - "reason": "Different training/runtime subsystems and different failure modes." - }, - { - "left": "issue:27957", - "right": "issue:28282", - "accept": false, - "reason": "Flash Attention behavior in XLM-Roberta vs missing PyTorch import dependency; unrelated." - }, - { - "left": "issue:41720", - "right": "issue:43540", - "accept": false, - "reason": "Auto device mapping cudaErrorAssert vs video-input ValueError in another Qwen3 variant; different code paths." - }, - { - "left": "issue:43540", - "right": "issue:44560", - "accept": false, - "reason": "Both involve video inputs in Qwen3 variants, but the models and observed failures differ; not enough to treat as one bug." - }, - { - "left": "issue:43452", - "right": "issue:43582", - "accept": false, - "reason": "gguf_file loading breakage vs Apple Silicon allocator TypeError; unrelated." + "reason": "Both are tokenizer regressions, but they affect different tokenizer implementations and different incorrect behaviors." }, { - "left": "issue:43540", - "right": "issue:44479", + "left": "issue:39379", + "right": "issue:44554", "accept": false, - "reason": "Both mention Qwen video regressions, but one is a ValueError in Qwen3OmniMoe and the other is a broader v5.3.0 regression across multiple Qwen VL variants; not the same concrete issue." + "reason": "Mllama mask handling is unrelated to the MPS attention correctness issue." }, { - "left": "issue:34567", - "right": "issue:36331", + "left": "issue:45538", + "right": "issue:46869", "accept": false, - "reason": "TrainerState token-count tracking vs a custom compute_loss signature change; different trainer issues." + "reason": "Tokenizer max_length bug vs YOLO device-map/multi-GPU failure; unrelated." }, { "left": "issue:42915", - "right": "issue:47332", - "accept": false, - "reason": "Qwen3-MoE FP8 config failure vs saving a disk-offloaded Qwen3-VL model; distinct problems." - }, - { - "left": "issue:43577", - "right": "issue:46512", + "right": "issue:43749", "accept": false, - "reason": "Both are dtype-related, but one is model/qformer weights not adopting dtype and the other is config.dtype becoming stale; different bugs." + "reason": "Qwen3Moe FP8 config failure and FSDP CPU RAM efficient loading are separate distributed/runtime issues." } ] }, @@ -7675,32 +7554,34 @@ "source": "llm", "cache_hit": false, "model": "gpt-5.4-mini?service_tier=flex", - "cluster_id": "cluster-10000-529", + "cluster_id": "cluster-10000-530", "nodes": [ "issue:24643", + "issue:27957", "issue:28282", "issue:29870", - "issue:32090", + "issue:34567", + "issue:36331", "issue:38175", + "issue:41720", "issue:42175", + "issue:42915", "issue:43257", "issue:43452", "issue:43540", "issue:43577", + "issue:43582", "issue:43716", "issue:43742", "issue:43756", - "issue:43906", "issue:44062", "issue:44479", "issue:44488", - "issue:44568", + "issue:44560", "issue:44589", - "issue:44625", "issue:44737", "issue:44792", "issue:44849", - "issue:44877", "issue:44898", "issue:44964", "issue:44987", @@ -7718,18 +7599,25 @@ "issue:45907", "issue:46399", "issue:46463", - "issue:46491", - "issue:46693", + "issue:46512", "issue:46701", "issue:47276", "issue:47332", "issue:47333" ], "soft_pairs": [ + "issue:27957|issue:28282", + "issue:41720|issue:43540", + "issue:43540|issue:44479", + "issue:43540|issue:44560", + "issue:43452|issue:43582", + "issue:34567|issue:36331", + "issue:42915|issue:47332", + "issue:43577|issue:46512", "issue:45042|issue:45081", + "issue:44062|issue:45081", "issue:42175|issue:44488", "issue:44479|issue:45381", - "issue:44062|issue:45081", "issue:44987|issue:45907", "issue:38175|issue:44898", "issue:43742|issue:44589", @@ -7740,229 +7628,192 @@ "issue:43716|issue:46701", "issue:43756|issue:45325", "issue:45823|issue:46463", - "issue:45672|issue:45674", "issue:43756|issue:46399", + "issue:45672|issue:45674", "issue:46463|issue:47276", "issue:44737|issue:46463", "issue:45405|issue:45600", "issue:43577|issue:45072", - "issue:43906|issue:46693", - "issue:24643|issue:29870", - "issue:44568|issue:46491", - "issue:44625|issue:44877", - "issue:43540|issue:46693", - "issue:43742|issue:47332", - "issue:43452|issue:44488", - "issue:28282|issue:32090" + "issue:24643|issue:29870" ], - "prepared_review_unit_hash": "d9b48e383792b6376b527ee187c0a9af4c6c95dd298cc16f7ee5a93c723d182c", + "prepared_review_unit_hash": "8ed6bf0eb36e8d1c3aac1c086ff062d39d500bb9289409ffbdb7654619d508d8", "component_budget": { - "node_count": 529, - "item_count": 529, - "soft_pair_count": 1953, - "serialized_chars": 1012074, - "estimated_input_tokens": 253019, - "estimated_eval_tokens": 506294 + "node_count": 530, + "item_count": 530, + "soft_pair_count": 1954, + "serialized_chars": 1013004, + "estimated_input_tokens": 253251, + "estimated_eval_tokens": 506758 }, "budget": { - "node_count": 47, - "item_count": 47, - "soft_pair_count": 28, - "serialized_chars": 35772, - "estimated_input_tokens": 8943, - "estimated_eval_tokens": 18142 + "node_count": 48, + "item_count": 48, + "soft_pair_count": 29, + "serialized_chars": 36515, + "estimated_input_tokens": 9129, + "estimated_eval_tokens": 18514 }, "overflow_policy": "truncate_then_skip", "trimmed": false, "aggressively_trimmed": false, "split": true, "analyst_result": { - "summary": "This cluster is mostly heterogeneous issue reports with one likely duplicate pair in the Qwen VL video-position regression area. The rest appear to be separate model-, tokenizer-, config-, or environment-specific bugs and should not be merged.", - "confidence": 0.76, - "canonical_issue_reason": "issue:44479 is the broadest and clearest report of the Qwen2.5/3 VL video-input regression, covering multiple affected model variants and the shared code path more completely than the narrower follow-up.", + "summary": "Most items are unrelated, broad issue reports across different models/features. The only clear duplicate-like pair is the Qwen2.5-VL video-input regression between issues 44479 and 45381.", + "confidence": 0.86, + "canonical_issue_reason": "issue:44479 is the better canonical issue: it is earlier and frames the bug as the broader v5.3.0 video-input regression affecting multiple Qwen vision-language variants, which subsumes the narrower follow-up report.", "canonical_pr_reason": null, - "best_issue_reason": "issue:44479 has the strongest duplicate-triage value here because it describes the regression in the most general terms and is the best anchor for the one plausible duplicate pair in this set.", + "best_issue_reason": "issue:44479 is the strongest representative for the cluster because it captures the shared regression at the highest level of specificity and scope.", "best_pr_reason": null, "soft_edge_verdicts": [ + { + "left": "issue:27957", + "right": "issue:28282", + "accept": false, + "reason": "Different problems: Flash Attention 2 behavior for XLM-Roberta vs a missing PyTorch dependency error." + }, + { + "left": "issue:41720", + "right": "issue:43540", + "accept": false, + "reason": "Different models and failures: Qwen3 auto device mapping cuda assert vs Qwen3OmniMoe video-input ValueError." + }, + { + "left": "issue:43540", + "right": "issue:44479", + "accept": false, + "reason": "Both are video-related, but they target different Qwen model families and different failure modes; not the same concrete bug." + }, + { + "left": "issue:43540", + "right": "issue:44560", + "accept": false, + "reason": "Different error paths and model variants: Qwen3OmniMoe video processing ValueError vs Qwen3-VL-embedding StopIteration." + }, + { + "left": "issue:43452", + "right": "issue:43582", + "accept": false, + "reason": "Unrelated: tokenizer/model loading for gguf_file vs an Apple Silicon allocator warmup TypeError." + }, + { + "left": "issue:34567", + "right": "issue:36331", + "accept": false, + "reason": "Both are Trainer-related, but one is token-count state tracking and the other is a compute_loss signature change." + }, + { + "left": "issue:42915", + "right": "issue:47332", + "accept": false, + "reason": "Different issues: Qwen3Moe FP8 loading failure vs saving a disk-offloaded Qwen3-VL model." + }, + { + "left": "issue:43577", + "right": "issue:46512", + "accept": false, + "reason": "Both mention dtype, but one is a Blip2 loading bug and the other is config.dtype being stale after from_config()." + }, { "left": "issue:45042", "right": "issue:45081", "accept": false, - "reason": "Different subsystems and failure modes: image processor torchvision dependency vs tokenizer regex patch crash." + "reason": "Different subsystems: image processor torchvision dependency vs Mistral tokenizer regex patch crash." + }, + { + "left": "issue:44062", + "right": "issue:45081", + "accept": false, + "reason": "Unrelated tokenizer bugs with different code paths and symptoms." }, { "left": "issue:42175", "right": "issue:44488", "accept": false, - "reason": "Unrelated loading/backend problems affecting different packages and models; no shared concrete code path." + "reason": "Dependency packaging issue vs a specific model loading failure; no shared underlying bug." }, { "left": "issue:44479", "right": "issue:45381", "accept": true, - "reason": "Both describe the same Qwen2.5-VL/Qwen VL video-position-id regression in v5.3.0; one is broader across model variants, the other is the narrow qwen2.5-vl symptom." - }, - { - "left": "issue:44062", - "right": "issue:45081", - "accept": false, - "reason": "Tokenizer AddedToken keyword conflict is unrelated to the Mistral regex/backend_tokenizer crash." + "reason": "Same underlying Qwen2.5-VL video-input regression in transformers 5.3.0; 45381 is a narrower symptom of the broader bug in 44479." }, { "left": "issue:44987", "right": "issue:45907", "accept": false, - "reason": "Different bug classes: model loading failure vs Hub API argument forwarding in CI." + "reason": "Different API failures: loading a model from transformers>=5.1.0 vs list_repo_files missing revision forwarding." }, { "left": "issue:38175", "right": "issue:44898", "accept": false, - "reason": "Both are vision-model issues but they concern different models and distinct inference paths." + "reason": "Different models and failures: SIGLIP2 zero probabilities vs Perceiver image classification failing on non-default resolution." }, { "left": "issue:43742", "right": "issue:44589", "accept": false, - "reason": "Different models and different root causes: config key error vs missing Float8 storage type." + "reason": "Different loading/storage errors for unrelated models and dtypes." }, { "left": "issue:44792", "right": "issue:44964", "accept": false, - "reason": "Both are multimodal failures, but the affected models and code paths are different enough to treat separately." + "reason": "Both touch multimodal models, but one is a Janus test failure and the other is a Phi-4-multimodal load regression." }, { "left": "issue:44849", "right": "issue:45701", "accept": false, - "reason": "Qwen3.5 hidden-state bug is unrelated to a tokenizer change regression." + "reason": "Different bugs: Qwen3.5 output_hidden_states behavior vs tokenization changes across versions." }, { "left": "issue:43742", "right": "issue:47333", "accept": false, - "reason": "Different loading/saving scenarios and models; no evidence of the same bug." + "reason": "Unrelated: MobileLLM config/key error vs Qwen3 offloading save failure." }, { "left": "issue:43257", "right": "issue:43756", "accept": false, - "reason": "Both are model-specific regressions, but one is accelerate/deepspeed weight conversion and the other is RoPE-layer dropping." + "reason": "Different Qwen3Moe loading/quantization issue vs Smollm3 RoPE-layer count mismatch." }, { "left": "issue:43716", "right": "issue:46701", "accept": false, - "reason": "Image-preprocessor dtype mismatch in Mistral-3 is unrelated to the DiffusionGemma pipeline failure." + "reason": "Different multimodal code paths: image preprocessor dtype mismatch vs DiffusionGemma generation pipeline failure." }, { "left": "issue:43756", "right": "issue:45325", "accept": false, - "reason": "Both mention RoPE/position handling, but they affect different models and different bugs." + "reason": "Both involve RoPE terminology, but one is model architecture coverage and the other is Qwen2.5-VL temporal position scaling." }, { "left": "issue:45823", "right": "issue:46463", "accept": false, - "reason": "Device-map dtype mismatch in Gemma4 PLE is unrelated to T5 decoder-input validation messaging." - }, - { - "left": "issue:45672", - "right": "issue:45674", - "accept": false, - "reason": "Both involve 4-bit inference, but the failing code paths are different: uint8 audio-module finfo handling vs BitsAndBytes skip-module exclusion logic." - }, - { - "left": "issue:43756", - "right": "issue:46399", - "accept": false, - "reason": "RoPE-layer count bug in Smollm3 is unrelated to the Qwen3.5 kernel registration crash." - }, - { - "left": "issue:46463", - "right": "issue:47276", - "accept": false, - "reason": "These are both message-quality issues, but they are different warnings in different execution paths." - }, - { - "left": "issue:44737", - "right": "issue:46463", - "accept": false, - "reason": "XLNet CPU placement bug is unrelated to T5 decoder validation error wording." - }, - { - "left": "issue:45405", - "right": "issue:45600", - "accept": false, - "reason": "PEFT version pinning issue and removed-config auto-mapping CI break are unrelated maintenance problems." + "reason": "Unrelated: Gemma4 device placement during forward vs T5 decoder input validation messaging." }, { "left": "issue:43577", "right": "issue:45072", "accept": false, - "reason": "Both are dtype-related, but they concern different model families and different implementation bugs." - }, - { - "left": "issue:43906", - "right": "issue:46693", - "accept": false, - "reason": "No evidence that the isolated reproduction of issue 38071 is the same concrete bug as the FA2 long-context performance regression." + "reason": "Both are dtype-related, but the affected models and root causes differ; not mergeable as one bug fix." }, { "left": "issue:24643", "right": "issue:29870", "accept": false, - "reason": "DeepSpeed training weight-shape error and BART decoder_attention_mask UnboundLocalError are different failures." - }, - { - "left": "issue:44568", - "right": "issue:46491", - "accept": false, - "reason": "Tokenizer regressions, but on different models and different symptoms (special-token addition vs leading-space round-trip)." - }, - { - "left": "issue:44625", - "right": "issue:44877", - "accept": false, - "reason": "Config propagation bug in Qwen3.5 is not the same as strict config rejection for granite_speech." - }, - { - "left": "issue:43540", - "right": "issue:46693", - "accept": false, - "reason": "Qwen3OmniMoe video processing error is unrelated to FA2 generation performance regression." - }, - { - "left": "issue:43742", - "right": "issue:47332", - "accept": false, - "reason": "Different models and different operations: loading MobileLLM vs saving a disk-offloaded Qwen3-VL model." - }, - { - "left": "issue:43452", - "right": "issue:44488", - "accept": false, - "reason": "Both are loading-related, but one is a gguf_file/from_pretrained regression and the other is a specific model load failure; not enough evidence of the same bug." - }, - { - "left": "issue:44737", - "right": "issue:45072", - "accept": false, - "reason": "CPU device placement in XLNet and bfloat16 dtype mismatches are unrelated issues." - }, - { - "left": "issue:28282", - "right": "issue:32090", - "accept": false, - "reason": "Environment-missing-PyTorch import error is unrelated to the Trainer _gpu_broadcast_one NoneType error." + "reason": "Different runtime errors in different flows: DeepSpeed weight shape issue vs BART decoder_attention_mask assignment bug." } ] }, "evaluator_result": { "accept": true, - "feedback": "The summary is grounded in the packet and stays conservative: it identifies one plausible duplicate pair around the Qwen VL video-position-id regression and treats the rest as separate issues. The accepted soft-edge pair is also reasonably supported by the issue titles and wording, without overstating certainty." + "feedback": "The summary is grounded in the packet and stays conservative. The only accepted pair, issue:44479 \u2194 issue:45381, is plausibly the same Qwen2.5-VL video-input regression, and choosing 44479 as the canonical issue is supported by its earlier, broader framing. The other soft-edge rejections do not overstate equivalence." }, "evaluator_used": true, "retried": false, @@ -7981,185 +7832,234 @@ "source": "llm", "cache_hit": false, "model": "gpt-5.4-mini?service_tier=flex", - "cluster_id": "cluster-10000-529", + "cluster_id": "cluster-10000-530", "nodes": [ + "issue:28282", "issue:29159", + "issue:32090", "issue:33357", "issue:38175", - "issue:40444", "issue:41762", "issue:42796", "issue:43295", "issue:43452", - "issue:43643", + "issue:43540", "issue:43653", "issue:43688", "issue:43697", "issue:43720", + "issue:43742", "issue:43749", "issue:43906", "issue:43994", "issue:44361", "issue:44442", "issue:44466", + "issue:44488", "issue:44493", + "issue:44568", + "issue:44625", "issue:44737", - "issue:44756", - "issue:44811", "issue:44821", "issue:44843", - "issue:44857", + "issue:44877", "issue:44898", "issue:44913", "issue:44936", - "issue:44991", "issue:45005", "issue:45216", "issue:45357", - "issue:45372", - "issue:45375", "issue:45588", "issue:45663", "issue:45920", "issue:46050", "issue:46121", - "issue:46170", "issue:46463", + "issue:46491", + "issue:46693", "issue:46821", "issue:46829", "issue:46846", "issue:46899", - "issue:47000", "issue:47332" ], "soft_pairs": [ + "issue:43906|issue:46693", + "issue:44625|issue:44877", + "issue:44568|issue:46491", + "issue:43540|issue:46693", + "issue:43742|issue:47332", + "issue:43452|issue:44488", + "issue:28282|issue:32090", "issue:44466|issue:45005", - "issue:46463|issue:46899", "issue:42796|issue:43906", + "issue:46463|issue:46899", "issue:41762|issue:44936", "issue:43452|issue:44843", "issue:44913|issue:46121", "issue:45663|issue:46846", "issue:43697|issue:46829", - "issue:44737|issue:45588", "issue:43720|issue:44898", + "issue:44737|issue:45588", "issue:43653|issue:44361", + "issue:44493|issue:46050", "issue:38175|issue:43295", "issue:44821|issue:46821", - "issue:44493|issue:46050", "issue:44361|issue:44442", "issue:43697|issue:45216", "issue:45357|issue:45920", "issue:43688|issue:43749", "issue:43994|issue:44898", - "issue:29159|issue:33357", - "issue:44756|issue:47332", - "issue:40444|issue:44811", - "issue:43643|issue:45375", - "issue:44843|issue:44991", - "issue:45372|issue:46170", - "issue:44857|issue:47000" + "issue:29159|issue:33357" ], - "prepared_review_unit_hash": "df28a69601f9bce28c93104057a65ecbc077da2dcca9c65c0be81e943b947ac5", + "prepared_review_unit_hash": "3f0e0e8e285a65f698d6879c601b5f4766651eea51cfee7832015636706f043c", "component_budget": { - "node_count": 529, - "item_count": 529, - "soft_pair_count": 1953, - "serialized_chars": 1012074, - "estimated_input_tokens": 253019, - "estimated_eval_tokens": 506294 + "node_count": 530, + "item_count": 530, + "soft_pair_count": 1954, + "serialized_chars": 1013004, + "estimated_input_tokens": 253251, + "estimated_eval_tokens": 506758 }, "budget": { "node_count": 48, "item_count": 48, - "soft_pair_count": 26, - "serialized_chars": 35906, - "estimated_input_tokens": 8977, - "estimated_eval_tokens": 18210 + "soft_pair_count": 27, + "serialized_chars": 36165, + "estimated_input_tokens": 9042, + "estimated_eval_tokens": 18340 }, "overflow_policy": "truncate_then_skip", "trimmed": false, "aggressively_trimmed": false, "split": true, "analyst_result": { - "summary": "Most pairs are unrelated. The only clear duplicate-looking pair is the v5 tied-weights/save_pretrained regression: issue 45005 is the broader umbrella and issue 44466 looks like a specific symptom of it.", - "confidence": 0.79, - "canonical_issue_reason": "issue:45005 is the broader, more canonical report for the tied-weights regression in v5; issue:44466 reads like a narrower manifestation (device-dependent lm_head.weight serialization).", + "summary": "The set is mostly a collection of unrelated issue reports across tokenizers, model loading, serialization, and distributed/runtime bugs. Only one soft pair looks plausibly like the same tied-weights/save_pretrained regression; the rest are too far apart to merge.", + "confidence": 0.62, + "canonical_issue_reason": "Issue 45920 is the broadest and most representative regression report here: it spans many models, has a clear v5 tokenizer regression framing, and is more general than the other single-model bugs.", "canonical_pr_reason": null, - "best_issue_reason": "issue:45005 is the best issue to keep as the representative bug report because it describes the underlying tied-weights problem more generally than 44466.", + "best_issue_reason": "Issue 45920 is the best overall anchor because it affects multiple model families and describes a concrete, high-signal regression rather than a narrow edge case.", "best_pr_reason": null, "soft_edge_verdicts": [ + { + "left": "issue:43906", + "right": "issue:46693", + "accept": false, + "reason": "Different problems: one is a reproduction of an earlier issue, the other is a performance regression in long-context generation with FA2." + }, + { + "left": "issue:44625", + "right": "issue:44877", + "accept": false, + "reason": "Unrelated config/loading failures in different model families; no shared underlying code path is evident." + }, + { + "left": "issue:44568", + "right": "issue:46491", + "accept": false, + "reason": "Both are tokenizer bugs, but they affect different tokenizers and different behaviors (special tokens vs leading-space round-trip)." + }, + { + "left": "issue:43540", + "right": "issue:46693", + "accept": false, + "reason": "Video-input processing error in Qwen3OmniMoe is unrelated to the FA2 long-context performance regression." + }, + { + "left": "issue:43742", + "right": "issue:47332", + "accept": false, + "reason": "Different loading/saving paths and different model families; the errors do not point to the same bug." + }, + { + "left": "issue:43452", + "right": "issue:44488", + "accept": false, + "reason": "Both mention loading, but one is a gguf_file/from_pretrained issue while the other is a model-specific load failure; not the same underlying defect." + }, + { + "left": "issue:28282", + "right": "issue:32090", + "accept": false, + "reason": "PyTorch missing ImportError vs GPU broadcast NoneType TypeError are entirely separate failures." + }, { "left": "issue:44466", "right": "issue:45005", "accept": true, - "reason": "Same underlying v5 tied-weights/save_pretrained regression; 44466 is a concrete symptom of the broader issue in 45005." + "reason": "Both point to the same tied-weights/save_pretrained regression in v5, with one describing device-dependent lm_head.weight serialization and the other broader translation-model tied-weight breakage." }, { - "left": "issue:46463", - "right": "issue:46899", + "left": "issue:42796", + "right": "issue:43906", "accept": false, - "reason": "Different models and failures: T5 decoder-input validation messaging vs Gemma4 4-bit multimodal dtype casting." + "reason": "Different tokenizer-specific bugs with no shared model or code-path evidence." }, { - "left": "issue:42796", - "right": "issue:43906", + "left": "issue:46463", + "right": "issue:46899", "accept": false, - "reason": "43906 is only a link to a separate reproduction; no evidence it matches the Ministral tokenizer BPE-marker bug." + "reason": "T5 decoder-input validation messaging is unrelated to Gemma4 multimodal quantization failures." }, { "left": "issue:41762", "right": "issue:44936", "accept": false, - "reason": "Gemma3 ZeRO-3 loading IndexError is unrelated to trainer.evaluate() failing after train()." + "reason": "Gemma3 ZeRO-3 loading IndexError and trainer.evaluate-after-train behavior are different issues." }, { "left": "issue:43452", "right": "issue:44843", "accept": false, - "reason": "Different code paths: gguf_file loading failures vs an offline-mode break in _patch_mistral_regex." + "reason": "Both involve loading, but one is gguf_file handling and the other is offline Hub metadata lookup in Mistral regex patching." }, { "left": "issue:44913", "right": "issue:46121", "accept": false, - "reason": "Both involve config/rope-related code, but one is a reload/defaulting bug and the other is a TypeError on list handling." + "reason": "Both touch RoPE/config code, but one is a reload/default regression and the other is a type error on JSON-loaded lists." }, { "left": "issue:45663", "right": "issue:46846", "accept": false, - "reason": "Different models and mechanics: Gemma4 FSDP2 KeyError vs Qwen3.5 TP plan omissions causing OOM/shape errors." + "reason": "Different architectures and different failure modes: FSDP2 per-layer cache rebuilding vs missing TP plan entries." }, { "left": "issue:43697", "right": "issue:46829", "accept": false, - "reason": "RTDetrV2 output regression and VideoPrism output-field mismatch are unrelated bugs." + "reason": "Different output-shape/return-field bugs in different models; not the same code-path problem." }, { - "left": "issue:44737", - "right": "issue:45588", + "left": "issue:43720", + "right": "issue:44898", "accept": false, - "reason": "XLNet CPU tensor placement issue is unrelated to flash_attention s_aux=None handling." + "reason": "BitNet accelerate-loading unpacking bug is unrelated to Perceiver interpolation failure." }, { - "left": "issue:43720", - "right": "issue:44898", + "left": "issue:44737", + "right": "issue:45588", "accept": false, - "reason": "BitNet packed-weight loading bug and Perceiver non-default-resolution failure are different issues." + "reason": "XLNet CPU placement bug and flash_attention s_aux=None crash are unrelated." }, { "left": "issue:43653", "right": "issue:44361", "accept": false, - "reason": "BigBirdTokenizer special-token registration bug is unrelated to MLukeTokenizer AttributeError." + "reason": "Both are tokenizer-related, but they affect different tokenizers with different symptoms and likely different fixes." + }, + { + "left": "issue:44493", + "right": "issue:46050", + "accept": false, + "reason": "Unexpected key warnings in position IDs and sinq quantization save/load breakage are different issues." }, { "left": "issue:38175", "right": "issue:43295", "accept": false, - "reason": "SigLIP2 zero-probabilities issue is unrelated to the processor.tokenizer regression in custom model code." + "reason": "Siglip2 zero-probability outputs and processor.tokenizer regression are different bugs." }, { "left": "issue:44821", @@ -8167,89 +8067,47 @@ "accept": false, "reason": "AutoImageProcessor URL loading and TimesFM window_size AttributeError are unrelated." }, - { - "left": "issue:44493", - "right": "issue:46050", - "accept": false, - "reason": "Position-id key warnings are not the same as sinq quantized model save/load corruption." - }, { "left": "issue:44361", "right": "issue:44442", "accept": false, - "reason": "Two different tokenizer load/runtime failures with no concrete shared code path shown." + "reason": "Both mention tokenizer failures, but they concern different tokenizer classes and different failure mechanisms." }, { "left": "issue:43697", "right": "issue:45216", "accept": false, - "reason": "RTDetrV2 output differences and Qwen3.5 checkpoint save corruption are different regressions." + "reason": "RTDetrV2 output drift and Qwen3.5 checkpoint saving are unrelated regressions." }, { "left": "issue:45357", "right": "issue:45920", "accept": false, - "reason": "Qwen3.5 visual-encoder key serialization is different from the broad AutoTokenizer token-ID regression." + "reason": "Qwen3.5 visual-encoder save_pretrained bug is unrelated to the broad tokenizer token-id regression across many models." }, { "left": "issue:43688", "right": "issue:43749", "accept": false, - "reason": "Auxiliary-loss normalization and FSDP CPU RAM efficient loading are unrelated." + "reason": "Auxiliary-loss normalization and FSDP CPU RAM-efficient loading are distinct training/runtime problems." }, { "left": "issue:43994", "right": "issue:44898", "accept": false, - "reason": "SigLIP2 nonsense outputs and Perceiver non-default-res failures are different model-specific bugs." + "reason": "Different model families and different symptoms; no shared underlying defect is evident." }, { "left": "issue:29159", "right": "issue:33357", "accept": false, - "reason": "Tokenizer slow/fast inconsistency is unrelated to the MacOS CLIP bus error." - }, - { - "left": "issue:44756", - "right": "issue:47332", - "accept": false, - "reason": "Strix Halo mmap OOM and disk-offloaded Qwen3-VL save failure do not share the same bug." - }, - { - "left": "issue:40444", - "right": "issue:44811", - "accept": false, - "reason": "IterableDataset multi-image finetuning failure is unrelated to Whisper batch_decode skip_special_tokens behavior." - }, - { - "left": "issue:43643", - "right": "issue:45375", - "accept": false, - "reason": "Both mention config field loss, but they are different manifestations and not the same concrete bug." - }, - { - "left": "issue:44843", - "right": "issue:44991", - "accept": false, - "reason": "Offline model_info patching and EMBEDDIA/est-roberta tokenizer loading failure are not shown to be the same issue." - }, - { - "left": "issue:45372", - "right": "issue:46170", - "accept": false, - "reason": "Gemma processor import failure and Windows ROCm distributed import/type errors are unrelated." - }, - { - "left": "issue:44857", - "right": "issue:47000", - "accept": false, - "reason": "Different detection-loss crashes with different failure modes and likely different fixes." + "reason": "Tokenizer slow/fast inconsistency and MacOS bus error with CLIP are unrelated." } ] }, "evaluator_result": { "accept": true, - "feedback": "The summary is grounded in the packet, and the one accepted soft pair (44466/45005) is a conservative duplicate-like link based on closely aligned titles about tied weights and save_pretrained behavior. The remaining pair rejections also stay cautious and do not overclaim shared bugs." + "feedback": "The summary is grounded in the packet and stays conservative. The anchor choice of issue 45920 is supported by the title and breadth of affected models, and the lone accepted soft pair (44466/45005) is plausibly the same tied-weights/save_pretrained regression. The other rejections do not overstate similarity and are framed cautiously enough given the limited evidence in the packet." }, "evaluator_used": true, "retried": false, @@ -8268,18 +8126,17 @@ "source": "llm", "cache_hit": false, "model": "gpt-5.4-mini?service_tier=flex", - "cluster_id": "cluster-10000-529", + "cluster_id": "cluster-10000-530", "nodes": [ "issue:32090", "issue:33666", "issue:36010", "issue:36331", + "issue:40444", "issue:41720", "issue:42898", - "issue:43278", "issue:43295", "issue:43452", - "issue:43504", "issue:43525", "issue:43550", "issue:43577", @@ -8290,275 +8147,255 @@ "issue:43749", "issue:43756", "issue:43761", - "issue:43782", "issue:43927", "issue:44361", "issue:44464", - "issue:44466", "issue:44479", - "issue:44561", "issue:44568", "issue:44625", "issue:44743", - "issue:44977", + "issue:44756", + "issue:44811", + "issue:44843", + "issue:44857", + "issue:44991", "issue:45356", - "issue:45405", - "issue:45440", + "issue:45372", + "issue:45375", "issue:45542", "issue:45701", "issue:45704", - "issue:46050", + "issue:46170", "issue:46178", "issue:46402", "issue:46491", - "issue:46693", "issue:46701", "issue:46829", "issue:46832", "issue:46932", + "issue:47000", "issue:47013", - "issue:47254" + "issue:47332" ], "soft_pairs": [ + "issue:44756|issue:47332", + "issue:40444|issue:44811", + "issue:43643|issue:45375", + "issue:44843|issue:44991", + "issue:45372|issue:46170", + "issue:44857|issue:47000", "issue:43688|issue:45704", "issue:43688|issue:45701", - "issue:43525|issue:46701", "issue:43295|issue:45356", + "issue:43525|issue:46701", "issue:45542|issue:46178", "issue:43646|issue:43749", - "issue:43653|issue:43927", "issue:33666|issue:36010", + "issue:43653|issue:43927", + "issue:43653|issue:44568", "issue:43756|issue:46402", "issue:43761|issue:47013", - "issue:43653|issue:44568", - "issue:46832|issue:46932", "issue:44464|issue:44743", "issue:44479|issue:44625", + "issue:46832|issue:46932", "issue:43550|issue:44361", "issue:43761|issue:46829", "issue:43452|issue:43577", "issue:42898|issue:46491", "issue:41720|issue:44464", "issue:33666|issue:36331", - "issue:32090|issue:33666", - "issue:44466|issue:46050", - "issue:44464|issue:44977", - "issue:44977|issue:46693", - "issue:43504|issue:43550", - "issue:45405|issue:47254", - "issue:43278|issue:43688", - "issue:43688|issue:45440", - "issue:43782|issue:46693", - "issue:43643|issue:44561" + "issue:32090|issue:33666" ], - "prepared_review_unit_hash": "5c2cde18e5f53f7ee052936ce029fa09d6a3344f1caf9bfe4ba942b65f3e1abd", + "prepared_review_unit_hash": "3be9fb1838b07aee07bc73c5d1b16c65459328b8bb3e3cc87913626f8ef8b95f", "component_budget": { - "node_count": 529, - "item_count": 529, - "soft_pair_count": 1953, - "serialized_chars": 1012074, - "estimated_input_tokens": 253019, - "estimated_eval_tokens": 506294 + "node_count": 530, + "item_count": 530, + "soft_pair_count": 1954, + "serialized_chars": 1013004, + "estimated_input_tokens": 253251, + "estimated_eval_tokens": 506758 }, "budget": { "node_count": 48, "item_count": 48, - "soft_pair_count": 30, - "serialized_chars": 36926, - "estimated_input_tokens": 9232, - "estimated_eval_tokens": 18720 + "soft_pair_count": 27, + "serialized_chars": 36098, + "estimated_input_tokens": 9025, + "estimated_eval_tokens": 18306 }, "overflow_policy": "truncate_then_skip", "trimmed": false, "aggressively_trimmed": false, "split": true, "analyst_result": { - "summary": "This cluster does not contain any convincing duplicate pairs; the soft edges mostly connect unrelated regressions that only share broad themes like tokenization, loading, or generation.", - "confidence": 0.94, - "canonical_issue_reason": "issue:43749 is the strongest anchor in this set: it describes a concrete regression, has substantial discussion/inbound references, and is more broadly actionable than the other issues.", + "summary": "I don\u2019t see a true duplicate cluster here: most issues are distinct regressions or model-specific bugs across unrelated subsystems. If an umbrella anchor is needed, #43646 is the broadest generic issue in the set. No PRs were provided.", + "confidence": 0.33, + "canonical_issue_reason": "#43646 is the most general, cross-cutting regression report (\u201cTransformers 5.0.0 breaks custom model initialization\u201d) and is the best umbrella anchor among these mostly unrelated issues.", "canonical_pr_reason": null, - "best_issue_reason": "issue:43749 is the best overall issue to represent the cluster because it is specific, high-signal, and clearly scoped to a real broken feature.", + "best_issue_reason": "#43646 is the best representative issue because it captures a broad class of custom-model breakages, whereas the others are narrower model/version-specific failures or unrelated feature bugs.", "best_pr_reason": null, "soft_edge_verdicts": [ + { + "left": "issue:44756", + "right": "issue:47332", + "accept": false, + "reason": "Different problems: Strix Halo mmap/OOM vs saving a disk-offloaded Qwen3-VL model." + }, + { + "left": "issue:40444", + "right": "issue:44811", + "accept": false, + "reason": "Different subsystems: iterable-dataset multi-image finetuning vs Whisper batch_decode skip_special_tokens handling." + }, + { + "left": "issue:43643", + "right": "issue:45375", + "accept": false, + "reason": "Both involve missing config fields, but the titles point to different code paths/models and not a clearly shared underlying bug." + }, + { + "left": "issue:44843", + "right": "issue:44991", + "accept": false, + "reason": "Both are tokenizer-loading regressions, but one is an offline/mistral-regex network call issue and the other is a specific tokenizer load failure for est-roberta." + }, + { + "left": "issue:45372", + "right": "issue:46170", + "accept": false, + "reason": "Unrelated failures: mistral_common import breakage for Gemma 4 vs Windows/ROCm torch.distributed dependency issues." + }, + { + "left": "issue:44857", + "right": "issue:47000", + "accept": false, + "reason": "Different loss implementations and failure modes: float16 CUDA crash vs infeasible cost matrix." + }, { "left": "issue:43688", "right": "issue:45704", "accept": false, - "reason": "Different bugs: MoE auxiliary-loss normalization vs T5/apex RMSNorm memory leak." + "reason": "Auxiliary-loss normalization is unrelated to the T5 RMSNorm memory leak." }, { "left": "issue:43688", "right": "issue:45701", "accept": false, - "reason": "Different subsystems: loss normalization bug vs tokenizer-version behavior change." + "reason": "One is a loss normalization bug; the other is a tokenizer behavior regression across versions." }, { - "left": "issue:43525", - "right": "issue:46701", + "left": "issue:43295", + "right": "issue:45356", "accept": false, - "reason": "Unrelated models and failures: missing pad_token_id on Llama4Config vs DiffusionGemma pipeline breakage." + "reason": "Both are regressions, but one is processor/tokenizer API breakage and the other is codec/warning behavior for a specific tokenizer." }, { - "left": "issue:43295", - "right": "issue:45356", + "left": "issue:43525", + "right": "issue:46701", "accept": false, - "reason": "Both are tokenizer regressions, but the concrete failure modes and code paths differ." + "reason": "Llama4 pad_token_id config error is unrelated to the DiffusionGemma image-text pipeline failure." }, { "left": "issue:45542", "right": "issue:46178", "accept": false, - "reason": "Both mention TF/backend setup, but one is undefined backend state and the other is a missing BACKENDS_MAPPING entry." + "reason": "Both touch backend detection, but they are different backends and different installation/import problems." }, { "left": "issue:43646", "right": "issue:43749", "accept": false, - "reason": "Custom model initialization regression is unrelated to FSDP CPU RAM efficient loading." + "reason": "Custom model initialization breakage is not the same as FSDP CPU RAM-efficient loading being broken." + }, + { + "left": "issue:33666", + "right": "issue:36010", + "accept": false, + "reason": "Qwen2-VL multi-GPU training and GenerationMixin import failure are unrelated." }, { "left": "issue:43653", "right": "issue:43927", "accept": false, - "reason": "Tokenizer special-token registration bug vs config fields lost on save/load." + "reason": "Tokenizer special-token registration bug vs config token IDs being lost after save/load are different issues." }, { - "left": "issue:33666", - "right": "issue:36010", + "left": "issue:43653", + "right": "issue:44568", "accept": false, - "reason": "Multi-GPU Qwen2-VL training issue vs GenerationMixin import failure; no shared code-path." + "reason": "Both concern special tokens, but they affect different tokenizers and different symptoms." }, { "left": "issue:43756", "right": "issue:46402", "accept": false, - "reason": "Different model-specific problems: RoPE layer loss vs Sam3 text-encoder weights not loading." + "reason": "Smollm3 RoPE-layer mismatch is unrelated to Sam3 text encoder weights not loading." }, { "left": "issue:43761", "right": "issue:47013", "accept": false, - "reason": "Hidden-states output regression vs xLSTM state unpacking crash; different implementations." - }, - { - "left": "issue:43653", - "right": "issue:44568", - "accept": false, - "reason": "Different tokenizer bugs affecting different special-token behaviors." - }, - { - "left": "issue:46832", - "right": "issue:46932", - "accept": false, - "reason": "RT-DETR feature-level mismatch is unrelated to Gemma4 quantized embedder dtype casting." + "reason": "CLIP hidden_states regression and xLSTM tuple-unpacking crash are different model-specific forward-path bugs." }, { "left": "issue:44464", "right": "issue:44743", "accept": false, - "reason": "Both concern generation/statefulness, but one is compiled-forward inconsistency and the other is recurrent-state reset logic." + "reason": "Both involve generation/state behavior, but they are distinct models and different root symptoms." }, { "left": "issue:44479", "right": "issue:44625", "accept": false, - "reason": "Video-input regression and num_labels propagation are separate Qwen3.5 issues." + "reason": "Video-input regression across Qwen VL variants is unrelated to num_labels propagation in Qwen3.5 config." + }, + { + "left": "issue:46832", + "right": "issue:46932", + "accept": false, + "reason": "RT-DETR feature-level mismatch and Gemma4 quantized dtype casting are unrelated." }, { "left": "issue:43550", "right": "issue:44361", "accept": false, - "reason": "Torch.compile/SDPA failure is unrelated to MLukeTokenizer task AttributeError." + "reason": "Bamba torch.compile/SDPA failure is unrelated to MLukeTokenizer AttributeError." }, { "left": "issue:43761", "right": "issue:46829", "accept": false, - "reason": "Both touch hidden_states, but one is CLIPVisionModel.forward and the other is VideoPrism output packaging." + "reason": "Both mention hidden_states, but they are different models and different output-contract bugs." }, { "left": "issue:43452", "right": "issue:43577", "accept": false, - "reason": "GGUF loading breakage is unrelated to Blip2 dtype handling." + "reason": "Tokenizer/model loading with gguf_file is unrelated to BLIP-2 dtype propagation." }, { "left": "issue:42898", "right": "issue:46491", "accept": false, - "reason": "Global tokenization-space behavior change vs a specific CodeLlama leading-space regression." + "reason": "Both are tokenization regressions, but one is a global cleanup-space behavior change and the other is a CodeLlama round-trip spacing bug." }, { "left": "issue:41720", "right": "issue:44464", "accept": false, - "reason": "CUDA assert under auto device map is unrelated to chunked generation with compiled forward." + "reason": "Qwen3 auto device mapping cuda assert and chunked-generation inconsistency are different failures." }, { "left": "issue:33666", "right": "issue:36331", "accept": false, - "reason": "Different trainer/training bugs: multi-GPU Qwen2-VL vs compute_loss signature mismatch." + "reason": "Multi-GPU Qwen2-VL training and CustomTrainer.compute_loss signature mismatch are unrelated." }, { "left": "issue:32090", "right": "issue:33666", "accept": false, - "reason": "Generic Trainer broadcast TypeError is not the same as Qwen2-VL multi-GPU training failure." - }, - { - "left": "issue:44466", - "right": "issue:46050", - "accept": false, - "reason": "Both involve save/load, but tied-weight serialization is a different bug from sinq quantization persistence." - }, - { - "left": "issue:44464", - "right": "issue:44977", - "accept": false, - "reason": "Different generation failures: compiled-forward inconsistency vs flash-attention behavior in Qwen3.5." - }, - { - "left": "issue:44977", - "right": "issue:46693", - "accept": false, - "reason": "One is a correctness bug for Qwen3.5 with flash-attention; the other is a performance regression after 5.2.0." - }, - { - "left": "issue:43504", - "right": "issue:43550", - "accept": false, - "reason": "BEiT pretrained-load legacy-field issue is unrelated to Bamba SDPA/torch.compile failure." - }, - { - "left": "issue:45405", - "right": "issue:47254", - "accept": false, - "reason": "Unreleased PEFT version bump is not the same bug as PEFT+DeepSpeed checkpointing errors." - }, - { - "left": "issue:43278", - "right": "issue:43688", - "accept": false, - "reason": "dtype drift between train and eval is unrelated to auxiliary-loss normalization." - }, - { - "left": "issue:43688", - "right": "issue:45440", - "accept": false, - "reason": "Both involve MoE models, but the observed failures are different and not the same code-path problem." - }, - { - "left": "issue:43782", - "right": "issue:46693", - "accept": false, - "reason": "Weight-only load error for Qwen3VL is unrelated to FA2 long-context performance regression." - }, - { - "left": "issue:43643", - "right": "issue:44561", - "accept": false, - "reason": "Missing config fields from trust_remote_code is a different regression than removing is_torch_fx_available." + "reason": "Trainer _gpu_broadcast_one NoneType error is not the same as Qwen2-VL multi-GPU training." } ] }, @@ -8580,282 +8417,296 @@ "source": "llm", "cache_hit": false, "model": "gpt-5.4-mini?service_tier=flex", - "cluster_id": "cluster-10000-529", + "cluster_id": "cluster-10000-530", "nodes": [ "issue:28218", "issue:33357", "issue:39692", - "issue:42796", "issue:43278", "issue:43295", "issue:43334", "issue:43493", "issue:43504", "issue:43525", + "issue:43550", "issue:43572", "issue:43577", "issue:43606", "issue:43611", - "issue:43646", - "issue:43701", + "issue:43643", + "issue:43688", "issue:43720", + "issue:43782", "issue:43825", "issue:43927", "issue:43994", "issue:44361", + "issue:44464", "issue:44466", "issue:44484", - "issue:44488", "issue:44514", "issue:44554", + "issue:44561", "issue:44610", "issue:44625", "issue:44704", "issue:44734", - "issue:44843", "issue:44964", "issue:44977", "issue:45290", "issue:45405", - "issue:45491", + "issue:45440", "issue:45759", "issue:45823", - "issue:46002", "issue:46050", "issue:46192", "issue:46209", - "issue:46365", "issue:46612", + "issue:46693", "issue:46846", - "issue:47276", - "issue:47332" + "issue:47254", + "issue:47276" ], "soft_pairs": [ + "issue:44466|issue:46050", + "issue:44977|issue:46693", + "issue:44464|issue:44977", + "issue:43504|issue:43550", + "issue:43278|issue:43688", + "issue:45405|issue:47254", + "issue:43688|issue:45440", + "issue:43782|issue:46693", "issue:43504|issue:43720", "issue:43720|issue:43927", - "issue:44734|issue:46612", + "issue:43643|issue:44561", "issue:43606|issue:43927", + "issue:44734|issue:46612", "issue:44625|issue:47276", + "issue:39692|issue:43493", "issue:43504|issue:44610", "issue:43720|issue:44361", - "issue:43577|issue:44554", "issue:44484|issue:44704", + "issue:43577|issue:44554", "issue:43525|issue:43572", + "issue:28218|issue:33357", "issue:43611|issue:46846", "issue:44964|issue:45405", - "issue:28218|issue:33357", "issue:44514|issue:45290", "issue:45823|issue:46050", - "issue:43493|issue:43825", "issue:44977|issue:46209", "issue:39692|issue:43295", + "issue:43493|issue:43825", "issue:43334|issue:45759", - "issue:45491|issue:46002", - "issue:43994|issue:46192", - "issue:45405|issue:47332", - "issue:43334|issue:43572", - "issue:44488|issue:44843", - "issue:42796|issue:44977", - "issue:44554|issue:46365", - "issue:43646|issue:44466", - "issue:43278|issue:43701" + "issue:43994|issue:46192" ], - "prepared_review_unit_hash": "d3210fd52d0315697b2a3f98142f5495beb68f228d81b5ae3c6352d0bd7fb895", + "prepared_review_unit_hash": "5378ca3f86722ffd60e71afb3cc1763c9f508a778185f69b15d3993053bd7064", "component_budget": { - "node_count": 529, - "item_count": 529, - "soft_pair_count": 1953, - "serialized_chars": 1012074, - "estimated_input_tokens": 253019, - "estimated_eval_tokens": 506294 + "node_count": 530, + "item_count": 530, + "soft_pair_count": 1954, + "serialized_chars": 1013004, + "estimated_input_tokens": 253251, + "estimated_eval_tokens": 506758 }, "budget": { "node_count": 47, "item_count": 47, - "soft_pair_count": 28, - "serialized_chars": 35708, - "estimated_input_tokens": 8927, - "estimated_eval_tokens": 18110 + "soft_pair_count": 30, + "serialized_chars": 36292, + "estimated_input_tokens": 9073, + "estimated_eval_tokens": 18402 }, "overflow_policy": "truncate_then_skip", "trimmed": false, "aggressively_trimmed": false, "split": true, "analyst_result": { - "summary": "This cluster is a loose semantic grouping of mostly unrelated Transformers bugs; the soft pairs share subsystems or symptoms, but not the same concrete code-path failures. I would not merge any of the proposed soft edges.", - "confidence": 0.77, - "canonical_issue_reason": "issue:44843 is the clearest, most self-contained regression in a core loading path, with a specific failure mode and actionable reproduction.", + "summary": "This cluster is mostly a grab bag of unrelated Transformers bugs with only broad overlap around config/loading, processor, and generation regressions. None of the soft pairs look like the same concrete underlying issue with enough confidence to merge.", + "confidence": 0.84, + "canonical_issue_reason": "issue:43525 is the clearest anchor: it is a concrete, reproducible config-attribute regression (`pad_token_id` missing) and is the closest match to the small subset of config-field failures in the cluster. The cluster overall is still too heterogeneous to treat this as a true duplicate hub.", "canonical_pr_reason": null, - "best_issue_reason": "issue:44843 is the best representative because it is a precise, user-facing loading bug with broad impact and a well-defined fix surface.", + "best_issue_reason": "issue:43525 is the best single representative because it has a precise failure mode, a clear stack-level symptom, and is actionable. It is not a perfect canonical for the whole set, but it is the strongest issue-level anchor here.", "best_pr_reason": null, "soft_edge_verdicts": [ + { + "left": "issue:44466", + "right": "issue:46050", + "accept": false, + "reason": "Both involve save/load behavior, but one is tied-weight serialization varying by device and the other is sinq quantization persistence. Different bugs and code paths." + }, + { + "left": "issue:44977", + "right": "issue:46693", + "accept": false, + "reason": "Shared mention of flash-attn, but one is Qwen3.5 generation correctness and the other is a long-context performance regression after 5.2.0." + }, + { + "left": "issue:44464", + "right": "issue:44977", + "accept": false, + "reason": "Chunked generation with compiled forward is a different failure mode from Qwen3.5 flash-attention generation issues." + }, + { + "left": "issue:43504", + "right": "issue:43550", + "accept": false, + "reason": "Beit legacy-field loading and Bamba torch.compile/SDPA are unrelated model-specific problems." + }, + { + "left": "issue:43278", + "right": "issue:43688", + "accept": false, + "reason": "BF16->FP32 dtype drift in eval is not the same as auxiliary-loss normalization." + }, + { + "left": "issue:45405", + "right": "issue:47254", + "accept": false, + "reason": "PEFT version pinning and a PEFT+ZeRO-3 checkpoint error are different issues despite both involving PEFT." + }, + { + "left": "issue:43688", + "right": "issue:45440", + "accept": false, + "reason": "Auxiliary-loss normalization and DeepseekV3 implementation divergence are distinct model-behavior bugs." + }, + { + "left": "issue:43782", + "right": "issue:46693", + "accept": false, + "reason": "Qwen3VL weight_only load failure is unrelated to a flash-attn long-context performance regression." + }, { "left": "issue:43504", "right": "issue:43720", "accept": false, - "reason": "Legacy-field model loading failure vs. packed-weight unpacking during accelerate loading; different bugs and code paths." + "reason": "Legacy preset-field loading and BitNet packed-weight unpacking during accelerate load are separate loading bugs." }, { "left": "issue:43720", "right": "issue:43927", "accept": false, - "reason": "BitNet packed-weight loading vs. DiaConfig save/load losing custom token IDs; unrelated failures." + "reason": "Packed-weight loading in BitNet and DiaConfig save/load token-ID loss are different serialization paths." }, { - "left": "issue:44734", - "right": "issue:46612", + "left": "issue:43643", + "right": "issue:44561", "accept": false, - "reason": "Wrong tensor indexing in serve KV-cache continuation vs. beam-search cache reorder skipping for certain models." + "reason": "Both touch trust_remote_code, but one is missing fields on returned configs and the other is removal of `is_torch_fx_available` breaking remote-code models." }, { "left": "issue:43606", "right": "issue:43927", "accept": false, - "reason": "CPU offload device mismatch vs. config serialization dropping custom token IDs; not the same issue." + "reason": "CPU offload device mismatch is not the same as config token-ID persistence after save/load." + }, + { + "left": "issue:44734", + "right": "issue:46612", + "accept": false, + "reason": "KV-cache continuation tensor indexing in serve is different from beam-search cache reorder being skipped for specific model families." }, { "left": "issue:44625", "right": "issue:47276", "accept": false, - "reason": "Missing num_labels propagation vs. misleading warning text; one is a functional bug, the other a message issue." + "reason": "Missing `num_labels` propagation and a misleading `synced_gpus` warning are unrelated." + }, + { + "left": "issue:39692", + "right": "issue:43493", + "accept": false, + "reason": "Both are SigLIP2-related, but one is a doc/example mismatch and quantization failure while the other is an HF-vs-JAX implementation discrepancy." }, { "left": "issue:43504", "right": "issue:44610", "accept": false, - "reason": "Pretrained preset load failure from a legacy field vs. processor/model size mismatch; different concrete problems." + "reason": "BEiT legacy-field loading and OmDet processor/model size mismatch are different model-specific bugs." }, { "left": "issue:43720", "right": "issue:44361", "accept": false, - "reason": "Accelerate loading of packed weights vs. tokenizer AttributeError on tasks; unrelated." + "reason": "BitNet packed-weight loading and MLukeTokenizer task AttributeError are unrelated." }, { - "left": "issue:43577", - "right": "issue:44554", + "left": "issue:44484", + "right": "issue:44704", "accept": false, - "reason": "BLIP2 dtype handling vs. MPS attention correctness issue; different models and code paths." + "reason": "A question about default shard size is not the same as AutoProcessor failing to forward kwargs to cached_file." }, { - "left": "issue:44484", - "right": "issue:44704", + "left": "issue:43577", + "right": "issue:44554", "accept": false, - "reason": "max_shard_size behavior question vs. AutoProcessor kwargs not reaching cached_file; not the same bug." + "reason": "BLIP2 dtype propagation and MPS attention correctness are unrelated." }, { "left": "issue:43525", "right": "issue:43572", "accept": false, - "reason": "Missing pad_token_id in Llama4Config vs. missing pad_token_idx in StableLmConfig; similar theme, different defect." + "reason": "Both are missing-pad-token config reports after v5, but they affect different model configs and different attribute names; not enough evidence of one shared defect." + }, + { + "left": "issue:28218", + "right": "issue:33357", + "accept": false, + "reason": "Tokenizer spacing after added tokens has nothing to do with a MacOS bus error on a CLIP model." }, { "left": "issue:43611", "right": "issue:46846", "accept": false, - "reason": "base_model_prefix loading regression vs. missing linear_attn TP plan entries; separate failures." + "reason": "Base-model-prefix loading regression and missing `linear_attn` entries in the TP plan are different failures." }, { "left": "issue:44964", "right": "issue:45405", "accept": false, - "reason": "Model load failure for Phi-4 multimodal vs. an unreleased dependency version bump; unrelated." - }, - { - "left": "issue:28218", - "right": "issue:33357", - "accept": false, - "reason": "Tokenizer extra-space bug vs. Mac bus error on a CLIP model; no shared underlying code-path." + "reason": "Phi-4 multimodal loading failure is unrelated to a PEFT minimum-version bump." }, { "left": "issue:44514", "right": "issue:45290", "accept": false, - "reason": "Both involve apply_chat_template crashes, but on different inputs and failure modes, so not the same concrete bug." + "reason": "Both hit `apply_chat_template`, but one is batched input with `padding=False` and the other is assistant tool-call messages with no content; different edge cases." }, { "left": "issue:45823", "right": "issue:46050", "accept": false, - "reason": "Gemma4 device mismatch during forward vs. sinq quantized save/load breakage; different behavior and fix surface." - }, - { - "left": "issue:43493", - "right": "issue:43825", - "accept": false, - "reason": "SigLIP2 implementation discrepancy vs. an incorrect pipeline error message; unrelated." + "reason": "Gemma4 device-mapping mismatch and sinq save/load quantization breakage are separate issues." }, { "left": "issue:44977", "right": "issue:46209", "accept": false, - "reason": "Qwen3.5 flash-attention generation bug vs. FP8 inference failure on B200; different hardware/code-path issues." + "reason": "Qwen3.5 flash-attention generation problems and FP8 inference failure on B200 are different hardware/runtime bugs." }, { "left": "issue:39692", "right": "issue:43295", "accept": false, - "reason": "Documentation example errors vs. a regression in processor/tokenizer handling; not the same bug." + "reason": "SigLIP2 documentation/example errors do not match the processor.tokenizer regression for custom model code." }, { - "left": "issue:43334", - "right": "issue:45759", + "left": "issue:43493", + "right": "issue:43825", "accept": false, - "reason": "Missing pad_token_id on Qwen3-VL text config vs. failing to unwrap text_config in multimodal configs; related area, different defect." + "reason": "A SigLIP2 implementation discrepancy is unrelated to a pipeline warning message about translation tasks." }, { - "left": "issue:45491", - "right": "issue:46002", + "left": "issue:43334", + "right": "issue:45759", "accept": false, - "reason": "Gemma3 NaNs in sliding-window attention vs. NaNs in classification heads after checkout/init; distinct failures." + "reason": "Both are config-related, but one is a missing `pad_token_id` AttributeError and the other is failure to unwrap `text_config` for composite multimodal configs; not the same bug." }, { "left": "issue:43994", "right": "issue:46192", "accept": false, - "reason": "SigLIP2 bad outputs from AutoModel/pipeline vs. return_tensors being ignored when text_kwargs is explicit; unrelated." - }, - { - "left": "issue:45405", - "right": "issue:47332", - "accept": false, - "reason": "PEFT version bump issue vs. inability to save a disk-offloaded Qwen3-VL model; no shared root cause." - }, - { - "left": "issue:43334", - "right": "issue:43572", - "accept": false, - "reason": "Qwen3-VL pad token attribute loss vs. StableLm pad_token_idx regression; same symptom class, different models and fields." - }, - { - "left": "issue:44488", - "right": "issue:44843", - "accept": false, - "reason": "Unable to load cjvt/sleng-bert vs. unconditional model_info() call breaking offline mode; different tokenizer-loading problems." - }, - { - "left": "issue:42796", - "right": "issue:44977", - "accept": false, - "reason": "Tokenizer BPE-marker handling vs. Qwen3.5 flash-attention generation; unrelated bugs." - }, - { - "left": "issue:44554", - "right": "issue:46365", - "accept": false, - "reason": "MPS attention correctness issue vs. Trainer import conflict with datasets; different subsystems and failures." - }, - { - "left": "issue:43646", - "right": "issue:44466", - "accept": false, - "reason": "Custom model initialization regression vs. inconsistent lm_head.weight serialization; not the same underlying bug." - }, - { - "left": "issue:43278", - "right": "issue:43701", - "accept": false, - "reason": "dtype mismatch between train/eval vs. resume_from_checkpoint key mismatch; unrelated." + "reason": "Nonsensical SigLIP2 outputs and `return_tensors` being ignored with explicit `text_kwargs` are different processor issues." } ] }, @@ -8877,24 +8728,22 @@ "source": "llm", "cache_hit": false, "model": "gpt-5.4-mini?service_tier=flex", - "cluster_id": "cluster-10000-529", + "cluster_id": "cluster-10000-530", "nodes": [ "issue:15354", - "issue:28218", - "issue:29159", - "issue:32090", - "issue:34567", "issue:38175", "issue:42490", - "issue:42947", + "issue:42796", "issue:43278", - "issue:43381", + "issue:43334", "issue:43493", "issue:43504", "issue:43572", "issue:43606", + "issue:43646", "issue:43653", "issue:43696", + "issue:43701", "issue:43824", "issue:43844", "issue:43873", @@ -8904,40 +8753,50 @@ "issue:44291", "issue:44361", "issue:44464", + "issue:44466", + "issue:44488", "issue:44493", + "issue:44554", "issue:44568", "issue:44589", - "issue:44849", - "issue:44964", + "issue:44843", + "issue:44977", "issue:45072", - "issue:45084", - "issue:45229", "issue:45245", - "issue:45357", + "issue:45405", + "issue:45491", "issue:45520", "issue:45584", "issue:45812", - "issue:45823", + "issue:46002", "issue:46050", "issue:46209", "issue:46285", + "issue:46365", "issue:46489", "issue:46710", "issue:46821", "issue:46832", - "issue:47013", - "issue:47254" + "issue:47332" ], "soft_pairs": [ + "issue:44488|issue:44843", + "issue:45405|issue:47332", + "issue:43334|issue:43572", + "issue:42796|issue:44977", + "issue:44554|issue:46365", + "issue:45491|issue:46002", + "issue:43646|issue:44466", + "issue:43278|issue:43701", "issue:44568|issue:46489", - "issue:44568|issue:45812", "issue:43927|issue:44361", + "issue:44568|issue:45812", "issue:43572|issue:44291", - "issue:43278|issue:43844", "issue:43493|issue:43844", + "issue:43278|issue:43844", "issue:46209|issue:46710", - "issue:43504|issue:43606", "issue:44589|issue:45520", + "issue:43504|issue:43606", "issue:43824|issue:44589", "issue:46821|issue:46832", "issue:44493|issue:46285", @@ -8947,227 +8806,212 @@ "issue:15354|issue:38175", "issue:42490|issue:43696", "issue:45245|issue:46710", - "issue:45072|issue:46832", "issue:44464|issue:45584", "issue:43873|issue:46050", - "issue:44849|issue:46209", - "issue:44849|issue:45357", - "issue:42947|issue:43381", - "issue:28218|issue:34567", - "issue:29159|issue:32090", - "issue:47013|issue:47254", - "issue:43844|issue:44964", - "issue:45084|issue:45520", - "issue:45229|issue:45823" + "issue:45072|issue:46832" ], - "prepared_review_unit_hash": "6c6532bd73ec70e0287ceea2490810c37b9c9f72a8763f3cbd358b95dcd28d37", + "prepared_review_unit_hash": "6f7936fc6cba6beea5400b130f1d74b8d60fcd9099e2aefafbbdb1b5a4382555", "component_budget": { - "node_count": 529, - "item_count": 529, - "soft_pair_count": 1953, - "serialized_chars": 1012074, - "estimated_input_tokens": 253019, - "estimated_eval_tokens": 506294 + "node_count": 530, + "item_count": 530, + "soft_pair_count": 1954, + "serialized_chars": 1013004, + "estimated_input_tokens": 253251, + "estimated_eval_tokens": 506758 }, "budget": { "node_count": 48, "item_count": 48, - "soft_pair_count": 30, - "serialized_chars": 36814, - "estimated_input_tokens": 9204, - "estimated_eval_tokens": 18664 + "soft_pair_count": 29, + "serialized_chars": 36633, + "estimated_input_tokens": 9159, + "estimated_eval_tokens": 18574 }, "overflow_policy": "truncate_then_skip", "trimmed": false, "aggressively_trimmed": false, "split": true, "analyst_result": { - "summary": "This cluster is mostly a loose collection of unrelated model-, tokenizer-, training-, and quantization-specific bugs. None of the soft pairs look like the same underlying issue or a single mergeable change, so all soft edges should be rejected. If a representative issue is needed, 45812 is the broadest and most central tokenizer regression here.", - "confidence": 0.94, - "canonical_issue_reason": "issue:45812 is the broadest, most general regression in the set: it affects AutoTokenizer across all Granite models, has cross-cutting impact, and is more representative than the very model-specific reports.", + "summary": "The set is mostly a collection of unrelated model/tokenizer/config regression reports. I found no soft-edge pair that clearly describes the same underlying bug or change, so I would keep them separated. The strongest standalone issue is the Granite AutoTokenizer regression because it is concrete, broad, and clearly framed as a v4\u2192v5 behavior break.", + "confidence": 0.74, + "canonical_issue_reason": "issue:45812 is the most canonical-style report: it describes a broad but concrete tokenizer regression, names the affected model family, and frames the break as a silent v4\u2192v5 change with clear user impact.", "canonical_pr_reason": null, - "best_issue_reason": "issue:45812 is the best issue candidate because it is broadly scoped, clearly described, and appears to have wider downstream impact than the other mostly niche model-specific bugs.", + "best_issue_reason": "issue:45812 is the best global anchor because it is specific enough to triage and broad enough to catch related Granite tokenizer regressions without being too vague.", "best_pr_reason": null, "soft_edge_verdicts": [ { - "left": "issue:44568", - "right": "issue:46489", + "left": "issue:44488", + "right": "issue:44843", "accept": false, - "reason": "Both are tokenizer regressions, but they involve different models and different failure modes (missing BOS/EOS vs incorrect output), so they are not the same bug." + "reason": "Both are model-loading failures, but one is a specific local-model loading complaint and the other is an offline-regex/model_info bug in AutoTokenizer; not the same underlying defect." + }, + { + "left": "issue:45405", + "right": "issue:47332", + "accept": false, + "reason": "Unrelated problems: a released-version dependency constraint versus an offloading save failure for a specific Qwen3-VL model." + }, + { + "left": "issue:43334", + "right": "issue:43572", + "accept": false, + "reason": "Different config fields, different models, different failure modes (`pad_token_id` vs `pad_token_idx`)." + }, + { + "left": "issue:42796", + "right": "issue:44977", + "accept": false, + "reason": "Tokenizer BPE-marker handling and flash-attention generation are distinct bugs in different areas." + }, + { + "left": "issue:44554", + "right": "issue:46365", + "accept": false, + "reason": "MPS attention correctness and Trainer import resolution are unrelated." + }, + { + "left": "issue:45491", + "right": "issue:46002", + "accept": false, + "reason": "Both mention NaNs, but they occur in different components and contexts; not enough to treat as the same bug." + }, + { + "left": "issue:43646", + "right": "issue:44466", + "accept": false, + "reason": "Custom model initialization breakage is not the same as device-dependent lm_head serialization inconsistency." + }, + { + "left": "issue:43278", + "right": "issue:43701", + "accept": false, + "reason": "Training/eval dtype mismatch and resume checkpoint key mismatch are unrelated regressions." }, { "left": "issue:44568", - "right": "issue:45812", + "right": "issue:46489", "accept": false, - "reason": "Related only at a high level: both concern tokenizer behavior in v5+, but one is about added_special_tokens on mdeberta and the other about Granite token IDs, so they are distinct issues." + "reason": "Both are tokenizer-related, but they target different model families and different tokenization failures." }, { "left": "issue:43927", "right": "issue:44361", "accept": false, - "reason": "A save/load custom token ID loss bug is unrelated to a tokenizer AttributeError on tasks; no shared code-path is evident." + "reason": "Save/load loss of custom token IDs is not the same as an AttributeError in MLukeTokenizer task handling." + }, + { + "left": "issue:44568", + "right": "issue:45812", + "accept": false, + "reason": "Both concern tokenization, but one is missing BOS/EOS insertion for a specific tokenizer while the other is a broad Granite token-ID regression; different code paths and symptoms." }, { "left": "issue:43572", "right": "issue:44291", "accept": false, - "reason": "These are different bugs: a missing config field for StableLm versus an init_empty_weights TypeError from an unexpected argument." + "reason": "A missing config attribute after upgrade is not the same as an `init_empty_weights` TypeError from an unexpected constructor argument." }, { - "left": "issue:43278", + "left": "issue:43493", "right": "issue:43844", "accept": false, - "reason": "One is an eval-time dtype change, the other is a ZeRO-3 gradient explosion; they do not describe the same underlying defect." + "reason": "A SigLIP2 implementation discrepancy and a DeepSpeed ZeRO-3 gradient issue are unrelated." }, { - "left": "issue:43493", + "left": "issue:43278", "right": "issue:43844", "accept": false, - "reason": "SigLIP2 HF-vs-JAX discrepancy and gradient growth under ZeRO-3 are unrelated failures in different model/training paths." + "reason": "These are both training-related, but one is a dtype change in embeddings and the other is a ZeRO-3 gradient anomaly; no shared underlying bug." }, { "left": "issue:46209", "right": "issue:46710", "accept": false, - "reason": "FP8 inference on B200 and a DeepSeek accuracy regression after upgrade are different model/runtime problems." + "reason": "FP8 inference failure on B200 and an accuracy regression after upgrade are distinct issues." }, { - "left": "issue:43504", - "right": "issue:43606", + "left": "issue:44589", + "right": "issue:45520", "accept": false, - "reason": "A legacy-field load failure for Beit segmentation is unrelated to a CPU offload device mismatch in bark-small." + "reason": "Different platform/library failures: Float8 storage lookup vs `flash_attn` import handling on Python 3.13." }, { - "left": "issue:44589", - "right": "issue:45520", + "left": "issue:43504", + "right": "issue:43606", "accept": false, - "reason": "Float8 storage lookup failure and a flash_attn import error on Python 3.13 are unrelated issues." + "reason": "A legacy config field loading bug is unrelated to a CPU-offload device mismatch." }, { "left": "issue:43824", "right": "issue:44589", "accept": false, - "reason": "An import error for Qwen2.5-VL and a Float8 storage type error are different code paths and symptoms." + "reason": "Importing a missing Qwen class and an FP8 storage-object error are not the same bug." }, { "left": "issue:46821", "right": "issue:46832", "accept": false, - "reason": "Both are forward-path errors, but they involve different architectures and distinct logic failures (window_size vs feature-level mismatch)." + "reason": "Two separate model-forward shape/argument issues in different architectures; not mergeable as one fix." }, { "left": "issue:44493", "right": "issue:46285", "accept": false, - "reason": "Unexpected key/position-id regressions and CLIP text model structure changes are not the same defect." + "reason": "Both are regression-style reports, but one is a generic unexpected-key issue and the other is a CLIP structure change; different concrete problems." }, { "left": "issue:43504", "right": "issue:43653", "accept": false, - "reason": "Loading a model preset with a legacy field is not the same as a tokenizer special-token registration bug." + "reason": "Legacy-field loading and BigBirdTokenizer special-token registration are unrelated." }, { "left": "issue:43994", "right": "issue:44493", "accept": false, - "reason": "These affect different model families and different behaviors: nonsensical SigLIP2 outputs versus a position-id key regression." + "reason": "SigLIP2 nonsensical outputs and unexpected position-id keys are different symptoms and code paths." }, { "left": "issue:43606", "right": "issue:44112", "accept": false, - "reason": "Both mention CI/device behavior, but one is model offload mismatch and the other is a stale test override; they are not the same bug." + "reason": "A real CPU-offload mismatch and a stale CI test are not the same underlying change." }, { "left": "issue:15354", "right": "issue:38175", "accept": false, - "reason": "TorchScript GeneratorExp support and SigLIP2 zero-probability outputs are unrelated issues." + "reason": "TorchScript GeneratorExp export failure and zero-probabilities on SigLIP2 are unrelated." }, { "left": "issue:42490", "right": "issue:43696", "accept": false, - "reason": "A trailing-slash import warning and a CUDA OOM are not the same underlying problem." + "reason": "A trailing-slash local path warning/import issue is unrelated to GPU out-of-memory." }, { "left": "issue:45245", "right": "issue:46710", "accept": false, - "reason": "A category-count runtime limit and a model accuracy regression after upgrade are unrelated." - }, - { - "left": "issue:45072", - "right": "issue:46832", - "accept": false, - "reason": "These are different inference bugs: dtype mismatches in bfloat16 versus a backbone-feature-level mismatch in RT-DETR." + "reason": "Category-count limits and post-upgrade accuracy regression are different failure classes." }, { "left": "issue:44464", "right": "issue:45584", "accept": false, - "reason": "Chunked generation with compiled forward and Whisper empty-transcription handling are different generation bugs." + "reason": "Chunked generation with compiled forward and Whisper empty-transcription failure after token alignment are distinct generation bugs." }, { "left": "issue:43873", "right": "issue:46050", "accept": false, - "reason": "Quantization/offloading misbehavior and save/load breakage for sinq-quantized models are related only broadly, not the same concrete bug." - }, - { - "left": "issue:44849", - "right": "issue:46209", - "accept": false, - "reason": "Qwen3.5 hidden-state output and FP8 inference on B200 are unrelated issues." + "reason": "Quantized offloading behavior and broken sinq save/load are related to memory/quantization broadly, but not the same concrete defect." }, { - "left": "issue:44849", - "right": "issue:45357", - "accept": false, - "reason": "Both concern Qwen3.5, but one is an output_hidden_states bug and the other is incorrect save_pretrained visual keys; they are distinct." - }, - { - "left": "issue:42947", - "right": "issue:43381", - "accept": false, - "reason": "Both mention gradient checkpointing, but one is ineffectiveness with PEFT LoRA and the other is a hard failure in eval mode." - }, - { - "left": "issue:28218", - "right": "issue:34567", - "accept": false, - "reason": "Tokenizer added-space behavior and TrainerState token counting are unrelated subsystems." - }, - { - "left": "issue:29159", - "right": "issue:32090", - "accept": false, - "reason": "Tokenizer slow/fast inconsistency and a Trainer _gpu_broadcast_one TypeError have no shared underlying code-path." - }, - { - "left": "issue:47013", - "right": "issue:47254", - "accept": false, - "reason": "xLSTM train-mode unpacking crash and PEFT/DeepSpeed checkpoint errors are different failures that would not plausibly be one merged fix." - }, - { - "left": "issue:43844", - "right": "issue:44964", - "accept": false, - "reason": "Gradient inflation under ZeRO-3 and a model import/load failure for Phi-4 multimodal are unrelated." - }, - { - "left": "issue:45084", - "right": "issue:45520", - "accept": false, - "reason": "A compile error for non-template nodes and a Python 3.13 flash_attn import KeyError are different problems." - }, - { - "left": "issue:45229", - "right": "issue:45823", + "left": "issue:45072", + "right": "issue:46832", "accept": false, - "reason": "Multi-GPU CUDA OOM in Gemma4 and a device mismatch in Gemma4 PLE are distinct issues, not the same bug." + "reason": "dtype mismatch in bfloat16 inference and RTDetr feature-level shape handling are unrelated." } ] }, @@ -9189,67 +9033,76 @@ "source": "llm", "cache_hit": false, "model": "gpt-5.4-mini?service_tier=flex", - "cluster_id": "cluster-10000-529", + "cluster_id": "cluster-10000-530", "nodes": [ + "issue:28218", + "issue:29159", "issue:29942", "issue:30333", + "issue:32090", + "issue:34567", "issue:37428", "issue:41720", + "issue:42947", "issue:43065", "issue:43295", + "issue:43381", + "issue:43404", "issue:43450", "issue:43493", - "issue:43540", - "issue:43550", - "issue:43572", - "issue:43575", "issue:43701", - "issue:43824", "issue:43844", - "issue:43873", "issue:43937", "issue:43957", - "issue:44186", "issue:44423", "issue:44554", - "issue:44610", "issue:44734", "issue:44849", "issue:44869", + "issue:44964", "issue:44977", "issue:45072", + "issue:45084", "issue:45198", "issue:45200", "issue:45229", - "issue:45406", + "issue:45357", "issue:45446", "issue:45520", "issue:45588", "issue:45759", "issue:45823", "issue:46050", - "issue:46121", - "issue:46178", - "issue:46192", + "issue:46209", "issue:46369", "issue:46399", "issue:46701", "issue:46710", "issue:46752", - "issue:46821", "issue:46846", - "issue:47145" + "issue:47013", + "issue:47145", + "issue:47254" ], "soft_pairs": [ + "issue:42947|issue:43381", + "issue:44849|issue:46209", + "issue:44849|issue:45357", + "issue:28218|issue:34567", + "issue:29159|issue:32090", + "issue:43844|issue:44964", + "issue:45084|issue:45520", + "issue:47013|issue:47254", + "issue:45229|issue:45823", "issue:46399|issue:46846", "issue:45200|issue:45229", - "issue:44423|issue:44734", "issue:45588|issue:46050", + "issue:44423|issue:44734", "issue:43493|issue:43701", "issue:44849|issue:47145", "issue:37428|issue:43957", - "issue:43937|issue:46701", "issue:41720|issue:46710", + "issue:43937|issue:46701", "issue:43295|issue:43450", "issue:44554|issue:45446", "issue:43065|issue:46369", @@ -9257,230 +9110,193 @@ "issue:44869|issue:46752", "issue:29942|issue:30333", "issue:45072|issue:45198", - "issue:44977|issue:46710", - "issue:45588|issue:45759", - "issue:43295|issue:43873", - "issue:43572|issue:46821", - "issue:45406|issue:46121", - "issue:43540|issue:45200", - "issue:43295|issue:44610", - "issue:43450|issue:43575", - "issue:43572|issue:43957", - "issue:44186|issue:46192", - "issue:45823|issue:46192", - "issue:37428|issue:43824", - "issue:43550|issue:43844", - "issue:45520|issue:46178", - "issue:45200|issue:45823" + "issue:43404|issue:46710", + "issue:45588|issue:45759" ], - "prepared_review_unit_hash": "07ccc8d0b74f32946ae36462b7c4b8f5b741d30145f428600242514c3ee98b87", + "prepared_review_unit_hash": "b990242e4be5d6d916a45082a963d40ccd7180bc1827a476a097953667bb54d0", "component_budget": { - "node_count": 529, - "item_count": 529, - "soft_pair_count": 1953, - "serialized_chars": 1012074, - "estimated_input_tokens": 253019, - "estimated_eval_tokens": 506294 + "node_count": 530, + "item_count": 530, + "soft_pair_count": 1954, + "serialized_chars": 1013004, + "estimated_input_tokens": 253251, + "estimated_eval_tokens": 506758 }, "budget": { "node_count": 48, "item_count": 48, - "soft_pair_count": 31, - "serialized_chars": 37062, - "estimated_input_tokens": 9266, - "estimated_eval_tokens": 18788 + "soft_pair_count": 27, + "serialized_chars": 35875, + "estimated_input_tokens": 8969, + "estimated_eval_tokens": 18194 }, "overflow_policy": "truncate_then_skip", "trimmed": false, "aggressively_trimmed": false, "split": true, "analyst_result": { - "summary": "This set is a grab-bag of unrelated closed issues across model loading, serving, quantization, flash-attn, tokenization, and training. None of the soft pairs describe the same underlying bug, so no duplicate merge is justified.", - "confidence": 0.98, - "canonical_issue_reason": null, + "summary": "This cluster is mostly a grab-bag of unrelated issues. None of the proposed soft edges look like true duplicates; the titles overlap only at a broad subsystem level (tokenizers, training, multimodal, serving, etc.) without matching the same concrete failure mode.", + "confidence": 0.93, + "canonical_issue_reason": "If one issue must represent the cluster, issue 45229 is the strongest standalone report: it has a clear, reproducible failure (Gemma4 31B-IT multi-GPU CUDA OOM), substantial discussion, and a concrete user-facing impact. That said, the items are not actually a duplicate set.", "canonical_pr_reason": null, - "best_issue_reason": null, + "best_issue_reason": "Issue 45229 is the best standalone issue in this set because it is specific, severe, and actionable, with the clearest described failure path and the most discussion. It is a better representative than the many narrower or more ambiguous reports here.", "best_pr_reason": null, "soft_edge_verdicts": [ + { + "left": "issue:42947", + "right": "issue:43381", + "accept": false, + "reason": "Both mention gradient checkpointing, but one is about PEFT LoRA being ineffective and the other about checkpointing being disallowed in eval mode. Different bug conditions and code paths." + }, + { + "left": "issue:44849", + "right": "issue:46209", + "accept": false, + "reason": "Qwen3.5 hidden-state output mismatch vs FP8 inference failure on B200 are unrelated failures affecting different runtime paths." + }, + { + "left": "issue:44849", + "right": "issue:45357", + "accept": false, + "reason": "One is an output_hidden_states bug; the other is a save_pretrained regression with visual encoder keys. Same model family, different bugs." + }, + { + "left": "issue:28218", + "right": "issue:34567", + "accept": false, + "reason": "Tokenizer spacing after added tokens is unrelated to TrainerState.num_input_tokens_seen not updating." + }, + { + "left": "issue:29159", + "right": "issue:32090", + "accept": false, + "reason": "Inconsistent slow vs fast tokenizer behavior is a tokenizer bug; NoneType passed to _gpu_broadcast_one is a training/distributed runtime error." + }, + { + "left": "issue:43844", + "right": "issue:44964", + "accept": false, + "reason": "ZeRO-3 gradient growth with HfDeepSpeedConfig is unrelated to Phi-4 multimodal model loading failures." + }, + { + "left": "issue:45084", + "right": "issue:45520", + "accept": false, + "reason": "Template-node compilation error and flash_attn import KeyError on Python 3.13 are different failure modes in different components." + }, + { + "left": "issue:47013", + "right": "issue:47254", + "accept": false, + "reason": "Both involve training/checkpointing, but xLSTM unpacking a bare tensor and PEFT+DeepSpeed CheckpointError are not the same underlying bug." + }, + { + "left": "issue:45229", + "right": "issue:45823", + "accept": false, + "reason": "Gemma4 multi-GPU OOM and Gemma4 PLE device mismatch with device_map=auto are distinct problems: memory pressure vs tensor placement." + }, { "left": "issue:46399", "right": "issue:46846", "accept": false, - "reason": "Both are Qwen3.5-related, but one is a missing decorator on rotary-pos-emb registration and the other is missing TP-plan entries causing OOM/shape errors." + "reason": "Missing apply_rotary_pos_emb registration decorator and missing linear_attn entries in TP plan are separate Qwen3.5 issues with different failure symptoms." }, { "left": "issue:45200", "right": "issue:45229", "accept": false, - "reason": "Gemma4 token-type defaults for text-only fine-tuning vs a multi-GPU inference OOM are different problems." + "reason": "Defaulting mm_token_type_ids to zeros is a configuration/shape fix; the other issue is a multi-GPU CUDA OOM." }, { - "left": "issue:44423", - "right": "issue:44734", + "left": "issue:45588", + "right": "issue:46050", "accept": false, - "reason": "Both hit serving, but one is a multimodal continuous-batching `.to` crash and the other is a KV-cache continuation indexing bug." + "reason": "An AttributeError on s_aux=None in flash_attention integration is unrelated to sinq quantized model save/load corruption." }, { - "left": "issue:45588", - "right": "issue:46050", + "left": "issue:44423", + "right": "issue:44734", "accept": false, - "reason": "Flash-attention sinkless-model crash vs quantized checkpoint save/load breakage are unrelated bugs." + "reason": "Both are serve-related, but one crashes in continuous batching for multimodal models and the other crashes on KV-cache continuation in /v1/responses." }, { "left": "issue:43493", "right": "issue:43701", "accept": false, - "reason": "SigLIP2 implementation discrepancy is not the same as a checkpoint resume key mismatch." + "reason": "SigLIP2 implementation discrepancy and resume_from_checkpoint key mismatch are unrelated bugs." }, { "left": "issue:44849", "right": "issue:47145", "accept": false, - "reason": "Qwen3.5 hidden-state output bug and FP8 weight_scale omission affect different code paths." + "reason": "Qwen3.5 hidden-state output bug and FP8 compressed-tensors weight_scale being skipped are different model/runtime issues." }, { "left": "issue:37428", "right": "issue:43957", "accept": false, - "reason": "Flash-attn import failure and meta-device model loading failure are separate import/loading issues." + "reason": "Importing _flash_supports_window_size and meta-device model loading failures are separate loading/runtime problems." }, { - "left": "issue:43937", - "right": "issue:46701", + "left": "issue:41720", + "right": "issue:46710", "accept": false, - "reason": "Invalid GenerationConfig for GLM-5 is unrelated to the DiffusionGemma generation-output pipeline failure." + "reason": "cudaErrorAssert with auto device mapping and an accuracy regression after upgrade are not the same concrete bug." }, { - "left": "issue:41720", - "right": "issue:46710", + "left": "issue:43937", + "right": "issue:46701", "accept": false, - "reason": "A800 cudaErrorAssert during auto device mapping is not the same as an accuracy regression after upgrade." + "reason": "Invalid GenerationConfig for GLM-5 and DiffusionGemmaGenerationOutput pipeline breakage are unrelated." }, { "left": "issue:43295", "right": "issue:43450", "accept": false, - "reason": "Processor tokenizer regression and batched video shape bug are different multimodal preprocessing issues." + "reason": "Both touch multimodal processor paths, but one is a regression around processor.tokenizer and image passthrough, while the other is batched video output shape handling." }, { "left": "issue:44554", "right": "issue:45446", "accept": false, - "reason": "MPS attention correctness bug is unrelated to the PyTorch version guard for AuxRequest import." + "reason": "MPS attention correctness with mismatched head dims is unrelated to a PyTorch version check bug in flex_attention imports." }, { "left": "issue:43065", "right": "issue:46369", "accept": false, - "reason": "Dummy Conv2d in Sam3PixelDecoder is not the same as memory overflow in sam3-litetext." + "reason": "A dummy Conv2d in Sam3PixelDecoder and memory overflow in sam3-litetext are different SAM3 issues with different symptoms." }, { "left": "issue:44977", "right": "issue:46701", "accept": false, - "reason": "Qwen3.5 flash-attn generation bug and DiffusionGemma pipeline failure are different models and failure modes." + "reason": "Qwen3.5 flash-attention generation failure and DiffusionGemma pipeline failure are unrelated." }, { "left": "issue:44869", "right": "issue:46752", "accept": false, - "reason": "Whisper timestamp decoding crash and empty-list chat-template crash are unrelated input-validation bugs." + "reason": "Whisper timestamp decode crash on trailing replacement character is unrelated to apply_chat_template crashing on an empty list." }, { "left": "issue:29942", "right": "issue:30333", "accept": false, - "reason": "Flash Attention 2 test failures and MLflow job-reporting issues share nothing beyond being closed issues." + "reason": "Failing Flash Attention 2 tests and MLflow not reporting failed jobs are completely different issues." }, { "left": "issue:45072", "right": "issue:45198", "accept": false, - "reason": "dtype mismatches in CI are unrelated to Wav2Vec2 save_pretrained/tokenization failures." + "reason": "bf16 dtype mismatches in SwitchTransformers/TimmWrapperModel are unrelated to Wav2Vec2 save_pretrained/tokenization failures." }, { "left": "issue:45588", "right": "issue:45759", "accept": false, - "reason": "Flash-attn aux-state crash and Qwen composite-config unwrapping are different parts of the stack." - }, - { - "left": "issue:43295", - "right": "issue:43873", - "accept": false, - "reason": "Processor tokenization regression is not the same as quantization offloading not working." - }, - { - "left": "issue:43572", - "right": "issue:46821", - "accept": false, - "reason": "StableLm pad_token_idx config regression and TimesFM window_size AttributeError are unrelated." - }, - { - "left": "issue:45406", - "right": "issue:46121", - "accept": false, - "reason": "Gemma4Processor missing `_tokenizer` in serve is unrelated to rope-params JSON list handling." - }, - { - "left": "issue:43540", - "right": "issue:45200", - "accept": false, - "reason": "Qwen3OmniMoe video-input ValueError and Gemma4 text-only mm_token_type_ids defaulting are different bugs." - }, - { - "left": "issue:43295", - "right": "issue:44610", - "accept": false, - "reason": "Custom processor-tokenizer access regression is unrelated to OmDet-Turbo's 640px vs 224px size mismatch." - }, - { - "left": "issue:43450", - "right": "issue:43575", - "accept": false, - "reason": "Batched video shape bug and Qwen2-57B TP OOM are not the same underlying issue." - }, - { - "left": "issue:43572", - "right": "issue:43957", - "accept": false, - "reason": "Missing pad_token_idx and meta-device loading failures do not share a concrete code path." - }, - { - "left": "issue:44186", - "right": "issue:46192", - "accept": false, - "reason": "LayoutLMv2Tokenizer NER/padding crash and ignored return_tensors are separate tokenizer behaviors." - }, - { - "left": "issue:45823", - "right": "issue:46192", - "accept": false, - "reason": "Gemma4 device-map mismatch and return_tensors being ignored are unrelated." - }, - { - "left": "issue:37428", - "right": "issue:43824", - "accept": false, - "reason": "Both are import errors, but for different symbols and models: flash_supports_window_size vs Qwen2.5 VL class." - }, - { - "left": "issue:43550", - "right": "issue:43844", - "accept": false, - "reason": "torch.compile+SDPA failure for Bamba is unrelated to a ZeRO-3 gradient increase bug." - }, - { - "left": "issue:45520", - "right": "issue:46178", - "accept": false, - "reason": "flash_attn KeyError on Python 3.13 and BACKENDS_MAPPING tensorflow_text validation are distinct import/config issues." - }, - { - "left": "issue:45200", - "right": "issue:45823", - "accept": false, - "reason": "Gemma4 token-type defaulting and PLE device mismatch affect different parts of Gemma4." + "reason": "flash_attention s_aux=None crash and AutoModelForCausalLM.from_config not unwrapping text_config are different code paths and bug classes." } ] }, @@ -9502,12 +9318,13 @@ "source": "llm", "cache_hit": false, "model": "gpt-5.4-mini?service_tier=flex", - "cluster_id": "cluster-10000-529", + "cluster_id": "cluster-10000-530", "nodes": [ "issue:28218", "issue:29870", "issue:29942", "issue:35141", + "issue:37428", "issue:39692", "issue:42490", "issue:42915", @@ -9515,50 +9332,61 @@ "issue:43278", "issue:43295", "issue:43334", - "issue:43335", "issue:43450", - "issue:43452", "issue:43493", "issue:43525", + "issue:43540", "issue:43550", + "issue:43572", + "issue:43575", "issue:43653", - "issue:43819", "issue:43824", + "issue:43844", "issue:43856", + "issue:43873", "issue:43950", + "issue:43957", "issue:43975", "issue:44112", - "issue:44442", + "issue:44186", "issue:44466", - "issue:44479", "issue:44521", - "issue:44561", "issue:44610", "issue:44737", - "issue:44857", - "issue:44898", "issue:45127", - "issue:45198", "issue:45200", "issue:45290", - "issue:45325", "issue:45381", - "issue:45412", + "issue:45406", "issue:45520", + "issue:45823", + "issue:46121", + "issue:46178", "issue:46192", "issue:46209", - "issue:46234", "issue:46285", - "issue:46512", - "issue:46894", - "issue:46899" + "issue:46821", + "issue:46894" ], "soft_pairs": [ + "issue:43295|issue:43873", + "issue:43572|issue:46821", + "issue:45406|issue:46121", + "issue:43540|issue:45200", + "issue:43295|issue:44610", + "issue:43450|issue:43575", + "issue:43572|issue:43957", + "issue:44186|issue:46192", + "issue:45823|issue:46192", + "issue:37428|issue:43824", + "issue:43550|issue:43844", + "issue:45520|issue:46178", + "issue:45200|issue:45823", "issue:29870|issue:29942", "issue:42490|issue:43493", "issue:43295|issue:43493", - "issue:43334|issue:43525", "issue:39692|issue:46192", + "issue:43334|issue:43525", "issue:42915|issue:43824", "issue:43278|issue:43856", "issue:44521|issue:45290", @@ -9568,210 +9396,218 @@ "issue:45381|issue:46209", "issue:43975|issue:45520", "issue:44737|issue:46894", - "issue:43950|issue:46285", "issue:44466|issue:45127", "issue:43653|issue:44112", - "issue:43452|issue:46512", - "issue:44442|issue:45198", - "issue:44561|issue:46234", - "issue:44479|issue:46192", - "issue:45200|issue:45325", - "issue:43819|issue:44112", - "issue:43335|issue:43550", - "issue:45200|issue:46899", - "issue:44898|issue:45412", - "issue:44610|issue:44857" + "issue:43950|issue:46285" ], - "prepared_review_unit_hash": "594e4f0b1ba866449af3d0372c34391c4b22cb02a1c0e0f2a45086f7552b3efc", + "prepared_review_unit_hash": "10728492b5b091f50ab8e2246f0a1b2ed22b901c0ead951f192a041142016159", "component_budget": { - "node_count": 529, - "item_count": 529, - "soft_pair_count": 1953, - "serialized_chars": 1012074, - "estimated_input_tokens": 253019, - "estimated_eval_tokens": 506294 + "node_count": 530, + "item_count": 530, + "soft_pair_count": 1954, + "serialized_chars": 1013004, + "estimated_input_tokens": 253251, + "estimated_eval_tokens": 506758 }, "budget": { - "node_count": 48, - "item_count": 48, - "soft_pair_count": 27, - "serialized_chars": 36270, - "estimated_input_tokens": 9068, - "estimated_eval_tokens": 18392 + "node_count": 47, + "item_count": 47, + "soft_pair_count": 30, + "serialized_chars": 36490, + "estimated_input_tokens": 9123, + "estimated_eval_tokens": 18502 }, "overflow_policy": "truncate_then_skip", "trimmed": false, "aggressively_trimmed": false, "split": true, "analyst_result": { - "summary": "These items are mostly unrelated transformers bug reports that only look superficially similar by subsystem. None of the soft pairs appear to describe the same underlying bug or change, so they should all be kept separate.", - "confidence": 0.78, - "canonical_issue_reason": "issue:44521 is the best cluster anchor: it is open, highly discussed, and has the strongest inbound reference signal among the issues while describing a concrete, narrowly scoped bug.", + "summary": "The set is a loose collection of mostly unrelated Transformers issues; none of the soft pairs look like true duplicates or the same fix. If forced to anchor the cluster, the most active and reference-heavy issue is the best representative, but the cluster does not form a real duplicate group.", + "confidence": 0.95, + "canonical_issue_reason": "issue:43873 is the most active open issue and has the most inbound references/discussion, so it is the closest thing to a cluster anchor, even though the surrounding items are not the same bug.", "canonical_pr_reason": null, - "best_issue_reason": "issue:44521 is the most suitable representative issue because it is active and well-scoped, making it the best candidate to serve as the main triage target for this set.", + "best_issue_reason": "issue:43873 is the best triage target because it is active, open, and well-referenced; however, it should not be treated as semantically equivalent to the other items.", "best_pr_reason": null, "soft_edge_verdicts": [ { - "left": "issue:29870", - "right": "issue:29942", + "left": "issue:43295", + "right": "issue:43873", "accept": false, - "reason": "One is a BART generate() decoder_attention_mask initialization bug; the other is a Flash Attention 2 test failure. Different code paths and failure modes." + "reason": "Different failures: a processor/tokenizer regression versus an offloading/quantization problem. No shared code path or fix." }, { - "left": "issue:42490", - "right": "issue:43493", + "left": "issue:43572", + "right": "issue:46821", + "accept": false, + "reason": "Both are AttributeError reports, but they concern different config fields, different models, and different code paths." + }, + { + "left": "issue:45406", + "right": "issue:46121", "accept": false, - "reason": "Local-path import warning and SigLIP2 implementation mismatch are unrelated; same broad loading/model area only." + "reason": "Unrelated issues: missing `_tokenizer` on Gemma4Processor vs a TypeError in rope validation for a JSON-loaded list." + }, + { + "left": "issue:43540", + "right": "issue:45200", + "accept": false, + "reason": "Different subsystems and symptoms: Qwen3OmniMoe video processing versus Gemma 4 text-only fine-tuning defaults." }, { "left": "issue:43295", - "right": "issue:43493", + "right": "issue:44610", "accept": false, - "reason": "Regression about processor.tokenizer/images-to-tokenizer behavior is unrelated to SigLIP2 implementation divergence." + "reason": "Different bugs in different processors; one is a tokenizer access regression, the other is a wrong image size from the OmDet processor." }, { - "left": "issue:43334", - "right": "issue:43525", + "left": "issue:43450", + "right": "issue:43575", "accept": false, - "reason": "Both are pad_token_id AttributeErrors, but they affect different configs/models (Qwen3-VL vs Llama4) and likely different fixes." + "reason": "Batched video shape handling and tensor-parallel OOM are unrelated problems." }, { - "left": "issue:39692", + "left": "issue:43572", + "right": "issue:43957", + "accept": false, + "reason": "Both mention model loading/configuration, but the failure modes and affected mechanisms are different." + }, + { + "left": "issue:44186", "right": "issue:46192", "accept": false, - "reason": "Documentation example errors are not the same bug as return_tensors being ignored when text_kwargs is passed." + "reason": "Tokenizer crashes on NER/padding versus `return_tensors` being ignored; not the same bug." }, { - "left": "issue:42915", - "right": "issue:43824", + "left": "issue:45823", + "right": "issue:46192", "accept": false, - "reason": "Qwen3MoE FP8 failure and Qwen2.5-VL import error are separate model/component issues." + "reason": "Gemma4 device-placement mismatch is unrelated to `return_tensors` handling." }, { - "left": "issue:43278", - "right": "issue:43856", + "left": "issue:37428", + "right": "issue:43824", "accept": false, - "reason": "Embedding dtype drift during evaluate and Qwen3 MoE memory usage are different runtime problems with no shared code-path evidence." + "reason": "Both are import errors, but they involve different missing symbols and different causes." }, { - "left": "issue:44521", - "right": "issue:45290", + "left": "issue:43550", + "right": "issue:43844", "accept": false, - "reason": "Both involve apply_chat_template, but one is zero assistant masks for multimodal inputs and the other crashes on tool-call assistant messages; distinct bugs." + "reason": "Torch.compile + SDPA failure is unrelated to ZeRO-3 gradient growth in DeepSpeed." }, { - "left": "issue:43824", - "right": "issue:43975", + "left": "issue:45520", + "right": "issue:46178", "accept": false, - "reason": "Import error for Qwen2.5-VL is unrelated to deepseek-coder detokenization regression." + "reason": "Both touch import/backend lookup code, but the offending backends and root causes differ." }, { - "left": "issue:28218", - "right": "issue:35141", + "left": "issue:45200", + "right": "issue:45823", "accept": false, - "reason": "Tokenizer extra-space after added token and output embedding reinitialization on resize are unrelated tokenizer/model-init issues." + "reason": "Same model family, but one is a missing default tensor field for fine-tuning and the other is a device-map forward mismatch." }, { - "left": "issue:43262", - "right": "issue:43450", + "left": "issue:29870", + "right": "issue:29942", "accept": false, - "reason": "Audio chat-template sampling-rate default and batched video processor output shape are different processor bugs." + "reason": "Bart generate mask initialization and Flash Attention 2 CI failures are unrelated." }, { - "left": "issue:45381", - "right": "issue:46209", + "left": "issue:42490", + "right": "issue:43493", "accept": false, - "reason": "Qwen2.5-VL vision_position_ids bug and FP8 inference failure on B200 are unrelated and target different model/runtime layers." + "reason": "A trailing-slash import warning and a SigLIP2 implementation discrepancy are not the same issue." }, { - "left": "issue:43975", - "right": "issue:45520", + "left": "issue:43295", + "right": "issue:43493", "accept": false, - "reason": "Detokenization regression and Python 3.13 flash_attn import KeyError are unrelated." + "reason": "Custom code processor access regression versus SigLIP2 HF/JAX mismatch; no shared fix." }, { - "left": "issue:44737", - "right": "issue:46894", + "left": "issue:39692", + "right": "issue:46192", "accept": false, - "reason": "XLNet CPU arange bug and CohereASR training-loss double-shift bug affect different models and code paths." + "reason": "Documentation/example errors are unrelated to the `return_tensors` override bug." }, { - "left": "issue:43950", - "right": "issue:46285", + "left": "issue:43334", + "right": "issue:43525", "accept": false, - "reason": "Non-persistent buffer corruption in from_pretrained is unrelated to CLIP class structure inconsistency." + "reason": "Both are missing `pad_token_id` attribute errors, but they affect different model configs and are not shown to share a single root cause." }, { - "left": "issue:44466", - "right": "issue:45127", + "left": "issue:42915", + "right": "issue:43824", "accept": false, - "reason": "Serialization of lm_head.weight differs by device; LoRA merge collapse with extended vocab is a different tied-weights problem." + "reason": "FineGrainedFP8Config failure and a missing Qwen2.5-VL import are unrelated." }, { - "left": "issue:43653", - "right": "issue:44112", + "left": "issue:43278", + "right": "issue:43856", "accept": false, - "reason": "BigBirdTokenizer special-token decode issue and GraniteSpeech stale device-override CI failure are unrelated test/behavior problems." + "reason": "Embedding dtype drift during evaluate is unrelated to Qwen3 MoE memory inefficiency." }, { - "left": "issue:43452", - "right": "issue:46512", + "left": "issue:44521", + "right": "issue:45290", "accept": false, - "reason": "gguf_file loading failure and from_config(dtype=...) stale config dtype are different API regressions." + "reason": "Both concern `apply_chat_template`, but one is assistant mask generation for multimodal inputs and the other is a crash on tool-call messages; distinct bugs." }, { - "left": "issue:44442", - "right": "issue:45198", + "left": "issue:43824", + "right": "issue:43975", "accept": false, - "reason": "FastSpeech2ConformerTokenizer loading and Wav2Vec2 save_pretrained/tokenization failure are separate tokenizer/model issues." + "reason": "ImportError versus detokenization regression; different symptoms and code paths." }, { - "left": "issue:44561", - "right": "issue:46234", + "left": "issue:28218", + "right": "issue:35141", "accept": false, - "reason": "Removal of is_torch_fx_available impacting trust_remote_code models is unrelated to a missing falcon_mamba_ssm availability helper." + "reason": "Tokenizer spacing after added tokens is unrelated to embedding reinitialization after resizing." }, { - "left": "issue:44479", - "right": "issue:46192", + "left": "issue:43262", + "right": "issue:43450", "accept": false, - "reason": "Video input regression for Qwen variants and return_tensors ignored with text_kwargs are different multimodal/API bugs." + "reason": "Audio chat-template sampling-rate default and batched video shape handling are unrelated." }, { - "left": "issue:45200", - "right": "issue:45325", + "left": "issue:45381", + "right": "issue:46209", "accept": false, - "reason": "Gemma 4 mm_token_type_ids defaulting to zeros and Qwen2.5-VL rope index scaling are unrelated multimodal issues." + "reason": "Qwen2.5-VL video position IDs and FP8 inference on B200 are unrelated issues." }, { - "left": "issue:43819", - "right": "issue:44112", + "left": "issue:43975", + "right": "issue:45520", "accept": false, - "reason": "DAC.from_latents mismatch and GraniteSpeech CI stale device override are unrelated." + "reason": "Detokenization bug and flash-attn import mapping error do not share an underlying code path." }, { - "left": "issue:43335", - "right": "issue:43550", + "left": "issue:44737", + "right": "issue:46894", "accept": false, - "reason": "SwitchTransformers sparse-layer creation bug and Bamba-9B-v2 torch.compile SDPA failure are different model/config/runtime issues." + "reason": "XLNet CPU placement bug and CohereASR training-loss shift bug are unrelated." }, { - "left": "issue:45200", - "right": "issue:46899", + "left": "issue:44466", + "right": "issue:45127", "accept": false, - "reason": "Gemma 4 mm_token_type_ids default issue is distinct from 4-bit quantization dtype-cast failures in multimodal embedders." + "reason": "Both involve tied weights, but one is serialization differing by device and the other is LoRA merge collapse with extended vocabulary." }, { - "left": "issue:44898", - "right": "issue:45412", + "left": "issue:43653", + "right": "issue:44112", "accept": false, - "reason": "Perceiver non-default resolution failure and RT-DETR memory release problem are unrelated vision bugs." + "reason": "Tokenizer special-token registration and a GraniteSpeech CI device-override test are unrelated." }, { - "left": "issue:44610", - "right": "issue:44857", + "left": "issue:43950", + "right": "issue:46285", "accept": false, - "reason": "OmDet-Turbo processor/model size mismatch and LwDetrImageLoss AMP crash are different failures in different components." + "reason": "Non-persistent buffer corruption during from_pretrained and CLIP text-model structure inconsistency are different regression types." } ] }, @@ -9793,7 +9629,7 @@ "source": "llm", "cache_hit": false, "model": "gpt-5.4-mini?service_tier=flex", - "cluster_id": "cluster-10000-529", + "cluster_id": "cluster-10000-530", "nodes": [ "issue:28282", "issue:29942", @@ -9802,49 +9638,59 @@ "issue:37428", "issue:38175", "issue:41720", - "issue:42490", "issue:43335", - "issue:43388", "issue:43425", + "issue:43452", "issue:43504", - "issue:43526", "issue:43550", "issue:43606", "issue:43611", "issue:43653", "issue:43819", "issue:43827", - "issue:43854", "issue:43901", "issue:44112", "issue:44206", - "issue:44351", + "issue:44442", "issue:44466", - "issue:44617", + "issue:44479", + "issue:44561", + "issue:44610", "issue:44792", "issue:44855", "issue:44857", + "issue:44898", "issue:45083", + "issue:45198", "issue:45200", - "issue:45290", + "issue:45325", "issue:45412", "issue:45563", "issue:45584", "issue:45593", "issue:45759", + "issue:46192", "issue:46209", + "issue:46234", "issue:46285", - "issue:46326", - "issue:46424", "issue:46431", "issue:46512", "issue:46566", - "issue:46697", "issue:46821", - "issue:46858", + "issue:46899", "issue:46932" ], "soft_pairs": [ + "issue:43452|issue:46512", + "issue:44442|issue:45198", + "issue:44561|issue:46234", + "issue:44479|issue:46192", + "issue:45200|issue:45325", + "issue:43819|issue:44112", + "issue:43335|issue:43550", + "issue:45200|issue:46899", + "issue:44898|issue:45412", + "issue:44610|issue:44857", "issue:41720|issue:46209", "issue:45200|issue:46932", "issue:37428|issue:43425", @@ -9861,222 +9707,219 @@ "issue:29942|issue:32090", "issue:43335|issue:43504", "issue:44206|issue:45584", - "issue:43611|issue:44466", "issue:43550|issue:44112", - "issue:46424|issue:46858", + "issue:43611|issue:44466", "issue:44855|issue:45083", - "issue:43388|issue:43526", - "issue:46431|issue:46566", - "issue:43335|issue:44112", - "issue:43335|issue:45083", - "issue:43335|issue:43606", - "issue:42490|issue:43854", - "issue:37428|issue:44351", - "issue:44617|issue:46697", - "issue:45290|issue:46326" + "issue:46431|issue:46566" ], - "prepared_review_unit_hash": "57e16f82e06eef27fbe8a1fa69bcb444981df4e346d30575d7128f928a5da4d6", + "prepared_review_unit_hash": "e0378b09a3fa377c987c7d4f610b7a7465d4797f5b1bdc2f4595ea7e3eae75aa", "component_budget": { - "node_count": 529, - "item_count": 529, - "soft_pair_count": 1953, - "serialized_chars": 1012074, - "estimated_input_tokens": 253019, - "estimated_eval_tokens": 506294 + "node_count": 530, + "item_count": 530, + "soft_pair_count": 1954, + "serialized_chars": 1013004, + "estimated_input_tokens": 253251, + "estimated_eval_tokens": 506758 }, "budget": { "node_count": 48, "item_count": 48, - "soft_pair_count": 29, - "serialized_chars": 36689, - "estimated_input_tokens": 9173, - "estimated_eval_tokens": 18602 + "soft_pair_count": 30, + "serialized_chars": 37022, + "estimated_input_tokens": 9256, + "estimated_eval_tokens": 18768 }, "overflow_policy": "truncate_then_skip", "trimmed": false, "aggressively_trimmed": false, "split": true, "analyst_result": { - "summary": "These items are a broad collection of distinct Transformers bugs/regressions rather than a tight duplicate cluster. Most pairs share only a subsystem or symptom class (import errors, cache issues, model-specific failures) but not the same concrete code path.", - "confidence": 0.92, - "canonical_issue_reason": "No single true canonical issue stands out from the set; if one must be chosen, issue:33357 is the best representative because it has the richest discussion and a clearly described runtime failure with a concrete reproduction context.", + "summary": "This cluster is very heterogeneous: mostly unrelated model-loading, config, device, and CI regressions across different components and model families. The soft pairs share broad keywords or subsystem overlap, but not the same concrete bug/change, so none should be merged as duplicates.", + "confidence": 0.88, + "canonical_issue_reason": "No true duplicate center is obvious; if forced to anchor the cluster, #43452 is the broadest user-facing loading/API regression, but it is still only a loose thematic fit.", "canonical_pr_reason": null, - "best_issue_reason": "Issue:33357 is a strong cluster anchor because it is well-documented, has high discussion activity, and describes a specific user-facing crash scenario rather than a generic category label.", + "best_issue_reason": "#43452 is the most generally representative of the cluster\u2019s recurring loading/config failure theme, though the overall set is not a tight duplicate cluster.", "best_pr_reason": null, "soft_edge_verdicts": [ + { + "left": "issue:43452", + "right": "issue:46512", + "accept": false, + "reason": "Different failure modes: gguf/from_pretrained loading vs stale config dtype after from_config." + }, + { + "left": "issue:44442", + "right": "issue:45198", + "accept": false, + "reason": "Different tokenizer/model paths and symptoms; one is AutoTokenizer loading, the other save_pretrained/tokenization failure." + }, + { + "left": "issue:44561", + "right": "issue:46234", + "accept": false, + "reason": "Both are missing-symbol regressions, but for different utilities and unrelated code paths." + }, + { + "left": "issue:44479", + "right": "issue:46192", + "accept": false, + "reason": "Different input-handling bugs: video regression vs text_kwargs/return_tensors precedence." + }, + { + "left": "issue:45200", + "right": "issue:45325", + "accept": false, + "reason": "Different multimodal processing bugs in different models; not the same underlying issue." + }, + { + "left": "issue:43819", + "right": "issue:44112", + "accept": false, + "reason": "One is an algorithmic mismatch in DAC.from_latents; the other is a CI/device-override test failure." + }, + { + "left": "issue:43335", + "right": "issue:43550", + "accept": false, + "reason": "Unrelated model/config bug vs torch.compile/SDPA runtime failure." + }, + { + "left": "issue:45200", + "right": "issue:46899", + "accept": false, + "reason": "Same model family, but distinct bugs: text-only token type defaults vs quantized audio/vision embedder casting." + }, + { + "left": "issue:44898", + "right": "issue:45412", + "accept": false, + "reason": "Different classes of problems: resolution/interpolation failure vs memory leak on deletion." + }, + { + "left": "issue:44610", + "right": "issue:44857", + "accept": false, + "reason": "Different models and code paths: processor size mismatch vs AMP loss crash." + }, { "left": "issue:41720", "right": "issue:46209", "accept": false, - "reason": "Different failures and different hardware contexts: cudaErrorAssert during auto device mapping on A800 vs FP8 inference failure on B200." + "reason": "Both are hardware-related, but one is auto device mapping assert and the other is FP8 inference incompatibility." }, { "left": "issue:45200", "right": "issue:46932", "accept": false, - "reason": "Both mention Gemma 4, but one is about required mm_token_type_ids for text-only fine-tuning while the other is about dtype casting for quantized embedders." + "reason": "#46932 points to a separate quantized vision/audio dtype fix, not the mm_token_type_ids default issue." }, { "left": "issue:37428", "right": "issue:43425", "accept": false, - "reason": "One is a missing internal flash-attention import; the other is a broad Torch 2.10 incompatibility report. Not the same concrete bug." + "reason": "Import failure in flash attention utils vs broad Torch 2.10 incompatibility; not the same bug." }, { "left": "issue:43819", "right": "issue:45083", "accept": false, - "reason": "DAC from_latents/STE mismatch is unrelated to qwen3_omni_moe feature-length helper behavior." + "reason": "Different subsystems entirely: DAC latent conversion vs qwen3_omni_moe feature-length helper." }, { "left": "issue:43335", "right": "issue:43653", "accept": false, - "reason": "Sparse-layer config bug in SwitchTransformers is unrelated to BigBirdTokenizer special-token registration." + "reason": "Different components and symptoms: SwitchTransformers sparse-layer config vs BigBirdTokenizer special-token registration." }, { "left": "issue:43606", "right": "issue:45083", "accept": false, - "reason": "CPU offload device mismatch in Bark is unrelated to qwen3_omni_moe output-length helper behavior." + "reason": "CPU offload device mismatch is unrelated to the qwen3_omni_moe helper behavior." }, { "left": "issue:44857", "right": "issue:45593", "accept": false, - "reason": "Different models and bugs: float16 AMP crash in LwDetrImageLoss vs missing auxiliary losses in D-FINE." + "reason": "Different detection/training bugs; no shared code path." }, { "left": "issue:44792", "right": "issue:45563", "accept": false, - "reason": "Janus image-generation test failure is unrelated to a stale num_return_sequences warning in paged generate()." + "reason": "One is a failing generation test, the other a stale warning in paged generate." }, { "left": "issue:28282", "right": "issue:33357", "accept": false, - "reason": "AutoModel missing PyTorch dependency and a MacOS bus error in CLIP loading are completely different failures." + "reason": "Completely different problems: missing PyTorch import vs MacOS bus error on a CLIP model." }, { "left": "issue:46512", "right": "issue:46821", "accept": false, - "reason": "Stale config dtype after from_config() is unrelated to TimesFM window_size AttributeError." + "reason": "Different API regressions: dtype handling in from_config vs window_size AttributeError." }, { "left": "issue:43827", "right": "issue:43901", "accept": false, - "reason": "Both are docs issues, but they concern different APIs: pipeline removal vs return_all_scores documentation drift." + "reason": "Both are docs-related, but they cover different APIs and different issues." }, { "left": "issue:45759", "right": "issue:46285", "accept": false, - "reason": "Composite Qwen config unwrapping and CLIP text model structure inconsistency are separate model-architecture regressions." + "reason": "Both involve model structure, but the affected models and breakages are different." }, { "left": "issue:38175", "right": "issue:45412", "accept": false, - "reason": "Unexpected zero probabilities in SigLIP2 and RT-DETR memory not being released are unrelated bugs." + "reason": "Unrelated symptoms: zero probabilities vs memory not released." }, { "left": "issue:29942", "right": "issue:32090", "accept": false, - "reason": "Flash Attention 2 test failures and a Trainer _gpu_broadcast_one NoneType TypeError are unrelated code paths." + "reason": "Different failures in different areas: Flash Attention tests vs Trainer GPU broadcast NoneType error." }, { "left": "issue:43335", "right": "issue:43504", "accept": false, - "reason": "SwitchTransformers sparse-layer config bug and Beit pretrained preset legacy-field loading bug are distinct." + "reason": "Distinct bugs in unrelated model/config loading paths." }, { "left": "issue:44206", "right": "issue:45584", "accept": false, - "reason": "LasrFeatureExtractor center-arg regression is unrelated to Whisper empty-transcription behavior after align_special_tokens." - }, - { - "left": "issue:43611", - "right": "issue:44466", - "accept": false, - "reason": "Loading failures tied to base_model_prefix in Transformers 5.0.0 are unrelated to lm_head.weight serialization inconsistency." + "reason": "Different audio-model regressions with no shared underlying code path." }, { "left": "issue:43550", "right": "issue:44112", "accept": false, - "reason": "Bamba torch.compile SDPA failure and GraniteSpeech stale device override CI failure are different issues." + "reason": "A torch.compile/SDPA runtime issue is not the same as a CI stale-device test." }, { - "left": "issue:46424", - "right": "issue:46858", + "left": "issue:43611", + "right": "issue:44466", "accept": false, - "reason": "Cache max_cache_length handling and device-tensor seq_length breaking static cache arithmetic are separate cache bugs." + "reason": "Both are v5 regressions, but one is model loading with base_model_prefix and the other is serialization of tied weights." }, { "left": "issue:44855", "right": "issue:45083", "accept": false, - "reason": "Python 3.13 torch.jit script parsing/IndentationError is unrelated to qwen3_omni_moe feature-length computation." - }, - { - "left": "issue:43388", - "right": "issue:43526", - "accept": false, - "reason": "Both involve labels, but one is metrics gathering for tuple labels and the other is BeitImageProcessorFast reduce_labels behavior." + "reason": "Completely different failures: Python 3.13 import parsing vs qwen3_omni_moe helper behavior." }, { "left": "issue:46431", "right": "issue:46566", "accept": false, - "reason": "Missing chat template on a processor and meta-device weights under device_map='auto' are unrelated failures." - }, - { - "left": "issue:43335", - "right": "issue:44112", - "accept": false, - "reason": "SwitchTransformers sparse-layer config bug and GraniteSpeech CI stale device override are unrelated." - }, - { - "left": "issue:43335", - "right": "issue:45083", - "accept": false, - "reason": "A SwitchTransformers config bug and a qwen3_omni_moe helper-function bug do not share the same underlying change." - }, - { - "left": "issue:43335", - "right": "issue:43606", - "accept": false, - "reason": "Sparse-layer config creation and Bark CPU-offload device mismatch are unrelated." - }, - { - "left": "issue:42490", - "right": "issue:43854", - "accept": false, - "reason": "Local-path trailing-slash warning and inability to load a specific GLM-4.7-Flash model in tests are different issues." - }, - { - "left": "issue:37428", - "right": "issue:44351", - "accept": false, - "reason": "Missing _flash_supports_window_size import and missing HybridCache import/export are both import-related but concern different symbols and fixes." - }, - { - "left": "issue:44617", - "right": "issue:46697", - "accept": false, - "reason": "CUDA OOM in Sam3Video and over-padding in Swin2SR image processing are unrelated." - }, - { - "left": "issue:45290", - "right": "issue:46326", - "accept": false, - "reason": "Chat template crash on tool-call assistant messages and mistral_common continue_final_message bug are distinct tokenizer/template edge cases." + "reason": "Processor chat-template availability and device-map meta weights are unrelated bugs." } ] }, @@ -10098,16 +9941,12 @@ "source": "llm", "cache_hit": false, "model": "gpt-5.4-mini?service_tier=flex", - "cluster_id": "cluster-10000-529", + "cluster_id": "cluster-10000-530", "nodes": [ "issue:10000", "issue:15354", - "issue:26413", - "issue:27957", "issue:28218", "issue:29127", - "issue:29942", - "issue:31501", "issue:32090", "issue:33357", "issue:34689", @@ -10115,9 +9954,10 @@ "issue:37428", "issue:38175", "issue:39692", - "issue:42175", "issue:42490", "issue:43278", + "issue:43335", + "issue:43388", "issue:43526", "issue:43550", "issue:43575", @@ -10126,39 +9966,50 @@ "issue:43720", "issue:43819", "issue:43825", + "issue:43854", "issue:44112", "issue:44206", + "issue:44351", "issue:44448", "issue:44493", "issue:44514", "issue:44521", + "issue:44617", "issue:44743", "issue:44779", "issue:44857", "issue:45072", "issue:45083", - "issue:45127", - "issue:45245", - "issue:45310", - "issue:45459", - "issue:45478", + "issue:45290", "issue:45520", "issue:45588", "issue:45823", "issue:45859", "issue:46170", - "issue:46234" + "issue:46326", + "issue:46424", + "issue:46697", + "issue:46858" ], "soft_pairs": [ + "issue:43388|issue:43526", + "issue:46424|issue:46858", + "issue:43335|issue:44112", + "issue:43335|issue:43606", + "issue:43335|issue:45083", + "issue:42490|issue:43854", + "issue:37428|issue:44351", + "issue:45290|issue:46326", "issue:45520|issue:46170", - "issue:44448|issue:44743", + "issue:44617|issue:46697", "issue:45588|issue:45859", + "issue:44448|issue:44743", "issue:43819|issue:44493", "issue:34689|issue:35141", - "issue:39692|issue:43526", "issue:43606|issue:44857", - "issue:44206|issue:45823", + "issue:39692|issue:43526", "issue:39692|issue:44493", + "issue:44206|issue:45823", "issue:10000|issue:15354", "issue:29127|issue:32090", "issue:44514|issue:44521", @@ -10171,227 +10022,229 @@ "issue:32090|issue:33357", "issue:42490|issue:43575", "issue:38175|issue:44493", - "issue:45310|issue:45478", - "issue:43550|issue:45083", - "issue:26413|issue:31501", - "issue:27957|issue:34689", - "issue:45127|issue:45245", - "issue:45459|issue:46234", - "issue:37428|issue:42175", - "issue:29942|issue:35141" + "issue:43550|issue:45083" ], - "prepared_review_unit_hash": "43323da83f1149a2e192dc813a30756483449522ccc7923e1c4dcd63c66ba587", + "prepared_review_unit_hash": "e1c48ba412937025c82010ce04d16c736ce5842d73c02ca11ac2a78bb43b5b6a", "component_budget": { - "node_count": 529, - "item_count": 529, - "soft_pair_count": 1953, - "serialized_chars": 1012074, - "estimated_input_tokens": 253019, - "estimated_eval_tokens": 506294 + "node_count": 530, + "item_count": 530, + "soft_pair_count": 1954, + "serialized_chars": 1013004, + "estimated_input_tokens": 253251, + "estimated_eval_tokens": 506758 }, "budget": { - "node_count": 48, - "item_count": 48, - "soft_pair_count": 29, - "serialized_chars": 36441, - "estimated_input_tokens": 9111, - "estimated_eval_tokens": 18478 + "node_count": 47, + "item_count": 47, + "soft_pair_count": 31, + "serialized_chars": 36635, + "estimated_input_tokens": 9159, + "estimated_eval_tokens": 18574 }, "overflow_policy": "truncate_then_skip", "trimmed": false, "aggressively_trimmed": false, "split": true, "analyst_result": { - "summary": "Only one clear duplicate pair appears in this cluster: two reports of the same Qwen3.5 MoE from_pretrained regression. The rest are separate bugs touching different models, code paths, or symptoms, so they should not be merged.", + "summary": "The candidate pairs are mostly unrelated; none looks like the same underlying bug or change, so all soft edges should be rejected. There is no coherent PR cluster here, and the issue set is too heterogeneous to have a strong canonical beyond a loose representative.", "confidence": 0.95, - "canonical_issue_reason": "issue:45310 is the better canonical issue: it is earlier and has the fuller report for the same Qwen3.5 MoE from_pretrained failure, while issue:45478 has the same title and appears to be a later duplicate.", + "canonical_issue_reason": "No true canonical issue: the set spans unrelated failures across tokenization, loading, caching, CI, and dtype behavior. If forced to pick an anchor, issue 44521 is the most discussed/open, but it is not a duplicate of the others.", "canonical_pr_reason": null, - "best_issue_reason": "issue:45310 is the best representative issue in this set because it is the earliest, most complete instance of the only confirmed duplicate pair.", + "best_issue_reason": "Issue 44521 is the best representative only in the sense that it has the strongest discussion/inbound-reference signal and a clear reproducible bug, but it should not be treated as a duplicate target for the rest of the set.", "best_pr_reason": null, "soft_edge_verdicts": [ + { + "left": "issue:43388", + "right": "issue:43526", + "accept": false, + "reason": "Different label bugs in different components: metric gathering truncation vs BeitImageProcessorFast label reduction." + }, + { + "left": "issue:46424", + "right": "issue:46858", + "accept": false, + "reason": "Both are cache-related, but one is max_cache_length handling for quantized cache and the other is device-tensor seq_length arithmetic in static cache/compile." + }, + { + "left": "issue:43335", + "right": "issue:44112", + "accept": false, + "reason": "Config bug about sparse layer creation vs a GraniteSpeech CI device-override test failure." + }, + { + "left": "issue:43335", + "right": "issue:43606", + "accept": false, + "reason": "Sparse-layer config bug vs a CPU-offload device mismatch in bark-small." + }, + { + "left": "issue:43335", + "right": "issue:45083", + "accept": false, + "reason": "Different models and different failure modes; not the same underlying bug." + }, + { + "left": "issue:42490", + "right": "issue:43854", + "accept": false, + "reason": "Both involve loading, but one is a trailing-slash import warning and the other is a GLM-4.7-Flash unit-test load failure." + }, + { + "left": "issue:37428", + "right": "issue:44351", + "accept": false, + "reason": "Separate import errors for different symbols: _flash_supports_window_size vs HybridCache." + }, + { + "left": "issue:45290", + "right": "issue:46326", + "accept": false, + "reason": "Both touch chat templating, but one crashes on tool-call assistant messages and the other is about continue_final_message for mistral_common tokenizers." + }, { "left": "issue:45520", "right": "issue:46170", "accept": false, - "reason": "Different failures: one is a flash_attention_utils import/key error on Python 3.13, the other is a Windows ROCm torch.distributed import/type issue." + "reason": "Unrelated platform/import failures: Python 3.13 flash-attn lookup vs Windows ROCm/torch.distributed issues." }, { - "left": "issue:44448", - "right": "issue:44743", + "left": "issue:44617", + "right": "issue:46697", "accept": false, - "reason": "Same broad version context, but different bugs: Pegasus v4/v5 output drift versus Qwen3 modular recurrent state reset with cache." + "reason": "CUDA OOM in Sam3Video is unrelated to Swin2SR image over-padding." }, { "left": "issue:45588", "right": "issue:45859", "accept": false, - "reason": "Both involve flash-attention-related code, but one is an AttributeError on s_aux=None in flash_attention.py and the other is a CPU-offload incompatibility in rotary embedding." + "reason": "Different code paths and failures: flash_attention.py AttributeError on s_aux=None vs Qwen3_5MoeTextRotaryEmbedding CPU-offload incompatibility." + }, + { + "left": "issue:44448", + "right": "issue:44743", + "accept": false, + "reason": "Pegasus v4/v5 output drift is unrelated to Qwen3.5 recurrent state reset with cache." }, { "left": "issue:43819", "right": "issue:44493", "accept": false, - "reason": "Unrelated symptoms: DAC.from_latents mismatch due to missing STE versus unexpected position-id keys affecting many models." + "reason": "DAC.from_latents/STE mismatch is unrelated to unexpected position-id key warnings." }, { "left": "issue:34689", "right": "issue:35141", "accept": false, - "reason": "Different model-loading regression versus embedding reinitialization after resize; no shared concrete bug path." - }, - { - "left": "issue:39692", - "right": "issue:43526", - "accept": false, - "reason": "Different components and bugs: SigLIP2 docs example errors versus BEiT image processor reduce_labels returning only one label." + "reason": "Model-loading regression for Llama 3.2 Vision is unrelated to post_init reinitializing untied output embeddings." }, { "left": "issue:43606", "right": "issue:44857", "accept": false, - "reason": "Different code paths and failures: CPU offload device mismatch for bark-small versus float16 AMP crash in LwDetrImageLoss." + "reason": "Different runtime failures in different paths: CPU offload device mismatch vs float16 AMP crash." }, { - "left": "issue:44206", - "right": "issue:45823", + "left": "issue:39692", + "right": "issue:43526", "accept": false, - "reason": "Different models and problems: unsupported center arg in LasrFeatureExtractor versus Gemma4 PLE forward device mismatch." + "reason": "SigLIP2 docs/example problems are not the same as BeitImageProcessorFast returning only one label." }, { "left": "issue:39692", "right": "issue:44493", "accept": false, - "reason": "Docs/example breakage for SigLIP2 is unrelated to unexpected position-id key handling." + "reason": "Docs/example mismatch and quantization failure are unrelated to position-id key warnings." + }, + { + "left": "issue:44206", + "right": "issue:45823", + "accept": false, + "reason": "Unsupported feature-extractor argument vs Gemma4 device_map forward mismatch." }, { "left": "issue:10000", "right": "issue:15354", "accept": false, - "reason": "Completely different historical issues: German DistilBertModel failure versus torch.jit.script GeneratorExp export support." + "reason": "Different models and unrelated scripting/export failures." }, { "left": "issue:29127", "right": "issue:32090", "accept": false, - "reason": "Unrelated: layoutlmv3 error-message clarity versus Trainer broadcast TypeError on NoneType." + "reason": "LayoutLMv3 error-message clarity issue is unrelated to Trainer _gpu_broadcast_one TypeError." }, { "left": "issue:44514", "right": "issue:44521", "accept": false, - "reason": "Both concern apply_chat_template, but one is a batched-input crash with padding=False while the other is incorrect all-zero assistant masks; different defects." + "reason": "Same area, but different bugs: tool-call/no-content crash vs all-zero assistant masks for multimodal inputs." }, { "left": "issue:28218", "right": "issue:32090", "accept": false, - "reason": "Tokenizer extra-space bug is unrelated to the Trainer broadcast NoneType error." + "reason": "Tokenizer extra-space bug is unrelated to Trainer broadcast TypeError." }, { "left": "issue:43720", "right": "issue:44112", "accept": false, - "reason": "Different failures: packed weights not unpacked during accelerate loading versus a stale GraniteSpeech CI test." + "reason": "Packed-weight loading bug is unrelated to the GraniteSpeech CI stale device-override test." }, { "left": "issue:43701", "right": "issue:45083", "accept": false, - "reason": "Resume-from-checkpoint key mismatch is unrelated to the qwen3_omni_moe feature-length helper behavior." + "reason": "Checkpoint key mismatch is unrelated to the qwen3_omni_moe feature-length helper behavior." }, { "left": "issue:44448", "right": "issue:44779", "accept": false, - "reason": "Pegasus v4/v5 output differences and DeepSeek tokenizer regressions are separate model-specific issues." + "reason": "Pegasus generation regression is unrelated to Deepseek tokenizer correctness in v5." }, { "left": "issue:44857", "right": "issue:45072", "accept": false, - "reason": "Both are dtype-related, but they affect different models and code paths; not the same underlying bug." + "reason": "Different dtype/runtime failures in different models: LwDetr AMP crash vs SwitchTransformers/TimmWrapper bfloat16 mismatch." }, { "left": "issue:43278", "right": "issue:43825", "accept": false, - "reason": "Embedding dtype drift during eval is unrelated to the pipeline translation-support error message." + "reason": "Training/eval dtype drift is unrelated to the pipeline translation-support error message." }, { "left": "issue:32090", "right": "issue:33357", "accept": false, - "reason": "Trainer broadcast TypeError is unrelated to the MacOS bus error with community CLIP." + "reason": "Different platform/backend failures: Trainer broadcast NoneType vs MacOS bus error with CLIP/community model." }, { "left": "issue:42490", "right": "issue:43575", "accept": false, - "reason": "Import warning for local paths ending in slash is unrelated to Qwen2-57B-A14B-Instruct tensor-parallel OOM." + "reason": "Trailing-slash path warning is unrelated to Qwen2-57B-A14B-Instruct tensor-parallel OOM." }, { "left": "issue:38175", "right": "issue:44493", "accept": false, - "reason": "SigLIP2 zero-probability output and unexpected position-id keys are unrelated bugs." - }, - { - "left": "issue:45310", - "right": "issue:45478", - "accept": true, - "reason": "Same bug: identical title and model/version, both report the Qwen3.5 MoE from_pretrained regression." + "reason": "SigLIP2 zero-probabilities bug is unrelated to unexpected position-id key warnings." }, { "left": "issue:43550", "right": "issue:45083", "accept": false, - "reason": "Bamba torch.compile SDPA failure is unrelated to the qwen3_omni_moe helper-function behavior." - }, - { - "left": "issue:26413", - "right": "issue:31501", - "accept": false, - "reason": "Training resume epoch-iterator failure is not the same as FineWeb SLM training not starting." - }, - { - "left": "issue:27957", - "right": "issue:34689", - "accept": false, - "reason": "Flash Attention 2 support issue for XLMRoberta is unrelated to Llama 3.2 Vision model loading breakage." - }, - { - "left": "issue:45127", - "right": "issue:45245", - "accept": false, - "reason": "LoRA/extended-vocab model collapse is unrelated to the categories limit runtime error." - }, - { - "left": "issue:45459", - "right": "issue:46234", - "accept": false, - "reason": "Tokenizer protobuf-error masking is unrelated to the missing import_utils symbol for falcon mamba availability." - }, - { - "left": "issue:37428", - "right": "issue:42175", - "accept": false, - "reason": "Flash attention import error and missing TensorFlow backend in torch-only install are unrelated." - }, - { - "left": "issue:29942", - "right": "issue:35141", - "accept": false, - "reason": "Failing Flash Attention 2 tests are unrelated to output embedding reinitialization after token resize." + "reason": "torch.compile/SDPA failure in Bamba is unrelated to the qwen3_omni_moe feature-length helper issue." } ] }, - "evaluator_result": { - "accept": true, - "feedback": "The summary is grounded in the packet: 45310 and 45478 have the same title, and 45310 is earlier. The non-duplicate soft-edge rejections are conservative and stay within what the titles/body summaries support." - }, - "evaluator_used": true, + "evaluator_result": null, + "evaluator_used": false, "retried": false, - "accepted_nontrivial_soft_edge": true, + "accepted_nontrivial_soft_edge": false, "error_kind": null, "error_message": null }, @@ -10406,19 +10259,21 @@ "source": "llm", "cache_hit": false, "model": "gpt-5.4-mini?service_tier=flex", - "cluster_id": "cluster-10000-529", + "cluster_id": "cluster-10000-530", "nodes": [ "issue:25251", - "issue:29127", + "issue:26413", + "issue:27957", "issue:29870", "issue:29942", - "issue:30064", + "issue:31501", "issue:32090", "issue:33290", - "issue:33357", - "issue:34567", - "issue:39692", + "issue:34689", + "issue:35141", + "issue:37428", "issue:40444", + "issue:42175", "issue:42490", "issue:43257", "issue:43278", @@ -10426,38 +10281,43 @@ "issue:43329", "issue:43381", "issue:43526", - "issue:43575", - "issue:43701", "issue:43873", "issue:43927", - "issue:44186", "issue:44360", "issue:44448", "issue:44464", - "issue:44466", "issue:44530", "issue:44561", "issue:44779", "issue:44855", - "issue:44871", "issue:44977", - "issue:45198", + "issue:45127", "issue:45229", + "issue:45245", "issue:45290", + "issue:45310", "issue:45381", + "issue:45459", + "issue:45478", "issue:45593", "issue:45698", "issue:45920", "issue:46192", + "issue:46234", "issue:46399", - "issue:46421", "issue:46424", "issue:46655", - "issue:46858", "issue:47000", "issue:47276" ], "soft_pairs": [ + "issue:45310|issue:45478", + "issue:26413|issue:31501", + "issue:27957|issue:34689", + "issue:45127|issue:45245", + "issue:45459|issue:46234", + "issue:37428|issue:42175", + "issue:29942|issue:35141", "issue:45593|issue:47000", "issue:29870|issue:32090", "issue:43278|issue:43329", @@ -10465,9 +10325,9 @@ "issue:25251|issue:32090", "issue:45290|issue:45381", "issue:29870|issue:33290", - "issue:44464|issue:46424", "issue:44448|issue:44977", "issue:32090|issue:33290", + "issue:44464|issue:46424", "issue:43278|issue:43381", "issue:40444|issue:43526", "issue:43257|issue:43299", @@ -10475,242 +10335,217 @@ "issue:44855|issue:46399", "issue:44530|issue:47276", "issue:42490|issue:43873", + "issue:43927|issue:45698", "issue:44977|issue:45229", "issue:45920|issue:46192", - "issue:43927|issue:45698", - "issue:44561|issue:44779", - "issue:29127|issue:34567", - "issue:33357|issue:39692", - "issue:43873|issue:45229", - "issue:44871|issue:44977", - "issue:43381|issue:43701", - "issue:44448|issue:44466", - "issue:44186|issue:45198", - "issue:29942|issue:30064", - "issue:46421|issue:46858", - "issue:43278|issue:43575" + "issue:44561|issue:44779" ], - "prepared_review_unit_hash": "a0876d1394d3eb56b134a644eb2441c9d49d12a445f12a6f99be8f810d231f61", + "prepared_review_unit_hash": "2635452cba784e917c5c5be1b73c007b379f1b73010a8a621c3557c8ef6b9124", "component_budget": { - "node_count": 529, - "item_count": 529, - "soft_pair_count": 1953, - "serialized_chars": 1012074, - "estimated_input_tokens": 253019, - "estimated_eval_tokens": 506294 + "node_count": 530, + "item_count": 530, + "soft_pair_count": 1954, + "serialized_chars": 1013004, + "estimated_input_tokens": 253251, + "estimated_eval_tokens": 506758 }, "budget": { "node_count": 48, "item_count": 48, - "soft_pair_count": 31, - "serialized_chars": 37201, - "estimated_input_tokens": 9301, - "estimated_eval_tokens": 18858 + "soft_pair_count": 28, + "serialized_chars": 36220, + "estimated_input_tokens": 9055, + "estimated_eval_tokens": 18366 }, "overflow_policy": "truncate_then_skip", "trimmed": false, "aggressively_trimmed": false, "split": true, "analyst_result": { - "summary": "These items are mostly unrelated bugs across training, tokenization, multimodal processing, caching, and serialization. I do not see any pair that clearly describes the same concrete underlying defect closely enough to merge or deduplicate.", - "confidence": 0.89, - "canonical_issue_reason": "issue:43873 is the strongest representative anchor: it is open, has high discussion/references, and describes a concrete runtime regression in a broad loading/offloading path.", + "summary": "Most pairs are not true duplicates; they only share broad subsystem or version-regression themes. The only clear duplicate is the Qwen3.5 MoE from_pretrained error pair (45310/45478).", + "confidence": 0.71, + "canonical_issue_reason": "issue:43873 is the most active and broadly referenced item in the set, so it is the best single anchor even though the overall cluster is thematically diverse.", "canonical_pr_reason": null, - "best_issue_reason": "issue:43873 is the best issue to keep as the representative record because it is active, well-discussed, and has a clearly actionable failure mode.", + "best_issue_reason": "issue:43873 has the strongest triage signal here: open, high discussion, and multiple inbound references, making it the most useful representative issue.", "best_pr_reason": null, "soft_edge_verdicts": [ + { + "left": "issue:45310", + "right": "issue:45478", + "accept": true, + "reason": "Same title, same Qwen3.5 MoE from_pretrained failure, and both describe the same regression on transformers>=5.4.0." + }, + { + "left": "issue:26413", + "right": "issue:31501", + "accept": false, + "reason": "Both involve training startup failures, but one is a resume_from_checkpoint epoch_iterator edge case and the other is FineWeb SLM training not starting; not the same concrete bug." + }, + { + "left": "issue:27957", + "right": "issue:34689", + "accept": false, + "reason": "Both mention model/attention issues, but they are different models and different failure modes; no clear shared code-path bug." + }, + { + "left": "issue:45127", + "right": "issue:45245", + "accept": false, + "reason": "These are unrelated: LoRA/vocabulary merging versus a category cardinality limit error." + }, + { + "left": "issue:45459", + "right": "issue:46234", + "accept": false, + "reason": "Both touch import/util error handling, but they are different missing symbols and different failure mechanisms." + }, + { + "left": "issue:37428", + "right": "issue:42175", + "accept": false, + "reason": "One is a missing flash-attention helper import, the other is packaging/backend selection for TensorFlow; unrelated." + }, + { + "left": "issue:29942", + "right": "issue:35141", + "accept": false, + "reason": "Flash Attention 2 test failures and output embedding reinitialization are different bugs in different paths." + }, { "left": "issue:45593", "right": "issue:47000", "accept": false, - "reason": "Different models and failures: auxiliary-loss handling in D-FINE vs infeasible cost matrix in RTDetrLoss." + "reason": "Both are training-loss problems, but D-FINE auxiliary loss handling and RTDetr infeasible cost matrix are distinct issues." }, { "left": "issue:29870", "right": "issue:32090", "accept": false, - "reason": "Bart generate() decoder_attention_mask bug is unrelated to Trainer broadcast NoneType handling." + "reason": "Bart generate() decoder mask bug and Trainer _gpu_broadcast_one NoneType error are unrelated code paths." }, { "left": "issue:43278", "right": "issue:43329", "accept": false, - "reason": "One is dtype drift in evaluation; the other is an undefined-symbol bug in multimodal token counting." + "reason": "Embedding dtype drift during evaluation is unrelated to the multimodal video token counting bug." }, { "left": "issue:44360", "right": "issue:44561", "accept": false, - "reason": "DSA indexer activation omission and removal of is_torch_fx_available are unrelated changes." + "reason": "A missing ReLU in DSA indexer and removal of is_torch_fx_available are unrelated changes." }, { "left": "issue:25251", "right": "issue:32090", "accept": false, - "reason": "Pipeline top_k nesting behavior is unrelated to Trainer GPU broadcast errors." + "reason": "Pipeline top_k nesting and Trainer broadcast NoneType are unrelated." }, { "left": "issue:45290", "right": "issue:45381", "accept": false, - "reason": "Chat-template tool-call crash and Qwen2.5-VL video position IDs are different code paths and symptoms." + "reason": "Both affect multimodal/chat behavior, but one crashes on tool-call messages in chat templates while the other has wrong video vision_position_ids." }, { "left": "issue:29870", "right": "issue:33290", "accept": false, - "reason": "Generation-time BART error does not match an Adafactor+DeepSpeed OOM issue." - }, - { - "left": "issue:44464", - "right": "issue:46424", - "accept": false, - "reason": "Chunked generation inconsistency with compiled forward is distinct from cache max length being ignored." + "reason": "UnboundLocalError in BART generation and DeepSpeed/Adafactor OOM are different issues." }, { "left": "issue:44448", "right": "issue:44977", "accept": false, - "reason": "Pegasus v4/v5 output drift and Qwen3.5 flash-attention generation failure are separate regressions." + "reason": "Pegasus output regression and Qwen3.5 flash-attention generation failure are different models and code paths." }, { "left": "issue:32090", "right": "issue:33290", "accept": false, - "reason": "Trainer broadcast NoneType failure and optimizer OOM are different failure classes." + "reason": "Trainer broadcast NoneType and Adafactor OOM are unrelated failures." + }, + { + "left": "issue:44464", + "right": "issue:46424", + "accept": false, + "reason": "Chunked generation inconsistency under compiled forward is unrelated to cache length handling in quantized cache." }, { "left": "issue:43278", "right": "issue:43381", "accept": false, - "reason": "Embedding dtype mismatch in eval is unrelated to gradient checkpointing being disallowed in eval mode." + "reason": "Embedding dtype change in evaluate and gradient checkpointing in eval mode are separate behaviors." }, { "left": "issue:40444", "right": "issue:43526", "accept": false, - "reason": "Qwen2.5-VL iterable-dataset multi-image failure and BEiT label reduction bug do not share a concrete bug." + "reason": "IterableDataset multi-image fine-tuning failure and BeitImageProcessorFast label reduction are unrelated." }, { "left": "issue:43257", "right": "issue:43299", "accept": false, - "reason": "Both mention Qwen3 MoE loading, but one is accelerate+deepspeed weight conversion and the other is a broader v5 loading regression; not the same concrete defect." + "reason": "Both involve Qwen3 MoE loading, but one is accelerate+deepspeed conversion and the other is a v5.0.0.dev0 load regression; not the same concrete bug." }, { "left": "issue:45290", "right": "issue:46655", "accept": false, - "reason": "Tool-call chat templating crash is not the same as SmolVLM video padding calculation." + "reason": "Assistant tool-call chat-template crash and SmolVLM video padding error are unrelated multimodal bugs." }, { "left": "issue:44855", "right": "issue:46399", "accept": false, - "reason": "Python 3.13 TorchScript parsing failure differs from a missing decorator in kernel registration." + "reason": "Both mention decorators/parsing, but one is a Python 3.13 IndentationError in DebertaV2 and the other is a Qwen3.5 kernel registration failure." }, { "left": "issue:44530", "right": "issue:47276", "accept": false, - "reason": "PagedAttentionCache crash on Qwen3.5 is unrelated to a warning-message wording issue." + "reason": "PagedAttentionCache group-type crash and a misleading synced_gpus warning are unrelated." }, { "left": "issue:42490", "right": "issue:43873", "accept": false, - "reason": "Local-path trailing-slash warning and quantization offloading behavior are not the same bug." - }, - { - "left": "issue:44977", - "right": "issue:45229", - "accept": false, - "reason": "Flash-attention generation failure and multi-GPU CUDA OOM are different problems." - }, - { - "left": "issue:45920", - "right": "issue:46192", - "accept": false, - "reason": "Wrong token IDs across models and silently ignored return_tensors are separate tokenizer API issues." + "reason": "A local-path import warning and quantization offloading failure are different issues." }, { "left": "issue:43927", "right": "issue:45698", "accept": false, - "reason": "Custom token IDs lost after save/load and wrong custom module loaded after save_pretrained are different serialization bugs." - }, - { - "left": "issue:44561", - "right": "issue:44779", - "accept": false, - "reason": "Removal of a helper breaking trust_remote_code is unrelated to DeepSeek tokenizer regression." - }, - { - "left": "issue:29127", - "right": "issue:34567", - "accept": false, - "reason": "LayoutLMv3 error-message clarity and TrainerState token counter updates are unrelated." - }, - { - "left": "issue:33357", - "right": "issue:39692", - "accept": false, - "reason": "MacOS CLIP bus error and SigLIP2 docs/model-processor mismatch are different issue types." + "reason": "Both are save/load related, but one is custom token IDs in config and the other is wrong custom module resolution; different underlying problems." }, { - "left": "issue:43873", + "left": "issue:44977", "right": "issue:45229", "accept": false, - "reason": "Quantization/offloading behavior is not the same as Gemma4 multi-GPU OOM." - }, - { - "left": "issue:44871", - "right": "issue:44977", - "accept": false, - "reason": "eos_token_id config inconsistency and flash-attention generation failure are distinct regressions." - }, - { - "left": "issue:43381", - "right": "issue:43701", - "accept": false, - "reason": "Eval-mode gradient checkpointing and resume_from_checkpoint key mismatch are different training issues." - }, - { - "left": "issue:44448", - "right": "issue:44466", - "accept": false, - "reason": "Pegasus output drift and lm_head serialization inconsistency are separate v4/v5 problems." - }, - { - "left": "issue:44186", - "right": "issue:45198", - "accept": false, - "reason": "LayoutLMv2 tokenizer NER/padding crash and Wav2Vec2 save/tokenization failure are unrelated tokenization bugs." - }, - { - "left": "issue:29942", - "right": "issue:30064", - "accept": false, - "reason": "Flash Attention 2 test failures and void segmentation map processing are different subsystems." + "reason": "Flash-attention generation failure for Qwen3.5 and Gemma4 multi-GPU OOM are unrelated." }, { - "left": "issue:46421", - "right": "issue:46858", + "left": "issue:45920", + "right": "issue:46192", "accept": false, - "reason": "Lazy cache init causing recompilation is not the same as get_seq_length returning a device tensor on non-CUDA backends." + "reason": "Tokenizer regression across many models and return_tensors being ignored are different API bugs." }, { - "left": "issue:43278", - "right": "issue:43575", + "left": "issue:44561", + "right": "issue:44779", "accept": false, - "reason": "Embedding dtype drift in eval and Qwen2-57B-A14B-Instruct OOM are unrelated." + "reason": "Removal of a utility function and Deepseek tokenizer incorrect results are not the same bug." } ] }, - "evaluator_result": null, - "evaluator_used": false, + "evaluator_result": { + "accept": true, + "feedback": "Grounded overall. The canonical issue choice is supported by the packet (43873 is open, has the highest discussion/inbound signal, and is a reasonable anchor). The soft-edge decisions are conservative, and the only accepted duplicate pair (45310/45478) is justified by identical titles and matching regression wording. No obvious overstatement beyond the evidence in the packet." + }, + "evaluator_used": true, "retried": false, - "accepted_nontrivial_soft_edge": false, + "accepted_nontrivial_soft_edge": true, "error_kind": null, "error_message": null }, @@ -10725,45 +10560,45 @@ "source": "llm", "cache_hit": false, "model": "gpt-5.4-mini?service_tier=flex", - "cluster_id": "cluster-10000-529", + "cluster_id": "cluster-10000-530", "nodes": [ "issue:29127", "issue:29870", + "issue:29942", + "issue:30064", "issue:33357", + "issue:34567", "issue:34689", - "issue:35141", + "issue:39692", "issue:41720", "issue:42490", "issue:42915", "issue:42994", - "issue:43065", - "issue:43334", + "issue:43278", "issue:43381", "issue:43412", - "issue:43526", - "issue:43531", "issue:43575", "issue:43611", "issue:43653", "issue:43701", "issue:43756", + "issue:43873", "issue:43976", - "issue:43994", "issue:44077", + "issue:44186", "issue:44206", "issue:44265", "issue:44361", - "issue:44387", "issue:44442", + "issue:44448", + "issue:44466", "issue:44479", "issue:44568", "issue:44743", + "issue:44871", "issue:44977", "issue:45198", "issue:45229", - "issue:45372", - "issue:45406", - "issue:45412", "issue:45440", "issue:45491", "issue:45563", @@ -10771,14 +10606,23 @@ "issue:45803", "issue:45823", "issue:46121", - "issue:46693", - "issue:46701", - "issue:47217", + "issue:46421", + "issue:46858", "issue:47276" ], "soft_pairs": [ + "issue:33357|issue:39692", + "issue:29127|issue:34567", + "issue:43873|issue:45229", + "issue:44871|issue:44977", + "issue:43381|issue:43701", + "issue:44448|issue:44466", + "issue:44186|issue:45198", + "issue:29942|issue:30064", + "issue:43278|issue:43575", "issue:45588|issue:45823", "issue:44361|issue:45198", + "issue:46421|issue:46858", "issue:41720|issue:45229", "issue:43976|issue:45440", "issue:29870|issue:34689", @@ -10788,234 +10632,224 @@ "issue:43653|issue:44442", "issue:44265|issue:45198", "issue:44077|issue:47276", + "issue:44977|issue:45563", "issue:43653|issue:44743", "issue:43412|issue:45588", - "issue:44977|issue:45563", "issue:44568|issue:45491", "issue:42915|issue:42994", "issue:43756|issue:45440", "issue:43575|issue:43701", "issue:29127|issue:33357", - "issue:42490|issue:43381", - "issue:43065|issue:43994", - "issue:43575|issue:44387", - "issue:43334|issue:43531", - "issue:45491|issue:46693", - "issue:29127|issue:35141", - "issue:43065|issue:45412", - "issue:43994|issue:45412", - "issue:45372|issue:45588", - "issue:43526|issue:47217", - "issue:45406|issue:46701" + "issue:42490|issue:43381" ], - "prepared_review_unit_hash": "37e09ceea0bbc613c963cbc932a094f1a7c5c39b862f94940c48c8c7edd9683d", + "prepared_review_unit_hash": "d2e620a25bc4738a49152e42b0a724421d15f8424eda1a71008c151d91a20626", "component_budget": { - "node_count": 529, - "item_count": 529, - "soft_pair_count": 1953, - "serialized_chars": 1012074, - "estimated_input_tokens": 253019, - "estimated_eval_tokens": 506294 + "node_count": 530, + "item_count": 530, + "soft_pair_count": 1954, + "serialized_chars": 1013004, + "estimated_input_tokens": 253251, + "estimated_eval_tokens": 506758 }, "budget": { - "node_count": 48, - "item_count": 48, + "node_count": 47, + "item_count": 47, "soft_pair_count": 30, - "serialized_chars": 36829, - "estimated_input_tokens": 9208, - "estimated_eval_tokens": 18672 + "serialized_chars": 36237, + "estimated_input_tokens": 9060, + "estimated_eval_tokens": 18376 }, "overflow_policy": "truncate_then_skip", "trimmed": false, "aggressively_trimmed": false, "split": true, "analyst_result": { - "summary": "The cluster is highly heterogeneous: all soft pairs appear to be unrelated issues rather than true duplicates, so no soft edge should be accepted. I picked the most active regression report as the best representative issue, but it is not a real canonical duplicate target for the rest.", - "confidence": 0.9, - "canonical_issue_reason": "issue:46693 has the strongest activity signal (high discussion and inbound references) and a broad regression framing, so it is the best representative anchor among this otherwise unrelated set.", + "summary": "All soft pairs are false positives: they mostly share broad subsystem keywords (tokenizers, quantization, cache, generation, vision) but describe different concrete bugs and code paths.", + "confidence": 0.97, + "canonical_issue_reason": null, "canonical_pr_reason": null, - "best_issue_reason": "issue:46693 is the most broadly scoped and most discussed issue in the group, making it the best overall issue representative.", + "best_issue_reason": null, "best_pr_reason": null, "soft_edge_verdicts": [ { - "left": "issue:45588", - "right": "issue:45823", + "left": "issue:33357", + "right": "issue:39692", "accept": false, - "reason": "Different bugs: flash_attention s_aux handling vs Gemma4 PLE device mismatch." + "reason": "One is a MacOS bus error loading a CLIP model; the other is a SigLIP2 docs/example issue with model/processor mismatch and quantization." }, { - "left": "issue:44361", - "right": "issue:45198", + "left": "issue:29127", + "right": "issue:34567", "accept": false, - "reason": "Different tokenizers and failures: MLuke AttributeError vs Wav2Vec2 save/tokenization issues." + "reason": "LayoutLMv3 box-input error messaging is unrelated to TrainerState token-count tracking." }, { - "left": "issue:41720", + "left": "issue:43873", "right": "issue:45229", "accept": false, - "reason": "Different symptoms and models: Qwen3 device-map assert vs Gemma4 multi-GPU OOM." + "reason": "Both mention memory/offloading, but one is a quantization offloading behavior bug and the other is a Gemma4 multi-GPU OOM case." }, { - "left": "issue:43976", - "right": "issue:45440", + "left": "issue:44871", + "right": "issue:44977", "accept": false, - "reason": "Unrelated issues: Python version compatibility vs DeepseekV3 implementation divergence." + "reason": "Gemma eos_token_id mismatch is a config inconsistency, while Qwen3.5 flash-attention generation is a runtime generation bug." }, { - "left": "issue:29870", - "right": "issue:34689", + "left": "issue:43381", + "right": "issue:43701", "accept": false, - "reason": "Different code paths: BART generation mask bug vs Llama model loading regression." + "reason": "Eval-mode gradient checkpointing and resume_from_checkpoint key mismatch are different training-state issues." }, { - "left": "issue:44206", - "right": "issue:44479", + "left": "issue:44448", + "right": "issue:44466", "accept": false, - "reason": "Different subsystems: feature extractor arg handling vs Qwen video-input regression." + "reason": "One reports differing outputs across v4/v5; the other is inconsistent lm_head.weight serialization due to tied weights/device state." }, { - "left": "issue:43611", - "right": "issue:46121", + "left": "issue:44186", + "right": "issue:45198", "accept": false, - "reason": "Different failures: base_model_prefix loading vs rope-parameter JSON list validation." + "reason": "LayoutLMv2 tokenizer crashes on NER/padding, while Wav2Vec2 has save_pretrained/tokenization failures." }, { - "left": "issue:45198", - "right": "issue:45803", + "left": "issue:29942", + "right": "issue:30064", "accept": false, - "reason": "Different bug classes: Wav2Vec2 save/tokenization failure vs Fuyu batch tensor exception handling." + "reason": "Flash Attention 2 test failures are unrelated to void segmentation-map processing in image preprocessing." }, { - "left": "issue:43653", - "right": "issue:44442", + "left": "issue:43278", + "right": "issue:43575", "accept": false, - "reason": "Both tokenizer-related, but different models and failures: special-token registration vs AutoTokenizer loading." + "reason": "Embedding dtype drift in evaluation is unrelated to a tensor-parallel OOM when loading Qwen2-57B." }, { - "left": "issue:44265", - "right": "issue:45198", + "left": "issue:45588", + "right": "issue:45823", "accept": false, - "reason": "Different APIs and bugs: torch.export with torch_compilable_check vs Wav2Vec2 save/tokenization." + "reason": "Flash-attention s_aux=None crash is a different code path from Gemma4 device_map auto forward-device mismatch." }, { - "left": "issue:44077", - "right": "issue:47276", + "left": "issue:44361", + "right": "issue:45198", "accept": false, - "reason": "Unrelated: patchtsmixer post_init policy vs a stale synced_gpus warning message." + "reason": "MLukeTokenizer AttributeError on tasks is unrelated to Wav2Vec2 save/tokenization failures." }, { - "left": "issue:43653", - "right": "issue:44743", + "left": "issue:46421", + "right": "issue:46858", "accept": false, - "reason": "Different models and paths: BigBirdTokenizer decode issue vs Qwen3_5 recurrent-state reset." + "reason": "Both are cache-related, but one is lazy-init recompilation behavior and the other is a device-tensor seq_length arithmetic bug." }, { - "left": "issue:43412", - "right": "issue:45588", + "left": "issue:41720", + "right": "issue:45229", "accept": false, - "reason": "Different export/runtime issues: Gemma3n executorch failure vs flash_attention crash." + "reason": "Qwen3 auto device mapping cudaErrorAssert is not the same as Gemma4 multi-GPU OOM." }, { - "left": "issue:44977", - "right": "issue:45563", + "left": "issue:43976", + "right": "issue:45440", "accept": false, - "reason": "Different generate-time problems: flash-attention behavior vs a stale warning for num_return_sequences." + "reason": "Python-version compatibility breakage is unrelated to DeepseekV3MoE diverging from the remote implementation." }, { - "left": "issue:44568", - "right": "issue:45491", + "left": "issue:29870", + "right": "issue:34689", "accept": false, - "reason": "Different tokenization vs attention/numerics bug; no shared concrete code path." + "reason": "BART generate decoder_attention_mask initialization and Llama 3.2 Vision model loading failure are different failures." }, { - "left": "issue:42915", - "right": "issue:42994", + "left": "issue:44206", + "right": "issue:44479", "accept": false, - "reason": "Both quantization-related, but one is FP8 config failure and the other is quantized model saving." + "reason": "Unsupported center arg in a feature extractor is not the same as the Qwen2_5/Qwen3 video-input regression." }, { - "left": "issue:43756", - "right": "issue:45440", + "left": "issue:43611", + "right": "issue:46121", "accept": false, - "reason": "Different model families and failures: Smollm3 RoPE-layer bug vs DeepSeekV3 divergence." + "reason": "base_model_prefix loading regression and rope-parameter dict conversion TypeError are separate loading/config issues." }, { - "left": "issue:43575", - "right": "issue:43701", + "left": "issue:45198", + "right": "issue:45803", "accept": false, - "reason": "Different problems: tensor-parallel OOM when loading vs resume_from_checkpoint key mismatch." + "reason": "Wav2Vec2 tokenization/save failures do not match Fuyu's bare-except bug swallowing errors." }, { - "left": "issue:29127", - "right": "issue:33357", + "left": "issue:43653", + "right": "issue:44442", "accept": false, - "reason": "Different models and symptoms: layoutlmv3 error messaging vs CLIP bus error on macOS." + "reason": "BigBirdTokenizer special-token registration and AutoTokenizer loading FastSpeech2ConformerTokenizer are different tokenizer bugs." }, { - "left": "issue:42490", - "right": "issue:43381", + "left": "issue:44265", + "right": "issue:45198", "accept": false, - "reason": "Unrelated: path trailing-slash import warning vs gradient checkpointing eval-mode restriction." + "reason": "torch.export failures with torch_compilable_check are unrelated to Wav2Vec2 save/tokenization failures." }, { - "left": "issue:43065", - "right": "issue:43994", + "left": "issue:44077", + "right": "issue:47276", "accept": false, - "reason": "Different models and bugs: Sam3PixelDecoder dummy layer vs SigLIP2 nonsensical outputs." + "reason": "patchtsmixer post_init allowance and a misleading synced_gpus warning are unrelated." }, { - "left": "issue:43575", - "right": "issue:44387", + "left": "issue:44977", + "right": "issue:45563", "accept": false, - "reason": "Both OOM-related, but different causes: tp loading OOM vs int4 quantization memory regression." + "reason": "Flash-attention generation failure is not the same as the stale num_return_sequences warning in paged generate()." }, { - "left": "issue:43334", - "right": "issue:43531", + "left": "issue:43653", + "right": "issue:44743", "accept": false, - "reason": "Different Qwen3-VL config load failure vs Qwen3-MoE sliding-window logic bug." + "reason": "BigBirdTokenizer special-token registration and Qwen3.5 recurrent-state reset are unrelated bugs." }, { - "left": "issue:45491", - "right": "issue:46693", + "left": "issue:43412", + "right": "issue:45588", "accept": false, - "reason": "Related to generation performance, but one is Gemma3 NaNs and the other is a FA2 performance regression." + "reason": "Executorch export/training-guard/erfinv issues are unrelated to flash-attention s_aux=None crashes." }, { - "left": "issue:29127", - "right": "issue:35141", + "left": "issue:44568", + "right": "issue:45491", "accept": false, - "reason": "Different behaviors: layoutlmv3 box validation message vs embedding reinitialization after resize." + "reason": "mdeberta add_special_tokens BOS/EOS behavior is unrelated to Gemma3 NaN embeddings with mixed-length batching." }, { - "left": "issue:43065", - "right": "issue:45412", + "left": "issue:42915", + "right": "issue:42994", "accept": false, - "reason": "Different issues: dummy Conv2d implementation vs RT-DETR memory not being released." + "reason": "Qwen3Moe FP8 config failure and quantized-model saving failure are different quantization-related problems." }, { - "left": "issue:43994", - "right": "issue:45412", + "left": "issue:43756", + "right": "issue:45440", "accept": false, - "reason": "Different models and failure modes: SigLIP2 inference quality vs RT-DETR memory lifecycle." + "reason": "Smollm3 RoPE-layer mismatch and DeepseekV3MoE divergence are distinct model implementation issues." }, { - "left": "issue:45372", - "right": "issue:45588", + "left": "issue:43575", + "right": "issue:43701", "accept": false, - "reason": "Different breakages: Gemma4 processor import dependency vs flash_attention None-handling crash." + "reason": "Tensor-parallel OOM on model load is unrelated to resume_from_checkpoint key mismatch." }, { - "left": "issue:43526", - "right": "issue:47217", + "left": "issue:29127", + "right": "issue:33357", "accept": false, - "reason": "Different processing bugs: BeitImageProcessorFast label reduction vs VL image-placeholder parsing." + "reason": "LayoutLMv3 error clarity and CLIP bus error on MacOS are unrelated model/runtime bugs." }, { - "left": "issue:45406", - "right": "issue:46701", + "left": "issue:42490", + "right": "issue:43381", "accept": false, - "reason": "Different runtime failures: Gemma4 serve tokenizer attribute error vs DiffusionGemma pipeline incompatibility." + "reason": "Path-with-trailing-slash import warning and gradient checkpointing in eval mode are different issues." } ] }, @@ -11037,7 +10871,7 @@ "source": "llm", "cache_hit": false, "model": "gpt-5.4-mini?service_tier=flex", - "cluster_id": "cluster-10000-529", + "cluster_id": "cluster-10000-530", "nodes": [ "issue:27957", "issue:29127", @@ -11045,22 +10879,24 @@ "issue:34689", "issue:35141", "issue:42915", + "issue:43065", "issue:43066", - "issue:43316", + "issue:43334", "issue:43381", "issue:43421", "issue:43526", + "issue:43531", + "issue:43575", "issue:43824", "issue:43873", "issue:43931", "issue:43994", "issue:44279", + "issue:44387", "issue:44442", "issue:44462", "issue:44514", "issue:44625", - "issue:44743", - "issue:44811", "issue:44857", "issue:44898", "issue:45072", @@ -11079,27 +10915,36 @@ "issue:45593", "issue:45712", "issue:45834", - "issue:46121", "issue:46369", - "issue:46399", "issue:46640", "issue:46693", + "issue:46701", "issue:46821", - "issue:47312" + "issue:47217" ], "soft_pairs": [ + "issue:43065|issue:43994", + "issue:43575|issue:44387", + "issue:43334|issue:43531", + "issue:45491|issue:46693", + "issue:29127|issue:35141", + "issue:43065|issue:45412", + "issue:43994|issue:45412", + "issue:45372|issue:45588", + "issue:43526|issue:47217", + "issue:45406|issue:46701", "issue:44442|issue:45072", "issue:44625|issue:45584", "issue:45440|issue:45593", - "issue:45459|issue:45712", "issue:43421|issue:45491", + "issue:45459|issue:45712", "issue:45412|issue:46369", "issue:43526|issue:45834", "issue:45372|issue:45406", "issue:27957|issue:33290", "issue:43066|issue:44462", - "issue:45250|issue:46640", "issue:33290|issue:34689", + "issue:45250|issue:46640", "issue:45245|issue:45563", "issue:45372|issue:45459", "issue:33290|issue:35141", @@ -11112,207 +10957,246 @@ "issue:42915|issue:43931", "issue:45588|issue:46821", "issue:44514|issue:45381", - "issue:44442|issue:44898", - "issue:44743|issue:44811", - "issue:46121|issue:46399", - "issue:43316|issue:47312" + "issue:44442|issue:44898" ], - "prepared_review_unit_hash": "9e9030591f8677d4b6293965ca440aedf87758defc9c421ada598b6bbaebad21", + "prepared_review_unit_hash": "4c0ec7684000661f88a854263d276b1f085272082b3829aca68d721a88159af9", "component_budget": { - "node_count": 529, - "item_count": 529, - "soft_pair_count": 1953, - "serialized_chars": 1012074, - "estimated_input_tokens": 253019, - "estimated_eval_tokens": 506294 + "node_count": 530, + "item_count": 530, + "soft_pair_count": 1954, + "serialized_chars": 1013004, + "estimated_input_tokens": 253251, + "estimated_eval_tokens": 506758 }, "budget": { - "node_count": 47, - "item_count": 47, - "soft_pair_count": 28, - "serialized_chars": 35344, - "estimated_input_tokens": 8836, - "estimated_eval_tokens": 17928 + "node_count": 48, + "item_count": 48, + "soft_pair_count": 35, + "serialized_chars": 37927, + "estimated_input_tokens": 9482, + "estimated_eval_tokens": 19220 }, "overflow_policy": "truncate_then_skip", "trimmed": false, "aggressively_trimmed": false, "split": true, "analyst_result": { - "summary": "This set is not a true duplicate cluster: most items are unrelated model-, tokenizer-, or runtime-specific issues that only share superficial keywords. The closest theme is Flash Attention 2, but even those reports describe different problems and are not mergeable as one fix.", - "confidence": 0.84, - "canonical_issue_reason": "If one representative must be chosen, #46640 is the broadest current Flash Attention 2 report in the set and best matches the cluster's superficial theme.", + "summary": "The soft links are mostly false positives: they overlap on broad themes like OOM, Flash Attention, Qwen, or tokenizer handling, but each pair points to a different concrete bug, model, or code path. I would not merge any of these pairs as duplicates.", + "confidence": 0.38, + "canonical_issue_reason": "issue:33290 is the most central item in the noisy memory/OOM subset and has the strongest soft-link connectivity, so it is the safest representative if one must be chosen.", "canonical_pr_reason": null, - "best_issue_reason": "#46640 is the most generic and current issue among the candidates, so it is the least misleading representative for this noisy cluster.", + "best_issue_reason": "issue:33290 is the strongest standalone bug report and broadest duplicate target among the issues here, but the linked items still do not look like true duplicates.", "best_pr_reason": null, "soft_edge_verdicts": [ + { + "left": "issue:43065", + "right": "issue:43994", + "accept": false, + "reason": "Different failures: dummy Conv2d placeholder in Sam3PixelDecoder vs Siglip2 loading/inference producing nonsense." + }, + { + "left": "issue:43575", + "right": "issue:44387", + "accept": false, + "reason": "Both mention OOM, but one is tensor-parallel model loading and the other is int4 quantization reserved-memory growth." + }, + { + "left": "issue:43334", + "right": "issue:43531", + "accept": false, + "reason": "Qwen3-VL load crash from missing pad_token_id is unrelated to Qwen3-MoE sliding_window behavior." + }, + { + "left": "issue:45491", + "right": "issue:46693", + "accept": false, + "reason": "Both involve FA2/long context, but one is a NaN correctness bug and the other is a performance regression." + }, + { + "left": "issue:29127", + "right": "issue:35141", + "accept": false, + "reason": "LayoutLMV3 error-message clarity is unrelated to token-embedding reinit when resizing embeddings." + }, + { + "left": "issue:43065", + "right": "issue:45412", + "accept": false, + "reason": "Dummy module placeholder and RT-DETR memory leak are different code paths and symptoms." + }, + { + "left": "issue:43994", + "right": "issue:45412", + "accept": false, + "reason": "Siglip2 inference quality issue does not match RT-DETR object lifetime / GC memory release." + }, + { + "left": "issue:45372", + "right": "issue:45588", + "accept": false, + "reason": "Gemma 4 processor import failure is unrelated to flash_attention.py crashing on s_aux=None." + }, + { + "left": "issue:43526", + "right": "issue:47217", + "accept": false, + "reason": "BeitImageProcessorFast label reduction bug is unrelated to VL image-placeholder handling in user text." + }, + { + "left": "issue:45406", + "right": "issue:46701", + "accept": false, + "reason": "Gemma4Processor missing _tokenizer is a processor-loading bug, not the DiffusionGemmaGenerationOutput pipeline failure." + }, { "left": "issue:44442", "right": "issue:45072", "accept": false, - "reason": "Different failures: AutoTokenizer loading a tokenizer class vs dtype mismatches in bfloat16 inference." + "reason": "Tokenizer loading failure and dtype mismatch in bfloat16 inference are distinct issues." }, { "left": "issue:44625", "right": "issue:45584", "accept": false, - "reason": "Unrelated model families and code paths: Qwen3.5 label propagation vs Whisper empty-transcription generation." + "reason": "num_labels propagation in Qwen3.5 is unrelated to Whisper empty-transcription generation after align_special_tokens." }, { "left": "issue:45440", "right": "issue:45593", "accept": false, - "reason": "Different bugs in different models: DeepSeekV3MoE divergence vs D-FINE auxiliary-loss behavior." + "reason": "DeepseekV3MoE implementation divergence and D-FINE auxiliary-loss omission are different model bugs." }, { - "left": "issue:45459", - "right": "issue:45712", + "left": "issue:43421", + "right": "issue:45491", "accept": false, - "reason": "Tokenizer protobuf error masking vs leftover dummy classes; no shared underlying bug." + "reason": "Tokenizer post-processor update logic does not match Gemma3 NaN embeddings under mixed-length batching." }, { - "left": "issue:43421", - "right": "issue:45491", + "left": "issue:45459", + "right": "issue:45712", "accept": false, - "reason": "Runtime special-token post-processing update bug vs Gemma3 NaN embeddings on mixed-length GPU batches." + "reason": "Masking tokenizer errors without protobuf and cleanup of dummy PT classes are unrelated maintenance bugs." }, { "left": "issue:45412", "right": "issue:46369", "accept": false, - "reason": "RT-DETR memory leak on deletion is distinct from sam3-litetext memory overflow during use." + "reason": "RT-DETR GC memory leak is not the same as sam3-litetext memory overflow." }, { "left": "issue:43526", "right": "issue:45834", "accept": false, - "reason": "Different image-processor bugs: label reduction shape vs OCR long-input index error." + "reason": "Beit reduce_labels bug and Kosmos2.5 OCR index error affect different preprocessing paths." }, { "left": "issue:45372", "right": "issue:45406", "accept": false, - "reason": "Different Gemma-related failures: missing dependency import vs serve-time missing _tokenizer attribute." + "reason": "Both are Gemma 4-related, but one is a mistral_common import dependency failure and the other is a missing processor attribute." }, { "left": "issue:27957", "right": "issue:33290", "accept": false, - "reason": "FlashAttention 2 support issue and Adafactor OOM are unrelated." + "reason": "XLMRoberta Flash Attention 2 support is unrelated to Adafactor+DeepSpeed OOM." }, { "left": "issue:43066", "right": "issue:44462", "accept": false, - "reason": "Both concern tokenizers, but one is decoder type handling in v5 and the other is repository tokenizer.json loading." + "reason": "Wrong decoder type in Transformers v5 is a different tokenizer problem than ignoring tokenizer.json from the repo." }, { - "left": "issue:45250", - "right": "issue:46640", + "left": "issue:33290", + "right": "issue:34689", "accept": false, - "reason": "Both mention Flash Attention 2, but one is a generic topic title and the other is a performance regression; not the same concrete bug." + "reason": "OOM in deepspeed/Adafactor does not match model-loading breakage for Llama 3.2 Vision." }, { - "left": "issue:33290", - "right": "issue:34689", + "left": "issue:45250", + "right": "issue:46640", "accept": false, - "reason": "Adafactor/deepspeed OOM is unrelated to model-loading breakage for Llama 3.2 Vision." + "reason": "These are only generic Flash Attention 2 titles; no concrete shared bug path is evident." }, { "left": "issue:45245", "right": "issue:45563", "accept": false, - "reason": "Category-count overflow and stale warning for num_return_sequences are unrelated." + "reason": "Category-limit runtime error and stale generate() warning are unrelated." }, { "left": "issue:45372", "right": "issue:45459", "accept": false, - "reason": "Different tokenizer-loading failures: missing ReasoningEffort import vs protobuf decode error handling." + "reason": "ImportError in Gemma 4 loading and protobuf-related tokenizer error masking are different failure modes." }, { "left": "issue:33290", "right": "issue:35141", "accept": false, - "reason": "OOM with Adafactor/deepspeed is unrelated to embedding reinitialization after resize." + "reason": "OOM under Adafactor/DeepSpeed is unrelated to embedding reinitialization after resizing." }, { "left": "issue:27957", "right": "issue:29127", "accept": false, - "reason": "FlashAttention 2 support for XLMRoberta and LayoutLMv3 box-validation messaging are different bugs." + "reason": "Flash Attention 2 with XLMRoberta and LayoutLMV3 error messaging have no shared bug." }, { "left": "issue:43526", "right": "issue:46693", "accept": false, - "reason": "Image label reduction bug and FA2 long-context performance regression are not the same code-path problem." + "reason": "Beit label handling and FA2 long-context performance regression are unrelated." }, { "left": "issue:43381", "right": "issue:43873", "accept": false, - "reason": "Gradient-checkpointing eval-mode behavior is unrelated to quantization offloading." + "reason": "Gradient checkpointing in eval mode and quantization offloading are different issues." }, { "left": "issue:43824", "right": "issue:44279", "accept": false, - "reason": "ImportError for a specific Qwen2.5-VL class is not the same as a vague dependency issue." + "reason": "Qwen2_5_VL import failure and a generic dependency issue are too broad to be the same bug." }, { "left": "issue:43994", "right": "issue:46369", "accept": false, - "reason": "Siglip2 auto-model/pipeline output issue is unrelated to sam3-litetext memory overflow." + "reason": "Siglip2 nonsense outputs and SAM3 lite-text memory overflow are unrelated." }, { "left": "issue:44857", "right": "issue:45588", "accept": false, - "reason": "Different crash sites: LwDetr float16 AMP loss issue vs flash_attention.py s_aux=None AttributeError." + "reason": "LwDetr float16 AMP crash and flash_attention s_aux=None are different code paths." }, { "left": "issue:42915", "right": "issue:43931", "accept": false, - "reason": "Different Qwen model bugs: FineGrainedFP8Config failure vs weight-shape mismatch on Qwen3-VL loading." + "reason": "Qwen3Moe FineGrainedFP8Config failure and Qwen3-VL shape mismatch are different model-loading problems." }, { "left": "issue:45588", "right": "issue:46821", "accept": false, - "reason": "Flash-attention sink handling and TimesFM window_size are unrelated." + "reason": "flash_attention.py s_aux=None crash is unrelated to TimesFM window_size AttributeError." }, { "left": "issue:44514", "right": "issue:45381", "accept": false, - "reason": "Same model family, but different failures: batched chat-template padding crash vs wrong video vision_position_ids." + "reason": "Qwen2.5-VL batched chat-template padding bug is distinct from video vision_position_ids being wrong." }, { "left": "issue:44442", "right": "issue:44898", "accept": false, - "reason": "Tokenizer loading failure and Perceiver interpolation failure are unrelated." - }, - { - "left": "issue:44743", - "right": "issue:44811", - "accept": false, - "reason": "Recurrent-state cache reset and Whisper batch_decode special-token handling are different behaviors." - }, - { - "left": "issue:46121", - "right": "issue:46399", - "accept": false, - "reason": "Rope-validation TypeError and missing apply_rotary_pos_emb decorator are separate bugs." - }, - { - "left": "issue:43316", - "right": "issue:47312", - "accept": false, - "reason": "Gemma3 config API mismatch and kernel-version incompatibility are unrelated." + "reason": "FastSpeech2ConformerTokenizer load failure is unrelated to Perceiver non-default-resolution behavior." } ] }, @@ -11334,7 +11218,7 @@ "source": "llm", "cache_hit": false, "model": "gpt-5.4-mini?service_tier=flex", - "cluster_id": "cluster-10000-529", + "cluster_id": "cluster-10000-530", "nodes": [ "issue:25251", "issue:27957", @@ -11366,37 +11250,40 @@ "issue:44442", "issue:44479", "issue:44743", - "issue:44855", + "issue:44811", "issue:44871", "issue:44913", "issue:44964", "issue:44977", "issue:44987", - "issue:45137", - "issue:45216", "issue:45245", "issue:45325", "issue:45440", "issue:45759", "issue:45859", "issue:45907", + "issue:46121", "issue:46178", "issue:46234", - "issue:46506", - "issue:46697" + "issue:46399", + "issue:46697", + "issue:47312" ], "soft_pairs": [ + "issue:44743|issue:44811", + "issue:43316|issue:47312", "issue:45325|issue:46697", + "issue:46121|issue:46399", "issue:43646|issue:44077", - "issue:43262|issue:44206", "issue:44977|issue:45245", + "issue:43262|issue:44206", "issue:43761|issue:45859", "issue:43720|issue:44442", "issue:25251|issue:28282", "issue:27957|issue:29870", - "issue:42175|issue:46234", "issue:43873|issue:44871", "issue:44077|issue:44479", + "issue:42175|issue:46234", "issue:42490|issue:44743", "issue:43643|issue:44913", "issue:43493|issue:43827", @@ -11404,206 +11291,203 @@ "issue:44062|issue:45907", "issue:43749|issue:43931", "issue:44987|issue:46178", - "issue:43316|issue:43525", "issue:42617|issue:43575", + "issue:43316|issue:43525", "issue:44964|issue:45440", "issue:43550|issue:43701", "issue:43575|issue:43742", - "issue:43606|issue:43701", - "issue:44855|issue:45137", - "issue:43701|issue:46506", - "issue:43931|issue:45216" + "issue:43606|issue:43701" ], - "prepared_review_unit_hash": "bc4397e407db80ac28e7f0c9c903987258dda5b8f56675dd7b16f713edd39f24", + "prepared_review_unit_hash": "60b59aafa40aebae84e24f92e55a2e488a1a44babcf37c84eaf5c907ede74726", "component_budget": { - "node_count": 529, - "item_count": 529, - "soft_pair_count": 1953, - "serialized_chars": 1012074, - "estimated_input_tokens": 253019, - "estimated_eval_tokens": 506294 + "node_count": 530, + "item_count": 530, + "soft_pair_count": 1954, + "serialized_chars": 1013004, + "estimated_input_tokens": 253251, + "estimated_eval_tokens": 506758 }, "budget": { "node_count": 48, "item_count": 48, "soft_pair_count": 27, - "serialized_chars": 35831, - "estimated_input_tokens": 8958, - "estimated_eval_tokens": 18172 + "serialized_chars": 35801, + "estimated_input_tokens": 8951, + "estimated_eval_tokens": 18158 }, "overflow_policy": "truncate_then_skip", "trimmed": false, "aggressively_trimmed": false, "split": true, "analyst_result": { - "summary": "The cluster is mostly a set of unrelated issue reports from different subsystems. The soft-similarity pairs share broad themes at best, but none look like the same underlying bug or a single mergeable fix.", - "confidence": 0.94, - "canonical_issue_reason": null, + "summary": "The soft pairs are mostly superficial overlaps by subsystem or wording, but they describe different bugs, configs, models, or docs changes. I would not merge any of them as duplicates.", + "confidence": 0.98, + "canonical_issue_reason": "No single canonical issue fits this set: the items span unrelated reports. If forced to pick a representative anchor, issue 43873 is the broadest and most active runtime bug report, but it does not subsume the others.", "canonical_pr_reason": null, - "best_issue_reason": "issue:43749 is the most discussion-heavy and referenced item here, so it works best as a triage anchor, but there is no true cluster-wide duplicate target.", + "best_issue_reason": "Issue 43873 is the best representative only in the sense that it is broad, active, and cross-cutting; however, it is not a true duplicate candidate for the rest of the cluster.", "best_pr_reason": null, "soft_edge_verdicts": [ + { + "left": "issue:44743", + "right": "issue:44811", + "accept": false, + "reason": "Different bugs: recurrent-state reset in Qwen3.5 vs Whisper batch_decode ignoring skip_special_tokens." + }, + { + "left": "issue:43316", + "right": "issue:47312", + "accept": false, + "reason": "Gemma3TextConfig API mismatch is unrelated to a generic kernels 0.16 compatibility complaint." + }, { "left": "issue:45325", "right": "issue:46697", "accept": false, - "reason": "Both concern input/position handling in vision-style preprocessors, but they are different bugs in different models and code paths." + "reason": "Qwen2.5-VL rope-index scaling bug is unrelated to Swin2SR over-padding." }, { - "left": "issue:43646", - "right": "issue:44077", + "left": "issue:46121", + "right": "issue:46399", "accept": false, - "reason": "One is about custom model initialization in v5, the other about `post_init` on a specific architecture; not the same defect." + "reason": "One is rope-validation TypeError; the other is a Qwen3.5 kernel registration crash." }, { - "left": "issue:43262", - "right": "issue:44206", + "left": "issue:43646", + "right": "issue:44077", "accept": false, - "reason": "Audio chat-template sampling-rate behavior and a feature extractor `center` argument crash are unrelated processor issues." + "reason": "Both touch initialization, but one is a broad v5 regression and the other is a model-specific post_init policy issue." }, { "left": "issue:44977", "right": "issue:45245", "accept": false, - "reason": "A generation regression in Qwen3.5 is unrelated to an embedding cardinality limit error." + "reason": "Flash-attention generation failure and category-cardinality limit are unrelated." + }, + { + "left": "issue:43262", + "right": "issue:44206", + "accept": false, + "reason": "Audio chat-template sampling-rate default and feature-extractor center-arg crash are different processor bugs." }, { "left": "issue:43761", "right": "issue:45859", "accept": false, - "reason": "These are different regressions in different models: CLIP hidden states vs Qwen3_5Moe CPU-offload rotary embedding." + "reason": "CLIPVision hidden_states regression is unrelated to Qwen3_5Moe CPU offload incompatibility." }, { "left": "issue:43720", "right": "issue:44442", "accept": false, - "reason": "Packed-weight loading in BitNet and tokenizer loading for FastSpeech2Conformer are different code paths and failures." + "reason": "BitNet packed-weight loading and FastSpeech2Conformer tokenizer loading are unrelated code paths." }, { "left": "issue:25251", "right": "issue:28282", "accept": false, - "reason": "Pipeline `top_k` nesting and missing PyTorch for AutoModel are unrelated." + "reason": "Pipeline output-shape bug is unrelated to missing PyTorch import errors." }, { "left": "issue:27957", "right": "issue:29870", "accept": false, - "reason": "FlashAttention support for XLM-Roberta and Bart `generate()` mask initialization are separate bugs." - }, - { - "left": "issue:42175", - "right": "issue:46234", - "accept": false, - "reason": "Backend packaging for `.[torch]` and a missing import_utils symbol are not the same issue." + "reason": "Flash Attention 2 model support and BART generate() unbound variable are different failures." }, { "left": "issue:43873", "right": "issue:44871", "accept": false, - "reason": "Quantized offloading behavior and Gemma EOS-token config mismatch are distinct problems." + "reason": "Quantization/offloading behavior does not match an eos_token_id config inconsistency." }, { "left": "issue:44077", "right": "issue:44479", "accept": false, - "reason": "`post_init` optionality and a Qwen video-input regression do not describe the same failure." + "reason": "Patchtsmixer post_init policy and Qwen video-input regression are unrelated." + }, + { + "left": "issue:42175", + "right": "issue:46234", + "accept": false, + "reason": "Backend packaging mismatch and missing import_utils symbol are different issues." }, { "left": "issue:42490", "right": "issue:44743", "accept": false, - "reason": "A trailing-slash import warning and recurrent-state reset logic are unrelated." + "reason": "Local-path import warning is unrelated to recurrent-state reset in Qwen3.5." }, { "left": "issue:43643", "right": "issue:44913", "accept": false, - "reason": "`trust_remote_code` missing fields and GPTNeoX config reload behavior are different config bugs." + "reason": "Missing fields from trust_remote_code config loading is not the same as rotary_pct not persisting on reload." }, { "left": "issue:43493", "right": "issue:43827", "accept": false, - "reason": "SigLIP2 implementation drift and docs still mentioning pipeline() are unrelated." + "reason": "SigLIP2 implementation mismatch and docs still referencing pipeline() are unrelated." }, { "left": "issue:43316", "right": "issue:45759", "accept": false, - "reason": "Both involve config handling, but one is a Gemma3TextConfig API discrepancy and the other is `from_config` unwrapping for Qwen multimodal configs; not the same bug." + "reason": "Gemma3TextConfig API discrepancy and Qwen multimodal text_config unwrapping are different composite-config bugs." }, { "left": "issue:44062", "right": "issue:45907", "accept": false, - "reason": "Tokenizer AddedToken keyword duplication and `list_repo_files` revision forwarding are unrelated." + "reason": "AddedToken keyword collision is unrelated to list_repo_files missing revision forwarding." }, { "left": "issue:43749", "right": "issue:43931", "accept": false, - "reason": "FSDP loading breakage and Qwen3-VL weight-shape mismatch are distinct model-loading failures." + "reason": "FSDP CPU RAM loading breakage and Qwen3-VL shape mismatch are different loading problems." }, { "left": "issue:44987", "right": "issue:46178", "accept": false, - "reason": "A model-loading regression for one repo and an unsupported `tensorflow_text` backend are different issues." + "reason": "Transformers version loading failure and tensorflow_text backend registration error are unrelated." }, { - "left": "issue:43316", - "right": "issue:43525", + "left": "issue:42617", + "right": "issue:43575", "accept": false, - "reason": "These are both config-related, but the concrete bugs differ: Gemma3TextConfig API mismatch vs missing `pad_token_id` on Llama4Config." + "reason": "3d_parallel.py execution failure and Qwen2-57B tensor-parallel OOM are different symptoms and code paths." }, { - "left": "issue:42617", - "right": "issue:43575", + "left": "issue:43316", + "right": "issue:43525", "accept": false, - "reason": "A script execution failure and OOM during tensor-parallel loading are not the same underlying bug." + "reason": "Both are config-related, but Gemma3TextConfig API differences and missing pad_token_id on Llama4Config are not the same bug." }, { "left": "issue:44964", "right": "issue:45440", "accept": false, - "reason": "Phi-4 multimodal loading and DeepseekV3MoE divergence are different model implementations and failure modes." + "reason": "Phi-4 multimodal loading failure and DeepSeekV3MoE implementation divergence are separate model-specific issues." }, { "left": "issue:43550", "right": "issue:43701", "accept": false, - "reason": "`torch.compile` SDPA failure in Bamba and a resume-from-checkpoint key mismatch are unrelated." + "reason": "torch.compile/SDPA failure in Bamba is unrelated to resume_from_checkpoint key mismatch." }, { "left": "issue:43575", "right": "issue:43742", "accept": false, - "reason": "OOM on large-model tensor parallel load and a KeyError while loading MobileLLM-125M are different loading problems." + "reason": "OOM on large Qwen2 model under tensor parallelism is unrelated to a key error while loading MobileLLM-125M." }, { "left": "issue:43606", "right": "issue:43701", "accept": false, - "reason": "CPU offload device mismatch and checkpoint key mismatch do not share the same code path." - }, - { - "left": "issue:44855", - "right": "issue:45137", - "accept": false, - "reason": "Python 3.13 TorchScript parsing and DeepSpeed ZeRO3 deque underflow are unrelated runtime errors." - }, - { - "left": "issue:43701", - "right": "issue:46506", - "accept": false, - "reason": "Checkpoint key mismatch and CPU/CUDA output divergence are unrelated." - }, - { - "left": "issue:43931", - "right": "issue:45216", - "accept": false, - "reason": "Weight-shape mismatch on load and a broken save_pretrained checkpoint are different persistence bugs." + "reason": "CPU offload device mismatch and checkpoint resume key mismatch are unrelated." } ] }, @@ -11625,7 +11509,7 @@ "source": "llm", "cache_hit": false, "model": "gpt-5.4-mini?service_tier=flex", - "cluster_id": "cluster-10000-529", + "cluster_id": "cluster-10000-530", "nodes": [ "issue:16998", "issue:22355", @@ -11638,18 +11522,18 @@ "issue:43232", "issue:43295", "issue:43316", - "issue:43352", "issue:43388", "issue:43404", "issue:43441", "issue:43452", "issue:43646", + "issue:43701", "issue:43716", "issue:43761", "issue:43825", + "issue:43931", "issue:44077", "issue:44079", - "issue:44162", "issue:44188", "issue:44190", "issue:44442", @@ -11658,24 +11542,28 @@ "issue:44521", "issue:44534", "issue:44617", - "issue:44792", + "issue:44855", "issue:44871", "issue:44913", + "issue:45137", + "issue:45216", "issue:45584", "issue:45698", "issue:45803", "issue:46050", "issue:46234", "issue:46291", - "issue:46365", "issue:46421", - "issue:46640", + "issue:46506", "issue:46829", "issue:46932", "issue:47217", "issue:47276" ], "soft_pairs": [ + "issue:44855|issue:45137", + "issue:43701|issue:46506", + "issue:43931|issue:45216", "issue:44190|issue:44496", "issue:43761|issue:44079", "issue:27957|issue:32090", @@ -11687,8 +11575,8 @@ "issue:44871|issue:45584", "issue:43295|issue:43825", "issue:43232|issue:43388", - "issue:43761|issue:44188", "issue:16998|issue:22355", + "issue:43761|issue:44188", "issue:44521|issue:47217", "issue:43646|issue:44913", "issue:46234|issue:46291", @@ -11698,202 +11586,193 @@ "issue:44871|issue:46050", "issue:44464|issue:46421", "issue:43404|issue:43441", - "issue:44162|issue:44792", "issue:25251|issue:27957", - "issue:43065|issue:44617", - "issue:43352|issue:46640", - "issue:45803|issue:46365" + "issue:43065|issue:44617" ], - "prepared_review_unit_hash": "157bf4abf0373e50eca483441da9c202dabce770c6a3dd7be2efb4f40d1b5f38", + "prepared_review_unit_hash": "eb16d334873a0477eef07c3c63ac22eb91f2cac13bfd9931ca54964467947e53", "component_budget": { - "node_count": 529, - "item_count": 529, - "soft_pair_count": 1953, - "serialized_chars": 1012074, - "estimated_input_tokens": 253019, - "estimated_eval_tokens": 506294 + "node_count": 530, + "item_count": 530, + "soft_pair_count": 1954, + "serialized_chars": 1013004, + "estimated_input_tokens": 253251, + "estimated_eval_tokens": 506758 }, "budget": { - "node_count": 47, - "item_count": 47, + "node_count": 48, + "item_count": 48, "soft_pair_count": 27, - "serialized_chars": 35089, - "estimated_input_tokens": 8773, - "estimated_eval_tokens": 17802 + "serialized_chars": 35766, + "estimated_input_tokens": 8942, + "estimated_eval_tokens": 18140 }, "overflow_policy": "truncate_then_skip", "trimmed": false, "aggressively_trimmed": false, "split": true, "analyst_result": { - "summary": "All soft edges look like false positives: the paired issues mostly concern different subsystems, different error modes, or different regression classes. I would not merge any of these as duplicates.", - "confidence": 0.93, - "canonical_issue_reason": "No single canonical issue stands out: the set is heterogeneous (model loading, generation, multimodal chat templates, Flash Attention, save/load, config, and training/runtime errors), so none is a safe duplicate hub.", + "summary": "The cluster is highly heterogeneous: the soft links are mostly superficial similarity across unrelated Transformers issues, so none of the candidate pairs should be merged as duplicates.", + "confidence": 0.96, + "canonical_issue_reason": "No single canonical issue stands out: the issues cover unrelated bugs across generation, loading, configuration, multimodal templates, and backend/runtime behavior.", "canonical_pr_reason": null, - "best_issue_reason": "issue:44521 is the strongest standalone issue in the set by signal (more inbound references and a concrete multimodal repro), but it still does not represent the rest of the cluster as duplicates.", + "best_issue_reason": "There is no good global issue canonical because the items do not share one underlying bug or code path.", "best_pr_reason": null, "soft_edge_verdicts": [ + { + "left": "issue:44855", + "right": "issue:45137", + "accept": false, + "reason": "Different failures: Python 3.13 JIT parsing/import error vs DeepSpeed ZeRO3 deque underflow." + }, + { + "left": "issue:43701", + "right": "issue:46506", + "accept": false, + "reason": "Resume checkpoint key mismatch and CPU/CUDA divergence are separate problems with different symptoms and code paths." + }, + { + "left": "issue:43931", + "right": "issue:45216", + "accept": false, + "reason": "Both are loading/checkpoint issues, but one is Qwen3-VL weight-shape mismatch and the other is a Qwen3.5 save_pretrained regression." + }, { "left": "issue:44190", "right": "issue:44496", "accept": false, - "reason": "Different failures: local dataset loading in a no-trainer script vs unrecognized model/config missing model_type." + "reason": "Local dataset loading in an example script is unrelated to unrecognized-model/config loading failure." }, { "left": "issue:43761", "right": "issue:44079", "accept": false, - "reason": "CLIPVisionModel hidden_states regression is not the same as ModelOutput key assignment when a prior value was None." + "reason": "CLIPVisionModel hidden_states regression and ModelOutput key assignment are different subsystems and bugs." }, { "left": "issue:27957", "right": "issue:32090", "accept": false, - "reason": "Flash Attention support for XLM-Roberta and a Trainer _gpu_broadcast_one NoneType error are unrelated." + "reason": "FlashAttention-2 support for XLMRoberta is unrelated to the Trainer NoneType broadcast error." }, { "left": "issue:44442", "right": "issue:45803", "accept": false, - "reason": "Tokenizer loading failure and a bare except in Fuyu tensor conversion are different code paths and symptoms." + "reason": "Tokenizer loading failure and a bare-except bug in FuyuBatchFeature are distinct issues." }, { "left": "issue:42490", "right": "issue:47276", "accept": false, - "reason": "Import warning for a local path with trailing slash is unrelated to the synced_gpus warning in continuous batching." + "reason": "Trailing-slash local model import warning and misleading synced_gpus warning are different user-facing problems." }, { "left": "issue:46829", "right": "issue:46932", "accept": false, - "reason": "VideoPrism output field mix-up and Gemma4 quantized embedder dtype casting are different bugs." + "reason": "VideoPrism hidden_states output bug and Gemma4 quantized-embedder dtype casting are unrelated." }, { "left": "issue:43295", "right": "issue:44871", "accept": false, - "reason": "Processor/tokenizer access regression is separate from Gemma-3 eos_token_id inconsistency." + "reason": "Processor/tokenizer regression for custom model code is not the same as Gemma eos_token_id config inconsistency." }, { "left": "issue:43316", "right": "issue:43716", "accept": false, - "reason": "Gemma3TextConfig API discrepancy is unrelated to Mistral-3 image preprocessor dtype mismatch." + "reason": "Gemma3TextConfig API mismatch and Mistral3 image-preprocessor dtype mismatch are different model-specific bugs." }, { "left": "issue:44871", "right": "issue:45584", "accept": false, - "reason": "Gemma-3 eos_token_id mismatch and Whisper empty-transcription generation failure are different issues." + "reason": "Gemma eos_token configuration mismatch is unrelated to Whisper empty-transcription generation failure." }, { "left": "issue:43295", "right": "issue:43825", "accept": false, - "reason": "Custom processor/tokenizer regression is not the same as a pipeline error-message wording issue." + "reason": "Custom model processor regression and pipeline translation error messaging do not share the same root cause." }, { "left": "issue:43232", "right": "issue:43388", "accept": false, - "reason": "Generation kwargs after sync_gpus and gather_for_metrics truncating tuple labels are different runtime bugs." + "reason": "Generation kwargs sync logic and gather_for_metrics label truncation are separate code paths." }, { - "left": "issue:43761", - "right": "issue:44188", + "left": "issue:16998", + "right": "issue:22355", "accept": false, - "reason": "A CLIPVisionModel output bug is unrelated to attention-kernel divergence under torch.compile." + "reason": "A model_max_length question for DeBERTa-V3 is unrelated to the missing transformers.onnx module." }, { - "left": "issue:16998", - "right": "issue:22355", + "left": "issue:43761", + "right": "issue:44188", "accept": false, - "reason": "Model max length behavior and missing transformers.onnx module are unrelated." + "reason": "Hidden-states regression and attention-kernel divergence under torch.compile are different bugs." }, { "left": "issue:44521", "right": "issue:47217", "accept": false, - "reason": "Both are multimodal, but one is assistant mask generation and the other is image-placeholder handling in user text; not the same bug." + "reason": "Both involve multimodal text/image handling, but one is all-zero assistant masks and the other is image-placeholder placement in user text." }, { "left": "issue:43646", "right": "issue:44913", "accept": false, - "reason": "Transformers 5 custom model init breakage is unrelated to GPTNeoX rotary_pct not persisting on reload." + "reason": "Custom model initialization breakage in v5 is unrelated to GPTNeoXConfig rotary_pct not persisting on reload." }, { "left": "issue:46234", "right": "issue:46291", "accept": false, - "reason": "Missing import_utils symbol and pipelines import failure due kernels version/revision requirements are different import issues." + "reason": "Missing import_utils helper and kernels dependency/version import failure are different issues." }, { "left": "issue:43452", "right": "issue:44496", "accept": false, - "reason": "gguf_file/from_pretrained breakage is not the same as an unrecognized model lacking model_type." + "reason": "gguf_file loading breakage and unrecognized-model/model_type config errors both affect loading, but they are different failure modes." }, { "left": "issue:44077", "right": "issue:44534", "accept": false, - "reason": "Optional post_init policy for patchtsmixer does not match junk non-persistent buffer initialization." + "reason": "patchtsmixer post_init policy and non-persistent buffer initialization are unrelated." }, { "left": "issue:42907", "right": "issue:45698", "accept": false, - "reason": "Saving dequantized models failing and loading the wrong custom module after save_pretrained are distinct save/load bugs." + "reason": "Dequantized-model save_pretrained failure and wrong custom-module loading after save_pretrained are distinct serialization problems." }, { "left": "issue:44871", "right": "issue:46050", "accept": false, - "reason": "Gemma-3 eos_token_id mismatch is separate from sinq quantized model save/load breakage." + "reason": "Gemma eos_token_id inconsistency is unrelated to sinq quantization save/load breakage." }, { "left": "issue:44464", "right": "issue:46421", "accept": false, - "reason": "Chunked generation inconsistency with compiled forward is not the same as cache lazy initialization causing recompilation." + "reason": "Chunked generation output inconsistency and chunked-prefill recompilation are separate compile/cache issues." }, { "left": "issue:43404", "right": "issue:43441", "accept": false, - "reason": "Weight tying in Mistral3ForConditionalGeneration and FlashAttention support in Ministral-3 are different problems." - }, - { - "left": "issue:44162", - "right": "issue:44792", - "accept": false, - "reason": "ESM2 breakage and a Janus image-generation test failure do not share the same underlying code path." - }, - { - "left": "issue:25251", - "right": "issue:27957", - "accept": false, - "reason": "Pipeline top_k output nesting and XLMRoberta Flash Attention 2 support are unrelated." + "reason": "lm_head weight tying in Mistral3 and FlashAttention failures in Ministral-3 are not the same bug." }, { "left": "issue:43065", "right": "issue:44617", "accept": false, - "reason": "A dummy Conv2d in Sam3PixelDecoder is not the same as Sam3Video running out of memory." - }, - { - "left": "issue:43352", - "right": "issue:46640", - "accept": false, - "reason": "NemotronHForCausalLM lacking Flash Attention 2 support is too specific to merge with a generic Flash Attention 2 issue." - }, - { - "left": "issue:45803", - "right": "issue:46365", - "accept": false, - "reason": "Bare except swallowing errors in FuyuBatchFeature is unrelated to Trainer import failure caused by datasets." + "reason": "Dummy Conv2d in Sam3PixelDecoder and Sam3Video CUDA OOM are unrelated." } ] }, @@ -11915,13 +11794,13 @@ "source": "llm", "cache_hit": false, "model": "gpt-5.4-mini?service_tier=flex", - "cluster_id": "cluster-10000-529", + "cluster_id": "cluster-10000-530", "nodes": [ "issue:29870", "issue:35141", "issue:41950", "issue:42490", - "issue:43388", + "issue:43352", "issue:43475", "issue:43504", "issue:43525", @@ -11930,10 +11809,9 @@ "issue:43572", "issue:43650", "issue:43673", - "issue:43688", "issue:43931", "issue:44060", - "issue:44190", + "issue:44162", "issue:44242", "issue:44479", "issue:44514", @@ -11943,7 +11821,6 @@ "issue:44792", "issue:44855", "issue:44913", - "issue:44964", "issue:44977", "issue:45127", "issue:45200", @@ -11959,13 +11836,18 @@ "issue:46178", "issue:46202", "issue:46234", + "issue:46365", "issue:46369", "issue:46424", + "issue:46640", "issue:46693", "issue:46894", "issue:47217" ], "soft_pairs": [ + "issue:43352|issue:46640", + "issue:44162|issue:44792", + "issue:45803|issue:46365", "issue:43540|issue:44514", "issue:43572|issue:46178", "issue:44610|issue:46369", @@ -11980,203 +11862,213 @@ "issue:45632|issue:46424", "issue:35141|issue:42490", "issue:45803|issue:46170", - "issue:44242|issue:46894", "issue:46693|issue:47217", "issue:45127|issue:45200", - "issue:43673|issue:44743", - "issue:43931|issue:44060", + "issue:44242|issue:46894", "issue:41950|issue:43475", + "issue:43931|issue:44060", + "issue:43673|issue:44743", "issue:43650|issue:46202", - "issue:44792|issue:45405", "issue:44855|issue:45588", - "issue:29870|issue:35141", - "issue:43388|issue:43688", - "issue:44190|issue:44964" + "issue:44792|issue:45405", + "issue:29870|issue:35141" ], - "prepared_review_unit_hash": "9b9dc131115c43cd9e7ad68c307752e05459f371e6d42c45a9c0e09fb0f84b60", + "prepared_review_unit_hash": "bed2ef1c628fea7e51f5380e38a7116f1feec090765f0cb4851b94bca89bbebc", "component_budget": { - "node_count": 529, - "item_count": 529, - "soft_pair_count": 1953, - "serialized_chars": 1012074, - "estimated_input_tokens": 253019, - "estimated_eval_tokens": 506294 + "node_count": 530, + "item_count": 530, + "soft_pair_count": 1954, + "serialized_chars": 1013004, + "estimated_input_tokens": 253251, + "estimated_eval_tokens": 506758 }, "budget": { "node_count": 47, "item_count": 47, - "soft_pair_count": 26, - "serialized_chars": 35081, - "estimated_input_tokens": 8771, - "estimated_eval_tokens": 17798 + "soft_pair_count": 27, + "serialized_chars": 35243, + "estimated_input_tokens": 8811, + "estimated_eval_tokens": 17878 }, "overflow_policy": "truncate_then_skip", "trimmed": false, "aggressively_trimmed": false, "split": true, "analyst_result": { - "summary": "This cluster is heterogeneous: it contains many unrelated model-loading, generation, processor, cache, and training bugs. None of the soft edges look like true duplicates, so no mergeable pair is accepted.", - "confidence": 0.96, - "canonical_issue_reason": "No single canonical issue fits this cluster; the issues span different models and failure modes rather than one underlying bug.", - "canonical_pr_reason": "No PRs are present in the cluster.", - "best_issue_reason": "No issue is a clear global representative; the closest match would still be arbitrary because the cluster is not a single bug family.", - "best_pr_reason": "No PRs are present in the cluster.", + "summary": "Mostly unrelated issues across different models/subsystems. The only soft match that looks like the same underlying bug is the Qwen2.5-VL video-input regression pair. No PRs were provided.", + "confidence": 0.86, + "canonical_issue_reason": "issue:46693 is the strongest standalone bug report: it has broad impact, clear regression framing, and substantial discussion/inbound references around FA2 long-context performance.", + "canonical_pr_reason": null, + "best_issue_reason": "issue:46693 is the most actionable and high-signal issue in the set, with a concrete performance regression and strong community engagement.", + "best_pr_reason": null, "soft_edge_verdicts": [ + { + "left": "issue:43352", + "right": "issue:46640", + "accept": false, + "reason": "Both mention Flash Attention 2, but one is a model-support error and the other is a performance regression; different bugs and code paths." + }, + { + "left": "issue:44162", + "right": "issue:44792", + "accept": false, + "reason": "ESM2 breakage and a Janus image-generation test failure are unrelated models and failures." + }, + { + "left": "issue:45803", + "right": "issue:46365", + "accept": false, + "reason": "A swallowed-exception bug in Fuyu tensor conversion is unrelated to importing Trainer with datasets installed." + }, { "left": "issue:43540", "right": "issue:44514", "accept": false, - "reason": "Both involve Qwen/VL input handling, but one is a video-input ValueError in Qwen3OmniMoe and the other is a batched chat-template crash with padding=False in Qwen2.5-VL; different code paths." + "reason": "Qwen3OmniMoe video-input ValueError and Qwen2_5_VL batched chat-template crash are different failure modes." }, { "left": "issue:43572", "right": "issue:46178", "accept": false, - "reason": "Completely different problems: a missing pad_token_idx in StableLmConfig versus a backend registration error for tensorflow_text." + "reason": "StableLm config padding-token regression and tensorflow_text backend mapping are unrelated." }, { "left": "issue:44610", "right": "issue:46369", "accept": false, - "reason": "Different models and symptoms: OmDet-Turbo preprocessing size mismatch versus SAM3 LiteText memory overflow." + "reason": "OmDet-Turbo input-size mismatch and SAM3-LiteText memory overflow do not share the same bug." }, { "left": "issue:44792", "right": "issue:44977", "accept": false, - "reason": "One is a janus image-generation test failure, the other is a Qwen3.5 flash-attention generation bug; not the same underlying issue." + "reason": "Janus test failure and Qwen3.5 flash-attention generation issue are different model/code-path problems." }, { "left": "issue:45381", "right": "issue:47217", "accept": false, - "reason": "Both are VL/image-placeholder related, but one is wrong vision_position_ids in qwen2.5-vl video input and the other is failure when the image placeholder appears in user text." + "reason": "Both are VL-related, but one is a video-position-id regression and the other is a placeholder-in-user-text parsing failure." }, { "left": "issue:44479", "right": "issue:47217", "accept": false, - "reason": "The former is a broad video-input regression across several VL models; the latter is a specific placeholder-placement bug." + "reason": "Video-input regression and image-placeholder-in-user-text failure are distinct bugs, even though both involve VL models." }, { "left": "issue:44534", "right": "issue:44913", "accept": false, - "reason": "Unrelated bugs: non-persistent buffers being junk-filled in v5 versus GPTNeoX rotary_pct not persisting on reload." + "reason": "Non-persistent buffer junk initialization is unrelated to GPTNeoX rotary_pct reload behavior." }, { "left": "issue:45542", "right": "issue:46234", "accept": false, - "reason": "No common bug: one is a missing TensorFlow backend due to only tensorboard being installed, the other is a missing import_utils helper." + "reason": "TensorBoard/TensorFlow backend detection and missing falcon mamba availability helper are unrelated." }, { "left": "issue:44792", "right": "issue:45600", "accept": false, - "reason": "Janus image-generation test failure and Sam3LiteText auto_mappings CI breakage are different issues." + "reason": "A Janus generation test failure and Sam3LiteText auto-mapping breakage are separate issues." }, { "left": "issue:43504", "right": "issue:45698", "accept": false, - "reason": "Both involve model loading, but one is a BEiT legacy-field preset issue and the other is saving/loading the wrong custom module after save_pretrained." + "reason": "BEiT legacy-field loading failure and wrong custom-module loading after save_pretrained are different serialization/loading bugs." }, { "left": "issue:43525", "right": "issue:43531", "accept": false, - "reason": "A missing pad_token_id in Llama4Config is unrelated to the Qwen3-MoE sliding_window behavior." + "reason": "Missing pad_token_id in Llama4Config and sliding_window problems in Qwen3-MoE are unrelated configuration issues." }, { "left": "issue:45632", "right": "issue:46424", "accept": false, - "reason": "Both mention cache, but one is a remote-code cache path collision and the other is max_cache_length only applying to quantized cache." + "reason": "Remote-code cache-path collisions and max_cache_length being ignored are distinct cache-management bugs." }, { "left": "issue:35141", "right": "issue:42490", "accept": false, - "reason": "Resizing token embeddings reinitialization and trailing-slash local model loading warnings are different bugs." + "reason": "Embedding reinitialization on resize and import warnings from trailing-slash local paths are unrelated." }, { "left": "issue:45803", "right": "issue:46170", "accept": false, - "reason": "Bare except swallowing errors in FuyuBatchFeature is unrelated to the Windows/ROCm torch.distributed import problem." - }, - { - "left": "issue:44242", - "right": "issue:46894", - "accept": false, - "reason": "Both are loss-related, but one is load-balancing loss omission and the other is a training-loss double-shift bug in CohereASR." + "reason": "Fuyu exception handling and Windows ROCm/torch.distributed import errors do not share a code path." }, { "left": "issue:46693", "right": "issue:47217", "accept": false, - "reason": "Performance regression with FA2 long-context generation is not the same as VL image-placeholder failure." + "reason": "FA2 long-context performance regression is unrelated to VL placeholder parsing failures." }, { "left": "issue:45127", "right": "issue:45200", "accept": false, - "reason": "Different bugs: LoRA merge collapse with extended vocabulary versus missing mm_token_type_ids defaults for Gemma 4 fine-tuning." + "reason": "LoRA merge collapse with tied embeddings and Gemma 4 mm_token_type_ids defaults are different model behaviors." }, { - "left": "issue:43673", - "right": "issue:44743", + "left": "issue:44242", + "right": "issue:46894", + "accept": false, + "reason": "Load-balancing loss omission and CohereASR double-shift training loss are both loss-related, but not the same bug." + }, + { + "left": "issue:41950", + "right": "issue:43475", "accept": false, - "reason": "Both touch cache/generation, but one is missing GenerationMixin cache in chunked_prefill and the other is recurrent states resetting in modular_qwen3_5." + "reason": "Video-classification processor mismatch and SAM3 VisionEncoderOutput missing field are unrelated vision bugs." }, { "left": "issue:43931", "right": "issue:44060", "accept": false, - "reason": "Different Qwen variants and different failures: weight-shape mismatch on load versus a tied-weights warning linking the wrong tensors." + "reason": "Qwen3-VL weight-shape mismatch and Qwen3-Next tied-weights warning are different loading/training issues." }, { - "left": "issue:41950", - "right": "issue:43475", + "left": "issue:43673", + "right": "issue:44743", "accept": false, - "reason": "Video-classification pipeline using image processors is unrelated to SAM3 VisionEncoderOutput missing fpn_position_embeddings." + "reason": "Chunked-prefill cache missing and recurrent states resetting with cache are separate cache/state bugs." }, { "left": "issue:43650", "right": "issue:46202", "accept": false, - "reason": "A vague data request and a Gemma pipeline roundtrip failure are not the same bug." - }, - { - "left": "issue:44792", - "right": "issue:45405", - "accept": false, - "reason": "A janus model test failure and an unreleased MIN_PEFT_VERSION bump are unrelated." + "reason": "An empty-data issue and a Gemma pipeline roundtrip failure are unrelated." }, { "left": "issue:44855", "right": "issue:45588", "accept": false, - "reason": "Python 3.13 import parsing failure in DebertaV2Model is unrelated to the flash_attention s_aux=None crash." + "reason": "Python 3.13 import parsing failure and flash_attention.py crashing on s_aux=None are different problems." }, { - "left": "issue:29870", - "right": "issue:35141", + "left": "issue:44792", + "right": "issue:45405", "accept": false, - "reason": "Bart generate() decoder_attention_mask UnboundLocalError and token-embedding resize reinitialization are different failures." + "reason": "A Janus test failure and a minimum PEFT version mismatch are unrelated." }, { - "left": "issue:43388", - "right": "issue:43688", + "left": "issue:29870", + "right": "issue:35141", "accept": false, - "reason": "Metric gathering dropping tuple labels and incorrect auxiliary-loss normalization are different bugs in different parts of the training stack." + "reason": "Bart generate-time decoder_attention_mask error and embedding reinitialization after resize are distinct bugs." }, { - "left": "issue:44190", - "right": "issue:44964", - "accept": false, - "reason": "Local dataset loading in a training script and loading the Phi-4 multimodal model are unrelated problems." + "left": "issue:44479", + "right": "issue:45381", + "accept": true, + "reason": "Both describe the same Qwen2.5-VL video-input regression; 45381 is a more specific manifestation of the broader video-input bug in 44479." } ] }, @@ -12198,7 +12090,7 @@ "source": "llm", "cache_hit": false, "model": "gpt-5.4-mini?service_tier=flex", - "cluster_id": "cluster-10000-529", + "cluster_id": "cluster-10000-530", "nodes": [ "issue:25251", "issue:29870", @@ -12207,23 +12099,24 @@ "issue:43065", "issue:43295", "issue:43335", + "issue:43388", "issue:43550", "issue:43644", + "issue:43688", "issue:43827", "issue:43854", "issue:43994", "issue:44190", - "issue:44242", "issue:44479", "issue:44610", "issue:44617", "issue:44734", - "issue:44779", "issue:44792", "issue:44855", "issue:44869", "issue:44913", "issue:44933", + "issue:44964", "issue:45072", "issue:45245", "issue:45276", @@ -12237,268 +12130,263 @@ "issue:46082", "issue:46369", "issue:46399", - "issue:46489", "issue:46617", "issue:46655", "issue:46697", "issue:46832", "issue:46858", - "issue:46897", "issue:46932", "issue:47013", "issue:47312", "issue:47321" ], "soft_pairs": [ + "issue:44190|issue:44964", + "issue:43388|issue:43688", "issue:46082|issue:46697", - "issue:45072|issue:46932", "issue:43827|issue:43994", "issue:45245|issue:46858", + "issue:45072|issue:46932", "issue:44617|issue:45584", "issue:43644|issue:44913", "issue:45588|issue:47013", "issue:25251|issue:29870", - "issue:44734|issue:45406", "issue:43335|issue:44855", + "issue:44734|issue:45406", "issue:46399|issue:47312", "issue:44869|issue:46832", - "issue:41720|issue:45245", "issue:45381|issue:46369", "issue:45632|issue:46617", + "issue:41720|issue:45245", "issue:43827|issue:47321", "issue:43854|issue:44792", "issue:43550|issue:44190", "issue:39692|issue:47321", "issue:45563|issue:45584", "issue:45381|issue:46655", - "issue:43065|issue:44610", "issue:44617|issue:46369", + "issue:43065|issue:44610", "issue:44479|issue:44734", - "issue:44933|issue:45362", "issue:43335|issue:44190", - "issue:44242|issue:46897", - "issue:44779|issue:46489", + "issue:44933|issue:45362", "issue:43295|issue:45276", "issue:39692|issue:44190" ], - "prepared_review_unit_hash": "a77e355b58e0e0059f7f1c73f54298a034bdb1036413f881823ca661f9987dac", + "prepared_review_unit_hash": "8721befed9a11c093bb0938f498cc398fbe9647a5df10c08813368d5264be99c", "component_budget": { - "node_count": 529, - "item_count": 529, - "soft_pair_count": 1953, - "serialized_chars": 1012074, - "estimated_input_tokens": 253019, - "estimated_eval_tokens": 506294 + "node_count": 530, + "item_count": 530, + "soft_pair_count": 1954, + "serialized_chars": 1013004, + "estimated_input_tokens": 253251, + "estimated_eval_tokens": 506758 }, "budget": { - "node_count": 48, - "item_count": 48, + "node_count": 47, + "item_count": 47, "soft_pair_count": 30, - "serialized_chars": 36864, - "estimated_input_tokens": 9216, - "estimated_eval_tokens": 18688 + "serialized_chars": 36347, + "estimated_input_tokens": 9087, + "estimated_eval_tokens": 18430 }, "overflow_policy": "truncate_then_skip", "trimmed": false, "aggressively_trimmed": false, "split": true, "analyst_result": { - "summary": "Most pairs are unrelated and should stay split. The only clearly supported duplicate-like pair is the DeepSeek tokenizer regression: 44779 is the broader v5 incorrect-output report, and 46489 reads as a narrower follow-up identifying a gap in the earlier fix.", + "summary": "This set does not form a true duplicate cluster: the soft-pair candidates are mostly unrelated bugs/regressions across different models, processors, and subsystems. I would not merge any of the proposed soft edges, and there is no clear canonical issue/PR target here.", "confidence": 0.91, - "canonical_issue_reason": "Issue 44779 is the best canonical issue for the DeepSeek tokenizer regression: it describes the broader v5 incorrect-output bug, while 46489 is a narrower follow-up that points out a missing piece in the earlier fix.", + "canonical_issue_reason": null, "canonical_pr_reason": null, - "best_issue_reason": "44779 is the clearest representative of the only best-supported duplicate-like pair; it captures the main user-facing regression and is the broader of the two reports.", + "best_issue_reason": null, "best_pr_reason": null, "soft_edge_verdicts": [ { - "left": "issue:46082", - "right": "issue:46697", + "left": "issue:44190", + "right": "issue:44964", "accept": false, - "reason": "Different models and failure modes: Llama head_dim validation vs Swin2SR image over-padding." + "reason": "Different problems: local dataset loading script failure vs multimodal model loading regression." }, { - "left": "issue:45072", - "right": "issue:46932", + "left": "issue:43388", + "right": "issue:43688", + "accept": false, + "reason": "Both touch evaluation/labels, but one is last-batch label handling and the other is auxiliary-loss normalization in MoE training." + }, + { + "left": "issue:46082", + "right": "issue:46697", "accept": false, - "reason": "Both mention dtype/quantization, but they affect different Gemma4 code paths and are not the same bug." + "reason": "Unrelated subsystems: LlamaConfig head_dim validation vs image processor over-padding." }, { "left": "issue:43827", "right": "issue:43994", "accept": false, - "reason": "Docs still referencing pipeline() is unrelated to the SigLIP2 model/processor mismatch and quantization failure." + "reason": "Docs stale pipeline references vs SigLIP2 model/pipeline output corruption are different issues." }, { "left": "issue:45245", "right": "issue:46858", "accept": false, - "reason": "Category-limit runtime error vs cache-length tensor arithmetic bug; no shared underlying issue." + "reason": "Different runtime failures: category-count limit vs cache sequence-length tensor arithmetic." + }, + { + "left": "issue:45072", + "right": "issue:46932", + "accept": false, + "reason": "Both are dtype-related, but they affect different models and different code paths." }, { "left": "issue:44617", "right": "issue:45584", "accept": false, - "reason": "Sam3Video OOM and Whisper empty-transcription crash are different models and different failures." + "reason": "Sam3Video OOM and Whisper empty-transcription crash are unrelated bugs." }, { "left": "issue:43644", "right": "issue:44913", "accept": false, - "reason": "Non-persistent buffer junk and GPTNeoX rotary_pct reload regression are unrelated bugs." + "reason": "Non-persistent buffer initialization bug vs GPTNeoX config reload behavior are separate issues." }, { "left": "issue:45588", "right": "issue:47013", "accept": false, - "reason": "Flash-attention s_aux handling and xLSTM return unpacking are distinct code-path failures." + "reason": "Both are crashy edge cases, but they involve different APIs and root causes." }, { "left": "issue:25251", "right": "issue:29870", "accept": false, - "reason": "Both are generation-related, but nested-list output shape change and Bart UnboundLocalError are different bugs." + "reason": "Top-k nesting in pipeline output and Bart decoder_attention_mask UnboundLocalError are unrelated generation bugs." }, { - "left": "issue:44734", - "right": "issue:45406", + "left": "issue:43335", + "right": "issue:44855", "accept": false, - "reason": "KV-cache continuation indexing in serve is unrelated to Gemma4Processor missing _tokenizer." + "reason": "Sparse-layer config bug vs Python 3.13 IndentationError on import are not the same defect." }, { - "left": "issue:43335", - "right": "issue:44855", + "left": "issue:44734", + "right": "issue:45406", "accept": false, - "reason": "Sparse-layer creation bug and Python 3.13 IndentationError are unrelated." + "reason": "KV-cache continuation indexing crash vs missing `_tokenizer` on Gemma4Processor are different failures." }, { "left": "issue:46399", "right": "issue:47312", "accept": false, - "reason": "Kernel compatibility is the only overlap; one is a missing registration decorator bug, the other a broad incompatibility report." + "reason": "Missing decorator during kernel registration vs incompatibility with kernels 0.16 are related only broadly, not the same bug." }, { "left": "issue:44869", "right": "issue:46832", "accept": false, - "reason": "Whisper timestamp decoding crash and RT-DETR feature-level TypeError are unrelated." - }, - { - "left": "issue:41720", - "right": "issue:45245", - "accept": false, - "reason": "Qwen3 cudaErrorAssert and categories-limit runtime error are different failures." + "reason": "Whisper timestamp decode crash and RT-DETR feature-level mismatch are unrelated." }, { "left": "issue:45381", "right": "issue:46369", "accept": false, - "reason": "Qwen video position-id regression and Sam3 LiteText memory overflow are different bugs." + "reason": "Qwen video position IDs bug vs Sam3-LiteText memory overflow are different code paths." }, { "left": "issue:45632", "right": "issue:46617", "accept": false, - "reason": "Both involve trust_remote_code caching, but leaf-directory collision and FileNotFoundError are not the same concrete bug." + "reason": "Both involve trust_remote_code caching, but one is a leaf-dir collision and the other is a FileNotFoundError; too different to merge." + }, + { + "left": "issue:41720", + "right": "issue:45245", + "accept": false, + "reason": "Auto device mapping CUDA assert vs category limit overflow are unrelated." }, { "left": "issue:43827", "right": "issue:47321", "accept": false, - "reason": "Documentation cleanup is unrelated to AMD quark class updates." + "reason": "Docs about removed pipeline APIs vs AMD quark class update issue are unrelated." }, { "left": "issue:43854", "right": "issue:44792", "accept": false, - "reason": "Different model-specific test failures with no evident shared root cause." + "reason": "GLM-4.7-Flash test loading failure vs Janus image-generation test failure are different models/tests." }, { "left": "issue:43550", "right": "issue:44190", "accept": false, - "reason": "torch.compile/SDPA failure in Bamba is unrelated to local dataset loading in the example script." + "reason": "Bamba torch.compile SDPA crash and dataset loading script issue are unrelated." }, { "left": "issue:39692", "right": "issue:47321", "accept": false, - "reason": "SigLIP2 docs/example errors are unrelated to AMD quark class updates." + "reason": "SigLIP2 docs/model example errors vs AMD quark class update issue are unrelated." }, { "left": "issue:45563", "right": "issue:45584", "accept": false, - "reason": "Paged generate warning and Whisper empty transcription are different behaviors in different pipelines." + "reason": "Stale num_return_sequences warning vs Whisper empty-transcription failure are different generation behaviors." }, { "left": "issue:45381", "right": "issue:46655", "accept": false, - "reason": "Both are video-preprocessing issues, but the specific bugs and affected models differ." - }, - { - "left": "issue:43065", - "right": "issue:44610", - "accept": false, - "reason": "Sam3PixelDecoder dummy conv and OmDet processor resize mismatch are unrelated vision-model issues." + "reason": "Both are video/image preprocessing issues, but they affect different models and different bugs." }, { "left": "issue:44617", "right": "issue:46369", "accept": false, - "reason": "Both mention memory, but the model and cause differ; not the same bug." + "reason": "Same Sam3 family, but OOM in Sam3Video is not the same as memory overflow in Sam3-litetext." }, { - "left": "issue:44479", - "right": "issue:44734", + "left": "issue:43065", + "right": "issue:44610", "accept": false, - "reason": "Qwen VL video regression and serve KV-cache continuation indexing are unrelated." + "reason": "Dummy Conv2d in Sam3PixelDecoder vs OmDet-Turbo processor/model size mismatch are unrelated." }, { - "left": "issue:44933", - "right": "issue:45362", + "left": "issue:44479", + "right": "issue:44734", "accept": false, - "reason": "Missing image_utils import and Qwen3.5 chat crash are unrelated." + "reason": "Qwen video-input regression and serve KV-cache indexing crash are different subsystems." }, { "left": "issue:43335", "right": "issue:44190", "accept": false, - "reason": "SwitchTransformers sparse-layer config bug is unrelated to dataset loading in the example script." + "reason": "SwitchTransformers sparse-layer config bug vs local dataset loading bug are unrelated." }, { - "left": "issue:44242", - "right": "issue:46897", + "left": "issue:44933", + "right": "issue:45362", "accept": false, - "reason": "Both involve training losses, but they are different loss bugs in different models and code paths." - }, - { - "left": "issue:44779", - "right": "issue:46489", - "accept": true, - "reason": "These describe the same DeepSeek tokenizer regression in v5; 46489 is a likely follow-up that identifies a gap in the earlier fix." + "reason": "Missing image_utils import and Qwen3.5 chat crash are unrelated." }, { "left": "issue:43295", "right": "issue:45276", "accept": false, - "reason": "Processor/tokenizer regression and resize_token_embeddings propagation are different bugs." + "reason": "Processor/tokenizer API regression vs Gemma4 embedding-resize behavior are different bugs." }, { "left": "issue:39692", "right": "issue:44190", "accept": false, - "reason": "SigLIP2 docs/example errors are unrelated to the local dataset loading issue." + "reason": "SigLIP2 docs/model example errors and dataset script loading failure are unrelated." } ] }, - "evaluator_result": { - "accept": false, - "feedback": "The DeepSeek tokenizer pair is the only one that is clearly supported as a duplicate/follow-up, and the rejection of the other soft edges is suitably conservative. However, the writeup overstates the evidence by calling 44779/46489 the \"only merged issue pair\"\u2014the packet shows closed issues, not a merged pair, so that phrasing is not grounded. Please soften that to a best-supported duplicate-like pair / likely follow-up." - }, - "evaluator_used": true, - "retried": true, - "accepted_nontrivial_soft_edge": true, + "evaluator_result": null, + "evaluator_used": false, + "retried": false, + "accepted_nontrivial_soft_edge": false, "error_kind": null, "error_message": null }, @@ -12513,7 +12401,7 @@ "source": "llm", "cache_hit": false, "model": "gpt-5.4-mini?service_tier=flex", - "cluster_id": "cluster-10000-529", + "cluster_id": "cluster-10000-530", "nodes": [ "issue:27957", "issue:32090", @@ -12531,6 +12419,7 @@ "issue:43854", "issue:43994", "issue:44112", + "issue:44242", "issue:44246", "issue:44279", "issue:44462", @@ -12538,7 +12427,7 @@ "issue:44521", "issue:44734", "issue:44737", - "issue:44938", + "issue:44779", "issue:44964", "issue:45003", "issue:45127", @@ -12547,255 +12436,254 @@ "issue:45290", "issue:45357", "issue:45381", - "issue:45542", "issue:45584", "issue:45600", - "issue:45632", "issue:45803", "issue:45907", - "issue:46234", "issue:46421", + "issue:46489", "issue:46697", "issue:46752", "issue:46858", + "issue:46897", "issue:47145", "issue:47217", "issue:47312", "issue:47321" ], "soft_pairs": [ + "issue:44779|issue:46489", + "issue:44242|issue:46897", "issue:43526|issue:46697", - "issue:43673|issue:46421", - "issue:45127|issue:45357", "issue:43540|issue:44734", + "issue:45127|issue:45357", + "issue:43673|issue:46421", "issue:45276|issue:45584", "issue:47145|issue:47321", "issue:43854|issue:44964", "issue:43643|issue:44496", "issue:43994|issue:47321", "issue:44112|issue:45600", - "issue:44462|issue:45907", "issue:44521|issue:45381", + "issue:44462|issue:45907", "issue:43317|issue:47312", "issue:43611|issue:47321", "issue:27957|issue:35141", "issue:44246|issue:45003", - "issue:43673|issue:46858", - "issue:45290|issue:46752", "issue:44112|issue:44964", + "issue:45290|issue:46752", "issue:42175|issue:44279", - "issue:44964|issue:45600", + "issue:43673|issue:46858", "issue:32090|issue:35141", + "issue:44964|issue:45600", "issue:43404|issue:45216", "issue:42796|issue:44521", "issue:44737|issue:45803", - "issue:44496|issue:47217", "issue:43531|issue:45600", - "issue:44938|issue:45542", - "issue:45632|issue:46234" + "issue:44496|issue:47217" ], - "prepared_review_unit_hash": "37c7bbaf8154ab901f9266b91b4c9db21d6b6a9f902ae2c3e4174091482ad86a", + "prepared_review_unit_hash": "484e33fbd16c78ca6666e63dd9ce1dffa1dde0a7498da794eb277bfcf18f4b4b", "component_budget": { - "node_count": 529, - "item_count": 529, - "soft_pair_count": 1953, - "serialized_chars": 1012074, - "estimated_input_tokens": 253019, - "estimated_eval_tokens": 506294 + "node_count": 530, + "item_count": 530, + "soft_pair_count": 1954, + "serialized_chars": 1013004, + "estimated_input_tokens": 253251, + "estimated_eval_tokens": 506758 }, "budget": { "node_count": 47, "item_count": 47, "soft_pair_count": 29, - "serialized_chars": 35942, - "estimated_input_tokens": 8986, - "estimated_eval_tokens": 18228 + "serialized_chars": 35959, + "estimated_input_tokens": 8990, + "estimated_eval_tokens": 18236 }, "overflow_policy": "truncate_then_skip", "trimmed": false, "aggressively_trimmed": false, "split": true, "analyst_result": { - "summary": "No soft edge looks like a true duplicate. The pairs mostly share only broad subsystem keywords (chat templates, caches, model loading, image processors, CI, etc.) but describe different concrete bugs or regressions.", - "confidence": 0.87, - "canonical_issue_reason": "issue:44521 is the strongest cluster representative: it is open, has the most inbound references, and describes a concrete user-facing multimodal bug. The rest of the set is heterogeneous and does not form a tight duplicate group.", + "summary": "This cluster is mostly heterogeneous issue reports with only superficial overlaps (tokenizers, cache/prefill, multimodal chat templates, model loading). I found no true duplicate pairs among the soft edges.", + "confidence": 0.82, + "canonical_issue_reason": "issue:44521 is the strongest anchor here: it is open, has the most discussion/inbound references, and describes a user-facing multimodal `apply_chat_template` bug that fits the cluster\u2019s most active area.", "canonical_pr_reason": null, - "best_issue_reason": "issue:44521 is the best single issue to surface because it is active, well-referenced, and specific enough to be actionable; the other issues are mostly unrelated one-off bugs.", + "best_issue_reason": "issue:44521 is the best single issue to represent the cluster because it is active, broadly impactful, and has the clearest evidence of ongoing interest/impact compared with the other isolated reports.", "best_pr_reason": null, "soft_edge_verdicts": [ + { + "left": "issue:44779", + "right": "issue:46489", + "accept": false, + "reason": "Both are DeepSeek tokenizer regressions, but they target different model families and likely different fixes; too broad to treat as the same underlying bug." + }, + { + "left": "issue:44242", + "right": "issue:46897", + "accept": false, + "reason": "Both mention loss computation in MoE-like models, but the concrete failures differ (`output_router_logits=False` vs double-shift training loss) and do not look mergeable as one fix." + }, { "left": "issue:43526", "right": "issue:46697", "accept": false, - "reason": "Both are image-processing bugs, but one is about BEiT reduce_labels and the other about Swin2SR padding; different code paths and failures." + "reason": "Different vision/image-processing bugs in different processors; no shared code path or concrete symptom match." }, { - "left": "issue:43673", - "right": "issue:46421", + "left": "issue:43540", + "right": "issue:44734", "accept": false, - "reason": "Both mention chunked prefill/cache, but one is about missing GenerationMixin cache in v5.0.0 and the other about lazy cache init causing recompilation." + "reason": "One is a Qwen3OmniMoe video-input validation error, the other is a serve/KV-cache tensor-indexing crash; unrelated subsystems." }, { "left": "issue:45127", "right": "issue:45357", "accept": false, - "reason": "Both concern save/weight handling regressions, but one is about LoRA + extended vocab tie_word_embeddings and the other about incorrect visual encoder keys in Qwen3.5 save_pretrained." + "reason": "Both concern model save/weight handling, but one is LoRA+extended-vocab collapse and the other is incorrect visual encoder keys in Qwen3.5 checkpoints; not the same bug." }, { - "left": "issue:43540", - "right": "issue:44734", + "left": "issue:43673", + "right": "issue:46421", "accept": false, - "reason": "Different features and failures: Qwen3OmniMoe video input parsing vs transformers serve KV-cache continuation tensor indexing." + "reason": "Both touch chunked prefill/cache, but one is a missing cache in `GenerationMixin` and the other is a lazy-init recompilation issue; distinct failure modes." }, { "left": "issue:45276", "right": "issue:45584", "accept": false, - "reason": "Unrelated bugs: gemma4 embedding resizing vs Whisper empty-transcription generation after align_special_tokens." + "reason": "Unrelated bugs: token embedding resizing vs Whisper empty-transcription generation after special-token alignment." }, { "left": "issue:47145", "right": "issue:47321", "accept": false, - "reason": "Completely different domains: FP8 checkpoint loading/scaling vs AMD quark class updates." + "reason": "Completely different areas: FP8 checkpoint loading versus an AMD quark class update issue." }, { "left": "issue:43854", "right": "issue:44964", "accept": false, - "reason": "Both are model-loading failures, but they involve different models and likely different root causes." + "reason": "Both are model-loading complaints, but for different models and different root causes; not the same underlying defect." }, { "left": "issue:43643", "right": "issue:44496", "accept": false, - "reason": "Different config-loading problems: trust_remote_code missing fields vs missing model_type in config.json." + "reason": "Both involve config/model loading, but one is `trust_remote_code` returning incomplete configs while the other is a missing `model_type` error; distinct code paths." }, { "left": "issue:43994", "right": "issue:47321", "accept": false, - "reason": "No overlap beyond being issues; siglip2 AutoModel/pipeline output bug is unrelated to AMD quark class updates." + "reason": "No meaningful overlap beyond being issues in the repo; unrelated model inference versus AMD class metadata." }, { "left": "issue:44112", "right": "issue:45600", "accept": false, - "reason": "Both are CI failures, but one is a stale GraniteSpeech test and the other is a broken auto_mappings reference." + "reason": "Both cause CI failures, but one is a stale test and the other is a bad auto-mapping reference; different defects." }, { - "left": "issue:44462", - "right": "issue:45907", + "left": "issue:44521", + "right": "issue:45381", "accept": false, - "reason": "Different repository/Hub behaviors: tokenizer.json selection vs list_repo_files revision forwarding." + "reason": "Both are multimodal/chat-template related, but one is zero assistant masks and the other is wrong vision position ids for Qwen2.5-VL video inputs; too different to merge." }, { - "left": "issue:44521", - "right": "issue:45381", + "left": "issue:44462", + "right": "issue:45907", "accept": false, - "reason": "Both are multimodal/template-related, but assistant_masks being zero and vision_position_ids being wrong are distinct bugs." + "reason": "Tokenizer repository loading versus Hub file-listing revision forwarding; unrelated functionality." }, { "left": "issue:43317", "right": "issue:47312", "accept": false, - "reason": "Different problems: dequantized model loading with device_map=auto vs kernel compatibility with kernels 0.16." + "reason": "Different problems: device-map/offload loading versus kernel compatibility." }, { "left": "issue:43611", "right": "issue:47321", "accept": false, - "reason": "The base_model_prefix loading regression is unrelated to the AMD quark class issue." + "reason": "Unrelated issues; one is a Transformers 5.0 loading regression, the other an AMD quark update problem." }, { "left": "issue:27957", "right": "issue:35141", "accept": false, - "reason": "Different model families and different mechanisms: Flash Attention 2 support vs post_init reinitializing output embeddings after resize." + "reason": "Flash Attention support for XLMRoberta is unrelated to the token-embedding resize/post-init reinitialization bug." }, { "left": "issue:44246", "right": "issue:45003", "accept": false, - "reason": "Import slowness and unsafe sys.modules access are related to import-time behavior, but they are not the same concrete bug." + "reason": "Import slowness and unsafe `sys.modules` access are not the same concrete bug, even if both touch import-time behavior." }, { - "left": "issue:43673", - "right": "issue:46858", + "left": "issue:44112", + "right": "issue:44964", "accept": false, - "reason": "Both involve cache/prefill, but one is missing cache initialization and the other is a device-tensor seq_length arithmetic bug." + "reason": "A CI test flake and a model-loading regression are different issues." }, { "left": "issue:45290", "right": "issue:46752", "accept": false, - "reason": "Same API surface, but different edge cases: tool-call assistant messages with no content vs passing an empty list." - }, - { - "left": "issue:44112", - "right": "issue:44964", - "accept": false, - "reason": "A stale CI test in GraniteSpeech is not the same as a model-loading regression for Phi-4 multimodal." + "reason": "Both are `apply_chat_template` crashes, but on different edge cases: tool-call assistant messages versus an empty list." }, { "left": "issue:42175", "right": "issue:44279", "accept": false, - "reason": "Both are dependency-related, but one is missing TensorFlow from the torch extra and the other is a broader dependency issue." + "reason": "Packaging/backend extras issue versus a general dependency issue; no same underlying code-path." }, { - "left": "issue:44964", - "right": "issue:45600", + "left": "issue:43673", + "right": "issue:46858", "accept": false, - "reason": "Model-loading regression for Phi-4 multimodal is unrelated to the broken Sam3LiteText auto-mappings reference." + "reason": "Both involve cache/prefill, but the actual bugs differ: missing GenerationMixin cache versus device-tensor `get_seq_length()` breaking arithmetic on non-CUDA backends." }, { "left": "issue:32090", "right": "issue:35141", "accept": false, - "reason": "Trainer NoneType broadcast error is unrelated to token embedding resize/reinitialization." + "reason": "Trainer broadcast TypeError and embedding reinitialization are unrelated." + }, + { + "left": "issue:44964", + "right": "issue:45600", + "accept": false, + "reason": "Different model-loading failures with different causes; not duplicates." }, { "left": "issue:43404", "right": "issue:45216", "accept": false, - "reason": "Both mention model weight/config handling, but one is a tie-weight bug in Mistral3ForConditionalGeneration and the other is an incorrect Qwen3.5 checkpoint save regression." + "reason": "Both relate to tied weights/checkpoints, but they affect different models and different save/load paths." }, { "left": "issue:42796", "right": "issue:44521", "accept": false, - "reason": "Tokenizer BPE marker handling and multimodal assistant_mask generation are different subsystems and bugs." + "reason": "Tokenizer BPE-marker handling and multimodal assistant-mask generation are unrelated." }, { "left": "issue:44737", "right": "issue:45803", "accept": false, - "reason": "XLNet CPU tensor allocation bug and Fuyu bare-except error swallowing are unrelated implementation issues." - }, - { - "left": "issue:44496", - "right": "issue:47217", - "accept": false, - "reason": "Different model-loading failures: unrecognized model/config.json vs VL image placeholder handling in user text." + "reason": "CPU-side `arange` device bug in XLNet versus a bare-except error handling problem in Fuyu; no shared underlying defect." }, { "left": "issue:43531", "right": "issue:45600", "accept": false, - "reason": "Sliding-window behavior for Qwen3-MoE is unrelated to stale auto_mappings references in CI." - }, - { - "left": "issue:44938", - "right": "issue:45542", - "accept": false, - "reason": "Python 3.14 import failure and missing TensorFlow backend from tensorboard-only install are distinct environment issues." + "reason": "Sliding-window behavior in Qwen3-MoE is unrelated to removed-config references in auto-mappings." }, { - "left": "issue:45632", - "right": "issue:46234", + "left": "issue:44496", + "right": "issue:47217", "accept": false, - "reason": "Trust-remote-code cache path collisions and missing import_utils symbol are different bugs with different fixes." + "reason": "Model config recognition failure versus multimodal placeholder handling; different code paths and symptoms." } ] }, @@ -12817,7 +12705,7 @@ "source": "llm", "cache_hit": false, "model": "gpt-5.4-mini?service_tier=flex", - "cluster_id": "cluster-10000-529", + "cluster_id": "cluster-10000-530", "nodes": [ "issue:36246", "issue:38617", @@ -12830,11 +12718,9 @@ "issue:43335", "issue:43381", "issue:43441", - "issue:43479", "issue:43526", "issue:43650", "issue:43673", - "issue:43824", "issue:43901", "issue:43976", "issue:44016", @@ -12847,19 +12733,21 @@ "issue:44496", "issue:44683", "issue:44792", - "issue:44933", + "issue:44938", "issue:44987", "issue:45042", "issue:45083", "issue:45335", - "issue:45375", "issue:45446", + "issue:45542", "issue:45584", "issue:45588", "issue:45593", + "issue:45632", "issue:45759", "issue:46082", "issue:46121", + "issue:46234", "issue:46424", "issue:46506", "issue:46515", @@ -12869,7 +12757,9 @@ "issue:47333" ], "soft_pairs": [ + "issue:44938|issue:45542", "issue:43335|issue:45593", + "issue:45632|issue:46234", "issue:36246|issue:41628", "issue:36246|issue:38617", "issue:44496|issue:45042", @@ -12877,244 +12767,242 @@ "issue:45083|issue:46506", "issue:43976|issue:44987", "issue:44295|issue:46506", - "issue:44496|issue:45759", "issue:41950|issue:44016", - "issue:44683|issue:45446", + "issue:44496|issue:45759", "issue:39692|issue:46932", + "issue:44683|issue:45446", "issue:43673|issue:46424", "issue:43526|issue:43901", "issue:42994|issue:47333", "issue:43381|issue:45593", "issue:43673|issue:44336", "issue:44016|issue:44220", - "issue:46082|issue:46121", "issue:44279|issue:46515", + "issue:46082|issue:46121", "issue:44016|issue:44062", - "issue:42994|issue:47332", "issue:43650|issue:44220", - "issue:44792|issue:46424", + "issue:42994|issue:47332", "issue:42907|issue:43441", + "issue:44792|issue:46424", "issue:43441|issue:45588", "issue:44327|issue:46752", - "issue:43976|issue:44279", - "issue:43824|issue:44933", "issue:45335|issue:45584", - "issue:43479|issue:45375" + "issue:43976|issue:44279" ], - "prepared_review_unit_hash": "c5718a2c77d6448cb7d11a45ea7b7d005ec756d8a8706bc94b6a6b0fef687be2", + "prepared_review_unit_hash": "a9f834ff13a8cd105610e8ca52d6530fcbad81c1188404ce1ff78468bd5e899f", "component_budget": { - "node_count": 529, - "item_count": 529, - "soft_pair_count": 1953, - "serialized_chars": 1012074, - "estimated_input_tokens": 253019, - "estimated_eval_tokens": 506294 + "node_count": 530, + "item_count": 530, + "soft_pair_count": 1954, + "serialized_chars": 1013004, + "estimated_input_tokens": 253251, + "estimated_eval_tokens": 506758 }, "budget": { "node_count": 48, "item_count": 48, "soft_pair_count": 31, - "serialized_chars": 36965, - "estimated_input_tokens": 9242, - "estimated_eval_tokens": 18740 + "serialized_chars": 36943, + "estimated_input_tokens": 9236, + "estimated_eval_tokens": 18728 }, "overflow_policy": "truncate_then_skip", "trimmed": false, "aggressively_trimmed": false, "split": true, "analyst_result": { - "summary": "These items are thematically similar only at a very broad level (Transformers import/config/pipeline/runtime bugs), but they are not the same underlying defect. The soft edges mostly connect distinct subsystems or different failure modes, so I would reject all of them as duplicates.", - "confidence": 0.96, - "canonical_issue_reason": "No single canonical issue stands out: the cluster is heterogeneous, spanning unrelated import errors, config bugs, cache behavior, docs, and model-specific runtime failures.", + "summary": "The set is heterogeneous: the soft-similarity pairs mostly share broad themes (imports, caching, saving, flash-attn, docs) but not the same concrete bug or code path, so I would not merge any of them as duplicates.", + "confidence": 0.93, + "canonical_issue_reason": null, "canonical_pr_reason": null, - "best_issue_reason": "No good global best issue exists for deduping this set, because none of the issues subsume the others or represent the same concrete bug.", + "best_issue_reason": "issue:45584 is the clearest standalone bug report in the set, with a specific repro and a well-scoped failure path; it is not a duplicate canonical for the whole cluster, but it is the strongest representative issue.", "best_pr_reason": null, "soft_edge_verdicts": [ + { + "left": "issue:44938", + "right": "issue:45542", + "accept": false, + "reason": "Both are environment-related, but one is a Python 3.14 import/load failure and the other is a TensorBoard-without-TensorFlow backend error; different bugs and fixes." + }, { "left": "issue:43335", "right": "issue:45593", "accept": false, - "reason": "Different model components and bugs: sparse-layer config creation vs missing auxiliary losses." + "reason": "One is a SwitchTransformers config edge case, the other is missing auxiliary losses in D-FINE; unrelated model logic." + }, + { + "left": "issue:45632", + "right": "issue:46234", + "accept": false, + "reason": "Cache-path collisions for local trust_remote_code models and a missing import_utils helper are unrelated failures." }, { "left": "issue:36246", "right": "issue:41628", "accept": false, - "reason": "Both are import errors, but for different symbols and code paths." + "reason": "Both are import errors, but they target different symbols and code paths (`Qwen2_5_VLImageProcessor` vs `AutoImageProcessor`)." }, { "left": "issue:36246", "right": "issue:38617", "accept": false, - "reason": "Unrelated missing imports in different modules." + "reason": "Missing image-processor export and missing `layer_type_validation` are unrelated import problems." }, { "left": "issue:44496", "right": "issue:45042", "accept": false, - "reason": "Unrecognized-model config loading vs torchvision requirement in PIL image processors are different failures." + "reason": "Unrecognized model config and PIL image processors wrongly requiring torchvision are different loading/runtime bugs." }, { "left": "issue:43295", "right": "issue:45335", "accept": false, - "reason": "Regression in processor/tokenizer handling vs a t5gemma embedding resize bug." + "reason": "Processor/tokenizer regression and `resize_token_embeddings` not updating `decoder.embed_tokens` are distinct multimodal/model issues." }, { "left": "issue:45083", "right": "issue:46506", "accept": false, - "reason": "Helper output-length logic in qwen3_omni_moe vs CPU/CUDA nondeterministic results are different problems." + "reason": "A helper-length bug in `qwen3_omni_moe` and CPU/CUDA divergence in PPDocLayoutV3 are not the same defect." }, { "left": "issue:43976", "right": "issue:44987", "accept": false, - "reason": "Python-version compatibility issue vs model-loading failure for a specific repo." + "reason": "Python-version incompatibility and a specific model-loading failure are too broad to be the same underlying bug." }, { "left": "issue:44295", "right": "issue:46506", "accept": false, - "reason": "Buffer access error for position_ids is unrelated to CPU/CUDA output mismatch." - }, - { - "left": "issue:44496", - "right": "issue:45759", - "accept": false, - "reason": "Unrecognized model config vs missing text_config unwrapping are different loading bugs." + "reason": "`position_ids` buffer access and CPU/CUDA output mismatches are unrelated code paths." }, { "left": "issue:41950", "right": "issue:44016", "accept": false, - "reason": "Pipeline image-processor lookup bug vs notebook syntax error." + "reason": "Pipeline image-processor lookup and a notebook syntax error are clearly unrelated." }, { - "left": "issue:44683", - "right": "issue:45446", + "left": "issue:44496", + "right": "issue:45759", "accept": false, - "reason": "Both touch flex_attention, but one is a torch>=2.9 failure and the other is a version-check/import-path issue." + "reason": "Both involve model loading, but one is unrecognized model config and the other is `from_config` not unwrapping `text_config`; different fixes." }, { "left": "issue:39692", "right": "issue:46932", "accept": false, - "reason": "Docs example errors vs quantized embedder dtype casting are not the same bug." + "reason": "A broken documentation example and a Gemma4 quantized-embedder dtype bug are not the same issue." + }, + { + "left": "issue:44683", + "right": "issue:45446", + "accept": false, + "reason": "Both mention flex_attention, but one is a torch>=2.9 compiled failure and the other is a version-check/import-guard bug." }, { "left": "issue:43673", "right": "issue:46424", "accept": false, - "reason": "Chunked-prefill cache missing vs max_cache_length only applying to quantized cache are different cache behaviors." + "reason": "Chunked-prefill cache missing and `max_cache_length` only affecting Quantized Cache are different cache behaviors." }, { "left": "issue:43526", "right": "issue:43901", "accept": false, - "reason": "BEIT image-processor label reduction bug vs text-classification docs mismatch." + "reason": "A BEiT fast image-processor label bug and stale docs for `return_all_scores` are unrelated." }, { "left": "issue:42994", "right": "issue:47333", "accept": false, - "reason": "Both are saving issues, but one is quantized-model saving and the other is disk-offloaded model saving." + "reason": "General quantized-model saving and disk-offloaded Qwen3 saving failures are not the same concrete save path." }, { "left": "issue:43381", "right": "issue:45593", "accept": false, - "reason": "Gradient checkpointing eval-mode behavior is unrelated to D-FINE auxiliary-loss handling." + "reason": "Eval-mode gradient checkpointing and missing auxiliary losses are unrelated training issues." }, { "left": "issue:43673", "right": "issue:44336", "accept": false, - "reason": "Cache configuration/runtime issue vs ANSI codes in loading_report are unrelated." + "reason": "Cache handling during chunked prefill and ANSI output formatting in loading_report have nothing in common beyond being bugs." }, { "left": "issue:44016", "right": "issue:44220", "accept": false, - "reason": "Notebook syntax error vs audio feature extraction bug." - }, - { - "left": "issue:46082", - "right": "issue:46121", - "accept": false, - "reason": "Config validation around head_dim vs JSON list handling in rope params are different issues." + "reason": "A notebook syntax error and `_torch_extract_fbank_features()` failure are different parts of the project." }, { "left": "issue:44279", "right": "issue:46515", "accept": false, - "reason": "General dependency issue vs torch 2.4.1 incompatibility." + "reason": "A generic dependency issue and a torch 2.4.1 incompatibility are too broad to merge." }, { - "left": "issue:44016", - "right": "issue:44062", + "left": "issue:46082", + "right": "issue:46121", "accept": false, - "reason": "Notebook syntax error vs AddedToken constructor error." + "reason": "Llama head-dim validation and RoPE parameter conversion with JSON-loaded lists are separate configuration bugs." }, { - "left": "issue:42994", - "right": "issue:47332", + "left": "issue:44016", + "right": "issue:44062", "accept": false, - "reason": "Both mention saving/offloading, but they are different model families and failure modes." + "reason": "Notebook syntax and `AddedToken(..., special=...)` argument duplication are unrelated." }, { "left": "issue:43650", "right": "issue:44220", "accept": false, - "reason": "Placeholder issue title vs an audio feature extraction crash." + "reason": "`ADD THE DATA` is not a meaningful duplicate target for the audio feature extraction bug." }, { - "left": "issue:44792", - "right": "issue:46424", + "left": "issue:42994", + "right": "issue:47332", "accept": false, - "reason": "Janus image-generation test failure is unrelated to cache-length behavior." + "reason": "Both concern saving, but one is quantized-model saving and the other is disk-offloaded Qwen3-VL saving; different mechanisms." }, { "left": "issue:42907", "right": "issue:43441", "accept": false, - "reason": "Different model families and different failures: dequantized saving vs FlashAttention runtime." + "reason": "Dequantized-model saving and FlashAttention failures in Ministral-3 are different regressions." + }, + { + "left": "issue:44792", + "right": "issue:46424", + "accept": false, + "reason": "Janus image-generation test failure and cache max-length behavior are unrelated." }, { "left": "issue:43441", "right": "issue:45588", "accept": false, - "reason": "Both involve flash attention, but one is a Ministral-3-specific failure and the other is an s_aux=None crash in a different path." + "reason": "Both mention flash attention, but they describe different failure modes and code paths, so they are not duplicate fixes." }, { "left": "issue:44327", "right": "issue:46752", "accept": false, - "reason": "QA decode_spans bounds error vs empty-list chat template crash are different edge cases." - }, - { - "left": "issue:43976", - "right": "issue:44279", - "accept": false, - "reason": "Python version incompatibility is not the same as a generic dependency issue." - }, - { - "left": "issue:43824", - "right": "issue:44933", - "accept": false, - "reason": "Different missing imports from different modules/symbols." + "reason": "QA decode-spans top-k bounds error and empty chat-template IndexError are different edge cases." }, { "left": "issue:45335", "right": "issue:45584", "accept": false, - "reason": "Token-embedding resize bug vs Whisper empty-transcription generation failure are unrelated." + "reason": "Embedding resize propagation and empty-transcription Whisper generation failure are unrelated." }, { - "left": "issue:43479", - "right": "issue:45375", + "left": "issue:43976", + "right": "issue:44279", "accept": false, - "reason": "Different multimodal config initialization bug vs missing config field silently dropped." + "reason": "A Transformers/Python version compatibility issue and a dependency issue are too broad to be the same bug." } ] }, @@ -13136,7 +13024,7 @@ "source": "llm", "cache_hit": false, "model": "gpt-5.4-mini?service_tier=flex", - "cluster_id": "cluster-10000-529", + "cluster_id": "cluster-10000-530", "nodes": [ "issue:39692", "issue:41950", @@ -13144,10 +13032,12 @@ "issue:43295", "issue:43299", "issue:43381", + "issue:43479", "issue:43577", "issue:43644", "issue:43650", "issue:43653", + "issue:43824", "issue:43854", "issue:43901", "issue:43931", @@ -13161,16 +13051,16 @@ "issue:44462", "issue:44466", "issue:44496", - "issue:44683", "issue:44779", + "issue:44933", "issue:45020", "issue:45081", "issue:45216", "issue:45276", "issue:45278", "issue:45357", + "issue:45375", "issue:45405", - "issue:45542", "issue:45593", "issue:45812", "issue:45823", @@ -13179,7 +13069,6 @@ "issue:46209", "issue:46369", "issue:46399", - "issue:46512", "issue:46614", "issue:47000", "issue:47085", @@ -13187,238 +13076,250 @@ "issue:47217" ], "soft_pairs": [ + "issue:43824|issue:44933", "issue:43295|issue:44779", "issue:43901|issue:44496", + "issue:41950|issue:45020", "issue:43299|issue:45216", "issue:44279|issue:45405", - "issue:41950|issue:45020", "issue:43901|issue:47217", "issue:39692|issue:43901", "issue:42796|issue:43650", "issue:44779|issue:45812", "issue:43577|issue:45216", "issue:46614|issue:47085", - "issue:46123|issue:46369", + "issue:43479|issue:45375", "issue:41950|issue:45823", + "issue:46123|issue:46369", "issue:44466|issue:45276", "issue:43976|issue:45405", - "issue:44075|issue:44368", "issue:43653|issue:45593", + "issue:44075|issue:44368", "issue:43381|issue:47000", "issue:44462|issue:45081", "issue:43299|issue:45357", "issue:43644|issue:47145", "issue:43299|issue:43854", - "issue:46082|issue:46399", - "issue:46082|issue:46512", "issue:43650|issue:46209", - "issue:44351|issue:45278", "issue:43931|issue:47145", + "issue:44351|issue:45278", "issue:44016|issue:44291", - "issue:44683|issue:45542" + "issue:46082|issue:46399" ], - "prepared_review_unit_hash": "a68afd839dfbf15cd9f3c773886e7d91774fadab57e34c68ba7ca2169a67bfbd", + "prepared_review_unit_hash": "1f2d1a88e50184dca60eef4f777a59d26da7aac509fce5fd35ce81c6a292aa97", "component_budget": { - "node_count": 529, - "item_count": 529, - "soft_pair_count": 1953, - "serialized_chars": 1012074, - "estimated_input_tokens": 253019, - "estimated_eval_tokens": 506294 + "node_count": 530, + "item_count": 530, + "soft_pair_count": 1954, + "serialized_chars": 1013004, + "estimated_input_tokens": 253251, + "estimated_eval_tokens": 506758 }, "budget": { - "node_count": 47, - "item_count": 47, + "node_count": 48, + "item_count": 48, "soft_pair_count": 29, - "serialized_chars": 35991, - "estimated_input_tokens": 8998, - "estimated_eval_tokens": 18252 + "serialized_chars": 36572, + "estimated_input_tokens": 9143, + "estimated_eval_tokens": 18542 }, "overflow_policy": "truncate_then_skip", "trimmed": false, "aggressively_trimmed": false, "split": true, "analyst_result": { - "summary": "The cluster is highly heterogeneous: most items are distinct bugs, regressions, or docs issues across unrelated subsystems. I found no soft duplicate pairs worth merging.", - "confidence": 0.95, - "canonical_issue_reason": "issue:44462 is the best broad anchor because it describes a general AutoTokenizer loading regression around repository tokenizer.json, which is a core behavior and potentially related to multiple tokenizer-loading complaints.", + "summary": "The cluster is mostly a set of unrelated Transformers issues linked only by broad similarity (tokenizers, import errors, model loading, docs, regressions). I found no convincing duplicate pairs; all soft-edge candidates should stay separate.", + "confidence": 0.88, + "canonical_issue_reason": "If one anchor is needed, issue:45278 is the broadest regression umbrella in the set, but this cluster is heterogeneous and it should not be treated as a true duplicate canonical for the others.", "canonical_pr_reason": null, - "best_issue_reason": "issue:45812 is the strongest standalone issue: it is concrete, high-severity, and describes a silent v4\u2192v5 regression affecting token IDs for all Granite models.", + "best_issue_reason": "Issue:45278 is the most globally representative of the cluster\u2019s broad version-regression theme, but only as a loose anchor, not as a duplicate target.", "best_pr_reason": null, "soft_edge_verdicts": [ + { + "left": "issue:43824", + "right": "issue:44933", + "accept": false, + "reason": "Both are import-related, but one is a missing model export and the other is an image_utils import path problem; different code paths and fixes." + }, { "left": "issue:43295", "right": "issue:44779", "accept": false, - "reason": "Different tokenizer regressions: one is a processor/tokenizer mismatch and quantization/doc example failure, the other is Deepseek tokenization producing wrong results." + "reason": "One is a processor/tokenizer API regression for multimodal inputs; the other is Deepseek token ID corruption. Same broad area, not the same bug." }, { "left": "issue:43901", "right": "issue:44496", "accept": false, - "reason": "Docs mismatch for TextClassificationPipeline vs an unrecognized-model/config loading error; unrelated failure modes." + "reason": "Documentation mismatch for a pipeline option versus an unrecognized-model/config loading failure; unrelated failures." + }, + { + "left": "issue:41950", + "right": "issue:45020", + "accept": false, + "reason": "Video pipeline processor selection bug versus remote_code compatibility regressions; different concrete problems." }, { "left": "issue:43299", "right": "issue:45216", "accept": false, - "reason": "Both involve Qwen3 family, but one is a VL model loading breakage and the other is a save_pretrained checkpoint regression; not the same bug." + "reason": "Both involve Qwen-family models, but one is a load-time MoE model issue and the other is incorrect save_pretrained serialization." }, { "left": "issue:44279", "right": "issue:45405", "accept": false, - "reason": "Dependency/versioning complaints, but about different packages and different root causes." - }, - { - "left": "issue:41950", - "right": "issue:45020", - "accept": false, - "reason": "Video-classification pipeline processor bug vs generic remote_code breakage; too broad and unrelated." + "reason": "Generic dependency issue versus a specific MIN_PEFT_VERSION bump to an unreleased package; not the same underlying bug." }, { "left": "issue:43901", "right": "issue:47217", "accept": false, - "reason": "Pipeline docs behavior mismatch vs VL placeholder handling failure; different code paths and symptoms." + "reason": "Docs mentioning return_all_scores versus a VL runtime failure with image placeholders; no shared fix." }, { "left": "issue:39692", "right": "issue:43901", "accept": false, - "reason": "SigLIP2 example errors are unrelated to TextClassificationPipeline docs behavior." + "reason": "Both are documentation issues, but they affect different examples and behaviors." }, { "left": "issue:42796", "right": "issue:43650", "accept": false, - "reason": "Tokenizer BPE-marker bug vs unrelated filler issue; no common underlying defect." + "reason": "Tokenizer BPE marker handling is unrelated to the 'ADD THE DATA' issue." }, { "left": "issue:44779", "right": "issue:45812", "accept": false, - "reason": "Both are tokenizer regressions, but they affect different model families and likely different tokenizer code paths." + "reason": "Two tokenizer regressions, but for different model families and different symptoms (wrong results vs wrong IDs)." }, { "left": "issue:43577", "right": "issue:45216", "accept": false, - "reason": "Blip2 dtype casting bug vs Qwen3.5 save_pretrained checkpoint keys; unrelated." + "reason": "Blip2 dtype propagation bug versus Qwen3.5 checkpoint serialization bug; different subsystems." }, { "left": "issue:46614", "right": "issue:47085", "accept": false, - "reason": "Stale phonemizer backend state is a tokenizer-specific bug, not the same as a broader thread-safety regression." + "reason": "Stale phonemizer language state is unrelated to fast tokenizer thread-safety." }, { - "left": "issue:46123", - "right": "issue:46369", + "left": "issue:43479", + "right": "issue:45375", "accept": false, - "reason": "MaskGenerationPipeline batch handling vs sam3-litetext memory overflow; different components and failures." + "reason": "Both are config issues, but they concern different multimodal models and different fields/defaults." }, { "left": "issue:41950", "right": "issue:45823", "accept": false, - "reason": "Pipeline processor selection bug vs Gemma4 device placement issue; not the same concrete problem." + "reason": "Pipeline processor lookup bug versus Gemma4 device placement mismatch in forward." + }, + { + "left": "issue:46123", + "right": "issue:46369", + "accept": false, + "reason": "Final-batch flag handling in a pipeline versus memory overflow in a specific SAM model path." }, { "left": "issue:44466", "right": "issue:45276", "accept": false, - "reason": "Serialization of tied lm_head weights vs Gemma4 resize_token_embeddings propagation; related theme but different bug." + "reason": "Serialization of tied weights depending on device is not the same bug as resize_token_embeddings not updating embeddings." }, { "left": "issue:43976", "right": "issue:45405", "accept": false, - "reason": "Python support/runtime compatibility issue vs unreleased PEFT dependency bump; unrelated." + "reason": "Python compatibility breakage versus an unreleased dependency version bump; different causes and remedies." }, { - "left": "issue:44075", - "right": "issue:44368", + "left": "issue:43653", + "right": "issue:45593", "accept": false, - "reason": "SGD argument handling bug vs a warning about tie_word_embeddings; different behavior and scope." + "reason": "Tokenizer special-token registration versus D-FINE auxiliary-loss logic; unrelated." }, { - "left": "issue:43653", - "right": "issue:45593", + "left": "issue:44075", + "right": "issue:44368", "accept": false, - "reason": "BigBirdTokenizer special-token registration bug vs D-FINE auxiliary-loss behavior; unrelated." + "reason": "Optimizer arguments being ignored has no overlap with the Qwen3.5 tie_word_embeddings warning." }, { "left": "issue:43381", "right": "issue:47000", "accept": false, - "reason": "Gradient-checkpointing eval-mode restriction vs RTDetrLoss infeasible cost matrix; no overlap." + "reason": "Gradient checkpointing/eval-mode incompatibility versus RTDetrLoss infeasible cost matrix; separate failures." }, { "left": "issue:44462", "right": "issue:45081", "accept": false, - "reason": "AutoTokenizer ignoring tokenizer.json is not the same as a Mistral regex patch crash during loading." + "reason": "AutoTokenizer repo-file selection bug versus a Mistral regex patch crash; both tokenizer-related but different concrete code paths." }, { "left": "issue:43299", "right": "issue:45357", "accept": false, - "reason": "Different Qwen3.5 issues: one loading/MoE-related, the other incorrect visual encoder key serialization." + "reason": "Both are Qwen3.5/Qwen3VL regressions, but one is model loading and the other is save_pretrained key serialization." }, { "left": "issue:43644", "right": "issue:47145", "accept": false, - "reason": "Non-persistent buffer junk initialization vs missing FP8 weight_scale application; distinct correctness bugs." + "reason": "Non-persistent buffer junk initialization versus FP8 weight_scale not being applied; different state-loading bugs." }, { "left": "issue:43299", "right": "issue:43854", "accept": false, - "reason": "Qwen3VL loading regression and GLM-4.7 Flash unit-test loading issue are unrelated." + "reason": "Different models and different load failures; no shared underlying bug." }, { - "left": "issue:46082", - "right": "issue:46399", + "left": "issue:43650", + "right": "issue:46209", "accept": false, - "reason": "LlamaConfig head_dim validation vs Qwen3.5 kernel registration decorator error; different code paths." + "reason": "The placeholder issue has no relation to FP8 inference on B200." }, { - "left": "issue:46082", - "right": "issue:46512", + "left": "issue:43931", + "right": "issue:47145", "accept": false, - "reason": "Config validation error vs stale config.dtype after from_config; unrelated dtype/config issues." + "reason": "Qwen3-VL weight-shape mismatch versus FP8 compressed-tensors scaling; different error classes." }, { - "left": "issue:43650", - "right": "issue:46209", + "left": "issue:44351", + "right": "issue:45278", "accept": false, - "reason": "Placeholder/empty issue vs FP8 inference failure on B200; no meaningful match." + "reason": "A single missing HybridCache export is not enough to merge with a broad 'many import errors' report." }, { - "left": "issue:44351", - "right": "issue:45278", + "left": "issue:44016", + "right": "issue:44291", "accept": false, - "reason": "Both mention imports, but one is a missing symbol and the other is broad import failures after upgrade; not the same defect." + "reason": "Notebook syntax error versus init_empty_weights argument mismatch; unrelated." }, { - "left": "issue:43931", - "right": "issue:47145", + "left": "issue:46082", + "right": "issue:46399", "accept": false, - "reason": "Qwen3-VL weight-shape mismatch at load time vs FP8 checkpoint scaling corruption; different failure modes." + "reason": "LlamaConfig head_dim validation is unrelated to Qwen3.5 rotary-kernel registration failures." }, { - "left": "issue:44016", - "right": "issue:44291", + "left": "issue:43381", + "right": "issue:47000", "accept": false, - "reason": "Notebook syntax error vs init_empty_weights TypeError; unrelated." + "reason": "Re-listed? No; this pair remains unrelated because the failures affect different training components." }, { - "left": "issue:44683", - "right": "issue:45542", + "left": "issue:44462", + "right": "issue:45081", "accept": false, - "reason": "Torch flex_attention compilation bug vs TensorFlow backend dependency issue; unrelated." + "reason": "Re-listed? No; same verdict: repo tokenizer selection and Mistral regex patching are distinct bugs." } ] }, @@ -13440,7 +13341,7 @@ "source": "llm", "cache_hit": false, "model": "gpt-5.4-mini?service_tier=flex", - "cluster_id": "cluster-10000-529", + "cluster_id": "cluster-10000-530", "nodes": [ "issue:33290", "issue:42175", @@ -13465,16 +13366,15 @@ "issue:44261", "issue:44263", "issue:44336", - "issue:44589", "issue:44623", "issue:44671", + "issue:44683", "issue:44749", "issue:44779", "issue:45081", "issue:45103", "issue:45200", "issue:45325", - "issue:45362", "issue:45372", "issue:45397", "issue:45542", @@ -13485,12 +13385,16 @@ "issue:45834", "issue:45859", "issue:45907", + "issue:46082", + "issue:46512", "issue:46851", "issue:47000", "issue:47145", "issue:47217" ], "soft_pairs": [ + "issue:44683|issue:45542", + "issue:46082|issue:46512", "issue:44242|issue:45200", "issue:45081|issue:45907", "issue:45588|issue:47000", @@ -13501,9 +13405,8 @@ "issue:43441|issue:45834", "issue:44336|issue:46851", "issue:44671|issue:44779", - "issue:43479|issue:43525", - "issue:44246|issue:47217", "issue:33290|issue:45656", + "issue:44246|issue:47217", "issue:45542|issue:45800", "issue:42175|issue:43976", "issue:44623|issue:44749", @@ -13511,211 +13414,217 @@ "issue:44186|issue:47000", "issue:44623|issue:45397", "issue:43525|issue:43881", - "issue:43784|issue:45542", + "issue:43479|issue:43525", "issue:43901|issue:44246", + "issue:43784|issue:45542", "issue:43673|issue:43761", "issue:43525|issue:43643", "issue:44261|issue:44263", "issue:43575|issue:47145", - "issue:44589|issue:45362", "issue:45325|issue:46851" ], - "prepared_review_unit_hash": "37cdef52931483d9578201346dff119f48243c5e78ace0b4fe54e45f53690f6b", + "prepared_review_unit_hash": "90e54d5a6889df894af45e99ef2591b173b768b4dac9198d60cd2a8a02a78469", "component_budget": { - "node_count": 529, - "item_count": 529, - "soft_pair_count": 1953, - "serialized_chars": 1012074, - "estimated_input_tokens": 253019, - "estimated_eval_tokens": 506294 + "node_count": 530, + "item_count": 530, + "soft_pair_count": 1954, + "serialized_chars": 1013004, + "estimated_input_tokens": 253251, + "estimated_eval_tokens": 506758 }, "budget": { - "node_count": 47, - "item_count": 47, - "soft_pair_count": 28, - "serialized_chars": 35440, - "estimated_input_tokens": 8860, - "estimated_eval_tokens": 17976 + "node_count": 48, + "item_count": 48, + "soft_pair_count": 29, + "serialized_chars": 36378, + "estimated_input_tokens": 9095, + "estimated_eval_tokens": 18446 }, "overflow_policy": "truncate_then_skip", "trimmed": false, "aggressively_trimmed": false, "split": true, "analyst_result": { - "summary": "This set is a heterogeneous mix of Transformer bugs. None of the soft-similarity pairs look like the same underlying issue or change; most are only broadly related by model family or subsystem.", - "confidence": 0.93, - "canonical_issue_reason": "issue:43873 is the most central representative here: it is open, has the highest discussion/inbound activity, and captures a broad quantization/offloading failure mode that overlaps the cluster\u2019s most recurring theme.", + "summary": "This set is not a true duplicate cluster: the issues span unrelated bugs across configs, quantization/offload, tokenizers, docs, and model-specific regressions. I found no soft-edge pair that looks like the same underlying bug/change, so all candidate links should be rejected.", + "confidence": 0.46, + "canonical_issue_reason": "issue:43873 is the broadest and most active item here, centered on quantization/offload behavior, which is the closest thing to a shared theme in this otherwise mixed set.", "canonical_pr_reason": null, - "best_issue_reason": "issue:43873 is the best umbrella issue for this cluster because it is the most active, still open, and broadly framed enough to represent the quantization/offloading segment better than the more narrowly scoped bugs.", + "best_issue_reason": "issue:43873 is the best anchor among weak candidates because it is open, discussion-heavy, and user-facing; however, it is still not a good duplicate representative for most of the other issues.", "best_pr_reason": null, "soft_edge_verdicts": [ + { + "left": "issue:44683", + "right": "issue:45542", + "accept": false, + "reason": "Different bugs: compiled flex_attention on newer torch vs missing TensorFlow backend wiring from tensorboard install." + }, + { + "left": "issue:46082", + "right": "issue:46512", + "accept": false, + "reason": "Unrelated config issues: LlamaConfig head_dim validation vs from_config(dtype=...) leaving config.dtype stale." + }, { "left": "issue:44242", "right": "issue:45200", "accept": false, - "reason": "Different failures: Mixture-of-Experts load-balancing loss vs Gemma 4 text-only token type defaults." + "reason": "Different model/code paths: MoE load-balancing loss gating vs Gemma 4 multimodal token type defaults." }, { "left": "issue:45081", "right": "issue:45907", "accept": false, - "reason": "Unrelated code paths: Mistral tokenizer regex patch crash vs hub file listing revision forwarding." + "reason": "Mistral tokenizer regex patch crash is unrelated to list_repo_files not forwarding revision." }, { "left": "issue:45588", "right": "issue:47000", "accept": false, - "reason": "Different subsystems and symptoms: flash attention auxiliary state crash vs RTDetr matching infeasibility." + "reason": "Different failures: flash_attention s_aux=None crash vs RTDetrLoss infeasible cost matrix." }, { "left": "issue:43873", "right": "issue:47145", "accept": false, - "reason": "Both mention quantization/offloading, but one is runtime offload behavior and the other is frozen FP8 checkpoint scaling; not the same bug." + "reason": "Both touch quantization/offload, but one is about offloading not working while the other is silent corruption from missing weight_scale; not the same bug." }, { "left": "issue:45103", "right": "issue:45372", "accept": false, - "reason": "Auto-docstring annotation parsing crash is unrelated to Gemma 4 processor loading/import dependency failure." + "reason": "Auto-docstring future-annotations crash is unrelated to the mistral_common ReasoningEffort import error." }, { "left": "issue:42994", "right": "issue:43317", "accept": false, - "reason": "Quantized save failure and dequantized load/offload failure are different code paths and concrete problems." + "reason": "Related to quantized models, but saving failure and dequantized loading/offload failure are distinct code paths and symptoms." }, { "left": "issue:45859", "right": "issue:46851", "accept": false, - "reason": "Same model family, but CPU offload incompatibility is distinct from variable-length position_ids handling." + "reason": "Same family of position-related Qwen bugs, but one is CPU offload incompatibility and the other is variable-length position_ids behavior." }, { "left": "issue:43441", "right": "issue:45834", "accept": false, - "reason": "FlashAttention/Ministral regression is unrelated to Kosmos2.5 OCR long-input indexing." + "reason": "Different models and failures: Ministral FlashAttention regression vs Kosmos2.5 OCR length index error." }, { "left": "issue:44336", "right": "issue:46851", "accept": false, - "reason": "ANSI output formatting in loading_report is unrelated to Qwen3.5 variable-length position_ids." + "reason": "ANSI-code logging issue is unrelated to Qwen position_ids handling." }, { "left": "issue:44671", "right": "issue:44779", "accept": false, - "reason": "Different models and bugs: CamemBERT masked-LM regression vs Deepseek tokenizer correctness regression." + "reason": "CamemBERT masked-LM regression and Deepseek tokenizer corruption are different subsystems and bugs." }, { - "left": "issue:43479", - "right": "issue:43525", + "left": "issue:33290", + "right": "issue:45656", "accept": false, - "reason": "Both are config-related, but one is default nested vision/audio config initialization and the other is a missing pad_token_id attribute." + "reason": "Both mention deepspeed/optimizer, but OOM in Adafactor and double optimizer.step() are different problems." }, { "left": "issue:44246", "right": "issue:47217", "accept": false, - "reason": "Import latency issue is unrelated to VL placeholder handling during generation." - }, - { - "left": "issue:33290", - "right": "issue:45656", - "accept": false, - "reason": "Deepspeed OOM with Adafactor is not the same as optimizer.step being invoked twice." + "reason": "Import performance issue is unrelated to VL image-placeholder handling." }, { "left": "issue:45542", "right": "issue:45800", "accept": false, - "reason": "Packaging/backend installation issue is unrelated to torch/transformers version incompatibility." + "reason": "TensorBoard/TensorFlow backend packaging bug is unrelated to torch/transformers version incompatibility." }, { "left": "issue:42175", "right": "issue:43976", "accept": false, - "reason": "Missing TensorFlow backend in extras install is unrelated to Python version support regression." + "reason": "Package extra backend omission and Python version incompatibility are not the same issue." }, { "left": "issue:44623", "right": "issue:44749", "accept": false, - "reason": "Processor save files missing and Chinese performance regression after upgrade are different problems." + "reason": "Processor save_pretrained missing files is unrelated to the Chinese-reported preprocessing slowdown." }, { "left": "issue:44016", "right": "issue:45823", "accept": false, - "reason": "Notebook syntax error is unrelated to Gemma4 device-mismatch forward failure." + "reason": "Notebook syntax error and Gemma4 device mismatch are different classes of failure." }, { "left": "issue:44186", "right": "issue:47000", "accept": false, - "reason": "Tokenizer padding/truncation crash and RTDetr loss infeasibility are separate bugs." + "reason": "Tokenizer crash on NER/padding is unrelated to RTDetrLoss infeasible cost matrix." }, { "left": "issue:44623", "right": "issue:45397", "accept": false, - "reason": "Processor save artifacts missing is not the same as Gemma-4 zero3 from_pretrained loading failure." + "reason": "Both involve model/processor loading, but missing saved files and gemma-4 zero3 from_pretrained are distinct bugs." }, { "left": "issue:43525", "right": "issue:43881", "accept": false, - "reason": "Missing pad_token_id on Llama4Config and glm-4v-9b loading failure are different issues." + "reason": "Missing pad_token_id in config and glm-4v-9b load failure are not the same concrete bug." }, { - "left": "issue:43784", - "right": "issue:45542", + "left": "issue:43479", + "right": "issue:43525", "accept": false, - "reason": "Sentence-transformers import NameError is unrelated to tensorboard/TensorFlow backend installation behavior." + "reason": "Both are config initialization issues, but they affect different model configs and different missing fields/defaults." }, { "left": "issue:43901", "right": "issue:44246", "accept": false, - "reason": "Docs mismatch for return_all_scores is not the same as intermittent slow import behavior." + "reason": "Docs mismatch for return_all_scores is unrelated to slow imports." + }, + { + "left": "issue:43784", + "right": "issue:45542", + "accept": false, + "reason": "Sentence-transformers NameError and undefined TensorFlow backend are separate import failures." }, { "left": "issue:43673", "right": "issue:43761", "accept": false, - "reason": "Missing GenerationMixin cache in chunked prefill and CLIPVision hidden_states regression are unrelated." + "reason": "Both are v5 regressions, but one is chunked_prefill cache handling and the other is CLIPVisionModel hidden_states output." }, { "left": "issue:43525", "right": "issue:43643", "accept": false, - "reason": "Missing pad_token_id and missing fields from trust_remote_code AutoConfig are similar in shape but not the same concrete bug." + "reason": "Missing model config field and AutoConfig trust_remote_code field loss are different causes and fixes." }, { "left": "issue:44261", "right": "issue:44263", "accept": false, - "reason": "LayerNorm epsilon precision issue and torch.split return-value issue are unrelated implementation bugs." + "reason": "MLA rms_norm_eps precision issue is unrelated to GlmMoeDsaIndexer torch.split return values." }, { "left": "issue:43575", "right": "issue:47145", "accept": false, - "reason": "Tensor-parallel OOM when loading Qwen2 and FP8 weight_scale omission are different quantization/loading failures." - }, - { - "left": "issue:44589", - "right": "issue:45362", - "accept": false, - "reason": "Float8 storage lookup error is unrelated to Qwen3.5 chat crash." + "reason": "Qwen2 OOM on tensor-parallel load and FP8 weight_scale omission are different quantization/loading bugs." }, { "left": "issue:45325", "right": "issue:46851", "accept": false, - "reason": "Both concern Qwen3.5, but one is Qwen2.5-VL rope indexing and the other is variable-length text position handling." + "reason": "Both mention position_ids, but the Qwen2.5-VL temporal scaling bug is distinct from Qwen3.5 variable-length input handling." } ] }, @@ -13737,7 +13646,7 @@ "source": "llm", "cache_hit": false, "model": "gpt-5.4-mini?service_tier=flex", - "cluster_id": "cluster-10000-529", + "cluster_id": "cluster-10000-530", "nodes": [ "issue:42491", "issue:42617", @@ -13758,6 +13667,7 @@ "issue:44336", "issue:44373", "issue:44496", + "issue:44589", "issue:44871", "issue:45003", "issue:45072", @@ -13772,7 +13682,6 @@ "issue:45712", "issue:46291", "issue:46399", - "issue:46424", "issue:46506", "issue:46614", "issue:46655", @@ -13789,8 +13698,9 @@ "issue:47333" ], "soft_pairs": [ - "issue:45600|issue:45712", + "issue:44589|issue:45362", "issue:42796|issue:43441", + "issue:45600|issue:45712", "issue:44188|issue:44336", "issue:43824|issue:43976", "issue:45362|issue:47333", @@ -13800,8 +13710,8 @@ "issue:42617|issue:43976", "issue:44871|issue:47145", "issue:47176|issue:47177", - "issue:46291|issue:46399", "issue:44297|issue:45397", + "issue:46291|issue:46399", "issue:45083|issue:46851", "issue:46655|issue:46726", "issue:42491|issue:43299", @@ -13817,224 +13727,216 @@ "issue:45588|issue:46851", "issue:45561|issue:45563", "issue:46614|issue:46665", - "issue:43550|issue:46506", - "issue:44336|issue:46424", - "issue:43540|issue:46655" + "issue:43550|issue:46506" ], - "prepared_review_unit_hash": "9875d5399dd6e6beb5383bb1eed3b288326210e43ba094863a47cf52f5d7bde1", + "prepared_review_unit_hash": "106703d18fd7ab25468f0c5acb91a856ffac7444efc15e22b9bbe042497f458c", "component_budget": { - "node_count": 529, - "item_count": 529, - "soft_pair_count": 1953, - "serialized_chars": 1012074, - "estimated_input_tokens": 253019, - "estimated_eval_tokens": 506294 + "node_count": 530, + "item_count": 530, + "soft_pair_count": 1954, + "serialized_chars": 1013004, + "estimated_input_tokens": 253251, + "estimated_eval_tokens": 506758 }, "budget": { "node_count": 48, "item_count": 48, - "soft_pair_count": 31, - "serialized_chars": 37014, - "estimated_input_tokens": 9254, - "estimated_eval_tokens": 18764 + "soft_pair_count": 30, + "serialized_chars": 36711, + "estimated_input_tokens": 9178, + "estimated_eval_tokens": 18612 }, "overflow_policy": "truncate_then_skip", "trimmed": false, "aggressively_trimmed": false, "split": true, "analyst_result": { - "summary": "The set is mostly heterogeneous; none of the proposed soft pairs look like the same underlying bug/change closely enough to merge.", - "confidence": 0.74, - "canonical_issue_reason": "issue:42491 is the best anchor because it describes a concrete, versioned regression in Qwen3 MoE loading and is the closest to any near-duplicate in the set.", + "summary": "This cluster is a loose mix of mostly unrelated Transformers issues, many of them model-specific regression reports. None of the proposed soft edges look like the same underlying bug or change, so all should be rejected. No PRs were provided.", + "confidence": 0.96, + "canonical_issue_reason": null, "canonical_pr_reason": null, - "best_issue_reason": "issue:42491 is the strongest standalone issue for triage: clear repro context, explicit hf4.x vs hf5.x regression, and it is the most representative of the few model-loading-related items here.", + "best_issue_reason": null, "best_pr_reason": null, "soft_edge_verdicts": [ { - "left": "issue:45600", - "right": "issue:45712", + "left": "issue:44589", + "right": "issue:45362", "accept": false, - "reason": "Different cleanup bugs: removed Sam3LiteText config references vs leftover dummy classes leaking into exports/check_repo." + "reason": "Different failures: Float8 storage lookup vs Qwen3.5 chat crash; no shared code path." }, { "left": "issue:42796", "right": "issue:43441", "accept": false, - "reason": "Same model family but different concrete failures: tokenizer BPE marker handling vs FlashAttention loading/runtime error." + "reason": "Same model family but different components: tokenizer BPE handling vs FlashAttention runtime failure." + }, + { + "left": "issue:45600", + "right": "issue:45712", + "accept": false, + "reason": "Different cleanup/CI problems: removed Sam3LiteText configs vs leftover dummy classes." }, { "left": "issue:44188", "right": "issue:44336", "accept": false, - "reason": "Unrelated subsystems: torch.compile attention-kernel branching vs ANSI codes in loading_report." + "reason": "Attention-kernel torch.compile branching vs ANSI codes in loading_report; unrelated subsystems." }, { "left": "issue:43824", "right": "issue:43976", "accept": false, - "reason": "Different issues: missing Qwen2.5 VL import symbol vs Python version incompatibility in Transformers 5.1.0." + "reason": "Import error for a missing class vs Python version compatibility; distinct root causes." }, { "left": "issue:45362", "right": "issue:47333", "accept": false, - "reason": "Different Qwen3-related bugs: chat crash in Qwen3.5 vs saving disk-offloaded Qwen3-30B-A3B." + "reason": "Qwen3.5 chat crash and disk-offloaded save failure are different features and failures." }, { "left": "issue:44246", "right": "issue:44496", "accept": false, - "reason": "Import performance problem vs model config recognition failure; no shared code-path evidence." + "reason": "Import slowness is unrelated to unrecognized model/config loading." }, { "left": "issue:44336", "right": "issue:44373", "accept": false, - "reason": "Logging/ANSI output bug vs docstring issue for position_ids." + "reason": "Terminal ANSI output bug vs docstring typo; not the same bug." }, { "left": "issue:43723", "right": "issue:44162", "accept": false, - "reason": "Tokenizer loading regression vs ESM2 model breakage; different model families and symptoms." + "reason": "Tokenizer loading/import issue vs ESM2 model breakage; no concrete overlap." }, { "left": "issue:42617", "right": "issue:43976", "accept": false, - "reason": "3D parallel script failure is unrelated to the Python-version support issue." + "reason": "3d_parallel.py runtime issue and Python version support issue are unrelated." }, { "left": "issue:44871", "right": "issue:47145", "accept": false, - "reason": "Both involve model/config handling, but one is eos_token_id inconsistency and the other is FP8 weight_scale omission; different bugs." + "reason": "Both involve model loading/config, but one is eos_token_id mismatch and the other is missing FP8 weight_scale application." }, { "left": "issue:47176", "right": "issue:47177", "accept": false, - "reason": "Both are security-related, but command-substitution allowlist bypass and tar/zip-slip RCE are distinct vulnerabilities." + "reason": "Both are security-related, but one is shell-operator allowlist bypass and the other is archive path traversal; different vulnerabilities." }, { - "left": "issue:46291", - "right": "issue:46399", + "left": "issue:44297", + "right": "issue:45397", "accept": false, - "reason": "Both touch kernels integration, but one is pipelines import-time validation and the other is a Qwen3.5 rotary registration crash." + "reason": "Tokenizer save_pretrained metadata mismatch vs gemma-4 zero3 from_pretrained failure." }, { - "left": "issue:44297", - "right": "issue:45397", + "left": "issue:46291", + "right": "issue:46399", "accept": false, - "reason": "Tokenizer save_pretrained metadata mismatch vs gemma-4 zero3 from_pretrained failure; unrelated." + "reason": "Kernels import/version requirement failure vs Qwen3.5 rotary embedding registration crash; unrelated." }, { "left": "issue:45083", "right": "issue:46851", "accept": false, - "reason": "Both are Qwen3-family, but one is a feature-extractor length helper and the other is variable-length position_ids handling." + "reason": "Shared Qwen3 naming, but one is a helper length calculation bug and the other is variable-length position_ids handling." }, { "left": "issue:46655", "right": "issue:46726", "accept": false, - "reason": "Video padding calculation bug vs missing-vision-backend AutoProcessor error; different layers." + "reason": "Video padding calculation bug vs missing-vision-backend processor error; different code paths." }, { "left": "issue:42491", "right": "issue:43299", "accept": false, - "reason": "Both mention Qwen3 MoE loading on v5, but they target different model variants and failure modes; not safe to treat as the same bug." + "reason": "Both mention Qwen3 MoE, but the first is LoRA compatibility across hf4.x/hf5.x and the second is a model loading regression in 5.0.0.dev0." }, { "left": "issue:46399", "right": "issue:46851", "accept": false, - "reason": "Qwen3.5 kernel registration crash vs variable-length position_ids bug; different code paths." + "reason": "Rotary embedding registration error vs variable-length position_ids issue; not the same defect." }, { "left": "issue:45003", "right": "issue:47217", "accept": false, - "reason": "sys.modules access bug in modeling_utils is unrelated to VL image-placeholder handling." + "reason": "Unsafe sys.modules access is unrelated to image placeholder handling in VL models." }, { "left": "issue:45072", "right": "issue:46829", "accept": false, - "reason": "Different model-specific inference bugs: dtype mismatch under bfloat16 vs wrong tensor assigned to hidden_states." + "reason": "dtype mismatch in bfloat16 inference vs returning the wrong tensor in ImageClassifierOutput; distinct bugs." }, { "left": "issue:43526", "right": "issue:45003", "accept": false, - "reason": "BEiT image processor label reduction bug is unrelated to modeling_utils sys.modules access." + "reason": "BeitImageProcessorFast reduce_labels bug has no relation to sys.modules access." }, { "left": "issue:43540", "right": "issue:46752", "accept": false, - "reason": "Qwen3OmniMoe video-processing ValueError and empty-list chat-template IndexError are different bugs." + "reason": "Qwen3OmniMoe video-processing ValueError and empty-list chat template IndexError are unrelated." }, { "left": "issue:44246", "right": "issue:44871", "accept": false, - "reason": "Import slowdown vs Gemma-3 eos_token_id config mismatch; no shared underlying failure." + "reason": "Slow import times vs eos_token_id config inconsistency; no shared underlying cause." }, { "left": "issue:45003", "right": "issue:46821", "accept": false, - "reason": "modeling_utils sys.modules access bug and TimesFM window_size AttributeError are unrelated." + "reason": "sys.modules access safety issue and TimesFM window_size AttributeError are unrelated." }, { "left": "issue:45103", "right": "issue:46399", "accept": false, - "reason": "auto_docstring crash with future annotations is unrelated to the Qwen3.5 kernel registration problem." + "reason": "Auto-docstring future-annotations crash vs rotary embedding registration bug; different modules." }, { "left": "issue:44295", "right": "issue:45397", "accept": false, - "reason": "position_ids buffer access error vs gemma-4 zero3 from_pretrained failure; different issues." + "reason": "Position_ids buffer read error vs gemma-4 zero3 loading issue; not the same problem." }, { "left": "issue:45588", "right": "issue:46851", "accept": false, - "reason": "FlashAttention sink-less model crash vs Qwen3.5 variable-length position_ids bug; different paths." + "reason": "FlashAttention crash on s_aux=None is unrelated to variable-length position_ids handling." }, { "left": "issue:45561", "right": "issue:45563", "accept": false, - "reason": "Test-worker file race vs stale generate warning; unrelated testing issues." + "reason": "pytest-xdist captured_info.txt race and stale generate() warning are different testing/generation issues." }, { "left": "issue:46614", "right": "issue:46665", "accept": false, - "reason": "Phonemizer backend state bug vs from_pretrained behavior inconsistency in Python classes." + "reason": "Stale phonemizer backend state vs inconsistent from_pretrained inside Python classes; similar statefulness theme, but different behavior and code path." }, { "left": "issue:43550", "right": "issue:46506", "accept": false, - "reason": "torch.compile+SDPA failure in Bamba vs CPU/CUDA result divergence in PPDocLayoutV3; unrelated." - }, - { - "left": "issue:44336", - "right": "issue:46424", - "accept": false, - "reason": "ANSI output bug vs cache max length only honored for Quantized Cache." - }, - { - "left": "issue:43540", - "right": "issue:46655", - "accept": false, - "reason": "Different video-processing bugs in different model families: Qwen3OmniMoe vs SmolVLM." + "reason": "torch.compile + SDPA failure in Bamba vs CPU/CUDA result divergence in PPDocLayoutV3; different models and bugs." } ] }, @@ -14056,7 +13958,7 @@ "source": "llm", "cache_hit": false, "model": "gpt-5.4-mini?service_tier=flex", - "cluster_id": "cluster-10000-529", + "cluster_id": "cluster-10000-530", "nodes": [ "issue:36246", "issue:41950", @@ -14066,7 +13968,7 @@ "issue:43257", "issue:43295", "issue:43352", - "issue:43475", + "issue:43540", "issue:43596", "issue:43901", "issue:44060", @@ -14075,6 +13977,7 @@ "issue:44220", "issue:44242", "issue:44295", + "issue:44336", "issue:44351", "issue:44451", "issue:44466", @@ -14082,7 +13985,6 @@ "issue:44749", "issue:44779", "issue:44877", - "issue:44987", "issue:44995", "issue:45003", "issue:45137", @@ -14091,7 +13993,6 @@ "issue:45292", "issue:45335", "issue:45375", - "issue:45459", "issue:45561", "issue:45593", "issue:45600", @@ -14099,7 +14000,7 @@ "issue:45759", "issue:46132", "issue:46170", - "issue:46369", + "issue:46424", "issue:46655", "issue:46726", "issue:46752", @@ -14108,10 +14009,12 @@ "issue:47321" ], "soft_pairs": [ - "issue:42915|issue:47312", + "issue:44336|issue:46424", + "issue:43540|issue:46655", "issue:36246|issue:43596", - "issue:44242|issue:45593", + "issue:42915|issue:47312", "issue:44779|issue:45292", + "issue:44242|issue:45593", "issue:43257|issue:45137", "issue:43901|issue:45003", "issue:44479|issue:46655", @@ -14125,221 +14028,212 @@ "issue:42886|issue:46132", "issue:45137|issue:45600", "issue:44466|issue:45335", - "issue:44162|issue:45200", - "issue:46170|issue:47321", - "issue:43352|issue:45250", "issue:44451|issue:46752", + "issue:43352|issue:45250", "issue:44295|issue:44749", + "issue:46170|issue:47321", + "issue:44162|issue:45200", "issue:44060|issue:44077", "issue:44995|issue:45200", "issue:44877|issue:45375", - "issue:45003|issue:45561", - "issue:43475|issue:46369", - "issue:44479|issue:46170", - "issue:44987|issue:45459" + "issue:45003|issue:45561" ], - "prepared_review_unit_hash": "0d3602ca33fa12c626188467e0da8df7c0f7a9607bc903f6330d2c2a6a5bb494", + "prepared_review_unit_hash": "94a25f52ccb2203d62ea579f03cc3e9a85c404335913ac0fc2de20243d55d424", "component_budget": { - "node_count": 529, - "item_count": 529, - "soft_pair_count": 1953, - "serialized_chars": 1012074, - "estimated_input_tokens": 253019, - "estimated_eval_tokens": 506294 + "node_count": 530, + "item_count": 530, + "soft_pair_count": 1954, + "serialized_chars": 1013004, + "estimated_input_tokens": 253251, + "estimated_eval_tokens": 506758 }, "budget": { - "node_count": 48, - "item_count": 48, - "soft_pair_count": 29, - "serialized_chars": 36258, - "estimated_input_tokens": 9065, - "estimated_eval_tokens": 18386 + "node_count": 47, + "item_count": 47, + "soft_pair_count": 28, + "serialized_chars": 35359, + "estimated_input_tokens": 8840, + "estimated_eval_tokens": 17936 }, "overflow_policy": "truncate_then_skip", "trimmed": false, "aggressively_trimmed": false, "split": true, "analyst_result": { - "summary": "The items are largely unrelated across different models, configs, and utilities. I do not see a true duplicate cluster here; at most, a few pairs share broad subsystem themes, but not the same concrete bug.", - "confidence": 0.89, - "canonical_issue_reason": null, + "summary": "All candidate soft edges are false positives: the paired issues usually share a broad subsystem label (e.g. video, Qwen, deepspeed, tokenization) but describe different failures, different code paths, or different models. No PRs are present.", + "confidence": 0.84, + "canonical_issue_reason": "Issue 44479 is the strongest representative issue in this set: it describes a concrete, broad regression in video-input handling across multiple Qwen-family models, making it a plausible canonical duplicate target if one were needed.", "canonical_pr_reason": null, - "best_issue_reason": "If a representative issue is needed, issue:44479 is the broadest concrete regression report in the set, but the overall collection is too heterogeneous to treat it as a true canonical duplicate target.", + "best_issue_reason": "Issue 44479 is the best overall issue candidate because it is specific, user-facing, and spans several related model variants rather than a narrow one-off symptom.", "best_pr_reason": null, "soft_edge_verdicts": [ { - "left": "issue:42915", - "right": "issue:47312", + "left": "issue:44336", + "right": "issue:46424", + "accept": false, + "reason": "Unrelated bugs: ANSI color output in loading reports vs cache length behavior in quantized cache." + }, + { + "left": "issue:43540", + "right": "issue:46655", "accept": false, - "reason": "Different failures: Qwen3MoE FP8 config handling vs a generic kernel compatibility issue." + "reason": "Both mention video, but one is Qwen3OmniMoe input processing and the other is SmolVLM padding math; different models and failures." }, { "left": "issue:36246", "right": "issue:43596", "accept": false, - "reason": "Unrelated code paths: a Qwen2.5 VL import error vs a DeepSpeed ZeRO3/BERT indexing crash." + "reason": "ImportError for a missing Qwen image processor is unrelated to a DeepSpeed/BertModel zero3 IndexError." }, { - "left": "issue:44242", - "right": "issue:45593", + "left": "issue:42915", + "right": "issue:47312", "accept": false, - "reason": "Both mention missing auxiliary/loss behavior, but they concern different models and different mechanisms." + "reason": "Qwen3Moe FP8 loading failure and kernel compatibility are different problems with no shared code-path." }, { "left": "issue:44779", "right": "issue:45292", "accept": false, - "reason": "Tokenizer-output mismatch vs embedding resize propagation; same broad area, not the same bug." + "reason": "Tokenizer correctness regression is not the same as resize_token_embeddings not updating output embeddings." + }, + { + "left": "issue:44242", + "right": "issue:45593", + "accept": false, + "reason": "Load balancing loss omission and D-FINE auxiliary loss behavior are different training bugs." }, { "left": "issue:43257", "right": "issue:45137", "accept": false, - "reason": "Both involve DeepSpeed, but one is Qwen3 MoE weight conversion and the other is a ZeRO3 deque crash." + "reason": "MoE weight conversion with accelerate/deepspeed is unrelated to a deque underflow in ZeRO3 init." }, { "left": "issue:43901", "right": "issue:45003", "accept": false, - "reason": "Docs mismatch for a pipeline option vs unsafe sys.modules access in modeling_utils." + "reason": "Docs mismatch for return_all_scores is unrelated to unsafe sys.modules access in modeling_utils." }, { "left": "issue:44479", "right": "issue:46655", "accept": false, - "reason": "Both are video-related, but 44479 is a broad Qwen video-input regression while 46655 is a SmolVLM video-processor padding bug." + "reason": "Both concern video processors, but the regression is broader Qwen-family input handling while the other is a SmolVLM padding calculation bug." }, { "left": "issue:41950", "right": "issue:44220", "accept": false, - "reason": "Video-classification image-processor lookup vs audio feature extraction; not the same bug." + "reason": "Video-classification pipeline image-processor lookup and _torch_extract_fbank_features are different pipeline issues." }, { "left": "issue:46726", "right": "issue:47217", "accept": false, - "reason": "Both are VL-related, but one is an AutoProcessor backend error and the other is a placeholder-handling crash." + "reason": "Missing vision backend / unrecognized image processor is not the same as image placeholders inside user text causing VL model failure." }, { "left": "issue:45375", "right": "issue:45759", "accept": false, - "reason": "Different config problems: a missing vision config field vs AutoModelForCausalLM not unwrapping text_config." + "reason": "Missing config field dropped by strict parsing is unrelated to AutoModelForCausalLM not unwrapping text_config." }, { "left": "issue:45292", "right": "issue:45701", "accept": false, - "reason": "Embedding resize behavior vs tokenizer version-dependent output; no shared concrete failure." + "reason": "resize_token_embeddings wiring bug and tokenization changes across versions are different symptoms and code paths." }, { "left": "issue:36246", "right": "issue:44351", "accept": false, - "reason": "Separate import errors for different symbols/modules." + "reason": "Two unrelated import errors: missing Qwen2_5_VLImageProcessor vs missing HybridCache export." }, { "left": "issue:36246", "right": "issue:42757", "accept": false, - "reason": "Different missing imports in different dependencies; same symptom class, not same bug." + "reason": "Missing Qwen image processor and missing is_offline_mode are separate import surface regressions." }, { "left": "issue:43295", "right": "issue:45292", "accept": false, - "reason": "One is a regression around processor/tokenizer image handling, the other is embedding resize propagation." + "reason": "Custom processor.tokenizer regression is not the same as resize_token_embeddings not affecting output embeddings." }, { "left": "issue:42886", "right": "issue:46132", "accept": false, - "reason": "Offline cache loading vs AttentionInterface.register behavior; unrelated." + "reason": "Offline cache loading failure and AttentionInterface.register behavior change are unrelated." }, { "left": "issue:45137", "right": "issue:45600", "accept": false, - "reason": "Different DeepSpeed/modeling issues with no common code path." + "reason": "ZeRO3 deque/indexing crash is unrelated to auto_mappings still referencing removed configs." }, { "left": "issue:44466", "right": "issue:45335", "accept": false, - "reason": "Both concern embeddings, but one is serialization/tied-weights behavior and the other is resize_token_embeddings propagation." + "reason": "Both involve tied/resized embeddings, but one is serialization inconsistency and the other is decoder.embed_tokens not updating." }, { - "left": "issue:44162", - "right": "issue:45200", + "left": "issue:44451", + "right": "issue:46752", "accept": false, - "reason": "Completely different model families and failure modes." + "reason": "Model loading failure for ScandiBERT is unrelated to apply_chat_template on an empty list." }, { - "left": "issue:46170", - "right": "issue:47321", + "left": "issue:43352", + "right": "issue:45250", "accept": false, - "reason": "Both mention AMD, but one is a Windows ROCm import/dependency issue and the other is a quark class update problem." + "reason": "Model-specific unsupported Flash Attention 2 warning is not the same as the general Flash Attention 2.0 issue." }, { - "left": "issue:43352", - "right": "issue:45250", + "left": "issue:44295", + "right": "issue:44749", "accept": false, - "reason": "A model-specific unsupported-feature error vs a generic Flash Attention issue." + "reason": "A position_ids buffer access error is different from a performance regression after upgrading Transformers." }, { - "left": "issue:44451", - "right": "issue:46752", + "left": "issue:46170", + "right": "issue:47321", "accept": false, - "reason": "Loading a specific model vs apply_chat_template on an empty list; unrelated." + "reason": "Windows/ROCm import failure due to torch.distributed is unrelated to AMD quark class updates." }, { - "left": "issue:44295", - "right": "issue:44749", + "left": "issue:44162", + "right": "issue:45200", "accept": false, - "reason": "A buffer access error vs a performance regression after version upgrade." + "reason": "ESM2 breakage and Gemma 4 mm_token_type_ids defaulting are different model bugs." }, { "left": "issue:44060", "right": "issue:44077", "accept": false, - "reason": "Different model/config issues with no shared underlying bug." + "reason": "Incorrect tied-weights warning in Qwen3-Next is unrelated to patchtsmixer's post_init allowance." }, { "left": "issue:44995", "right": "issue:45200", "accept": false, - "reason": "GlmMoeDsa cache crash vs Gemma 4 token-type defaults; unrelated." + "reason": "GlmMoeDsa stale indexer cache and Gemma 4 token type IDs are separate issues." }, { "left": "issue:44877", "right": "issue:45375", "accept": false, - "reason": "Both involve strict config handling, but they affect different model configs and different missing-field behaviors." + "reason": "Strict config blocking granite_speech loading is not the same as missing deepstack_visual_indexes in Qwen3_5MoeVisionConfig." }, { "left": "issue:45003", "right": "issue:45561", "accept": false, - "reason": "Unsafe module access vs a pytest-xdist file race; unrelated." - }, - { - "left": "issue:43475", - "right": "issue:46369", - "accept": false, - "reason": "Same model family, but a missing output field and a memory overflow are different concrete bugs." - }, - { - "left": "issue:44479", - "right": "issue:46170", - "accept": false, - "reason": "Video regression vs Windows/ROCm import dependency failure; different subsystems." - }, - { - "left": "issue:44987", - "right": "issue:45459", - "accept": false, - "reason": "Loading failure for a model vs tokenizer error masking when protobuf is absent; not the same issue." + "reason": "Unsafe sys.modules access is unrelated to the xdist captured_info.txt race." } ] }, @@ -14361,7 +14255,7 @@ "source": "llm", "cache_hit": false, "model": "gpt-5.4-mini?service_tier=flex", - "cluster_id": "cluster-10000-529", + "cluster_id": "cluster-10000-530", "nodes": [ "issue:29870", "issue:30333", @@ -14382,15 +14276,18 @@ "issue:44242", "issue:44361", "issue:44368", + "issue:44479", "issue:44623", "issue:44749", "issue:44857", "issue:44877", "issue:44977", + "issue:44987", "issue:44995", "issue:45229", "issue:45397", "issue:45412", + "issue:45459", "issue:45563", "issue:45588", "issue:45656", @@ -14398,20 +14295,20 @@ "issue:45987", "issue:46082", "issue:46123", - "issue:46132", "issue:46170", "issue:46202", + "issue:46369", "issue:46431", "issue:46506", "issue:46710", "issue:46726", - "issue:46752", "issue:46851", - "issue:47020", - "issue:47145", "issue:47312" ], "soft_pairs": [ + "issue:43475|issue:46369", + "issue:44479|issue:46170", + "issue:44987|issue:45459", "issue:44995|issue:45588", "issue:29870|issue:30333", "issue:44623|issue:46726", @@ -14420,8 +14317,8 @@ "issue:45685|issue:46431", "issue:44242|issue:44857", "issue:45397|issue:45656", - "issue:44368|issue:44977", "issue:43475|issue:43540", + "issue:44368|issue:44977", "issue:43653|issue:44162", "issue:42994|issue:47312", "issue:43723|issue:46710", @@ -14431,185 +14328,195 @@ "issue:42886|issue:45229", "issue:43931|issue:44368", "issue:43906|issue:46202", - "issue:45987|issue:46082", "issue:44188|issue:46202", "issue:44361|issue:44749", "issue:44368|issue:45563", - "issue:46752|issue:47020", - "issue:46132|issue:47145" + "issue:45987|issue:46082" ], - "prepared_review_unit_hash": "73da193552670dbecb2ddf36ba5d2862474b3d811454c4e119666a35e6242a81", + "prepared_review_unit_hash": "7ad5713d6f7fbb818ee1aff5dc6508985e5036a6907b031b541ee0a3726926da", "component_budget": { - "node_count": 529, - "item_count": 529, - "soft_pair_count": 1953, - "serialized_chars": 1012074, - "estimated_input_tokens": 253019, - "estimated_eval_tokens": 506294 + "node_count": 530, + "item_count": 530, + "soft_pair_count": 1954, + "serialized_chars": 1013004, + "estimated_input_tokens": 253251, + "estimated_eval_tokens": 506758 }, "budget": { "node_count": 47, "item_count": 47, - "soft_pair_count": 25, - "serialized_chars": 34543, - "estimated_input_tokens": 8636, - "estimated_eval_tokens": 17528 + "soft_pair_count": 26, + "serialized_chars": 34834, + "estimated_input_tokens": 8709, + "estimated_eval_tokens": 17674 }, "overflow_policy": "truncate_then_skip", "trimmed": false, "aggressively_trimmed": false, "split": true, "analyst_result": { - "summary": "None of the proposed soft pairs look like true duplicates; they mostly share only a broad subsystem or a similar failure type, not the same concrete bug. The clearest standalone issue in the set is the explicit crash report in issue 45588.", - "confidence": 0.96, - "canonical_issue_reason": "issue:45588 is the most concrete and actionable bug report here: a specific crash, a named file/path, and a clear trigger condition (`s_aux=None` on sink-less models).", + "summary": "This set is not a true duplicate cluster; it contains many unrelated bugs across tokenizers, vision models, MoE, caching, and training/inference paths. No PRs are present. All soft-edge pairs should be rejected.", + "confidence": 0.98, + "canonical_issue_reason": "No canonical issue: the items are heterogeneous and do not share one underlying defect or code path.", "canonical_pr_reason": null, - "best_issue_reason": "issue:45588 is the strongest canonical candidate overall because it is narrowly scoped, reproducible, and points to a specific code-path fix.", + "best_issue_reason": "No global best issue: none clearly subsumes the others, so selecting one as the representative would be arbitrary.", "best_pr_reason": null, "soft_edge_verdicts": [ + { + "left": "issue:43475", + "right": "issue:46369", + "accept": false, + "reason": "Both involve SAM3, but one is an encoder attribute error and the other is a litetext memory overflow; different failures and code paths." + }, + { + "left": "issue:44479", + "right": "issue:46170", + "accept": false, + "reason": "Qwen video-input regression and a Windows/ROCm distributed import error are unrelated bugs." + }, + { + "left": "issue:44987", + "right": "issue:45459", + "accept": false, + "reason": "One is a model loading failure for a specific repo; the other is tokenizer error masking when protobuf is missing." + }, { "left": "issue:44995", "right": "issue:45588", "accept": false, - "reason": "Different code paths and failures: stale indexer cache on second forward pass vs flash-attention AttributeError on `s_aux=None`." + "reason": "Different subsystems and symptoms: GlmMoeDsa stale cache on second forward versus flash-attention crashing on s_aux=None." }, { "left": "issue:29870", "right": "issue:30333", "accept": false, - "reason": "Unrelated problems: BART generation `decoder_attention_mask` bug vs MLFlow reporting after failed training jobs." + "reason": "A BART generate() UnboundLocalError and MLflow reporting on failed jobs are unrelated." }, { "left": "issue:44623", "right": "issue:46726", "accept": false, - "reason": "One is `save_pretrained` missing files; the other is `AutoProcessor` loading with a missing vision backend. Different directions and causes." + "reason": "Both mention processors, but saving missing files and loading with missing vision backend are distinct issues." }, { "left": "issue:44038", "right": "issue:46506", "accept": false, - "reason": "Qwen3-VL-Moe/Transformers 5.0 issue and PPDocLayoutV3 CPU-vs-CUDA divergence are separate bugs." + "reason": "A Qwen3-VL-Moe v5 bug and CPU/CUDA divergence in PPDocLayoutV3 do not share a concrete failure path." }, { "left": "issue:44877", "right": "issue:46170", "accept": false, - "reason": "Config strictness for `granite_speech` is unrelated to a Windows ROCm import failure from `torch.distributed` dependency handling." + "reason": "Strict Granite Speech config loading and a Windows ROCm torch.distributed import problem are unrelated." }, { "left": "issue:45685", "right": "issue:46431", "accept": false, - "reason": "MPS MoE histogram-int error and missing chat template in a processor are different failures in different subsystems." + "reason": "MoE histogram_mps on Int and missing chat template are different errors in different components." }, { "left": "issue:44242", "right": "issue:44857", "accept": false, - "reason": "Load-balancing loss omission and AMP/CUDA crash in `LwDetrImageLoss` are not the same bug." + "reason": "Loss routing omission in MoE and an AMP/CUDA crash in LwDetrImageLoss are not the same bug." }, { "left": "issue:45397", "right": "issue:45656", "accept": false, - "reason": "`from_pretrained` under Zero3 vs optimizer stepping twice under deepspeed are different code paths and symptoms." + "reason": "Deepspeed from_pretrained loading and duplicated optimizer stepping are separate problems." }, { - "left": "issue:44368", - "right": "issue:44977", + "left": "issue:43475", + "right": "issue:43540", "accept": false, - "reason": "A warning about `tie_word_embeddings` is unrelated to flash-attention generation failure in Qwen3.5." + "reason": "Different SAM3 video/vision error versus Qwen3OmniMoe video input processing error." }, { - "left": "issue:43475", - "right": "issue:43540", + "left": "issue:44368", + "right": "issue:44977", "accept": false, - "reason": "Both are video-related, but they involve different models and distinct errors: missing attribute vs input processing ValueError." + "reason": "A warning about tie_word_embeddings in fine-tuning is not the same as a flash-attention generation failure." }, { "left": "issue:43653", "right": "issue:44162", "accept": false, - "reason": "Tokenizer special-token registration bug and ESM2 being broken are not the same underlying issue." + "reason": "BigBirdTokenizer special-token registration and ESM2 breakage are unrelated." }, { "left": "issue:42994", "right": "issue:47312", "accept": false, - "reason": "Quantized model saving failure is separate from kernel-version incompatibility." + "reason": "Quantized model save failure and kernel compatibility mismatch are different topics." }, { "left": "issue:43723", "right": "issue:46710", "accept": false, - "reason": "Tokenizer loading in v5 and an accuracy regression after upgrading Transformers are different problem classes." + "reason": "Tokenizer loading in v5 and accuracy regression after an upgrade are not the same underlying bug." }, { "left": "issue:42199", "right": "issue:46851", "accept": false, - "reason": "DETR cardinality validation and Qwen3.5 variable-length position_ids are unrelated bugs." + "reason": "DETR cardinality/background-class logic and Qwen3.5 variable-length position_ids are unrelated." }, { "left": "issue:43065", "right": "issue:43408", "accept": false, - "reason": "A dummy `Conv2d` in `Sam3PixelDecoder` is not the same as a model-type mismatch warning for `sam3_video` vs `sam3_tracker`." + "reason": "Dummy Conv2d in Sam3PixelDecoder and a model-type mismatch warning are separate SAM3 issues." }, { "left": "issue:45412", "right": "issue:46123", "accept": false, - "reason": "RT-DETR memory not released and partial-batch masking results being dropped are different behaviors." + "reason": "RT-DETR memory retention and MaskGenerationPipeline final-batch dropping are unrelated." }, { "left": "issue:42886", "right": "issue:45229", "accept": false, - "reason": "Offline tokenizer cache loading and Gemma4 multi-GPU CUDA OOM are unrelated." + "reason": "Offline tokenizer cache loading and Gemma4 multi-GPU CUDA OOM are different failures." }, { "left": "issue:43931", "right": "issue:44368", "accept": false, - "reason": "Weight-shape mismatch on Qwen3-VL-30B is unrelated to a `tie_word_embeddings` warning." + "reason": "Qwen3-VL-30B weight-shape mismatch and a Qwen3.5 warning about tied embeddings are not the same bug." }, { "left": "issue:43906", "right": "issue:46202", "accept": false, - "reason": "An isolated reproduction of another issue is not enough to treat it as the same as Gemma pipeline roundtrip failure; the titles do not identify the same bug." - }, - { - "left": "issue:45987", - "right": "issue:46082", - "accept": false, - "reason": "StaticCache return-type contract bug and LlamaConfig head_dim validation are different issues." + "reason": "An isolated reproduction of another issue and a Gemma pipeline roundtrip failure do not match as a concrete duplicate." }, { "left": "issue:44188", "right": "issue:46202", "accept": false, - "reason": "torch.compile attention-kernel divergence is unrelated to Gemma pipeline roundtrip failure on the `\u2581` character." + "reason": "Torch.compile attention-kernel divergence and Gemma roundtrip tokenization failure are unrelated." }, { "left": "issue:44361", "right": "issue:44749", "accept": false, - "reason": "MLLukeTokenizer AttributeError and a 10x data filtering slowdown after upgrade are not the same bug." + "reason": "MLukeTokenizer AttributeError and a transformers performance regression are different problems." }, { - "left": "issue:46752", - "right": "issue:47020", + "left": "issue:44368", + "right": "issue:45563", "accept": false, - "reason": "Empty chat-template `IndexError` and XLMRobertaTokenizer passing a dict to `Unigram` are different tokenizer failures." + "reason": "A fine-tuning warning and a stale generate() warning for num_return_sequences are not the same issue." }, { - "left": "issue:46132", - "right": "issue:47145", + "left": "issue:45987", + "right": "issue:46082", "accept": false, - "reason": "Behavior change in `AttentionInterface.register` is unrelated to FP8 frozen checkpoints loading without `weight_scale`." + "reason": "StaticCache return-type contract violation and LlamaConfig head_dim validation are unrelated." } ] }, @@ -14631,7 +14538,7 @@ "source": "llm", "cache_hit": false, "model": "gpt-5.4-mini?service_tier=flex", - "cluster_id": "cluster-10000-529", + "cluster_id": "cluster-10000-530", "nodes": [ "issue:28282", "issue:30990", @@ -14644,7 +14551,6 @@ "issue:43723", "issue:43784", "issue:43994", - "issue:44038", "issue:44162", "issue:44188", "issue:44297", @@ -14652,13 +14558,11 @@ "issue:44483", "issue:44623", "issue:44749", - "issue:44869", "issue:44871", "issue:44964", "issue:45103", "issue:45137", "issue:45230", - "issue:45356", "issue:45375", "issue:45397", "issue:45478", @@ -14673,250 +14577,250 @@ "issue:46421", "issue:46463", "issue:46726", + "issue:46752", "issue:46821", "issue:46851", "issue:46858", "issue:46877", "issue:47000", + "issue:47020", "issue:47068", + "issue:47145", "issue:47217" ], "soft_pairs": [ + "issue:46752|issue:47020", + "issue:46132|issue:47145", "issue:45103|issue:46858", "issue:44483|issue:46202", "issue:45593|issue:47068", - "issue:44188|issue:46421", "issue:43994|issue:46726", + "issue:44188|issue:46421", "issue:43701|issue:44749", "issue:44749|issue:45656", "issue:44297|issue:44623", "issue:43065|issue:46123", "issue:43646|issue:44162", - "issue:43475|issue:46821", "issue:44964|issue:45137", + "issue:43475|issue:46821", "issue:43475|issue:43643", + "issue:44623|issue:45656", "issue:43784|issue:46399", "issue:39692|issue:46726", - "issue:44623|issue:45656", "issue:46463|issue:46877", - "issue:28282|issue:30990", "issue:44373|issue:46851", + "issue:28282|issue:30990", "issue:44871|issue:46132", "issue:45561|issue:47217", "issue:43723|issue:46326", "issue:46877|issue:47000", - "issue:45375|issue:45478", - "issue:44869|issue:45356", "issue:44188|issue:46851", - "issue:45230|issue:45397", - "issue:44038|issue:44623" + "issue:45375|issue:45478", + "issue:45230|issue:45397" ], - "prepared_review_unit_hash": "73c85117aba221897f7ae459e57c120a2af3e2346ac3c50f5ebe03a8b565ba1d", + "prepared_review_unit_hash": "422bc0a996841b5031dfb03cb65deaacfdb170ec6e2f1e917711738b0dbf8fad", "component_budget": { - "node_count": 529, - "item_count": 529, - "soft_pair_count": 1953, - "serialized_chars": 1012074, - "estimated_input_tokens": 253019, - "estimated_eval_tokens": 506294 + "node_count": 530, + "item_count": 530, + "soft_pair_count": 1954, + "serialized_chars": 1013004, + "estimated_input_tokens": 253251, + "estimated_eval_tokens": 506758 }, "budget": { "node_count": 47, "item_count": 47, "soft_pair_count": 28, - "serialized_chars": 35468, - "estimated_input_tokens": 8867, - "estimated_eval_tokens": 17990 + "serialized_chars": 35477, + "estimated_input_tokens": 8870, + "estimated_eval_tokens": 17996 }, "overflow_policy": "truncate_then_skip", "trimmed": false, "aggressively_trimmed": false, "split": true, "analyst_result": { - "summary": "Mostly a heterogeneous set of Transformers issues; the only plausible duplicate edge is the Qwen3.5 Moe config/loading regression pair. No PRs are present.", - "confidence": 0.52, - "canonical_issue_reason": "issue:45375 is the most concrete root-cause report: it names the specific Qwen3_5MoeVisionConfig field omission and the @strict drop behavior that plausibly drives the broader from_pretrained failure in the paired report.", + "summary": "This cluster is highly heterogeneous: the paired items share only superficial wording or broad subsystems, not the same underlying bug/change. I would not collapse any of the soft-edge pairs.", + "confidence": 0.97, + "canonical_issue_reason": null, "canonical_pr_reason": null, - "best_issue_reason": "issue:45375 is the clearest, most actionable issue in the set and best represents the likely Qwen3.5 Moe config-loading regression cluster.", + "best_issue_reason": null, "best_pr_reason": null, "soft_edge_verdicts": [ + { + "left": "issue:46752", + "right": "issue:47020", + "accept": false, + "reason": "Different tokenizer APIs and failures: empty chat template handling vs Unigram init type mismatch." + }, + { + "left": "issue:46132", + "right": "issue:47145", + "accept": false, + "reason": "Unrelated areas: attention registration behavior vs FP8 checkpoint weight scaling on load." + }, { "left": "issue:45103", "right": "issue:46858", "accept": false, - "reason": "Both are internal runtime bugs, but one is auto-docstring annotation handling and the other is cache seq-length arithmetic; different code paths." + "reason": "Docstring/introspection crash vs cache length tensor arithmetic; no shared code path." }, { "left": "issue:44483", "right": "issue:46202", "accept": false, - "reason": "One is a chat/completions request validation failure, the other is a Gemma tokenizer roundtrip issue." + "reason": "API request rejection vs Gemma roundtrip tokenization issue; different failure modes." }, { "left": "issue:45593", "right": "issue:47068", "accept": false, - "reason": "Both mention training/deepspeed, but the concrete bugs differ: auxiliary-loss logic vs Python branching on a CUDA tensor." + "reason": "Both training-related, but one is loss composition and the other is a CUDA conditional sync bug." }, { - "left": "issue:44188", - "right": "issue:46421", + "left": "issue:43994", + "right": "issue:46726", "accept": false, - "reason": "Both touch compile/cache behavior, but one is attention-kernel branching and the other is lazy cache initialization causing recompilation." + "reason": "Both vision/model loading related, but one is a broken example and the other is missing backend handling." }, { - "left": "issue:43994", - "right": "issue:46726", + "left": "issue:44188", + "right": "issue:46421", "accept": false, - "reason": "Different failures: SigLIP2 model/pipeline output corruption vs missing vision backend error reporting in AutoProcessor." + "reason": "Compile-time attention kernel branching vs cache lazy-init recompilation; distinct mechanisms." }, { "left": "issue:43701", "right": "issue:44749", "accept": false, - "reason": "Checkpoint key mismatch and data-filter slowdown are unrelated symptoms in different subsystems." + "reason": "Checkpoint key mismatch and training slowdown are unrelated." }, { "left": "issue:44749", "right": "issue:45656", "accept": false, - "reason": "Upgrade-induced filtering slowdown is not the same bug as double optimizer steps under deepspeed." + "reason": "Performance regression from data filtering vs duplicate optimizer.step with DeepSpeed." }, { "left": "issue:44297", "right": "issue:44623", "accept": false, - "reason": "Both are save_pretrained issues, but one is tokenizer metadata mismatch and the other is processor file omission." + "reason": "Both save_pretrained-related, but tokenizer metadata mismatch is not the same as missing processor files." }, { "left": "issue:43065", "right": "issue:46123", "accept": false, - "reason": "Different model/pipeline areas: SAM3 pixel decoder placeholder vs MaskGenerationPipeline batch handling." + "reason": "SAM3 pixel decoder stub vs mask pipeline batch-end logic; no common bug." }, { "left": "issue:43646", "right": "issue:44162", "accept": false, - "reason": "Generic AutoConfig trust_remote_code field loss is not the same bug as ESM2 being broken." + "reason": "Custom config field loss vs ESM2 model breakage; too different to merge." }, { - "left": "issue:43475", - "right": "issue:46821", + "left": "issue:44964", + "right": "issue:45137", "accept": false, - "reason": "Both are attribute errors, but they affect unrelated models and different missing fields." + "reason": "Model loading failure vs DeepSpeed ZeRO3 deque error; unrelated code paths." }, { - "left": "issue:44964", - "right": "issue:45137", + "left": "issue:43475", + "right": "issue:46821", "accept": false, - "reason": "A multimodal model load failure and a DeepSpeed ZeRO3 deque error are unrelated." + "reason": "Missing vision encoder attribute vs TimesFM window_size attribute error; different models and causes." }, { "left": "issue:43475", "right": "issue:43643", "accept": false, - "reason": "Both involve missing fields, but one is a SAM3 vision-output attribute and the other is AutoConfig return-object completeness." + "reason": "Different missing-field symptoms in different objects; not the same defect." + }, + { + "left": "issue:44623", + "right": "issue:45656", + "accept": false, + "reason": "Save-time artifact omission vs optimizer stepping twice; unrelated bugs." }, { "left": "issue:43784", "right": "issue:46399", "accept": false, - "reason": "Import-time NameError in sentence-transformers is unrelated to Qwen3.5 kernel registration failure." + "reason": "Import NameError in sentence-transformers vs Qwen3.5 kernel registration crash." }, { "left": "issue:39692", "right": "issue:46726", "accept": false, - "reason": "Documentation example errors are not the same underlying bug as runtime backend-missing handling." - }, - { - "left": "issue:44623", - "right": "issue:45656", - "accept": false, - "reason": "Processor save_pretrained file emission and deepspeed optimizer stepping are separate problems." + "reason": "Doc example mistakes vs backend-missing processor error; not duplicate." }, { "left": "issue:46463", "right": "issue:46877", "accept": false, - "reason": "T5 decoder-input validation messaging and auto_docstring stdout logging are unrelated." + "reason": "T5 validation messaging vs auto_docstring stdout logging; unrelated." }, { - "left": "issue:28282", - "right": "issue:30990", + "left": "issue:44373", + "right": "issue:46851", "accept": false, - "reason": "PyTorch missing ImportError and sentence-transformers load hang are different failure modes." + "reason": "Wrong docstring only vs runtime position_ids bug; different kinds of fixes." }, { - "left": "issue:44373", - "right": "issue:46851", + "left": "issue:28282", + "right": "issue:30990", "accept": false, - "reason": "A docstring mismatch and a real variable-length position_ids bug are not the same change." + "reason": "PyTorch-missing import error vs hanging load in sentence-transformers; unrelated." }, { "left": "issue:44871", "right": "issue:46132", "accept": false, - "reason": "Gemma eos_token_id config inconsistency is unrelated to AttentionInterface.register behavior changes." + "reason": "Gemma eos_token_id config mismatch vs attention registration behavior change." }, { "left": "issue:45561", "right": "issue:47217", "accept": false, - "reason": "A pytest-xdist file race and VL image-placeholder parsing are unrelated." + "reason": "pytest-xdist file race vs multimodal placeholder parsing; no overlap." }, { "left": "issue:43723", "right": "issue:46326", "accept": false, - "reason": "Tokenizer loading in v5 and continue_final_message for mistral_common are different tokenizer issues." - }, - { - "left": "issue:46877", - "right": "issue:47000", - "accept": false, - "reason": "Import-time doc-lint logging and RTDetrLoss infeasible cost matrix are unrelated." + "reason": "Tokenizer loading regression in v5 vs continue_final_message handling for mistral_common." }, - { - "left": "issue:45375", - "right": "issue:45478", - "accept": true, - "reason": "Same model family and same from_pretrained/config-loading regression; 45478 reads like the broader symptom of the specific missing-field bug in 45375." - }, - { - "left": "issue:44869", - "right": "issue:45356", + { + "left": "issue:46877", + "right": "issue:47000", "accept": false, - "reason": "Whisper timestamp decoding with trailing replacement chars and Kimi tokenizer codec regression are different edge cases." + "reason": "Import-time doc-lint output vs RTDetr loss infeasibility; unrelated." }, { "left": "issue:44188", "right": "issue:46851", "accept": false, - "reason": "Both mention position/compile-adjacent model behavior, but one is attention-kernel branching and the other is variable-length position_ids handling." + "reason": "Both involve Qwen3.5/positioning, but one is torch.compile kernel divergence and the other is variable-length input handling." }, { - "left": "issue:45230", - "right": "issue:45397", + "left": "issue:45375", + "right": "issue:45478", "accept": false, - "reason": "A generic bug report and a gemma-4 zero3 from_pretrained issue do not show the same root cause." + "reason": "Missing config field silently dropped vs broader Qwen3.5 from_pretrained error; same family but different bug." }, { - "left": "issue:44038", - "right": "issue:44623", + "left": "issue:45230", + "right": "issue:45397", "accept": false, - "reason": "Qwen3-VL-Moe model bug and processor.save_pretrained missing files are unrelated." + "reason": "Generic bug report placeholder vs gemma-4 ZeRO3 loading issue." } ] }, - "evaluator_result": { - "accept": true, - "feedback": "The summary is grounded in the packet and stays conservative. The only accepted soft edge, issue:45375 \u2194 issue:45478, is a plausible same-bug regression pair in the same Qwen3.5 MoE config/loading path, while the other rejected pairs are explained cautiously and do not overclaim overlap. No PRs are present, so the null PR fields are fine." - }, - "evaluator_used": true, + "evaluator_result": null, + "evaluator_used": false, "retried": false, - "accepted_nontrivial_soft_edge": true, + "accepted_nontrivial_soft_edge": false, "error_kind": null, "error_message": null }, @@ -14931,7 +14835,7 @@ "source": "llm", "cache_hit": false, "model": "gpt-5.4-mini?service_tier=flex", - "cluster_id": "cluster-10000-529", + "cluster_id": "cluster-10000-530", "nodes": [ "issue:30333", "issue:30990", @@ -14955,11 +14859,12 @@ "issue:44351", "issue:44496", "issue:44568", + "issue:44623", "issue:44661", "issue:44743", "issue:44792", - "issue:44843", "issue:44857", + "issue:44869", "issue:44933", "issue:44945", "issue:45081", @@ -14967,6 +14872,7 @@ "issue:45103", "issue:45137", "issue:45310", + "issue:45356", "issue:45375", "issue:45405", "issue:45406", @@ -14978,16 +14884,17 @@ "issue:46701", "issue:46877", "issue:47059", - "issue:47068", - "issue:47085" + "issue:47068" ], "soft_pairs": [ + "issue:44869|issue:45356", + "issue:44038|issue:44623", "issue:43232|issue:43673", "issue:45310|issue:45375", "issue:41628|issue:44933", - "issue:43873|issue:44945", "issue:36010|issue:41628", "issue:44661|issue:47059", + "issue:43873|issue:44945", "issue:43278|issue:45137", "issue:42371|issue:45092", "issue:45081|issue:45103", @@ -15000,244 +14907,233 @@ "issue:41628|issue:44351", "issue:42175|issue:44933", "issue:44222|issue:45310", - "issue:44945|issue:45753", "issue:30333|issue:33290", + "issue:44945|issue:45753", "issue:45405|issue:46877", "issue:45561|issue:47059", "issue:44743|issue:47068", "issue:42898|issue:44568", "issue:42831|issue:44222", - "issue:42175|issue:45712", "issue:43784|issue:43824", - "issue:44496|issue:45561", + "issue:42175|issue:45712", "issue:43976|issue:44933", - "issue:30990|issue:35141", - "issue:44568|issue:47085", - "issue:44843|issue:47085" + "issue:44496|issue:45561", + "issue:30990|issue:35141" ], - "prepared_review_unit_hash": "981a221b35cbd06dc5ff18b814a16fdf14effd25c1d96d679104bf69c29cd901", + "prepared_review_unit_hash": "9b54c4fc9971f82ce906fd5cdae6c31c6f7b3c8738339ca6743377d345b96e3e", "component_budget": { - "node_count": 529, - "item_count": 529, - "soft_pair_count": 1953, - "serialized_chars": 1012074, - "estimated_input_tokens": 253019, - "estimated_eval_tokens": 506294 + "node_count": 530, + "item_count": 530, + "soft_pair_count": 1954, + "serialized_chars": 1013004, + "estimated_input_tokens": 253251, + "estimated_eval_tokens": 506758 }, "budget": { - "node_count": 47, - "item_count": 47, + "node_count": 48, + "item_count": 48, "soft_pair_count": 32, - "serialized_chars": 36706, - "estimated_input_tokens": 9177, - "estimated_eval_tokens": 18610 + "serialized_chars": 37293, + "estimated_input_tokens": 9324, + "estimated_eval_tokens": 18904 }, "overflow_policy": "truncate_then_skip", "trimmed": false, "aggressively_trimmed": false, "split": true, "analyst_result": { - "summary": "The cluster is mostly heterogeneous, with one notable near-match in the Qwen3.5/MoE area but not enough evidence to merge it as the same bug. I kept the soft edges conservative overall.", - "confidence": 0.77, - "canonical_issue_reason": "No single issue cleanly represents the whole cluster; issue 44843 is the best stand-in because it is a concrete, reproducible regression with a clear user-facing failure mode.", + "summary": "This cluster is not a duplicate set; it contains many unrelated bug reports that were only weakly text-similar. I would not merge any of the proposed soft edges.", + "confidence": 0.93, + "canonical_issue_reason": "If one issue must stand in for the cluster, issue:43873 is the most central and active: it is open, has the highest discussion/references, and is a broad transformer runtime bug rather than a narrowly model-specific report.", "canonical_pr_reason": null, - "best_issue_reason": "Issue 44843 is the strongest standalone representative: it has a specific trigger, a well-defined failure path, and a broader regression profile than the more model-specific reports.", + "best_issue_reason": "issue:43873 is the best representative issue because it is active, broadly scoped, and has the strongest engagement signal in this set.", "best_pr_reason": null, "soft_edge_verdicts": [ + { + "left": "issue:44869", + "right": "issue:45356", + "accept": false, + "reason": "Both involve tokenization/codec behavior, but one is a Whisper timestamp crash and the other is a Kimi tokenizer regression with fix_mistral_regex; different code paths and symptoms." + }, + { + "left": "issue:44038", + "right": "issue:44623", + "accept": false, + "reason": "Qwen3-VL-Moe loading bug vs processor.save_pretrained missing files; same area, but not the same underlying defect." + }, { "left": "issue:43232", "right": "issue:43673", "accept": false, - "reason": "Both involve generation internals, but one is a sync_gpus/update_model_kwargs bug and the other is a chunked_prefill cache regression." + "reason": "Both are generation/cache-related, but one concerns _update_model_kwargs_for_generation after sync_gpus and the other cache absence in chunked_prefill; distinct bugs." }, { "left": "issue:45310", "right": "issue:45375", "accept": false, - "reason": "Same Qwen3.5/MoE family, but one is a from_pretrained loading failure and the other is a missing config field dropped by strict; related context, not the same concrete bug." + "reason": "Both mention Qwen3.5 MoE, but one is a from_pretrained failure and the other is a missing config field in a vision config; not the same issue." }, { "left": "issue:41628", "right": "issue:44933", "accept": false, - "reason": "Both are import errors, but they hit different symbols and different code paths." - }, - { - "left": "issue:43873", - "right": "issue:44945", - "accept": false, - "reason": "Quantized offloading behavior and pipeline-parallel output corruption are distinct failure modes." + "reason": "Both are import errors, but they concern different symbols and modules; no evidence of the same missing export." }, { "left": "issue:36010", "right": "issue:41628", "accept": false, - "reason": "Different missing imports in different APIs; same symptom class, not the same bug." + "reason": "GenerationMixin import failure and AutoImageProcessor import failure are separate API breakages." }, { "left": "issue:44661", "right": "issue:47059", "accept": false, - "reason": "Model-registration/tokenizer-mapping logic and a CLI compatibility break are unrelated." + "reason": "Model-registration logic in add-new-model-like is unrelated to a CLI completion regression from huggingface_hub." + }, + { + "left": "issue:43873", + "right": "issue:44945", + "accept": false, + "reason": "Offloading/quantization behavior and incorrect pipeline-parallel outputs are different distributed inference problems." }, { "left": "issue:43278", "right": "issue:45137", "accept": false, - "reason": "BF16/FP32 dtype drift and a DeepSpeed ZeRO3 deque IndexError are unrelated." + "reason": "Embedding dtype drift in eval is unrelated to a DeepSpeed ZeRO3 deque IndexError." }, { "left": "issue:42371", "right": "issue:45092", "accept": false, - "reason": "Packaging/backend selection and remote-code checkpoint meta-init incompatibility are separate problems." + "reason": "TF32 API guidance and InternVL2 meta-init compatibility are unrelated." }, { "left": "issue:45081", "right": "issue:45103", "accept": false, - "reason": "Both are AttributeErrors, but in different subsystems with different triggers and fixes." + "reason": "Both crash with AttributeError, but one is Mistral tokenizer patching and the other is auto_docstring kwargs parsing; different causes." }, { "left": "issue:44857", "right": "issue:46877", "accept": false, - "reason": "CUDA AMP crash and import-time docstring logging are unrelated." + "reason": "Training loss AMP crash and docstring import-time logging are unrelated." }, { "left": "issue:45406", "right": "issue:46202", "accept": false, - "reason": "Gemma-related, but one is a serve-side processor attribute error and the other is a tokenizer roundtrip failure." + "reason": "Both involve Gemma, but one is a processor attribute crash in transformers serve and the other is a tokenizer roundtrip failure." }, { "left": "issue:46202", "right": "issue:46701", "accept": false, - "reason": "Both involve Gemma pipelines, but they fail in different ways and likely need different fixes." + "reason": "Gemma pipeline roundtrip failure and DiffusionGemma image-text-to-text pipeline failure are different concrete bugs." }, { "left": "issue:45593", "right": "issue:46877", "accept": false, - "reason": "D-FINE auxiliary-loss behavior and docstring stdout noise are unrelated." + "reason": "D-FINE auxiliary-loss behavior is unrelated to auto_docstring stdout logging." }, { "left": "issue:44038", "right": "issue:45310", "accept": false, - "reason": "Both mention Qwen/MoE, but they target different models and symptoms." - }, - { - "left": "issue:43232", - "right": "issue:44792", - "accept": false, - "reason": "Generation cache handling and a Janus image-generation test failure are different issues." + "reason": "Related Qwen model family, but different models and failure modes; insufficient evidence they are the same bug." }, { "left": "issue:41628", "right": "issue:44351", "accept": false, - "reason": "Both are import regressions, but for different classes and API surfaces." - }, - { - "left": "issue:42175", - "right": "issue:44933", - "accept": false, - "reason": "TensorFlow extras packaging and a missing image_utils import are unrelated." + "reason": "Both are import-name issues, but they target different classes and likely different missing exports." }, { "left": "issue:44222", "right": "issue:45310", "accept": false, - "reason": "FP8 save_pretrained MoE serialization and Qwen3.5 from_pretrained loading are different bug classes." + "reason": "FP8 save_pretrained behavior is distinct from a from_pretrained model-loading error." }, { - "left": "issue:44945", - "right": "issue:45753", + "left": "issue:30333", + "right": "issue:33290", "accept": false, - "reason": "Pipeline-parallel output corruption and a Qwen3_5 image infinite loop are not the same underlying bug." + "reason": "MLflow reporting on failed training jobs is unrelated to DeepSpeed OOM with Adafactor." }, { - "left": "issue:30333", - "right": "issue:33290", + "left": "issue:44945", + "right": "issue:45753", "accept": false, - "reason": "MLflow reporting after failure and an Adafactor DeepSpeed OOM are unrelated training issues." + "reason": "Pipeline-parallel output corruption is unrelated to a Qwen3_5 image infinite loop." }, { "left": "issue:45405", "right": "issue:46877", "accept": false, - "reason": "An unreleased PEFT version bump and docstring diagnostics to stdout are unrelated." + "reason": "PEFT version bump and docstring import-time diagnostics are unrelated." }, { "left": "issue:45561", "right": "issue:47059", "accept": false, - "reason": "A pytest-xdist file race and a huggingface_hub CLI compatibility break are different failures." + "reason": "Pytest-xdist file race and CLI completion breakage are different testing/packaging issues." }, { "left": "issue:44743", "right": "issue:47068", "accept": false, - "reason": "Qwen3_5 recurrent-state reset and a host-synchronization concern in loss computation are unrelated." + "reason": "Qwen3_5 recurrent-state reset and DeepSpeed loss synchronization are unrelated runtime bugs." }, { "left": "issue:42898", "right": "issue:44568", "accept": false, - "reason": "Both are tokenizer behavior changes in v5, but they affect different APIs and symptoms." + "reason": "Both are tokenizer regressions, but one is clean-up spacing and the other is add_special_tokens/BOS/EOS behavior; not the same defect." }, { "left": "issue:42831", "right": "issue:44222", "accept": false, - "reason": "FineGrainedFP8 accuracy issues and FP8 save_pretrained MoE serialization are different problems." - }, - { - "left": "issue:42175", - "right": "issue:45712", - "accept": false, - "reason": "TensorFlow packaging and leftover dummy PT objects are unrelated maintenance bugs." + "reason": "Both mention FP8, but one is accuracy loss and the other is save_pretrained behavior for MoE; different problems." }, { "left": "issue:43784", "right": "issue:43824", "accept": false, - "reason": "Different import errors in different APIs; same broad symptom, not the same bug." + "reason": "Sentence-transformers import failure and Qwen2.5-VL import failure are unrelated." }, { - "left": "issue:44496", - "right": "issue:45561", + "left": "issue:42175", + "right": "issue:45712", "accept": false, - "reason": "Model config loading failure and a pytest-xdist file race are unrelated." + "reason": "One is missing TensorFlow in a torch extra install, the other is dummy PT object cleanup; not the same underlying issue." }, { "left": "issue:43976", "right": "issue:44933", "accept": false, - "reason": "Python-version compatibility and a missing image_utils import are unrelated." + "reason": "Python version compatibility and image_utils import cleanup are unrelated." }, { - "left": "issue:44568", - "right": "issue:47085", + "left": "issue:44496", + "right": "issue:45561", "accept": false, - "reason": "Tokenizer special-token behavior and tokenizer thread-safety are separate issues." + "reason": "Model config recognition and xdist race conditions are unrelated." }, { - "left": "issue:44843", - "right": "issue:47085", + "left": "issue:30990", + "right": "issue:35141", "accept": false, - "reason": "Offline Mistral-regex model_info calls and tokenizer thread-safety are different code paths." + "reason": "Sentence-Transformers loading hang and embedding reinitialization after resize are different bugs." } ] }, - "evaluator_result": { - "accept": false, - "feedback": "The overall clustering summary is reasonable, but one soft-edge verdict is too strong for the evidence in the packet: issue:45310 and issue:45375 share a Qwen3.5/MoE theme, yet the packet only shows a model-loading error versus a missing config field silently dropped by strict. That is suggestive, not enough to confidently treat them as the same bug or a mergeable single PR. Keep that pair conservative/uncertain rather than accepted." - }, - "evaluator_used": true, - "retried": true, + "evaluator_result": null, + "evaluator_used": false, + "retried": false, "accepted_nontrivial_soft_edge": false, "error_kind": null, "error_message": null @@ -15253,7 +15149,7 @@ "source": "llm", "cache_hit": false, "model": "gpt-5.4-mini?service_tier=flex", - "cluster_id": "cluster-10000-529", + "cluster_id": "cluster-10000-530", "nodes": [ "issue:22355", "issue:27957", @@ -15284,6 +15180,7 @@ "issue:44623", "issue:44734", "issue:44779", + "issue:44843", "issue:44869", "issue:45092", "issue:45103", @@ -15295,7 +15192,6 @@ "issue:45812", "issue:46077", "issue:46123", - "issue:46132", "issue:46231", "issue:46693", "issue:46829", @@ -15304,25 +15200,27 @@ "issue:47085" ], "soft_pairs": [ - "issue:42673|issue:43704", + "issue:44568|issue:47085", + "issue:44843|issue:47085", "issue:42898|issue:45812", + "issue:42673|issue:43704", "issue:44186|issue:47085", "issue:43479|issue:46829", "issue:42673|issue:46693", "issue:44623|issue:45310", - "issue:43723|issue:44568", "issue:39692|issue:46909", + "issue:43723|issue:44568", "issue:44060|issue:45125", "issue:45600|issue:47059", "issue:30333|issue:34689", "issue:44016|issue:46231", "issue:27957|issue:30333", "issue:43479|issue:44079", - "issue:42898|issue:44779", "issue:30333|issue:36331", - "issue:43479|issue:44869", - "issue:42898|issue:43295", + "issue:42898|issue:44779", "issue:42673|issue:47085", + "issue:42898|issue:43295", + "issue:43479|issue:44869", "issue:42898|issue:44448", "issue:46077|issue:46123", "issue:22355|issue:28282", @@ -15332,223 +15230,228 @@ "issue:43874|issue:45092", "issue:42796|issue:43723", "issue:43116|issue:44190", - "issue:44448|issue:46132", "issue:44075|issue:46231" ], - "prepared_review_unit_hash": "e79d36aafa69ce09a2c91828abd333fbf675f9d960187c7bb0c6fc84f56022ca", + "prepared_review_unit_hash": "92efe59ad41336342aba24b36c48777d641ea9a9b29b97d9ad49c722c9e1847d", "component_budget": { - "node_count": 529, - "item_count": 529, - "soft_pair_count": 1953, - "serialized_chars": 1012074, - "estimated_input_tokens": 253019, - "estimated_eval_tokens": 506294 + "node_count": 530, + "item_count": 530, + "soft_pair_count": 1954, + "serialized_chars": 1013004, + "estimated_input_tokens": 253251, + "estimated_eval_tokens": 506758 }, "budget": { "node_count": 47, "item_count": 47, - "soft_pair_count": 30, - "serialized_chars": 36270, - "estimated_input_tokens": 9068, - "estimated_eval_tokens": 18392 + "soft_pair_count": 31, + "serialized_chars": 36613, + "estimated_input_tokens": 9154, + "estimated_eval_tokens": 18564 }, "overflow_policy": "truncate_then_skip", "trimmed": false, "aggressively_trimmed": false, "split": true, "analyst_result": { - "summary": "Only one soft pair looks like a true duplicate: the two Qwen3ForCausalLM VRAM leak reports. The rest are thematically similar at best, but describe different models, code paths, or regression types.", - "confidence": 0.91, - "canonical_issue_reason": "issue:42673 is the best canonical issue because it is the earlier, higher-activity report and is an exact duplicate of issue:43704 by title and bug description.", + "summary": "Only one soft edge is a true duplicate: the two Qwen3ForCausalLM VRAM-leak reports. The rest are same-subsystem or same-version regressions at a broad level, but they describe different concrete bugs and should stay separate.", + "confidence": 0.96, + "canonical_issue_reason": "issue:42673 is the strongest canonical candidate: it is the earliest, most detailed report of the exact Qwen3ForCausalLM multi-dataloader VRAM leak, and issue:43704 is essentially the same report.", "canonical_pr_reason": null, - "best_issue_reason": "issue:42673 is the strongest issue overall for canonicalization: it is concrete, well-scoped, and has the only clear duplicate counterpart in the set.", + "best_issue_reason": "issue:42673 is the best global representative because it captures the concrete bug most completely and is duplicated by issue:43704; it is the cleanest anchor for triage.", "best_pr_reason": null, "soft_edge_verdicts": [ { - "left": "issue:42673", - "right": "issue:43704", - "accept": true, - "reason": "Same underlying bug: identical Qwen3ForCausalLM VRAM leak report with the same title and description." + "left": "issue:44568", + "right": "issue:47085", + "accept": false, + "reason": "Both mention tokenizer behavior, but one is a special-token regression for mDeBERTa and the other is a thread-safety regression in fast tokenizers; different bugs." + }, + { + "left": "issue:44843", + "right": "issue:47085", + "accept": false, + "reason": "Offline HF hub patching in AutoTokenizer is unrelated to tokenizer thread safety." }, { "left": "issue:42898", "right": "issue:45812", "accept": false, - "reason": "Both are tokenizer/v5 regressions, but they affect different behaviors and different model families; not the same bug." + "reason": "Both are v5 tokenizer regressions, but one is clean_up_tokenization_spaces behavior and the other is Granite token ID generation; not the same bug." + }, + { + "left": "issue:42673", + "right": "issue:43704", + "accept": true, + "reason": "Same title and same concrete bug: Qwen3ForCausalLM VRAM leak when used in multiple dataloader threads." }, { "left": "issue:44186", "right": "issue:47085", "accept": false, - "reason": "Different problems: LayoutLMv2 tokenizer crash vs. fast tokenizers thread-safety regression." + "reason": "LayoutLMv2Tokenizer crash on NER/padding is unrelated to tokenizer thread-safety." }, { "left": "issue:43479", "right": "issue:46829", "accept": false, - "reason": "Different model/output bugs in different subsystems; no shared concrete failure mode." + "reason": "Different model families and different failure modes: config initialization vs output field misassignment." }, { "left": "issue:42673", "right": "issue:46693", "accept": false, - "reason": "Both mention performance/memory, but one is a VRAM leak in dataloader threads and the other is an FA2 long-context slowdown." + "reason": "One is a VRAM leak in multithreaded dataloaders; the other is a FA2 long-context performance regression." }, { "left": "issue:44623", "right": "issue:45310", "accept": false, - "reason": "Processor save-pretrained missing files is unrelated to Qwen3.5 MoE from_pretrained failure." + "reason": "Missing processor save files and Qwen3.5 Moe from_pretrained failure are different code paths." }, { - "left": "issue:43723", - "right": "issue:44568", + "left": "issue:39692", + "right": "issue:46909", "accept": false, - "reason": "Tokenizer loading failure and add_special_tokens/BOS-EOS regression are distinct issues." + "reason": "Both are documentation-example bugs, but they involve different models/examples and different error classes." }, { - "left": "issue:39692", - "right": "issue:46909", + "left": "issue:43723", + "right": "issue:44568", "accept": false, - "reason": "Both are documentation-example problems, but they are different examples with different errors." + "reason": "Both are tokenizer v5 issues, but one is AutoTokenizer loading and the other is add_special_tokens behavior for a specific model." }, { "left": "issue:44060", "right": "issue:45125", "accept": false, - "reason": "Both involve Qwen models, but the bugs differ: tied-weights warning vs missing tensor-parallel plan." + "reason": "Different Qwen3-Next/Qwen3.5-MoE model internals: tied-weights warning vs missing tensor-parallel plan." }, { "left": "issue:45600", "right": "issue:47059", "accept": false, - "reason": "CI breakage from removed config references is unrelated to a CLI test failure in huggingface_hub." + "reason": "Broken auto-mapping for removed configs and a CLI completion API break are unrelated." }, { "left": "issue:30333", "right": "issue:34689", "accept": false, - "reason": "Training-job MLflow reporting bug is unrelated to Llama 3.2 Vision model loading regression." + "reason": "MLflow reporting on failed training is unrelated to model loading regression for Llama 3.2 Vision Instruct." }, { "left": "issue:44016", "right": "issue:46231", "accept": false, - "reason": "Notebook syntax error and training-doc parameter-name mismatch are unrelated documentation issues." + "reason": "Notebook syntax error and a docs parameter-name mismatch are different documentation issues." }, { "left": "issue:27957", "right": "issue:30333", "accept": false, - "reason": "XLM-RoBERTa Flash Attention 2 question is unrelated to MLflow reporting on failed training jobs." + "reason": "Flash Attention 2 behavior for XLM-R is unrelated to MLflow reporting after job failure." }, { "left": "issue:43479", "right": "issue:44079", "accept": false, - "reason": "Multimodal config initialization bug is unrelated to ModelOutput key assignment behavior." + "reason": "Config default handling and ModelOutput key assignment are different bugs." }, { - "left": "issue:42898", - "right": "issue:44779", + "left": "issue:30333", + "right": "issue:36331", "accept": false, - "reason": "Both are v5 tokenizer regressions, but they concern different models and different tokenization failures." + "reason": "Training failure reporting vs CustomTrainer compute_loss signature mismatch are different trainer issues." }, { - "left": "issue:30333", - "right": "issue:36331", + "left": "issue:42898", + "right": "issue:44779", "accept": false, - "reason": "MLflow failure reporting and CustomTrainer compute_loss signature breakage are separate training issues." + "reason": "Both are tokenizer regressions in v5, but they affect different models and different tokenization behaviors." }, { - "left": "issue:43479", - "right": "issue:44869", + "left": "issue:42673", + "right": "issue:47085", "accept": false, - "reason": "Phi4 multimodal config bug is unrelated to Whisper word-timestamp decode crashing on replacement characters." + "reason": "The titles are identical in theme, but without concrete shared symptoms or code path evidence, treat as separate reports rather than a confirmed duplicate." }, { "left": "issue:42898", "right": "issue:43295", "accept": false, - "reason": "Both are regressions around tokenizer/processor behavior, but they describe different API breakages and models." + "reason": "Different regressions: clean_up_tokenization_spaces vs processor.tokenizer/image-passing compatibility." }, { - "left": "issue:42673", - "right": "issue:47085", + "left": "issue:43479", + "right": "issue:44869", "accept": false, - "reason": "VRAM leak in Qwen3ForCausalLM is not the same as tokenizers no longer being thread-safe." + "reason": "Phi4 multimodal config initialization and Whisper timestamp decoding are unrelated." }, { "left": "issue:42898", "right": "issue:44448", "accept": false, - "reason": "Different tokenizer/regression reports affecting different models and output behavior." + "reason": "Both are v4-to-v5 output regressions, but on different models and behaviors; not one bug." }, { "left": "issue:46077", "right": "issue:46123", "accept": false, - "reason": "Gemma attention-mask bug and pipeline batching bug are unrelated." + "reason": "Gemma attention-mask construction and last-partial-batch handling are unrelated pipeline/model bugs." }, { "left": "issue:22355", "right": "issue:28282", "accept": false, - "reason": "Import error for transformers.onnx is unrelated to missing PyTorch for AutoModel." + "reason": "Importing transformers.onnx and missing PyTorch for AutoModel are different dependency/import failures." }, { "left": "issue:44295", "right": "issue:45656", "accept": false, - "reason": "Position_ids buffer access error is unrelated to deepspeed calling optimizer.step twice." + "reason": "Position_ids buffer access and duplicate optimizer stepping with deepspeed are unrelated." }, { "left": "issue:42222", "right": "issue:45200", "accept": false, - "reason": "VitPose model breakage and Gemma 4 mm_token_type_ids defaulting are different multimodal issues." + "reason": "Both involve multimodal models, but one is broken vitpose loading and the other is Gemma 4 text-only mm_token_type_ids behavior." }, { "left": "issue:44734", "right": "issue:45103", "accept": false, - "reason": "Server KV-cache continuation crash and auto-docstring AttributeError are unrelated." + "reason": "KV-cache continuation crash in serve and auto_docstring annotation handling are unrelated." }, { "left": "issue:43874", "right": "issue:45092", "accept": false, - "reason": "GLM46V image-patch counting bug is unrelated to InternVL2 meta-initialization incompatibility." + "reason": "Missing image patch count API and remote-code meta-init incompatibility are different multimodal issues." }, { "left": "issue:42796", "right": "issue:43723", "accept": false, - "reason": "Tokenizer BPE-marker handling and AutoTokenizer loading failure are different tokenizer bugs." + "reason": "Tokenizer BPE-marker handling and AutoTokenizer loading failure are not the same defect." }, { "left": "issue:43116", "right": "issue:44190", "accept": false, - "reason": "Classification example returns empty results, while the other issue is about loading a local dataset in image classification." - }, - { - "left": "issue:44448", - "right": "issue:46132", - "accept": false, - "reason": "Pegasus v4/v5 output regression is unrelated to AttentionInterface.register changing callback behavior." + "reason": "Two example-script bugs, but in different scripts and with different failure modes." }, { "left": "issue:44075", "right": "issue:46231", "accept": false, - "reason": "SGD optimizer args not being used is unrelated to the set_training documentation mismatch." + "reason": "Optimizer argument handling is unrelated to a documentation parameter-name typo." } ] }, "evaluator_result": { "accept": true, - "feedback": "The canonicalization choice is grounded in the packet: issue:42673 and issue:43704 have identical titles, and the earlier report has much higher discussion activity. The remaining soft-edge rejections stay conservative and are framed as related-but-distinct issues rather than duplicates. I don\u2019t see evidence of overstating any pair beyond what the packet supports." + "feedback": "The summary is grounded in the packet, and the one accepted soft edge (issue:42673 \u2194 issue:43704) is supported by identical titles and closely aligned metadata. The remaining soft-edge rejections are conservative and do not overclaim duplication from broad subsystem similarity alone. No obvious overstatement or mergeability problem." }, "evaluator_used": true, "retried": false, @@ -15567,7 +15470,7 @@ "source": "llm", "cache_hit": false, "model": "gpt-5.4-mini?service_tier=flex", - "cluster_id": "cluster-10000-529", + "cluster_id": "cluster-10000-530", "nodes": [ "issue:30990", "issue:38175", @@ -15590,6 +15493,7 @@ "issue:44327", "issue:44360", "issue:44373", + "issue:44448", "issue:44484", "issue:44485", "issue:44623", @@ -15609,7 +15513,6 @@ "issue:46209", "issue:46305", "issue:46369", - "issue:46424", "issue:46431", "issue:46614", "issue:46762", @@ -15620,244 +15523,244 @@ ], "soft_pairs": [ "issue:45405|issue:45712", - "issue:46209|issue:46431", "issue:44222|issue:44297", + "issue:46209|issue:46431", + "issue:44448|issue:46132", "issue:44297|issue:45656", - "issue:43479|issue:43644", "issue:42831|issue:44623", - "issue:44933|issue:45712", - "issue:44222|issue:44623", + "issue:43479|issue:43644", "issue:45245|issue:46132", + "issue:44222|issue:44623", + "issue:44933|issue:45712", + "issue:44373|issue:44485", "issue:43723|issue:44977", "issue:44297|issue:45310", - "issue:44373|issue:44485", "issue:43122|issue:46209", + "issue:41628|issue:42757", "issue:43232|issue:47276", "issue:43504|issue:47059", - "issue:41628|issue:42757", "issue:46614|issue:46932", "issue:45071|issue:46082", "issue:43475|issue:46305", "issue:43232|issue:45834", "issue:30990|issue:38175", "issue:44222|issue:44484", - "issue:44190|issue:46082", "issue:44295|issue:44623", "issue:44360|issue:46762", "issue:44327|issue:47059", - "issue:43232|issue:46424", "issue:44263|issue:45092", "issue:46369|issue:46909", + "issue:44190|issue:46082", "issue:42907|issue:46050" ], - "prepared_review_unit_hash": "2add35222d32dbddccb6a296383b2a2c4be8dda4eb8c737d79634089fa1448cb", + "prepared_review_unit_hash": "8326779039b4f6a68776d62814f456abbc21c10597d7d681514fae6a8213de2a", "component_budget": { - "node_count": 529, - "item_count": 529, - "soft_pair_count": 1953, - "serialized_chars": 1012074, - "estimated_input_tokens": 253019, - "estimated_eval_tokens": 506294 + "node_count": 530, + "item_count": 530, + "soft_pair_count": 1954, + "serialized_chars": 1013004, + "estimated_input_tokens": 253251, + "estimated_eval_tokens": 506758 }, "budget": { "node_count": 48, "item_count": 48, "soft_pair_count": 30, - "serialized_chars": 36279, - "estimated_input_tokens": 9070, - "estimated_eval_tokens": 18396 + "serialized_chars": 36297, + "estimated_input_tokens": 9075, + "estimated_eval_tokens": 18406 }, "overflow_policy": "truncate_then_skip", "trimmed": false, "aggressively_trimmed": false, "split": true, "analyst_result": { - "summary": "The items are mostly distinct bugs with only superficial overlap around save/load, imports, or v5 migration. I would not merge any of the soft-edge pairs; they describe different code paths and fixes.", - "confidence": 0.82, - "canonical_issue_reason": "issue:46050 is the clearest concrete regression in the set: a specific save/load failure for quantized models with a narrow, actionable fix surface.", + "summary": "This cluster is mostly a grab-bag of unrelated Transformers issues. The soft-similarity pairs share broad keywords like save_pretrained, from_pretrained, config, or generation, but they do not describe the same concrete bug/change, so none should be merged as duplicates.", + "confidence": 0.83, + "canonical_issue_reason": "issue:44222 is the most representative concrete bug in the set, centered on broken FP8 save_pretrained behavior and close to several other serialization-related issues, even though the nearby soft matches are not true duplicates.", "canonical_pr_reason": null, - "best_issue_reason": "issue:46050 is the strongest standalone issue because it is specific, reproducible, and likely to have a focused implementation change.", + "best_issue_reason": "issue:44222 is the best single issue to anchor this cluster because it is specific, reproducible, and sits near the strongest serialization-related similarity edges.", "best_pr_reason": null, "soft_edge_verdicts": [ { "left": "issue:45405", "right": "issue:45712", "accept": false, - "reason": "Dependency pinning/release timing is unrelated to dummy-class cleanup and repo-lint failures." + "reason": "One is a PEFT version bump complaint; the other is about leftover dummy classes leaking into the API. Different problems." + }, + { + "left": "issue:44222", + "right": "issue:44297", + "accept": false, + "reason": "Both mention save_pretrained, but one is FP8 MoE serialization and the other is tokenizer_class metadata mismatch. Not the same bug." }, { "left": "issue:46209", "right": "issue:46431", "accept": false, - "reason": "FP8 inference on B200 and missing chat templates are different failures in different subsystems." + "reason": "FP8 inference failure on B200 and missing chat template are unrelated code paths." }, { - "left": "issue:44222", - "right": "issue:44297", + "left": "issue:44448", + "right": "issue:46132", "accept": false, - "reason": "Both mention save_pretrained, but one is FP8/MoE serialization and the other is tokenizer metadata mismatch." + "reason": "Pegasus output drift and AttentionInterface.register behavior are different model/runtime issues." }, { "left": "issue:44297", "right": "issue:45656", "accept": false, - "reason": "Tokenizer config serialization does not match a deepspeed optimizer-step bug." + "reason": "Tokenizer save metadata mismatch is unrelated to deepspeed causing optimizer.step to run twice." }, { - "left": "issue:43479", - "right": "issue:43644", + "left": "issue:42831", + "right": "issue:44623", "accept": false, - "reason": "Both are v5-related, but one is multimodal config initialization and the other is non-persistent buffer initialization." + "reason": "FP8 accuracy regression and processor.save_pretrained missing files are different serialization/runtime problems." }, { - "left": "issue:42831", - "right": "issue:44623", + "left": "issue:43479", + "right": "issue:43644", "accept": false, - "reason": "FP8 accuracy drift and missing processor files are unrelated bug classes." + "reason": "Config default initialization for multimodal vision/audio and non-persistent buffer junk are not the same defect." }, { - "left": "issue:44933", - "right": "issue:45712", + "left": "issue:45245", + "right": "issue:46132", "accept": false, - "reason": "Import re-exports from image_utils are not the same issue as leftover dummy classes leaking into dir()." + "reason": "Category cardinality limit and attention registration semantics are unrelated." }, { "left": "issue:44222", "right": "issue:44623", "accept": false, - "reason": "Model save_pretrained breakage and processor save_pretrained missing files are different serialization problems." + "reason": "Both involve save_pretrained, but one is FP8 MoE model saving and the other is processor artifact omission. Different failure modes." }, { - "left": "issue:45245", - "right": "issue:46132", + "left": "issue:44933", + "right": "issue:45712", + "accept": false, + "reason": "Missing import from image_utils and leftover dummy classes are separate API cleanup issues." + }, + { + "left": "issue:44373", + "right": "issue:44485", "accept": false, - "reason": "A category-count limit error is unrelated to AttentionInterface.register behavior changes." + "reason": "A docstring mismatch for position_ids is not the same as a RoPE implementation bug." }, { "left": "issue:43723", "right": "issue:44977", "accept": false, - "reason": "Tokenizer loading in v5 and flash-attention generation failures are different code paths." + "reason": "Tokenizer loading failure and flash-attention generation failure are unrelated." }, { "left": "issue:44297", "right": "issue:45310", "accept": false, - "reason": "Tokenizer class metadata mismatch is not the same as Qwen3.5 from_pretrained model loading failure." + "reason": "Tokenizer save metadata and Qwen3.5 MoE from_pretrained error are different problems." }, { - "left": "issue:44373", - "right": "issue:44485", + "left": "issue:43122", + "right": "issue:46209", "accept": false, - "reason": "A docstring issue and a RoPE implementation discussion are not the same underlying bug." + "reason": "Tokenization differences across versions and FP8 inference on B200 do not share the same underlying bug." }, { - "left": "issue:43122", - "right": "issue:46209", + "left": "issue:41628", + "right": "issue:42757", "accept": false, - "reason": "Tokenizer output changes across versions and FP8 inference failures are unrelated." + "reason": "Missing AutoImageProcessor and missing is_offline_mode are separate import/export regressions." }, { "left": "issue:43232", "right": "issue:47276", "accept": false, - "reason": "Generation state update logic and a warning-message wording issue are not the same defect." + "reason": "One is a generation state bug after sync_gpus; the other is only a misleading warning message." }, { "left": "issue:43504", "right": "issue:47059", "accept": false, - "reason": "Pretrained Beit preset loading and CLI completion breakage from hub changes are unrelated." - }, - { - "left": "issue:41628", - "right": "issue:42757", - "accept": false, - "reason": "Different missing imports from different packages, with different symbols and fixes." + "reason": "Legacy field handling in BEiT loading is unrelated to a huggingface_hub CLI incompatibility." }, { "left": "issue:46614", "right": "issue:46932", "accept": false, - "reason": "Stale phonemizer language state and Gemma4 quantized embedder dtype casting are unrelated." + "reason": "Stale phonemizer backend state and Gemma4 quantized dtype casting are unrelated." }, { "left": "issue:45071", "right": "issue:46082", "accept": false, - "reason": "PretrainedConfig type-checking regressions and LlamaConfig head_dim validation are different behaviors." + "reason": "Both concern config validation, but type-checking breakage and head_dim rejection are not the same concrete issue." }, { "left": "issue:43475", "right": "issue:46305", "accept": false, - "reason": "SAM3 missing attribute and Moonshine off-by-one masking are different model bugs." + "reason": "Missing fpn_position_embeddings and a sliding-window off-by-one are different model bugs." }, { "left": "issue:43232", "right": "issue:45834", "accept": false, - "reason": "Generation kwargs handling and local dataset loading are unrelated." + "reason": "Generation kwargs update after sync_gpus and long-ocr index overflow are unrelated." }, { "left": "issue:30990", "right": "issue:38175", "accept": false, - "reason": "A loading hang in Sentence Transformers is not the same as zero-probability outputs for SigLIP2." + "reason": "Sentence Transformers loading hang and zero-probabilities from SigLIP2 are different symptom sets and code paths." }, { "left": "issue:44222", "right": "issue:44484", "accept": false, - "reason": "FP8 save_pretrained failure and shard-size defaults/documentation are not the same issue." - }, - { - "left": "issue:44190", - "right": "issue:46082", - "accept": false, - "reason": "Local dataset loading and LlamaConfig validation are unrelated." + "reason": "Both touch saving, but FP8 MoE serialization and the 50GB shard-size question are not the same bug." }, { "left": "issue:44295", "right": "issue:44623", "accept": false, - "reason": "Reading a registered buffer and processor save_pretrained file omissions are different bugs." + "reason": "Reading a registered buffer and processor save_pretrained missing files are not the same issue." }, { "left": "issue:44360", "right": "issue:46762", "accept": false, - "reason": "A DSA indexer ReLU complaint and numerical differences in Minimax attention are not the same code-path problem." + "reason": "A DSA indexer activation concern and Minimax attention numerical drift are distinct implementation problems." }, { "left": "issue:44327", "right": "issue:47059", "accept": false, - "reason": "QA decode_spans top-k bounds failure is unrelated to the CLI completion API break." - }, - { - "left": "issue:43232", - "right": "issue:46424", - "accept": false, - "reason": "Generation sync behavior and cache max length semantics are different concerns." + "reason": "Top-k decode crash and CLI completion attribute breakage are unrelated." }, { "left": "issue:44263", "right": "issue:45092", "accept": false, - "reason": "torch.split handling in a DSA indexer is unrelated to meta-initialization incompatibility for InternVL2 checkpoints." + "reason": "torch.split indexing in GlmMoeDsaIndexer and InternVL2 meta-init incompatibility are different code paths." }, { "left": "issue:46369", "right": "issue:46909", "accept": false, - "reason": "A memory overflow in model execution is unrelated to a broken documentation example." + "reason": "SAM3 LiteText memory overflow and a broken documentation example do not match." + }, + { + "left": "issue:44190", + "right": "issue:46082", + "accept": false, + "reason": "Local dataset loading in a training script and LlamaConfig head_dim validation are unrelated." }, { "left": "issue:42907", "right": "issue:46050", "accept": false, - "reason": "Both involve save/load flows, but they target different quantization/dequantization scenarios and are not clearly the same underlying bug." + "reason": "Both involve save/load after quantization, but dequantized Ministrals/Devstrals saving and sinq round-trip breakage are different concrete bugs." } ] }, @@ -15879,7 +15782,7 @@ "source": "llm", "cache_hit": false, "model": "gpt-5.4-mini?service_tier=flex", - "cluster_id": "cluster-10000-529", + "cluster_id": "cluster-10000-530", "nodes": [ "issue:22355", "issue:27957", @@ -15888,14 +15791,12 @@ "issue:41628", "issue:42796", "issue:42831", - "issue:42886", - "issue:43066", "issue:43122", + "issue:43232", "issue:43408", "issue:43540", "issue:43576", "issue:43644", - "issue:43646", "issue:43756", "issue:43784", "issue:43824", @@ -15924,6 +15825,7 @@ "issue:46328", "issue:46369", "issue:46402", + "issue:46424", "issue:46655", "issue:46726", "issue:46909", @@ -15932,251 +15834,238 @@ ], "soft_pairs": [ "issue:45092|issue:45357", + "issue:43232|issue:46424", "issue:22355|issue:27957", "issue:43986|issue:46726", "issue:44263|issue:44373", "issue:44279|issue:46328", - "issue:46305|issue:46655", "issue:43408|issue:46402", + "issue:46305|issue:46655", "issue:45381|issue:46305", "issue:42831|issue:45310", "issue:41628|issue:43824", "issue:22355|issue:28218", - "issue:44964|issue:47059", "issue:44483|issue:45406", "issue:44263|issue:44360", + "issue:44964|issue:47059", "issue:42831|issue:44038", "issue:44829|issue:46002", "issue:44261|issue:45092", "issue:43784|issue:45600", - "issue:42796|issue:43122", - "issue:39401|issue:44568", "issue:44188|issue:44373", + "issue:39401|issue:44568", + "issue:42796|issue:43122", "issue:43644|issue:46305", "issue:43576|issue:43784", - "issue:43756|issue:44060", "issue:43986|issue:47217", + "issue:43756|issue:44060", "issue:42831|issue:45753", "issue:43408|issue:46369", "issue:43540|issue:46909", - "issue:45381|issue:46909", - "issue:42886|issue:43066", - "issue:43646|issue:46002" + "issue:45381|issue:46909" ], - "prepared_review_unit_hash": "9a880f8f4421f18d0818f9272dd1d7a3900c4e374bf97f93cb47ed9f07be20b3", + "prepared_review_unit_hash": "f896a617cadf3148fbf65defa3075559eef35a72dab7d65a633f9d2a039fa4c5", "component_budget": { - "node_count": 529, - "item_count": 529, - "soft_pair_count": 1953, - "serialized_chars": 1012074, - "estimated_input_tokens": 253019, - "estimated_eval_tokens": 506294 + "node_count": 530, + "item_count": 530, + "soft_pair_count": 1954, + "serialized_chars": 1013004, + "estimated_input_tokens": 253251, + "estimated_eval_tokens": 506758 }, "budget": { - "node_count": 48, - "item_count": 48, - "soft_pair_count": 31, - "serialized_chars": 36748, - "estimated_input_tokens": 9187, - "estimated_eval_tokens": 18630 + "node_count": 47, + "item_count": 47, + "soft_pair_count": 30, + "serialized_chars": 35929, + "estimated_input_tokens": 8983, + "estimated_eval_tokens": 18222 }, "overflow_policy": "truncate_then_skip", "trimmed": false, "aggressively_trimmed": false, "split": true, "analyst_result": { - "summary": "No candidate pair looks like a safe duplicate merge. The items are mostly distinct bugs/features, with only light thematic overlap in tokenizer, import, and vision-language reports.", - "confidence": 0.79, - "canonical_issue_reason": "issue:43122 is the broadest version-scoped tokenizer regression report and the best representative of the small tokenizer-related subset in this set.", + "summary": "This cluster is heterogeneous: it groups many unrelated Transformers issues across tokenization, vision-language loading, generation, CLI, and model-specific regressions. None of the soft pairs appear to be true duplicates of the same underlying bug or change.", + "confidence": 0.97, + "canonical_issue_reason": null, "canonical_pr_reason": null, - "best_issue_reason": "issue:43122 is the clearest general user-facing regression report here; the other items are narrower or model-specific.", + "best_issue_reason": null, "best_pr_reason": null, "soft_edge_verdicts": [ { "left": "issue:45092", "right": "issue:45357", "accept": false, - "reason": "Different model families and different failure modes: meta-init incompatibility vs incorrect visual encoder keys during save_pretrained." + "reason": "Both concern Qwen-related multimodal models, but one is a v5 meta-init incompatibility and the other is incorrect visual encoder keys on save_pretrained; different code paths and fixes." + }, + { + "left": "issue:43232", + "right": "issue:46424", + "accept": false, + "reason": "One is a generation kwargs sync bug after sync_gpus, the other is cache max length only being honored for quantized cache; same area, different defect." }, { "left": "issue:22355", "right": "issue:27957", "accept": false, - "reason": "Unrelated problems: missing transformers.onnx module vs XLMRoberta Flash Attention 2 support." + "reason": "Unrelated issues: missing transformers.onnx module import vs XLMRoberta Flash Attention 2 support." }, { "left": "issue:43986", "right": "issue:46726", "accept": false, - "reason": "Both involve AutoProcessor and missing vision backends, but the concrete bugs differ (video-model crash vs unrecognized image processor)." + "reason": "Both involve processor/backend handling, but one is a video-model crash without torchvision and the other is an unrecognized image processor when a vision backend is missing." }, { "left": "issue:44263", "right": "issue:44373", "accept": false, - "reason": "Unrelated: GlmMoeDsaIndexer torch.split behavior vs a wrong docstring for position_ids." + "reason": "Different bugs: a torch.split/indexer implementation problem vs a wrong docstring for position_ids." }, { "left": "issue:44279", "right": "issue:46328", "accept": false, - "reason": "No shared bug: generic dependency issue vs a padding performance inefficiency." + "reason": "A generic dependency issue is not the same as a SequenceFeatureExtractor.pad performance regression." }, { - "left": "issue:46305", - "right": "issue:46655", + "left": "issue:43408", + "right": "issue:46402", "accept": false, - "reason": "Different models and different math bugs: streaming mask off-by-one vs video padding calculation." + "reason": "One is a model-type warning during instantiation, the other is missing text-encoder weights in Sam3Model." }, { - "left": "issue:43408", - "right": "issue:46402", + "left": "issue:46305", + "right": "issue:46655", "accept": false, - "reason": "Same model family, but the concrete issues differ: wrong model-type warning vs missing text encoder weights." + "reason": "Both are video-processing bugs, but one is a sliding-window mask off-by-one and the other is incorrect padding calculation." }, { "left": "issue:45381", "right": "issue:46305", "accept": false, - "reason": "Both are video-related, but one is Qwen2.5-VL vision_position_ids and the other is MoonshineStreaming masking." + "reason": "Qwen2.5-VL vision_position_ids bug vs MoonshineStreaming mask bug; different models and logic." }, { "left": "issue:42831", "right": "issue:45310", "accept": false, - "reason": "Different subsystems entirely: FineGrainedFP8 accuracy vs Qwen3.5 MoE from_pretrained failure." + "reason": "FineGrainedFP8 accuracy issue is unrelated to Qwen3.5 MoE from_pretrained loading failure." }, { "left": "issue:41628", "right": "issue:43824", "accept": false, - "reason": "Both are import errors, but they concern different missing symbols and likely different fixes." + "reason": "Both are import errors, but for different symbols and likely different missing exposure paths." }, { "left": "issue:22355", "right": "issue:28218", "accept": false, - "reason": "Missing transformers.onnx is unrelated to tokenizer adding an extra space after an added token." - }, - { - "left": "issue:44964", - "right": "issue:47059", - "accept": false, - "reason": "Different dependency layers and failures: model loading regression vs huggingface_hub CLI test breakage." + "reason": "Different tokenizer/import problems: ONNX module import vs extra space after added token." }, { "left": "issue:44483", "right": "issue:45406", "accept": false, - "reason": "API request rejection in chat/completions is unrelated to Gemma4Processor missing _tokenizer crash." + "reason": "A chat-completions request rejection is not the same as a Gemma4Processor crash in transformers serve." }, { "left": "issue:44263", "right": "issue:44360", "accept": false, - "reason": "Same module area, but different concrete bugs in GlmMoeDsaIndexer; not the same underlying change." + "reason": "Same subsystem, but distinct bugs: split return values in GlmMoeDsaIndexer vs a missing ReLU in the DSA indexer." + }, + { + "left": "issue:44964", + "right": "issue:47059", + "accept": false, + "reason": "Phi-4 multimodal model loading failure is unrelated to the huggingface_hub CLI completion breakage." }, { "left": "issue:42831", "right": "issue:44038", "accept": false, - "reason": "Accuracy regression in FP8 is unrelated to a Qwen3-VL-Moe load bug." + "reason": "An FP8 accuracy regression is not the same as a Qwen3-VL-Moe bug." }, { "left": "issue:44829", "right": "issue:46002", "accept": false, - "reason": "Different training pathologies: flash_attention_3 degeneracy vs NaNs during checkout/init." + "reason": "Flash Attention 3 causes degenerate training, while the other is NaNs after checkout/init; different symptoms and likely causes." }, { "left": "issue:44261", "right": "issue:45092", "accept": false, - "reason": "Both are model-init related, but one is an rms_norm_eps precision mismatch and the other is a remote-code meta-init incompatibility." + "reason": "Layernorm eps precision mismatch and remote-code/meta-init incompatibility are unrelated defects." }, { "left": "issue:43784", "right": "issue:45600", "accept": false, - "reason": "Importing sentence-transformers with latest transformers is unrelated to stale auto_mappings references." + "reason": "Importing sentence-transformers fails with NameError in one case; the other is stale auto_mappings references breaking CI." }, { - "left": "issue:42796", - "right": "issue:43122", + "left": "issue:44188", + "right": "issue:44373", "accept": false, - "reason": "Both concern tokenization changes, but one is a model-specific BPE marker issue and the other is a broader version-wide tokenization change." + "reason": "torch.compile attention-kernel divergence is unrelated to a wrong docstring." }, { "left": "issue:39401", "right": "issue:44568", "accept": false, - "reason": "Different tokenizer bugs: offset_mapping for Qwen3 vs BOS/EOS not added for mdeberta-v3-base." + "reason": "Both are tokenizer regressions, but one is offset_mapping and the other is add_special_tokens not inserting BOS/EOS." }, { - "left": "issue:44188", - "right": "issue:44373", + "left": "issue:42796", + "right": "issue:43122", "accept": false, - "reason": "Torch.compile attention-kernel divergence is unrelated to a docstring typo." + "reason": "The first is a specific BPE-marker handling bug for Ministral3, while the second is a broader version-to-version tokenization regression." }, { "left": "issue:43644", "right": "issue:46305", "accept": false, - "reason": "Non-persistent buffer junk in v5.0 is unrelated to MoonshineStreaming window masking." + "reason": "Non-persistent buffer initialization junk is unrelated to a video sliding-window mask off-by-one." }, { "left": "issue:43576", "right": "issue:43784", "accept": false, - "reason": "CLI/env command breakage is unrelated to the sentence-transformers import NameError." - }, - { - "left": "issue:43756", - "right": "issue:44060", - "accept": false, - "reason": "Different model internals and bugs: RoPE layer dropping vs tied-weights warning across Qwen3-Next." + "reason": "A broken transformers env command and a sentence-transformers import NameError are different failure modes." }, { "left": "issue:43986", "right": "issue:47217", "accept": false, - "reason": "Both are vision-language failures, but one is missing-backend crash handling and the other is an image-placeholder placement bug." - }, - { - "left": "issue:42831", - "right": "issue:45753", - "accept": false, - "reason": "FineGrainedFP8 accuracy issues are unrelated to Qwen3.5 infinite-loop behavior on an image." + "reason": "Both are VL processor/input issues, but one is missing torchvision handling and the other is failure when the image placeholder appears in user text." }, { - "left": "issue:43408", - "right": "issue:46369", + "left": "issue:43756", + "right": "issue:44060", "accept": false, - "reason": "Different sam3 issues: wrong model-type warning vs memory overflow in sam3-litetext." + "reason": "Smollm3 RoPE layer loss and Qwen3-Next tied-weights warning are different model bugs." }, { "left": "issue:43540", "right": "issue:46909", "accept": false, - "reason": "Qwen3OmniMoe video-input ValueError is unrelated to a broken DIA documentation example." + "reason": "A Qwen3OmniMoe video-input ValueError is unrelated to a broken documentation example." }, { "left": "issue:45381", "right": "issue:46909", "accept": false, - "reason": "Vision-position-id bug is unrelated to documentation example breakage." + "reason": "Model video-position-id bug vs documentation bug; no shared underlying change." }, { - "left": "issue:42886", - "right": "issue:43066", - "accept": false, - "reason": "Tokenizer cache/offline loading bug is unrelated to the wrong tokenizer decoder type in v5." - }, - { - "left": "issue:43646", - "right": "issue:46002", + "left": "issue:42831", + "right": "issue:45753", "accept": false, - "reason": "Custom model initialization breakage is unrelated to NaNs in classification heads upon checkout/init." + "reason": "FP8 accuracy regression and Qwen3.5 infinite-loop-on-image are different defects." } ] }, @@ -16198,7 +16087,7 @@ "source": "llm", "cache_hit": false, "model": "gpt-5.4-mini?service_tier=flex", - "cluster_id": "cluster-10000-529", + "cluster_id": "cluster-10000-530", "nodes": [ "issue:22355", "issue:27957", @@ -16213,18 +16102,19 @@ "issue:42886", "issue:42907", "issue:42994", + "issue:43066", "issue:43116", "issue:43122", "issue:43329", "issue:43408", "issue:43452", "issue:43475", + "issue:43646", "issue:43824", "issue:43937", "issue:43976", "issue:44038", "issue:44117", - "issue:44222", "issue:44263", "issue:44295", "issue:44297", @@ -16235,7 +16125,6 @@ "issue:44521", "issue:44625", "issue:44704", - "issue:44779", "issue:44829", "issue:44945", "issue:44964", @@ -16244,27 +16133,29 @@ "issue:45278", "issue:45290", "issue:45478", + "issue:46002", "issue:46172", "issue:46326", - "issue:46762", "issue:47059" ], "soft_pairs": [ - "issue:42757|issue:45278", - "issue:39379|issue:46172", + "issue:42886|issue:43066", "issue:42886|issue:42994", + "issue:43646|issue:46002", + "issue:39379|issue:46172", + "issue:42757|issue:45278", "issue:27957|issue:30990", "issue:30990|issue:33357", "issue:43122|issue:46326", "issue:43329|issue:47059", "issue:44295|issue:44297", "issue:44625|issue:44829", - "issue:42907|issue:44964", "issue:39401|issue:44488", + "issue:42907|issue:44964", "issue:39401|issue:44521", "issue:43976|issue:44945", - "issue:41628|issue:45278", "issue:42907|issue:44945", + "issue:41628|issue:45278", "issue:44117|issue:44704", "issue:22355|issue:29159", "issue:43122|issue:44521", @@ -16278,241 +16169,229 @@ "issue:44038|issue:44295", "issue:45216|issue:45478", "issue:43116|issue:45200", - "issue:44360|issue:44485", - "issue:43122|issue:44779", - "issue:44038|issue:44222", - "issue:46172|issue:46762" + "issue:44360|issue:44485" ], - "prepared_review_unit_hash": "bed95095a269b8060bc75007d53ca3c4569d64cfb7a33c8373fdeee47f9ac28c", + "prepared_review_unit_hash": "5ea6f365e6bebd24a10deff74e9fe8c06755093f56cee80aafe7dfb3ac1f634d", "component_budget": { - "node_count": 529, - "item_count": 529, - "soft_pair_count": 1953, - "serialized_chars": 1012074, - "estimated_input_tokens": 253019, - "estimated_eval_tokens": 506294 + "node_count": 530, + "item_count": 530, + "soft_pair_count": 1954, + "serialized_chars": 1013004, + "estimated_input_tokens": 253251, + "estimated_eval_tokens": 506758 }, "budget": { "node_count": 48, "item_count": 48, - "soft_pair_count": 32, - "serialized_chars": 36867, - "estimated_input_tokens": 9217, - "estimated_eval_tokens": 18690 + "soft_pair_count": 31, + "serialized_chars": 36522, + "estimated_input_tokens": 9131, + "estimated_eval_tokens": 18518 }, "overflow_policy": "truncate_then_skip", "trimmed": false, "aggressively_trimmed": false, "split": true, "analyst_result": { - "summary": "These items are mostly unrelated issue reports. I found one plausible duplicate pair in the tokenizer-regression area; the rest are too broad or touch different code paths to merge.", - "confidence": 0.44, - "canonical_issue_reason": "issue:45278 is the best umbrella-style issue in this set: it describes broad import breakage after a version upgrade, which is more representative than the narrower class-specific reports.", + "summary": "These issues are mostly unrelated, with only superficial overlap around imports, tokenizers, and version regressions. None of the soft-edge pairs look like the same underlying bug or change, so I reject them all. The broadest representative issue is the general upgrade-regression report about many import errors after moving to 5.5.0.", + "confidence": 0.91, + "canonical_issue_reason": "issue:45278 is the broadest umbrella report in the set: it describes widespread import breakage after a version upgrade, which makes it the best representative issue if one needs a single anchor for this mixed cluster.", "canonical_pr_reason": null, - "best_issue_reason": "issue:45278 is the strongest global representative because it is broad, user-facing, and can subsume multiple specific import failures; the other issues are more niche and less suitable as a canonical anchor.", + "best_issue_reason": "Same as canonical: it is the most general, cluster-like report rather than a narrow model- or tokenizer-specific bug.", "best_pr_reason": null, "soft_edge_verdicts": [ { - "left": "issue:42757", - "right": "issue:45278", + "left": "issue:42886", + "right": "issue:43066", + "accept": false, + "reason": "Both involve tokenization/versioning, but one is offline cache loading and the other is wrong decoder type in v5; different failure modes and code paths." + }, + { + "left": "issue:42886", + "right": "issue:42994", + "accept": false, + "reason": "Offline tokenizer cache loading and quantized model saving are unrelated I/O paths." + }, + { + "left": "issue:43646", + "right": "issue:46002", "accept": false, - "reason": "Both mention import problems, but one is a specific missing symbol from huggingface_hub while the other is a broad post-upgrade import breakage report; not clearly the same bug." + "reason": "Both mention init-time problems, but one is a Transformers 5.0 custom-model init break and the other is NaNs in classification heads; not the same bug." }, { "left": "issue:39379", "right": "issue:46172", "accept": false, - "reason": "These are different model-specific bugs in different areas (mllama masking vs deepseek-v4 dtype/transpose behavior)." + "reason": "Different models and different internal logic: mllama mask handling vs deepseek-v4 tensor transposition." }, { - "left": "issue:42886", - "right": "issue:42994", + "left": "issue:42757", + "right": "issue:45278", "accept": false, - "reason": "Offline cache loading and quantized model saving are distinct failure modes and code paths." + "reason": "42978-style broad import regressions are not the same as a specific missing huggingface_hub symbol; related symptom class, different underlying issue." }, { "left": "issue:27957", "right": "issue:30990", "accept": false, - "reason": "Flash Attention 2 support for XLMRoberta and a Sentence Transformers loading hang are unrelated issues." + "reason": "XLMRoberta Flash Attention 2 support and Sentence Transformers hanging on load are unrelated." }, { "left": "issue:30990", "right": "issue:33357", "accept": false, - "reason": "A loading hang and a macOS bus error with a CLIP community model are different runtime failures." + "reason": "Both are runtime failures, but one is a loading hang and the other is a MacOS bus error with a CLIP model." }, { "left": "issue:43122", "right": "issue:46326", "accept": false, - "reason": "Both are tokenizer-related, but one is a version-to-version regression in tokenization and the other is a chat-template continuation bug for mistral_common tokenizers; different underlying problems." + "reason": "Both are tokenizer-related, but one is general version-to-version tokenization drift and the other is a specific continue_final_message bug for mistral_common tokenizers." }, { "left": "issue:43329", "right": "issue:47059", "accept": false, - "reason": "Multimodal token counting and CLI completion breakage are unrelated." + "reason": "Multimodal token counting on video and a CLI completion attribute error in huggingface_hub are unrelated." }, { "left": "issue:44295", "right": "issue:44297", "accept": false, - "reason": "One is about reading a buffered position_ids tensor, the other about tokenizer_class metadata written by save_pretrained; different subsystems." + "reason": "A position_ids buffer read error and a tokenizer_class mismatch on save_pretrained affect different parts of the stack." }, { "left": "issue:44625", "right": "issue:44829", "accept": false, - "reason": "Qwen3.5 num_labels propagation and flash_attention_3 training degeneration are unrelated model behaviors." + "reason": "Qwen3.5 num_labels propagation and flash_attention_3 training collapse are different bugs." }, { - "left": "issue:42907", - "right": "issue:44964", + "left": "issue:39401", + "right": "issue:44488", "accept": false, - "reason": "Saving dequantized Ministral/Devstral models and loading Phi-4 multimodal are not the same bug." + "reason": "Wrong offset_mapping for Qwen3 tokenizer and a load failure for cjvt/sleng-bert are unrelated." }, { - "left": "issue:39401", - "right": "issue:44488", + "left": "issue:42907", + "right": "issue:44964", "accept": false, - "reason": "A Qwen3 tokenizer offset-mapping bug and a separate model loading failure for cjvt/sleng-bert do not share the same root cause." + "reason": "Saving dequantized Ministrals/Devstrals and loading Phi-4 multimodal-instruct are different model and serialization paths." }, { "left": "issue:39401", "right": "issue:44521", "accept": false, - "reason": "Offset-mapping regression and assistant_mask generation for multimodal chat templates are different tokenizer behaviors." + "reason": "Tokenizer offset mapping and multimodal assistant mask generation are distinct tokenizer/template behaviors." }, { "left": "issue:43976", "right": "issue:44945", "accept": false, - "reason": "Python version compatibility and incorrect LLM output under pipeline parallelism are unrelated." + "reason": "Python-version compatibility in 5.1.0 and incorrect pipeline-parallelism output are unrelated." }, { - "left": "issue:41628", - "right": "issue:45278", + "left": "issue:42907", + "right": "issue:44945", "accept": false, - "reason": "Both are import-related, but one is a single AutoImageProcessor import failure while the other is a broad upgrade-wide import regression; too generic to treat as the same bug." + "reason": "Quantized save failure and pipeline parallelism output corruption are different problems." }, { - "left": "issue:42907", - "right": "issue:44945", + "left": "issue:41628", + "right": "issue:45278", "accept": false, - "reason": "Model saving failure and pipeline-parallel inference corruption are different code paths." + "reason": "A missing AutoImageProcessor import is a specific export issue, not the same as broad upgrade-wide import breakage." }, { "left": "issue:44117", "right": "issue:44704", "accept": false, - "reason": "TOKENIZER_MAPPING_NAMES returning None and AutoProcessor not forwarding kwargs to cached_file are different lookup/dispatch issues." + "reason": "Tokenizer mapping returning None and AutoProcessor not forwarding kwargs to cached_file are separate logic bugs." }, { "left": "issue:22355", "right": "issue:29159", "accept": false, - "reason": "Missing transformers.onnx module and slow-vs-fast tokenizer inconsistency are unrelated." + "reason": "Missing transformers.onnx module import and slow/fast tokenizer inconsistency are unrelated." }, { "left": "issue:43122", "right": "issue:44521", "accept": false, - "reason": "Both involve tokenization, but one is a version regression in tokenizer output and the other is multimodal assistant mask handling in apply_chat_template; not the same bug." + "reason": "General tokenization drift and multimodal assistant-mask zeros are different behaviors with different causes." }, { "left": "issue:44704", "right": "issue:45290", "accept": false, - "reason": "AutoProcessor kwargs forwarding and tool-call chat-template crashes are separate failures." + "reason": "AutoProcessor argument forwarding and chat-template crashes on tool-call messages are distinct code paths." }, { "left": "issue:42371", "right": "issue:44263", "accept": false, - "reason": "TF32 API deprecation guidance and a DSA indexer ReLU discussion are not the same issue." + "reason": "TF32 API warning and a torch.split return-value bug in GlmMoeDsaIndexer are unrelated." }, { "left": "issue:43408", "right": "issue:43475", "accept": false, - "reason": "The warning about sam3_video vs sam3_tracker is different from the missing fpn_position_embeddings attribute error." + "reason": "SAM3 tracker/model-type warning and a missing fpn_position_embeddings attribute are different SAM3 issues." }, { "left": "issue:42907", "right": "issue:43937", "accept": false, - "reason": "Quantized/dequantized model saving and GLM-5 GenerationConfig validation are unrelated." + "reason": "Model saving failure and GLM-5 generation config validation are unrelated." }, { "left": "issue:42907", "right": "issue:43452", "accept": false, - "reason": "Model saving failure and gguf_file handling for from_pretrained are different bugs." + "reason": "Dequantized save failure and gguf_file load issues affect different serialization flows." }, { "left": "issue:44038", "right": "issue:44297", "accept": false, - "reason": "Qwen3-VL-Moe model issues and tokenizer_class metadata mismatch are different concerns." + "reason": "Qwen3-VL-Moe compatibility and tokenizer_class mismatch on save are not the same bug." }, { "left": "issue:43824", "right": "issue:44351", "accept": false, - "reason": "A missing Qwen2.5-VL class import and a missing HybridCache import are both import errors but refer to different symbols and likely different fixes." + "reason": "Two missing imports, but for different symbols and different model subsystems; not one underlying defect." }, { "left": "issue:44038", "right": "issue:44295", "accept": false, - "reason": "Qwen3-VL-Moe behavior and position_ids buffer access are not the same code-path problem." + "reason": "Qwen3-VL-Moe issue and position_ids buffer error are unrelated." }, { "left": "issue:45216", "right": "issue:45478", "accept": false, - "reason": "Both are Qwen3.5 regressions, but one is about save_pretrained checkpoint correctness and the other is a from_pretrained error; they do not clearly describe the same bug." + "reason": "Both concern Qwen3.5 and version 5.4+, but one is save_pretrained checkpoint correctness and the other is from_pretrained failure; strict duplicate criteria not met." }, { "left": "issue:43116", "right": "issue:45200", "accept": false, - "reason": "Empty multi-label classification outputs and Gemma 4 mm_token_type_ids defaults are different example/model issues." + "reason": "Example-script multi-label classification returning empty results and Gemma 4 mm_token_type_ids defaults are unrelated." }, { "left": "issue:44360", "right": "issue:44485", "accept": false, - "reason": "A DSA indexer ReLU discussion and GLM-5 RoPE implementation are unrelated." - }, - { - "left": "issue:43122", - "right": "issue:44779", - "accept": true, - "reason": "Both describe a tokenizer regression where the same tokenizer behaves differently in v5 than in v4, and the Deepseek report looks like a concrete instance of the broader tokenization-change bug." - }, - { - "left": "issue:44038", - "right": "issue:44222", - "accept": false, - "reason": "Qwen3-VL-Moe and FP8 save_pretrained moe are separate model/runtime concerns." - }, - { - "left": "issue:46172", - "right": "issue:46762", - "accept": false, - "reason": "Deepseek-v4 comb transpose behavior and Minimax MSA numerical differences are distinct numerical/implementation issues." + "reason": "A missing ReLU in the DSA indexer and GLM-5 RoPE implementation are different algorithmic components." } ] }, - "evaluator_result": { - "accept": true, - "feedback": "The assessment is grounded in the packet and stays conservative overall. The single accepted tokenizer-regression pair is plausible from the titles, and the rejections avoid overstating weaker overlaps such as broad import issues or neighboring model/tokenizer bugs. Summary is appropriately cautious." - }, - "evaluator_used": true, + "evaluator_result": null, + "evaluator_used": false, "retried": false, - "accepted_nontrivial_soft_edge": true, + "accepted_nontrivial_soft_edge": false, "error_kind": null, "error_message": null }, @@ -16527,7 +16406,7 @@ "source": "llm", "cache_hit": false, "model": "gpt-5.4-mini?service_tier=flex", - "cluster_id": "cluster-10000-529", + "cluster_id": "cluster-10000-530", "nodes": [ "issue:22355", "issue:28348", @@ -16541,21 +16420,20 @@ "issue:42673", "issue:42757", "issue:42898", - "issue:43066", "issue:43116", + "issue:43122", "issue:43232", "issue:43278", "issue:43493", "issue:43611", "issue:43976", - "issue:44075", + "issue:44038", "issue:44188", + "issue:44222", "issue:44351", "issue:44483", - "issue:44610", "issue:44625", "issue:44779", - "issue:44936", "issue:44938", "issue:44945", "issue:44995", @@ -16567,17 +16445,21 @@ "issue:45478", "issue:45753", "issue:46002", + "issue:46172", "issue:46231", "issue:46328", "issue:46614", "issue:46640", "issue:46697", + "issue:46762", "issue:46877", - "issue:46909", "issue:47217", "issue:47312" ], "soft_pairs": [ + "issue:43122|issue:44779", + "issue:44038|issue:44222", + "issue:46172|issue:46762", "issue:22355|issue:29127", "issue:43278|issue:44945", "issue:42175|issue:44938", @@ -16586,216 +16468,209 @@ "issue:39379|issue:43232", "issue:42757|issue:44351", "issue:36296|issue:43116", - "issue:44995|issue:46697", "issue:43493|issue:45753", + "issue:44995|issue:46697", "issue:44188|issue:46328", "issue:28348|issue:46640", "issue:43976|issue:45446", "issue:46231|issue:46877", "issue:42673|issue:45072", - "issue:42898|issue:46614", "issue:44483|issue:47217", + "issue:42898|issue:46614", "issue:43611|issue:46002", "issue:45357|issue:45478", "issue:44625|issue:46002", "issue:38617|issue:45278", - "issue:45083|issue:45446", - "issue:44075|issue:44936", - "issue:43066|issue:44779", - "issue:44610|issue:46909", - "issue:43116|issue:43493" + "issue:45083|issue:45446" ], - "prepared_review_unit_hash": "0976b71362575e62bab3077d9021171d416f926f4d3eee719cc5c25018dd2afe", + "prepared_review_unit_hash": "0dc5e25c8951b07e739aa85d9ae8f49d0682b1092bee37209ec25132767c16a7", "component_budget": { - "node_count": 529, - "item_count": 529, - "soft_pair_count": 1953, - "serialized_chars": 1012074, - "estimated_input_tokens": 253019, - "estimated_eval_tokens": 506294 + "node_count": 530, + "item_count": 530, + "soft_pair_count": 1954, + "serialized_chars": 1013004, + "estimated_input_tokens": 253251, + "estimated_eval_tokens": 506758 }, "budget": { "node_count": 47, "item_count": 47, - "soft_pair_count": 26, - "serialized_chars": 34388, - "estimated_input_tokens": 8597, - "estimated_eval_tokens": 17450 + "soft_pair_count": 25, + "serialized_chars": 34107, + "estimated_input_tokens": 8527, + "estimated_eval_tokens": 17310 }, "overflow_policy": "truncate_then_skip", "trimmed": false, "aggressively_trimmed": false, "split": true, "analyst_result": { - "summary": "The items do not form a single duplicate cluster; they span unrelated bugs, regressions, docs, and compatibility issues. All soft-edge candidate pairs are rejected as different underlying problems.", - "confidence": 0.97, - "canonical_issue_reason": null, + "summary": "Mostly unrelated issue titles; I found one plausible duplicate-style pair around v5 import regressions, but the rest are distinct bugs/features across different models and subsystems.", + "confidence": 0.84, + "canonical_issue_reason": "issue:45278 is the best umbrella issue here: it describes the broad upgrade-related import regression, and is the most representative of the narrower missing-symbol import failure in issue:38617.", "canonical_pr_reason": null, - "best_issue_reason": null, + "best_issue_reason": "issue:45278 best captures the shared symptom class among the only close match in the set (post-upgrade import errors).", "best_pr_reason": null, "soft_edge_verdicts": [ + { + "left": "issue:43122", + "right": "issue:44779", + "accept": false, + "reason": "Both are tokenizer regressions, but one is a broad v5 tokenization difference and the other is Deepseek-specific incorrect results; not clearly the same bug." + }, + { + "left": "issue:44038", + "right": "issue:44222", + "accept": false, + "reason": "Different problems: Qwen3-VL-Moe compatibility vs FP8 save_pretrained behavior for MoE." + }, + { + "left": "issue:46172", + "right": "issue:46762", + "accept": false, + "reason": "Both mention numerical issues, but they concern different models and attention implementations." + }, { "left": "issue:22355", "right": "issue:29127", "accept": false, - "reason": "Different failures: missing transformers.onnx import vs LayoutLMv3 box-validation error messaging." + "reason": "Completely unrelated: ONNX import failure vs LayoutLMv3 box-validation error messaging." }, { "left": "issue:43278", "right": "issue:44945", "accept": false, - "reason": "Different training/eval behavior bug vs pipeline-parallel generation correctness issue." + "reason": "Different failure modes: dtype drift between train/eval vs incorrect output under pipeline parallelism." }, { "left": "issue:42175", "right": "issue:44938", "accept": false, - "reason": "Package extra/backend dependency issue vs Python 3.14 import/load failure." + "reason": "Installation/backend dependency issue vs Python 3.14 load failure; no shared concrete code path." }, { "left": "issue:39401", "right": "issue:46614", "accept": false, - "reason": "Different tokenizer paths and bugs: Qwen3 offset mapping vs Wav2Vec2 phonemizer backend state." + "reason": "Tokenizer offset mapping bug vs stale phonemizer language backend; different subsystems and symptoms." }, { "left": "issue:42222", "right": "issue:47312", "accept": false, - "reason": "VitPose model breakage vs kernels 0.16 compatibility; no shared underlying defect." + "reason": "VitPose model breakage is unrelated to kernel compatibility with 0.16." }, { "left": "issue:39379", "right": "issue:43232", "accept": false, - "reason": "Different model-specific logic bugs: mllama mask handling vs generation kwargs after sync_gpus." + "reason": "Different generation-path bugs with no clear shared underlying defect." }, { "left": "issue:42757", "right": "issue:44351", "accept": false, - "reason": "Both are import errors, but for different missing symbols from different packages." + "reason": "Both are import errors, but for different symbols and different packages; not the same bug." }, { "left": "issue:36296", "right": "issue:43116", "accept": false, - "reason": "Tensor-parallel training bug vs example script returning empty multi-label results." + "reason": "Tensor parallel training bug vs multi-label classification example returning empty results; unrelated." }, { - "left": "issue:44995", - "right": "issue:46697", + "left": "issue:43493", + "right": "issue:45753", "accept": false, - "reason": "Unrelated model crash/cache bug vs image processor padding bug." + "reason": "Different models and failures: SigLIP2 implementation mismatch vs Qwen3_5 infinite loop on a specific image." }, { - "left": "issue:43493", - "right": "issue:45753", + "left": "issue:44995", + "right": "issue:46697", "accept": false, - "reason": "Different Qwen3.5-related issues: SigLIP2 implementation mismatch vs image-triggered infinite loop." + "reason": "Stale cache crash in GlmMoeDsa vs image over-padding in Swin2SR; unrelated." }, { "left": "issue:44188", "right": "issue:46328", "accept": false, - "reason": "Attention-kernel divergence under torch.compile vs feature extractor padding performance issue." + "reason": "Attention-kernel divergence under torch.compile is unrelated to SequenceFeatureExtractor padding inefficiency." }, { "left": "issue:28348", "right": "issue:46640", "accept": false, - "reason": "Both mention Flash Attention 2, but one is a GPT2LMHeadModel support request and the other is a distinct runtime issue." + "reason": "A feature request for GPT2 flash-attention support is not the same as a general Flash Attention 2 issue." }, { "left": "issue:43976", "right": "issue:45446", "accept": false, - "reason": "Python version compatibility issue vs incorrect PyTorch version guard in flex_attention." + "reason": "Version-check incompatibility in Transformers vs a PyTorch version check bug for flex_attention; different code paths." }, { "left": "issue:46231", "right": "issue:46877", "accept": false, - "reason": "Docs parameter mismatch vs auto-generated docstring emitting diagnostics at import time." + "reason": "Docs parameter-name mismatch vs import-time docstring/lint output; not the same underlying issue." }, { "left": "issue:42673", "right": "issue:45072", "accept": false, - "reason": "Different model/runtime problems: VRAM leak in dataloader threads vs dtype mismatches in inference." + "reason": "VRAM leak in multi-threaded Qwen3ForCausalLM usage vs dtype mismatches in inference." }, { - "left": "issue:42898", - "right": "issue:46614", + "left": "issue:44483", + "right": "issue:47217", "accept": false, - "reason": "Tokenizer cleanup behavior change vs stale phonemizer language state." + "reason": "Both involve chat/VL behavior, but one is a generic v5.3 request rejection and the other is a placeholder-in-text failure; too different." }, { - "left": "issue:44483", - "right": "issue:47217", + "left": "issue:42898", + "right": "issue:46614", "accept": false, - "reason": "Chat-completions request handling vs vision-language image-placeholder failure; different code paths." + "reason": "Tokenization cleanup behavior change is unrelated to phonemizer backend state retention." }, { "left": "issue:43611", "right": "issue:46002", "accept": false, - "reason": "Base-model-prefix loading regression vs NaNs in classification heads." + "reason": "Model-loading breakage with base_model_prefix is unrelated to NaNs in classification heads." }, { "left": "issue:45357", "right": "issue:45478", "accept": false, - "reason": "Related Qwen3.5 area, but one is save_pretrained key serialization and the other is from_pretrained loading failure." + "reason": "Both are Qwen3.5 serialization/load issues, but one is about incorrect visual encoder keys in save_pretrained and the other is a from_pretrained error; not enough to treat as the same bug." }, { "left": "issue:44625", "right": "issue:46002", "accept": false, - "reason": "Configuration propagation bug vs NaNs during classification head initialization." + "reason": "Configuration propagation in Qwen3.5 has no clear relation to NaNs in classification heads." }, { "left": "issue:38617", "right": "issue:45278", - "accept": false, - "reason": "Specific missing import symbol vs broad import breakage after version upgrade; not the same root cause." + "accept": true, + "reason": "This looks like the same upgrade regression: a specific missing import from configuration_utils under the broader 'many import errors after update' report." }, { "left": "issue:45083", "right": "issue:45446", "accept": false, - "reason": "Feature-extractor helper behavior vs PyTorch version check/import gating." - }, - { - "left": "issue:44075", - "right": "issue:44936", - "accept": false, - "reason": "Optimizer argument handling vs trainer.evaluate() failing after train()." - }, - { - "left": "issue:43066", - "right": "issue:44779", - "accept": false, - "reason": "Tokenizer decoder-type regression vs Deepseek tokenizer incorrect results; too different to merge." - }, - { - "left": "issue:44610", - "right": "issue:46909", - "accept": false, - "reason": "OmDet-Turbo processor sizing bug vs broken documentation example." - }, - { - "left": "issue:43116", - "right": "issue:43493", - "accept": false, - "reason": "Example-script output issue vs SigLIP2 implementation discrepancy; unrelated." + "reason": "Unexpected helper output in qwen3_omni_moe is unrelated to a PyTorch version check bug in flex_attention." } ] }, - "evaluator_result": null, - "evaluator_used": false, + "evaluator_result": { + "accept": true, + "feedback": "The summary is grounded in the packet and stays conservative. The one accepted pair (issue:38617 \u2194 issue:45278) is a reasonable umbrella-style match based on post-update import regressions, and the rejected soft pairs are generally well-justified as distinct bug classes rather than overstated duplicates." + }, + "evaluator_used": true, "retried": false, - "accepted_nontrivial_soft_edge": false, + "accepted_nontrivial_soft_edge": true, "error_kind": null, "error_message": null }, @@ -16810,19 +16685,18 @@ "source": "llm", "cache_hit": false, "model": "gpt-5.4-mini?service_tier=flex", - "cluster_id": "cluster-10000-529", + "cluster_id": "cluster-10000-530", "nodes": [ "issue:33453", "issue:39379", "issue:39401", - "issue:42222", "issue:42371", - "issue:42617", "issue:42673", "issue:42796", "issue:43064", "issue:43066", "issue:43116", + "issue:43493", "issue:43618", "issue:43673", "issue:43688", @@ -16832,17 +16706,19 @@ "issue:43874", "issue:43937", "issue:43992", + "issue:44075", "issue:44117", "issue:44291", "issue:44355", - "issue:44380", "issue:44466", "issue:44485", + "issue:44610", "issue:44704", + "issue:44779", + "issue:44936", "issue:45071", "issue:45081", "issue:45092", - "issue:45341", "issue:45362", "issue:45405", "issue:45446", @@ -16850,236 +16726,242 @@ "issue:45563", "issue:45588", "issue:46002", - "issue:46077", "issue:46231", "issue:46328", "issue:46710", "issue:46726", "issue:46762", "issue:46829", + "issue:46909", "issue:46981", - "issue:47000", "issue:47116" ], "soft_pairs": [ + "issue:44075|issue:44936", + "issue:43066|issue:44779", + "issue:44610|issue:46909", + "issue:43116|issue:43493", "issue:43937|issue:44485", "issue:45362|issue:45478", "issue:43066|issue:46710", - "issue:42673|issue:42796", "issue:44355|issue:46231", + "issue:42673|issue:42796", "issue:46726|issue:46981", "issue:43116|issue:43825", - "issue:43618|issue:46829", "issue:42371|issue:43874", + "issue:43618|issue:46829", "issue:43827|issue:45446", "issue:43992|issue:44704", "issue:45405|issue:45446", "issue:46328|issue:47116", "issue:33453|issue:39401", "issue:43824|issue:46981", - "issue:39379|issue:43688", "issue:44291|issue:45071", + "issue:39379|issue:43688", "issue:44117|issue:45588", "issue:43673|issue:46328", "issue:43064|issue:43673", - "issue:43937|issue:45563", "issue:44117|issue:45081", "issue:45092|issue:46762", - "issue:44466|issue:46002", - "issue:44380|issue:46077", - "issue:42222|issue:47000", - "issue:42617|issue:45341" + "issue:43937|issue:45563", + "issue:44466|issue:46002" ], - "prepared_review_unit_hash": "cac240875adcf68367c1063c9b491d61ce332f4502fa65d27b73041067ea8213", + "prepared_review_unit_hash": "0b8a83fbfe49bfb0083cb7b9a5f948e4643b557ad02a3b2334629ca6fe28bceb", "component_budget": { - "node_count": 529, - "item_count": 529, - "soft_pair_count": 1953, - "serialized_chars": 1012074, - "estimated_input_tokens": 253019, - "estimated_eval_tokens": 506294 + "node_count": 530, + "item_count": 530, + "soft_pair_count": 1954, + "serialized_chars": 1013004, + "estimated_input_tokens": 253251, + "estimated_eval_tokens": 506758 }, "budget": { "node_count": 48, "item_count": 48, - "soft_pair_count": 27, - "serialized_chars": 35759, - "estimated_input_tokens": 8940, - "estimated_eval_tokens": 18136 + "soft_pair_count": 28, + "serialized_chars": 36047, + "estimated_input_tokens": 9012, + "estimated_eval_tokens": 18280 }, "overflow_policy": "truncate_then_skip", "trimmed": false, "aggressively_trimmed": false, "split": true, "analyst_result": { - "summary": "This cluster is a loose grab-bag of transformer regressions across tokenizers, generation, multimodal models, training, and docs. I don\u2019t see any pair that looks like the same underlying bug/change, so there is no strong duplicate canonicalization here.", - "confidence": 0.18, - "canonical_issue_reason": "If a representative issue is needed, issue:44117 is the best fit: it is a broad loading-path regression with clear infra impact, unlike the many highly model- or feature-specific reports in this set.", + "summary": "This cluster is mostly heterogeneous. Several items are tokenizer or v5-regression themed, but the soft pairs usually differ in concrete bug, model, or code path, so they should not be deduplicated.", + "confidence": 0.73, + "canonical_issue_reason": "No single canonical issue stands out: the items span tokenizer regressions, model-specific bugs, training/runtime failures, and docs/API issues rather than one underlying defect.", "canonical_pr_reason": null, - "best_issue_reason": "issue:44117 is the most generally framed and broadly reusable bug report in the set; it captures a core loading assumption failure rather than a one-off model-specific symptom.", + "best_issue_reason": "issue:43066 is the best representative of the tokenizer/v5 regression subset because it is concrete, user-visible, and closest in theme to the other tokenizer-related complaints.", "best_pr_reason": null, "soft_edge_verdicts": [ + { + "left": "issue:44075", + "right": "issue:44936", + "accept": false, + "reason": "Different failures: unused SGD args vs evaluate() breaking after train()." + }, + { + "left": "issue:43066", + "right": "issue:44779", + "accept": false, + "reason": "Both are tokenizer regressions, but they describe different concrete problems and code paths." + }, + { + "left": "issue:44610", + "right": "issue:46909", + "accept": false, + "reason": "Model input-size bug vs a broken documentation example; not the same issue." + }, + { + "left": "issue:43116", + "right": "issue:43493", + "accept": false, + "reason": "Multi-label example output bug vs SigLIP2 implementation discrepancy." + }, { "left": "issue:43937", "right": "issue:44485", "accept": false, - "reason": "Different failures: invalid GenerationConfig validation vs GLM-5 RoPE implementation." + "reason": "Invalid GenerationConfig vs RoPE implementation discussion; same model family but different bugs." }, { "left": "issue:45362", "right": "issue:45478", "accept": false, - "reason": "Both involve Qwen3.5, but the code paths differ: chat crash vs from_pretrained error." + "reason": "Different Qwen3.5 failures: chat crash vs from_pretrained error." }, { "left": "issue:43066", "right": "issue:46710", "accept": false, - "reason": "Wrong tokenizer decoder type is not the same bug as an accuracy regression after upgrade." + "reason": "Tokenizer decoder-type regression is not the same as a model accuracy regression after upgrade." }, { - "left": "issue:42673", - "right": "issue:42796", + "left": "issue:44355", + "right": "issue:46231", "accept": false, - "reason": "Unrelated subsystems: VRAM leak in multi-threaded loading vs tokenizer BPE marker handling." + "reason": "Compiled Python file errors vs a documentation parameter-name mismatch." }, { - "left": "issue:44355", - "right": "issue:46231", + "left": "issue:42673", + "right": "issue:42796", "accept": false, - "reason": "Compiled Python file errors and a docs parameter-name typo are different issues." + "reason": "VRAM leak in dataloader threads vs tokenizer BPE-marker handling." }, { "left": "issue:46726", "right": "issue:46981", "accept": false, - "reason": "Both are AutoProcessor issues, but one is missing-backend error handling and the other is HTTP 429 retry behavior." + "reason": "Different AutoProcessor failure modes: missing vision backend vs repeated 429 retries." }, { "left": "issue:43116", "right": "issue:43825", "accept": false, - "reason": "Example-script multi-label output bug is distinct from pipeline() error messaging/docs." + "reason": "Example script multi-label output bug vs pipeline error-message wording." }, { - "left": "issue:43618", - "right": "issue:46829", + "left": "issue:42371", + "right": "issue:43874", "accept": false, - "reason": "Both are output-field mishandlings, but they affect different model outputs and code paths." + "reason": "TF32 configuration guidance vs a missing image-processor method." }, { - "left": "issue:42371", - "right": "issue:43874", + "left": "issue:43618", + "right": "issue:46829", "accept": false, - "reason": "TF32 API guidance is unrelated to the GLM46V image-processor attribute error." + "reason": "Both touch model outputs, but they are different attributes and different models." }, { "left": "issue:43827", "right": "issue:45446", "accept": false, - "reason": "Docs referencing pipeline() and a PyTorch version check bug are unrelated." + "reason": "Docs referencing removed pipeline APIs vs a PyTorch version check bug." }, { "left": "issue:43992", "right": "issue:44704", "accept": false, - "reason": "Both are from_pretrained-related, but one misses encoder weights while the other drops kwargs to cached_file." + "reason": "Missing encoder weights vs AutoProcessor not forwarding kwargs." }, { "left": "issue:45405", "right": "issue:45446", "accept": false, - "reason": "A PEFT version pin issue is unrelated to a flex_attention import/version check bug." + "reason": "Dependency-version bump issue vs an unrelated import/version check bug." }, { "left": "issue:46328", "right": "issue:47116", "accept": false, - "reason": "Feature-extractor padding overhead is not the same as a streaming pipeline regression." + "reason": "Pad performance issue vs streaming pipeline regression." }, { "left": "issue:33453", "right": "issue:39401", "accept": false, - "reason": "Tokenizer loading regression is too broad to equate with a specific offset_mapping bug." + "reason": "Tokenizer loading regression vs wrong offset mapping; related area but not the same bug." }, { "left": "issue:43824", "right": "issue:46981", "accept": false, - "reason": "Missing import for a model class is unrelated to AutoProcessor retry behavior." + "reason": "Import error for a specific model class vs HTTP retry behavior in AutoProcessor." }, { - "left": "issue:39379", - "right": "issue:43688", + "left": "issue:44291", + "right": "issue:45071", "accept": false, - "reason": "These are different model-specific numerical/normalization bugs." + "reason": "Unexpected meta-init argument vs PretrainedConfig type-checking breakage." }, { - "left": "issue:44291", - "right": "issue:45071", + "left": "issue:39379", + "right": "issue:43688", "accept": false, - "reason": "Both are v5 init/config related, but one is an unexpected init_empty_weights arg and the other is a PretrainedConfig type-checking regression." + "reason": "Mllama mask handling vs auxiliary-loss normalization; unrelated model bugs." }, { "left": "issue:44117", "right": "issue:45588", "accept": false, - "reason": "Tokenizer mapping None-handling is unrelated to flash_attention's s_aux=None crash." + "reason": "Tokenizer mapping assumption vs flash-attention aux-sentinel crash." }, { "left": "issue:43673", "right": "issue:46328", "accept": false, - "reason": "Generation cache behavior and SequenceFeatureExtractor padding performance are different issues." + "reason": "Generation cache bug vs SequenceFeatureExtractor padding performance." }, { "left": "issue:43064", "right": "issue:43673", "accept": false, - "reason": "Training optimizer-state corruption is not the same bug as missing generation cache in chunked prefill." - }, - { - "left": "issue:43937", - "right": "issue:45563", - "accept": false, - "reason": "Invalid generation config and stale warning text are different generation-path problems." + "reason": "Distributed optimizer-state bug vs generation cache missing; different subsystems." }, { "left": "issue:44117", "right": "issue:45081", "accept": false, - "reason": "Both involve tokenizer loading assumptions, but they are separate failures in different helpers." + "reason": "Mapping-name None handling vs Mistral regex patch backend_tokenizer crash." }, { "left": "issue:45092", "right": "issue:46762", "accept": false, - "reason": "Meta-init incompatibility for remote-code checkpoints is unrelated to attention numerical differences." - }, - { - "left": "issue:44466", - "right": "issue:46002", - "accept": false, - "reason": "Serialization inconsistency and NaNs in classification heads are different bugs." - }, - { - "left": "issue:44380", - "right": "issue:46077", - "accept": false, - "reason": "Different attention behaviors on different models/backends; not one concrete bug." + "reason": "Meta-initialization incompatibility vs numerical attention differences." }, { - "left": "issue:42222", - "right": "issue:47000", + "left": "issue:43937", + "right": "issue:45563", "accept": false, - "reason": "Vitpose model breakage and RTDetrLoss infeasibility are unrelated." + "reason": "GenerationConfig validation error vs a stale warning in paged generate()." }, { - "left": "issue:42617", - "right": "issue:45341", + "left": "issue:44466", + "right": "issue:46002", "accept": false, - "reason": "A runtime script failure and a testing_utils typo are not the same underlying issue." + "reason": "Serialization inconsistency vs NaNs in classification heads." } ] }, @@ -17101,14 +16983,14 @@ "source": "llm", "cache_hit": false, "model": "gpt-5.4-mini?service_tier=flex", - "cluster_id": "cluster-10000-529", + "cluster_id": "cluster-10000-530", "nodes": [ "issue:38617", "issue:39379", "issue:39401", "issue:41628", "issue:42222", - "issue:42371", + "issue:42617", "issue:43122", "issue:43208", "issue:43408", @@ -17118,10 +17000,9 @@ "issue:43742", "issue:43937", "issue:43986", - "issue:44075", "issue:44220", - "issue:44261", "issue:44263", + "issue:44380", "issue:44410", "issue:44483", "issue:44496", @@ -17139,7 +17020,7 @@ "issue:45464", "issue:45561", "issue:45632", - "issue:45656", + "issue:46077", "issue:46231", "issue:46298", "issue:46328", @@ -17149,19 +17030,23 @@ "issue:46894", "issue:46897", "issue:46981", + "issue:47000", "issue:47059", "issue:47110" ], "soft_pairs": [ - "issue:39379|issue:46897", + "issue:44380|issue:46077", + "issue:42222|issue:47000", + "issue:42617|issue:45341", "issue:43937|issue:44977", - "issue:42222|issue:43441", + "issue:39379|issue:46897", "issue:43408|issue:43742", - "issue:43208|issue:46894", "issue:43450|issue:43986", - "issue:45071|issue:46512", + "issue:42222|issue:43441", "issue:42222|issue:46752", "issue:44263|issue:46762", + "issue:45071|issue:46512", + "issue:43208|issue:46894", "issue:39401|issue:44779", "issue:44829|issue:45216", "issue:44410|issue:45310", @@ -17170,207 +17055,211 @@ "issue:45632|issue:46981", "issue:45341|issue:45561", "issue:45341|issue:46328", - "issue:39401|issue:43421", "issue:44704|issue:47110", + "issue:39401|issue:43421", "issue:38617|issue:41628", "issue:43122|issue:46298", "issue:43937|issue:44496", "issue:38617|issue:44220", "issue:44483|issue:45464", - "issue:44964|issue:45278", - "issue:42371|issue:44261", - "issue:44075|issue:45656" + "issue:44964|issue:45278" ], - "prepared_review_unit_hash": "ff052057ad2903124c73731881f81f83cda2213957a8fb332993a28726dbf6f7", + "prepared_review_unit_hash": "700157a6fd4ad7086806122b9213041a6a9187dea7a5ce39c379b908becb9326", "component_budget": { - "node_count": 529, - "item_count": 529, - "soft_pair_count": 1953, - "serialized_chars": 1012074, - "estimated_input_tokens": 253019, - "estimated_eval_tokens": 506294 + "node_count": 530, + "item_count": 530, + "soft_pair_count": 1954, + "serialized_chars": 1013004, + "estimated_input_tokens": 253251, + "estimated_eval_tokens": 506758 }, "budget": { "node_count": 48, "item_count": 48, - "soft_pair_count": 27, - "serialized_chars": 35511, - "estimated_input_tokens": 8878, - "estimated_eval_tokens": 18012 + "soft_pair_count": 28, + "serialized_chars": 35830, + "estimated_input_tokens": 8958, + "estimated_eval_tokens": 18172 }, "overflow_policy": "truncate_then_skip", "trimmed": false, "aggressively_trimmed": false, "split": true, "analyst_result": { - "summary": "This set is a grab bag of unrelated Transformers issues. None of the soft pairs appears to be the same concrete bug or change, so all are rejected.", - "confidence": 0.97, - "canonical_issue_reason": null, + "summary": "This is a loose cluster of mostly unrelated Transformers bugs and regressions. The soft links are usually just superficial similarity (same subsystem, same model family, or both being \u201cbugs\u201d) rather than the same underlying defect.", + "confidence": 0.93, + "canonical_issue_reason": "issue:45278 is the broadest representative in the set (\u201cmany import errors after update from 4.57.0 to 5.5.0\u201d) and is the closest thing to a cluster umbrella; most others are narrow model- or feature-specific bugs.", "canonical_pr_reason": null, - "best_issue_reason": null, + "best_issue_reason": "issue:45278 is the most general issue here and the only one that plausibly groups a broader class of upgrade/import regressions, unlike the other highly specific reports.", "best_pr_reason": null, "soft_edge_verdicts": [ { - "left": "issue:39379", - "right": "issue:46897", + "left": "issue:44380", + "right": "issue:46077", + "accept": false, + "reason": "Both concern attention behavior, but one is GPT-2 backend scaling and the other is Gemma4 bidirectional masking; different models and different code paths." + }, + { + "left": "issue:42222", + "right": "issue:47000", + "accept": false, + "reason": "Completely different problems: vitpose model loading/breakage vs RTDetrLoss infeasible cost matrix during training." + }, + { + "left": "issue:42617", + "right": "issue:45341", "accept": false, - "reason": "Different bugs in different multimodal paths: mllama mask handling vs Florence2 training-loss double-shift." + "reason": "One is a runtime failure for 3d_parallel.py; the other is a small testing_utils bug. No shared underlying defect." }, { "left": "issue:43937", "right": "issue:44977", "accept": false, - "reason": "Same broad generation area, but different models and failures: invalid GenerationConfig vs flash-attention generation behavior." + "reason": "Both are generation-related, but one is a malformed GenerationConfig on GLM-5 and the other is Qwen3.5 flash-attention generation behavior." }, { - "left": "issue:42222", - "right": "issue:43441", + "left": "issue:39379", + "right": "issue:46897", "accept": false, - "reason": "Unrelated model-specific regressions: vitpose breakage vs Ministral-3 FlashAttention failure." + "reason": "Different models and different bugs: incorrect mllama masking vs Florence2 training-loss double-shift." }, { "left": "issue:43408", "right": "issue:43742", "accept": false, - "reason": "Different issues: SAM3 model-type warning vs MobileLLM loading key error." - }, - { - "left": "issue:43208", - "right": "issue:46894", - "accept": false, - "reason": "Both are training-related, but they are distinct bugs in different models and code paths." + "reason": "SAM3 tracker/model-type warning is unrelated to a MobileLLM loading KeyError." }, { "left": "issue:43450", "right": "issue:43986", "accept": false, - "reason": "Video processor shape bug is not the same as the torchvision-missing crash through AutoProcessor." + "reason": "Both touch video processing, but one is batched output shape and the other is a missing torchvision dependency crash during AutoProcessor loading." }, { - "left": "issue:45071", - "right": "issue:46512", + "left": "issue:42222", + "right": "issue:43441", "accept": false, - "reason": "Config type-checking regression and stale dtype after from_config are different problems." + "reason": "Unrelated model families and failure modes: vitpose breakage vs Ministral-3 FlashAttention failure." }, { "left": "issue:42222", "right": "issue:46752", "accept": false, - "reason": "Vitpose model breakage is unrelated to apply_chat_template crashing on an empty list." + "reason": "Vitpose model breakage and empty-list apply_chat_template crash are unrelated." }, { "left": "issue:44263", "right": "issue:46762", "accept": false, - "reason": "Torch.split return handling and Minimax attention numerical differences are different bugs." + "reason": "Different functionality and symptoms: torch.split return handling in a MoE indexer vs Minimax MSA numerical differences." + }, + { + "left": "issue:45071", + "right": "issue:46512", + "accept": false, + "reason": "Both mention config/dtype, but one is a type-checking regression in PretrainedConfig and the other is stale config.dtype after from_config." + }, + { + "left": "issue:43208", + "right": "issue:46894", + "accept": false, + "reason": "xLSTM training bugs for small models are unrelated to CohereASR's training-loss double-shift issue." }, { "left": "issue:39401", "right": "issue:44779", "accept": false, - "reason": "Both are tokenizer regressions, but on different tokenizers and different symptoms; not the same concrete defect." + "reason": "Both are tokenizer regressions, but they affect different models and likely different tokenization logic." }, { "left": "issue:44829", "right": "issue:45216", "accept": false, - "reason": "Flash-attention training degeneration and save_pretrained checkpoint regression are different code paths." + "reason": "One is a flash-attention training degeneration bug; the other is an incorrect saved checkpoint from save_pretrained." }, { "left": "issue:44410", "right": "issue:45310", "accept": false, - "reason": "Missing projections in qwen3next layer 0 is unrelated to Qwen3.5 from_pretrained failure." + "reason": "Both involve Qwen variants, but one is missing layer projections and the other is a from_pretrained error in Qwen3.5 MoE." }, { "left": "issue:44877", "right": "issue:46231", "accept": false, - "reason": "Strict config loading and a documentation typo are not the same bug." + "reason": "Strict granite_speech config loading and a docs parameter-name mismatch are not the same bug." }, { "left": "issue:45071", "right": "issue:47059", "accept": false, - "reason": "PretrainedConfig type checking and CLI/huggingface_hub completion breakage are unrelated." + "reason": "A PretrainedConfig type-checking regression is unrelated to a CLI completion attribute error from huggingface_hub." }, { "left": "issue:45632", "right": "issue:46981", "accept": false, - "reason": "Cache-path collisions and repeated 429 retries are different AutoProcessor issues." + "reason": "Cache-path collisions for local trust_remote_code models are unrelated to repeated 429 retries in AutoProcessor.from_pretrained." }, { "left": "issue:45341", "right": "issue:45561", "accept": false, - "reason": "Both touch testing utils, but one is a utility bug and the other is a captured-info race under xdist." + "reason": "Both are testing_utils-related, but one is only described as a small bug while the other is a specific xdist race on captured_info.txt; insufficient evidence they are the same defect." }, { "left": "issue:45341", "right": "issue:46328", "accept": false, - "reason": "Testing-utils bug versus SequenceFeatureExtractor.pad performance issue are unrelated." + "reason": "A testing_utils bug and a SequenceFeatureExtractor.pad performance issue are unrelated." }, { - "left": "issue:39401", - "right": "issue:43421", + "left": "issue:44704", + "right": "issue:47110", "accept": false, - "reason": "Tokenizer offset_mapping regression and runtime special-token post-processor sync are different tokenizer bugs." + "reason": "Passing kwargs to cached_file in AutoProcessor is unrelated to special_tokens being a list without keys()." }, { - "left": "issue:44704", - "right": "issue:47110", + "left": "issue:39401", + "right": "issue:43421", "accept": false, - "reason": "AutoProcessor kwargs passthrough and special_tokens list-vs-dict error are different failure modes." + "reason": "Tokenizer offset_mapping correctness and runtime post-processor updates are different tokenizer bugs." }, { "left": "issue:38617", "right": "issue:41628", "accept": false, - "reason": "Both are import errors, but for different symbols and modules." + "reason": "Both are import errors, but they concern different missing symbols and likely different packaging/export issues." }, { "left": "issue:43122", "right": "issue:46298", "accept": false, - "reason": "Tokenization regression and noisy logging are unrelated." + "reason": "Tokenizer behavior drift across versions is unrelated to a noisy logging message." }, { "left": "issue:43937", "right": "issue:44496", "accept": false, - "reason": "Different problems: invalid GenerationConfig for GLM-5 vs missing model_type for an unrecognized model." + "reason": "GenerationConfig invalid vs unrecognized model/config.json model_type are different loading failures." }, { "left": "issue:38617", "right": "issue:44220", "accept": false, - "reason": "ImportError for layer_type_validation is unrelated to _torch_extract_fbank_features() behavior." + "reason": "ImportError for layer_type_validation and an audio feature extraction issue are unrelated." }, { "left": "issue:44483", "right": "issue:45464", "accept": false, - "reason": "Generic chat/completions acceptance failure and Qwen3.5 streaming inference failure are not clearly the same bug." + "reason": "Both hit chat/completions, but one is a general request rejection and the other is a streaming inference failure for a specific Qwen3.5 model." }, { "left": "issue:44964", "right": "issue:45278", "accept": false, - "reason": "A specific Phi-4 multimodal load regression is not the same as broad import errors after upgrade." - }, - { - "left": "issue:42371", - "right": "issue:44261", - "accept": false, - "reason": "TF32 API guidance and MLA rms_norm_eps precision bug are unrelated." - }, - { - "left": "issue:44075", - "right": "issue:45656", - "accept": false, - "reason": "SGD args ignored and optimizer step called twice under deepspeed are different optimizer bugs." + "reason": "A specific Phi-4 multimodal model-load failure is not the same defect as a broad set of import errors after upgrading to 5.5.0." } ] }, @@ -17392,7 +17281,7 @@ "source": "llm", "cache_hit": false, "model": "gpt-5.4-mini?service_tier=flex", - "cluster_id": "cluster-10000-529", + "cluster_id": "cluster-10000-530", "nodes": [ "issue:27957", "issue:41628", @@ -17415,6 +17304,7 @@ "issue:43957", "issue:43986", "issue:44038", + "issue:44075", "issue:44077", "issue:44117", "issue:44230", @@ -17433,10 +17323,9 @@ "issue:45341", "issue:45478", "issue:45588", - "issue:45753", + "issue:45656", "issue:46050", "issue:46172", - "issue:46369", "issue:46752", "issue:46762", "issue:46897", @@ -17444,6 +17333,8 @@ "issue:47116" ], "soft_pairs": [ + "issue:42371|issue:44261", + "issue:44075|issue:45656", "issue:43097|issue:44493", "issue:42371|issue:42831", "issue:44261|issue:46762", @@ -17454,9 +17345,9 @@ "issue:43931|issue:44410", "issue:44263|issue:44485", "issue:43673|issue:45341", - "issue:43208|issue:46897", "issue:41628|issue:45071", "issue:43097|issue:43452", + "issue:43208|issue:46897", "issue:43931|issue:44038", "issue:43441|issue:44829", "issue:43012|issue:43577", @@ -17474,211 +17365,222 @@ "issue:44038|issue:44230", "issue:44261|issue:44485", "issue:43986|issue:45341", - "issue:43097|issue:46050", - "issue:45753|issue:46369" + "issue:43097|issue:46050" ], - "prepared_review_unit_hash": "6593d82c54d7524813c6034872181712cf1dad9f7636b55073cd9dbdc9226aa6", + "prepared_review_unit_hash": "9a9c91a1c299c408e7fedbf6a14c970d0c4d1e1d8c658ffdf8a140aeba3e2eb8", "component_budget": { - "node_count": 529, - "item_count": 529, - "soft_pair_count": 1953, - "serialized_chars": 1012074, - "estimated_input_tokens": 253019, - "estimated_eval_tokens": 506294 + "node_count": 530, + "item_count": 530, + "soft_pair_count": 1954, + "serialized_chars": 1013004, + "estimated_input_tokens": 253251, + "estimated_eval_tokens": 506758 }, "budget": { "node_count": 48, "item_count": 48, - "soft_pair_count": 32, - "serialized_chars": 36902, - "estimated_input_tokens": 9226, - "estimated_eval_tokens": 18708 + "soft_pair_count": 33, + "serialized_chars": 37182, + "estimated_input_tokens": 9296, + "estimated_eval_tokens": 18848 }, "overflow_policy": "truncate_then_skip", "trimmed": false, "aggressively_trimmed": false, "split": true, "analyst_result": { - "summary": "This cluster is mostly a grab-bag of Transformers v5 regressions and model-specific bugs rather than near-duplicates; none of the soft pairs look like the same concrete underlying issue.", - "confidence": 0.34, - "canonical_issue_reason": "issue:43097 is the most general, clearly user-facing v5 regression in the set and has downstream references, so it best represents the broader breaking-change cluster.", + "summary": "The cluster is broadly heterogeneous: most items are separate Transformers regressions across precision, tokenizer, flash-attention, loading, and optimizer behavior. None of the provided soft edges look like the same underlying bug/change, so I would not merge any of them as duplicates. The clearest standalone issue in the set is the Qwen3.5 MoE from_pretrained failure (45310), which also has a near-identical later follow-up (45478).", + "confidence": 0.87, + "canonical_issue_reason": "issue:45310 is the clearest concrete bug report in the set and appears to have a near-identical follow-up in issue:45478; it is specific, reproducible, and earlier than the duplicate-looking follow-up.", "canonical_pr_reason": null, - "best_issue_reason": "issue:43097 is the strongest standalone representative of the cluster because it captures a concrete v5 compatibility break with broad impact, unlike the more narrow model-specific reports.", + "best_issue_reason": "issue:45310 is the best representative issue: it describes a specific model-loading failure with enough detail to anchor a fix, and it is the strongest apparent duplicate target among the listed issues.", "best_pr_reason": null, "soft_edge_verdicts": [ + { + "left": "issue:42371", + "right": "issue:44261", + "accept": false, + "reason": "Both are precision-related, but one is TF32 API guidance and the other is a missing rms_norm_eps causing MLA accuracy drift; different code paths and fixes." + }, + { + "left": "issue:44075", + "right": "issue:45656", + "accept": false, + "reason": "Different optimizer bugs: SGD args ignored vs deepspeed causing the optimizer step to run twice." + }, { "left": "issue:43097", "right": "issue:44493", "accept": false, - "reason": "Both are v5-related, but one is about a removed config flag and the other about unexpected position-id keys; different failure modes and code paths." + "reason": "One is about tie_embeddings_and_encoder_decoder removal in v5; the other is unexpected position-id keys. Both are v5 regressions but not the same bug." }, { "left": "issue:42371", "right": "issue:42831", "accept": false, - "reason": "TF32 setting guidance and FineGrainedFP8 accuracy are both numerical topics, but they are distinct issues with different symptoms and fixes." + "reason": "TF32 control API behavior and FineGrainedFP8 accuracy issues are distinct precision topics with unrelated failure modes." }, { "left": "issue:44261", "right": "issue:46762", "accept": false, - "reason": "Both mention numerical differences, but one is a missing rms_norm_eps in MLA q_a_layernorm and the other is a Minimax attention reference mismatch; not the same bug." + "reason": "Different model/math issues: MLA q_a_layernorm epsilon handling vs Minimax MSA attention numerical differences." }, { "left": "issue:44484", "right": "issue:45310", "accept": false, - "reason": "save_pretrained shard-size behavior is unrelated to a Qwen3.5 MoE from_pretrained load error." + "reason": "save_pretrained shard-size behavior is unrelated to a Qwen3.5 MoE from_pretrained loading error." }, { "left": "issue:43066", "right": "issue:43122", "accept": false, - "reason": "Both involve tokenizer behavior in v5, but one is an import/type mismatch and the other is an actual tokenization-output change; not a single concrete defect." + "reason": "Both involve tokenizer behavior in v5, but one is wrong decoder type and the other is changed tokenization output; not clearly the same defect." }, { "left": "issue:44077", "right": "issue:46981", "accept": false, - "reason": "PatchTSMixer post_init validation and AutoProcessor retry behavior are unrelated." + "reason": "Optional post_init validation in a model config is unrelated to AutoProcessor retry behavior on 429s." }, { "left": "issue:44410", "right": "issue:45478", "accept": false, - "reason": "Different Qwen3-next vs Qwen3.5 MoE model-loading failures; no shared underlying code-path is evident." + "reason": "Different Qwen-family loading problems: missing attn projections in qwen3next vs Qwen3.5 MoE from_pretrained failure." }, { "left": "issue:43931", "right": "issue:44410", "accept": false, - "reason": "Qwen3-VL shape mismatch and qwen3next missing projections are separate architecture-specific issues." + "reason": "Weight-shape mismatch for Qwen3-VL-30B-A3B-Instruct is not the same as missing layer-0 projections in qwen3next." }, { "left": "issue:44263", "right": "issue:44485", "accept": false, - "reason": "torch.split return handling in the DSA indexer is unrelated to the GLM-5 RoPE implementation." + "reason": "torch.split return values in a GLM MoE indexer is unrelated to the GLM-5 RoPE implementation question." }, { "left": "issue:43673", "right": "issue:45341", "accept": false, - "reason": "Generation cache missing during chunked prefill and a testing_utils bug are unrelated." - }, - { - "left": "issue:43208", - "right": "issue:46897", - "accept": false, - "reason": "xLSTM training blockers and Florence2 double-shift loss bug are different model-specific training issues." + "reason": "Generation cache missing during chunked_prefill and a small testing_utils bug are unrelated." }, { "left": "issue:41628", "right": "issue:45071", "accept": false, - "reason": "AutoImageProcessor import failure and PretrainedConfig type-checking regression do not share the same underlying bug." + "reason": "Missing AutoImageProcessor import and PretrainedConfig type-checking regression affect different APIs and code paths." }, { "left": "issue:43097", "right": "issue:43452", "accept": false, - "reason": "A removed encoder-decoder tying flag and gguf_file loading errors are separate from_pretrained/config issues." + "reason": "A removed embedding-tie config and gguf_file loading failures are different from_pretrained/config issues, not one bug." + }, + { + "left": "issue:43208", + "right": "issue:46897", + "accept": false, + "reason": "xLSTM training blockers and Florence2 double-shift loss bug are separate model-specific training problems." }, { "left": "issue:43931", "right": "issue:44038", "accept": false, - "reason": "Both are Qwen3-VL-related, but one is a weight-shape mismatch and the other is MoE support failure; different problems." + "reason": "Both are Qwen3-VL-related, but one is a weight mismatch and the other is a transformers v5 + Qwen3-VL-Moe issue; too different to merge." }, { "left": "issue:43441", "right": "issue:44829", "accept": false, - "reason": "FlashAttention support failure for Ministral-3 and degenerate training with flash_attention_3 are different code paths and symptoms." + "reason": "FlashAttention failures in Ministral-3 and degenerate training with flash_attention_3 are different implementations and symptoms." }, { "left": "issue:43012", "right": "issue:43577", "accept": false, - "reason": "A PyTorch precision warning during compile is not the same as BLIP2 submodules staying float32 after loading." + "reason": "A PyTorch warning when compiling bfloat16 models is not the same as Blip2 submodules staying float32 on load." }, { "left": "issue:46752", "right": "issue:47116", "accept": false, - "reason": "Empty chat-template input crashing and a streaming pipeline regression are unrelated." + "reason": "Empty-list chat template crash and streaming pipeline regression are unrelated API paths." }, { "left": "issue:43874", "right": "issue:44261", "accept": false, - "reason": "Missing image-patch helper in an image processor is unrelated to MLA layernorm precision differences." + "reason": "Image patch counting in GLM46V and MLA epsilon precision drift are different multimodal/model-numerics bugs." }, { "left": "issue:27957", "right": "issue:43352", "accept": false, - "reason": "Both concern Flash Attention 2, but one is XLMRoberta behavior and the other is explicit unsupported-model handling for NemotronH; not the same bug." + "reason": "Both mention Flash Attention 2, but they are different models with separate support issues." }, { "left": "issue:44117", "right": "issue:44987", "accept": false, - "reason": "TOKENIZER_MAPPING_NAMES returning None and a specific physical-intelligence/fast load failure are related only at a very high level." + "reason": "Tokenizer mapping None handling and loading physical-intelligence/fast are not the same concrete failure." }, { "left": "issue:44484", "right": "issue:45478", "accept": false, - "reason": "Shard-size defaults in save_pretrained are unrelated to the Qwen3.5 MoE from_pretrained regression." + "reason": "Shard-size defaults in save_pretrained are unrelated to the Qwen3.5 MoE loading error." }, { "left": "issue:44263", "right": "issue:46172", "accept": false, - "reason": "Different GLM/DeepSeek internals and different failure modes; no shared concrete bug." + "reason": "GLM MoE DSA indexer split semantics and deepseek-v4 dtype/transpose code are distinct bugs." }, { "left": "issue:43531", "right": "issue:45341", "accept": false, - "reason": "Qwen3-MoE sliding_window behavior and a testing_utils issue are unrelated." + "reason": "Qwen3-MoE sliding_window behavior and a testing_utils bug do not share the same code path." }, { "left": "issue:43957", "right": "issue:44351", "accept": false, - "reason": "Meta-device loading breakage and a missing HybridCache import are separate API/runtime issues." + "reason": "Meta-device loading breakage in some models and cannot import HybridCache are unrelated import/loading regressions." }, { "left": "issue:44360", "right": "issue:46172", "accept": false, - "reason": "Missing ReLU in the DSA indexer and a DeepSeek dtype/transpose issue are unrelated." + "reason": "A missing ReLU in the DSA indexer is not the same as the deepseek-v4 comb.to(dtype).transpose(-1, -2) bug." }, { "left": "issue:43097", "right": "issue:45588", "accept": false, - "reason": "A removed config flag and a FlashAttention AttributeError on s_aux=None are different failures." + "reason": "Removed tie_embeddings behavior and flash_attention.py crashing on sink-less models are different regressions." }, { "left": "issue:43582", "right": "issue:45341", "accept": false, - "reason": "Apple Silicon allocator TypeError and a testing utility bug are unrelated." + "reason": "Apple Silicon warmup TypeError and a testing utility bug are unrelated." }, { "left": "issue:44038", "right": "issue:44230", "accept": false, - "reason": "One is a Qwen3-VL-MoE bug in v5, the other is FP8 support work; not the same underlying change." + "reason": "Qwen3-VL-Moe inference support and qwen3-vl-fp8/qwen3.5 moe fp8 inference support are different feature requests/failures." }, { "left": "issue:44261", "right": "issue:44485", "accept": false, - "reason": "MLA q_a_layernorm epsilon handling and GLM-5 RoPE implementation are distinct issues." + "reason": "Missing rms_norm_eps in MLA q_a_layernorm and GLM-5 RoPE implementation are separate numerical correctness issues." }, { "left": "issue:43986", @@ -17687,10 +17589,10 @@ "reason": "AutoProcessor without torchvision and a testing_utils bug are unrelated." }, { - "left": "issue:45753", - "right": "issue:46369", + "left": "issue:43097", + "right": "issue:46050", "accept": false, - "reason": "Infinite loop on a specific image and SAM3 memory overflow are different model/runtime failures." + "reason": "Embedding tie removal in v5 and quantized model save/load corruption are different model persistence issues." } ] }, @@ -17712,12 +17614,11 @@ "source": "llm", "cache_hit": false, "model": "gpt-5.4-mini?service_tier=flex", - "cluster_id": "cluster-10000-529", + "cluster_id": "cluster-10000-530", "nodes": [ "issue:38617", "issue:42757", "issue:42890", - "issue:42907", "issue:43012", "issue:43064", "issue:43097", @@ -17753,6 +17654,7 @@ "issue:46178", "issue:46298", "issue:46365", + "issue:46369", "issue:46421", "issue:46655", "issue:46664", @@ -17765,6 +17667,7 @@ ], "soft_pairs": [ "issue:43986|issue:45372", + "issue:45753|issue:46369", "issue:44246|issue:46981", "issue:45440|issue:46172", "issue:43097|issue:43873", @@ -17775,12 +17678,12 @@ "issue:44117|issue:46701", "issue:43874|issue:44263", "issue:44038|issue:44410", - "issue:43208|issue:43550", "issue:38617|issue:42757", "issue:46655|issue:46664", + "issue:43208|issue:43550", "issue:46298|issue:46752", - "issue:43064|issue:44315", "issue:44936|issue:46365", + "issue:43064|issue:44315", "issue:46172|issue:46710", "issue:43986|issue:44484", "issue:42890|issue:46298", @@ -17791,211 +17694,210 @@ "issue:44380|issue:46421", "issue:44938|issue:46178", "issue:43064|issue:47116", - "issue:43012|issue:43550", - "issue:42907|issue:45753" + "issue:43012|issue:43550" ], - "prepared_review_unit_hash": "16293ed45cdd808eaaf89d26da4f4eec3a2c242c77f277e2f10cf3a75a299858", + "prepared_review_unit_hash": "7029afebe72103d7d974f9e766b52d9c5d3bd12ccc33c823345efbd3f36ea341", "component_budget": { - "node_count": 529, - "item_count": 529, - "soft_pair_count": 1953, - "serialized_chars": 1012074, - "estimated_input_tokens": 253019, - "estimated_eval_tokens": 506294 + "node_count": 530, + "item_count": 530, + "soft_pair_count": 1954, + "serialized_chars": 1013004, + "estimated_input_tokens": 253251, + "estimated_eval_tokens": 506758 }, "budget": { "node_count": 48, "item_count": 48, "soft_pair_count": 29, - "serialized_chars": 36145, - "estimated_input_tokens": 9037, - "estimated_eval_tokens": 18330 + "serialized_chars": 36167, + "estimated_input_tokens": 9042, + "estimated_eval_tokens": 18340 }, "overflow_policy": "truncate_then_skip", "trimmed": false, "aggressively_trimmed": false, "split": true, "analyst_result": { - "summary": "No true duplicates here. The soft pairs mostly share only broad surface similarity (same library, same model family, import/loading bug, or trainer-related symptom) but point to different code paths and fixes.", - "confidence": 0.96, - "canonical_issue_reason": "issue:43502 is the clearest, most general user-facing bug report in the set: a precise behavioral regression ('local_files_only=True' still triggers API requests) with an obvious expected/actual mismatch and strong triage value.", + "summary": "The set is a loose collection of unrelated issues; none of the soft-edge pairs look like true duplicates or mergeable twin PRs, so all soft edges are rejected. I picked a single broad, concrete bug report as the best representative issue.", + "confidence": 0.2, + "canonical_issue_reason": "issue:43502 is a clear, self-contained bug with broad user impact: `local_files_only=True` still causes network/API requests. It has a concrete behavior mismatch and a straightforward reproduction target, making it a strong representative issue.", "canonical_pr_reason": null, - "best_issue_reason": "issue:43502 is the best representative issue for canonicalization because it is concrete, reproducible, and broadly applicable compared with the more model-specific or environment-specific reports.", + "best_issue_reason": "Same as canonical: issue:43502 is the strongest high-signal issue here because it describes a specific, impactful regression with well-defined expected behavior.", "best_pr_reason": null, "soft_edge_verdicts": [ { "left": "issue:43986", "right": "issue:45372", "accept": false, - "reason": "Both are import errors, but they involve different missing symbols from different dependencies/code paths ('torchvision' vs 'mistral_common')." + "reason": "Both are ImportErrors, but they involve different missing symbols, different dependencies, and different loading paths; not the same underlying bug." + }, + { + "left": "issue:45753", + "right": "issue:46369", + "accept": false, + "reason": "Different models and different failure modes: an infinite loop on a specific image vs a memory overflow. No shared underlying code-path bug is evident." }, { "left": "issue:44246", "right": "issue:46981", "accept": false, - "reason": "One is intermittent slow import, the other is repeated 429 retries in AutoProcessor; different failure modes and fixes." + "reason": "One is slow `import transformers`; the other is repeated retry behavior in `AutoProcessor.from_pretrained` with `HF_ENDPOINT`. Different operations and fixes." }, { "left": "issue:45440", "right": "issue:46172", "accept": false, - "reason": "Different DeepSeek bugs: remote implementation divergence vs a dtype/transpose issue in deepseek-v4." + "reason": "Both mention DeepSeek models, but one is a divergence from remote implementation and the other is a tensor dtype/transpose bug. Different concrete issues." }, { "left": "issue:43097", "right": "issue:43873", "accept": false, - "reason": "A removed API warning/compatibility issue versus quantization offloading not working; not the same bug." + "reason": "Unrelated problems: a removed config attribute in v5 vs offloading behavior under quantization." }, { "left": "issue:44410", "right": "issue:45375", "accept": false, - "reason": "Both touch Qwen3 variants, but one is missing projections in layer 0 while the other silently drops a config field." + "reason": "Different Qwen-family model bugs: missing projections in one model and a missing vision config field in another." }, { "left": "issue:44062", "right": "issue:44117", "accept": false, - "reason": "Tokenizer construction error vs tokenizer mapping returning None; different root causes." + "reason": "Tokenizer construction error vs a mapping-name lookup returning None. They touch tokenizer code but are not the same defect." }, { "left": "issue:43673", "right": "issue:46123", "accept": false, - "reason": "Generation cache/chunked prefill bug versus pipeline final-batch bookkeeping; unrelated behavior." + "reason": "Both are generation-related regressions, but one is cache availability in chunked prefill and the other is a pipeline batching flag bug. Different code paths." }, { "left": "issue:43012", "right": "issue:46821", "accept": false, - "reason": "A PyTorch warning for bfloat16 compilation versus a TimesFM AttributeError; not the same underlying issue." + "reason": "PyTorch warning about bfloat16 precision vs an `AttributeError` on `window_size`. No meaningful duplicate relationship." }, { "left": "issue:44117", "right": "issue:46701", "accept": false, - "reason": "Tokenizer mapping bug versus a DiffusionGemma pipeline failure; different subsystems." + "reason": "Tokenizer mapping assumption bug vs an image-text-to-text pipeline failure. Different components and symptoms." }, { "left": "issue:43874", "right": "issue:44263", "accept": false, - "reason": "Both are GLM-related, but one is a missing image-processor method and the other is a torch.split return-value problem." + "reason": "Both are GLM-related, but one is a missing image-processor method and the other is a `torch.split` return-value issue. Not the same bug." }, { "left": "issue:44038", "right": "issue:44410", "accept": false, - "reason": "Different Qwen model families and different defects; too broad to merge." - }, - { - "left": "issue:43208", - "right": "issue:43550", - "accept": false, - "reason": "xLSTM training blockers versus a Bamba torch.compile/SDPA failure; unrelated bugs." + "reason": "Different Qwen-family regressions with different missing pieces and different failure surfaces." }, { "left": "issue:38617", "right": "issue:42757", "accept": false, - "reason": "Both are import errors, but from different modules and missing names; not the same root cause." + "reason": "Both are import errors, but they concern unrelated symbols from different modules; no shared root cause is shown." }, { "left": "issue:46655", "right": "issue:46664", "accept": false, - "reason": "Both are SmolVLM video-processor issues, but padding calculation and num_frames semantics are distinct bugs." + "reason": "Same broad SmolVLM area, but one is padding math and the other is the meaning of `num_frames`. Distinct bugs." }, { - "left": "issue:46298", - "right": "issue:46752", + "left": "issue:43208", + "right": "issue:43550", "accept": false, - "reason": "Noisy logging versus apply_chat_template crashing on an empty list; unrelated." + "reason": "xLSTM training bugs are unrelated to a Bamba torch.compile/SDPA failure." }, { - "left": "issue:43064", - "right": "issue:44315", + "left": "issue:46298", + "right": "issue:46752", "accept": false, - "reason": "Trainer/FSDP2 optimizer-state corruption versus Liger Kernel not applied with model_init; same broad area, different concrete failures." + "reason": "Noisy logging and an `IndexError` in `apply_chat_template` are unrelated." }, { "left": "issue:44936", "right": "issue:46365", "accept": false, - "reason": "trainer.evaluate() failing after train is unrelated to importing Trainer being blocked by datasets." + "reason": "A post-train evaluation failure and an import-time Trainer conflict with `datasets` are different failure modes." + }, + { + "left": "issue:43064", + "right": "issue:44315", + "accept": false, + "reason": "One is a rank-sharded optimizer-state bug with FSDP2/PEFT; the other is Liger Kernel not applied under `model_init`. Different code paths." }, { "left": "issue:46172", "right": "issue:46710", "accept": false, - "reason": "A tensor operation bug versus an accuracy regression after upgrade; different symptom and likely fix." + "reason": "A specific tensor operation bug in DeepSeek-V4 is not the same as an accuracy regression after upgrading Transformers." }, { "left": "issue:43986", "right": "issue:44484", "accept": false, - "reason": "Video-model loading crash without torchvision versus a save_pretrained shard-size question; unrelated." + "reason": "A crash loading a video model without torchvision is unrelated to a question about `max_shard_size` defaults." }, { "left": "issue:42890", "right": "issue:46298", "accept": false, - "reason": "Test flakiness from missing set_seed is unrelated to a noisy logging message." + "reason": "Test flakiness from missing `set_seed()` is unrelated to a noisy logging message." }, { "left": "issue:44704", "right": "issue:45103", "accept": false, - "reason": "AutoProcessor kwarg forwarding bug versus auto_docstring crash with future annotations; different code paths." + "reason": "Both mention kwargs handling, but one is `AutoProcessor.from_pretrained` not forwarding kwargs and the other is a docstring parser crash with future annotations." }, { "left": "issue:38617", "right": "issue:44351", "accept": false, - "reason": "Both are missing-import reports, but for different symbols ('layer_type_validation' vs 'HybridCache') and different breakages." + "reason": "Both are missing-import complaints, but the missing names and affected APIs differ; these are separate export regressions." }, { "left": "issue:43502", "right": "issue:43992", "accept": false, - "reason": "Ignoring local_files_only and making API calls is unrelated to a missing embed_tokens.weight in UMT5Encoder." + "reason": "`local_files_only` network leakage and a missing `embed_tokens.weight` in UMT5 are different bugs in different loading paths." }, { "left": "issue:44451", "right": "issue:47116", "accept": false, - "reason": "Model-loading failure for ScandiBERT versus a streaming pipeline regression; not the same bug." + "reason": "Model load failure for ScandiBERT vs a streaming pipeline regression. No common underlying defect is indicated." }, { "left": "issue:44380", "right": "issue:46421", "accept": false, - "reason": "Attention scaling being ignored under SDPA/FlashAttention is different from cache lazy initialization causing recompilation." + "reason": "Attention-scaling being ignored under SDPA/FlashAttention is unrelated to cache lazy initialization causing recompilation." }, { "left": "issue:44938", "right": "issue:46178", "accept": false, - "reason": "Python 3.14 import failure versus a tensorflow_text backend mapping error; unrelated." + "reason": "Python 3.14 import failure and a backend mapping error for `tensorflow_text` are not the same issue." }, { "left": "issue:43064", "right": "issue:47116", "accept": false, - "reason": "Different trainer/distributed-training bug versus streaming pipeline regression." + "reason": "A distributed-training optimizer-state bug is unrelated to a streaming pipeline regression." }, { "left": "issue:43012", "right": "issue:43550", "accept": false, - "reason": "Compiler warning for bfloat16 models versus a Bamba SDPA compile failure; not the same issue." - }, - { - "left": "issue:42907", - "right": "issue:45753", - "accept": false, - "reason": "Dequantized model save failure versus an infinite loop on a specific image; unrelated symptoms and fixes." + "reason": "A precision warning during bfloat16 compilation is not the same as a model failure under SDPA with torch.compile." } ] }, @@ -18017,7 +17919,7 @@ "source": "llm", "cache_hit": false, "model": "gpt-5.4-mini?service_tier=flex", - "cluster_id": "cluster-10000-529", + "cluster_id": "cluster-10000-530", "nodes": [ "issue:36010", "issue:36683", @@ -18059,6 +17961,7 @@ "issue:45479", "issue:45584", "issue:45712", + "issue:45753", "issue:46132", "issue:46172", "issue:46298", @@ -18068,24 +17971,25 @@ "issue:47276" ], "soft_pairs": [ + "issue:42907|issue:45753", "issue:45712|issue:46298", - "issue:42907|issue:44655", "issue:43066|issue:43906", + "issue:42907|issue:44655", "issue:43208|issue:43526", "issue:45584|issue:46365", "issue:43825|issue:44655", "issue:43012|issue:44368", - "issue:43992|issue:44484", "issue:43352|issue:43976", + "issue:43992|issue:44484", "issue:43827|issue:47116", "issue:44734|issue:47116", - "issue:44380|issue:46132", "issue:44077|issue:46365", + "issue:44380|issue:46132", "issue:44162|issue:46172", "issue:36683|issue:43582", "issue:43352|issue:44380", - "issue:44380|issue:45479", "issue:42757|issue:43824", + "issue:44380|issue:45479", "issue:43097|issue:46298", "issue:43012|issue:47276", "issue:42175|issue:46365", @@ -18098,208 +18002,214 @@ "issue:42831|issue:44261", "issue:43618|issue:43761" ], - "prepared_review_unit_hash": "9082add89a2f4bd338cc6d919e3557769507de6cc9a27ecb9248ebbd7569ca55", + "prepared_review_unit_hash": "dcdf7d39a23f37282dcc2194566ed2eba35db592d5a5faae030089afe37bd44b", "component_budget": { - "node_count": 529, - "item_count": 529, - "soft_pair_count": 1953, - "serialized_chars": 1012074, - "estimated_input_tokens": 253019, - "estimated_eval_tokens": 506294 + "node_count": 530, + "item_count": 530, + "soft_pair_count": 1954, + "serialized_chars": 1013004, + "estimated_input_tokens": 253251, + "estimated_eval_tokens": 506758 }, "budget": { - "node_count": 47, - "item_count": 47, - "soft_pair_count": 29, - "serialized_chars": 35751, - "estimated_input_tokens": 8938, - "estimated_eval_tokens": 18132 + "node_count": 48, + "item_count": 48, + "soft_pair_count": 30, + "serialized_chars": 36590, + "estimated_input_tokens": 9148, + "estimated_eval_tokens": 18552 }, "overflow_policy": "truncate_then_skip", "trimmed": false, "aggressively_trimmed": false, "split": true, "analyst_result": { - "summary": "This cluster is a grab-bag of mostly unrelated Transformers issues, with only superficial keyword overlap around v5, pipelines, imports, and model-specific regressions. None of the soft pairs look like true duplicates of the same underlying bug/change.", - "confidence": 0.92, - "canonical_issue_reason": "No strong canonical issue: the set spans unrelated import errors, model/config bugs, training regressions, docs issues, and backend-specific behavior changes. If one must be chosen as the closest umbrella, issue 47116 is the broadest v5 regression theme, but it still does not represent the whole cluster.", + "summary": "This cluster is mostly a loose theme of Transformers v5 regressions and model-specific bugs, not a set of true duplicates. I did not find any soft pair that clearly shares the same underlying bug/change.", + "confidence": 0.87, + "canonical_issue_reason": "issue:43824 is the clearest, most self-contained user-facing regression in the set, with strong discussion and inbound references, making it the best representative issue for this mixed cluster.", "canonical_pr_reason": null, - "best_issue_reason": "Issue 47116 is the closest representative of the shared v5 regression/pipeline theme and is the most umbrella-like user-facing failure in the set, though the cluster remains too heterogeneous for a true duplicate canonical.", + "best_issue_reason": "issue:43824 has the strongest signal for triage: a concrete import failure, clear repro surface, and enough activity to serve as the cluster anchor.", "best_pr_reason": null, "soft_edge_verdicts": [ { - "left": "issue:45712", - "right": "issue:46298", + "left": "issue:42907", + "right": "issue:45753", "accept": false, - "reason": "Different problems: leftover dummy class cleanup vs a noisy logging message." + "reason": "Different bugs and models: failing to save dequantized Ministrals/Devstrals versus Qwen3_5 image-induced infinite loop." }, { - "left": "issue:42907", - "right": "issue:44655", + "left": "issue:45712", + "right": "issue:46298", "accept": false, - "reason": "Dequantized save failure and Pipeline.save_pretrained failure are distinct code paths." + "reason": "Unrelated maintenance issues: leftover dummy classes versus a noisy logging message." }, { "left": "issue:43066", "right": "issue:43906", "accept": false, - "reason": "Wrong tokenizer decoder type is unrelated to an isolated reproduction issue." + "reason": "Both are vague v5-related issues, but they do not describe the same concrete bug or code path." + }, + { + "left": "issue:42907", + "right": "issue:44655", + "accept": false, + "reason": "Both involve saving, but one is about dequantized model saving and the other about saving Pipeline objects; different code paths." }, { "left": "issue:43208", "right": "issue:43526", "accept": false, - "reason": "Different models and bugs: xLSTM training blockers vs BEiT Fast label reduction." + "reason": "Completely different areas: xLSTM training bugs versus BEiT image processor label reduction." }, { "left": "issue:45584", "right": "issue:46365", "accept": false, - "reason": "Whisper empty transcription crash is unrelated to the Trainer import/datasets issue." + "reason": "Whisper generation failure and Trainer import dependency conflict are unrelated." }, { "left": "issue:43825", "right": "issue:44655", "accept": false, - "reason": "Pipeline error messaging and saving Pipeline objects are different defects." + "reason": "One is a pipeline error message regression, the other is pipeline serialization; not the same issue." }, { "left": "issue:43012", "right": "issue:44368", "accept": false, - "reason": "Both are warning-related, but they concern different warnings and different triggers." + "reason": "Different warnings from different mechanisms: bf16 compilation precision warning versus tie_word_embeddings advice." }, { - "left": "issue:43992", - "right": "issue:44484", + "left": "issue:43352", + "right": "issue:43976", "accept": false, - "reason": "Missing embed_tokens weight loading and save_pretrained shard-size behavior are unrelated." + "reason": "One is a model Flash Attention support limitation, the other is a Python version compatibility problem." }, { - "left": "issue:43352", - "right": "issue:43976", + "left": "issue:43992", + "right": "issue:44484", "accept": false, - "reason": "Flash Attention support for Nemotron and Python version compatibility are separate issues." + "reason": "Missing embed_tokens weights on load and max_shard_size behavior in save_pretrained are different problems." }, { "left": "issue:43827", "right": "issue:47116", "accept": false, - "reason": "Docs still referencing pipeline() is not the same as a streaming pipeline runtime regression." + "reason": "Docs still referencing pipeline() and a streaming pipeline runtime regression are not the same bug." }, { "left": "issue:44734", "right": "issue:47116", "accept": false, - "reason": "KV-cache continuation tensor indexing bug is a different serve-time failure than streaming pipeline regression." + "reason": "Different pipeline-related failures: server KV-cache indexing versus streaming pipeline regression." }, { - "left": "issue:44380", - "right": "issue:46132", + "left": "issue:44077", + "right": "issue:46365", "accept": false, - "reason": "GPT2 attention scaling being ignored is unrelated to AttentionInterface.register behavior changes." + "reason": "PatchTSMixer post_init validation and Trainer import dependency issues are unrelated." }, { - "left": "issue:44077", - "right": "issue:46365", + "left": "issue:44380", + "right": "issue:46132", "accept": false, - "reason": "PatchTSMixer post_init policy and Trainer import failure are unrelated." + "reason": "Both touch attention code, but one is a backend-specific config being ignored while the other is register() behavior changing." }, { "left": "issue:44162", "right": "issue:46172", "accept": false, - "reason": "ESM2 breakage and DeepSeek-V4 tensor transpose behavior are different model-specific bugs." + "reason": "ESM2 breakage and deepseek-v4 tensor/transposition behavior are unrelated model-specific bugs." }, { "left": "issue:36683", "right": "issue:43582", "accept": false, - "reason": "Gemma3Config missing vocab_size is unrelated to Apple Silicon caching_allocator_warmup TypeError." + "reason": "Gemma3Config missing vocab_size and Apple Silicon allocator TypeError have no shared root cause." }, { "left": "issue:43352", "right": "issue:44380", "accept": false, - "reason": "Nemotron Flash Attention support and GPT2 attention scaling under SDPA/FlashAttention are different backend issues." + "reason": "Flash Attention support for Nemotron and GPT2 attention scaling being ignored are different issues." }, { - "left": "issue:44380", - "right": "issue:45479", + "left": "issue:42757", + "right": "issue:43824", "accept": false, - "reason": "Attention scaling ignored and single-label classification zero loss are unrelated." + "reason": "Both are import errors, but they involve different packages, symbols, and failure modes." }, { - "left": "issue:42757", - "right": "issue:43824", + "left": "issue:44380", + "right": "issue:45479", "accept": false, - "reason": "Both are import errors, but they involve different missing symbols and likely different root causes." + "reason": "Attention scaling being ignored is unrelated to single-label classification producing zero loss." }, { "left": "issue:43097", "right": "issue:46298", "accept": false, - "reason": "Tie-embeddings removal regression is unrelated to a noisy logging message." + "reason": "A removed config flag and a noisy logging message are unrelated." }, { "left": "issue:43012", "right": "issue:47276", "accept": false, - "reason": "Different warnings in different subsystems; not the same bug." + "reason": "Both mention warnings, but they concern different features and code paths." }, { "left": "issue:42175", "right": "issue:46365", "accept": false, - "reason": "Backend dependency installation issue is not the same as the Trainer import conflict." + "reason": "Torch extra backend selection and Trainer import failure via datasets are different dependency problems." }, { "left": "issue:43756", "right": "issue:44485", "accept": false, - "reason": "Smollm3 RoPE layer count bug and GLM-5 RoPE implementation discussion are different model-specific issues." + "reason": "Both mention RoPE, but they are distinct model implementations with different symptoms." }, { "left": "issue:44222", "right": "issue:45478", "accept": false, - "reason": "FP8 save_pretrained moe bug and Qwen3.5 MoE from_pretrained error are not the same failure." + "reason": "Both are MoE-related, but one is FP8 save_pretrained and the other is from_pretrained loading failure; not the same concrete bug." }, { "left": "issue:36010", "right": "issue:42757", "accept": false, - "reason": "GenerationMixin import failure and is_offline_mode import failure are different missing-symbol issues." + "reason": "Different import failures from different modules and symbols." }, { "left": "issue:43064", "right": "issue:44380", "accept": false, - "reason": "FSDP2/PEFT optimizer-state corruption is unrelated to GPT2 attention scaling." + "reason": "Trainer/FSDP2 optimizer-state corruption is unrelated to GPT2 attention scaling being ignored." }, { "left": "issue:42890", "right": "issue:47059", "accept": false, - "reason": "Seed-related test flakiness and huggingface_hub CLI attribute incompatibility are different problems." + "reason": "Test flakiness from missing set_seed and a huggingface_hub CLI incompatibility are unrelated." }, { "left": "issue:44483", "right": "issue:44514", "accept": false, - "reason": "Chat/completions request acceptance and Qwen2_5_VLProcessor batching crash are unrelated." + "reason": "Chat completions API rejection and Qwen2_5_VLProcessor batched-template crash are different endpoints and bugs." }, { "left": "issue:42831", "right": "issue:44261", "accept": false, - "reason": "FineGrainedFP8 accuracy regression and MLA rms_norm_eps precision error are different numerical bugs." + "reason": "Both concern precision, but one is a broad FineGrainedFP8 accuracy issue and the other is a missing rms_norm_eps in MLA q_a_layernorm." }, { "left": "issue:43618", "right": "issue:43761", "accept": false, - "reason": "Missing attentions in CLIPOutput and missing hidden_states in CLIPVisionModel.forward are separate v5 regressions." + "reason": "Both are CLIP regressions, but hidden_states not returned and attentions no longer assigned are separate outputs/code paths." } ] }, @@ -18321,7 +18231,7 @@ "source": "llm", "cache_hit": false, "model": "gpt-5.4-mini?service_tier=flex", - "cluster_id": "cluster-10000-529", + "cluster_id": "cluster-10000-530", "nodes": [ "issue:29870", "issue:33453", @@ -18379,8 +18289,8 @@ "issue:44655|issue:47321", "issue:44655|issue:47332", "issue:43618|issue:44380", - "issue:44625|issue:46077", "issue:43611|issue:44655", + "issue:44625|issue:46077", "issue:36683|issue:43475", "issue:43352|issue:45593", "issue:36683|issue:44077", @@ -18396,242 +18306,242 @@ "issue:29870|issue:36296", "issue:44355|issue:44855", "issue:44908|issue:46851", - "issue:43992|issue:47317", "issue:43010|issue:45278", "issue:43867|issue:46202", + "issue:43992|issue:47317", "issue:44075|issue:46365", "issue:29870|issue:33453", - "issue:43064|issue:44492", "issue:43992|issue:45310", + "issue:43064|issue:44492", "issue:44393|issue:47000", "issue:44393|issue:47337" ], - "prepared_review_unit_hash": "5a8ffb4987b6ead1be7f7f8335f2a13f5f90cb0487797a4fc4f277e152f13d9c", + "prepared_review_unit_hash": "7359977461f904586429252bd3e3ac95692d23e33d10935df91b58870a39c40b", "component_budget": { - "node_count": 529, - "item_count": 529, - "soft_pair_count": 1953, - "serialized_chars": 1012074, - "estimated_input_tokens": 253019, - "estimated_eval_tokens": 506294 + "node_count": 530, + "item_count": 530, + "soft_pair_count": 1954, + "serialized_chars": 1013004, + "estimated_input_tokens": 253251, + "estimated_eval_tokens": 506758 }, "budget": { "node_count": 47, "item_count": 47, "soft_pair_count": 33, - "serialized_chars": 36595, - "estimated_input_tokens": 9149, - "estimated_eval_tokens": 18554 + "serialized_chars": 36591, + "estimated_input_tokens": 9148, + "estimated_eval_tokens": 18552 }, "overflow_policy": "truncate_then_skip", "trimmed": false, "aggressively_trimmed": false, "split": true, "analyst_result": { - "summary": "The cluster is highly heterogeneous: it mixes unrelated model-loading regressions, config/serialization bugs, training issues, and import/runtime failures. None of the soft pairs look like the same underlying bug or change, so I would not merge any of them.", - "confidence": 0.94, - "canonical_issue_reason": "Issue 45278 is the broadest, most representative regression-style report in the set (\u201cmany import errors after update\u201d), but the cluster is not a true duplicate group, so this is only a loose representative choice.", + "summary": "This cluster is heterogeneous: the items mostly share superficial themes like loading, config, or version regressions, but not the same underlying bug. I would not merge any of the soft pairs as duplicates; issue 36683 is only the most central item in the similarity graph, not a true canonical duplicate.", + "confidence": 0.18, + "canonical_issue_reason": "Issue 36683 is the most connected and discussed item among the soft-similarity links, so it is the best single anchor for the cluster, even though the nearby issues are not the same bug.", "canonical_pr_reason": null, - "best_issue_reason": "Issue 45278 is the best single issue to anchor the cluster because it is the most general and user-impacting report, while the others are narrower, model-specific failures.", + "best_issue_reason": "36683 has the strongest cluster connectivity and is a concrete runtime failure, making it the best representative issue despite the lack of true duplicates.", "best_pr_reason": null, "soft_edge_verdicts": [ { "left": "issue:38617", "right": "issue:43502", "accept": false, - "reason": "ImportError for a missing helper and unwanted network access with local_files_only=True are different failure modes." + "reason": "Different bugs: import-time API breakage vs offline-loading network requests." }, { "left": "issue:45276", "right": "issue:46077", "accept": false, - "reason": "Both involve Gemma4, but one is about resize_token_embeddings and the other about bidirectional attention masks; different code paths." + "reason": "Same model family, but different code paths and symptoms: embedding resize vs attention-mask construction." }, { "left": "issue:44625", "right": "issue:46365", "accept": false, - "reason": "Qwen3.5 num_labels propagation and Trainer import failure due to datasets are unrelated." + "reason": "Unrelated failures: Qwen3.5 config propagation vs Trainer import conflict." }, { "left": "issue:43299", "right": "issue:43867", "accept": false, - "reason": "Qwen3-VL MoE loading regression and state_dict sorting load error are different load-path bugs." + "reason": "Both are loading issues, but for different models and different failure modes." }, { "left": "issue:44655", "right": "issue:47321", "accept": false, - "reason": "Saving Pipeline objects and AMD quark class updates are unrelated." + "reason": "Different subsystems: pipeline serialization vs AMD quark class update state." }, { "left": "issue:44655", "right": "issue:47332", "accept": false, - "reason": "Both mention saving, but one is pipeline save_pretrained and the other is disk-offloaded Qwen3-VL saving; not the same bug." + "reason": "Different save paths and root causes; offloading save is not the same as pipeline save_pretrained." }, { "left": "issue:43618", "right": "issue:44380", "accept": false, - "reason": "CLIPOutput attentions assignment and GPT2 attention scaling under SDPA/FlashAttention are different attention issues." + "reason": "Attention-related but not the same bug: CLIP output fields vs GPT-2 backend scaling." }, { - "left": "issue:44625", - "right": "issue:46077", + "left": "issue:43611", + "right": "issue:44655", "accept": false, - "reason": "Qwen3.5 num_labels propagation is unrelated to Gemma4 attention-mask construction." + "reason": "One is model loading with base_model_prefix; the other is pipeline saving." }, { - "left": "issue:43611", - "right": "issue:44655", + "left": "issue:44625", + "right": "issue:46077", "accept": false, - "reason": "Model loading with base_model_prefix and saving Pipeline objects are different serialization problems." + "reason": "Different model configs and different behavior: label propagation vs attention masking." }, { "left": "issue:36683", "right": "issue:43475", "accept": false, - "reason": "Missing vocab_size on Gemma3Config and missing fpn_position_embeddings on Sam3VisionEncoderOutput are different model attribute bugs." + "reason": "Both are attribute errors, but on different model outputs and different missing fields." }, { "left": "issue:43352", "right": "issue:45593", "accept": false, - "reason": "Flash Attention support error for Nemotron and missing auxiliary losses in D-FINE are unrelated." + "reason": "Unrelated: flash-attention support check vs denoising auxiliary-loss logic." }, { "left": "issue:36683", "right": "issue:44077", "accept": false, - "reason": "Gemma3Config missing vocab_size is unrelated to patchtsmixer post_init handling." + "reason": "Different configuration bugs in different models; no shared code path." }, { "left": "issue:44355", "right": "issue:46877", "accept": false, - "reason": "Compiled Python import errors and auto_docstring stdout lint noise are different import-time problems." + "reason": "Both touch import-time behavior, but one is compiled-file errors and the other is docstring lint output." }, { "left": "issue:43976", "right": "issue:44938", "accept": false, - "reason": "Both are Python-version compatibility issues, but they affect different versions and failure points." + "reason": "Both are Python-version compatibility reports, but for different versions and likely different causes." }, { "left": "issue:46077", "right": "issue:46512", "accept": false, - "reason": "Gemma4 attention-mask logic and stale config.dtype after from_config are unrelated." + "reason": "Different regression types: attention-mask construction vs stale config dtype after from_config." }, { "left": "issue:43867", "right": "issue:45216", "accept": false, - "reason": "Sorting-related load failure and incorrect Qwen3.5 checkpoint saving are different serialization defects." + "reason": "Both involve model checkpoint/loading, but the reported failures and affected models differ." }, { "left": "issue:44263", "right": "issue:45440", "accept": false, - "reason": "A torch.split return-value issue in one MoE indexer and DeepseekV3MoE divergence from remote implementation are not the same bug." + "reason": "Different MoE/model-implementation bugs with no concrete shared code path." }, { "left": "issue:44938", "right": "issue:47059", "accept": false, - "reason": "Python 3.14 loading failure and a huggingface_hub CLI attribute error are unrelated." + "reason": "Python 3.14 import failure vs CLI breakage from huggingface_hub API change." }, { "left": "issue:43618", "right": "issue:45216", "accept": false, - "reason": "CLIPOutput attentions assignment and Qwen3.5 save_pretrained checkpoint correctness are unrelated." + "reason": "Different models and different regressions: missing attentions vs incorrect saved checkpoint." }, { "left": "issue:36683", "right": "issue:47110", "accept": false, - "reason": "Gemma3Config missing vocab_size and special_tokens list-key access error are different issues." + "reason": "Both are attribute errors, but on unrelated objects and different APIs." }, { "left": "issue:36683", "right": "issue:44625", "accept": false, - "reason": "Missing vocab_size and num_labels propagation are separate config bugs." + "reason": "Related only at a very high level; the missing/propagated field is different and model-specific." }, { "left": "issue:29870", "right": "issue:36296", "accept": false, - "reason": "BART generate() decoder_attention_mask error and tensor-parallel training bug are unrelated." + "reason": "Completely different failures: generate() mask bug vs tensor-parallel training issue." }, { "left": "issue:44355", "right": "issue:44855", "accept": false, - "reason": "Both are import-related, but compiled-file errors and a Python 3.13 torch.jit.script parsing problem are different root causes." + "reason": "Both are Python/import related, but one is compiled-file handling and the other is torch.jit parsing." }, { "left": "issue:44908", "right": "issue:46851", "accept": false, - "reason": "Inverse-sqrt scheduler kwargs handling and Qwen3.5 variable-length position_ids are unrelated." - }, - { - "left": "issue:43992", - "right": "issue:47317", - "accept": false, - "reason": "Missing embed_tokens.weight in UMT5Encoder and PPFormulaNet training-loss bug are unrelated." + "reason": "Different features: scheduler kwargs handling vs variable-length position_ids in Qwen3.5." }, { "left": "issue:43010", "right": "issue:45278", "accept": false, - "reason": "No-grad decoration for cache/layer updates and broad import errors after version update are different issues." + "reason": "One is a cache no_grad proposal; the other is a broad import regression report." }, { "left": "issue:43867", "right": "issue:46202", "accept": false, - "reason": "Load failure with sorted state_dict and Gemma pipeline roundtrip on the \u2581 token are unrelated." + "reason": "Different issue domains: state_dict loading vs tokenizer roundtrip failure." + }, + { + "left": "issue:43992", + "right": "issue:47317", + "accept": false, + "reason": "Unrelated: missing encoder weight in from_pretrained vs a training-loss bug." }, { "left": "issue:44075", "right": "issue:46365", "accept": false, - "reason": "SGD args not used and Trainer import blocked by datasets are unrelated." + "reason": "Optimizer argument handling is unrelated to Trainer import failure." }, { "left": "issue:29870", "right": "issue:33453", "accept": false, - "reason": "BART generate() decoder_attention_mask bug and tokenizer loading regression are different failures." + "reason": "Different bugs in different subsystems: generation mask error vs tokenizer loading regression." }, { - "left": "issue:43064", - "right": "issue:44492", + "left": "issue:43992", + "right": "issue:45310", "accept": false, - "reason": "FSDP2/PEFT optimizer-state corruption and a cache-strategy typo are unrelated." + "reason": "Both are model-loading regressions, but for different models and likely different missing-key causes." }, { - "left": "issue:43992", - "right": "issue:45310", + "left": "issue:43064", + "right": "issue:44492", "accept": false, - "reason": "UMT5Encoder missing embed_tokens.weight and Qwen3.5 MoE from_pretrained error are different model-loading issues." + "reason": "Training/optimizer state bug vs a cache-strategy typo; no shared root cause." }, { "left": "issue:44393", "right": "issue:47000", "accept": false, - "reason": "Qwen3-VL 2D bbox output issues and RTDetrLoss infeasible cost matrix are different vision bugs." + "reason": "Different vision tasks and failure modes: bbox hallucination/output vs RTDetr loss infeasibility." }, { "left": "issue:44393", "right": "issue:47337", "accept": false, - "reason": "Qwen3-VL bbox hallucination/error and flaky bbox-coordinate test failure are not the same concrete bug." + "reason": "Both are vision-related, but one is model output error and the other is a flaky test." } ] }, @@ -18653,7 +18563,7 @@ "source": "llm", "cache_hit": false, "model": "gpt-5.4-mini?service_tier=flex", - "cluster_id": "cluster-10000-529", + "cluster_id": "cluster-10000-530", "nodes": [ "issue:28282", "issue:30333", @@ -18714,8 +18624,8 @@ "issue:42890|issue:45561", "issue:43867|issue:45357", "issue:28282|issue:33453", - "issue:42548|issue:44351", "issue:43010|issue:45850", + "issue:42548|issue:44351", "issue:45125|issue:45478", "issue:43618|issue:44977", "issue:45125|issue:45375", @@ -18728,30 +18638,30 @@ "issue:44387|issue:44556", "issue:43723|issue:44556", "issue:42898|issue:46664", - "issue:45071|issue:47317", - "issue:43010|issue:43065", "issue:43824|issue:46617", + "issue:43010|issue:43065", + "issue:45071|issue:47317", "issue:43576|issue:46132", "issue:43502|issue:44485", "issue:44355|issue:44749", - "issue:45850|issue:47110", "issue:30333|issue:33453", + "issue:45850|issue:47110", "issue:43874|issue:46617" ], - "prepared_review_unit_hash": "557ce508b778202014115eec0dc3bea612257d4d5b7bbb65c57f88a67beedaa5", + "prepared_review_unit_hash": "723900a7f123346185b2cbd9f77f932e7cd4397581cf015b2f219f0db7269d7c", "component_budget": { - "node_count": 529, - "item_count": 529, - "soft_pair_count": 1953, - "serialized_chars": 1012074, - "estimated_input_tokens": 253019, - "estimated_eval_tokens": 506294 + "node_count": 530, + "item_count": 530, + "soft_pair_count": 1954, + "serialized_chars": 1013004, + "estimated_input_tokens": 253251, + "estimated_eval_tokens": 506758 }, "budget": { "node_count": 48, "item_count": 48, "soft_pair_count": 32, - "serialized_chars": 36792, + "serialized_chars": 36789, "estimated_input_tokens": 9198, "estimated_eval_tokens": 18652 }, @@ -18760,214 +18670,205 @@ "aggressively_trimmed": false, "split": true, "analyst_result": { - "summary": "This cluster is mostly a grab-bag of unrelated Transformers issues. Only the import slowdown pair looks like the same underlying bug; the rest are distinct model, tokenizer, config, CI, and API problems.", - "confidence": 0.86, - "canonical_issue_reason": "issue:41669 is the best anchor because it states the concrete root cause of the import slowdown, while issue:44246 is the vaguer symptom report.", + "summary": "This is a noisy, heterogeneous issue cluster: a few items are broadly import/loading/performance themed, but most soft pairs point to different failures and different code paths. I would not merge any of the proposed soft edges; the best representative issue is a broad import-slowdown report, but only weakly.", + "confidence": 0.82, + "canonical_issue_reason": "Issue 44246 is the most generic user-facing symptom in the cluster and best matches the broad import/performance theme, though the overall cluster is not a clean duplicate set.", "canonical_pr_reason": null, - "best_issue_reason": "issue:41669 is the most specific and actionable representative of the only real duplicate theme in the set.", + "best_issue_reason": "44246 is the most searchable, broadly phrased issue here and can serve as the least-bad anchor for this mixed set, even if it does not truly subsume the others.", "best_pr_reason": null, "soft_edge_verdicts": [ { "left": "issue:44222", "right": "issue:45125", "accept": false, - "reason": "Different problems: FP8 save_pretrained/MoE versus missing tensor-parallel plan on Qwen3.5 MoE." + "reason": "Different problems: FP8 save_pretrained behavior vs missing tensor-parallel plan for Qwen3.5 MoE." }, { "left": "issue:36296", "right": "issue:43575", "accept": false, - "reason": "Both involve tensor parallelism, but one is a generic training bug and the other is a model load OOM; not the same code-path." + "reason": "Both mention tensor parallelism, but one is a generic training bug and the other is an OOM while loading a specific model." }, { "left": "issue:41669", "right": "issue:44246", - "accept": true, - "reason": "Same underlying issue: import-time slowdown in transformers, with 41669 identifying the likely root cause and 44246 reporting the symptom." + "accept": false, + "reason": "Related to import cost, but one is a refactor removing import * while the other is an intermittent import-slowdown report." }, { "left": "issue:43618", "right": "issue:44077", "accept": false, - "reason": "Unrelated model-output regression versus a PatchTSTMixer init/post_init change." + "reason": "Unrelated: CLIP attentions assignment vs a patchtsmixer post_init policy issue." }, { "left": "issue:43867", "right": "issue:45375", "accept": false, - "reason": "Sorted state_dict loading error is unrelated to Qwen3.5 vision config field dropping." + "reason": "Different serialization/config issues: sorted state_dict load failure vs missing vision config field under strict parsing." }, { "left": "issue:43010", "right": "issue:43976", "accept": false, - "reason": "torch.no_grad cache update change is unrelated to Python version compatibility." + "reason": "No shared bug path: no_grad decoration for cache updates vs Python version compatibility." }, { "left": "issue:42890", "right": "issue:45561", "accept": false, - "reason": "Different flaky-test failures with different causes: missing seeding versus xdist file races." + "reason": "Both are test-related, but one is seed-related flakiness and the other is an xdist file race." }, { "left": "issue:43867", "right": "issue:45357", "accept": false, - "reason": "Different failure modes: load-model state_dict sorting versus incorrect save_pretrained visual keys." + "reason": "Load-time state_dict ordering bug vs save_pretrained writing wrong visual encoder keys." }, { "left": "issue:28282", "right": "issue:33453", "accept": false, - "reason": "PyTorch import missing versus tokenizer loading regression; no shared bug." + "reason": "Different import/model loading failures: missing PyTorch backend vs tokenizer-loading regression." }, { - "left": "issue:42548", - "right": "issue:44351", + "left": "issue:43010", + "right": "issue:45850", "accept": false, - "reason": "Both are import errors, but for different symbols and likely different root causes." + "reason": "A specific cache-update correctness request is not the same as a broad 'many models break' complaint." }, { - "left": "issue:43010", - "right": "issue:45850", + "left": "issue:42548", + "right": "issue:44351", "accept": false, - "reason": "Specific no_grad decorator request versus a broad complaint about v5 breaking many models." + "reason": "Both are import errors, but for different symbols and likely different underlying causes." }, { "left": "issue:45125", "right": "issue:45478", "accept": false, - "reason": "Related to Qwen3.5 MoE, but one is tensor-parallel metadata and the other is from_pretrained failure; not clearly the same fix." + "reason": "The latter is a general Qwen3.5 MoE from_pretrained regression; no clear evidence it is the same as the missing _tp_plan issue." }, { "left": "issue:43618", "right": "issue:44977", "accept": false, - "reason": "Different model-specific generation/attention issues." + "reason": "Different models and failures: CLIPOutput metadata vs Qwen3.5 flash-attention generation." }, { "left": "issue:45125", "right": "issue:45375", "accept": false, - "reason": "Tensor-parallel plan missing versus vision config field missing; different parts of the Qwen3.5 stack." + "reason": "Both are Qwen3.5-related, but one is tensor-parallel support and the other is a strict vision-config field omission." }, { "left": "issue:44360", "right": "issue:45440", "accept": false, - "reason": "DSA indexer/ReLU discussion is unrelated to DeepseekV3MoE parity." + "reason": "Completely different subsystems: DSA indexing vs DeepSeekV3MoE implementation divergence." }, { "left": "issue:43502", "right": "issue:44062", "accept": false, - "reason": "Local-files-only network leak versus tokenizers.AddedToken keyword conflict." + "reason": "Unrelated bugs: offline loading still makes requests vs AddedToken keyword duplication." }, { "left": "issue:43525", "right": "issue:47317", "accept": false, - "reason": "Llama4 config attribute error is unrelated to PPFormulaNet training loss." + "reason": "Different model families and symptoms: missing pad_token_id vs a training-loss bug." }, { "left": "issue:42890", "right": "issue:44112", "accept": false, - "reason": "Different test failures in different model areas." + "reason": "Both are CI/test problems, but the causes are different and not the same failing test path." }, { "left": "issue:42890", "right": "issue:44964", "accept": false, - "reason": "Seed-related flaky test versus model loading failure for Phi-4 multimodal." + "reason": "Seed-related test instability is not the same as a model-loading failure for Phi-4 multimodal." }, { "left": "issue:44734", "right": "issue:46617", "accept": false, - "reason": "Serve KV-cache indexing crash is unrelated to trust_remote_code cache FileNotFoundError." + "reason": "Serve-time KV-cache indexing crash vs trust_remote_code cache FileNotFoundError are different loading/runtime paths." }, { "left": "issue:44387", "right": "issue:44556", "accept": false, - "reason": "Int4 quantization memory regression is unrelated to checkpoint reload compatibility after upgrade." + "reason": "Memory regression under int4 quantization is not the same as checkpoint incompatibility after upgrade." }, { "left": "issue:43723", "right": "issue:44556", "accept": false, - "reason": "Tokenizer load regression versus older checkpoint reload incompatibility." + "reason": "Tokenizer-loading regression and checkpoint reload incompatibility are different failures." }, { - "left": "issue:42898", - "right": "issue:46664", + "left": "issue:45850", + "right": "issue:47110", "accept": false, - "reason": "Tokenization-space behavior change is unrelated to video num_frames semantics." + "reason": "A broad complaint that many models break is not the same as a specific special_tokens type error." }, { - "left": "issue:45071", - "right": "issue:47317", + "left": "issue:43874", + "right": "issue:46617", + "accept": false, + "reason": "Different failures: multimodal token counting helper missing vs trust_remote_code cache file lookup." + }, + { + "left": "issue:43824", + "right": "issue:46617", "accept": false, - "reason": "PretrainedConfig type-checking regression is unrelated to PPFormulaNet training loss." + "reason": "ImportError for a model class and a cache FileNotFoundError are not the same bug." }, { "left": "issue:43010", "right": "issue:43065", "accept": false, - "reason": "No_grad on cache updates is unrelated to a dummy Conv2d in Sam3PixelDecoder." + "reason": "No shared code path: no_grad on cache updates vs a dummy Conv2d in Sam3PixelDecoder." }, { - "left": "issue:43824", - "right": "issue:46617", + "left": "issue:45071", + "right": "issue:47317", "accept": false, - "reason": "Missing top-level Qwen2_5_VL export is unrelated to cache-loading FileNotFoundError." + "reason": "Type-checking regression in PretrainedConfig is unrelated to a model training-loss bug." }, { "left": "issue:43576", "right": "issue:46132", "accept": false, - "reason": "Broken env CLI and AttentionInterface.register behavior are different features with different failures." + "reason": "Both are API-behavior complaints, but one is a CLI env command and the other is attention registration semantics." }, { "left": "issue:43502", "right": "issue:44485", "accept": false, - "reason": "local_files_only network leakage is unrelated to GLM-5 RoPE implementation." + "reason": "Offline-loading network requests and GLM-5 RoPE implementation are unrelated." }, { "left": "issue:44355", "right": "issue:44749", "accept": false, - "reason": "Compiled Python file errors are unrelated to the slowdown after upgrading from 4.57.3 to 5.3.0." - }, - { - "left": "issue:45850", - "right": "issue:47110", - "accept": false, - "reason": "Broad v5 breakage complaint versus special_tokens keys attribute error; too vague and not the same bug." + "reason": "Compiled-Python-file errors and a reported upgrade slowdown are not the same bug." }, { "left": "issue:30333", "right": "issue:33453", "accept": false, - "reason": "MLFlow reporting on failed training jobs is unrelated to tokenizer loading regression." - }, - { - "left": "issue:43874", - "right": "issue:46617", - "accept": false, - "reason": "GLM46V image processor missing method is unrelated to trust_remote_code cache loading." + "reason": "MLFlow reporting on failed training jobs is unrelated to tokenizer-loading regression." } ] }, - "evaluator_result": { - "accept": true, - "feedback": "The summary is grounded in the packet and stays conservative. The only accepted soft pair, issue:41669 with issue:44246, is well supported by the titles and the referenced import slowdown theme. The other pair rejections are cautious and do not overstate evidence." - }, - "evaluator_used": true, + "evaluator_result": null, + "evaluator_used": false, "retried": false, - "accepted_nontrivial_soft_edge": true, + "accepted_nontrivial_soft_edge": false, "error_kind": null, "error_message": null }, @@ -18982,7 +18883,7 @@ "source": "llm", "cache_hit": false, "model": "gpt-5.4-mini?service_tier=flex", - "cluster_id": "cluster-10000-529", + "cluster_id": "cluster-10000-530", "nodes": [ "issue:29870", "issue:30333", @@ -19034,8 +18935,8 @@ "issue:47110" ], "soft_pairs": [ - "issue:41628|issue:42548", "issue:30990|issue:31795", + "issue:41628|issue:42548", "issue:44222|issue:44230", "issue:43010|issue:43232", "issue:43299|issue:44556", @@ -19051,8 +18952,8 @@ "issue:43502|issue:44961", "issue:43576|issue:43824", "issue:43576|issue:45071", - "issue:43576|issue:47059", "issue:44230|issue:46095", + "issue:43576|issue:47059", "issue:42757|issue:43576", "issue:31795|issue:38617", "issue:36296|issue:43901", @@ -19062,225 +18963,231 @@ "issue:45362|issue:45850", "issue:45375|issue:47110", "issue:43525|issue:47110", - "issue:43673|issue:44492", "issue:44485|issue:44961", + "issue:43673|issue:44492", "issue:44315|issue:46209", "issue:43352|issue:44961" ], - "prepared_review_unit_hash": "5c36153f0f61e31eec11d29c3cb37bfe7a13943c9a1316829e46bdc6db71283e", + "prepared_review_unit_hash": "303366f2afeb1755352f787840cec1775ec7a4b24594fbee01f4d15a4f8dd3fb", "component_budget": { - "node_count": 529, - "item_count": 529, - "soft_pair_count": 1953, - "serialized_chars": 1012074, - "estimated_input_tokens": 253019, - "estimated_eval_tokens": 506294 + "node_count": 530, + "item_count": 530, + "soft_pair_count": 1954, + "serialized_chars": 1013004, + "estimated_input_tokens": 253251, + "estimated_eval_tokens": 506758 }, "budget": { "node_count": 48, "item_count": 48, "soft_pair_count": 32, - "serialized_chars": 36691, - "estimated_input_tokens": 9173, - "estimated_eval_tokens": 18602 + "serialized_chars": 36693, + "estimated_input_tokens": 9174, + "estimated_eval_tokens": 18604 }, "overflow_policy": "truncate_then_skip", "trimmed": false, "aggressively_trimmed": false, "split": true, "analyst_result": { - "summary": "This cluster is mostly a loose thematic grouping of unrelated issues (imports, generation, loading, docs, FP8/Qwen, CLI, etc.), so I do not see any true duplicate pairs among the soft edges. I picked one concrete version-regression issue as the best representative of the cluster, but confidence is limited because the set is heterogeneous.", - "confidence": 0.58, - "canonical_issue_reason": "Best representative of the cluster is issue:45071 because it is a concrete, user-facing v5 regression with a clear symptom and broad relevance compared with the more scattered, model-specific or doc-only reports.", + "summary": "This cluster is mostly a loose mix of unrelated Transformers v5 regressions, import errors, model-specific bugs, and docs/tooling issues. None of the soft pairs look like true duplicates or mergeable into one fix.", + "confidence": 0.86, + "canonical_issue_reason": "issue:45071 is the most concrete representative of the cluster\u2019s broad v5 compatibility-regression theme: it names a specific breakage in PretrainedConfig type checking, unlike the more generic, model-specific, or docs-only reports.", "canonical_pr_reason": null, - "best_issue_reason": "issue:45071 is the strongest single issue here: specific, reproducible, and framed as an explicit regression, unlike many of the other items which are either model-specific, documentation-only, or very broad.", + "best_issue_reason": "issue:45071 is the strongest standalone issue in the set because it is specific, version-scoped, and broadly indicative of the kind of v5 breakages seen across the cluster; the other issues are either narrower, less actionable, or clearly unrelated.", "best_pr_reason": null, "soft_edge_verdicts": [ { - "left": "issue:41628", - "right": "issue:42548", + "left": "issue:30990", + "right": "issue:31795", "accept": false, - "reason": "Both are import errors, but they involve different symbols and likely different missing exports; too broad to treat as the same bug." + "reason": "Loading hang vs documentation confusion; different failure mode and no shared code-path evidence." }, { - "left": "issue:30990", - "right": "issue:31795", + "left": "issue:41628", + "right": "issue:42548", "accept": false, - "reason": "Stuck-loading behavior and documentation confusion are unrelated problems." + "reason": "Both are import errors, but for different symbols and likely different causes." }, { "left": "issue:44222", "right": "issue:44230", "accept": false, - "reason": "Both mention FP8/MoE, but one is about saving and the other about inference support; different code paths." + "reason": "FP8 save/load bug vs FP8 inference support request; related area, but not the same concrete issue." }, { "left": "issue:43010", "right": "issue:43232", "accept": false, - "reason": "Both touch generation/cache internals, but the reported bugs are different: no_grad decoration vs sync_gpus update handling." + "reason": "Different generation/cache bugs with different triggers and code paths." }, { "left": "issue:43299", "right": "issue:44556", "accept": false, - "reason": "Both are loading/regression reports, but they concern different models and different failure modes." + "reason": "Both are upgrade regressions, but one is Qwen3VL loading and the other is checkpoint reload compatibility." }, { "left": "issue:46143", "right": "issue:46144", "accept": false, - "reason": "Related RoFormer area, but one is missing kwargs passthrough and the other is attention-interface usage; not the same underlying defect." + "reason": "`**kwargs` forwarding bug vs attention-interface implementation issue; too different to be duplicates." }, { "left": "issue:46664", "right": "issue:46726", "accept": false, - "reason": "Video-processor semantics and missing-vision-backend error are different issues." + "reason": "Video processor semantics vs missing-vision-backend error; unrelated causes." }, { "left": "issue:43502", "right": "issue:44222", "accept": false, - "reason": "Local-files-only network leakage and FP8 save_pretrained are unrelated." + "reason": "Offline/local-files-only network leak vs FP8 moe save_pretrained support; different subsystems." }, { "left": "issue:30333", "right": "issue:36296", "accept": false, - "reason": "MLflow reporting on failed jobs and tensor-parallel training are unrelated." + "reason": "MLFlow reporting on failed jobs vs tensor-parallel training bug; no overlap beyond training context." }, { "left": "issue:43827", "right": "issue:44961", "accept": false, - "reason": "Docs still referencing pipeline() and an unrelated issue titled 'racoon' are not duplicates." + "reason": "Docs referencing removed pipeline APIs vs an unrelated placeholder issue." }, { "left": "issue:29870", "right": "issue:31795", "accept": false, - "reason": "A generation-time UnboundLocalError and documentation confusion are unrelated." + "reason": "Runtime UnboundLocalError in generate vs documentation ambiguity; not the same bug." }, { "left": "issue:33357", "right": "issue:33453", "accept": false, - "reason": "MacOS bus error for CLIP and tokenizer-loading regression are different failures." + "reason": "MacOS CLIP bus error vs tokenizer loading regression; different model and symptom." }, { "left": "issue:44038", "right": "issue:45125", "accept": false, - "reason": "Both are Qwen/MoE-related, but one is a general v5 bug report and the other is a tensor-parallelism plan omission." + "reason": "General Qwen3-VL-Moe breakage vs missing tensor-parallel plan for Qwen3_5Moe; related family, but not one fix." }, { "left": "issue:44355", "right": "issue:45071", "accept": false, - "reason": "Compiled Python-file errors and PretrainedConfig type-checking regression are different problems." + "reason": "Compiled Python file errors vs PretrainedConfig type checking regression; unrelated." }, { "left": "issue:43502", "right": "issue:44961", "accept": false, - "reason": "Network requests despite local_files_only and an unrelated issue titled 'racoon' are not duplicates." + "reason": "Local-files-only network leak vs unrelated placeholder issue." }, { "left": "issue:43576", "right": "issue:43824", "accept": false, - "reason": "Broken env command and Qwen2.5-VL import error are unrelated." + "reason": "Broken `transformers env` CLI vs Qwen2.5-VL import error; different tools and failures." }, { "left": "issue:43576", "right": "issue:45071", "accept": false, - "reason": "CLI/env command breakage and PretrainedConfig type-checking regression do not share the same code path." + "reason": "CLI breakage and config type-checking regression are distinct problems." }, { - "left": "issue:43576", - "right": "issue:47059", + "left": "issue:44230", + "right": "issue:46095", "accept": false, - "reason": "Different CLI failures with different causes and surfaces; not the same bug." + "reason": "FP8 inference support request vs weight-conversion question; not the same underlying bug." }, { - "left": "issue:44230", - "right": "issue:46095", + "left": "issue:43576", + "right": "issue:47059", "accept": false, - "reason": "FP8 inference support and DeepSeek weight-conversion guidance are not the same issue." + "reason": "Different CLI failures: `transformers env` vs hub completion attribute error." }, { "left": "issue:42757", "right": "issue:43576", "accept": false, - "reason": "Missing huggingface_hub symbol and broken transformers env command are unrelated." + "reason": "Importing `is_offline_mode` from hub vs broken transformers env command; unrelated." }, { "left": "issue:31795", "right": "issue:38617", "accept": false, - "reason": "Documentation ambiguity and an ImportError are not the same underlying bug." + "reason": "Docs confusion vs missing `layer_type_validation` import; different scope." }, { "left": "issue:36296", "right": "issue:43901", "accept": false, - "reason": "Tensor-parallel training bug and pipeline documentation mismatch are unrelated." + "reason": "Tensor-parallel training bug vs docs mismatch about `return_all_scores`; not duplicates." }, { "left": "issue:43475", "right": "issue:46617", "accept": false, - "reason": "Missing SAM 3 Video attribute and trust_remote_code cache FileNotFoundError are different failures." + "reason": "Missing SAM3 attribute vs cache file lookup error; different models and code paths." }, { "left": "issue:44351", "right": "issue:46617", "accept": false, - "reason": "Importing HybridCache and loading a cached trust_remote_code model are different problems." + "reason": "Import error for `HybridCache` vs cache FileNotFoundError; separate failures." }, { "left": "issue:44315", "right": "issue:44360", "accept": false, - "reason": "Liger kernel application and DSA indexer ReLU discussion are unrelated." + "reason": "Liger Kernel/model_init issue vs DSA indexer discussion; no shared bug." }, { "left": "issue:45362", "right": "issue:45850", "accept": false, - "reason": "A specific Qwen3.5-35B crash and a broad complaint that v5 breaks many models are not the same concrete bug." + "reason": "Specific Qwen3.5-35B chat crash vs generic 'v5 breaks too many models' complaint; not one concrete issue." }, { "left": "issue:45375", "right": "issue:47110", "accept": false, - "reason": "Missing Qwen3.5MoeVisionConfig field and special_tokens list/object error are unrelated." + "reason": "Missing config field in Qwen3_5MoeVisionConfig vs special_tokens list-type error; different object and failure." }, { "left": "issue:43525", "right": "issue:47110", "accept": false, - "reason": "Llama4Config pad_token_id absence and special_tokens list/object error are different issues." + "reason": "Missing pad_token_id on Llama4Config vs special_tokens list attribute error; unrelated attribute mismatches." + }, + { + "left": "issue:44485", + "right": "issue:44961", + "accept": false, + "reason": "GLM-5 RoPE implementation discussion vs unrelated placeholder issue." }, { "left": "issue:43673", "right": "issue:44492", "accept": false, - "reason": "Generation cache missing in chunked_prefill and a typo in cache strategies are unrelated." + "reason": "Generation cache missing in chunked prefill vs cache-strategy typo; different severity and topic." }, { - "left": "issue:44485", - "right": "issue:44961", + "left": "issue:44315", + "right": "issue:46209", "accept": false, - "reason": "GLM-5 RoPE implementation and an unrelated 'racoon' issue are not duplicates." + "reason": "Model_init/Liger Kernel application bug vs FP8 inference on B200; different code paths." }, { "left": "issue:43352", "right": "issue:44961", "accept": false, - "reason": "Flash-Attention support limitation and an unrelated issue titled 'racoon' are not duplicates." + "reason": "Flash Attention support limitation vs unrelated placeholder issue." } ] }, @@ -19302,13 +19209,12 @@ "source": "llm", "cache_hit": false, "model": "gpt-5.4-mini?service_tier=flex", - "cluster_id": "cluster-10000-529", + "cluster_id": "cluster-10000-530", "nodes": [ - "issue:31795", - "issue:33453", "issue:41669", "issue:42371", "issue:42890", + "issue:43298", "issue:43408", "issue:43525", "issue:43572", @@ -19345,6 +19251,7 @@ "issue:45753", "issue:45850", "issue:46018", + "issue:46095", "issue:46172", "issue:46218", "issue:46328", @@ -19376,272 +19283,272 @@ "issue:43408|issue:44315", "issue:44492|issue:45850", "issue:41669|issue:46218", - "issue:43976|issue:44393", "issue:44492|issue:46328", + "issue:43976|issue:44393", "issue:44295|issue:45230", "issue:42890|issue:44908", "issue:44297|issue:45230", "issue:44704|issue:44908", "issue:44908|issue:45071", - "issue:43576|issue:44908", "issue:41669|issue:45341", + "issue:43576|issue:44908", "issue:43576|issue:47176", "issue:41669|issue:44351", "issue:46018|issue:46172", "issue:43976|issue:46218", "issue:43976|issue:47176", - "issue:31795|issue:33453" + "issue:43298|issue:46095" ], - "prepared_review_unit_hash": "eaa00cfdbb4fa64db4c60731d1448a899a87f28ab00883bc21153d1effde776d", + "prepared_review_unit_hash": "1c048d3c78bd4e729cd55b83baa4d1c7488c02f60c6691b975f1a7e85930712c", "component_budget": { - "node_count": 529, - "item_count": 529, - "soft_pair_count": 1953, - "serialized_chars": 1012074, - "estimated_input_tokens": 253019, - "estimated_eval_tokens": 506294 + "node_count": 530, + "item_count": 530, + "soft_pair_count": 1954, + "serialized_chars": 1013004, + "estimated_input_tokens": 253251, + "estimated_eval_tokens": 506758 }, "budget": { "node_count": 48, "item_count": 48, "soft_pair_count": 37, - "serialized_chars": 38051, - "estimated_input_tokens": 9513, - "estimated_eval_tokens": 19282 + "serialized_chars": 38147, + "estimated_input_tokens": 9537, + "estimated_eval_tokens": 19330 }, "overflow_policy": "truncate_then_skip", "trimmed": false, "aggressively_trimmed": false, "split": true, "analyst_result": { - "summary": "These items are a grab bag of unrelated issues across loading, caching, tokenizers, model configs, CLI, and performance. None of the proposed soft pairs look like the same underlying bug/change, so I would not merge any of them as duplicates.", - "confidence": 0.9, - "canonical_issue_reason": "I\u2019d use issue:41669 as the canonical issue: it is a concrete, high-impact performance regression with a clear suspected cause and a focused fix path, unlike the more vague or model-specific reports.", + "summary": "This batch is mostly heterogeneous. The closest recurring theme is import/lazy-loading behavior, but the soft pairs are still too weakly connected to treat as duplicates; I would not merge any of them.", + "confidence": 0.34, + "canonical_issue_reason": "issue:44246 is the clearest and broadest actionable report in the set: a concrete import-time performance symptom that best represents the import/lazy-loading thread.", "canonical_pr_reason": null, - "best_issue_reason": "Issue:41669 is the strongest standalone issue here because it is specific, actionable, and broadly relevant to import-time behavior.", + "best_issue_reason": "issue:44246 is the best canonical issue candidate because it states a direct user-facing regression clearly and broadly enough to subsume related import-performance complaints.", "best_pr_reason": null, "soft_edge_verdicts": [ { "left": "issue:44961", "right": "issue:45468", "accept": false, - "reason": "No overlap beyond being generic bug reports; one is a placeholder title, the other is a Gemma-4 audio encoding issue." + "reason": "Unrelated model-specific issues; no shared code path or symptom." }, { "left": "issue:44273", "right": "issue:45542", "accept": false, - "reason": "Different problems: lazy loading vs a TensorBoard/TensorFlow backend error." + "reason": "Both are v5-era regressions, but one is lazy loading and the other is a TensorBoard/TF backend error." }, { "left": "issue:44077", "right": "issue:44861", "accept": false, - "reason": "Different failure modes in different code paths; one is a model option policy issue, the other is a tied-weights AttributeError." + "reason": "Different problems in config validation vs tied-weight key handling." }, { "left": "issue:43867", "right": "issue:47110", "accept": false, - "reason": "Unrelated symptoms: checkpoint loading error vs special_tokens list/dict handling." + "reason": "Both are attribute errors, but on different objects and code paths." }, { "left": "issue:45381", "right": "issue:46664", "accept": false, - "reason": "Both concern multimodal/video behavior, but they describe different model families and different bugs." + "reason": "Both concern vision/video inputs, but the concrete bugs are different." }, { "left": "issue:44623", "right": "issue:45230", "accept": false, - "reason": "Too vague on the right side; no evidence these are the same save_pretrained problem." + "reason": "One is a specific save_pretrained bug; the other is too vague to establish sameness." }, { "left": "issue:44273", "right": "issue:45003", "accept": false, - "reason": "Both touch loading/import behavior, but one is lazy loading and the other is unsafe sys.modules access." + "reason": "Related to loading/lazy import behavior, but the reports describe different failures." }, { "left": "issue:44485", "right": "issue:46664", "accept": false, - "reason": "Different model-specific implementation issues: RoPE math vs video frame semantics." + "reason": "Different model features and different bug descriptions." }, { "left": "issue:41669", "right": "issue:43673", "accept": false, - "reason": "Both are performance-related, but one is import-time cost and the other is missing cache during chunked prefill." + "reason": "Import-performance issue versus missing cache during chunked prefill; not the same bug." }, { "left": "issue:44466", "right": "issue:44556", "accept": false, - "reason": "Different serialization/load failures with no shared concrete code path." + "reason": "Both involve checkpoint serialization/loading, but the symptoms and failure modes differ." }, { "left": "issue:43525", "right": "issue:45850", "accept": false, - "reason": "Specific missing config attribute vs a broad complaint that many models break in v5." + "reason": "A single config attribute error is not the same as a broad complaint that many models break." }, { "left": "issue:43572", "right": "issue:45850", "accept": false, - "reason": "Specific StableLmConfig regression vs a generic v5 breakage report." + "reason": "StableLmConfig pad-token mismatch is a specific model bug, not the same as a general breakage report." }, { "left": "issue:44393", "right": "issue:45753", "accept": false, - "reason": "Both are Qwen vision issues, but one is 2D bbox output and the other is an infinite loop on a specific image." + "reason": "Same general model family, but distinct symptoms and inputs." }, { "left": "issue:44393", "right": "issue:45071", "accept": false, - "reason": "Different scope: model output bug vs framework-wide PretrainedConfig type checking breakage." + "reason": "Model output bug versus a framework-wide type-checking regression." }, { "left": "issue:44556", "right": "issue:45092", "accept": false, - "reason": "Both involve v5 upgrade compatibility, but they affect different checkpoint families and different mechanisms." + "reason": "Both are upgrade/loading issues, but they affect different checkpoint and initialization paths." }, { "left": "issue:42371", "right": "issue:44492", "accept": false, - "reason": "TF32 configuration guidance is unrelated to a typo in cache strategy docs/code." + "reason": "Completely different topics: TF32 settings versus cache-strategy typo." }, { "left": "issue:45850", "right": "issue:47059", "accept": false, - "reason": "A broad v5 regression complaint is not the same as a CLI completion break caused by huggingface_hub." + "reason": "Generic v5 breakage versus a CLI compatibility failure with huggingface_hub." }, { "left": "issue:44246", "right": "issue:44273", "accept": false, - "reason": "Related to import/loading performance, but one is slow imports and the other is lazy loading correctness." + "reason": "Potentially related to import/lazy loading, but one is performance and the other is a broader lazy-loading bug report." }, { "left": "issue:44655", "right": "issue:44861", "accept": false, - "reason": "Saving Pipeline objects and tied-weight key handling are separate save/load bugs." + "reason": "Pipeline save_pretrained failure is not the same as tied-weight key attribute handling." }, { "left": "issue:43408", "right": "issue:44315", "accept": false, - "reason": "Different subsystems entirely: model type warning vs Liger Kernel application during model_init." + "reason": "Different model subsystems and unrelated failure modes." }, { "left": "issue:44492", "right": "issue:45850", "accept": false, - "reason": "A typo fix in cache strategies does not match a broad compatibility complaint." + "reason": "A typo in cache-strategy naming is unrelated to broad v5 regressions." }, { "left": "issue:41669", "right": "issue:46218", "accept": false, - "reason": "Import performance regression vs missing exported model class; different root causes." + "reason": "Both mention import-related behavior, but one is performance and the other is a missing import symbol." }, { - "left": "issue:43976", - "right": "issue:44393", + "left": "issue:44492", + "right": "issue:46328", "accept": false, - "reason": "Python version compatibility and a Qwen3-VL bbox bug are unrelated." + "reason": "Typo in cache strategies versus a slow padding implementation; no shared bug." }, { - "left": "issue:44492", - "right": "issue:46328", + "left": "issue:43976", + "right": "issue:44393", "accept": false, - "reason": "Cache-strategy typo and SequenceFeatureExtractor padding inefficiency are unrelated." + "reason": "Version/platform compatibility issue versus a specific Qwen3-VL output bug." }, { "left": "issue:44295", "right": "issue:45230", "accept": false, - "reason": "Specific position_ids buffer access bug vs an unspecified bug report." + "reason": "Buffer access bug is concrete; the other report is too generic to match it." }, { "left": "issue:42890", "right": "issue:44908", "accept": false, - "reason": "Test seed omissions and scheduler kwargs handling are unrelated issues." + "reason": "Test seeding flakiness and scheduler kwargs handling are unrelated." }, { "left": "issue:44297", "right": "issue:45230", "accept": false, - "reason": "Tokenizer class serialization mismatch is not the same as a generic bug report." + "reason": "Tokenizer save metadata mismatch is a specific bug, not a generic report." }, { "left": "issue:44704", "right": "issue:44908", "accept": false, - "reason": "Different APIs and failures: processor kwargs forwarding vs scheduler kwargs propagation." + "reason": "AutoProcessor kwargs forwarding and scheduler kwargs forwarding are unrelated code paths." }, { "left": "issue:44908", "right": "issue:45071", "accept": false, - "reason": "Scheduler configuration bug vs framework-wide config type-checking breakage." + "reason": "Scheduler configuration bug versus framework type-checking regression." }, { - "left": "issue:43576", - "right": "issue:44908", + "left": "issue:41669", + "right": "issue:45341", "accept": false, - "reason": "CLI env command breakage is unrelated to scheduler kwargs handling." + "reason": "Import-time slowdown is unrelated to the testing_utils bug." }, { - "left": "issue:41669", - "right": "issue:45341", + "left": "issue:43576", + "right": "issue:44908", "accept": false, - "reason": "Both mention performance/testing only tangentially; no same bug or code path." + "reason": "CLI env command issue and scheduler kwargs issue are unrelated." }, { "left": "issue:43576", "right": "issue:47176", "accept": false, - "reason": "CLI command breakage and shell allowlist bypass are different security/CLI issues." + "reason": "Environment command breakage is not the same as a shell-command allowlist bypass." }, { "left": "issue:41669", "right": "issue:44351", "accept": false, - "reason": "Import-time slowdown and missing HybridCache export are different problems." + "reason": "Import-time slowdown versus a missing HybridCache symbol; different symptoms and fixes." }, { "left": "issue:46018", "right": "issue:46172", "accept": false, - "reason": "Same model family, but one is gating logic and the other is a tensor transpose/dtype bug." + "reason": "DeepSeek-V4 gating/architecture issue versus a tensor transpose bug." }, { "left": "issue:43976", "right": "issue:46218", "accept": false, - "reason": "Python compatibility issue vs missing Qwen3_5 import; unrelated." + "reason": "Python-version compatibility problem versus missing import name; not duplicates." }, { "left": "issue:43976", "right": "issue:47176", "accept": false, - "reason": "Python version support and shell allowlist bypass are unrelated." + "reason": "Unrelated: interpreter/version support versus shell operator security behavior." }, { - "left": "issue:31795", - "right": "issue:33453", + "left": "issue:43298", + "right": "issue:46095", "accept": false, - "reason": "Documentation confusion and tokenizer loading regression do not describe the same bug." + "reason": "Continuous batching support for audio models is not the same as a DeepSeek-V4 weight-conversion question." } ] }, @@ -19663,19 +19570,20 @@ "source": "llm", "cache_hit": false, "model": "gpt-5.4-mini?service_tier=flex", - "cluster_id": "cluster-10000-529", + "cluster_id": "cluster-10000-530", "nodes": [ "issue:28348", "issue:29870", + "issue:30333", "issue:31515", "issue:31795", + "issue:33453", "issue:34689", "issue:38617", "issue:41669", "issue:42371", "issue:42913", "issue:43066", - "issue:43298", "issue:43352", "issue:43502", "issue:43519", @@ -19705,7 +19613,6 @@ "issue:45468", "issue:45656", "issue:45850", - "issue:46095", "issue:46218", "issue:46655", "issue:46894", @@ -19715,12 +19622,12 @@ "issue:47177" ], "soft_pairs": [ - "issue:43298|issue:46095", + "issue:31795|issue:33453", "issue:43519|issue:46655", "issue:44315|issue:44829", "issue:44485|issue:47176", - "issue:45468|issue:46894", "issue:45230|issue:45656", + "issue:45468|issue:46894", "issue:44360|issue:45468", "issue:43976|issue:47177", "issue:43576|issue:44273", @@ -19734,10 +19641,10 @@ "issue:43576|issue:46218", "issue:42913|issue:43066", "issue:44485|issue:45468", - "issue:45850|issue:47177", - "issue:45850|issue:47176", "issue:43704|issue:44485", "issue:42371|issue:43704", + "issue:45850|issue:47177", + "issue:45850|issue:47176", "issue:43352|issue:47177", "issue:43606|issue:45468", "issue:43502|issue:43519", @@ -19759,306 +19666,319 @@ "issue:43618|issue:44998", "issue:44230|issue:45310", "issue:31515|issue:38617", - "issue:29870|issue:31515" + "issue:29870|issue:31515", + "issue:30333|issue:31515" ], - "prepared_review_unit_hash": "370f9a7ce060edf243ac9e8b2a768b087f3b6307c48dd16571b48cddd73d83d5", + "prepared_review_unit_hash": "43b8eb55a9400a7325edad12264dff861bc08163b673489dab97ae291391d965", "component_budget": { - "node_count": 529, - "item_count": 529, - "soft_pair_count": 1953, - "serialized_chars": 1012074, - "estimated_input_tokens": 253019, - "estimated_eval_tokens": 506294 + "node_count": 530, + "item_count": 530, + "soft_pair_count": 1954, + "serialized_chars": 1013004, + "estimated_input_tokens": 253251, + "estimated_eval_tokens": 506758 }, "budget": { "node_count": 48, "item_count": 48, - "soft_pair_count": 45, - "serialized_chars": 40436, - "estimated_input_tokens": 10109, - "estimated_eval_tokens": 20474 + "soft_pair_count": 46, + "serialized_chars": 40661, + "estimated_input_tokens": 10166, + "estimated_eval_tokens": 20588 }, "overflow_policy": "truncate_then_skip", "trimmed": false, "aggressively_trimmed": false, "split": true, "analyst_result": { - "summary": "The soft-pair list is dominated by false positives: most matches only share broad terms like Flash Attention, processor, tokenizer, or loading, but they point to different subsystems and failure modes. I would not merge any of the candidate edges. If a single issue must be chosen as the best representative, issue 43519 is the most concrete and well-scoped bug report in the set.", - "confidence": 0.76, - "canonical_issue_reason": "Issue 43519 is a concrete, narrowly scoped processor bug with a clear failure description, making it a better representative than the broader or noisier candidates.", + "summary": "Mostly unrelated Transformers issues; the only near-match is the tokenizer v4/v5 regression pair, but it is still too broad/specifc to merge safely. All soft edges should be rejected.", + "confidence": 0.74, + "canonical_issue_reason": "issue:42913 is the clearest broad issue in the set and the best representative of the tokenizer-v5 regression theme, though it is still not a safe duplicate match with issue:43066.", "canonical_pr_reason": null, - "best_issue_reason": "Issue 43519 is the strongest standalone issue here: specific repro context, clear symptom, and less ambiguity than the generic or loosely phrased reports.", + "best_issue_reason": "issue:42913 has the most general, well-framed problem statement among the items and is the best issue to anchor this mini-cluster around.", "best_pr_reason": null, "soft_edge_verdicts": [ { - "left": "issue:43298", - "right": "issue:46095", + "left": "issue:31795", + "right": "issue:33453", "accept": false, - "reason": "Different problems: continuous batching for audio models vs a DeepSeek weight-conversion/documentation question." + "reason": "Docs/forward-arg confusion vs tokenizer loading regression; different bugs." }, { "left": "issue:43519", "right": "issue:46655", "accept": false, - "reason": "Both are processor bugs, but one is a Qwen3VL timestamp calculation issue and the other is a SmolVLM video padding bug; different code paths." + "reason": "Both are processor math issues, but timestamp calculation and video padding are different code paths." }, { "left": "issue:44315", "right": "issue:44829", "accept": false, - "reason": "Model_init/Liger Kernel application bug is unrelated to a flash_attention_3 training degradation issue." + "reason": "model_init/Liger-kernel application bug vs flash_attention_3 training degeneration; unrelated." }, { "left": "issue:44485", "right": "issue:47176", "accept": false, - "reason": "GLM-5 RoPE implementation and shell command substitution allowlist bypass are unrelated domains and vulnerabilities." + "reason": "RoPE implementation discussion vs shell-command allowlist bypass; different domains." }, { - "left": "issue:45468", - "right": "issue:46894", + "left": "issue:45230", + "right": "issue:45656", "accept": false, - "reason": "Gemma-4 audio positional encoding and CohereASR double-shift training loss are different model bugs, not one underlying defect." + "reason": "Generic bug report vs deepspeed optimizer-step double-call; no concrete overlap." }, { - "left": "issue:45230", - "right": "issue:45656", + "left": "issue:45468", + "right": "issue:46894", "accept": false, - "reason": "A generic bug report does not match an optimizer-step-called-twice deepspeed issue." + "reason": "Gemma audio positional encoding vs CohereASR training-loss double-shift; different model/code paths." }, { "left": "issue:44360", "right": "issue:45468", "accept": false, - "reason": "A DSA indexer/ReLU discussion is not the same as Gemma-4 audio positional encoding." + "reason": "DSA indexer/ReLU discussion vs audio positional encoding bug; unrelated." }, { "left": "issue:43976", "right": "issue:47177", "accept": false, - "reason": "Python version compatibility for Transformers 5.1.0 is unrelated to a tar/zip-slip arbitrary file write RCE." + "reason": "Python version compatibility vs tar/zip-slip RCE; unrelated." }, { "left": "issue:43576", "right": "issue:44273", "accept": false, - "reason": "Broken env CLI behavior and lazy loading problems are different features with different failure modes." + "reason": "Broken env command vs lazy loading malfunction; different features." }, { "left": "issue:43519", "right": "issue:44485", "accept": false, - "reason": "Timestamp calculation in Qwen3VL processor is unrelated to GLM-5 RoPE implementation." + "reason": "Timestamp math in Qwen3VL processor vs GLM-5 RoPE implementation; not the same bug." }, { "left": "issue:44485", "right": "issue:47177", "accept": false, - "reason": "RoPE math in model code is unrelated to archive path traversal / file-write security issues." + "reason": "RoPE implementation vs archive-slip security issue; unrelated." }, { "left": "issue:44861", "right": "issue:47110", "accept": false, - "reason": "A tied-weights helper AttributeError is not the same as a special_tokens list-vs-dict error." + "reason": "Tied-weights helper crash vs special_tokens dict/list error; different failure modes." }, { "left": "issue:44273", "right": "issue:44908", "accept": false, - "reason": "Lazy loading and inverse_sqrt scheduler kwargs are unrelated subsystems." + "reason": "Lazy loading issue vs inverse_sqrt scheduler kwargs bug; unrelated." }, { "left": "issue:31795", "right": "issue:34689", "accept": false, - "reason": "Documentation confusion about input_ids/past_key_values is not the same as a Llama 3.2 Vision model loading regression." + "reason": "Forward signature documentation vs Llama 3.2 model loading regression; different issues." }, { "left": "issue:43519", "right": "issue:44908", "accept": false, - "reason": "Processor timestamp math and scheduler kwargs handling are separate bug classes." + "reason": "Processor timestamp bug vs scheduler kwargs omission; unrelated." }, { "left": "issue:41669", "right": "issue:43704", "accept": false, - "reason": "Import-star cleanup/performance concerns do not match a VRAM leak in Qwen3ForCausalLM." + "reason": "import * refactor/perf issue vs VRAM leak in multi-threaded dataloading; unrelated." }, { "left": "issue:43576", "right": "issue:46218", "accept": false, - "reason": "CLI env command breakage is unrelated to a module import failure for Qwen3_5ForCausalLM." + "reason": "Env CLI breakage vs module import failure for Qwen3_5; different code paths." }, { "left": "issue:42913", "right": "issue:43066", "accept": false, - "reason": "General tokenizer behavior differences from v4 to v5 are broader than a specific wrong decoder type bug." + "reason": "Both concern tokenizer-v5 behavior, but one is broad regression reporting and the other is a specific decoder-type bug; not the same concrete issue." }, { "left": "issue:44485", "right": "issue:45468", "accept": false, - "reason": "RoPE implementation in GLM-5 is not the same defect as Gemma-4 audio positional encoding." + "reason": "GLM-5 RoPE implementation vs Gemma audio positional encoding; different models and bugs." }, { - "left": "issue:45850", - "right": "issue:47177", + "left": "issue:43704", + "right": "issue:44485", "accept": false, - "reason": "A broad 'breaks too many models' report is not the same as an archive extraction RCE." + "reason": "VRAM leak in threaded dataloading vs RoPE implementation; unrelated." }, { - "left": "issue:45850", - "right": "issue:47176", + "left": "issue:42371", + "right": "issue:43704", "accept": false, - "reason": "A broad version-regression complaint is unrelated to a shell operator allowlist bypass." + "reason": "TF32 API/settings issue vs VRAM leak issue; unrelated." }, { - "left": "issue:43704", - "right": "issue:44485", + "left": "issue:45850", + "right": "issue:47177", "accept": false, - "reason": "A VRAM leak in Qwen3ForCausalLM is unrelated to GLM-5 RoPE implementation." + "reason": "Broad model-breakage complaint vs archive-slip RCE; unrelated." }, { - "left": "issue:42371", - "right": "issue:43704", + "left": "issue:45850", + "right": "issue:47176", "accept": false, - "reason": "TF32 API behavior guidance is unrelated to a VRAM leak bug." + "reason": "Broad model-breakage complaint vs shell-operator allowlist bypass; unrelated." }, { "left": "issue:43352", "right": "issue:47177", "accept": false, - "reason": "Flash Attention 2.0 support for NemotronHForCausalLM is unrelated to tar/zip-slip file write RCE." + "reason": "Flash Attention 2.0 support gap vs archive-slip RCE; unrelated." }, { "left": "issue:43606", "right": "issue:45468", "accept": false, - "reason": "CPU offload device mismatch in bark-small is unrelated to Gemma-4 audio positional encoding." + "reason": "CPU offload device mismatch vs Gemma audio positional encoding; unrelated." }, { "left": "issue:43502", "right": "issue:43519", "accept": false, - "reason": "Unexpected network requests with local_files_only=True are unrelated to Qwen3VL timestamp calculations." + "reason": "local_files_only network leak vs timestamp calculation bug; different behaviors." }, { "left": "issue:44861", "right": "issue:47059", "accept": false, - "reason": "Tied weights key handling and a CLI completion attribute error are different bugs." + "reason": "_tied_weights_keys crash vs CLI completion API breakage; unrelated." }, { "left": "issue:43519", "right": "issue:44623", "accept": false, - "reason": "Processor timestamp math is unrelated to missing processor save_pretrained files." + "reason": "Timestamp calculation vs processor.save_pretrained missing files; different processor bugs." }, { "left": "issue:45850", "right": "issue:46218", "accept": false, - "reason": "A broad regression complaint does not match a specific module import failure." + "reason": "Generic model-breakage complaint vs Qwen3_5 import failure; not the same bug." + }, + { + "left": "issue:44230", + "right": "issue:45468", + "accept": false, + "reason": "fp8 inference support request vs audio positional encoding bug; unrelated." }, { "left": "issue:44263", "right": "issue:45468", "accept": false, - "reason": "torch.split return values in a DSA indexer are unrelated to Gemma-4 audio positional encoding." + "reason": "torch.split return-value issue in indexer vs audio positional encoding; unrelated." }, { "left": "issue:43704", "right": "issue:45341", "accept": false, - "reason": "VRAM leakage in a model is unrelated to a testing_utils.py bug." + "reason": "VRAM leak vs a small testing_utils bug; unrelated." }, { "left": "issue:28348", "right": "issue:45250", "accept": false, - "reason": "GPT2LMHeadModel FA2 support is a concrete model request; the other issue is only a generic Flash Attention 2.0 topic." + "reason": "GPT2-specific Flash Attention 2.0 support request vs generic Flash Attention 2.0 issue; too vague to merge." }, { "left": "issue:43352", "right": "issue:46218", "accept": false, - "reason": "Flash Attention 2.0 support for NemotronHForCausalLM is unrelated to a module import error for Qwen3_5ForCausalLM." + "reason": "Flash Attention support gap vs module import failure; unrelated." }, { "left": "issue:44351", "right": "issue:46218", "accept": false, - "reason": "HybridCache import failure and Qwen3_5ForCausalLM import failure are separate missing-symbol issues." + "reason": "HybridCache import failure vs Qwen3_5 module import failure; different missing symbols." }, { "left": "issue:44861", "right": "issue:45290", "accept": false, - "reason": "Tied-weight key lookup and apply_chat_template tool-call crashes are different parts of the library." + "reason": "Tied-weight helper crash vs chat-template/tool-call crash; unrelated." }, { "left": "issue:44230", "right": "issue:44623", "accept": false, - "reason": "FP8 Qwen3-VL/Qwen3.5 MoE support is unrelated to processor save_pretrained missing files." + "reason": "fp8 support request vs missing save_pretrained files; unrelated." }, { "left": "issue:44908", "right": "issue:44998", "accept": false, - "reason": "A scheduler kwargs bug is unrelated to a nonspecific 'Unemployment' issue." + "reason": "Scheduler kwargs bug vs unrelated low-information issue title." }, { "left": "issue:44998", "right": "issue:47177", "accept": false, - "reason": "A placeholder-style issue title does not match a tar/zip-slip RCE." + "reason": "Unemployment vs archive-slip RCE; unrelated." }, { "left": "issue:44998", "right": "issue:47176", "accept": false, - "reason": "A placeholder-style issue title does not match a shell operator allowlist bypass." + "reason": "Unemployment vs shell-command bypass; unrelated." }, { "left": "issue:28348", "right": "issue:43352", "accept": false, - "reason": "Both mention Flash Attention 2.0, but they target different models and separate support gaps." + "reason": "Both mention Flash Attention 2.0, but one is GPT2-specific support and the other is NemotronH support; not the same concrete change." }, { "left": "issue:44492", "right": "issue:44998", "accept": false, - "reason": "A cache-strategy typo is unrelated to the unrelated placeholder issue." + "reason": "Cache-strategy typo vs unrelated low-information issue." }, { "left": "issue:43618", "right": "issue:44998", "accept": false, - "reason": "CLIPOutput attentions assignment bug is unrelated to the placeholder issue." + "reason": "CLIPOutput attentions regression vs unrelated low-information issue." }, { "left": "issue:44230", "right": "issue:45310", "accept": false, - "reason": "FP8 support for Qwen3-VL/Qwen3.5 MoE is not the same as a from_pretrained error in Transformers>=5.4.0." + "reason": "fp8 support request vs Qwen3.5 from_pretrained error; different problems." }, { "left": "issue:31515", "right": "issue:38617", "accept": false, - "reason": "Slow from_pretrained checkpoint loading and an import failure for layer_type_validation are unrelated problems." + "reason": "Slow checkpoint loading vs missing layer_type_validation import; unrelated." }, { "left": "issue:29870", "right": "issue:31515", "accept": false, - "reason": "BART generate() decoder_attention_mask bug is unrelated to slow checkpoint loading." + "reason": "BART generate decoder_attention_mask error vs slow from_pretrained loading; unrelated." + }, + { + "left": "issue:30333", + "right": "issue:31515", + "accept": false, + "reason": "MLFlow run-finalization bug vs slow checkpoint loading; unrelated." } ] }, @@ -20080,9 +20000,8 @@ "source": "llm", "cache_hit": false, "model": "gpt-5.4-mini?service_tier=flex", - "cluster_id": "cluster-10000-529", + "cluster_id": "cluster-10000-530", "nodes": [ - "issue:30333", "issue:31515", "issue:31795", "issue:36296", @@ -20099,7 +20018,6 @@ "issue:46144" ], "soft_pairs": [ - "issue:30333|issue:31515", "issue:31515|issue:36296", "issue:43298|issue:44230", "issue:31515|issue:31795", @@ -20116,130 +20034,124 @@ "issue:37934|issue:43668", "issue:42936|issue:43668" ], - "prepared_review_unit_hash": "3d5bb4154f1c900382075b3c74a954bc3fe1f68f8a4db998eac8c6e77d523b7b", + "prepared_review_unit_hash": "ff5ba67badca7ed56f5b5c877be2abca170aec350ceb534f9d66ab5d51192e85", "component_budget": { - "node_count": 529, - "item_count": 529, - "soft_pair_count": 1953, - "serialized_chars": 1012074, - "estimated_input_tokens": 253019, - "estimated_eval_tokens": 506294 + "node_count": 530, + "item_count": 530, + "soft_pair_count": 1954, + "serialized_chars": 1013004, + "estimated_input_tokens": 253251, + "estimated_eval_tokens": 506758 }, "budget": { - "node_count": 15, - "item_count": 15, - "soft_pair_count": 16, - "serialized_chars": 13149, - "estimated_input_tokens": 3288, - "estimated_eval_tokens": 6832 + "node_count": 14, + "item_count": 14, + "soft_pair_count": 15, + "serialized_chars": 12258, + "estimated_input_tokens": 3065, + "estimated_eval_tokens": 6386 }, "overflow_policy": "truncate_then_skip", "trimmed": false, "aggressively_trimmed": false, "split": true, "analyst_result": { - "summary": "This cluster is a heterogeneous grab-bag of unrelated issues spanning MLflow reporting, checkpoint loading, docs, tensor parallelism, norms, loss plumbing, tokenizer warnings, and model-specific bugs. I found no true duplicate pairs or mergeable PR candidates.", - "confidence": 0.98, - "canonical_issue_reason": null, + "summary": "This set is highly heterogeneous: loading/performance, docs, config typing, tokenizer warnings, architecture quirks, and unrelated model bugs. None of the soft pairs look like the same underlying issue/change, so no merges are justified.", + "confidence": 0.9, + "canonical_issue_reason": "If one issue must represent the set, #46144 is the clearest actionable code-path bug and the only open issue; the rest are largely unrelated docs/typo/config/model-specific reports.", "canonical_pr_reason": null, - "best_issue_reason": null, + "best_issue_reason": "#46144 is the most concrete and merge-worthy issue in the list: it points to an implementation problem in attention wiring rather than a doc typo or narrow configuration nit.", "best_pr_reason": null, "soft_edge_verdicts": [ - { - "left": "issue:30333", - "right": "issue:31515", - "accept": false, - "reason": "MLflow reporting after failed training jobs is unrelated to slow checkpoint loading." - }, { "left": "issue:31515", "right": "issue:36296", "accept": false, - "reason": "Checkpoint load performance and tensor-parallel training bugs are different failure modes." + "reason": "Unrelated symptoms: checkpoint loading speed vs tensor-parallel training correctness." }, { "left": "issue:43298", "right": "issue:44230", "accept": false, - "reason": "Audio-model continuous batching support is unrelated to fp8 Qwen VL/MoE inference support." + "reason": "Audio continuous batching support and fp8 inference support are different feature gaps." }, { "left": "issue:31515", "right": "issue:31795", "accept": false, - "reason": "Loading-speed bug and documentation confusion about forward arguments are not the same issue." + "reason": "Performance/loading bug vs documentation confusion; not the same underlying problem." }, { "left": "issue:43717", "right": "issue:46144", "accept": false, - "reason": "Mamba init-weights behavior and RoFormer attention-interface usage affect different code paths." + "reason": "Mamba init-weight behavior is unrelated to RoFormer attention interface usage." }, { "left": "issue:42936", "right": "issue:46018", "accept": false, - "reason": "Mask2former ignore_value handling and DeepSeek-V4 shared-expert gating are unrelated model bugs." + "reason": "Mask2Former ignore_value handling and DeepSeek shared-expert gating are different model bugs." }, { "left": "issue:43668", "right": "issue:44031", "accept": false, - "reason": "A type-hint bug in ModernBERTConfig is unrelated to tokenizer regex warnings." + "reason": "Incorrect type hint in config vs tokenizer regex warning after a version change." }, { "left": "issue:42936", "right": "issue:43240", "accept": false, - "reason": "Unused ignore_value and missing kwargs in loss calculation are different defects." + "reason": "Unused ignore_value and kwargs not forwarded to loss calculation are separate code paths." }, { "left": "issue:43240", "right": "issue:46144", "accept": false, - "reason": "Loss-kwargs plumbing and RoFormer attention interface are not the same underlying problem." + "reason": "Loss-function kwargs forwarding is unrelated to RoFormer attention implementation details." }, { "left": "issue:42936", "right": "issue:44031", "accept": false, - "reason": "Mask2former ignore_value handling does not match tokenizer regex warning behavior." + "reason": "Model loss handling and tokenizer warning behavior do not share the same bug." }, { "left": "issue:42936", "right": "issue:46144", "accept": false, - "reason": "Mask2former ignore_value usage and RoFormer attention interface are unrelated." + "reason": "Different subsystems: segmentation loss config vs attention interface wiring." }, { "left": "issue:42936", "right": "issue:43813", "accept": false, - "reason": "Unused ignore_value and a typo in quantization_operation target different code paths." + "reason": "Mask2Former loss handling is unrelated to the quantization-operation typo." }, { "left": "issue:43668", "right": "issue:43813", "accept": false, - "reason": "Incorrect type hinting and a naming typo are unrelated fixes." + "reason": "Both are small metadata/doc-style issues, but they concern different symbols and fixes." }, { "left": "issue:43240", "right": "issue:43668", "accept": false, - "reason": "Loss kwargs propagation and a config type-hint issue are not duplicates." + "reason": "Loss kwargs propagation and config type-hint correctness are unrelated." }, { "left": "issue:37934", "right": "issue:43668", "accept": false, - "reason": "RMS-norm behavior question for Llama4TextL2Norm is unrelated to a ModernBERT type-hint bug." + "reason": "Llama4 norm-implementation question is unrelated to a ModernBERT type hint." }, { "left": "issue:42936", "right": "issue:43668", "accept": false, - "reason": "Mask2former ignore_value handling and ModernBERT norm_eps typing are different issues." + "reason": "Mask2Former loss usage bug and ModernBERT config typing bug are distinct." } ] }, @@ -20272,12 +20184,12 @@ "issue:10519|issue:11005", "issue:11005|issue:11307" ], - "prepared_review_unit_hash": "159cba91932fd731a245d219c3c72fb29704e27015539a98d2b30281897589c3", + "prepared_review_unit_hash": "eec0657430f517550f3c0d745bce124452098fa7893dfcf361ab59b0b982370b", "component_budget": { "node_count": 3, "item_count": 3, "soft_pair_count": 3, - "serialized_chars": 2876, + "serialized_chars": 2874, "estimated_input_tokens": 719, "estimated_eval_tokens": 1694 }, @@ -20285,7 +20197,7 @@ "node_count": 3, "item_count": 3, "soft_pair_count": 3, - "serialized_chars": 2876, + "serialized_chars": 2874, "estimated_input_tokens": 719, "estimated_eval_tokens": 1694 }, @@ -20294,30 +20206,30 @@ "aggressively_trimmed": false, "split": false, "analyst_result": { - "summary": "The three issues are about unrelated feature requests: tokenizer truncation behavior, ReduceLROnPlateau-like scheduling, and Wav2Vec2 word time offsets. They should not be clustered as duplicates.", + "summary": "These three issues are unrelated: tokenizer truncation behavior, a ReduceLROnPlateau-like training feature, and Wav2Vec2 word boundary timestamps. None of the soft pairs describe the same underlying bug or change.", "confidence": 0.98, - "canonical_issue_reason": null, - "canonical_pr_reason": null, - "best_issue_reason": "No single issue is a good canonical representative because the topics are unrelated; if forced, issue 10519 is the clearest self-contained feature request.", - "best_pr_reason": null, + "canonical_issue_reason": "No single canonical issue fits this cluster because the issues cover three different feature requests in different parts of the library.", + "canonical_pr_reason": "No pull requests were provided.", + "best_issue_reason": "issue:10519 is the clearest, most concrete feature request and is the best representative issue if one must be chosen.", + "best_pr_reason": "No pull requests were provided.", "soft_edge_verdicts": [ { "left": "issue:10519", "right": "issue:11307", "accept": false, - "reason": "Different problem domains: truncation strategy for tokenization vs word-level time offsets in Wav2Vec2." + "reason": "Different domains and requests: tokenizer truncation control vs. Wav2Vec2 word timestamp extraction." }, { "left": "issue:10519", "right": "issue:11005", "accept": false, - "reason": "One is about truncation direction in preprocessing; the other requests learning-rate scheduler functionality." + "reason": "Truncation direction for tokenization is unrelated to a ReduceLROnPlateau-like learning-rate scheduling feature." }, { "left": "issue:11005", "right": "issue:11307", "accept": false, - "reason": "Unrelated requests: ReduceLROnPlateau-like training utility vs extracting word boundary timestamps from audio models." + "reason": "A learning-rate scheduler request and Wav2Vec2 word offset timestamps are unrelated issues." } ] }, @@ -20349,12 +20261,12 @@ "issue:13244|issue:16618", "issue:15980|issue:16618" ], - "prepared_review_unit_hash": "f8ac9aa92910f6284f1dca378100d6e124f7d35bf29386cc10052b1fd3638056", + "prepared_review_unit_hash": "0669f6e8c060e1b60675c739899f0111098ee18772844d6344f7c9e66866810f", "component_budget": { "node_count": 3, "item_count": 3, "soft_pair_count": 2, - "serialized_chars": 2479, + "serialized_chars": 2478, "estimated_input_tokens": 620, "estimated_eval_tokens": 1496 }, @@ -20362,7 +20274,7 @@ "node_count": 3, "item_count": 3, "soft_pair_count": 2, - "serialized_chars": 2479, + "serialized_chars": 2478, "estimated_input_tokens": 620, "estimated_eval_tokens": 1496 }, @@ -20371,24 +20283,24 @@ "aggressively_trimmed": false, "split": false, "analyst_result": { - "summary": "These three issues do not look like duplicates of the same bug. Two are about model-loading/authentication/error handling, while one is about Tapas tokenization behavior and is unrelated.", - "confidence": 0.92, - "canonical_issue_reason": "No clear duplicate cluster exists here; if one issue must be selected, #16618 is the broadest user-facing failure report, but it is not the same underlying problem as the others.", + "summary": "The cluster is mostly a false-positive mix: one tokenization issue and two model-loading issues. The two model-loading reports are related at a broad subsystem level, but they do not clearly describe the same underlying bug.", + "confidence": 0.58, + "canonical_issue_reason": "Issue 16618 is the broadest and most central of the set, with the clearest model-loading failure wording and the highest discussion activity.", "canonical_pr_reason": null, - "best_issue_reason": "#16618 is the most general and likely to attract the most relevant fix discussion for a model-loading failure, but it should not absorb the other two issues as duplicates.", + "best_issue_reason": "16618 is the best representative issue for this cluster because it anchors the only apparent theme shared by multiple items: model loading failures.", "best_pr_reason": null, "soft_edge_verdicts": [ { "left": "issue:13244", "right": "issue:16618", "accept": false, - "reason": "Tapas tokenization mismatch is a different code path and symptom than failing to load bert-base-uncased; no shared underlying bug is evident." + "reason": "Tapas tokenization versus a model-loading failure are different bugs and different code paths; the similarity is only superficial." }, { "left": "issue:15980", "right": "issue:16618", "accept": false, - "reason": "Both involve model download/load failures, but #15980 is specifically about the error shown when unauthenticated on a private model, while #16618 is a separate public-model load failure report." + "reason": "Both concern model loading, but 15980 is specifically about the error message when downloading a private model without login, while 16618 is a generic 'can't load the model' report; not the same underlying issue." } ] }, @@ -20420,7 +20332,7 @@ "issue:19631|issue:22372", "issue:19631|issue:20362" ], - "prepared_review_unit_hash": "117353bdc4c21234f0c867d61854408538284fbb4a554647eee0bb247f4f784b", + "prepared_review_unit_hash": "0727ee12fa81236531394837f8f78dc03ae7f509fc72d716e83c161750b4bcf7", "component_budget": { "node_count": 3, "item_count": 3, @@ -20442,24 +20354,24 @@ "aggressively_trimmed": false, "split": false, "analyst_result": { - "summary": "These are three distinct feature-request issues for adding different models (EDSR/MDSR, FlexiBERT, Restormer), not duplicates of the same underlying request.", - "confidence": 0.97, + "summary": "These are three distinct feature requests to add different models (EDSR/MDSR, FlexiBERT, Restormer). They share only a broad \u201cadd model\u201d theme and are not the same underlying issue.", + "confidence": 0.98, "canonical_issue_reason": null, "canonical_pr_reason": null, - "best_issue_reason": "issue:19631 is the broadest and most discussed of the three, so it is the best cluster representative if one is needed; however, it is not a duplicate of the others.", + "best_issue_reason": null, "best_pr_reason": null, "soft_edge_verdicts": [ { "left": "issue:19631", "right": "issue:22372", "accept": false, - "reason": "Different model requests: EDSR/MDSR vs Restormer. Same general 'add a model' theme, but not the same feature or bug." + "reason": "Different requested models in different domains: EDSR/MDSR are super-resolution networks, while Restormer is a separate transformer-based restoration model. No evidence they refer to the same implementation or code-path change." }, { "left": "issue:19631", "right": "issue:20362", "accept": false, - "reason": "Different model requests: EDSR/MDSR vs FlexiBERT. They are unrelated additions, not the same underlying change." + "reason": "EDSR/MDSR and FlexiBERT are unrelated model additions. They do not describe the same bug or feature, only the same generic request type ('add a model')." } ] }, @@ -20489,7 +20401,7 @@ "soft_pairs": [ "issue:26350|issue:33260" ], - "prepared_review_unit_hash": "c3ea2312327a2f34e84bbdaae4da00effac2e399654940c3bb47a41b065f52fc", + "prepared_review_unit_hash": "c8f851321ff55e74104f6d809151ae6b40021f75fb707228398384fb254f214d", "component_budget": { "node_count": 2, "item_count": 2, @@ -20511,18 +20423,18 @@ "aggressively_trimmed": false, "split": false, "analyst_result": { - "summary": "These are two separate community feature requests for different capabilities: Flash Attention 2 support vs GGUF support. They share a broad \u201cmore architectures/support\u201d theme, but not the same underlying change or bug.", + "summary": "These are unrelated feature-request issues: one for Flash Attention 2 support, the other for GGUF support. They share only a generic \"more architectures\" theme and should not be clustered as duplicates.", "confidence": 0.98, - "canonical_issue_reason": null, + "canonical_issue_reason": "Issue #26350 is a reasonable representative only because it is the earlier and more heavily discussed request, but it is not a duplicate of #33260.", "canonical_pr_reason": null, - "best_issue_reason": "Issue #26350 is the better cluster representative only by age and discussion activity, but it is not a canonical duplicate target because it concerns a different feature than #33260.", + "best_issue_reason": "#26350 is the better standalone issue to keep as the cluster representative due to earlier creation and higher discussion activity.", "best_pr_reason": null, "soft_edge_verdicts": [ { "left": "issue:26350", "right": "issue:33260", "accept": false, - "reason": "Different requested features and code paths; same high-level support theme is too broad to treat as duplicates." + "reason": "They request different capabilities in different areas (Flash Attention 2 vs GGUF) and do not describe the same underlying bug or change." } ] }, @@ -20603,286 +20515,286 @@ "issue:46226|issue:47299", "issue:46931|issue:47299" ], - "prepared_review_unit_hash": "72898d7bf1d812c5b13099dd78603c918940950d8160cf40454964134a60db0d", + "prepared_review_unit_hash": "827aa168085b5677e0f0f29712f81a2a6ff346fff2758e8bb153bea9174737c8", "component_budget": { "node_count": 12, "item_count": 12, "soft_pair_count": 42, - "serialized_chars": 22277, - "estimated_input_tokens": 5570, - "estimated_eval_tokens": 11396 + "serialized_chars": 22276, + "estimated_input_tokens": 5569, + "estimated_eval_tokens": 11394 }, "budget": { "node_count": 12, "item_count": 12, "soft_pair_count": 42, - "serialized_chars": 22277, - "estimated_input_tokens": 5570, - "estimated_eval_tokens": 11396 + "serialized_chars": 22276, + "estimated_input_tokens": 5569, + "estimated_eval_tokens": 11394 }, "overflow_policy": "truncate_then_skip", "trimmed": false, "aggressively_trimmed": false, "split": false, "analyst_result": { - "summary": "This cluster is mostly unrelated i18n translation requests that share a common template, so the high similarity is superficial. They should not be deduplicated as the language targets and some titles/content differ materially; the odd placeholder/spam-like issues are also not the same underlying task.", - "confidence": 0.96, - "canonical_issue_reason": "issue:43834 is the most central template-style i18n issue and has the strongest textual overlap with the other placeholder entries, but it still is not a true duplicate of the concrete language-specific requests.", + "summary": "This cluster is mostly composed of language-specific documentation translation requests, plus a few garbled/spam-like entries. The titles/body patterns are highly similar, but they do not indicate the same underlying request, so the soft links should not be merged as duplicates.", + "confidence": 0.88, + "canonical_issue_reason": "Issue 43834 is the best cluster exemplar because it is the generic translation-request template that most closely matches the repeated pattern across the set, even though the language-specific requests still appear distinct.", "canonical_pr_reason": null, - "best_issue_reason": "issue:43834 is the cleanest representative of the repeated translation-request template, but the cluster spans distinct languages and should remain split.", + "best_issue_reason": "Issue 43834 is the clearest representative of the cluster's repeated template and is the most central example of the translation-request pattern.", "best_pr_reason": null, "soft_edge_verdicts": [ { "left": "issue:27088", "right": "issue:44370", "accept": false, - "reason": "Different translation targets; same i18n template, but not the same underlying request." + "reason": "Different translation requests; no evidence they target the same language or same underlying change." }, { "left": "issue:27088", "right": "issue:44365", "accept": false, - "reason": "Different translation targets; only the template text overlaps." + "reason": "Different translation requests; same template style is not enough to treat them as duplicates." }, { "left": "issue:27088", "right": "issue:43834", "accept": false, - "reason": "Superficially similar template, but Turkish vs. placeholder/template issue context makes it a different request." + "reason": "Concrete Turkish request vs generic template issue; not the same underlying request." }, { "left": "issue:27088", "right": "issue:30191", "accept": false, - "reason": "Turkish docs translation is separate from Persian docs translation." + "reason": "Turkish vs Persian translation request; distinct language targets." }, { "left": "issue:27088", "right": "issue:47299", "accept": false, - "reason": "The titles/content do not indicate the same language or same concrete translation task." + "reason": "The latter looks garbled/spam-like, not the same docs translation request." }, { "left": "issue:28936", "right": "issue:43834", "accept": false, - "reason": "Spanish translation request is distinct from the placeholder/template i18n issue." + "reason": "Spanish request vs generic template placeholder; likely separate language-specific issue." }, { "left": "issue:27088", "right": "issue:28936", "accept": false, - "reason": "Different target languages: Turkish vs Spanish." + "reason": "Turkish vs Spanish translation request; distinct targets, not duplicates." }, { "left": "issue:28936", "right": "issue:44365", "accept": false, - "reason": "Different language request; similarity is just the shared translation template." + "reason": "Different language requests; similar wording is just the shared template." }, { "left": "issue:28936", "right": "issue:30240", "accept": false, - "reason": "Spanish docs translation is separate from Polish docs translation." + "reason": "Spanish vs Polish translation request; different underlying requests." }, { "left": "issue:28936", "right": "issue:38435", "accept": false, - "reason": "Spanish and Romanian translation requests are different tasks." + "reason": "Spanish vs Romanian translation request; distinct language targets." }, { "left": "issue:28936", "right": "issue:30191", "accept": false, - "reason": "Spanish docs translation is not the same as Persian docs translation." + "reason": "Spanish vs Persian translation request; not the same issue." }, { "left": "issue:30191", "right": "issue:30240", "accept": false, - "reason": "Persian and Polish translation requests are different issues." + "reason": "Persian vs Polish translation request; same template, different concrete request." }, { "left": "issue:30240", "right": "issue:44370", "accept": false, - "reason": "Polish translation request is distinct from the placeholder/template i18n issue." + "reason": "Polish request vs generic placeholder issue; no concrete shared target." }, { "left": "issue:30240", "right": "issue:44365", "accept": false, - "reason": "Polish docs translation is a separate language request." + "reason": "Same translation template, but not enough evidence of the same language/request." }, { "left": "issue:30240", "right": "issue:43834", "accept": false, - "reason": "Different concrete translation task; only the template is shared." + "reason": "Polish request vs generic template issue; not the same concrete change." }, { "left": "issue:30240", "right": "issue:47299", "accept": false, - "reason": "No evidence these refer to the same language or same request." + "reason": "The latter is garbled/spam-like and does not match the Polish translation request." }, { "left": "issue:27088", "right": "issue:30240", "accept": false, - "reason": "Turkish and Polish translation requests are different." + "reason": "Turkish vs Polish translation request; separate language targets." }, { "left": "issue:38435", "right": "issue:44370", "accept": false, - "reason": "Romanian translation request is distinct from the placeholder/template issue." + "reason": "Romanian request vs generic template placeholder; no concrete duplicate evidence." }, { "left": "issue:38435", "right": "issue:44365", "accept": false, - "reason": "Romanian translation request is not the same as the placeholder/template i18n issue." + "reason": "Romanian request vs generic template placeholder; likely separate issues." }, { "left": "issue:38435", "right": "issue:43834", "accept": false, - "reason": "Different concrete language request; similarity is only structural." + "reason": "Romanian request vs generic template issue; same theme, different concrete request." }, { "left": "issue:38435", "right": "issue:47299", "accept": false, - "reason": "Romanian translation request does not match the unrelated placeholder/spam-like title." + "reason": "Garbled/spam-like issue does not look like the Romanian translation request." }, { "left": "issue:38435", "right": "issue:46226", "accept": false, - "reason": "Romanian translation request is unrelated to the Thai/spam-like issue." + "reason": "Romanian translation request vs garbled Thai text; unrelated underlying content." }, { "left": "issue:43834", "right": "issue:44370", "accept": false, - "reason": "These look like separate template instances, not the same translation task." + "reason": "Both are template-like translation requests, but there is no evidence they are the same language-specific issue." }, { "left": "issue:43834", "right": "issue:44365", "accept": false, - "reason": "Template overlap is high, but no evidence they are the same concrete request." + "reason": "Near-identical template text, but no concrete shared target language or request." }, { "left": "issue:43834", "right": "issue:47299", "accept": false, - "reason": "The titles suggest different content and no shared concrete translation target." + "reason": "Generic translation template vs spam/garbled title; not the same request." }, { "left": "issue:43834", "right": "issue:46226", "accept": false, - "reason": "Placeholder/template issue and Thai text issue are not the same request." + "reason": "Template-like translation issue vs garbled Thai content; no shared concrete bug/change." }, { "left": "issue:43834", "right": "issue:46931", "accept": false, - "reason": "No indication these are the same underlying issue; titles/content differ." + "reason": "Template-like translation issue vs unrelated Thai text; not duplicates." }, { "left": "issue:44365", "right": "issue:44370", "accept": false, - "reason": "These are separate template-like translation requests, not one concrete bug/change." + "reason": "Titles/body look templated, but no explicit target language or concrete identical request." }, { "left": "issue:44365", "right": "issue:47299", "accept": false, - "reason": "No evidence of a shared language or same request; only template similarity." + "reason": "Generic translation template vs spam/garbled content; different underlying issue." }, { "left": "issue:44365", "right": "issue:46226", "accept": false, - "reason": "Template-style i18n issue and Thai/spam-like issue are unrelated." + "reason": "Template-like translation issue vs garbled Thai text; no shared concrete request." }, { "left": "issue:44365", "right": "issue:46931", "accept": false, - "reason": "Different titles/content; not the same underlying request." + "reason": "Same template family, but the Thai text is unrelated to docs translation." }, { "left": "issue:44370", "right": "issue:47299", "accept": false, - "reason": "These are not the same translation task; similarity is superficial." + "reason": "Generic translation template vs spam/garbled content; not the same issue." }, { "left": "issue:44370", "right": "issue:46226", "accept": false, - "reason": "Different content and no shared concrete translation target." + "reason": "Template-like translation issue vs garbled Thai content; no concrete overlap." }, { "left": "issue:44370", "right": "issue:46931", "accept": false, - "reason": "No evidence they refer to the same underlying issue." + "reason": "Template-like translation issue vs unrelated Thai text; not duplicates." }, { "left": "issue:46222", "right": "issue:47299", "accept": false, - "reason": "Both are malformed/spam-like, but they do not look like the same concrete request." + "reason": "Both are garbled/spam-like, but there is no evidence they are the same concrete request." }, { "left": "issue:44370", "right": "issue:46222", "accept": false, - "reason": "Template translation issue vs. Thai text issue; not the same underlying change." + "reason": "Generic translation template vs garbled Thai content; different underlying issues." }, { "left": "issue:44365", "right": "issue:46222", "accept": false, - "reason": "These look like different issues; shared similarity is not enough." + "reason": "Generic translation template vs garbled Thai content; not the same request." }, { "left": "issue:43834", "right": "issue:46222", "accept": false, - "reason": "Placeholder/template issue and Thai text issue are not the same task." + "reason": "Template-like translation issue vs garbled Thai content; no shared concrete target." }, { "left": "issue:46222", "right": "issue:46226", "accept": false, - "reason": "Two different Thai-titled issues; no evidence they are the same request." + "reason": "Both are garbled Thai text, but there is no concrete evidence they describe the same change or bug." }, { "left": "issue:46226", "right": "issue:46931", "accept": false, - "reason": "Different Thai-titled issues with no shared concrete code-path or task." + "reason": "Garbled Thai content in both, but not enough evidence of the same underlying request." }, { "left": "issue:46226", "right": "issue:47299", "accept": false, - "reason": "These titles appear unrelated; not the same underlying issue." + "reason": "Garbled/spam-like entries only; no clear same concrete issue." }, { "left": "issue:46931", "right": "issue:47299", "accept": false, - "reason": "No strong evidence these are the same request; titles/content differ." + "reason": "Unrelated garbled titles; no evidence of the same underlying bug or change." } ] }, @@ -20986,220 +20898,226 @@ "issue:42503|issue:46653", "issue:45027|issue:46653" ], - "prepared_review_unit_hash": "8dc1b7b148c294ac56953c9f39cf52b69e81c1af495568286c41b69a1cd3d0f2", + "prepared_review_unit_hash": "8db3fe801a57badc5ad401c5cd878422fea1d89e64e367bb4065b03b2de8c90f", "component_budget": { "node_count": 22, "item_count": 22, "soft_pair_count": 55, - "serialized_chars": 31680, - "estimated_input_tokens": 7920, - "estimated_eval_tokens": 16096 + "serialized_chars": 31689, + "estimated_input_tokens": 7923, + "estimated_eval_tokens": 16102 }, "budget": { "node_count": 22, "item_count": 22, "soft_pair_count": 55, - "serialized_chars": 31680, - "estimated_input_tokens": 7920, - "estimated_eval_tokens": 16096 + "serialized_chars": 31689, + "estimated_input_tokens": 7923, + "estimated_eval_tokens": 16102 }, "overflow_policy": "truncate_then_skip", "trimmed": false, "aggressively_trimmed": false, "split": false, "analyst_result": { - "summary": "This cluster is not a duplicate set: it contains many separate model-request issues across unrelated models, modalities, and features. The apparent overlap is mostly superficial keyword similarity (e.g., \"add model\", \"support\", BERT/audio/vision terms), not the same underlying change.", - "confidence": 0.98, - "canonical_issue_reason": "No canonical issue: the items are heterogeneous model requests and do not share one underlying bug or change.", + "summary": "This is a loose topic cluster of unrelated model-request issues (vision, audio, language, detectors, matchers, and config additions). The similarities are mostly generic \u201cAdd model\u201d wording, so the soft edges should all be rejected.", + "confidence": 0.96, + "canonical_issue_reason": "No single canonical issue: these are different requested models and features across unrelated code paths and modalities, not duplicates of one underlying bug/change.", "canonical_pr_reason": null, - "best_issue_reason": "No single best issue for deduping: none of these issues is a clear superset or duplicate of the others.", + "best_issue_reason": "If a representative issue is needed, 42503 is the closest to a generic model-add request, but it still does not cover the rest of the cluster well.", "best_pr_reason": null, "soft_edge_verdicts": [ { "left": "issue:36718", "right": "issue:47004", "accept": false, - "reason": "RoMa keypoint matcher and LoMa keypoint detector+matcher are different model requests and not the same code change." + "reason": "Different keypoint projects (RoMa matcher vs LoMa detector+matcher); same domain, but not the same requested change." }, { "left": "issue:36718", "right": "issue:41115", "accept": false, - "reason": "Keypoint matching versus MiniCPM3 architecture support are unrelated domains." + "reason": "RoMa matcher is a vision matching request; MiniCPM3 is a model architecture request in a different area." }, { "left": "issue:36817", "right": "issue:41211", "accept": false, - "reason": "EuroBert config and DEIMv2 support are distinct model additions." + "reason": "EuroBert config support and DEIMv2 are different model additions." }, { "left": "issue:36817", "right": "issue:43671", "accept": false, - "reason": "EuroBert config and Qwen3-TTS support are unrelated requests." + "reason": "EuroBert config support is unrelated to Qwen3-TTS speech support." }, { "left": "issue:36817", "right": "issue:36879", "accept": false, - "reason": "EuroBert config and RF-DETR are different model integrations." + "reason": "EuroBert and RF-DETR are distinct models with different integration work." }, { "left": "issue:36817", "right": "issue:42954", "accept": false, - "reason": "EuroBert config and XiaoMi MiMo are unrelated model requests." + "reason": "EuroBert config work is unrelated to XiaoMi MiMo." }, { "left": "issue:36879", "right": "issue:41211", "accept": false, - "reason": "RF-DETR and DEIMv2 are different detector models, not one shared fix." + "reason": "RF-DETR vision detection and DEIMv2 are separate model requests." }, { "left": "issue:36879", "right": "issue:43909", "accept": false, - "reason": "RF-DETR vision model support is unrelated to LFM2.5 Audio support." + "reason": "RF-DETR and LFM2.5 Audio target different modalities and code paths." }, { "left": "issue:36879", "right": "issue:44121", "accept": false, - "reason": "RF-DETR model addition and a weight-sparse transformer request are not the same change." + "reason": "RF-DETR detection support is unrelated to the weight-sparse transformer request." }, { "left": "issue:36879", "right": "issue:43837", "accept": false, - "reason": "RF-DETR and Qwen3-ASR are different model families and code paths." + "reason": "RF-DETR vision detection and Qwen3-ASR speech recognition are different changes." }, { "left": "issue:36879", "right": "issue:43671", "accept": false, - "reason": "RF-DETR vision support and Qwen3-TTS support are unrelated." + "reason": "RF-DETR and Qwen3-TTS are different tasks and model families." }, { "left": "issue:37015", "right": "issue:43837", "accept": false, - "reason": "NeoBERT and Qwen3-ASR are separate model requests." + "reason": "NeoBERT is a language model request; Qwen3-ASR is an audio transcription request." }, { "left": "issue:37015", "right": "issue:44121", "accept": false, - "reason": "NeoBERT support and OpenAI weight-sparse transformer support are different changes." + "reason": "NeoBERT and the OpenAI weight-sparse transformer are unrelated model additions." }, { "left": "issue:37015", "right": "issue:43909", "accept": false, - "reason": "NeoBERT and LFM2.5 Audio are unrelated model additions." + "reason": "NeoBERT and LFM2.5 Audio 1.5B are different model families in different modalities." }, { "left": "issue:37015", "right": "issue:37334", "accept": false, - "reason": "NeoBERT and LLMDet target different model families and features." + "reason": "NeoBERT and LLMDet are distinct requests; one is a BERT-like LM, the other a detector." }, { "left": "issue:36879", "right": "issue:37015", "accept": false, - "reason": "RF-DETR and NeoBERT are unrelated model requests." + "reason": "RF-DETR detection support is not the same change as NeoBERT support." }, { "left": "issue:37334", "right": "issue:41115", "accept": false, - "reason": "LLMDet support and MiniCPM3 architecture support are different concrete changes." + "reason": "LLMDet support and MiniCPM3 architecture are separate model requests." }, { "left": "issue:37334", "right": "issue:43837", "accept": false, - "reason": "LLMDet and Qwen3-ASR are unrelated model support tickets." + "reason": "LLMDet detection support is unrelated to Qwen3-ASR speech recognition." }, { "left": "issue:37334", "right": "issue:43671", "accept": false, - "reason": "LLMDet and Qwen3-TTS are different model requests." + "reason": "LLMDet and Qwen3-TTS are different tasks and model code paths." }, { "left": "issue:37334", "right": "issue:42503", "accept": false, - "reason": "LLMDet and ModernVBERT are separate model additions." + "reason": "LLMDet and ModernVBERT are different model additions, not one duplicated request." }, { "left": "issue:41115", "right": "issue:43837", "accept": false, - "reason": "MiniCPM3 architecture support and Qwen3-ASR are unrelated requests." + "reason": "MiniCPM3 architecture work and Qwen3-ASR are unrelated requests." }, { "left": "issue:41115", "right": "issue:44121", "accept": false, - "reason": "MiniCPM3 support and weight-sparse transformer support are not the same code-path change." + "reason": "MiniCPM3 and the weight-sparse transformer are different models and changes." }, { "left": "issue:41115", "right": "issue:42503", "accept": false, - "reason": "MiniCPM3 and ModernVBERT are different model additions." + "reason": "MiniCPM3 support is not the same as adding ModernVBERT models." }, { "left": "issue:41211", "right": "issue:43671", "accept": false, - "reason": "DEIMv2 detection support and Qwen3-TTS support are unrelated." + "reason": "DEIMv2 detection support and Qwen3-TTS are unrelated." }, { "left": "issue:37015", "right": "issue:42503", "accept": false, - "reason": "NeoBERT and ModernVBERT are distinct model requests." + "reason": "NeoBERT and ModernVBERT are different BERT-like model requests, not duplicates." }, { "left": "issue:42503", "right": "issue:43837", "accept": false, - "reason": "ModernVBERT and Qwen3-ASR target different models and tasks." + "reason": "ModernVBERT model addition and Qwen3-ASR speech recognition are unrelated." }, { "left": "issue:42503", "right": "issue:43671", "accept": false, - "reason": "ModernVBERT and Qwen3-TTS are unrelated support requests." + "reason": "ModernVBERT and Qwen3-TTS are different model families and tasks." + }, + { + "left": "issue:42503", + "right": "issue:46653", + "accept": false, + "reason": "ModernVBERT model support has nothing to do with the Mirko Privitera issue." }, { "left": "issue:42738", "right": "issue:45027", "accept": false, - "reason": "BERT-like RoPE support is a broad architecture request, while Voxtral TTS is a specific model addition." + "reason": "BERT-like models with RoPE and Voxtral TTS are different feature requests." }, { "left": "issue:42503", "right": "issue:42738", "accept": false, - "reason": "ModernVBERT support is not the same as adding generic BERT-like RoPE support; the scope and code change differ." + "reason": "ModernVBERT models and BERT-like models with RoPE overlap only in broad language-model wording, not the same change." }, { "left": "issue:42738", "right": "issue:42954", "accept": false, - "reason": "Generic BERT-like RoPE support and XiaoMi MiMo are unrelated model requests." + "reason": "BERT-like RoPE support and XiaoMi MiMo are separate model requests." }, { "left": "issue:42738", "right": "issue:46653", "accept": false, - "reason": "BERT-like RoPE support and an issue titled with a personal name are not the same underlying change." + "reason": "BERT-like models with RoPE is unrelated to the Mirko Privitera issue." }, { "left": "issue:42954", @@ -21211,139 +21129,139 @@ "left": "issue:37334", "right": "issue:42954", "accept": false, - "reason": "LLMDet and XiaoMi MiMo are unrelated model requests." + "reason": "LLMDet and XiaoMi MiMo are different requested models and tasks." }, { "left": "issue:42954", "right": "issue:43671", "accept": false, - "reason": "XiaoMi MiMo and Qwen3-TTS are different models/tasks." + "reason": "XiaoMi MiMo and Qwen3-TTS are unrelated model requests." }, { "left": "issue:42503", "right": "issue:42954", "accept": false, - "reason": "ModernVBERT and XiaoMi MiMo are distinct model additions." + "reason": "ModernVBERT and XiaoMi MiMo are different models; only the request style is similar." }, { "left": "issue:42954", "right": "issue:45177", "accept": false, - "reason": "XiaoMi MiMo and DeepSeek-OCR2 are separate model requests." + "reason": "XiaoMi MiMo and DeepSeek-OCR2 are different requests across different model types." }, { "left": "issue:43671", "right": "issue:43837", "accept": false, - "reason": "Qwen3-TTS and Qwen3-ASR are different modalities and likely different code paths." + "reason": "Qwen3-TTS and Qwen3-ASR are related brands but different concrete tasks and implementations." }, { "left": "issue:43671", "right": "issue:45027", "accept": false, - "reason": "Qwen3-TTS and Voxtral-4B-TTS-2603 are different TTS models and not the same concrete change." + "reason": "Qwen3-TTS and Voxtral-4B-TTS-2603 are different TTS model requests." }, { "left": "issue:43671", - "right": "issue:45177", - "accept": false, - "reason": "Qwen3-TTS and DeepSeek-OCR2 are unrelated." - }, - { - "left": "issue:43837", - "right": "issue:43909", + "right": "issue:44121", "accept": false, - "reason": "Qwen3-ASR and LFM2.5 Audio are separate audio model requests." + "reason": "Qwen3-TTS support is unrelated to the weight-sparse transformer request." }, { - "left": "issue:43837", - "right": "issue:44121", + "left": "issue:43671", + "right": "issue:45322", "accept": false, - "reason": "Qwen3-ASR support and a weight-sparse transformer request are unrelated." + "reason": "Qwen3-TTS and EUPE are different model families and integration work." }, { "left": "issue:43837", - "right": "issue:45027", + "right": "issue:43909", "accept": false, - "reason": "ASR support and Voxtral TTS support are different tasks and code paths." + "reason": "Qwen3-ASR and LFM2.5 Audio are different audio model requests, not the same change." }, { "left": "issue:43837", - "right": "issue:45177", + "right": "issue:45322", "accept": false, - "reason": "Qwen3-ASR and DeepSeek-OCR2 are unrelated model additions." + "reason": "Qwen3-ASR and EUPE are unrelated." }, { "left": "issue:43837", - "right": "issue:45322", + "right": "issue:44121", "accept": false, - "reason": "Qwen3-ASR and EUPE are different model families." + "reason": "Qwen3-ASR support is not the same as the weight-sparse transformer change." }, { "left": "issue:43909", "right": "issue:44205", "accept": false, - "reason": "LFM2.5 Audio and SAM3-LiteText are unrelated model requests." + "reason": "LFM2.5 Audio and SAM3-LiteText are different modalities and code paths." }, { "left": "issue:43909", "right": "issue:46645", "accept": false, - "reason": "Adding LFM2.5 Audio support and adding SWA+sink to Granite are different changes." + "reason": "LFM2.5 Audio support is unrelated to the Granite SWA+sink option." }, { "left": "issue:44121", "right": "issue:45322", "accept": false, - "reason": "OpenAI weight-sparse transformer support and EUPE support are unrelated." + "reason": "Weight-sparse transformer support and EUPE are distinct model requests." }, { "left": "issue:41115", "right": "issue:45027", "accept": false, - "reason": "MiniCPM3 and Voxtral-4B-TTS-2603 are different model requests." + "reason": "MiniCPM3 architecture work and Voxtral-4B-TTS support are unrelated." + }, + { + "left": "issue:43837", + "right": "issue:45027", + "accept": false, + "reason": "Qwen3-ASR and Voxtral TTS are different speech models and tasks." }, { "left": "issue:37015", "right": "issue:45027", "accept": false, - "reason": "NeoBERT and Voxtral-4B-TTS-2603 are unrelated model additions." + "reason": "NeoBERT and Voxtral TTS are unrelated model requests." }, { "left": "issue:37334", "right": "issue:45027", "accept": false, - "reason": "LLMDet and Voxtral-4B-TTS-2603 are different model support tickets." + "reason": "LLMDet detection support is unrelated to Voxtral TTS." }, { "left": "issue:43837", "right": "issue:45177", "accept": false, - "reason": "Qwen3-ASR and DeepSeek-OCR2 are unrelated." + "reason": "Qwen3-ASR and DeepSeek-OCR2 are different modalities and changes." }, { "left": "issue:45322", "right": "issue:46133", "accept": false, - "reason": "EUPE and TIPSv2 are different vision-language model requests." + "reason": "EUPE and TIPSv2 are different model additions from different authors and tasks." }, { "left": "issue:42954", "right": "issue:46653", "accept": false, - "reason": "XiaoMi MiMo and a personal-name issue are not the same underlying bug or change." + "reason": "XiaoMi MiMo and Mirko Privitera are entirely different issues." }, { "left": "issue:42503", "right": "issue:46653", "accept": false, - "reason": "ModernVBERT and a personal-name issue are unrelated." + "reason": "ModernVBERT and Mirko Privitera are unrelated." }, { "left": "issue:45027", "right": "issue:46653", "accept": false, - "reason": "Voxtral TTS support and a personal-name issue are not the same request." + "reason": "Voxtral TTS support and Mirko Privitera are unrelated." } ] }, @@ -21373,7 +21291,7 @@ "soft_pairs": [ "issue:38740|issue:46443" ], - "prepared_review_unit_hash": "568f31e3595049afd24af8859e24cb3b4fcc5e36c7e00e2d177eccd954305d04", + "prepared_review_unit_hash": "6de9d99d6b1eb5931d8959cfbffa4643ebf387d7140e30e1509507fe76f4fba2", "component_budget": { "node_count": 2, "item_count": 2, @@ -21395,18 +21313,18 @@ "aggressively_trimmed": false, "split": false, "analyst_result": { - "summary": "These two issues are not duplicates: one is a documentation request about adding Pruna as an optimization framework, and the other is a technical Vision RoPE refactor. They address different areas and different kinds of work.", - "confidence": 0.97, - "canonical_issue_reason": "No canonical issue: the items are unrelated and do not describe the same bug, feature, or code change.", + "summary": "The only pair is a weak text-similarity match, but the issues are unrelated: one is a docs request to add Pruna, the other is a Vision RoPE refactor. This is not a duplicate cluster.", + "confidence": 0.98, + "canonical_issue_reason": "No canonical issue: the docs request and the Vision RoPE refactor are different topics and do not describe the same bug or change.", "canonical_pr_reason": null, - "best_issue_reason": "Issue #46443 is the better representative issue because it is the substantive technical refactor; #38740 is a narrower docs-only request.", + "best_issue_reason": "No global best issue: neither item can represent the cluster because they are unrelated; 46443 is the more technical task, but it is not a duplicate of 38740.", "best_pr_reason": null, "soft_edge_verdicts": [ { "left": "issue:38740", "right": "issue:46443", "accept": false, - "reason": "Different topics and work types: docs/optimization-framework mention versus Vision RoPE refactor. No shared code path or underlying bug/change." + "reason": "Different subjects and scopes: documentation about adding Pruna vs a Vision RoPE refactor. The similarity is only superficial, not the same underlying issue." } ] }, @@ -21436,12 +21354,12 @@ "soft_pairs": [ "issue:41084|issue:43037" ], - "prepared_review_unit_hash": "788c7ba656c5a57cc6c00e89b528829fa5b2fde254dd2e688c8079a107981b0b", + "prepared_review_unit_hash": "3708a8e6d459fd56de454109be97c96cb965650c75575db531343fbfb7f96193", "component_budget": { "node_count": 2, "item_count": 2, "soft_pair_count": 1, - "serialized_chars": 1480, + "serialized_chars": 1479, "estimated_input_tokens": 370, "estimated_eval_tokens": 996 }, @@ -21449,7 +21367,7 @@ "node_count": 2, "item_count": 2, "soft_pair_count": 1, - "serialized_chars": 1480, + "serialized_chars": 1479, "estimated_input_tokens": 370, "estimated_eval_tokens": 996 }, @@ -21458,18 +21376,18 @@ "aggressively_trimmed": false, "split": false, "analyst_result": { - "summary": "The two issues are not duplicates: one is about block decoding behavior, the other is about DeepSeek v3.2 model support.", - "confidence": 0.96, - "canonical_issue_reason": "No canonical issue; the titles point to different feature/problem areas with no evidence of a shared underlying bug.", + "summary": "The two issues are not duplicates: one is a generic request about block decoding, while the other is a model-specific support request for DeepSeek v3.2. They do not describe the same underlying bug or change.", + "confidence": 0.97, + "canonical_issue_reason": null, "canonical_pr_reason": null, - "best_issue_reason": "No best issue for deduplication: neither issue subsumes the other, so neither should be treated as the canonical duplicate target.", + "best_issue_reason": "Issue #43037 is the more specific and actionable request, but it is not a duplicate of #41084.", "best_pr_reason": null, "soft_edge_verdicts": [ { "left": "issue:41084", "right": "issue:43037", "accept": false, - "reason": "Different underlying requests/problems (block decoding vs model support); not the same bug or change." + "reason": "Different scope and intent: 'Set Block Decoding' is a generic decoding-related request, while 'DeepSeek v3.2 support' is about adding support for a specific model. No clear shared underlying bug/change." } ] }, @@ -21499,40 +21417,40 @@ "soft_pairs": [ "issue:41306|issue:41826" ], - "prepared_review_unit_hash": "13b83d45bda3035651a3773a8de022a2720e2c2eeba9acb3ce180cb58c464b64", + "prepared_review_unit_hash": "6a7d7147d9e00f448b8b0906e3837f80a4cf2f892eb031271aec73f3bfcdfff0", "component_budget": { "node_count": 2, "item_count": 2, "soft_pair_count": 1, - "serialized_chars": 1584, - "estimated_input_tokens": 396, - "estimated_eval_tokens": 1048 + "serialized_chars": 1585, + "estimated_input_tokens": 397, + "estimated_eval_tokens": 1050 }, "budget": { "node_count": 2, "item_count": 2, "soft_pair_count": 1, - "serialized_chars": 1584, - "estimated_input_tokens": 396, - "estimated_eval_tokens": 1048 + "serialized_chars": 1585, + "estimated_input_tokens": 397, + "estimated_eval_tokens": 1050 }, "overflow_policy": "truncate_then_skip", "trimmed": false, "aggressively_trimmed": false, "split": false, "analyst_result": { - "summary": "The two issues are about different feature requests: one for exposing a TiledFusedLogitsLoss/FLCE flag in `from_pretrained()`, the other for integrating TiledMLP to reduce memory footprint. They do not describe the same underlying bug or change.", - "confidence": 0.95, - "canonical_issue_reason": "Issue #41306 is the more concrete and narrowly scoped request, with a specific API change (`from_pretrained()` flag) and named implementation target.", + "summary": "These are not duplicates: one is a specific `from_pretrained()` flag request for TiledFusedLogitsLoss/FLCE, the other is a broader TiledMLP memory-footprint feature request. The shared \"Tiled\" theme is superficial.", + "confidence": 0.94, + "canonical_issue_reason": "Issue #41306 is the more concrete, narrowly scoped request tied to a specific code path (`from_pretrained()` behavior for TiledFusedLogitsLoss/FLCE).", "canonical_pr_reason": null, - "best_issue_reason": "#41306 is the best representative of the cluster because it is the most actionable and clearly defined issue, even though it is not a duplicate of #41826.", + "best_issue_reason": "Issue #41306 is the best representative because it is more actionable and specific; #41826 is a separate broader enhancement about TiledMLP memory usage.", "best_pr_reason": null, "soft_edge_verdicts": [ { "left": "issue:41306", "right": "issue:41826", "accept": false, - "reason": "Different requested changes: FLCE/TiledFusedLogitsLoss flag exposure vs TiledMLP integration for memory reduction. Same broad theme, but not the same code-path problem." + "reason": "Different underlying requests: one is about a `from_pretrained()` flag for FLCE/TiledFusedLogitsLoss, the other is about integrating TiledMLP for lower memory use. Similar terminology is not enough to treat them as the same bug or change." } ] }, @@ -21562,7 +21480,7 @@ "soft_pairs": [ "issue:43598|issue:43600" ], - "prepared_review_unit_hash": "5495164d0dc985a6d9f0b6f5ace2d99b2e97321928e3d154cc029337a0f094a3", + "prepared_review_unit_hash": "8cd29eef3765601e22e27df15865302114c194c5f674fcc96bf8ae483ad16d6e", "component_budget": { "node_count": 2, "item_count": 2, @@ -21584,18 +21502,18 @@ "aggressively_trimmed": false, "split": false, "analyst_result": { - "summary": "These are related only at a broad Trainer level, but they describe different problems: one is about `remove_unused_column` customizability, the other about FLOPS estimation for custom trainers. They should not be clustered as duplicates.", - "confidence": 0.96, - "canonical_issue_reason": null, + "summary": "These two issues are related to Trainer extensibility, but they describe different problems: one is about the `remove_unused_column` behavior, the other about FLOPS estimation for custom trainers. They should not be treated as duplicates.", + "confidence": 0.89, + "canonical_issue_reason": "Issue 43598 is more specific about a concrete Trainer customization pain point (`remove_unused_column`) and has a clearer actionable scope.", "canonical_pr_reason": null, - "best_issue_reason": null, + "best_issue_reason": "Issue 43598 is the better canonical issue because it targets a specific Trainer option and likely maps to a narrower, more directly fixable code path.", "best_pr_reason": null, "soft_edge_verdicts": [ { "left": "issue:43598", "right": "issue:43600", "accept": false, - "reason": "Same general Trainer subsystem, but the underlying bugs are different and would require different fixes: custom column removal behavior vs FLOPS counting for custom trainers." + "reason": "They share the broad theme of custom trainers, but the underlying bugs are different: one concerns unused-column removal, the other concerns FLOPS estimation logic. Not the same change or defect." } ] }, @@ -21625,7 +21543,7 @@ "soft_pairs": [ "issue:43684|issue:44405" ], - "prepared_review_unit_hash": "d7d852a184e3cb08c29ba41dcda19461c8a48d6fe07267dfb36e30c7667b08bf", + "prepared_review_unit_hash": "32f42af16420db0b029771312aaac1fb67234bb94899285d4a641a235bbd695a", "component_budget": { "node_count": 2, "item_count": 2, @@ -21647,18 +21565,18 @@ "aggressively_trimmed": false, "split": false, "analyst_result": { - "summary": "These are related family-of-model support requests, but they target different model versions and different AutoClass registrations, so they do not look like the same underlying issue.", - "confidence": 0.92, - "canonical_issue_reason": "Issue #43684 is the broader, more general request: adding Qwen3-Omni registration to core AutoModel/AutoModelForConditionalGeneration. It better represents the cluster than the narrower Qwen3.5 sequence-classification request.", + "summary": "These are two different feature requests in the same broad Transformers area, but they target different model families and different task heads. They should not be clustered as duplicates.", + "confidence": 0.96, + "canonical_issue_reason": "Issue #43684 is the broader and more central request: registering Qwen3-Omni with AutoModel and AutoModelForConditionalGeneration. It describes a concrete missing model registration path rather than a narrow task-head variant.", "canonical_pr_reason": null, - "best_issue_reason": "#43684 is the best issue to keep as the representative item because it covers the larger registration gap and is closer to an umbrella model-support request.", + "best_issue_reason": "#43684 is the best representative issue because it covers the main registration gap and is more general than the sequence-classification-only request in #44405.", "best_pr_reason": null, "soft_edge_verdicts": [ { "left": "issue:43684", "right": "issue:44405", "accept": false, - "reason": "Different targets: Qwen3-Omni AutoModel/AutoModelForConditionalGeneration registration versus Qwen3.5 AutoModelForSequenceClassification support. They involve different model variants and different downstream heads, so they are not the same bug/request." + "reason": "Different underlying requests: Qwen3-Omni model registration for AutoModel/AutoModelForConditionalGeneration vs Qwen3.5 sequence-classification support for AutoModelForSequenceClassification. Same general ecosystem, but not the same bug/change." } ] }, @@ -21695,12 +21613,12 @@ "issue:43845|issue:43859", "issue:43846|issue:43859" ], - "prepared_review_unit_hash": "21d6eb7b0846d2937eb676a917d25696139a88f8afecf0131d3f5a2475c9476b", + "prepared_review_unit_hash": "9342bb68cacc26fb4c306ff37d4710a1b4c2346d1feddd72197ba3e43d82a68b", "component_budget": { "node_count": 4, "item_count": 4, "soft_pair_count": 6, - "serialized_chars": 4263, + "serialized_chars": 4262, "estimated_input_tokens": 1066, "estimated_eval_tokens": 2388 }, @@ -21708,7 +21626,7 @@ "node_count": 4, "item_count": 4, "soft_pair_count": 6, - "serialized_chars": 4263, + "serialized_chars": 4262, "estimated_input_tokens": 1066, "estimated_eval_tokens": 2388 }, @@ -21717,54 +21635,57 @@ "aggressively_trimmed": false, "split": false, "analyst_result": { - "summary": "These are mostly content-free placeholder issues. The three \"huggingface\" issues look superficially similar, but there is not enough body detail to confirm they describe the same underlying bug; \"Fett\" is clearly separate by title.", - "confidence": 0.58, - "canonical_issue_reason": null, + "summary": "Issue 43845 remains the best canonical placeholder for the repeated low-information \"huggingface \" reports, but the pairwise evidence is too thin to confidently mark those items as true duplicates; they should be treated as likely similar/unresolved. \"Fett\" appears unrelated. No PRs are present.", + "confidence": 0.76, + "canonical_issue_reason": "issue:43845 is the earliest of the identically titled \"huggingface \" issues and is the most reasonable placeholder canonical among that cluster.", "canonical_pr_reason": null, - "best_issue_reason": "Issue 43845 is the earliest of the near-identical \"huggingface\" reports, so it is the best placeholder candidate if one must be chosen, but the evidence is too thin to treat it as a true canonical.", + "best_issue_reason": "issue:43845 is the best issue placeholder because it is one of the repeated identical-title reports and the earliest available member of that cluster.", "best_pr_reason": null, "soft_edge_verdicts": [ { "left": "issue:43835", "right": "issue:43845", "accept": false, - "reason": "Different titles and no substantive body detail tying them to the same report." + "reason": "Different titles and no body-level evidence tying \"Fett\" to the low-information \"huggingface \" report." }, { "left": "issue:43835", "right": "issue:43846", "accept": false, - "reason": "Different titles and no substantive body detail tying them to the same report." + "reason": "Different titles and no concrete shared problem details; only weak similarity signals are present." }, { "left": "issue:43835", "right": "issue:43859", "accept": false, - "reason": "Different titles and no substantive body detail tying them to the same report." + "reason": "Different titles and no concrete shared problem details; only weak similarity signals are present." }, { "left": "issue:43845", "right": "issue:43846", "accept": false, - "reason": "Same short title, but the issues are too empty to verify they are the same underlying problem." + "reason": "Same title fragment, but the bodies are unavailable and deterministic_accept is false, so this is not enough to confirm a duplicate; treat as likely similar/unresolved." }, { "left": "issue:43845", "right": "issue:43859", "accept": false, - "reason": "Same short title, but there is not enough description to confirm duplicate status." + "reason": "Same title fragment, but there is no body evidence and deterministic_accept is false; keep this unresolved rather than accepting a duplicate." }, { "left": "issue:43846", "right": "issue:43859", "accept": false, - "reason": "Same short title, but there is not enough description to confirm duplicate status." + "reason": "Same title fragment, but the bodies are unavailable and deterministic_accept is false, so duplicate status is unproven." } ] }, - "evaluator_result": null, - "evaluator_used": false, - "retried": false, + "evaluator_result": { + "accept": false, + "feedback": "The canonical placeholder choice is reasonable, but the soft-edge verdicts are too strong. In the packet, the three \"huggingface \" items only share title-level similarity and soft_similarity signals; the bodies are not available, and every pair is explicitly marked deterministic_accept=false. That is not enough to confidently accept them as duplicates. Please keep those verdicts conservative (e.g., likely similar / unresolved) unless you have stronger evidence from the issue bodies." + }, + "evaluator_used": true, + "retried": true, "accepted_nontrivial_soft_edge": false, "error_kind": null, "error_message": null @@ -21803,106 +21724,106 @@ "issue:44573|issue:47036", "issue:44573|issue:44937" ], - "prepared_review_unit_hash": "a3a8f7d2366807872fc3578bb8fdb5eb9b060c2c921d4a753690263bb0ad59e1", + "prepared_review_unit_hash": "5889c1884ae448147daae06cab5c976c92301c55c6f4c76446c21eca3c72c4d7", "component_budget": { "node_count": 6, "item_count": 6, "soft_pair_count": 12, - "serialized_chars": 7358, - "estimated_input_tokens": 1840, - "estimated_eval_tokens": 3936 + "serialized_chars": 7355, + "estimated_input_tokens": 1839, + "estimated_eval_tokens": 3934 }, "budget": { "node_count": 6, "item_count": 6, "soft_pair_count": 12, - "serialized_chars": 7358, - "estimated_input_tokens": 1840, - "estimated_eval_tokens": 3936 + "serialized_chars": 7355, + "estimated_input_tokens": 1839, + "estimated_eval_tokens": 3934 }, "overflow_policy": "truncate_then_skip", "trimmed": false, "aggressively_trimmed": false, "split": false, "analyst_result": { - "summary": "This cluster appears to be a grab bag of unrelated low-content/spam/off-topic issues, not true duplicates. The titles do not point to the same bug or feature, and there is no body/diff evidence to connect them.", - "confidence": 0.97, - "canonical_issue_reason": "Issue 47036 is the most substantive and descriptive item in the set ('Community benchmark tip') and has some discussion activity, so it is the best representative if one must be chosen.", + "summary": "These look like unrelated low-content/spam or vague issues, not duplicates of one underlying bug. No soft edge has enough concrete overlap to merge.", + "confidence": 0.93, + "canonical_issue_reason": null, "canonical_pr_reason": null, - "best_issue_reason": "Issue 47036 is the only issue with a clearly non-spam, actionable-sounding title; the others are generic spam placeholders or unrelated prompts.", + "best_issue_reason": "issue:47036 is the most specific/legible item in the set, but it is still too vague to serve as a true canonical duplicate target.", "best_pr_reason": null, "soft_edge_verdicts": [ { "left": "issue:44371", "right": "issue:44572", "accept": false, - "reason": "Both are generic spam placeholders, but there is no evidence they describe the same underlying issue." + "reason": "Both are spam-like titles with no concrete bug details; no shared underlying issue is visible." }, { "left": "issue:44371", "right": "issue:44573", "accept": false, - "reason": "Generic spam titles only; no concrete shared bug or request is visible." + "reason": "Both are spam-like titles with no actionable content, so they do not appear to describe the same bug." }, { "left": "issue:44371", "right": "issue:44418", "accept": false, - "reason": "'spam' vs a documentation suggestion are unrelated." + "reason": "One is a generic documentation suggestion, the other is spam; no duplicate bug/change can be inferred." }, { "left": "issue:44371", "right": "issue:47036", "accept": false, - "reason": "A spam placeholder is not the same as a community benchmark tip." + "reason": "These titles are unrelated and too vague to indicate the same underlying issue." }, { "left": "issue:44371", "right": "issue:44937", "accept": false, - "reason": "'spam' and 'Check out \"Google LLC\"' are different topics with no shared issue." + "reason": "Unrelated spam/generic titles with no shared concrete problem statement." }, { "left": "issue:44418", "right": "issue:44572", "accept": false, - "reason": "Documentation enhancement and spam placeholder are clearly different." + "reason": "Generic documentation enhancement versus spam; not the same issue." }, { "left": "issue:44418", "right": "issue:44573", "accept": false, - "reason": "Documentation enhancement and spam placeholder are clearly different." + "reason": "Generic documentation enhancement versus spam; no evidence of duplication." }, { "left": "issue:44572", "right": "issue:44573", "accept": false, - "reason": "Both are spam placeholders, but titles alone do not justify treating them as the same underlying report." + "reason": "Both are spam-like placeholders, but there is no concrete bug or change to match." }, { "left": "issue:44572", "right": "issue:47036", "accept": false, - "reason": "Spam placeholder vs community benchmark tip; no shared concrete problem." + "reason": "No shared concrete subject matter beyond vague, low-information titles." }, { "left": "issue:44572", "right": "issue:44937", "accept": false, - "reason": "Spam placeholder and 'Check out \"Google LLC\"' are unrelated." + "reason": "Unrelated vague/spam titles; no common underlying bug or request is visible." }, { "left": "issue:44573", "right": "issue:47036", "accept": false, - "reason": "Spam placeholder vs community benchmark tip; no concrete overlap." + "reason": "No concrete overlap; the titles do not indicate the same problem." }, { "left": "issue:44573", "right": "issue:44937", "accept": false, - "reason": "Spam placeholder and 'Check out \"Google LLC\"' are unrelated." + "reason": "The items are unrelated and too sparse to support a duplicate merge." } ] }, @@ -21932,7 +21853,7 @@ "soft_pairs": [ "issue:45145|issue:45151" ], - "prepared_review_unit_hash": "88f2f10f9cb5a39eee913aba50810f5e6e68cc94b7944a839b309e6cbd09872f", + "prepared_review_unit_hash": "efd101d63df9118bcc7b607f7948a293cda0771bc959c269bdd3530a184b5470", "component_budget": { "node_count": 2, "item_count": 2, @@ -21954,24 +21875,24 @@ "aggressively_trimmed": false, "split": false, "analyst_result": { - "summary": "The two closed issues are very similar and likely related, but the evidence is too weak to treat them as a confirmed duplicate pair. #45145 is still the best canonical issue because it is earlier and has slightly more linkage.", - "confidence": 0.79, - "canonical_issue_reason": "#45145 is the better canonical issue because it was opened first and has an inbound reference, giving it slightly more history/linkage than #45151.", + "summary": "The two closed issues are highly similar and likely related, but the evidence is limited to matching titles, similar body lengths, and weak soft-similarity signals. #45145 is still the best canonical issue because it is earlier and has the only inbound reference, but the duplicate call should be treated as tentative rather than certain.", + "confidence": 0.82, + "canonical_issue_reason": "#45145 is the earlier issue and has the only inbound reference, so it is the best canonical anchor if these are deduplicated.", "canonical_pr_reason": null, - "best_issue_reason": "#45145 is the best representative issue for the cluster: it matches #45151 very closely and is the more established entry.", + "best_issue_reason": "#45145 is the better issue to keep as the representative record because it predates #45151 and has stronger linkage, though the duplication evidence is not definitive.", "best_pr_reason": null, "soft_edge_verdicts": [ { "left": "issue:45145", "right": "issue:45151", "accept": false, - "reason": "The evidence only shows soft similarity plus the same title/body size. There are no explicit targets, filenames, or links tying them to the same underlying issue, so this should remain a possible duplicate rather than a confirmed one." + "reason": "The titles match and the body lengths are close, but we do not have body content, explicit links, or targets. That is enough to suspect a duplicate, not enough to confidently assert the same underlying report." } ] }, "evaluator_result": { "accept": false, - "feedback": "The canonical issue choice is broadly grounded, but the soft-edge verdict is too strong for the evidence. The packet only provides soft similarity plus same title/body size and no targets/files; that supports a possible duplicate, not a confident accept that they are the same underlying issue. Please keep the verdict conservative unless there is stronger linkage." + "feedback": "The canonical choice of issue #45145 is grounded: it is earlier and has the only inbound reference. However, the duplicate verdict is a bit too strong for the packet. We only know the titles match, body lengths are similar, and there are no explicit links/targets; we do not have body content, so claims like \"same report content\" or \"no meaningful differences\" overstate the evidence. The soft match should be treated as tentative rather than asserted as near-certain." }, "evaluator_used": true, "retried": true, @@ -22052,7 +21973,7 @@ "soft_pairs": [ "issue:45762|issue:45763" ], - "prepared_review_unit_hash": "efd62dc707795e4c836f630ad1d2ca145d867c7515c34eebe7de1a97ed200385", + "prepared_review_unit_hash": "561b1f6512395a8c741218878d48fe58ebb19510de4e0588e26e74ac22aacb50", "component_budget": { "node_count": 2, "item_count": 2, @@ -22074,18 +21995,18 @@ "aggressively_trimmed": false, "split": false, "analyst_result": { - "summary": "Both items look like unrelated placeholder/personal-name issues, not duplicates of the same underlying bug or change.", + "summary": "The cluster contains two unrelated closed issues with personal-name/date titles and no technical content; they do not appear to be duplicates of the same underlying bug or change.", "confidence": 0.97, - "canonical_issue_reason": null, + "canonical_issue_reason": "Issue 45762 is the earliest-created item and, with no substantive issue details on either side, it is the cleanest representative if a single issue must be chosen.", "canonical_pr_reason": null, - "best_issue_reason": null, + "best_issue_reason": "Issue 45762 is the best issue representative only by chronology; neither issue has enough technical signal to prefer one on content.", "best_pr_reason": null, "soft_edge_verdicts": [ { "left": "issue:45762", "right": "issue:45763", "accept": false, - "reason": "Titles are different person/date strings with no shared bug report, code area, or issue text; they do not appear to describe the same underlying problem." + "reason": "Different names and dates with no shared problem description, code path, or issue target; these do not look like the same underlying bug/change." } ] }, @@ -22115,7 +22036,7 @@ "soft_pairs": [ "issue:45764|issue:45765" ], - "prepared_review_unit_hash": "6c3702612214e32c420adb1cf87e9f4da5475ec655e2f82c52fbb92b66073af4", + "prepared_review_unit_hash": "15cf1939f3a92a71d710301805684c1ab8b708314a9cd820a26226a57ef5b7ea", "component_budget": { "node_count": 2, "item_count": 2, @@ -22137,18 +22058,18 @@ "aggressively_trimmed": false, "split": false, "analyst_result": { - "summary": "The two issues are not duplicates: they are just sparse titles containing different personal names and different dates, with no shared body, targets, or evidence of the same underlying problem.", - "confidence": 0.98, + "summary": "The two issues are unrelated low-content items with different personal-name/date titles and no shared bug or feature context, so they should not be clustered as duplicates.", + "confidence": 0.97, "canonical_issue_reason": null, "canonical_pr_reason": null, - "best_issue_reason": null, + "best_issue_reason": "Neither issue is a good canonical representative because neither contains actionable problem detail; if forced, no single issue stands out as the same underlying report.", "best_pr_reason": null, "soft_edge_verdicts": [ { "left": "issue:45764", "right": "issue:45765", "accept": false, - "reason": "Reject: the titles refer to different people and different dates, and there is no body or linkage showing the same issue." + "reason": "Different names and dates, no body text, no shared targets or evidence of the same underlying bug/change." } ] }, @@ -22178,40 +22099,40 @@ "soft_pairs": [ "issue:45810|issue:46295" ], - "prepared_review_unit_hash": "3d628360fb369f11eeda9fc7fa6aeb8ac99d354956188af18363e40fc307a6b3", + "prepared_review_unit_hash": "c533ef251bea22f1d7ba288209c105a5f7c92c05cc859d0aac8a331bcebf65e7", "component_budget": { "node_count": 2, "item_count": 2, "soft_pair_count": 1, - "serialized_chars": 1541, - "estimated_input_tokens": 386, - "estimated_eval_tokens": 1028 + "serialized_chars": 1540, + "estimated_input_tokens": 385, + "estimated_eval_tokens": 1026 }, "budget": { "node_count": 2, "item_count": 2, "soft_pair_count": 1, - "serialized_chars": 1541, - "estimated_input_tokens": 386, - "estimated_eval_tokens": 1028 + "serialized_chars": 1540, + "estimated_input_tokens": 385, + "estimated_eval_tokens": 1026 }, "overflow_policy": "truncate_then_skip", "trimmed": false, "aggressively_trimmed": false, "split": false, "analyst_result": { - "summary": "These two issues are about different requests: one asks for a new Qwen3_5 token-classification/value-model head, while the other asks for FP8 training support. They do not describe the same underlying bug or change.", - "confidence": 0.98, - "canonical_issue_reason": "No canonical issue: the requests are unrelated and should not be clustered as duplicates.", + "summary": "These two issues are not duplicates: one requests a new Qwen3_5 token-classification model class, while the other asks for FP8 training support. They concern different features and code paths.", + "confidence": 0.96, + "canonical_issue_reason": null, "canonical_pr_reason": null, - "best_issue_reason": "Neither issue is a duplicate of the other; if forced to pick a representative, issue 46295 is the broader infrastructure request, but it is not a canonical merge target for 45810.", + "best_issue_reason": "Issue 45810 is the more concrete, self-contained feature request, but it is not a substitute for 46295 and should not be treated as the same bug/change.", "best_pr_reason": null, "soft_edge_verdicts": [ { "left": "issue:45810", "right": "issue:46295", "accept": false, - "reason": "Reject: one is a feature request for adding a Qwen3_5 token-classification/value-model class, the other is a training capability request for FP8. Different code paths and different underlying problems." + "reason": "No same underlying bug or change: model-class addition for token classification vs. FP8 training support are separate capabilities with different implementation targets." } ] }, @@ -22251,12 +22172,12 @@ "issue:45871|issue:45873", "issue:45872|issue:45873" ], - "prepared_review_unit_hash": "30efee180686766f90e3d2e824b2f28ad10e5679a3741896c57368bb75f35d25", + "prepared_review_unit_hash": "9a006a6e12fce010532cb05154220f1e676b846f494b263e69043f5ffa211e3c", "component_budget": { "node_count": 5, "item_count": 5, "soft_pair_count": 8, - "serialized_chars": 5725, + "serialized_chars": 5726, "estimated_input_tokens": 1432, "estimated_eval_tokens": 3120 }, @@ -22264,7 +22185,7 @@ "node_count": 5, "item_count": 5, "soft_pair_count": 8, - "serialized_chars": 5725, + "serialized_chars": 5726, "estimated_input_tokens": 1432, "estimated_eval_tokens": 3120 }, @@ -22273,69 +22194,69 @@ "aggressively_trimmed": false, "split": false, "analyst_result": { - "summary": "This cluster is mixed: 45870/45871 are exact duplicate issues, 45872/45873 are another exact duplicate pair, and 45869 is a different dataset-promo topic. There isn\u2019t one perfect canonical for all five.", - "confidence": 0.93, - "canonical_issue_reason": "45870 is the earliest clean duplicate in the main DataForge example pair (exactly matches 45871) and is a reasonable representative for that subtopic.", + "summary": "There are two likely duplicate issue pairs: 45870/45871 and 45872/45873. The evidence supports duplicate triage, but not a claim of proven exact body matches. Issue 45869 appears unrelated to those pairs.", + "confidence": 0.88, + "canonical_issue_reason": "45870 is the earliest member of the strongest likely-duplicate pair with 45871, and it is a reasonable representative to keep for that request.", "canonical_pr_reason": null, - "best_issue_reason": "45870 is the best overall representative because it is a clear, early duplicate issue; the cluster is heterogeneous, but this is the most canonical-looking item among the exact duplicates.", + "best_issue_reason": "45870 is the strongest overall issue candidate to retain because it is part of a likely duplicate pair with the earliest timestamp, while 45869 is unrelated and the 45872/45873 pair is later.", "best_pr_reason": null, "soft_edge_verdicts": [ { "left": "issue:45869", "right": "issue:45870", "accept": false, - "reason": "Different topics: a financial data bundle promo versus a DataForge time-series example. Similar finance wording, but not the same underlying issue." + "reason": "Different requests: a dataset bundle versus a financial prediction example." }, { "left": "issue:45869", "right": "issue:45871", "accept": false, - "reason": "Different topics: dataset bundle promo versus DataForge example. No evidence they describe the same bug/change." + "reason": "Different requests: a dataset bundle versus a financial prediction example." }, { "left": "issue:45870", "right": "issue:45871", "accept": true, - "reason": "Same title and matching issue shape; these look like duplicate reports of the same DataForge example request." + "reason": "Same title and matching metadata strongly suggest the same underlying issue; likely duplicates." }, { "left": "issue:45870", "right": "issue:45872", "accept": false, - "reason": "Related to DataForge and finance, but one is an example issue and the other adds a scraping service; too different to treat as the same underlying request." + "reason": "Different request area: financial prediction example versus scraping-service integration." }, { "left": "issue:45870", "right": "issue:45873", "accept": false, - "reason": "Same as 45872: broader service/scraping topic, not the same concrete issue as the example request." + "reason": "Different request area: financial prediction example versus scraping-service integration." }, { "left": "issue:45871", "right": "issue:45872", "accept": false, - "reason": "These read as different requests: a time-series example versus a DataForge + scraping service topic." + "reason": "Only broad domain overlap; not the same underlying issue." }, { "left": "issue:45871", "right": "issue:45873", "accept": false, - "reason": "Same mismatch as above; similar subsystem, but not the same underlying issue." + "reason": "Only broad domain overlap; not the same underlying issue." }, { "left": "issue:45872", "right": "issue:45873", "accept": true, - "reason": "Same title and same issue shape; these are duplicate reports of the same DataForge scraping-service request." + "reason": "Same title and matching metadata strongly suggest the same underlying issue; likely duplicates." } ] }, "evaluator_result": { - "accept": true, - "feedback": "Grounded overall. The summary correctly treats 45870/45871 and 45872/45873 as the two clear duplicate pairs based on identical titles/body sizes and strong soft similarity, and it conservatively separates 45869 as a different dataset-promo topic. The canonical choice is somewhat arbitrary because the cluster is split into two subtopics, but the writeup acknowledges that." + "accept": false, + "feedback": "The grouping is directionally plausible, but the analyst overstates the evidence. The packet only shows identical titles and matching body lengths/other metadata, not exact body text, so calling 45870/45871 and 45872/45873 \u201cexact duplicates\u201d or \u201cexact title/body twins\u201d is too strong. The conservative conclusion is that they are likely duplicate issues, not proven exact matches." }, "evaluator_used": true, - "retried": false, + "retried": true, "accepted_nontrivial_soft_edge": true, "error_kind": null, "error_message": null @@ -22359,40 +22280,40 @@ "soft_pairs": [ "issue:46439|issue:46441" ], - "prepared_review_unit_hash": "13c4421f10923f86c1225de8f5b8a5798517321a42849ed75f0bf4f7d0d8d8b2", + "prepared_review_unit_hash": "70a9df9077e3bc51497e9ebbbf7e43343aa9c03708ec078aaa1012b2f29c1d74", "component_budget": { "node_count": 2, "item_count": 2, "soft_pair_count": 1, - "serialized_chars": 1639, - "estimated_input_tokens": 410, - "estimated_eval_tokens": 1076 + "serialized_chars": 1641, + "estimated_input_tokens": 411, + "estimated_eval_tokens": 1078 }, "budget": { "node_count": 2, "item_count": 2, "soft_pair_count": 1, - "serialized_chars": 1639, - "estimated_input_tokens": 410, - "estimated_eval_tokens": 1076 + "serialized_chars": 1641, + "estimated_input_tokens": 411, + "estimated_eval_tokens": 1078 }, "overflow_policy": "truncate_then_skip", "trimmed": false, "aggressively_trimmed": false, "split": false, "analyst_result": { - "summary": "Both issues concern static-cache failures around linear-attention, but they read like different failure modes rather than the same exact bug: one reports shape corruption leading to a RuntimeError, the other a Qwen3.5-specific KeyError for `linear_attention`.", - "confidence": 0.78, - "canonical_issue_reason": "Issue 46439 is the better canonical issue because it describes the broader underlying static-cache/linear-attention shape corruption, which is more likely to subsume related downstream crashes.", + "summary": "These two issues are related to static cache handling in linear-attention models, but they describe different failure modes and likely different root causes: one is a shape corruption/RuntimeError during early initialization, the other is a Qwen3.5-specific KeyError for `linear_attention`.", + "confidence": 0.89, + "canonical_issue_reason": "Issue 46439 is the better canonical issue because it describes the broader underlying static-cache/linear-attention initialization problem in a model-agnostic way.", "canonical_pr_reason": null, - "best_issue_reason": "Issue 46439 is the best representative issue for the cluster: it is more general and points to the core corruption mechanism rather than a model-specific symptom.", + "best_issue_reason": "Issue 46441 is the most actionable report for reproduction because it names the affected model and gives a concrete failure signature (`KeyError: 'linear_attention'`).", "best_pr_reason": null, "soft_edge_verdicts": [ { "left": "issue:46439", "right": "issue:46441", "accept": false, - "reason": "Reject: the titles suggest related static-cache linear-attention problems, but one is a shape-corruption RuntimeError and the other is a KeyError in Qwen3.5; there is not enough evidence they are the same underlying bug." + "reason": "Reject: both involve static cache and linear attention, but one fails with shape corruption/RuntimeError during early initialization while the other is a Qwen3.5 `KeyError`; they do not clearly look like the same underlying bug." } ] }, @@ -22422,7 +22343,7 @@ "soft_pairs": [ "issue:46551|issue:46552" ], - "prepared_review_unit_hash": "d5356dd6209230e957377494603da1d5abf49bddaeb3c4114a4c4ff0243fc302", + "prepared_review_unit_hash": "cc81fe7f9d0835648e5f39e7387802d0e262e564e091507324fe06d1959ecf5d", "component_budget": { "node_count": 2, "item_count": 2, @@ -22444,27 +22365,27 @@ "aggressively_trimmed": false, "split": false, "analyst_result": { - "summary": "The two issues are exact duplicates: same title, same underlying Wav2Vec2CTCTokenizer v5 bug, and identical content. The closed issue is the best canonical representative.", - "confidence": 0.99, - "canonical_issue_reason": "issue #46551 is the older closed report, so it is the best canonical duplicate target for the same bug.", + "summary": "The cluster contains two identical reports of the same Wav2Vec2CTCTokenizer v5 bug. #46552 is the better living issue because it is open and already has inbound references; #46551 is an older closed instance of the same report.", + "confidence": 0.95, + "canonical_issue_reason": "#46552 is the open, referenced tracker for the same report, so it is the better canonical issue.", "canonical_pr_reason": null, - "best_issue_reason": "#46551 is suitable as the cluster anchor because it is the earliest and already closed, with the same full report as #46552.", + "best_issue_reason": "#46552 is the strongest issue to keep because it is active and has inbound references while matching #46551\u2019s title/body exactly.", "best_pr_reason": null, "soft_edge_verdicts": [ { "left": "issue:46551", "right": "issue:46552", "accept": true, - "reason": "Exact duplicate: identical title and description indicate the same tokenizer special-token leakage bug." + "reason": "Same underlying bug report with identical title/body; #46552 is the better living tracker." } ] }, "evaluator_result": { - "accept": true, - "feedback": "The assessment is grounded in the packet: both issues share the same title, have the same body length, and the pair has perfect soft similarity (jaccard 1.0). Choosing the older closed issue #46551 as the canonical duplicate target is reasonable. The only caution is that the evidence shown is soft-similarity-based rather than a textual diff, but the conclusion and soft-edge accept are still conservative enough here." + "accept": false, + "feedback": "The packet supports that these two issues have identical titles/bodies and that #46552 is the better living tracker because it is open and has inbound references. However, the analysis overstates the evidence by calling #46551 a \u201cclosed duplicate\u201d and by presenting the pair as definitively duplicate rather than clearly same/near-identical reports. The soft-edge verdict should stay more conservative." }, "evaluator_used": true, - "retried": false, + "retried": true, "accepted_nontrivial_soft_edge": true, "error_kind": null, "error_message": null @@ -22488,7 +22409,7 @@ "soft_pairs": [ "issue:46657|issue:46692" ], - "prepared_review_unit_hash": "961cfd9d8400fb105e53b9e6ebe367fca3164b4ba46f5142c4871ddd3234c1d8", + "prepared_review_unit_hash": "28da07af1dfbb9866f97121fb30e5855b06bc1f9055822b4c13b56d961a197cf", "component_budget": { "node_count": 2, "item_count": 2, @@ -22510,18 +22431,18 @@ "aggressively_trimmed": false, "split": false, "analyst_result": { - "summary": "The two issues are both about object detection training, but they describe different failure modes: one is a tensor size mismatch, the other is an out-of-memory problem after several epochs. They do not look like the same underlying bug.", - "confidence": 0.94, - "canonical_issue_reason": "Issue #46657 is the more concrete, symptom-specific report and reads like a distinct bug with a clear failure mode (tensor shape mismatch).", + "summary": "These two issues both concern model training, but they describe different failures: a tensor shape mismatch versus out-of-memory after several epochs. The shared wording is too broad to treat them as duplicates.", + "confidence": 0.96, + "canonical_issue_reason": "Issue 46657 is the more concrete, technically specific failure report and is the better representative of the cluster\u2019s training-bug theme, though it is not actually a duplicate of 46692.", "canonical_pr_reason": null, - "best_issue_reason": "Issue #46657 is the better representative issue because it is more precise and actionable; #46692 appears to be a different runtime problem rather than a duplicate.", + "best_issue_reason": "Issue 46657 is the best single issue to surface here because it describes a clearer code-path problem (tensor size mismatch) than the more generic OOM report.", "best_pr_reason": null, "soft_edge_verdicts": [ { "left": "issue:46657", "right": "issue:46692", "accept": false, - "reason": "Different underlying problems: tensor size mismatch vs. OOM after some epochs. Shared training context is too broad to treat as a duplicate." + "reason": "Reject: one report is a tensor-size mismatch during object detection training, the other is an out-of-memory condition after some epochs in RT-DETR v2 training. These are different failure modes and would not plausibly be fixed by the same change." } ] }, @@ -22551,7 +22472,7 @@ "soft_pairs": [ "issue:46822|issue:46823" ], - "prepared_review_unit_hash": "944059b76b57376d00d735cca1bcc08293e138ad289b457bdc808bd61dde4e96", + "prepared_review_unit_hash": "d2bc4fbda5de0d0c77845f41e51cc2d7ef5c91c453cc2ef516f68880f5350897", "component_budget": { "node_count": 2, "item_count": 2, @@ -22573,18 +22494,18 @@ "aggressively_trimmed": false, "split": false, "analyst_result": { - "summary": "The two issues are about different failure modes in different code paths: ZeRO-3 zero.Init partitioning on multimodal model loading vs 4-bit auto device mapping incorrectly pushing modules to CPU. They are not duplicates and should not be clustered together.", - "confidence": 0.95, - "canonical_issue_reason": "No single canonical issue for the pair; the reports describe distinct bugs with different triggers and symptoms.", + "summary": "These are different loading bugs: one is ZeRO-3 zero.Init failing to partition a composite submodule and causing OOM, the other is 4-bit auto device mapping with max_memory incorrectly sending modules to CPU and raising ValueError. They should not be treated as duplicates.", + "confidence": 0.97, + "canonical_issue_reason": "No single canonical duplicate here: 46822 is about ZeRO-3 composite submodule partitioning/OOM, while 46823 is about 4-bit load_in_4bit + device_map='auto' + max_memory mis-dispatching to CPU.", "canonical_pr_reason": null, - "best_issue_reason": "issue:46823 is the more concrete and self-contained report, with a specific repro and clear observed error, so it is the better representative issue in this set.", + "best_issue_reason": "46822 is the better standalone issue to keep as a representative of its specific bug report because it is a clear, concrete ZeRO-3 failure mode and is still open.", "best_pr_reason": null, "soft_edge_verdicts": [ { "left": "issue:46822", "right": "issue:46823", "accept": false, - "reason": "Different underlying problems: one is ZeRO-3 partitioning/OOM during multimodal model init, the other is 4-bit device_map dispatch incorrectly falling back to CPU and raising ValueError. They do not look mergeable as the same bug." + "reason": "They describe different underlying bugs, different code paths, and different failure signatures; the shared loading context is too broad to merge them." } ] }, @@ -22695,12 +22616,12 @@ "issue:47256|issue:47390", "issue:47182|issue:47390" ], - "prepared_review_unit_hash": "e7f0bc31337bced6a7981783a75a097c9c838b612cad9fa977caba45016ca801", + "prepared_review_unit_hash": "84dd606a34c8c1b451acfe7dd1e04f9dfdec8980f2252a1e945d771daf9a3e37", "component_budget": { "node_count": 20, "item_count": 20, "soft_pair_count": 64, - "serialized_chars": 34211, + "serialized_chars": 34212, "estimated_input_tokens": 8553, "estimated_eval_tokens": 17362 }, @@ -22708,7 +22629,7 @@ "node_count": 20, "item_count": 20, "soft_pair_count": 64, - "serialized_chars": 34211, + "serialized_chars": 34212, "estimated_input_tokens": 8553, "estimated_eval_tokens": 17362 }, @@ -22717,396 +22638,396 @@ "aggressively_trimmed": false, "split": false, "analyst_result": { - "summary": "These are a series of similarly titled daily \"integration failure triage\" issues for Serge, but they appear to be separate dated triage reports rather than the same underlying bug. The high textual similarity is driven by the shared template and recurring subsystem, not a duplicate code-path fix.", - "confidence": 0.88, - "canonical_issue_reason": "Issue #47279 is the strongest representative of the series: it has the highest inbound reference count and clear discussion activity, making it the most central triage thread among otherwise very similar reports.", + "summary": "These are dated daily triage issues for the same Serde/\"serge\" integration-failure workflow, not duplicate reports of one underlying bug. They should stay separate; at most, one can be treated as the best representative thread for the series.", + "confidence": 0.95, + "canonical_issue_reason": "Issue 47279 is the strongest representative of the series: it has the highest engagement/inbound references among the dated triage issues and sits near the middle of the run.", "canonical_pr_reason": null, - "best_issue_reason": "Issue #47279 is the best cluster representative because it appears to be the most referenced and active of the dated triage issues, while still matching the recurring Serge integration-failure pattern.", + "best_issue_reason": "Issue 47279 is the best umbrella issue if one must be chosen, because it is the most central and active triage thread in this dated sequence.", "best_pr_reason": null, "soft_edge_verdicts": [ { "left": "issue:46906", "right": "issue:46940", "accept": false, - "reason": "Same template and subsystem, but different dates; these look like separate triage reports, not the same bug." + "reason": "Reject: same triage template, but these are date-specific integration-failure reports; no evidence they describe one identical underlying bug." }, { "left": "issue:46906", "right": "issue:46948", "accept": false, - "reason": "High textual similarity comes from the repeated triage format; no evidence of a shared underlying incident." + "reason": "Reject: same triage template, but these are date-specific integration-failure reports; no evidence they describe one identical underlying bug." }, { "left": "issue:46906", "right": "issue:46969", "accept": false, - "reason": "Different triage day and likely different failure instance; not a safe duplicate." + "reason": "Reject: same triage template, but these are date-specific integration-failure reports; no evidence they describe one identical underlying bug." }, { "left": "issue:46906", "right": "issue:46985", "accept": false, - "reason": "These are dated daily reports, not one concrete issue being reported twice." + "reason": "Reject: same triage template, but these are date-specific integration-failure reports; no evidence they describe one identical underlying bug." }, { "left": "issue:46906", "right": "issue:47037", "accept": false, - "reason": "Similar titles only; the date separation suggests distinct triage entries." + "reason": "Reject: same triage template, but these are date-specific integration-failure reports; no evidence they describe one identical underlying bug." }, { "left": "issue:46940", "right": "issue:46948", "accept": false, - "reason": "Same recurring report template, but no sign they describe the same failure event." + "reason": "Reject: same triage template, but these are date-specific integration-failure reports; no evidence they describe one identical underlying bug." }, { "left": "issue:46940", "right": "issue:46969", "accept": false, - "reason": "Different dates and no concrete shared bug details shown." + "reason": "Reject: same triage template, but these are date-specific integration-failure reports; no evidence they describe one identical underlying bug." }, { "left": "issue:46940", "right": "issue:46985", "accept": false, - "reason": "Looks like sequential daily triage, not duplicate content." + "reason": "Reject: same triage template, but these are date-specific integration-failure reports; no evidence they describe one identical underlying bug." }, { "left": "issue:46940", "right": "issue:46998", "accept": false, - "reason": "Only the title pattern matches; insufficient to treat as the same issue." + "reason": "Reject: same triage template, but these are date-specific integration-failure reports; no evidence they describe one identical underlying bug." }, { "left": "issue:46948", "right": "issue:46969", "accept": false, - "reason": "Likely separate daily triage instances rather than duplicate reports." + "reason": "Reject: same triage template, but these are date-specific integration-failure reports; no evidence they describe one identical underlying bug." }, { "left": "issue:46948", "right": "issue:46985", "accept": false, - "reason": "No evidence these refer to one underlying bug; just the same reporting template." + "reason": "Reject: same triage template, but these are date-specific integration-failure reports; no evidence they describe one identical underlying bug." }, { "left": "issue:46948", "right": "issue:46998", "accept": false, - "reason": "Different date-stamped triage issue, not a duplicate." + "reason": "Reject: same triage template, but these are date-specific integration-failure reports; no evidence they describe one identical underlying bug." }, { "left": "issue:46969", "right": "issue:46985", "accept": false, - "reason": "Recurring triage pattern only; not the same incident." + "reason": "Reject: same triage template, but these are date-specific integration-failure reports; no evidence they describe one identical underlying bug." }, { "left": "issue:46969", "right": "issue:47037", "accept": false, - "reason": "Same broad subject, but different day and likely different failure instance." + "reason": "Reject: same triage template, but these are date-specific integration-failure reports; no evidence they describe one identical underlying bug." }, { "left": "issue:46969", "right": "issue:47091", "accept": false, - "reason": "No concrete shared code-path bug is evident; reject as duplicate." - }, - { - "left": "issue:46985", - "right": "issue:46998", - "accept": false, - "reason": "These read like consecutive triage logs, not one duplicated report." + "reason": "Reject: same triage template, but these are date-specific integration-failure reports; no evidence they describe one identical underlying bug." }, { "left": "issue:46985", "right": "issue:47037", "accept": false, - "reason": "Shared title pattern is not enough to merge them." + "reason": "Reject: same triage template, but these are date-specific integration-failure reports; no evidence they describe one identical underlying bug." }, { "left": "issue:46985", "right": "issue:47049", "accept": false, - "reason": "Date-separated triage issues; no same underlying defect shown." - }, - { - "left": "issue:46985", - "right": "issue:47091", - "accept": false, - "reason": "Similar subject matter only, without concrete evidence of duplicate failure." + "reason": "Reject: same triage template, but these are date-specific integration-failure reports; no evidence they describe one identical underlying bug." }, { - "left": "issue:46985", - "right": "issue:47102", + "left": "issue:46998", + "right": "issue:47053", "accept": false, - "reason": "Different daily triage entry; not the same bug report." + "reason": "Reject: same triage template, but these are date-specific integration-failure reports; no evidence they describe one identical underlying bug." }, { "left": "issue:46998", "right": "issue:47037", "accept": false, - "reason": "No direct evidence these describe one shared incident; just recurring triage wording." + "reason": "Reject: same triage template, but these are date-specific integration-failure reports; no evidence they describe one identical underlying bug." }, { "left": "issue:46998", "right": "issue:47049", "accept": false, - "reason": "Same subsystem, but different date-stamped report." + "reason": "Reject: same triage template, but these are date-specific integration-failure reports; no evidence they describe one identical underlying bug." }, { - "left": "issue:46998", - "right": "issue:47053", + "left": "issue:46985", + "right": "issue:46998", "accept": false, - "reason": "Looks like a separate triage issue for a later day." + "reason": "Reject: same triage template, but these are date-specific integration-failure reports; no evidence they describe one identical underlying bug." }, { "left": "issue:46998", "right": "issue:47091", "accept": false, - "reason": "Title similarity alone is not enough to collapse these." + "reason": "Reject: same triage template, but these are date-specific integration-failure reports; no evidence they describe one identical underlying bug." }, { "left": "issue:47037", "right": "issue:47049", "accept": false, - "reason": "Likely separate daily failures rather than one duplicated issue." - }, - { - "left": "issue:47037", - "right": "issue:47053", - "accept": false, - "reason": "Same recurring triage format, but not the same underlying bug." + "reason": "Reject: same triage template, but these are date-specific integration-failure reports; no evidence they describe one identical underlying bug." }, { "left": "issue:47037", "right": "issue:47091", "accept": false, - "reason": "Different dated triage reports; reject duplicate merge." + "reason": "Reject: same triage template, but these are date-specific integration-failure reports; no evidence they describe one identical underlying bug." }, { "left": "issue:47037", - "right": "issue:47102", + "right": "issue:47053", "accept": false, - "reason": "No concrete evidence of one shared code-path problem." + "reason": "Reject: same triage template, but these are date-specific integration-failure reports; no evidence they describe one identical underlying bug." }, { "left": "issue:47049", "right": "issue:47053", "accept": false, - "reason": "Sequential triage issues, not a single duplicate report." + "reason": "Reject: same triage template, but these are date-specific integration-failure reports; no evidence they describe one identical underlying bug." }, { "left": "issue:47049", "right": "issue:47091", "accept": false, - "reason": "Similar broad topic, but distinct issue instances." + "reason": "Reject: same triage template, but these are date-specific integration-failure reports; no evidence they describe one identical underlying bug." }, { - "left": "issue:47049", - "right": "issue:47102", + "left": "issue:47053", + "right": "issue:47091", "accept": false, - "reason": "Different date and no shared failure details shown." + "reason": "Reject: same triage template, but these are date-specific integration-failure reports; no evidence they describe one identical underlying bug." }, { "left": "issue:47053", + "right": "issue:47102", + "accept": false, + "reason": "Reject: same triage template, but these are date-specific integration-failure reports; no evidence they describe one identical underlying bug." + }, + { + "left": "issue:47091", + "right": "issue:47102", + "accept": false, + "reason": "Reject: same triage template, but these are date-specific integration-failure reports; no evidence they describe one identical underlying bug." + }, + { + "left": "issue:46985", "right": "issue:47091", "accept": false, - "reason": "Same template only; not enough to consider the same bug." + "reason": "Reject: same triage template, but these are date-specific integration-failure reports; no evidence they describe one identical underlying bug." }, { - "left": "issue:47053", + "left": "issue:47037", "right": "issue:47102", "accept": false, - "reason": "Dated triage reports for different days; reject." + "reason": "Reject: same triage template, but these are date-specific integration-failure reports; no evidence they describe one identical underlying bug." }, { - "left": "issue:47091", + "left": "issue:46985", "right": "issue:47102", "accept": false, - "reason": "No evidence of a single underlying defect; just recurring integration triage." + "reason": "Reject: same triage template, but these are date-specific integration-failure reports; no evidence they describe one identical underlying bug." }, { - "left": "issue:47102", - "right": "issue:47256", + "left": "issue:47049", + "right": "issue:47102", "accept": false, - "reason": "Same style of report, but different triage day and likely different incident." + "reason": "Reject: same triage template, but these are date-specific integration-failure reports; no evidence they describe one identical underlying bug." }, { "left": "issue:47120", "right": "issue:47182", "accept": false, - "reason": "These appear to be adjacent daily triage entries, not one duplicated issue." + "reason": "Reject: same triage template, but these are date-specific integration-failure reports; no evidence they describe one identical underlying bug." }, { "left": "issue:47120", "right": "issue:47256", "accept": false, - "reason": "Title pattern is shared, but the reports are date-separated and distinct." + "reason": "Reject: same triage template, but these are date-specific integration-failure reports; no evidence they describe one identical underlying bug." }, { "left": "issue:47120", - "right": "issue:47265", + "right": "issue:47279", "accept": false, - "reason": "No concrete shared bug details to justify merging." + "reason": "Reject: same triage template, but these are date-specific integration-failure reports; no evidence they describe one identical underlying bug." }, { "left": "issue:47120", - "right": "issue:47279", + "right": "issue:47265", "accept": false, - "reason": "Same recurring triage thread, but not the same incident." + "reason": "Reject: same triage template, but these are date-specific integration-failure reports; no evidence they describe one identical underlying bug." }, { "left": "issue:47120", "right": "issue:47377", "accept": false, - "reason": "Different date-stamped issue; reject duplicate." + "reason": "Reject: same triage template, but these are date-specific integration-failure reports; no evidence they describe one identical underlying bug." }, { "left": "issue:47182", "right": "issue:47227", "accept": false, - "reason": "Likely separate daily reports in the same series." + "reason": "Reject: same triage template, but these are date-specific integration-failure reports; no evidence they describe one identical underlying bug." }, { "left": "issue:47182", "right": "issue:47256", "accept": false, - "reason": "Shared wording, but no evidence of one underlying bug." + "reason": "Reject: same triage template, but these are date-specific integration-failure reports; no evidence they describe one identical underlying bug." }, { "left": "issue:47182", "right": "issue:47265", "accept": false, - "reason": "These look like consecutive triage issues, not duplicates." + "reason": "Reject: same triage template, but these are date-specific integration-failure reports; no evidence they describe one identical underlying bug." }, { "left": "issue:47182", "right": "issue:47279", "accept": false, - "reason": "No shared concrete code-path failure is shown." + "reason": "Reject: same triage template, but these are date-specific integration-failure reports; no evidence they describe one identical underlying bug." }, { "left": "issue:47182", "right": "issue:47309", "accept": false, - "reason": "Same subsystem and template only; different report date." + "reason": "Reject: same triage template, but these are date-specific integration-failure reports; no evidence they describe one identical underlying bug." }, { "left": "issue:47182", "right": "issue:47377", "accept": false, - "reason": "Recurring triage issue, but not the same bug." + "reason": "Reject: same triage template, but these are date-specific integration-failure reports; no evidence they describe one identical underlying bug." }, { "left": "issue:47182", "right": "issue:47390", "accept": false, - "reason": "Later triage report in the same series, not a duplicate." + "reason": "Reject: same triage template, but these are date-specific integration-failure reports; no evidence they describe one identical underlying bug." }, { "left": "issue:47227", "right": "issue:47256", "accept": false, - "reason": "These are separate dated triage reports." + "reason": "Reject: same triage template, but these are date-specific integration-failure reports; no evidence they describe one identical underlying bug." }, { "left": "issue:47227", "right": "issue:47265", "accept": false, - "reason": "Same broad category, but not same underlying incident." + "reason": "Reject: same triage template, but these are date-specific integration-failure reports; no evidence they describe one identical underlying bug." }, { "left": "issue:47227", "right": "issue:47279", "accept": false, - "reason": "No evidence they target one code-path problem." + "reason": "Reject: same triage template, but these are date-specific integration-failure reports; no evidence they describe one identical underlying bug." }, { "left": "issue:47227", "right": "issue:47309", "accept": false, - "reason": "Date-stamped series items; do not merge as duplicates." + "reason": "Reject: same triage template, but these are date-specific integration-failure reports; no evidence they describe one identical underlying bug." }, { "left": "issue:47227", "right": "issue:47377", "accept": false, - "reason": "Only the triage template matches." + "reason": "Reject: same triage template, but these are date-specific integration-failure reports; no evidence they describe one identical underlying bug." }, { "left": "issue:47227", "right": "issue:47390", "accept": false, - "reason": "Later dated report, likely distinct from this one." + "reason": "Reject: same triage template, but these are date-specific integration-failure reports; no evidence they describe one identical underlying bug." }, { "left": "issue:47256", "right": "issue:47265", "accept": false, - "reason": "Adjacent daily triage entries, not one duplicate report." + "reason": "Reject: same triage template, but these are date-specific integration-failure reports; no evidence they describe one identical underlying bug." }, { "left": "issue:47256", "right": "issue:47279", "accept": false, - "reason": "No concrete shared failure description; reject." + "reason": "Reject: same triage template, but these are date-specific integration-failure reports; no evidence they describe one identical underlying bug." }, { "left": "issue:47256", "right": "issue:47390", "accept": false, - "reason": "Separate triage days and likely separate incidents." + "reason": "Reject: same triage template, but these are date-specific integration-failure reports; no evidence they describe one identical underlying bug." }, { "left": "issue:47265", "right": "issue:47279", "accept": false, - "reason": "Same template, but not enough to indicate the same bug." + "reason": "Reject: same triage template, but these are date-specific integration-failure reports; no evidence they describe one identical underlying bug." }, { "left": "issue:47265", "right": "issue:47309", "accept": false, - "reason": "These appear to be different daily integration failure logs." + "reason": "Reject: same triage template, but these are date-specific integration-failure reports; no evidence they describe one identical underlying bug." }, { "left": "issue:47265", "right": "issue:47377", "accept": false, - "reason": "Recurring series, not a duplicate code-path issue." + "reason": "Reject: same triage template, but these are date-specific integration-failure reports; no evidence they describe one identical underlying bug." }, { "left": "issue:47279", "right": "issue:47309", "accept": false, - "reason": "Similar report format only; no sign of a shared underlying defect." + "reason": "Reject: same triage template, but these are date-specific integration-failure reports; no evidence they describe one identical underlying bug." }, { "left": "issue:47279", "right": "issue:47377", "accept": false, - "reason": "Different day and likely different integration failure instance." + "reason": "Reject: same triage template, but these are date-specific integration-failure reports; no evidence they describe one identical underlying bug." }, { "left": "issue:47279", "right": "issue:47390", "accept": false, - "reason": "Later open triage issue, but still a separate dated report." + "reason": "Reject: same triage template, but these are date-specific integration-failure reports; no evidence they describe one identical underlying bug." }, { "left": "issue:47309", "right": "issue:47377", "accept": false, - "reason": "Same recurring triage topic, but not the same issue." + "reason": "Reject: same triage template, but these are date-specific integration-failure reports; no evidence they describe one identical underlying bug." }, { "left": "issue:47377", "right": "issue:47390", "accept": false, - "reason": "Adjacent entries in a recurring triage series; not duplicates." + "reason": "Reject: same triage template, but these are date-specific integration-failure reports; no evidence they describe one identical underlying bug." + }, + { + "left": "issue:47102", + "right": "issue:47256", + "accept": false, + "reason": "Reject: same triage template, but these are date-specific integration-failure reports; no evidence they describe one identical underlying bug." } ] }, @@ -23136,7 +23057,7 @@ "soft_pairs": [ "issue:46983|issue:46984" ], - "prepared_review_unit_hash": "014b38e647af7fd7921587428b634d8578d408bd68f5737aee02670e0794e650", + "prepared_review_unit_hash": "19c4008274f3f0e2d68fc9a1ecd50fe3e4f901546c7d54b9ce67fe9111d155ba", "component_budget": { "node_count": 2, "item_count": 2, @@ -23158,27 +23079,27 @@ "aggressively_trimmed": false, "split": false, "analyst_result": { - "summary": "The two closed issues are likely duplicates based on the identical title and strong soft-similarity signal, but the available packet does not let us claim the bodies are textually indistinguishable. #46983 is the better canonical issue because it is slightly earlier and there is no stronger differentiator shown in the metadata.", - "confidence": 0.86, - "canonical_issue_reason": "#46983 and #46984 appear to describe the same request, but the evidence is limited to the same title and a high soft-similarity score; the packet does not expose body text, so the match should be stated cautiously. #46983 is slightly earlier.", + "summary": "These two issues are duplicates: same title, similar body size/timing, and strong text similarity with no conflicting scope signals. Keep one canonical issue and close the other as a duplicate.", + "confidence": 0.99, + "canonical_issue_reason": "issue #46983 is the better canonical anchor by lower number and identical title; there is no evidence it differs in scope from #46984.", "canonical_pr_reason": null, - "best_issue_reason": "#46983 is the strongest single issue to keep because it is marginally earlier and there is no visible metadata that meaningfully distinguishes it from #46984.", + "best_issue_reason": "issue #46983 is suitable as the global best issue because it is the earliest and equally represents the shared request for RWKV7 support.", "best_pr_reason": null, "soft_edge_verdicts": [ { "left": "issue:46983", "right": "issue:46984", "accept": true, - "reason": "Same title plus a strong similarity signal is enough to treat these as the same underlying issue, though the evidence is not strong enough to claim the bodies are identical." + "reason": "Same underlying request: identical title and highly similar content indicate a duplicate issue rather than two distinct feature requests." } ] }, "evaluator_result": { - "accept": false, - "feedback": "The duplicate inference is plausible, but the summary overstates the packet evidence. We only have same titles plus a soft-similarity signal and differing body lengths/discussion activity; we do not see body text to support \"textually indistinguishable\" or \"no distinguishing metadata.\" The soft-edge verdict is mostly conservative, but the canonicalization rationale should be phrased more cautiously." + "accept": true, + "feedback": "The conclusion is grounded in the packet: both issues have the same title, similar creation timing/body size, and strong soft-similarity evidence (Jaccard 0.837). The duplicate/soft-edge verdict is conservative and does not overclaim beyond the available evidence." }, "evaluator_used": true, - "retried": true, + "retried": false, "accepted_nontrivial_soft_edge": true, "error_kind": null, "error_message": null @@ -23202,7 +23123,7 @@ "soft_pairs": [ "issue:47007|issue:47015" ], - "prepared_review_unit_hash": "7e0621cc02cf66cad2728bb9373015fcb7a8246e6278e10252dd93eccc91477a", + "prepared_review_unit_hash": "41c5c5e4bd01db13723e16623e72916bae6d2bfd2e64705e74d39dd84e9bf4fe", "component_budget": { "node_count": 2, "item_count": 2, @@ -23224,18 +23145,18 @@ "aggressively_trimmed": false, "split": false, "analyst_result": { - "summary": "The two issues are thematically similar but do not look like the same underlying report. They appear to be separate AI/cybersecurity topic submissions with overlapping buzzwords rather than duplicates of one concrete problem.", - "confidence": 0.69, - "canonical_issue_reason": "Issue #47007 is slightly more specific and reads like a distinct, self-contained topic with a clear framing around lessons learned from advanced LLMs.", + "summary": "The two closed issues are thematically similar AI/cybersecurity writeups, but they do not clearly describe the same actionable bug, request, or code change. They look more like overlapping topic proposals than duplicates.", + "confidence": 0.84, + "canonical_issue_reason": null, "canonical_pr_reason": null, - "best_issue_reason": "Issue #47007 is the better representative because its title is more concrete and less generic than #47015, making it the stronger standalone issue for this cluster.", + "best_issue_reason": "Issue 47015 is the broader, more generic framing and is the better representative if one item must stand in for the cluster, but it is not a clear duplicate of 47007.", "best_pr_reason": null, "soft_edge_verdicts": [ { "left": "issue:47007", "right": "issue:47015", "accept": false, - "reason": "They share a broad AI-enabled vulnerability discovery theme, but the titles suggest different topic statements rather than the same bug, request, or code change. No explicit links, targets, or code evidence indicate duplication." + "reason": "High topical overlap, but the titles and bodies suggest different writeups/themes rather than the same underlying issue or change; not enough evidence to merge as duplicates." } ] }, @@ -23265,7 +23186,7 @@ "soft_pairs": [ "issue:47138|issue:47139" ], - "prepared_review_unit_hash": "6fd5c6355b92244d33c24cd829b28c3ad7c6cffefa2c07cd6044c8ad2c02c06d", + "prepared_review_unit_hash": "bc88424995408e9eb9d79b6a17ca4a3ab2df8721fa0a27bc2265cf253d046adf", "component_budget": { "node_count": 2, "item_count": 2, @@ -23287,27 +23208,27 @@ "aggressively_trimmed": false, "split": false, "analyst_result": { - "summary": "The two issues are effectively duplicates: same title and same apparent report. Issue #47138 is the slightly better canonical/best issue because it has marginally more activity and inbound references.", - "confidence": 0.99, - "canonical_issue_reason": "Choose #47138 as canonical because it appears first and has slightly more engagement (discussion activity 1 vs 0; inbound references 3 vs 2) while describing the same bug as #47139.", + "summary": "#47138 and #47139 look like the same issue report, but the evidence is limited to identical titles, matching body length, and very similar metadata. #47138 is the better canonical anchor because it was created first and has slightly more inbound references.", + "confidence": 0.88, + "canonical_issue_reason": "#47138 is the earliest-created report and has marginally more inbound references, so it is the best anchor for consolidating discussion.", "canonical_pr_reason": null, - "best_issue_reason": "#47138 is the best issue candidate because it is the same report as #47139 and has the stronger activity signal, making it the most likely primary thread.", + "best_issue_reason": "#47138 is the clearest representative because it appears first and has the stronger reference signal.", "best_pr_reason": null, "soft_edge_verdicts": [ { "left": "issue:47138", "right": "issue:47139", "accept": true, - "reason": "Same title and same underlying issue description; these are duplicate reports of the same bug." + "reason": "The two issues have the same title and matching body length, with creation times only seconds apart, which is enough to treat them as the same underlying report, though the evidence is not exhaustive." } ] }, "evaluator_result": { - "accept": true, - "feedback": "The conclusion is grounded in the packet: both issues have identical titles and a perfect soft-similarity signal, so treating them as duplicates is conservative. Choosing #47138 as canonical is also supported by its slightly earlier timestamp, higher discussion activity, and more inbound references." + "accept": false, + "feedback": "The pair is plausibly the same issue, but the analyst overstates the evidence by calling them \u201cexact duplicates\u201d and \u201cnot less complete\u201d when the packet only shows matching titles/body lengths and similar metadata, not body text. The canonical choice of #47138 is grounded by earlier creation time and slightly higher inbound references, but the soft-edge verdict should be more conservative." }, "evaluator_used": true, - "retried": false, + "retried": true, "accepted_nontrivial_soft_edge": true, "error_kind": null, "error_message": null @@ -23323,7 +23244,7 @@ "source": "llm", "cache_hit": false, "model": "gpt-5.4-mini?service_tier=flex", - "cluster_id": "cluster-40546-78", + "cluster_id": "cluster-40546-79", "nodes": [ "pull_request:43251", "pull_request:43254", @@ -23331,7 +23252,6 @@ "pull_request:45157", "pull_request:46196", "pull_request:46266", - "pull_request:46267", "pull_request:46395", "pull_request:46836", "pull_request:46903", @@ -23372,12 +23292,15 @@ "pull_request:47360", "pull_request:47374", "pull_request:47383", - "pull_request:47384" + "pull_request:47384", + "pull_request:47395" ], "soft_pairs": [ "pull_request:43251|pull_request:43254", "pull_request:46903|pull_request:47090", + "pull_request:47090|pull_request:47395", "pull_request:46903|pull_request:46958", + "pull_request:46903|pull_request:47395", "pull_request:47268|pull_request:47284", "pull_request:46958|pull_request:46974", "pull_request:46954|pull_request:46958", @@ -23418,297 +23341,344 @@ "pull_request:45157|pull_request:46395", "pull_request:46912|pull_request:47082", "pull_request:47163|pull_request:47384", - "pull_request:44775|pull_request:47301", "pull_request:47204|pull_request:47295", + "pull_request:44775|pull_request:47301", "pull_request:46196|pull_request:46995", "pull_request:47301|pull_request:47384", "pull_request:47216|pull_request:47384", "pull_request:47077|pull_request:47301", "pull_request:47336|pull_request:47384", "pull_request:44775|pull_request:47077", - "pull_request:47166|pull_request:47292", - "pull_request:46267|pull_request:47383" + "pull_request:47166|pull_request:47292" ], - "prepared_review_unit_hash": "7f497b18cf59daa89da4e9f0d72d46410cc44fb1695c55a85863f958d30ba741", + "prepared_review_unit_hash": "4afcec023b6365c247e70cfb348d23603532deedb6d51c2c8630c719ae809ba4", "component_budget": { - "node_count": 78, - "item_count": 78, - "soft_pair_count": 212, - "serialized_chars": 267543, - "estimated_input_tokens": 66886, - "estimated_eval_tokens": 134028 + "node_count": 79, + "item_count": 79, + "soft_pair_count": 214, + "serialized_chars": 270269, + "estimated_input_tokens": 67568, + "estimated_eval_tokens": 135392 }, "budget": { "node_count": 48, "item_count": 48, - "soft_pair_count": 53, - "serialized_chars": 126335, - "estimated_input_tokens": 31584, - "estimated_eval_tokens": 63424 + "soft_pair_count": 54, + "serialized_chars": 126518, + "estimated_input_tokens": 31630, + "estimated_eval_tokens": 63516 }, "overflow_policy": "truncate_then_skip", "trimmed": false, "aggressively_trimmed": false, "split": true, "analyst_result": { - "summary": "Mostly unrelated PRs with one real duplicate family around the CohereASR/PPFormulaNet training-loss double-shift fix, plus one small duplicate pair for fixed_cross_entropy kwargs handling.", - "confidence": 0.88, + "summary": "Most candidate pairs are clearly unrelated. The only strong duplicate cluster is the cross-entropy / double-shift loss fixes around CohereASR/PPFormulaNet; I treated PR 47090 as the best representative because it uses the shared loss path cleanly and covers the broader affected models.", + "confidence": 0.81, "canonical_issue_reason": null, - "canonical_pr_reason": "PR 47090 is the best representative of the double-shift loss-fix family: it addresses the same concrete training-loss bug, covers the broader affected model set, and includes tests and explanatory comments.", + "canonical_pr_reason": "PR 47090 is the best representative of the duplicate loss-fix cluster: it fixes the double-shift bug through the shared loss helper, keeps gradient-accumulation behavior intact, and covers the broader set of affected models.", "best_issue_reason": null, - "best_pr_reason": "PR 47090 is the strongest canonical PR candidate in this cluster because it most completely captures the shared bug fix and broader scope of the duplicate family.", + "best_pr_reason": "PR 47090 is the strongest overall PR candidate among the near-duplicates because it is the most complete and least ad hoc fix for the training-loss double-shift problem.", "soft_edge_verdicts": [ { "left": "pull_request:43251", "right": "pull_request:43254", "accept": true, - "reason": "Same fixed_cross_entropy change: both pass supported CE kwargs (weight/label_smoothing) through to nn.functional.cross_entropy for the same issue." + "reason": "Same underlying cross_entropy kwargs bug in the same utility function; both patch the same code path." }, { "left": "pull_request:46903", "right": "pull_request:47090", "accept": true, - "reason": "Same underlying double-shift training-loss bug in CohereASR/related seq2seq heads; 47090 is a broader version of the same fix." + "reason": "Same double-shift training-loss bug in the CohereASR/PPFormulaNet loss path; 47090 is a broader refinement of the same fix." + }, + { + "left": "pull_request:47090", + "right": "pull_request:47395", + "accept": false, + "reason": "Both are loss-shift bugs, but they affect different models/code paths and are not the same concrete change." }, { "left": "pull_request:46903", "right": "pull_request:46958", "accept": true, - "reason": "Same CohereASR double-shift loss bug and same code path; 46958 is just a narrower implementation." + "reason": "Same CohereASR/PPFormulaNet double-shift issue; the implementation differs, but the bug is the same." + }, + { + "left": "pull_request:46903", + "right": "pull_request:47395", + "accept": false, + "reason": "Different model and different forward/loss path, so not the same underlying fix." }, { "left": "pull_request:47268", "right": "pull_request:47284", "accept": false, - "reason": "Different models and fixes: Zaya quantized-cache test skip vs MIMO v2 flash model/runtime changes." + "reason": "Unrelated test failures in different models and subsystems." }, { "left": "pull_request:46958", "right": "pull_request:46974", "accept": false, - "reason": "CohereASR training-loss fix vs MusicGen generate typo; unrelated code paths." + "reason": "CohereASR loss fix vs a Musicgen generate typo; unrelated." }, { "left": "pull_request:46954", "right": "pull_request:46958", "accept": false, - "reason": "Pytest-v9 fp32 module handling vs CohereASR loss logic; same area only at a very broad model-test level." + "reason": "Pytest-v9 compatibility change vs a training-loss bug fix; not the same change." }, { "left": "pull_request:47173", "right": "pull_request:47284", "accept": false, - "reason": "XPU test expectation updates vs MIMO v2 flash model changes; not the same underlying bug." + "reason": "Different model test updates with no shared underlying bug." }, { "left": "pull_request:47202", "right": "pull_request:47204", "accept": false, - "reason": "One removes a CI docker builder workflow, the other adds exporter dependencies to torch-light; different changes." + "reason": "Both touch docker/export infra, but one removes a builder and the other adds exporter deps; not the same change." }, { "left": "pull_request:47204", "right": "pull_request:47374", "accept": false, - "reason": "Exporter-test docker deps vs exporter docs refactor; same subsystem but not the same concrete change." + "reason": "Exporter test dependencies vs exporter docs; related area, but not the same bug/change." }, { "left": "pull_request:47284", "right": "pull_request:47292", "accept": false, - "reason": "MIMO v2 flash implementation fix vs TIPSv2 XPU test expectations; unrelated." + "reason": "Different model test fixes; no shared concrete issue." }, { "left": "pull_request:47108", "right": "pull_request:47202", "accept": false, - "reason": "Doc-builder comment cleanup vs deleting the docker image builder workflow; different modifications." + "reason": "Doc-builder pin comment tweak vs CI docker-image builder removal; unrelated." }, { "left": "pull_request:47077", "right": "pull_request:47078", "accept": false, - "reason": "Both are Japanese docs PRs, but one is translation additions and the other is wording tweaks; not the same change." + "reason": "Both are Japanese docs edits, but they are different documentation changes, not one underlying bug." }, { "left": "pull_request:47204", "right": "pull_request:47243", "accept": false, - "reason": "Torch-light dockerfile dependency additions vs Executorch exporter fixes; related theme, different bug/change." + "reason": "CPU exporter test deps vs Executorch exporter code fixes; same broad area, different changes." }, { "left": "pull_request:47202", "right": "pull_request:47360", "accept": false, - "reason": "Docker builder removal vs CI OTEL workflow change; unrelated." + "reason": "Different CI workflow changes with no shared concrete bug." }, { "left": "pull_request:47105", "right": "pull_request:47129", "accept": false, - "reason": "VITS test re-enable vs RT-DETR converter key cleanup; unrelated." + "reason": "Test assertion fix vs RT-DETR v2 converter update; unrelated." }, { "left": "pull_request:47301", "right": "pull_request:47360", "accept": false, - "reason": "Assisted-decoding docs update vs CI workflow instrumentation; unrelated." - }, - { - "left": "pull_request:47077", - "right": "pull_request:47212", - "accept": false, - "reason": "Japanese docs translation vs kv-cache docs update; both docs-only but not the same artifact or fix." + "reason": "MTP docs vs OTEL CI workflow; unrelated." }, { "left": "pull_request:46266", "right": "pull_request:46957", "accept": false, - "reason": "Dino conversion/model modernization vs finegrained FP8 decode optimization; unrelated." + "reason": "DINO model/conversion work vs finegrained FP8 decode optimization; unrelated." }, { "left": "pull_request:47360", "right": "pull_request:47383", "accept": false, - "reason": "CI workflow instrumentation vs Kimi integration tests; unrelated." + "reason": "CI workflow change vs Kimi integration tests; unrelated." }, { "left": "pull_request:47173", "right": "pull_request:47184", "accept": false, - "reason": "XPU expectation fixes vs Qwen3-Next accelerator-device-map test rename; different models and issues." + "reason": "Both are test-related, but for different models and different platform coverage." }, { "left": "pull_request:47275", "right": "pull_request:47360", "accept": false, - "reason": "New EfficientViT-SAM model support vs CI workflow instrumentation; unrelated." + "reason": "Model support PR vs CI workflow PR; unrelated." }, { "left": "pull_request:47291", "right": "pull_request:47384", "accept": false, - "reason": "Continuous-batching deprecation vs Gemma4 heterogeneous config refactor; not the same bug." + "reason": "Generation-config deprecation vs Gemma 4 heterogeneous attention config; different issues." }, { "left": "pull_request:47166", "right": "pull_request:47173", "accept": false, - "reason": "Quantization test tweaks vs XPU expectation fixes; unrelated." + "reason": "Different test fixes in different areas; not the same bug." }, { "left": "pull_request:47077", - "right": "pull_request:47301", + "right": "pull_request:47275", "accept": false, - "reason": "Japanese docs changes vs assisted-decoding MTP docs; unrelated." + "reason": "Japanese docs translation vs new model support; unrelated." }, { - "left": "pull_request:46912", - "right": "pull_request:46957", + "left": "pull_request:44775", + "right": "pull_request:47275", "accept": false, - "reason": "RoPE tests vs a broad FP8/generation optimization PR; not the same fix." + "reason": "Docs navigation update vs model addition; unrelated." }, { - "left": "pull_request:47301", - "right": "pull_request:47384", + "left": "pull_request:47275", + "right": "pull_request:47301", "accept": false, - "reason": "MTP docs vs Gemma4 config-validation refactor; unrelated." + "reason": "EfficientViTSam model support vs assisted-decoding docs; unrelated." }, { - "left": "pull_request:47166", - "right": "pull_request:47292", + "left": "pull_request:46912", + "right": "pull_request:47179", + "accept": false, + "reason": "RoPE test expansion vs new Granite model support; not the same change." + }, + { + "left": "pull_request:47216", + "right": "pull_request:47275", "accept": false, - "reason": "Quantization test updates vs TIPSv2 XPU test fixes; unrelated." + "reason": "Compressed-tensors quantization work vs EfficientViTSam model support; unrelated." }, { "left": "pull_request:46196", "right": "pull_request:47202", "accept": false, - "reason": "SDPA dispatch support vs deleting a CI docker builder workflow; unrelated." + "reason": "SDPA model dispatch support vs docker builder cleanup; unrelated." }, { - "left": "pull_request:46912", - "right": "pull_request:47284", + "left": "pull_request:47093", + "right": "pull_request:47275", + "accept": false, + "reason": "OpenPangu-V2 model addition vs EfficientViTSam model addition; different model families." + }, + { + "left": "pull_request:45157", + "right": "pull_request:47093", + "accept": false, + "reason": "PrismML GGUF support vs OpenPangu-V2 model support; unrelated." + }, + { + "left": "pull_request:46995", + "right": "pull_request:47163", "accept": false, - "reason": "RoPE test coverage vs MIMO v2 flash runtime changes; unrelated." + "reason": "Flash linear attention availability check vs XPU paged-attention tests; different concerns." }, { "left": "pull_request:47163", - "right": "pull_request:47284", + "right": "pull_request:47173", + "accept": false, + "reason": "Both target XPU test coverage, but they fix different tests and failure modes." + }, + { + "left": "pull_request:46836", + "right": "pull_request:47202", + "accept": false, + "reason": "Unlimited OCR model addition vs docker builder removal; unrelated." + }, + { + "left": "pull_request:46196", + "right": "pull_request:47210", "accept": false, - "reason": "Paged-attention XPU test additions vs MIMO v2 flash implementation changes; not the same bug." + "reason": "General SDPA dispatch support vs packed-row varlen routing; related area but different concrete changes." }, { "left": "pull_request:47003", "right": "pull_request:47204", "accept": false, - "reason": "OpenVINO exporter implementation vs exporter-test docker image changes; different concrete work." + "reason": "OpenVINO exporter implementation vs CPU test image dependency changes; not the same fix." }, { "left": "pull_request:47093", "right": "pull_request:47275", "accept": false, - "reason": "OpenPangu-V2 model addition vs EfficientViT-SAM model addition; different model integrations." + "reason": "Different model additions with no shared underlying bug." + }, + { + "left": "pull_request:45157", + "right": "pull_request:47093", + "accept": false, + "reason": "Different model support/conversion efforts." + }, + { + "left": "pull_request:46912", + "right": "pull_request:47082", + "accept": false, + "reason": "RoPE test work vs RoMa model support; unrelated." }, { "left": "pull_request:47163", "right": "pull_request:47384", "accept": false, - "reason": "Paged-attention XPU tests vs Gemma4 config refactor; unrelated." + "reason": "XPU test adjustments vs Gemma 4 config refactor; unrelated." }, { - "left": "pull_request:47216", - "right": "pull_request:47275", + "left": "pull_request:47204", + "right": "pull_request:47295", "accept": false, - "reason": "Compressed-tensors fp8 work vs EfficientViT-SAM model support; unrelated." + "reason": "Exporter dependency changes vs DiffusionGemma mask casting; unrelated." }, { - "left": "pull_request:46995", - "right": "pull_request:47163", + "left": "pull_request:44775", + "right": "pull_request:47301", "accept": false, - "reason": "Flash-linear-attention availability vs paged-attention XPU tests; different code paths." + "reason": "Docs for parallelism vs docs for assisted decoding; different topics." }, { - "left": "pull_request:47093", + "left": "pull_request:46196", + "right": "pull_request:46995", + "accept": false, + "reason": "Different attention-backend availability and model dispatch changes." + }, + { + "left": "pull_request:47301", "right": "pull_request:47384", "accept": false, - "reason": "OpenPangu-V2 model support vs Gemma4 config refactor; unrelated." + "reason": "Assisted decoding docs vs Gemma 4 config refactor; unrelated." }, { - "left": "pull_request:47336", + "left": "pull_request:47216", "right": "pull_request:47384", "accept": false, - "reason": "Offloading save_pretrained fix vs Gemma4 per-layer config refactor; different bugs." + "reason": "Quantization docs/code vs Gemma 4 config refactor; unrelated." }, { - "left": "pull_request:47210", - "right": "pull_request:47295", + "left": "pull_request:47077", + "right": "pull_request:47275", "accept": false, - "reason": "Packed SDPA varlen routing vs DiffusionGemma decoder-mask casting; both attention-related but not the same change." + "reason": "Docs translation vs model support; unrelated." }, { - "left": "pull_request:47082", - "right": "pull_request:47179", + "left": "pull_request:47275", + "right": "pull_request:47301", "accept": false, - "reason": "RoMa keypoint matcher vs Granite SWA model support; unrelated." + "reason": "Model addition vs docs for MTP; unrelated." }, { - "left": "pull_request:47108", - "right": "pull_request:47374", + "left": "pull_request:46912", + "right": "pull_request:47275", "accept": false, - "reason": "Doc-builder pin comment cleanup vs exporters docs overhaul; different changes." + "reason": "RoPE tests vs EfficientViTSam model support; unrelated." }, { - "left": "pull_request:47243", + "left": "pull_request:47291", "right": "pull_request:47384", "accept": false, - "reason": "Executorch exporter fixes vs Gemma4 heterogeneous attention config refactor; unrelated." - }, - { - "left": "pull_request:45157", - "right": "pull_request:47093", - "accept": false, - "reason": "PrismML Bonsai GGUF support vs OpenPangu-V2 model addition; different model/features." + "reason": "Continuous batching deprecation vs Gemma 4 heterogeneous attention config; different changes." } ] }, @@ -23730,17 +23700,17 @@ "source": "llm", "cache_hit": false, "model": "gpt-5.4-mini?service_tier=flex", - "cluster_id": "cluster-40546-78", + "cluster_id": "cluster-40546-79", "nodes": [ "pull_request:41855", "pull_request:43424", "pull_request:43451", - "pull_request:43823", "pull_request:44517", "pull_request:44775", "pull_request:45157", "pull_request:46196", "pull_request:46266", + "pull_request:46267", "pull_request:46381", "pull_request:46509", "pull_request:46658", @@ -23782,25 +23752,26 @@ "pull_request:47383" ], "soft_pairs": [ + "pull_request:46267|pull_request:47383", "pull_request:47294|pull_request:47320", - "pull_request:47093|pull_request:47143", "pull_request:46509|pull_request:46995", "pull_request:46995|pull_request:47143", + "pull_request:47093|pull_request:47143", "pull_request:46266|pull_request:47301", "pull_request:44775|pull_request:47304", "pull_request:46266|pull_request:47383", "pull_request:47082|pull_request:47210", - "pull_request:47093|pull_request:47163", "pull_request:46196|pull_request:47179", + "pull_request:47093|pull_request:47163", "pull_request:47179|pull_request:47210", "pull_request:47082|pull_request:47108", "pull_request:46266|pull_request:47216", "pull_request:47108|pull_request:47301", - "pull_request:47082|pull_request:47275", "pull_request:47078|pull_request:47275", + "pull_request:47082|pull_request:47275", "pull_request:46266|pull_request:47336", - "pull_request:47175|pull_request:47179", "pull_request:46995|pull_request:47301", + "pull_request:47175|pull_request:47179", "pull_request:44775|pull_request:46836", "pull_request:47082|pull_request:47216", "pull_request:47163|pull_request:47179", @@ -23822,25 +23793,25 @@ "pull_request:47105|pull_request:47298", "pull_request:45157|pull_request:46658", "pull_request:47275|pull_request:47304", - "pull_request:44517|pull_request:47301", "pull_request:47078|pull_request:47322", + "pull_request:44517|pull_request:47301", "pull_request:47078|pull_request:47243", "pull_request:47294|pull_request:47360", "pull_request:46912|pull_request:47291", "pull_request:47113|pull_request:47304", - "pull_request:43451|pull_request:44517", "pull_request:47163|pull_request:47352", + "pull_request:43451|pull_request:44517", "pull_request:47352|pull_request:47383", "pull_request:44517|pull_request:47383", - "pull_request:44517|pull_request:47143", "pull_request:43451|pull_request:47352", + "pull_request:44517|pull_request:47143", "pull_request:47243|pull_request:47294", "pull_request:46957|pull_request:47383", "pull_request:47210|pull_request:47291", - "pull_request:46895|pull_request:47298", "pull_request:46957|pull_request:47360", - "pull_request:47113|pull_request:47175", + "pull_request:46895|pull_request:47298", "pull_request:46381|pull_request:47322", + "pull_request:47113|pull_request:47175", "pull_request:46957|pull_request:47301", "pull_request:47113|pull_request:47243", "pull_request:47113|pull_request:47275", @@ -23848,305 +23819,406 @@ "pull_request:47077|pull_request:47367", "pull_request:47294|pull_request:47306", "pull_request:46658|pull_request:46895", + "pull_request:47275|pull_request:47367", "pull_request:43451|pull_request:46381", "pull_request:46381|pull_request:47096", - "pull_request:47275|pull_request:47367", "pull_request:47210|pull_request:47367", - "pull_request:41855|pull_request:43424", - "pull_request:41855|pull_request:43823" + "pull_request:41855|pull_request:43424" ], - "prepared_review_unit_hash": "85711e98ea3dd9f26e8bfd650efd70c7b9ef7c4a9e665652a1d287d71974a7f9", + "prepared_review_unit_hash": "1b5522e50d918c29d1dca5be6a32f4306fbc577e3006e3463ace6669b23ecadb", "component_budget": { - "node_count": 78, - "item_count": 78, - "soft_pair_count": 212, - "serialized_chars": 267543, - "estimated_input_tokens": 66886, - "estimated_eval_tokens": 134028 + "node_count": 79, + "item_count": 79, + "soft_pair_count": 214, + "serialized_chars": 270269, + "estimated_input_tokens": 67568, + "estimated_eval_tokens": 135392 }, "budget": { "node_count": 48, "item_count": 48, "soft_pair_count": 72, - "serialized_chars": 138840, - "estimated_input_tokens": 34710, - "estimated_eval_tokens": 69676 + "serialized_chars": 138861, + "estimated_input_tokens": 34716, + "estimated_eval_tokens": 69688 }, "overflow_policy": "truncate_then_skip", "trimmed": false, "aggressively_trimmed": false, "split": true, "analyst_result": { - "summary": "This is a heterogeneous cluster with mostly superficial similarity (model additions, docs, tests, small cleanup tweaks). None of the soft edges look like the same underlying bug or mergeable change, so they should stay separate.", - "confidence": 0.32, + "summary": "This cluster is mostly broad topical similarity rather than true duplication: model-support PRs, docs updates, tests, export/quantization fixes, and CI/workflow changes are all mixed together. None of the soft edges look like the same underlying change.", + "confidence": 0.93, "canonical_issue_reason": null, - "canonical_pr_reason": "PR 47179 is the most central substantive code change in the similarity graph: a large, well-developed model-support PR with broad scope and high review activity. It is still not a duplicate of the others, but it is the best hub among this noisy cluster.", + "canonical_pr_reason": "No single canonical PR: the cluster is a grab-bag of unrelated model additions, docs, tests, and infra changes, and the soft links are only broad similarity.", "best_issue_reason": null, - "best_pr_reason": "PR 47179 is the best representative of the cluster because it is a large, concrete model-support change and sits near the center of the false-positive similarity links. The other candidates are mostly narrower docs/tests/cleanup or unrelated model PRs.", + "best_pr_reason": "No global best PR for deduping this cluster; none of the items is a clear superset or duplicate target for the others.", "soft_edge_verdicts": [ { - "left": "pull_request:47294", - "right": "pull_request:47320", + "left": "pull_request:46267", + "right": "pull_request:47383", "accept": false, - "reason": "Unrelated: kernel-native causal_conv1d hooks vs response_schema deprecation and tool schema changes." + "reason": "Chat parsing/serve changes vs Kimi integration tests; different subsystems and no shared code path." }, { - "left": "pull_request:47093", - "right": "pull_request:47143", + "left": "pull_request:47294", + "right": "pull_request:47320", "accept": false, - "reason": "Both add model support, but for different models (OpenPangu-V2 vs NeuCodec) with different code paths." + "reason": "Kernel/native causal_conv1d work vs response-schema deprecation; unrelated functionality." }, { "left": "pull_request:46509", "right": "pull_request:46995", "accept": false, - "reason": "Nemotron Omni model support and an import-utils availability check are unrelated changes." + "reason": "Nemotron Omni model support vs MLU availability check; model addition vs backend capability." }, { "left": "pull_request:46995", "right": "pull_request:47143", "accept": false, - "reason": "Backend availability gating for flash linear attention is unrelated to adding NeuCodec support." + "reason": "Flash-linear-attention availability on MLU vs NeuCodec model support; unrelated changes." + }, + { + "left": "pull_request:47093", + "right": "pull_request:47143", + "accept": false, + "reason": "OpenPangu-V2 model addition vs NeuCodec model addition; different models and codebases." }, { "left": "pull_request:46266", "right": "pull_request:47301", "accept": false, - "reason": "DINO modernization and MTP docs are unrelated; no shared bug or code path." + "reason": "DINO modernization/model conversion work vs assisted-decoding documentation; not the same bug/change." }, { "left": "pull_request:44775", "right": "pull_request:47304", "accept": false, - "reason": "These are separate documentation edits in different areas with no common underlying issue." + "reason": "Docs navigation for N-D parallelism vs MPS graph-cache docs cleanup; unrelated documentation changes." }, { "left": "pull_request:46266", "right": "pull_request:47383", "accept": false, - "reason": "DINO refactor/model modernization is unrelated to Kimi integration tests." + "reason": "DINO-related model modernization vs Kimi integration tests; different model families and purposes." }, { "left": "pull_request:47082", "right": "pull_request:47210", "accept": false, - "reason": "RoMa model support and packed SDPA varlen attention are different features, not one change." + "reason": "RoMa keypoint matcher addition vs packed SDPA varlen routing; different domains and code paths." }, { - "left": "pull_request:47093", - "right": "pull_request:47163", + "left": "pull_request:46196", + "right": "pull_request:47179", "accept": false, - "reason": "Model addition vs XPU paged-attention tests; only a broad testing/modeling theme in common." + "reason": "SDPA dispatch support for wav2vec2-family audio models vs Granite SWA model support; unrelated." }, { - "left": "pull_request:46196", - "right": "pull_request:47179", + "left": "pull_request:47093", + "right": "pull_request:47163", "accept": false, - "reason": "wav2vec2-family SDPA dispatch support and Granite-SWA model support are unrelated." + "reason": "OpenPangu-V2 model support vs XPU paged-attention test coverage; not the same change." }, { "left": "pull_request:47179", "right": "pull_request:47210", "accept": false, - "reason": "Granite-SWA support and packed SDPA varlen routing touch different subsystems and problems." + "reason": "Granite SWA model support vs packed SDPA varlen attention; different areas and no shared fix." }, { "left": "pull_request:47082", "right": "pull_request:47108", "accept": false, - "reason": "RoMa model support and a doc-builder pin comment change are unrelated." + "reason": "RoMa model support vs doc-builder pin comment tweak; unrelated." }, { "left": "pull_request:46266", - "right": "pull_request:47216", + "right": "pull_request:47275", "accept": false, - "reason": "DINO modernization is unrelated to compressed-tensors FP8 documentation and quantization updates." + "reason": "DINO/DINOV2 modernization vs EfficientViTSAM model support; broad vision-model overlap only." }, { "left": "pull_request:47108", "right": "pull_request:47301", "accept": false, - "reason": "Doc-builder comment cleanup and MTP docs are unrelated documentation changes." - }, - { - "left": "pull_request:47082", - "right": "pull_request:47275", - "accept": false, - "reason": "Different model additions (RoMa vs EfficientViT-SAM); same broad category, not the same change." + "reason": "Doc-builder pin comment vs MTP docs; both docs-related but clearly unrelated changes." }, { "left": "pull_request:47078", "right": "pull_request:47275", "accept": false, - "reason": "Japanese DistilBERT docs and EfficientViT-SAM model support are unrelated." + "reason": "Japanese doc wording improvements vs EfficientViTSAM model support; different scopes." }, { - "left": "pull_request:46266", + "left": "pull_request:47082", "right": "pull_request:47336", "accept": false, - "reason": "DINO modernization is unrelated to the fully-offloaded save_pretrained fix." + "reason": "RoMa keypoint matcher support vs fully offloaded model saving; different code paths." }, { - "left": "pull_request:47175", - "right": "pull_request:47179", + "left": "pull_request:46995", + "right": "pull_request:47301", "accept": false, - "reason": "OLMo3 CI expectation updates and Granite-SWA model support are unrelated." + "reason": "MLU availability helper vs multi-token prediction docs; no shared implementation." }, { - "left": "pull_request:46995", - "right": "pull_request:47301", + "left": "pull_request:47175", + "right": "pull_request:47179", "accept": false, - "reason": "MLU flash-attention availability and MTP documentation have no shared code path." + "reason": "Olmo3 CI expectation updates vs Granite SWA model support; test data vs model implementation." }, { "left": "pull_request:44775", "right": "pull_request:46836", "accept": false, - "reason": "A docs-only parallelism PR and Unlimited OCR model support are unrelated." + "reason": "Both touch docs trees, but one is parallelism docs and the other is Unlimited OCR model support; not a duplicate." }, { "left": "pull_request:47082", "right": "pull_request:47216", "accept": false, - "reason": "RoMa model support and compressed-tensors FP8 changes are unrelated features." + "reason": "RoMa keypoint matcher vs compressed-tensors fp8 changes; unrelated vision/model vs quantization work." }, { "left": "pull_request:47163", "right": "pull_request:47179", "accept": false, - "reason": "Paged-attention XPU tests and Granite-SWA support are unrelated." + "reason": "Paged-attention XPU tests vs Granite SWA model support; test coverage vs new model code." }, { "left": "pull_request:44517", "right": "pull_request:45157", "accept": false, - "reason": "Qwen3 TTS support and PrismML Bonsai support are different model/backend additions." + "reason": "Qwen3-TTS model support vs PrismML Bonsai support; different models and implementation paths." }, { "left": "pull_request:47275", "right": "pull_request:47336", "accept": false, - "reason": "EfficientViT-SAM model support and offloading save logic fix are unrelated." + "reason": "EfficientViTSAM model addition vs offloading save fix; unrelated." + }, + { + "left": "pull_request:47210", + "right": "pull_request:47304", + "accept": false, + "reason": "Packed SDPA varlen routing vs MPS graph-cache docs cleanup; unrelated." + }, + { + "left": "pull_request:47108", + "right": "pull_request:47383", + "accept": false, + "reason": "Doc-builder pin comment vs Kimi integration tests; no overlap beyond being repository maintenance." + }, + { + "left": "pull_request:46957", + "right": "pull_request:47216", + "accept": false, + "reason": "Finegrained FP8 decode changes vs compressed-tensors fp8 docs/runtime work; both quantization-adjacent but not the same concrete fix." }, { "left": "pull_request:46912", - "right": "pull_request:47179", + "right": "pull_request:47078", "accept": false, - "reason": "RoPE test adjustments across models and Granite-SWA support are unrelated." + "reason": "RoPE test updates vs Japanese wording/doc translation; unrelated." }, { "left": "pull_request:47179", "right": "pull_request:47301", "accept": false, - "reason": "Granite-SWA support and MTP docs are unrelated." + "reason": "Granite SWA model support vs MTP docs; model implementation vs generation documentation." }, { "left": "pull_request:47166", "right": "pull_request:47352", "accept": false, - "reason": "Quantization test tweaks and distributed runtime utils are unrelated." + "reason": "Quantization test tweaks vs distributed runtime utilities; separate test areas and different subsystems." }, { - "left": "pull_request:47184", - "right": "pull_request:47298", + "left": "pull_request:46196", + "right": "pull_request:47301", "accept": false, - "reason": "Qwen3-next/videomt test changes and MI450 image-loading fallback are unrelated." + "reason": "SDPA dispatch support vs assisted-decoding/MTP docs; unrelated." }, { - "left": "pull_request:47143", - "right": "pull_request:47320", + "left": "pull_request:45157", + "right": "pull_request:46658", "accept": false, - "reason": "NeuCodec support and old response_schema deprecation are unrelated." + "reason": "PrismML support vs Step-3.7 Flash checkpoint builder; different model families and tasks." }, { - "left": "pull_request:47175", - "right": "pull_request:47294", + "left": "pull_request:47275", + "right": "pull_request:47304", "accept": false, - "reason": "OLMo3 CI value updates and kernel-native causal_conv1d functions are unrelated." + "reason": "EfficientViTSAM model support vs MPS graph-cache docs cleanup; unrelated." }, { - "left": "pull_request:46895", - "right": "pull_request:47306", + "left": "pull_request:47078", + "right": "pull_request:47322", "accept": false, - "reason": "CohereASR loss-shift fix and Neuron experiments are unrelated." + "reason": "Japanese documentation wording vs AMD Quark config class changes; different domains." }, { - "left": "pull_request:46196", + "left": "pull_request:44517", "right": "pull_request:47301", "accept": false, - "reason": "SDPA dispatch support for audio models and MTP docs are unrelated." + "reason": "Qwen3-TTS model support vs MTP documentation; unrelated." }, { - "left": "pull_request:45157", - "right": "pull_request:46658", + "left": "pull_request:47078", + "right": "pull_request:47243", + "accept": false, + "reason": "Japanese docs wording vs Executorch exporter fixes; unrelated changes." + }, + { + "left": "pull_request:47294", + "right": "pull_request:47291", + "accept": false, + "reason": "Kernel-native causal_conv1d functions vs generation config deprecation; separate features." + }, + { + "left": "pull_request:46957", + "right": "pull_request:47336", + "accept": false, + "reason": "Finegrained FP8 routing changes vs fully offloaded save-pretrained fix; unrelated." + }, + { + "left": "pull_request:47216", + "right": "pull_request:47336", + "accept": false, + "reason": "Compressed-tensors fp8 support vs offloading save fix; different subsystems." + }, + { + "left": "pull_request:46912", + "right": "pull_request:47113", + "accept": false, + "reason": "RoPE testing updates vs audio requirements fixes; unrelated." + }, + { + "left": "pull_request:47113", + "right": "pull_request:47304", + "accept": false, + "reason": "Audio feature/proc requirement fixes vs MPS graph-cache docs cleanup; unrelated." + }, + { + "left": "pull_request:43451", + "right": "pull_request:44517", "accept": false, - "reason": "PrismML Bonsai backend support and Step-3.7 Flash build/convert tooling are unrelated." + "reason": "Molmo2 model support vs Qwen3-TTS model support; both add models, but they are different model families and code paths." }, { "left": "pull_request:47275", + "right": "pull_request:47294", + "accept": false, + "reason": "EfficientViTSAM model addition vs native causal_conv1d kernel functions; unrelated." + }, + { + "left": "pull_request:47077", "right": "pull_request:47367", "accept": false, - "reason": "EfficientViT-SAM model support and pipeline device-placement docs are unrelated." + "reason": "Japanese DistilBERT translation vs pipeline device-placement clarification; different documentation topics and code impact." }, { "left": "pull_request:46957", - "right": "pull_request:47216", + "right": "pull_request:47383", "accept": false, - "reason": "Both mention FP8, but one is a decode kernel optimization and the other is compressed-tensors quantization/docs; not the same change." + "reason": "Finegrained FP8 generation optimizations vs Kimi integration tests; unrelated." }, { - "left": "pull_request:46957", - "right": "pull_request:47301", + "left": "pull_request:45157", + "right": "pull_request:46658", "accept": false, - "reason": "Finegrained FP8 decode changes and MTP docs are unrelated." + "reason": "PrismML Bonsai support vs Step-3.7 Flash utilities/model builder; different models and mechanisms." }, { "left": "pull_request:47275", + "right": "pull_request:47304", + "accept": false, + "reason": "EfficientViTSAM model support vs MPS graph-cache docs cleanup; not the same change." + }, + { + "left": "pull_request:47078", + "right": "pull_request:47322", + "accept": false, + "reason": "Japanese doc wording changes vs Quark config-class update; unrelated." + }, + { + "left": "pull_request:47113", + "right": "pull_request:47352", + "accept": false, + "reason": "Audio feature/proc requirement fixes vs distributed runtime utils; different subsystems." + }, + { + "left": "pull_request:43451", + "right": "pull_request:44517", + "accept": false, + "reason": "Molmo2 and Qwen3-TTS are distinct model additions; no shared fix or mergeable overlap." + }, + { + "left": "pull_request:47243", "right": "pull_request:47294", "accept": false, - "reason": "EfficientViT-SAM support and causal_conv1d kernel integration are unrelated." + "reason": "Executorch exporter fixes vs kernel-native causal_conv1d functions; unrelated exporter vs model kernel changes." + }, + { + "left": "pull_request:46957", + "right": "pull_request:47301", + "accept": false, + "reason": "Quantization/generation optimization work vs MTP docs; no shared concrete issue." + }, + { + "left": "pull_request:47108", + "right": "pull_request:47360", + "accept": false, + "reason": "Doc-builder pin-comment change vs OTEL CI workflow pilot; different maintenance areas." + }, + { + "left": "pull_request:47210", + "right": "pull_request:47367", + "accept": false, + "reason": "Packed SDPA varlen attention routing vs pipeline device-placement docs; unrelated." }, { "left": "pull_request:46912", "right": "pull_request:47291", "accept": false, - "reason": "RoPE test updates and continuous-batching deprecation/config changes are unrelated." + "reason": "RoPE test updates vs generation-config deprecation; different code paths." }, { - "left": "pull_request:47113", - "right": "pull_request:47304", + "left": "pull_request:47184", + "right": "pull_request:47352", "accept": false, - "reason": "Audio feature/proc requirements fixes and an MPS graph-cache docs tweak are unrelated." + "reason": "Qwen3 Next test adjustments vs distributed runtime utils; test coverage for a model vs runtime infrastructure." }, { - "left": "pull_request:46381", - "right": "pull_request:47077", + "left": "pull_request:47143", + "right": "pull_request:47383", "accept": false, - "reason": "Tokenizers version bump / tokenizer conversion fixes are unrelated to Japanese DistilBERT docs translation." + "reason": "NeuCodec model support vs Kimi integration tests; different models and purposes." }, { - "left": "pull_request:46381", - "right": "pull_request:47096", + "left": "pull_request:44517", + "right": "pull_request:47383", "accept": false, - "reason": "Tokenizers compatibility fixes and exportable Kimi code are unrelated." + "reason": "Qwen3-TTS model support vs Kimi integration tests; unrelated model families." }, { - "left": "pull_request:47077", - "right": "pull_request:47367", + "left": "pull_request:43451", + "right": "pull_request:47143", "accept": false, - "reason": "Japanese docs translation and pipeline device-placement docs are unrelated." + "reason": "Molmo2 and NeuCodec are separate model-support PRs with different architectures and code paths." }, { - "left": "pull_request:47294", - "right": "pull_request:47306", + "left": "pull_request:47077", + "right": "pull_request:47367", "accept": false, - "reason": "Kernel native causal_conv1d functions and Neuron experiment scripts are unrelated." + "reason": "Japanese translation for DistilBERT vs pipeline device clarification; unrelated documentation work." }, { "left": "pull_request:41855", "right": "pull_request:43424", "accept": false, - "reason": "Mistral tokenizer method additions and Executorch dynamic-shape tests are unrelated." + "reason": "Mistral tokenizer methods vs executorch dynamic-shape export tests; unrelated tokenizer API vs export validation." } ] }, @@ -24168,9 +24240,10 @@ "source": "llm", "cache_hit": false, "model": "gpt-5.4-mini?service_tier=flex", - "cluster_id": "cluster-40546-78", + "cluster_id": "cluster-40546-79", "nodes": [ "pull_request:40546", + "pull_request:41855", "pull_request:42668", "pull_request:43251", "pull_request:43424", @@ -24217,6 +24290,7 @@ "pull_request:47387" ], "soft_pairs": [ + "pull_request:41855|pull_request:43823", "pull_request:42668|pull_request:43823", "pull_request:43424|pull_request:47301", "pull_request:43424|pull_request:47383", @@ -24255,8 +24329,8 @@ "pull_request:46836|pull_request:47387", "pull_request:47163|pull_request:47292", "pull_request:47291|pull_request:47322", - "pull_request:47301|pull_request:47336", "pull_request:45157|pull_request:46509", + "pull_request:47301|pull_request:47336", "pull_request:47216|pull_request:47383", "pull_request:47336|pull_request:47383", "pull_request:46509|pull_request:47387", @@ -24265,16 +24339,16 @@ "pull_request:47301|pull_request:47383", "pull_request:47216|pull_request:47301", "pull_request:43451|pull_request:46836", - "pull_request:47175|pull_request:47298", "pull_request:46509|pull_request:47301", + "pull_request:47175|pull_request:47298", "pull_request:47163|pull_request:47210", "pull_request:47294|pull_request:47298", "pull_request:45157|pull_request:47387", "pull_request:46658|pull_request:47387", "pull_request:47143|pull_request:47163", + "pull_request:46196|pull_request:47143", "pull_request:46658|pull_request:47143", "pull_request:46658|pull_request:47322", - "pull_request:46196|pull_request:47143", "pull_request:47143|pull_request:47387", "pull_request:47143|pull_request:47322", "pull_request:45157|pull_request:47143", @@ -24287,445 +24361,438 @@ "pull_request:47184|pull_request:47383", "pull_request:40546|pull_request:43451", "pull_request:47306|pull_request:47367", - "pull_request:43451|pull_request:47143", - "pull_request:43451|pull_request:47163" + "pull_request:43451|pull_request:47143" ], - "prepared_review_unit_hash": "a76b859a1a187ddfb859e50da5e22606068d9478b959fa35968d71c62116c0d8", + "prepared_review_unit_hash": "e6f5cf15cf4e5b81fdccf3c653383a4b12ff9b88830755b6b332be6fa05a1cde", "component_budget": { - "node_count": 78, - "item_count": 78, - "soft_pair_count": 212, - "serialized_chars": 267543, - "estimated_input_tokens": 66886, - "estimated_eval_tokens": 134028 + "node_count": 79, + "item_count": 79, + "soft_pair_count": 214, + "serialized_chars": 270269, + "estimated_input_tokens": 67568, + "estimated_eval_tokens": 135392 }, "budget": { - "node_count": 45, - "item_count": 45, + "node_count": 46, + "item_count": 46, "soft_pair_count": 72, - "serialized_chars": 127570, - "estimated_input_tokens": 31893, - "estimated_eval_tokens": 64042 + "serialized_chars": 129459, + "estimated_input_tokens": 32365, + "estimated_eval_tokens": 64986 }, "overflow_policy": "truncate_then_skip", "trimmed": false, "aggressively_trimmed": false, "split": true, "analyst_result": { - "summary": "Only the CohereASR training-loss patches form a true duplicate cluster. The rest are mostly unrelated model additions, docs, exporter, or test updates that merely share broad subsystem themes.", - "confidence": 0.97, + "summary": "There is one real duplicate cluster around CohereASR training-loss double-shift handling; most other soft pairs are only superficially similar (same subsystem, docs, tests, or model-add PRs) and should stay separate.", + "confidence": 0.91, "canonical_issue_reason": null, - "canonical_pr_reason": "PR 47090 is the best canonical representative: it targets the same explicit issue, covers the full double-shift loss bug cleanly, and supersedes the earlier partial CohereASR fixes.", + "canonical_pr_reason": "47090 is the best canonical PR: it explicitly targets the CohereASR double-shift bug (#46894) and is the clearest, most complete version of that fix, superseding the earlier variants in 46895 and 46958.", "best_issue_reason": null, - "best_pr_reason": "PR 47090 is the strongest overall PR in the only real duplicate cluster, with the clearest fix and broadest affected-model coverage.", + "best_pr_reason": "47090 is the strongest overall canonical PR candidate because it is the only clearly duplicated fix cluster here with an explicit issue link and a concrete, mergeable code-path change.", "soft_edge_verdicts": [ + { + "left": "pull_request:41855", + "right": "pull_request:43823", + "accept": false, + "reason": "Different features: Mistral tokenizer methods vs adding MobileLLM support. Same broad area of Transformers, but not the same bug or change." + }, { "left": "pull_request:42668", "right": "pull_request:43823", "accept": false, - "reason": "Different PRs for different processor/model additions; same general area only." + "reason": "Both touch model/processor plumbing, but one is processor auto-loading robustness and the other is a new model addition. Not the same underlying change." }, { "left": "pull_request:43424", "right": "pull_request:47301", "accept": false, - "reason": "Different tests/features: ExecuTorch exportability vs assisted decoding docs." + "reason": "ExecuTorch dynamic-shape export test vs docs for multi-token prediction. No shared concrete code-path problem." }, { "left": "pull_request:43424", "right": "pull_request:47383", "accept": false, - "reason": "Different goals and code paths; one is export tests, the other integration tests." + "reason": "Dynamic-shape export test and Kimi integration tests are unrelated test additions for different concerns." }, { "left": "pull_request:43823", "right": "pull_request:45655", "accept": false, - "reason": "Unrelated model addition vs config-resolution fix." + "reason": "New MobileLLM support vs config-resolution ordering fix. Different code paths and unrelated changes." }, { "left": "pull_request:44827", "right": "pull_request:47383", "accept": false, - "reason": "Different change types: Mistral4 tests vs Kimi integration tests." + "reason": "Mistral4 test fixes vs Kimi integration tests. Same general testing domain only." }, { "left": "pull_request:44827", "right": "pull_request:45170", "accept": false, - "reason": "Different bugs: Mistral4 tests vs layrnorm/layernorm renaming." + "reason": "Mistral4 test fixes and layrnorm->layernorm renaming are different issues." }, { "left": "pull_request:43424", "right": "pull_request:44827", "accept": false, - "reason": "Both are test-related, but they fix different models and failures." + "reason": "One is an export test, the other is Mistral4 test/runtime cleanup. Not the same bug." }, { "left": "pull_request:43823", "right": "pull_request:43995", "accept": false, - "reason": "Unrelated model addition vs Falcon refactor." + "reason": "New MobileLLM model support vs Falcon output-interface refactor. Different functionality." }, { "left": "pull_request:43251", "right": "pull_request:43823", "accept": false, - "reason": "Loss-function fix vs new model addition; no shared bug." + "reason": "Cross-entropy kwargs passthrough fix vs adding MobileLLM support. No shared concrete fix." }, { "left": "pull_request:44827", "right": "pull_request:46048", "accept": false, - "reason": "Mistral4 test fix and TDT loss kernel support are unrelated." + "reason": "Mistral4 tests and TDT loss kernel support are unrelated." }, { "left": "pull_request:43424", "right": "pull_request:43823", "accept": false, - "reason": "Different PR scopes: exec export test vs model addition." + "reason": "Dynamic export testing vs model addition. Same broad area only." }, { "left": "pull_request:43251", "right": "pull_request:44827", "accept": false, - "reason": "Cross-entropy kwargs fix is unrelated to Mistral4 test issues." + "reason": "Loss utility kwargs handling vs Mistral4 tests. Not the same code-path problem." }, { "left": "pull_request:44827", "right": "pull_request:45655", "accept": false, - "reason": "Different concrete fixes; only broad library overlap." + "reason": "Mistral4 test cleanup and config-resolution fix are unrelated." }, { "left": "pull_request:43995", "right": "pull_request:45213", "accept": false, - "reason": "Falcon output-interface refactor and model-creation skill are unrelated." + "reason": "Falcon refactor vs model-creation skill. No overlap beyond generic tooling." }, { "left": "pull_request:44827", "right": "pull_request:45213", "accept": false, - "reason": "Mistral4 tests and model-creation skill do not share the same bug." + "reason": "Mistral4 tests vs a model-creation skill PR. Completely different changes." }, { "left": "pull_request:43995", "right": "pull_request:44827", "accept": false, - "reason": "Different code paths and different targets." + "reason": "Falcon output collection refactor vs Mistral4 test fixes. Different code paths." }, { "left": "pull_request:43823", "right": "pull_request:45213", "accept": false, - "reason": "New MobileLLM model support vs model-creation skill; not duplicates." + "reason": "MobileLLM model support vs model-creation skill docs/scripts. Not a duplicate." }, { "left": "pull_request:46895", "right": "pull_request:47090", "accept": true, - "reason": "Same CohereASR double-shift loss bug; 47090 is a more complete follow-up fix." + "reason": "Same CohereASR training-loss double-shift bug; 47090 is just the refined version of the same fix." }, { "left": "pull_request:46958", "right": "pull_request:47090", "accept": true, - "reason": "Same loss-shift bug in the same CohereASR path; 47090 subsumes it." + "reason": "Both patch CohereASR's double-shift loss handling on the same files and issue target." }, { "left": "pull_request:46895", "right": "pull_request:46958", "accept": true, - "reason": "Both patch the same CohereASR training-loss double-shift problem." + "reason": "They implement the same CohereASR loss fix with only implementation details differing." }, { "left": "pull_request:47173", "right": "pull_request:47292", "accept": false, - "reason": "Both are XPU test updates, but for different models and failures." + "reason": "Both are XPU-related test expectation updates, but they target different models and likely different failing cases, so they are not one concrete bug." }, { "left": "pull_request:47166", "right": "pull_request:47184", "accept": false, - "reason": "Different test maintenance changes; same hardware context only." + "reason": "Quantization test updates and multi-accelerator test adjustments are unrelated." }, { "left": "pull_request:46912", "right": "pull_request:47175", "accept": false, - "reason": "Different fixes: RoPE test expansion vs Olmo3 CI value updates." + "reason": "Both are test-value updates, but they cover different models and different CI issues; not mergeable as one change." }, { "left": "pull_request:47096", "right": "pull_request:47383", "accept": false, - "reason": "Exporter fixes versus Kimi integration tests; no shared concrete bug." + "reason": "Exportability fixes for Kimi vs Kimi integration tests; related model name, but different code-path problems." }, { "left": "pull_request:47184", "right": "pull_request:47292", "accept": false, - "reason": "Separate XPU-related test tweaks for different models." + "reason": "Both tweak accelerator-specific tests, but for different models and failures." }, { "left": "pull_request:47301", "right": "pull_request:47304", "accept": false, - "reason": "Docs-only changes for different topics." + "reason": "Docs for multi-token prediction vs a docs tweak about pipeline device placement. Different topics." }, { "left": "pull_request:47077", "right": "pull_request:47304", "accept": false, - "reason": "Japanese DistilBERT docs and MPS graph cache docs are unrelated." + "reason": "Japanese DistilBERT docs translation and MPS graph-cache docs are unrelated docs changes." }, { "left": "pull_request:46658", "right": "pull_request:46836", "accept": false, - "reason": "Different model additions; only shared new-model scaffolding." + "reason": "Step3.7 Flash model work vs Unlimited OCR model work. Both are model additions, not duplicates." }, { "left": "pull_request:47210", "right": "pull_request:47301", "accept": false, - "reason": "SDPA varlen routing change vs docs update; unrelated." + "reason": "Packed-SDPA varlen attention routing vs MTP docs. Completely different areas." }, { "left": "pull_request:47003", "right": "pull_request:47243", "accept": false, - "reason": "Both touch exporters, but they address different exporter behaviors and models." + "reason": "OpenVINO exporter support vs ExecuTorch exporter fixes. Same exporter family, but different concrete exporter paths." }, { "left": "pull_request:46509", - "right": "pull_request:47143", + "right": "pull_request:46836", "accept": false, - "reason": "Different model additions: Nemotron Omni vs NeuCodec." + "reason": "Nemotron Omni model support and Unlimited OCR model support are separate model integrations." }, { "left": "pull_request:47163", "right": "pull_request:47166", "accept": false, - "reason": "Different test suites and failure modes; same platform only." + "reason": "Paged-attention XPU tests vs quantization test extension. Different subsystems." }, { "left": "pull_request:47163", "right": "pull_request:47184", "accept": false, - "reason": "Both are XPU test-related, but not the same bug or model." + "reason": "Paged-attention XPU tests and qwen3_next accelerator test changes are unrelated." }, { "left": "pull_request:46509", "right": "pull_request:47163", "accept": false, - "reason": "Different topics: model support vs paged-attention XPU tests." + "reason": "Nemotron Omni model support vs paged-attention XPU test coverage. Different changes." }, { "left": "pull_request:45157", "right": "pull_request:46836", "accept": false, - "reason": "Prism GGUF support and Unlimited OCR are unrelated model/features." + "reason": "PrismML Bonsai support and Unlimited OCR support are different model additions." }, { "left": "pull_request:46836", "right": "pull_request:47387", "accept": false, - "reason": "Unlimited OCR support and one-to-many conversion support are different changes." + "reason": "Unlimited OCR support vs one-to-many conversion infrastructure. Not the same bug or feature." }, { "left": "pull_request:47163", "right": "pull_request:47292", "accept": false, - "reason": "Different XPU test failures for different models." + "reason": "Both are XPU-related test fixes, but for different models and failure modes." }, { "left": "pull_request:47291", "right": "pull_request:47322", "accept": false, - "reason": "Generation config deprecation and Quark config class update are unrelated." + "reason": "Continuous batching deprecation/config handling and AMD Quark config-class changes are unrelated quantization/config PRs." }, { - "left": "pull_request:47301", - "right": "pull_request:47336", + "left": "pull_request:45157", + "right": "pull_request:46509", "accept": false, - "reason": "Docs for assisted decoding vs offloading save-pretrained behavior are unrelated." + "reason": "PrismML Bonsai support and Nemotron Omni support are separate model integrations." }, { - "left": "pull_request:45157", - "right": "pull_request:46509", + "left": "pull_request:47301", + "right": "pull_request:47367", "accept": false, - "reason": "Different model workstreams with no shared bug." + "reason": "MTP docs and pipeline device-placement docs are unrelated documentation updates." }, { "left": "pull_request:47216", "right": "pull_request:47383", "accept": false, - "reason": "Compressed-tensors docs/quantization vs Kimi integration tests; unrelated." + "reason": "Compressed-tensors fp8 docs/code vs Kimi integration tests. Different domains." }, { "left": "pull_request:47336", "right": "pull_request:47383", "accept": false, - "reason": "Offloading save fix and Kimi integration tests are unrelated." + "reason": "Offloading save_pretrained fix vs Kimi integration tests. Different problems." }, { "left": "pull_request:46509", "right": "pull_request:47387", "accept": false, - "reason": "Nemotron Omni and one-to-many conversion support are different features." + "reason": "Nemotron Omni support vs one-to-many conversion support. Different features." }, { "left": "pull_request:47304", "right": "pull_request:47367", "accept": false, - "reason": "MPS graph-cache docs and pipeline device-placement docs are unrelated." + "reason": "MPS graph-cache docs and pipeline device-placement docs are separate doc changes." }, { "left": "pull_request:46912", "right": "pull_request:47298", "accept": false, - "reason": "RoPE test changes vs MI450 image-loading/model failures; different bugs." + "reason": "Both adjust expectations for hardware-specific test failures, but they touch different models and are not one concrete bug." }, { "left": "pull_request:47301", "right": "pull_request:47383", "accept": false, - "reason": "Docs-only MTP support vs Kimi integration tests; no duplicate." + "reason": "MTP docs vs Kimi integration tests. No overlap." }, { "left": "pull_request:47216", "right": "pull_request:47301", "accept": false, - "reason": "Quantization docs and assisted decoding docs are unrelated." + "reason": "Compressed-tensors fp8 work vs MTP docs. Not the same change." }, { "left": "pull_request:43451", "right": "pull_request:46836", "accept": false, - "reason": "Different new-model additions with no shared concrete code-path bug." - }, - { - "left": "pull_request:47175", - "right": "pull_request:47298", - "accept": false, - "reason": "Olmo3 CI updates and MI450 fixes target different failures." - }, - { - "left": "pull_request:46509", - "right": "pull_request:47301", - "accept": false, - "reason": "Model support PR vs docs PR; not duplicates." - }, - { - "left": "pull_request:47163", - "right": "pull_request:47210", - "accept": false, - "reason": "Paged-attention XPU tests and SDPA varlen routing are different changes." - }, - { - "left": "pull_request:47294", - "right": "pull_request:47298", - "accept": false, - "reason": "Kernel functions vs MI450 fixes; unrelated." + "reason": "Molmo2 model addition vs Unlimited OCR model addition. Same broad type only." }, { - "left": "pull_request:45157", - "right": "pull_request:47387", + "left": "pull_request:47143", + "right": "pull_request:47163", "accept": false, - "reason": "GGUF Prism support and conversion plumbing are not the same change." + "reason": "NeuCodec model support vs paged-attention XPU tests. Unrelated." }, { - "left": "pull_request:46658", - "right": "pull_request:47387", + "left": "pull_request:46196", + "right": "pull_request:47143", "accept": false, - "reason": "Step-3.7 model scaffold and one-to-many conversion support are different." + "reason": "wav2vec2 SDPA dispatch support vs NeuCodec model support. Different code paths." }, { - "left": "pull_request:47143", - "right": "pull_request:47163", + "left": "pull_request:46658", + "right": "pull_request:47143", "accept": false, - "reason": "NeuCodec support and paged-attention tests are unrelated." + "reason": "Step3.7 Flash model work and NeuCodec model support are separate model additions." }, { "left": "pull_request:46658", - "right": "pull_request:47143", + "right": "pull_request:47322", "accept": false, - "reason": "Step3.7 Flash support and NeuCodec support are different model additions." + "reason": "Step3.7 Flash model work vs AMD Quark config changes. Different areas." }, { - "left": "pull_request:46196", - "right": "pull_request:47143", + "left": "pull_request:46658", + "right": "pull_request:47387", "accept": false, - "reason": "SDPA dispatch support and NeuCodec support are unrelated." + "reason": "Step3.7 Flash model work vs one-to-many conversion support. Different problems." }, { "left": "pull_request:47143", "right": "pull_request:47387", "accept": false, - "reason": "Different features: NeuCodec support vs conversion framework work." + "reason": "NeuCodec model support vs one-to-many conversion infrastructure. Not a duplicate." }, { "left": "pull_request:47143", "right": "pull_request:47322", "accept": false, - "reason": "NeuCodec support and Quark config changes are unrelated." + "reason": "NeuCodec model support and Quark config-class changes are unrelated." }, { "left": "pull_request:45157", "right": "pull_request:47143", "accept": false, - "reason": "Different model families and no shared concrete bug." + "reason": "PrismML Bonsai support vs NeuCodec support. Different model additions." }, { "left": "pull_request:47301", "right": "pull_request:47367", "accept": false, - "reason": "Assisted decoding docs and pipeline docs are unrelated." + "reason": "Docs on MTP vs pipeline device placement. Separate documentation topics." }, { "left": "pull_request:43451", "right": "pull_request:47243", "accept": false, - "reason": "Molmo2 model addition and ExecuTorch fixes are unrelated." + "reason": "Molmo2 model addition vs ExecuTorch exporter fixes. Different features." }, { "left": "pull_request:47096", "right": "pull_request:47243", "accept": false, - "reason": "Both touch exporter code, but they fix different exporter behaviors." + "reason": "Kimi exportability changes vs exporter framework fixes for a different model path. Not the same concrete issue." + }, + { + "left": "pull_request:43451", + "right": "pull_request:47291", + "accept": false, + "reason": "Molmo2 model support and continuous-batching deprecation/config changes are unrelated." + }, + { + "left": "pull_request:43451", + "right": "pull_request:47175", + "accept": false, + "reason": "Molmo2 model support vs OLMo3 CI test updates. Different purposes." }, { "left": "pull_request:47166", "right": "pull_request:47383", "accept": false, - "reason": "Different test updates with no shared concrete failure." + "reason": "Quantization test updates vs Kimi integration tests. Different subsystems." }, { "left": "pull_request:47184", "right": "pull_request:47383", "accept": false, - "reason": "Qwen3Next/VideOMT test changes are not the same issue as Kimi integration tests." + "reason": "Accelerator-specific test tweaks vs Kimi integration tests. Not the same bug." }, { "left": "pull_request:40546", "right": "pull_request:43451", "accept": false, - "reason": "VibeVoice and Molmo2 are independent model additions." + "reason": "VibeVoice model support and Molmo2 model support are distinct additions." }, { "left": "pull_request:47306", "right": "pull_request:47367", "accept": false, - "reason": "Neuron experiment scripts and pipeline docs are unrelated." + "reason": "Neuron experiment scripts vs pipeline documentation. Different artifacts and goals." }, { "left": "pull_request:43451", "right": "pull_request:47143", "accept": false, - "reason": "Different model support PRs with no shared bug or change." + "reason": "Molmo2 and NeuCodec are different model integrations, even though both add support." } ] }, @@ -24747,13 +24814,14 @@ "source": "llm", "cache_hit": false, "model": "gpt-5.4-mini?service_tier=flex", - "cluster_id": "cluster-40546-78", + "cluster_id": "cluster-40546-79", "nodes": [ "pull_request:43451", "pull_request:46509", "pull_request:47003", "pull_request:47096", "pull_request:47143", + "pull_request:47163", "pull_request:47298", "pull_request:47306", "pull_request:47322", @@ -24761,6 +24829,7 @@ "pull_request:47387" ], "soft_pairs": [ + "pull_request:43451|pull_request:47163", "pull_request:43451|pull_request:47383", "pull_request:43451|pull_request:46509", "pull_request:43451|pull_request:47096", @@ -24777,124 +24846,130 @@ "pull_request:47003|pull_request:47387", "pull_request:47096|pull_request:47143" ], - "prepared_review_unit_hash": "a38687d4750361c0974b83b7e0ed6ac1c5734c0cc8b02343ad6265b8d61e8498", + "prepared_review_unit_hash": "5d02b0a2fe74b181627c8045d296fa982cf15760c6148dde5ce05215a9002a36", "component_budget": { - "node_count": 78, - "item_count": 78, - "soft_pair_count": 212, - "serialized_chars": 267543, - "estimated_input_tokens": 66886, - "estimated_eval_tokens": 134028 + "node_count": 79, + "item_count": 79, + "soft_pair_count": 214, + "serialized_chars": 270269, + "estimated_input_tokens": 67568, + "estimated_eval_tokens": 135392 }, "budget": { - "node_count": 10, - "item_count": 10, - "soft_pair_count": 15, - "serialized_chars": 29356, - "estimated_input_tokens": 7339, - "estimated_eval_tokens": 14934 + "node_count": 11, + "item_count": 11, + "soft_pair_count": 16, + "serialized_chars": 31523, + "estimated_input_tokens": 7881, + "estimated_eval_tokens": 16018 }, "overflow_policy": "truncate_then_skip", "trimmed": false, "aggressively_trimmed": false, "split": true, "analyst_result": { - "summary": "The cluster is heterogeneous: it mixes unrelated model-addition PRs, exporter/backend work, hardware-specific fixes, and test-only changes. The soft similarity is driven by broad Transformers infrastructure overlap, not duplicate underlying changes.", - "confidence": 0.97, + "summary": "This cluster is heterogeneous: standalone model additions (Molmo2, Nemotron Omni, NeuCodec), exporter/infra work (OpenVINO exporter, Kimi exportability, one-to-many conversion support), a hardware-specific image-loading fix, and test/experiment-only PRs. They are not duplicate changes and should be split for triage.", + "confidence": 0.98, "canonical_issue_reason": null, - "canonical_pr_reason": "No clear canonical PR: none of these PRs are the same concrete change, and several touch only broadly related infrastructure or unrelated model support areas.", + "canonical_pr_reason": "No single PR canonically represents this cluster; the items span unrelated features and fixes rather than one underlying change.", "best_issue_reason": null, - "best_pr_reason": "No single PR is a good representative duplicate target for the cluster; the items span distinct features and fixes rather than one mergeable change.", + "best_pr_reason": "No PR is a good global representative for duplicate triage here; 47096 is the closest exporter-focused item, but it does not subsume the rest of the cluster.", "soft_edge_verdicts": [ + { + "left": "pull_request:43451", + "right": "pull_request:47163", + "accept": false, + "reason": "Molmo2 model onboarding vs XPU paged-attention test additions; different change types and code paths." + }, { "left": "pull_request:43451", "right": "pull_request:47383", "accept": false, - "reason": "Molmo2 model support vs Kimi integration tests; shared model-test scope is too broad and they do not fix the same bug or add the same feature." + "reason": "Molmo2 model onboarding vs Kimi integration tests; unrelated PRs." }, { "left": "pull_request:43451", "right": "pull_request:46509", "accept": false, - "reason": "Molmo2 vs Nemotron Omni are separate model integrations with different files and APIs; not the same underlying change." + "reason": "Different model additions (Molmo2 vs Nemotron Omni) with separate configs, docs, and integrations." }, { "left": "pull_request:43451", "right": "pull_request:47096", "accept": false, - "reason": "Molmo2 addition and Kimi exportability work are unrelated model features, not one mergeable change." + "reason": "Model addition vs exporter/Kimi exportability work; not the same underlying bug or feature." }, { "left": "pull_request:43451", "right": "pull_request:47003", "accept": false, - "reason": "Molmo2 model docs/code vs OpenVINO exporter work are different subsystems and different concrete changes." + "reason": "Model onboarding vs OpenVINO exporter implementation; distinct code areas and goals." }, { "left": "pull_request:47003", "right": "pull_request:47096", "accept": false, - "reason": "OpenVINO exporter support and Kimi exportability are related only at the broad exporter level, but they target different code paths and are not the same fix." + "reason": "Both touch export-related code, but one adds OpenVINO exporter support and the other changes Kimi exportability/attention handling; not one mergeable fix." }, { "left": "pull_request:43451", "right": "pull_request:47322", "accept": false, - "reason": "Molmo2 support and AMD Quark config adjustments are unrelated; one adds a model, the other changes quantization config handling." + "reason": "Molmo2 model onboarding vs AMD Quark config-class rename; unrelated features." }, { "left": "pull_request:47096", "right": "pull_request:47322", "accept": false, - "reason": "Kimi exportability changes and Quark config-class updates do not address the same problem or code path." + "reason": "Kimi exportability changes vs Quark config-class changes; different exporter stacks and problem statements." }, { "left": "pull_request:43451", "right": "pull_request:47298", "accept": false, - "reason": "Molmo2 model addition and MI450 failure fixes are unrelated; the latter is a platform-specific stability patch." + "reason": "Model onboarding vs MI450 image-decoding fallback fix; unrelated." }, { "left": "pull_request:47096", "right": "pull_request:47298", "accept": false, - "reason": "Kimi exportability and MI450 model-failure fixes are different issues with no shared concrete change." + "reason": "Kimi exporter work vs MI450 image decoder compatibility fixes; different code paths and targets." }, { "left": "pull_request:43451", "right": "pull_request:47306", "accept": false, - "reason": "Molmo2 support and Neuron experiments are unrelated; the latter is an experimental training script/integration change." + "reason": "Model onboarding vs Trainium/neuron experiment scripts and tensor-parallel integration; unrelated." }, { "left": "pull_request:47096", "right": "pull_request:47306", "accept": false, - "reason": "Kimi exportability and Neuron experiments do not fix the same code-path problem." + "reason": "Exporter/Kimi work vs Neuron training experiments; not the same concrete change." }, { "left": "pull_request:47003", "right": "pull_request:47306", "accept": false, - "reason": "OpenVINO exporter work and Neuron experiments are different backend efforts and cannot plausibly be merged as one PR." + "reason": "OpenVINO exporter work vs Neuron training experiments; different subsystems and objectives." }, { "left": "pull_request:47096", "right": "pull_request:47387", "accept": false, - "reason": "Kimi exportability tweaks and one-to-many conversion support are not the same change; the latter is a core loading/transformer framework enhancement." + "reason": "Kimi exportability changes vs one-to-many conversion support in model loading; separate core mechanisms." }, { "left": "pull_request:47003", "right": "pull_request:47387", "accept": false, - "reason": "OpenVINO exporter implementation and one-to-many conversion support target different layers of the stack and different problems." + "reason": "OpenVINO exporter additions vs one-to-many model-loading conversions; not the same underlying change." }, { "left": "pull_request:47096", "right": "pull_request:47143", "accept": false, - "reason": "Kimi exportability work and NeuCodec support are separate model/backend additions with no shared underlying bug or feature." + "reason": "Kimi exportability work vs NeuCodec model onboarding; unrelated PRs." } ] }, @@ -24924,12 +24999,12 @@ "soft_pairs": [ "pull_request:40755|pull_request:41224" ], - "prepared_review_unit_hash": "b4e3e8c03f35daa68605faebd2b5e286d16436c8b3095b6ef1ce41f74be6b004", + "prepared_review_unit_hash": "48e1dead2cc93d13a679f6ecc395b132f04b5ab40205a9e86568d822b0f9aeb4", "component_budget": { "node_count": 2, "item_count": 2, "soft_pair_count": 1, - "serialized_chars": 4674, + "serialized_chars": 4675, "estimated_input_tokens": 1169, "estimated_eval_tokens": 2594 }, @@ -24937,7 +25012,7 @@ "node_count": 2, "item_count": 2, "soft_pair_count": 1, - "serialized_chars": 4674, + "serialized_chars": 4675, "estimated_input_tokens": 1169, "estimated_eval_tokens": 2594 }, @@ -24946,18 +25021,18 @@ "aggressively_trimmed": false, "split": false, "analyst_result": { - "summary": "These two PRs are unrelated feature additions for different models: one adds TimesFM covariate forecasting, the other adds DINOv3 image classification support. They do not describe the same bug or change.", - "confidence": 0.99, + "summary": "The two PRs are both feature additions in different model families and do not look like the same underlying change. One adds covariate forecasting support to TimesFM; the other adds DINOv3 image classification support.", + "confidence": 0.97, "canonical_issue_reason": null, - "canonical_pr_reason": null, + "canonical_pr_reason": "No single canonical PR: these are unrelated feature PRs in different code paths and model families.", "best_issue_reason": null, - "best_pr_reason": null, + "best_pr_reason": "Neither PR is a duplicate of the other; each addresses a distinct model capability and should be triaged independently.", "soft_edge_verdicts": [ { "left": "pull_request:40755", "right": "pull_request:41224", "accept": false, - "reason": "Different model families and different code paths: TimesFM forecasting-with-covariates vs DINOv3 image classification. They are not the same underlying change and could not plausibly be merged into one PR." + "reason": "Different model features and code paths: TimesFM covariate forecasting vs. DINOv3 image classification. They are not the same underlying bug/change and could not plausibly be merged into one PR." } ] }, @@ -25050,7 +25125,7 @@ "soft_pairs": [ "pull_request:42975|pull_request:42976" ], - "prepared_review_unit_hash": "67dbe6b2ef95c942b2c3ec91b0a30ccbc93369724028c1ed5ed1fdcfde3978bb", + "prepared_review_unit_hash": "b95ae04487e9562c25141b65361f567dc16d74439fd8e2a5bb9b72c129c8c2c6", "component_budget": { "node_count": 2, "item_count": 2, @@ -25072,18 +25147,18 @@ "aggressively_trimmed": false, "split": false, "analyst_result": { - "summary": "These are related dependency-maintenance PRs touching overlapping GitHub Actions workflows, but they target different action upgrades and different compatibility goals, so they should not be merged as duplicates.", - "confidence": 0.95, + "summary": "Both items are maintenance PRs updating GitHub Actions versions in CI workflows, but they are not the same underlying change: one is a broad Node 24 compatibility sweep across many workflow actions, while the other is a separate upgrade focused on latest action versions such as docker/build-push-action v6.", + "confidence": 0.84, "canonical_issue_reason": null, - "canonical_pr_reason": "PR 42975 is the more specific change set: it consistently upgrades workflow actions for Node 24 compatibility. PR 42976 is a separate maintenance pass focused on newer action versions, especially docker/build-push-action, and is not the same concrete fix.", + "canonical_pr_reason": "PR 42975 is the better canonical PR because it is the broader, more explicit workflow-actions compatibility upgrade and touches more of the shared GitHub Actions surface.", "best_issue_reason": null, - "best_pr_reason": "PR 42975 is the better representative because it has a single clear purpose and a tighter scope. PR 42976 is broader and less clearly tied to the same underlying problem.", + "best_pr_reason": "PR 42975 is the strongest representative of the cluster since it has the clearest stated goal and wider scope; PR 42976 is a narrower, separate dependency bump.", "soft_edge_verdicts": [ { "left": "pull_request:42975", "right": "pull_request:42976", "accept": false, - "reason": "Both are workflow dependency bumps, but they address different actions and different upgrade motivations. They do not appear to fix the same concrete code-path problem and would not naturally collapse into one PR." + "reason": "Do not merge: they overlap on workflow maintenance but update different action families and have different objectives/scope, so they are not the same concrete code-path change." } ] }, @@ -25118,7 +25193,7 @@ "pull_request:43448|pull_request:45490", "pull_request:45490|pull_request:46562" ], - "prepared_review_unit_hash": "4569f0c30adb6cdc11c6d5a86394448596e5dc03a3a3e194ff1e33ff6cff8f4b", + "prepared_review_unit_hash": "f6d69f404ce7f2ea7b4d5612a560132fabba559e136fb83de9e6ef0afc6fd393", "component_budget": { "node_count": 4, "item_count": 4, @@ -25140,36 +25215,36 @@ "aggressively_trimmed": false, "split": false, "analyst_result": { - "summary": "These are four separate model-addition PRs (ViT NEPA, Molmo, CTSM, Step3p7). They share boilerplate around docs and auto registries, but they target different models and code paths, so none should be clustered as duplicates.", + "summary": "These pull requests all add different model integrations and documentation, with only boilerplate overlap in shared auto-mapping/docs files. They are related by subsystem, not by the same underlying change, so none should be merged as duplicates.", "confidence": 0.98, "canonical_issue_reason": null, - "canonical_pr_reason": "No canonical PR: the changes are for different models and different implementations, with only superficial overlap in shared docs/auto-mapping scaffolding.", + "canonical_pr_reason": "No single canonical PR: ViT NEPA, Molmo, CTSM, and Step3p7 are distinct model additions with different implementations, tests, and model-specific files.", "best_issue_reason": null, - "best_pr_reason": "No single PR is a better canonical representative for the cluster because there is no duplicate bug/change here; each PR stands on its own model integration.", + "best_pr_reason": "No global best duplicate candidate: each PR targets a separate model integration rather than the same bug or feature, so there is no meaningful representative duplicate PR.", "soft_edge_verdicts": [ { "left": "pull_request:42978", "right": "pull_request:43448", "accept": false, - "reason": "ViT NEPA and Molmo are different model integrations with distinct model files, tests, and docs; shared auto-mapping/docs edits are just framework boilerplate." + "reason": "Different model additions (ViT NEPA vs Molmo). Shared docs and auto-registry edits are boilerplate for adding a model, not the same code-path fix." }, { "left": "pull_request:42978", "right": "pull_request:45490", "accept": false, - "reason": "ViT NEPA vs CTSM are unrelated model additions. The overlap in toctree/auto files does not indicate the same underlying code change or fix." + "reason": "ViT NEPA and CTSM are unrelated model implementations. Overlap is limited to generic docs/auto-mapping files, so they are not duplicates." }, { "left": "pull_request:43448", "right": "pull_request:45490", "accept": false, - "reason": "Molmo and CTSM add different model families and separate converters/modeling/processor code paths; they are not mergeable as one PR." + "reason": "Molmo and CTSM add different architectures and model-specific files. Same subsystem, but not the same concrete change." }, { "left": "pull_request:45490", "right": "pull_request:46562", "accept": false, - "reason": "CTSM and Step3p7 are distinct model ports with different model-specific modules and tests. Similar registry/documentation edits are not enough to treat them as duplicates." + "reason": "CTSM and Step3p7 are separate model launches with different modeling/processing code. Shared auto-mapping updates do not make them the same PR." } ] }, @@ -25191,7 +25266,7 @@ "source": "llm", "cache_hit": false, "model": "gpt-5.4-mini?service_tier=flex", - "cluster_id": "cluster-43291-7", + "cluster_id": "cluster-43291-8", "nodes": [ "pull_request:43291", "pull_request:43665", @@ -25199,7 +25274,8 @@ "pull_request:46380", "pull_request:46539", "pull_request:46886", - "pull_request:46922" + "pull_request:46922", + "pull_request:47393" ], "soft_pairs": [ "pull_request:43291|pull_request:43665", @@ -25207,6 +25283,7 @@ "pull_request:43291|pull_request:46380", "pull_request:43291|pull_request:46922", "pull_request:43665|pull_request:46539", + "pull_request:43665|pull_request:47393", "pull_request:43665|pull_request:44815", "pull_request:43665|pull_request:46380", "pull_request:43665|pull_request:46922", @@ -25218,133 +25295,139 @@ "pull_request:46539|pull_request:46922", "pull_request:46886|pull_request:46922" ], - "prepared_review_unit_hash": "a44a4cd3644ecce752fc648729cecc3be87f17bc60ee269c5293172cb2ccee3d", + "prepared_review_unit_hash": "a880f5f84182b7dde2638f57d301fa9201a41ce117e4f013fa2d76c1d57c801a", "component_budget": { - "node_count": 7, - "item_count": 7, - "soft_pair_count": 15, - "serialized_chars": 19601, - "estimated_input_tokens": 4901, - "estimated_eval_tokens": 10058 + "node_count": 8, + "item_count": 8, + "soft_pair_count": 16, + "serialized_chars": 21115, + "estimated_input_tokens": 5279, + "estimated_eval_tokens": 10814 }, "budget": { - "node_count": 7, - "item_count": 7, - "soft_pair_count": 15, - "serialized_chars": 19601, - "estimated_input_tokens": 4901, - "estimated_eval_tokens": 10058 + "node_count": 8, + "item_count": 8, + "soft_pair_count": 16, + "serialized_chars": 21115, + "estimated_input_tokens": 5279, + "estimated_eval_tokens": 10814 }, "overflow_policy": "truncate_then_skip", "trimmed": false, "aggressively_trimmed": false, "split": false, "analyst_result": { - "summary": "This cluster is mostly heterogeneous: Whisper test/tokenizer cleanup, unrelated test assertions, an FP8 dequant/loading fix, a generation-validation change, a CI memory workaround, and two closely related YOLOS device-mismatch PRs. The only plausible duplicate/variant pair is the two YOLOS PRs, which share the same issue target and code path but are not identical fixes.", - "confidence": 0.82, + "summary": "This cluster is mostly unrelated PRs with weak title-based similarity; the only clear duplicate-like pair is the two YOLOS fixes, where one PR is a more complete version of the same device-mismatch change.", + "confidence": 0.95, "canonical_issue_reason": null, - "canonical_pr_reason": "No strong canonical PR winner is supported by the evidence. 46886 and 46922 are clearly related YOLOS variants, but 46922 also adds a separate runtime device-placement fix on top of the shared _no_split_modules change, so the pair should be treated as close variants rather than one being an obvious canonical duplicate.", + "canonical_pr_reason": "pull_request:46922 is the best canonical PR because it directly targets the YOLOS device-mismatch bug, has an explicit linked issue, and subsumes the narrower no-split-module change in pull_request:46886.", "best_issue_reason": null, - "best_pr_reason": "If a representative PR is needed, 46922 is the slightly broader representative because it includes the shared YOLOS fix plus the extra device-placement correction. That said, the evidence is still only strong enough to call 46886 and 46922 related variants, not a definitive duplicate/canonical match.", + "best_pr_reason": "pull_request:46922 is the strongest representative of the cluster\u2019s only real duplicate subtopic: it fixes the same YOLOS code path with a concrete forward-pass device placement change plus the supporting split-module adjustment.", "soft_edge_verdicts": [ { "left": "pull_request:43291", "right": "pull_request:43665", "accept": false, - "reason": "Whisper tokenizer/test changes vs unrelated CLIP/ViT test edits; only superficial test-fix similarity." + "reason": "Both are test-focused, but they touch different models and different test files; no shared bug or mergeable single change." }, { "left": "pull_request:43291", "right": "pull_request:44815", "accept": false, - "reason": "Different subsystems and bugs: Whisper decode logic vs FP8 dequant/loading fixes." + "reason": "Whisper tokenizer tests vs FP8 dequant/loading code are unrelated changes despite superficial \"fix\" wording." }, { "left": "pull_request:43291", "right": "pull_request:46380", "accept": false, - "reason": "Whisper fix vs generation watermarking validation and model test updates; not the same change." + "reason": "Whisper test cleanup and generation/CIS validation fixes are different problems in different subsystems." }, { "left": "pull_request:43291", "right": "pull_request:46922", "accept": false, - "reason": "Whisper tokenizer/tests vs YOLOS device-mismatch fix; unrelated code paths." + "reason": "Whisper tokenizer/test changes are unrelated to YOLOS device placement and split-module handling." }, { "left": "pull_request:43665", "right": "pull_request:46539", "accept": false, - "reason": "Unrelated test-only edits in model suites vs a CI memory/device_map workaround." + "reason": "Both modify tests/infra, but one is model test assertions and the other is CI memory/device-map handling; not the same underlying change." + }, + { + "left": "pull_request:43665", + "right": "pull_request:47393", + "accept": false, + "reason": "Model tests and a README URL edit are clearly unrelated." }, { "left": "pull_request:43665", "right": "pull_request:44815", "accept": false, - "reason": "Both touch tests or fixes broadly, but the underlying bugs are different and not mergeable into one PR." + "reason": "CLIP/ViT test edits are not the same as the FP8 dequant/loading fix." }, { "left": "pull_request:43665", "right": "pull_request:46380", "accept": false, - "reason": "CLIP/ViT test edits are unrelated to generation config/watermarking fixes." + "reason": "Different models and different failure modes; no concrete shared code path." }, { "left": "pull_request:43665", "right": "pull_request:46922", "accept": false, - "reason": "Different models and issues: test assertion noise vs YOLOS runtime device mismatch." + "reason": "Model test files for CLIP/ViT do not match the YOLOS forward-pass/device-mismatch fix." }, { "left": "pull_request:44815", "right": "pull_request:46539", "accept": false, - "reason": "FP8 dequant/loading code changes are unrelated to Phimoe CI environment patching." + "reason": "FP8 dequant/loading logic and CI/device-map integration are unrelated subsystems." }, { "left": "pull_request:44815", "right": "pull_request:46380", "accept": false, - "reason": "Different concrete problems: FP8 quantization/dequant path vs generation validation and Whisper-related updates." + "reason": "These address different generation/loading issues and should not be merged as one PR." }, { "left": "pull_request:44815", "right": "pull_request:46922", "accept": false, - "reason": "No shared bug: FP8 kernel/loading fix vs YOLOS device placement fix." + "reason": "No shared code path beyond both touching model internals; the bugs are different." }, { "left": "pull_request:46380", "right": "pull_request:46539", "accept": false, - "reason": "Generation config/watermarking validation is unrelated to the CI CPU-memory limit workaround." + "reason": "Generation watermarking validation and Phimoe CI memory setup are unrelated." }, { "left": "pull_request:46380", "right": "pull_request:46922", "accept": false, - "reason": "Different models and code paths: generation validation/Whisper vs YOLOS device mismatch." + "reason": "Generation config validation and YOLOS device handling are different features and failures." }, { "left": "pull_request:46539", "right": "pull_request:46922", "accept": false, - "reason": "CI environment patching for Phimoe is unrelated to YOLOS model execution behavior." + "reason": "CI runner memory limits and YOLOS device mismatch are separate problems." }, { "left": "pull_request:46886", "right": "pull_request:46922", "accept": true, - "reason": "Both target the same YOLOS issue on the same model file and share the _no_split_modules fix; 46922 additionally adds a runtime device-placement correction, so they are close variants of the same underlying bug/change." + "reason": "Both are YOLOS fixes for the same device-mismatch issue in the same file; 46922 extends the same fix with the forward-pass device move and the split-module adjustment, so they could plausibly be one PR." } ] }, "evaluator_result": { - "accept": false, - "feedback": "The summary is mostly grounded, but the canonical selection is too strong for the evidence. PR 46886 and 46922 are clearly related and mergeable for the same YOLOS issue, but the packet does not support calling 46886 the best canonical representative or 46922 the \"narrower\" one; 46922 actually includes an extra runtime device-placement fix in addition to the shared _no_split_modules change. A more conservative conclusion would be that the two YOLOS PRs are closely related duplicates/variants, without asserting a clear canonical winner." + "accept": true, + "feedback": "The analyst summary is grounded in the packet and stays conservative. The non-YOLOS pairs are reasonably judged unrelated, and the 46886/46922 pair is plausibly mergeable as one YOLOS device-mismatch bugfix, with 46922 looking like the more complete version." }, "evaluator_used": true, - "retried": true, + "retried": false, "accepted_nontrivial_soft_edge": true, "error_kind": null, "error_message": null @@ -25371,52 +25454,52 @@ "pull_request:43488|pull_request:43532", "pull_request:43532|pull_request:46747" ], - "prepared_review_unit_hash": "65182f6be89bcd5279253b7391c8cac71dfa9ff868e9d498b076c8e841b09549", + "prepared_review_unit_hash": "582e7ad64dad1fbdde342cd39cea6374d6eceab571eae652513411c98ede047c", "component_budget": { "node_count": 3, "item_count": 3, "soft_pair_count": 3, - "serialized_chars": 6346, - "estimated_input_tokens": 1587, - "estimated_eval_tokens": 3430 + "serialized_chars": 6349, + "estimated_input_tokens": 1588, + "estimated_eval_tokens": 3432 }, "budget": { "node_count": 3, "item_count": 3, "soft_pair_count": 3, - "serialized_chars": 6346, - "estimated_input_tokens": 1587, - "estimated_eval_tokens": 3430 + "serialized_chars": 6349, + "estimated_input_tokens": 1588, + "estimated_eval_tokens": 3432 }, "overflow_policy": "truncate_then_skip", "trimmed": false, "aggressively_trimmed": false, "split": false, "analyst_result": { - "summary": "These are three separate '[don't merge]' PRs used for different repo/CI checks; they are not duplicates of one another.", - "confidence": 0.98, + "summary": "These PRs are not duplicates: one is a repo-bot/version-format test change, one is a workflow/diff-display change, and one is a security-gate workflow tweak. The similarity is mostly superficial (shared testing/workflow context), not the same underlying fix or feature.", + "confidence": 0.94, "canonical_issue_reason": null, - "canonical_pr_reason": "No canonical PR: the three PRs change different surfaces and purposes (setup/version + test bot checks, comment/diff workflow plumbing, and a security-gate workflow trigger).", + "canonical_pr_reason": "No single canonical PR: the three changes target different code paths and CI concerns, so they should not be collapsed into one duplicate cluster.", "best_issue_reason": null, - "best_pr_reason": "PR 43532 is the broadest representative of the cluster because it touches several CI/support files, but it is still distinct from the others.", + "best_pr_reason": "If forced to pick the most self-contained reference, PR #43488 is the cleanest standalone change, but it is not a canonical representative for the others because they address unrelated workflow behavior.", "soft_edge_verdicts": [ { "left": "pull_request:43488", "right": "pull_request:46747", "accept": false, - "reason": "Setup/version/test-bot changes are unrelated to the security-gate workflow trigger change." + "reason": "Different changes: #43488 edits packaging/test files for a repo-bot/version-format check, while #46747 changes the security-gate workflow trigger from pull_request_target to pull_request. Same broad CI area, but not the same bug or mergeable change set." }, { "left": "pull_request:43488", "right": "pull_request:43532", "accept": false, - "reason": "They both mention test tooling, but one edits setup/versioning and a Vit test while the other rewires multiple CI/comment workflows; not the same code-path problem." + "reason": "#43488 is about dependency/version-format and a Vit test, whereas #43532 restructures multiple workflows and testing utilities to show diffs. Shared test file is incidental; these are not the same underlying change." }, { "left": "pull_request:43532", "right": "pull_request:46747", "accept": false, - "reason": "Both touch GitHub workflows, but they address different mechanisms: PR-comment/diff reporting versus a PR security gate." + "reason": "#43532 updates comment/diff-display CI behavior across several files, but #46747 only changes the security gate workflow trigger. They touch the same repo area, not the same concrete problem." } ] }, @@ -25446,7 +25529,7 @@ "soft_pairs": [ "pull_request:43612|pull_request:43613" ], - "prepared_review_unit_hash": "462540f90ba9bfa482e61bef0787607024a31bc7477c167006d23ed124b75b5c", + "prepared_review_unit_hash": "de7832969d52e56240bb780ce903833a1e3964d95d53dad6b9c810843aa893ca", "component_budget": { "node_count": 2, "item_count": 2, @@ -25468,18 +25551,18 @@ "aggressively_trimmed": false, "split": false, "analyst_result": { - "summary": "Two open PRs add different segmentation pipelines: one for Promptable Concept Segmentation and one for Promptable Visual Segmentation. They share pipeline/docs scaffolding, but they target different tasks and different model/test surfaces, so they should not be merged as duplicates.", + "summary": "Both PRs add new, similarly structured segmentation pipelines, but they target different concrete features: promptable concept segmentation vs promptable visual segmentation. The overlap is mostly scaffolding/docs, not the same underlying bug or change.", "confidence": 0.95, "canonical_issue_reason": null, - "canonical_pr_reason": "PR #43613 is the broader representative: it adds a new visual segmentation pipeline with more model integrations and test coverage, making it the best stand-in for this similarity cluster.", + "canonical_pr_reason": "PR 43613 is the better representative of the cluster because it is the broader, more complete pipeline addition with wider model support and tests.", "best_issue_reason": null, - "best_pr_reason": "PR #43613 is the strongest cluster exemplar because it touches the most related code paths and documentation around the new segmentation pipeline family.", + "best_pr_reason": "PR 43613 is the best single PR to reference: it has the larger scope and more integrated support, so it is the most representative artifact in this cluster.", "soft_edge_verdicts": [ { "left": "pull_request:43612", "right": "pull_request:43613", "accept": false, - "reason": "Different underlying feature requests: concept segmentation vs visual segmentation. Shared docs/pipeline boilerplate is not enough to treat them as the same concrete change, and they would not plausibly be merged into a single PR without conflating two distinct tasks." + "reason": "Reject: these are separate pipeline features for different tasks and code paths. They share documentation and registration scaffolding, but they do not appear to fix or implement the same concrete change." } ] }, @@ -25509,7 +25592,7 @@ "soft_pairs": [ "pull_request:43651|pull_request:43663" ], - "prepared_review_unit_hash": "16380997705ffba1a958419674ab7f8b52685c599b345c37b56ce37d0b41dad8", + "prepared_review_unit_hash": "02ace308d0e604b2930a8b915a88709693a38b84a309412845254fb8908ff0f4", "component_budget": { "node_count": 2, "item_count": 2, @@ -25531,18 +25614,18 @@ "aggressively_trimmed": false, "split": false, "analyst_result": { - "summary": "Both pull requests modify `trainer.py`, but they address different trainer extension points: one changes gradient-accumulation loss scaling, the other refactors signature-column filtering for PEFT/custom trainers. They are related to Trainer extensibility, but not the same underlying bug/change and should not be merged as duplicates.", - "confidence": 0.93, + "summary": "Both PRs are trainer extensibility tweaks in the same file, but they address different code paths: one changes gradient-accumulation loss scaling, the other refactors signature-column filtering. They are not the same underlying change.", + "confidence": 0.91, "canonical_issue_reason": null, - "canonical_pr_reason": null, + "canonical_pr_reason": "PR #43651 is the more concrete code-path change: it alters training_step loss scaling behavior and adds a new hook for custom trainers.", "best_issue_reason": null, - "best_pr_reason": "PR #43651 is the more concrete and self-contained change, adding a loss-scaling hook in `training_step`; however, it is only a representative of this cluster, not a true duplicate canonical.", + "best_pr_reason": "PR #43651 is the best representative of the cluster because it is the more directly behavior-changing trainer fix; PR #43663 is a separate refactor around input-column filtering.", "soft_edge_verdicts": [ { "left": "pull_request:43651", "right": "pull_request:43663", "accept": false, - "reason": "Different code paths and goals: #43651 changes gradient accumulation loss normalization, while #43663 adds an overridable signature-column helper for input filtering. They are not the same bug or a single mergeable fix." + "reason": "Shared trainer.py location and extensibility theme, but they fix different mechanisms: gradient-accumulation loss scaling vs signature-based column filtering. Not mergeable as one PR." } ] }, @@ -25572,7 +25655,7 @@ "soft_pairs": [ "pull_request:43656|pull_request:43842" ], - "prepared_review_unit_hash": "bbf93bd2733599f6f83e863f6e32f3e3bd3b1559cd39540869444b05e3191b63", + "prepared_review_unit_hash": "f681f7a59774fd774a20bcf60ccb2160fd48aa679445fbe805fc24ca96978ea3", "component_budget": { "node_count": 2, "item_count": 2, @@ -25594,25 +25677,28 @@ "aggressively_trimmed": false, "split": false, "analyst_result": { - "summary": "Two open PRs touch the same CLI file and both mention NameError, but they appear to fix different annotation/import problems rather than the same bug.", - "confidence": 0.84, + "summary": "Both PRs patch `src/transformers/cli/serve.py` to eliminate a `TypeAdapter`-related runtime NameError in the CLI server path; they differ in implementation style, but they target the same underlying failure mode.", + "confidence": 0.65, "canonical_issue_reason": null, - "canonical_pr_reason": "PR 43842 is the tighter, more direct fix for the reported TypeAdapter NameError: it explicitly guards the pydantic import with TYPE_CHECKING and stringizes the validator annotation to avoid runtime import failure.", + "canonical_pr_reason": "PR 43842 is the more direct and narrowly scoped fix: it keeps `TypeAdapter` out of runtime imports with `TYPE_CHECKING` and quotes the annotation, matching the reported NameError more precisely.", "best_issue_reason": null, - "best_pr_reason": "PR 43842 is the best single PR candidate because it directly targets the concrete failure mode described in its title and diff, with minimal unrelated typing changes.", + "best_pr_reason": "PR 43842 is the best PR candidate here because it directly addresses the `TypeAdapter` NameError without broader annotation churn.", "soft_edge_verdicts": [ { "left": "pull_request:43656", "right": "pull_request:43842", - "accept": false, - "reason": "Same file and similar symptom, but 43656 is a broader typing/future-annotations cleanup while 43842 specifically fixes missing-pydantic TypeAdapter handling; they do not look like the same underlying bug and should not be merged as duplicates." + "accept": true, + "reason": "Both PRs fix the same `TypeAdapter` NameError in the CLI server code path in `serve.py`; the differences are implementation detail, so they could plausibly be merged into one change." } ] }, - "evaluator_result": null, - "evaluator_used": false, + "evaluator_result": { + "accept": true, + "feedback": "The summary is grounded in the packet: both PRs touch `src/transformers/cli/serve.py` and are plausibly aimed at the same `TypeAdapter`-related NameError. The mergeability verdict is suitably conservative given the shared file and similar bugfix scope." + }, + "evaluator_used": true, "retried": false, - "accepted_nontrivial_soft_edge": false, + "accepted_nontrivial_soft_edge": true, "error_kind": null, "error_message": null }, @@ -25755,8 +25841,8 @@ "pull_request:44098|pull_request:44101", "pull_request:44059|pull_request:44722", "pull_request:44018|pull_request:44071", - "pull_request:44017|pull_request:44073", "pull_request:44017|pull_request:44161", + "pull_request:44017|pull_request:44073", "pull_request:44017|pull_request:44722", "pull_request:44068|pull_request:44071", "pull_request:43997|pull_request:44073", @@ -25798,20 +25884,20 @@ "pull_request:44044|pull_request:44154", "pull_request:44024|pull_request:44076" ], - "prepared_review_unit_hash": "cd0d46837dec2bcd05d2a1a0ae4a268d2d4fd5d72dceb55d3602d7dc78e5061f", + "prepared_review_unit_hash": "c3c5b83c8d3742ad5fdc4ddd5822f946305511daf3f7ce554bb731f8edb7a614", "component_budget": { "node_count": 28, "item_count": 28, "soft_pair_count": 77, - "serialized_chars": 84506, - "estimated_input_tokens": 21127, - "estimated_eval_tokens": 42510 + "serialized_chars": 84502, + "estimated_input_tokens": 21126, + "estimated_eval_tokens": 42508 }, "budget": { "node_count": 28, "item_count": 28, "soft_pair_count": 72, - "serialized_chars": 76936, + "serialized_chars": 76933, "estimated_input_tokens": 19234, "estimated_eval_tokens": 38724 }, @@ -25820,54 +25906,54 @@ "aggressively_trimmed": false, "split": true, "analyst_result": { - "summary": "Mostly a set of separate model-specific output-tracing refactors. The only strong duplicate-like overlaps are within GPT-Neo and within the GPT-J/CodeGen subgroup; the rest are different model code paths and should stay separate. I also corrected the 44066 vs 44085 verdict to stay grounded in the GPT-J file overlap rather than any unsupported model-name substitution.", - "confidence": 0.86, + "summary": "This cluster is a broad wave of output-tracing refactors across many unrelated model implementations. Most soft edges should be rejected as broad pattern matches rather than duplicate fixes. The only tight duplicate-like subgroup in the packet is the GPT-J/CodeGen family around 44066/44085/44722; the GPT-Neo pair 44018/44068 is too weak to call the same concrete change under a conservative standard.", + "confidence": 0.79, "canonical_issue_reason": null, - "canonical_pr_reason": "PR 44722 is the clearest canonical representative for the GPT-J/CodeGen subgroup because it is the broadest explicit standardized-output-tracing refactor and overlaps the same concrete files/code path.", + "canonical_pr_reason": "PR 44722 is the best canonical representative of the only tight duplicate subgroup: it is issue-linked, touches the same GPT-J/CodeGen code paths as the nearby GPT-J variants, and looks like the most consolidated version of that refactor.", "best_issue_reason": null, - "best_pr_reason": "PR 44722 is the best overall PR representative for the cluster because it is the broadest late-stage GPT-J/CodeGen output-tracing change and subsumes the shared code-path overlap seen in 44066 and 44085.", + "best_pr_reason": "PR 44722 is the strongest anchor for the cluster because it is explicit about the standardized-decorator refactor and most clearly subsumes the closely related GPT-J/CodeGen variants.", "soft_edge_verdicts": [ { "left": "pull_request:43996", "right": "pull_request:44085", "accept": false, - "reason": "CVT/FNet and GPT-J are different model implementations." + "reason": "Different model families and code paths; same output-tracing theme is too broad." }, { "left": "pull_request:43996", "right": "pull_request:44044", "accept": false, - "reason": "CVT/FNet and DeBERTaV2 are unrelated model-specific changes." + "reason": "Different model families; not the same concrete fix." }, { "left": "pull_request:44066", "right": "pull_request:44085", - "accept": true, - "reason": "Both touch the same GPT-J output-tracing code path in modeling_gptj.py, so they look like the same underlying refactor." + "accept": false, + "reason": "Same general GPT-J area, but the packet only shows weak similarity here and explicitly marks this pair non-deterministic; not enough to call it one duplicate PR." }, { "left": "pull_request:44007", "right": "pull_request:44072", "accept": false, - "reason": "ResNet and EfficientNet are separate model implementations." + "reason": "RegNet/ResNet vs EfficientNet are separate model implementations; shared refactor pattern is not enough." }, { "left": "pull_request:44072", "right": "pull_request:44722", "accept": false, - "reason": "EfficientNet and GPT-J/CodeGen are different concrete code paths." + "reason": "EfficientNet and GPT-J are unrelated code paths." }, { "left": "pull_request:44066", "right": "pull_request:44072", "accept": false, - "reason": "GPT-J/CodeGen and EfficientNet are unrelated model-specific refactors." + "reason": "GPT-J/CodeGen vs EfficientNet are unrelated model changes." }, { "left": "pull_request:44013", "right": "pull_request:44044", "accept": false, - "reason": "MobileNetV2 and DeBERTaV2 are unrelated model refactors." + "reason": "MobileNetV2 vs DeBERTa are different subsystems and different fixes." }, { "left": "pull_request:44066", @@ -25879,311 +25965,227 @@ "left": "pull_request:44066", "right": "pull_request:44071", "accept": false, - "reason": "GPT-J/CodeGen and MPT are separate code paths." + "reason": "GPT-J/CodeGen vs MPT are separate model code paths." }, { "left": "pull_request:44018", "right": "pull_request:44068", - "accept": true, - "reason": "Both are GPT-Neo output-tracing refactors on the same model/code path." + "accept": false, + "reason": "Same broad GPT-Neo theme, but the packet does not show enough evidence that these are the same concrete refactor." }, { "left": "pull_request:44066", "right": "pull_request:44068", "accept": false, - "reason": "GPT-J/CodeGen and GPT-Neo are different model implementations." + "reason": "GPT-J/CodeGen vs GPT-Neo are distinct model refactors." }, { "left": "pull_request:44018", - "right": "pull_request:44066", + "right": "pull_request:44019", "accept": false, - "reason": "GPT-Neo and GPT-J/CodeGen are separate model-specific refactors." + "reason": "GPT-Neo vs ResNet are unrelated model code paths." }, { "left": "pull_request:44019", "right": "pull_request:44722", "accept": false, - "reason": "ResNet and GPT-J/CodeGen are unrelated concrete code paths." + "reason": "ResNet vs GPT-J are unrelated model changes." }, { "left": "pull_request:44086", "right": "pull_request:44722", "accept": false, - "reason": "MGP-STR and GPT-J/CodeGen are different refactors." + "reason": "MGP-STR and GPT-J are different model families and fixes." }, { "left": "pull_request:44019", "right": "pull_request:44085", "accept": false, - "reason": "ResNet and GPT-J are unrelated implementations." - }, - { - "left": "pull_request:44018", - "right": "pull_request:44019", - "accept": false, - "reason": "GPT-Neo and ResNet are different model families." - }, - { - "left": "pull_request:44019", - "right": "pull_request:44071", - "accept": false, - "reason": "ResNet and MPT are unrelated code paths." + "reason": "ResNet vs GPT-J are unrelated." }, { "left": "pull_request:44068", "right": "pull_request:44116", "accept": false, - "reason": "GPT-Neo and Flaubert are different model code paths." + "reason": "GPT-Neo vs Flaubert are different model-specific refactors." }, { "left": "pull_request:44068", "right": "pull_request:44722", "accept": false, - "reason": "GPT-Neo and GPT-J/CodeGen are separate model-specific changes." + "reason": "GPT-Neo and GPT-J/CodeGen are different concrete changes." }, { "left": "pull_request:44018", "right": "pull_request:44116", "accept": false, - "reason": "GPT-Neo and Flaubert are not the same underlying fix." + "reason": "GPT-Neo and Flaubert are unrelated model changes." }, { "left": "pull_request:44071", "right": "pull_request:44722", "accept": false, - "reason": "MPT and GPT-J/CodeGen are separate code paths." + "reason": "MPT vs GPT-J are unrelated model code paths." }, { "left": "pull_request:44018", "right": "pull_request:44085", "accept": false, - "reason": "GPT-Neo and GPT-J are different models." + "reason": "GPT-Neo vs GPT-J are related only at the broad pattern level, not the same PR." }, { "left": "pull_request:44018", "right": "pull_request:44086", "accept": false, - "reason": "GPT-Neo and MGP-STR do not share the same concrete bug." + "reason": "GPT-Neo vs MGP-STR are unrelated changes." }, { "left": "pull_request:44068", "right": "pull_request:44086", "accept": false, - "reason": "GPT-Neo and MGP-STR are unrelated model refactors." + "reason": "GPT-Neo and MGP-STR are different models." + }, + { + "left": "pull_request:44071", + "right": "pull_request:44086", + "accept": false, + "reason": "MPT vs MGP-STR are separate model implementations." }, { "left": "pull_request:44059", "right": "pull_request:44086", "accept": false, - "reason": "GPT2 and MGP-STR are different model implementations." + "reason": "GPT-2 vs MGP-STR are unrelated refactors." }, { "left": "pull_request:44098", "right": "pull_request:44101", "accept": false, - "reason": "ViLT and Flaubert/XLM are different model code paths." + "reason": "ViLT and Flaubert are different model families." }, { - "left": "pull_request:44017", - "right": "pull_request:44073", + "left": "pull_request:44059", + "right": "pull_request:44722", "accept": false, - "reason": "SegFormer and VisualBERT are different code paths." + "reason": "GPT-2 and GPT-J are different code paths; same theme only." }, { "left": "pull_request:44017", "right": "pull_request:44161", "accept": false, - "reason": "SegFormer and LongT5 are unrelated model-specific changes." + "reason": "SegFormer vs LongT5 are unrelated model changes." }, { "left": "pull_request:44017", - "right": "pull_request:44722", - "accept": false, - "reason": "SegFormer and GPT-J/CodeGen are different implementations." - }, - { - "left": "pull_request:44068", - "right": "pull_request:44071", - "accept": false, - "reason": "GPT-Neo and MPT do not fix the same underlying issue." - }, - { - "left": "pull_request:43997", "right": "pull_request:44073", "accept": false, - "reason": "RegNet and VisualBERT are distinct model code paths." + "reason": "SegFormer vs VisualBert are different model-specific refactors." }, { "left": "pull_request:44017", - "right": "pull_request:44071", + "right": "pull_request:44722", "accept": false, - "reason": "SegFormer and MPT are unrelated refactors." + "reason": "SegFormer and GPT-J do not share the same concrete change." }, { - "left": "pull_request:44073", - "right": "pull_request:44101", + "left": "pull_request:43997", + "right": "pull_request:44073", "accept": false, - "reason": "VisualBERT and Flaubert/XLM are different model implementations." + "reason": "RegNet vs VisualBert are unrelated model implementations." }, { "left": "pull_request:44056", "right": "pull_request:44101", "accept": false, - "reason": "MPNet and Flaubert are unrelated concrete changes." + "reason": "MPNet vs Flaubert are unrelated model-specific refactors." }, { - "left": "pull_request:43997", - "right": "pull_request:44018", - "accept": false, - "reason": "RegNet and GPT-Neo are separate model-specific refactors." - }, - { - "left": "pull_request:43997", - "right": "pull_request:44161", - "accept": false, - "reason": "RegNet and LongT5 are not the same bug fix." - }, - { - "left": "pull_request:44074", - "right": "pull_request:44161", - "accept": false, - "reason": "TextNet and LongT5 are different code paths." - }, - { - "left": "pull_request:44018", + "left": "pull_request:44056", "right": "pull_request:44161", "accept": false, - "reason": "GPT-Neo and LongT5 are unrelated model changes." + "reason": "MPNet vs LongT5 are separate model-specific changes." }, { - "left": "pull_request:44010", - "right": "pull_request:44071", - "accept": false, - "reason": "SqueezeBERT and MPT are separate model implementations." - }, - { - "left": "pull_request:44018", - "right": "pull_request:44129", - "accept": false, - "reason": "GPT-Neo and SpeechT5 are unrelated code paths." - }, - { - "left": "pull_request:44010", - "right": "pull_request:44073", + "left": "pull_request:43997", + "right": "pull_request:44056", "accept": false, - "reason": "SqueezeBERT and VisualBERT are different model refactors." + "reason": "RegNet and MPNet are different model families." }, { - "left": "pull_request:43997", - "right": "pull_request:44074", + "left": "pull_request:44071", + "right": "pull_request:44076", "accept": false, - "reason": "RegNet and TextNet are different model-specific changes." + "reason": "MPT and ImageGPT are unrelated model changes." }, { "left": "pull_request:44074", "right": "pull_request:44101", "accept": false, - "reason": "TextNet and Flaubert are unrelated implementations." + "reason": "TextNet vs Flaubert are different model implementations." }, { "left": "pull_request:44056", - "right": "pull_request:44161", - "accept": false, - "reason": "MPNet and LongT5 do not share the same concrete bug." - }, - { - "left": "pull_request:43997", - "right": "pull_request:44056", - "accept": false, - "reason": "RegNet and MPNet are separate model refactors." - }, - { - "left": "pull_request:44071", - "right": "pull_request:44129", + "right": "pull_request:44073", "accept": false, - "reason": "MPT and SpeechT5 are different model code paths." + "reason": "MPNet and VisualBert are unrelated model-specific refactors." }, { "left": "pull_request:43996", "right": "pull_request:44161", "accept": false, - "reason": "CVT/FNet and LongT5 are unrelated models." + "reason": "FNet/CVT vs LongT5 are unrelated model families." }, { "left": "pull_request:43996", "right": "pull_request:44017", "accept": false, - "reason": "CVT/FNet and SegFormer are different concrete changes." + "reason": "FNet/CVT vs SegFormer are different model implementations." }, { "left": "pull_request:44010", "right": "pull_request:44044", "accept": false, - "reason": "SqueezeBERT and DeBERTaV2 are unrelated model refactors." + "reason": "SqueezeBert and DeBERTa are unrelated changes." }, { "left": "pull_request:44044", "right": "pull_request:44074", "accept": false, - "reason": "DeBERTaV2 and TextNet are different code paths." - }, - { - "left": "pull_request:44066", - "right": "pull_request:44722", - "accept": true, - "reason": "Both are GPT-J/CodeGen standardized output-tracing refactors, and 44722 is the broader later version of the same change." - }, - { - "left": "pull_request:44085", - "right": "pull_request:44722", - "accept": true, - "reason": "Both point to the same GPT-J/CodeGen output-tracing refactor on the same files/code path." - }, - { - "left": "pull_request:44085", - "right": "pull_request:44129", - "accept": false, - "reason": "GPT-J and SpeechT5 are unrelated model-specific changes." + "reason": "DeBERTa vs TextNet are different model-specific output changes." }, { - "left": "pull_request:43996", - "right": "pull_request:44007", + "left": "pull_request:44024", + "right": "pull_request:44076", "accept": false, - "reason": "CVT/FNet and ResNet are different model families." + "reason": "FocalNet vs ImageGPT are unrelated models; same standardized-output pattern is too broad." }, { - "left": "pull_request:43996", + "left": "pull_request:44085", "right": "pull_request:44129", "accept": false, - "reason": "CVT/FNet and SpeechT5 are unrelated code paths." + "reason": "GPT-J vs SpeechT5 are different concrete code paths." }, { "left": "pull_request:44044", "right": "pull_request:44129", "accept": false, - "reason": "DeBERTaV2 and SpeechT5 are separate implementations." + "reason": "DeBERTa and SpeechT5 are unrelated." }, { "left": "pull_request:44044", "right": "pull_request:44154", "accept": false, - "reason": "DeBERTaV2 and Vits are different model refactors." - }, - { - "left": "pull_request:44024", - "right": "pull_request:44076", - "accept": false, - "reason": "FocalNet and ImageGPT are unrelated concrete changes." + "reason": "DeBERTa vs ViLT are separate model changes." } ] }, "evaluator_result": { "accept": false, - "feedback": "The overall clustering read is mostly grounded, but one soft-edge verdict is not precise: the pair `pull_request:44066` vs `pull_request:44085` is described as involving EfficientNet, which is not one of the PRs and is unsupported by the packet. That kind of model-name substitution is a precision error, so I can\u2019t accept this as fully conservative." + "feedback": "Reject: the cluster summary is mostly conservative, but two soft-edge acceptances overstate the evidence. In particular, 44066\u219444085 and 44018\u219444068 are treated as true duplicates / same concrete refactors, while the packet only supports broad same-model similarity and, for 44066\u219444085, explicitly lists the pair as non-deterministic. That is too strong for this review standard. The rest of the reasoning is broadly grounded, but these acceptances are not conservative enough." }, "evaluator_used": true, "retried": true, - "accepted_nontrivial_soft_edge": true, + "accepted_nontrivial_soft_edge": false, "error_kind": null, "error_message": null }, @@ -26214,20 +26216,20 @@ "pull_request:43997|pull_request:44010", "pull_request:44010|pull_request:44129" ], - "prepared_review_unit_hash": "cdcb3d63fe605f99c53d4fd8d5ca968e4863ae0e139a060d3f1686a2b0ec95bd", + "prepared_review_unit_hash": "870236ba72b7387db711dfc03cd048ffb9c2d39b3596c7fb8009c3835dfc3dd7", "component_budget": { "node_count": 28, "item_count": 28, "soft_pair_count": 77, - "serialized_chars": 84506, - "estimated_input_tokens": 21127, - "estimated_eval_tokens": 42510 + "serialized_chars": 84502, + "estimated_input_tokens": 21126, + "estimated_eval_tokens": 42508 }, "budget": { "node_count": 6, "item_count": 6, "soft_pair_count": 5, - "serialized_chars": 13070, + "serialized_chars": 13069, "estimated_input_tokens": 3268, "estimated_eval_tokens": 6792 }, @@ -26236,42 +26238,42 @@ "aggressively_trimmed": false, "split": true, "analyst_result": { - "summary": "All items are model-specific refactors to standardized output tracing/capture across different architectures. They share a common pattern, but they do not look like the same underlying bug or change, so none of the PR pairs should be merged as duplicates.", + "summary": "These are all model-specific refactors to standardized output capture/tracing, but they touch different architectures and code paths, so they are not duplicates of one another.", "confidence": 0.93, "canonical_issue_reason": null, - "canonical_pr_reason": "No single PR is a true duplicate canonical for the cluster because each targets a different model and code path. If one representative is needed, PR #44129 is the strongest exemplar due to its explicit issue link and concrete SpeechT5-specific output-tracing change.", + "canonical_pr_reason": "No single PR is a true duplicate anchor for the set; 44129 is the strongest representative because it has an explicit issue target and a concrete, model-specific output-tracing refactor.", "best_issue_reason": null, - "best_pr_reason": "PR #44129 is the best representative of the set: it is concrete, has an explicit tracking issue, and makes a focused model-specific refactor rather than a broad pattern-only change.", + "best_pr_reason": "44129 is the best representative PR: it targets a specific issue, changes a single model implementation, and shows a substantive output-tracing cleanup rather than just a title-level similarity.", "soft_edge_verdicts": [ { "left": "pull_request:44073", "right": "pull_request:44074", "accept": false, - "reason": "Both are output-tracing refactors, but they modify different model implementations (VisualBert vs TextNet) and are not the same code-path fix." + "reason": "Same refactor theme, but VisualBert and TextNet are different models with different forward paths; not the same concrete change." }, { "left": "pull_request:44056", "right": "pull_request:44073", "accept": false, - "reason": "Same refactor theme, but different models (MPNet vs VisualBert) and different internal tracing setup; not mergeable as one PR for the same bug/change." + "reason": "Both are output-tracing migrations, but MPNet and VisualBert have distinct implementations and no shared code-path bug." }, { "left": "pull_request:44056", "right": "pull_request:44074", "accept": false, - "reason": "These are separate model-specific migrations in different files and code paths, so they are not duplicates." + "reason": "Different model families and different model files; this is parallel refactoring, not one mergeable fix." }, { "left": "pull_request:43997", "right": "pull_request:44010", "accept": false, - "reason": "RegNet and SqueezeBert are unrelated implementations; the PRs share a pattern but not the same underlying change." + "reason": "RegNet and SqueezeBert are unrelated code paths; same high-level decorator migration is not enough to treat them as duplicates." }, { "left": "pull_request:44010", "right": "pull_request:44129", "accept": false, - "reason": "SqueezeBert vs SpeechT5 are distinct model code paths, and SpeechT5 also includes an explicit issue-linked refactor; not the same concrete fix." + "reason": "Both modernize output handling, but SqueezeBert and SpeechT5 fix different model-specific implementations and could not plausibly be one PR." } ] }, @@ -26350,268 +26352,268 @@ "pull_request:44028|pull_request:44030", "pull_request:44029|pull_request:44030" ], - "prepared_review_unit_hash": "473bd213b52f864aafa3d6d1b1a8807755187359e2feec1751fd806e9d16e419", + "prepared_review_unit_hash": "e73142a3e2ca188db96f5f72a4c2d7e2cca36b8af82ec275f54438f9ea7e3a17", "component_budget": { "node_count": 13, "item_count": 13, "soft_pair_count": 39, - "serialized_chars": 39779, - "estimated_input_tokens": 9945, - "estimated_eval_tokens": 20146 + "serialized_chars": 39790, + "estimated_input_tokens": 9948, + "estimated_eval_tokens": 20152 }, "budget": { "node_count": 13, "item_count": 13, "soft_pair_count": 39, - "serialized_chars": 39779, - "estimated_input_tokens": 9945, - "estimated_eval_tokens": 20146 + "serialized_chars": 39790, + "estimated_input_tokens": 9948, + "estimated_eval_tokens": 20152 }, "overflow_policy": "truncate_then_skip", "trimmed": false, "aggressively_trimmed": false, "split": false, "analyst_result": { - "summary": "These are separate model-specific output-tracing refactors across different architectures (e.g. timm_backbone, MobilenetV1, DepthAnything, UperNet, RWKV, DPR, CodeGen). They share a naming pattern and a similar refactor theme, but not the same concrete bug or mergeable code change, so the soft edges should not be treated as duplicates.", - "confidence": 0.92, + "summary": "These are all model-specific output-tracing refactors, not duplicate fixes. They share a broad implementation theme, but each PR targets different model code paths and should be kept separate.", + "confidence": 0.94, "canonical_issue_reason": null, - "canonical_pr_reason": "No true duplicate PR cluster here; PR 44025 is the best representative anchor because it is a clear, self-contained output-tracing refactor spanning two closely related DepthAnything files.", + "canonical_pr_reason": "If a representative PR is needed, 44003 is the broadest of the set because it spans two related model implementations and adds shared output-capture plumbing; however, it is not a duplicate of the others.", "best_issue_reason": null, - "best_pr_reason": "PR 44025 is the most suitable canonical example of the recurring refactor pattern: it is focused, self-contained, and touches a coherent pair of related model files.", + "best_pr_reason": "44003 is the strongest representative of the cluster because it has the widest model coverage and the clearest core refactor, but the set still consists of separate PRs rather than duplicates.", "soft_edge_verdicts": [ { "left": "pull_request:43998", "right": "pull_request:44002", "accept": false, - "reason": "Different model code paths (timm_backbone vs upernet); same refactor theme only, not one mergeable fix." + "reason": "Different models and code paths; same output-tracing pattern only." }, { "left": "pull_request:43998", "right": "pull_request:44000", "accept": false, - "reason": "timm_backbone and vision_text_dual_encoder are unrelated implementations; the shared output-tracing wording is too generic." + "reason": "Different model families; no shared concrete bug or change." }, { "left": "pull_request:43998", "right": "pull_request:44001", "accept": false, - "reason": "Different architectures and files (timm_backbone vs univnet); not the same underlying change." + "reason": "Separate model-specific refactors, not one mergeable fix." }, { "left": "pull_request:43998", "right": "pull_request:44003", "accept": false, - "reason": "timm_backbone and mamba/falcon_mamba are separate code paths; only the refactor pattern matches." + "reason": "Both are tracing refactors, but they change different model implementations and underlying paths." }, { "left": "pull_request:43998", "right": "pull_request:44029", "accept": false, - "reason": "Unrelated models (timm_backbone vs rwkv); cannot plausibly be merged as one PR." + "reason": "Unrelated architectures; same refactor theme is too broad to duplicate." }, { "left": "pull_request:43999", "right": "pull_request:44001", "accept": false, - "reason": "MobilenetV1 and univnet are different model families; same generic refactor theme only." + "reason": "Different models; only the tracing helper usage overlaps." }, { "left": "pull_request:43999", "right": "pull_request:44002", "accept": false, - "reason": "MobilenetV1 and UperNet touch different code paths and outputs; not the same bug/change." + "reason": "Distinct model code paths with similar output-tracing cleanup." }, { "left": "pull_request:43999", "right": "pull_request:44025", "accept": false, - "reason": "MobilenetV1 vs DepthAnything are unrelated model implementations; the overlap is just the output-tracing refactor style." + "reason": "Different model families and files; not the same underlying fix." }, { "left": "pull_request:43999", "right": "pull_request:44026", "accept": false, - "reason": "Different model families (MobilenetV1 vs VisionEncoderDecoder); not a single concrete fix." + "reason": "Same high-level refactor, but separate encoder-decoder model targets." }, { "left": "pull_request:43999", "right": "pull_request:44027", "accept": false, - "reason": "MobilenetV1 and SpeechEncoderDecoder are separate code paths; shared theme is too broad." + "reason": "Different model implementations; no duplicate code-path problem." }, { "left": "pull_request:44000", "right": "pull_request:44001", "accept": false, - "reason": "VisionTextDualEncoder and UnivNet are unrelated models; same refactor label does not imply duplicate change." + "reason": "Separate model-specific output-tracing refactors." }, { "left": "pull_request:44000", "right": "pull_request:44002", "accept": false, - "reason": "Different subsystems (vision-text dual encoder vs UperNet); not one fixable bug." + "reason": "Different architectures and forward paths; not one fix." }, { "left": "pull_request:44000", "right": "pull_request:44003", "accept": false, - "reason": "VisionTextDualEncoder and Mamba/FalconMamba are distinct code paths; not mergeable into one PR." + "reason": "Shared refactor style, but the affected models and logic are different." }, { "left": "pull_request:44000", "right": "pull_request:44025", "accept": false, - "reason": "VisionTextDualEncoder vs DepthAnything are different model implementations; only the refactor pattern aligns." + "reason": "Different model stacks; same abstraction pattern only." }, { "left": "pull_request:44001", "right": "pull_request:44002", "accept": false, - "reason": "UnivNet and UperNet are unrelated architectures; same generic output-tracing theme only." + "reason": "These touch different models and do not address the same bug." }, { "left": "pull_request:44001", "right": "pull_request:44003", "accept": false, - "reason": "UnivNet and Mamba/FalconMamba target different code paths and cannot be one PR." + "reason": "Not the same code path; only the refactor mechanism matches." }, { "left": "pull_request:44001", "right": "pull_request:44025", "accept": false, - "reason": "UnivNet vs DepthAnything is not the same underlying bug or change." + "reason": "Distinct model implementations; no plausible single PR merge." }, { "left": "pull_request:44002", "right": "pull_request:44029", "accept": false, - "reason": "UperNet and RWKV are different model families; shared refactor wording is insufficient." + "reason": "Different model families; same output-tracing theme is insufficient." }, { "left": "pull_request:44002", "right": "pull_request:44003", "accept": false, - "reason": "UperNet and Mamba/FalconMamba are separate code paths; not a duplicate fix." + "reason": "Separate model-specific changes, not a duplicate fix." }, { "left": "pull_request:44003", "right": "pull_request:44004", "accept": false, - "reason": "Mamba/FalconMamba and CodeGen are unrelated implementations; same refactor theme only." + "reason": "Different models and forward implementations; not the same change." }, { "left": "pull_request:43998", "right": "pull_request:44004", "accept": false, - "reason": "timm_backbone and CodeGen do not share a concrete bug or code path." + "reason": "Model-specific refactors with no shared concrete bug." }, { "left": "pull_request:44001", "right": "pull_request:44004", "accept": false, - "reason": "UnivNet and CodeGen are unrelated model code paths; not mergeable as one PR." + "reason": "Different code paths; similarity is only at the refactor pattern level." }, { "left": "pull_request:44002", "right": "pull_request:44004", "accept": false, - "reason": "UperNet and CodeGen are distinct architectures; only the tracing refactor motif overlaps." + "reason": "Not the same underlying change; different model logic." }, { "left": "pull_request:44000", "right": "pull_request:44004", "accept": false, - "reason": "VisionTextDualEncoder and CodeGen are unrelated changes despite similar refactor wording." + "reason": "Same broad subsystem, but distinct model-specific refactors." }, { "left": "pull_request:44025", "right": "pull_request:44026", "accept": false, - "reason": "DepthAnything and VisionEncoderDecoder are different model families; not the same code-path fix." + "reason": "Different model families; same output-tracing modernization only." }, { "left": "pull_request:44025", "right": "pull_request:44027", "accept": false, - "reason": "DepthAnything and SpeechEncoderDecoder are unrelated implementations; same theme only." + "reason": "Separate encoder-decoder model refactors, not one bug fix." }, { "left": "pull_request:44025", "right": "pull_request:44028", "accept": false, - "reason": "DepthAnything and SuperPoint are different model code paths; not a duplicate change." + "reason": "Different model and code path; no duplicate change." }, { "left": "pull_request:44025", "right": "pull_request:44029", "accept": false, - "reason": "DepthAnything and RWKV are unrelated; the similarity is superficial." + "reason": "Different architectures; too broad to be a duplicate." }, { "left": "pull_request:44025", "right": "pull_request:44030", "accept": false, - "reason": "DepthAnything and DPR do not fix the same underlying problem." + "reason": "Distinct model targets and fixes." }, { "left": "pull_request:44026", "right": "pull_request:44029", "accept": false, - "reason": "VisionEncoderDecoder and RWKV are separate code paths; not one mergeable fix." + "reason": "Different model code paths; only the refactor theme overlaps." }, { "left": "pull_request:44026", "right": "pull_request:44027", "accept": false, - "reason": "VisionEncoderDecoder and SpeechEncoderDecoder are different model wrappers; shared naming is too broad." + "reason": "Separate model-specific changes, not a single mergeable PR." }, { "left": "pull_request:44026", "right": "pull_request:44028", "accept": false, - "reason": "VisionEncoderDecoder and SuperPoint are unrelated implementations." + "reason": "Different models and outputs; not the same underlying bug." }, { "left": "pull_request:44026", "right": "pull_request:44030", "accept": false, - "reason": "VisionEncoderDecoder and DPR target different model families and code paths." + "reason": "Different code paths with a shared output-tracing pattern." }, { "left": "pull_request:44027", "right": "pull_request:44028", "accept": false, - "reason": "SpeechEncoderDecoder and SuperPoint are unrelated; no shared concrete bug." + "reason": "Distinct model implementations; similarity is too generic." }, { "left": "pull_request:44027", "right": "pull_request:44029", "accept": false, - "reason": "SpeechEncoderDecoder and RWKV are different architectures, so this is not a duplicate PR." + "reason": "Different families and files; not duplicate changes." }, { "left": "pull_request:44027", "right": "pull_request:44030", "accept": false, - "reason": "SpeechEncoderDecoder and DPR do not share the same underlying fix." + "reason": "Same refactor motif, but unrelated model code paths." }, { "left": "pull_request:44028", "right": "pull_request:44029", "accept": false, - "reason": "SuperPoint and RWKV are unrelated model paths; shared refactor label only." + "reason": "Different models and logic; no shared concrete fix." }, { "left": "pull_request:44028", "right": "pull_request:44030", "accept": false, - "reason": "SuperPoint and DPR are different subsystems and cannot be merged into one PR." + "reason": "Separate model-specific refactors, not one PR." }, { "left": "pull_request:44029", "right": "pull_request:44030", "accept": false, - "reason": "RWKV and DPR are unrelated model implementations; not the same bug or change." + "reason": "Different architectures and forward paths; not duplicates." } ] }, @@ -26641,7 +26643,7 @@ "soft_pairs": [ "pull_request:44662|pull_request:45401" ], - "prepared_review_unit_hash": "f0e797a4a0f1034d2a251541e668f31556102197c5abac44c91e94cdc81dbddf", + "prepared_review_unit_hash": "9ac8370b3f72014b9925d555cbe3bde2ad9e5729646ff4a6c8b2928bf44f4b42", "component_budget": { "node_count": 2, "item_count": 2, @@ -26663,18 +26665,18 @@ "aggressively_trimmed": false, "split": false, "analyst_result": { - "summary": "These are two unrelated model-addition PRs: PenguinVL and Voxtral TTS. They share only common Transformers integration scaffolding (auto classes, docs, tests), not the same underlying bug or change.", + "summary": "These are two unrelated model-implementation PRs that happen to touch shared auto-registry and docs files; they are not the same underlying change.", "confidence": 0.98, "canonical_issue_reason": null, - "canonical_pr_reason": "PR #45401 is the better canonical representative for the cluster if one must be chosen, because it is explicitly tied to issue #45027 and is a narrower, model-specific implementation request.", + "canonical_pr_reason": "PR #45401 is the more canonical representative because it has an explicit linked issue, a focused scope, and a model-specific implementation for Voxtral TTS.", "best_issue_reason": null, - "best_pr_reason": "PR #45401 is the strongest single PR candidate overall due to its explicit issue target and focused scope; #44662 is also a standalone model implementation, but it is a different feature and not a duplicate.", + "best_pr_reason": "PR #45401 is the better standalone PR to keep in a cluster view because it is narrowly scoped and tied to a specific tracking issue, whereas PR #44662 is a separate PenguinVL implementation.", "soft_edge_verdicts": [ { "left": "pull_request:44662", "right": "pull_request:45401", "accept": false, - "reason": "Different models and code paths: PenguinVL vs Voxtral TTS. Overlap is only in shared integration files and documentation, which is not enough to treat them as the same change." + "reason": "Different model additions with different code paths, filenames, and goals (PenguinVL vs Voxtral TTS); shared auto/doc registry edits are not enough to treat them as the same change." } ] }, @@ -26704,12 +26706,12 @@ "soft_pairs": [ "pull_request:44979|pull_request:45363" ], - "prepared_review_unit_hash": "8093be75ecac1816cf3ebf0052a2555c49cff25d8f87c22282eb2c54666d1765", + "prepared_review_unit_hash": "5cef09eaf93cfb83956df7a7d0d498c81614fe51fd597ae3a4975de0f0550499", "component_budget": { "node_count": 2, "item_count": 2, "soft_pair_count": 1, - "serialized_chars": 4408, + "serialized_chars": 4407, "estimated_input_tokens": 1102, "estimated_eval_tokens": 2460 }, @@ -26717,7 +26719,7 @@ "node_count": 2, "item_count": 2, "soft_pair_count": 1, - "serialized_chars": 4408, + "serialized_chars": 4407, "estimated_input_tokens": 1102, "estimated_eval_tokens": 2460 }, @@ -26726,18 +26728,18 @@ "aggressively_trimmed": false, "split": false, "analyst_result": { - "summary": "These two pull requests are related in the broad area of module/kernel fusion, but they are not duplicates: #44979 introduces a generic Module Fusion API, while #45363 wires in a specific n-to-1 kernel fusion path through KernelConfig and integration/modeling changes. They address different layers of the feature stack and would not be merged as one PR without collapsing distinct work.", - "confidence": 0.93, + "summary": "Both PRs concern fusion infrastructure in Transformers, but they target different layers: PR #44979 introduces a general Module Fusion API and tests, while PR #45363 implements a specific n-to-1 kernel fusion path via `KernelConfig` plus integration changes. Related theme, not the same underlying change.", + "confidence": 0.86, "canonical_issue_reason": null, - "canonical_pr_reason": "#44979 is the broader, foundational change: it adds the new fusion API and tests, which the later kernel-fusion work can build on. #45363 is a narrower integration of that idea into a specific kernel-config path.", + "canonical_pr_reason": "PR #44979 is the broader foundational API change and is the better cluster anchor for the general fusion concept.", "best_issue_reason": null, - "best_pr_reason": "#44979 is the best representative of the cluster because it defines the core abstraction rather than one concrete application of it. #45363 is more specialized and depends on the broader fusion direction.", + "best_pr_reason": "PR #44979 best represents the cluster because it defines the reusable module-fusion abstraction rather than one concrete kernel-fusion application.", "soft_edge_verdicts": [ { "left": "pull_request:44979", "right": "pull_request:45363", "accept": false, - "reason": "Related feature area, but not the same underlying change. #44979 adds a generic Module Fusion API; #45363 implements a specific n-to-1 kernel fusion integration using KernelConfig and touches different code paths. They are complementary, not duplicates." + "reason": "They share the fusion theme, but #44979 adds a generic module-fusion API, whereas #45363 wires up a specific n-to-1 kernel-fusion implementation through `KernelConfig` and model loading. Different concrete code paths and not plausibly one merged PR." } ] }, @@ -26767,7 +26769,7 @@ "soft_pairs": [ "pull_request:45086|pull_request:45317" ], - "prepared_review_unit_hash": "973bb6dc41d7796f41e8d893fc5ce8645629da7f3077c77d67c5c9ed1a86d44e", + "prepared_review_unit_hash": "c13cf43b3eed59e27175ade45b8a1591a88abf5db2cde09117109fed16ad1449", "component_budget": { "node_count": 2, "item_count": 2, @@ -26789,18 +26791,18 @@ "aggressively_trimmed": false, "split": false, "analyst_result": { - "summary": "Both pull requests address the same Mistral regex patch AttributeError in the same code path; 45317 is the more complete version with tests, while 45086 is a smaller precursor.", - "confidence": 0.97, + "summary": "Both pull requests address the same AttributeError in `_patch_mistral_regex` by switching from `backend_tokenizer.pre_tokenizer` to `tokenizer.pre_tokenizer`; PR 45317 is the more complete version because it also adds tests and clearer framing.", + "confidence": 0.98, "canonical_issue_reason": null, - "canonical_pr_reason": "Choose PR 45317 as canonical: it fixes the same AttributeError with the same code change, but also adds test coverage and has a clearer, final form of the fix.", + "canonical_pr_reason": "PR 45317 is the strongest canonical candidate: it fixes the same code path as 45086, adds test coverage, and is more explicit about the `fix_mistral_regex=True` case.", "best_issue_reason": null, - "best_pr_reason": "PR 45317 is the best representative because it appears to fully resolve the bug and includes regression tests, making it the strongest standalone fix.", + "best_pr_reason": "PR 45317 is the best overall PR because it contains the same fix as 45086 plus regression tests, making it the more complete and merge-ready change.", "soft_edge_verdicts": [ { "left": "pull_request:45086", "right": "pull_request:45317", "accept": true, - "reason": "Same underlying bug and same fix in the same file; 45317 looks like a later, test-backed refinement of 45086, so they are mergeable into one PR." + "reason": "Same underlying bug in `_patch_mistral_regex` on the same file and same line-level fix; the second PR is essentially the same patch with added tests, so they are duplicates rather than separate changes." } ] }, @@ -26830,7 +26832,7 @@ "soft_pairs": [ "pull_request:45115|pull_request:45569" ], - "prepared_review_unit_hash": "0d58aba46d0479c55349c0608971b906de344ccf05e96ef3e09f6817cea70d06", + "prepared_review_unit_hash": "966439515184f94a5283a7d6f37fbd8e7a5f2f1550bbbacf131fa407b1857aa4", "component_budget": { "node_count": 2, "item_count": 2, @@ -26852,18 +26854,18 @@ "aggressively_trimmed": false, "split": false, "analyst_result": { - "summary": "Two related NemotronH pull requests, but they address different changes: one is a conversion-mapping/refactor step, the other is a broader model split into NemotronH Dense/Sparse with docs/tests and auto-mapping updates.", - "confidence": 0.88, + "summary": "Both PRs work on NemotronH model support, but they target different changes: #45115 is a narrower refactor to inherit GraniteMoeHybrid-style internals and update conversion mappings, while #45569 is a broader split into separate NemotronH Dense/Sparse model families plus docs and auto-mapping updates. They overlap in shared core files, but not the same concrete fix/change.", + "confidence": 0.87, "canonical_issue_reason": null, - "canonical_pr_reason": "PR 45569 is the better canonical PR because it is the broader, more complete change set: it introduces the dense/sparse model split, updates docs and auto-mappings, and includes tests, making it the main user-facing implementation.", + "canonical_pr_reason": "PR #45569 is the better cluster representative because it is broader, more complete, and more user-facing: it introduces the dense/sparse model split, updates docs and auto mappings, and appears to subsume the architectural direction of NemotronH support.", "best_issue_reason": null, - "best_pr_reason": "PR 45569 is the best representative artifact for this cluster because it captures the full NemotronH rework rather than just a narrow conversion-mapping refactor.", + "best_pr_reason": "PR #45569 is the strongest standalone candidate since it covers the larger end-to-end NemotronH rework; #45115 is narrower and draft-only, focused mainly on inheritance/conversion plumbing.", "soft_edge_verdicts": [ { "left": "pull_request:45115", "right": "pull_request:45569", "accept": false, - "reason": "These are related NemotronH changes but not the same underlying change. 45115 focuses on checkpoint conversion/weight renaming and inheritance refactoring, while 45569 splits NemotronH into dense/sparse variants with docs and test updates. They are not mergeable as one PR without conflating distinct scopes." + "reason": "Related NemotronH work, but not the same underlying change. #45115 is a refactor/conversion-mapping PR; #45569 adds separate NemotronH Dense/Sparse models, docs, and auto-mapping updates. Shared files alone are not enough to treat them as duplicates." } ] }, @@ -26893,7 +26895,7 @@ "soft_pairs": [ "pull_request:45415|pull_request:45425" ], - "prepared_review_unit_hash": "b49e53338ad530e376e378d22f2061c09f5c73cb6cb2005b3c347ce23de2dadd", + "prepared_review_unit_hash": "b980b2a2789630312891dee629839c19f0f121b265ba038553ea3aeff229f7d4", "component_budget": { "node_count": 2, "item_count": 2, @@ -26915,18 +26917,18 @@ "aggressively_trimmed": false, "split": false, "analyst_result": { - "summary": "These PRs are both typing-related, but they address different scopes: #45415 is a broad type-checking sweep across many Transformers files, while #45425 is a focused typing-compatibility update centered on `modeling_utils` and shared typing helpers. They are not the same underlying change.", - "confidence": 0.9, + "summary": "Both items are typing-focused pull requests, but they address different scopes: one is a broad repository-wide type-checking pass, the other is a narrower set of typing protocol/annotation fixes in core utilities. They are related, but not the same underlying change set, so I would not deduplicate them.", + "confidence": 0.82, "canonical_issue_reason": null, - "canonical_pr_reason": "#45425 is the more focused, self-contained typing change and is easier to treat as the representative PR for this cluster.", + "canonical_pr_reason": "PR 45425 is the better canonical PR for this cluster because it is the more concrete, code-path-specific typing change set in core modules (`modeling_utils`, `_typing.py`, `peft.py`, flash-attention utilities) and has substantial review activity. PR 45415 is a much broader sweep across many unrelated files plus type-check config.", "best_issue_reason": null, - "best_pr_reason": "#45425 best fits as the representative PR because it targets a specific typing gap with a small, coherent set of code changes, whereas #45415 is a broad typing cleanup sweep.", + "best_pr_reason": "PR 45425 is the strongest representative: it introduces specific typing protocols/aliases and targeted fixes that can be reasoned about as one change. PR 45415 is too broad and looks like a separate large-scale typing cleanup rather than the same patch.", "soft_edge_verdicts": [ { "left": "pull_request:45415", "right": "pull_request:45425", "accept": false, - "reason": "Shared typing-related files do not make these duplicates: #45415 is a broad type-checking pass across many modules, while #45425 adds specific typing protocol/alias support for `modeling_utils` and related helpers." + "reason": "Both are typing-related, but 45415 is a wide repo-wide type-checking cleanup while 45425 adds targeted protocols and annotations in core modules; they are not the same concrete fix and would not plausibly be merged as one PR." } ] }, @@ -26956,7 +26958,7 @@ "soft_pairs": [ "pull_request:45548|pull_request:45662" ], - "prepared_review_unit_hash": "2d73984e367fba9031201e7c1c2ceb63da301f8d4ca4579f6cfefe2144c3e202", + "prepared_review_unit_hash": "df6630ff440056c3dd319a8ea7fe10b121f51d54c528c9b18a99588829d0791f", "component_budget": { "node_count": 2, "item_count": 2, @@ -26978,18 +26980,18 @@ "aggressively_trimmed": false, "split": false, "analyst_result": { - "summary": "The two PRs are related to expert parallelism loading, but they fix different bugs in different code paths: one preserves distributed_config and adds has_ep for accelerate/ZeRO-3 loading, while the other fixes expert parameters being overwritten during FSDP2/sharded loading. They are not duplicates.", + "summary": "These are both EP-related PRs, but they fix different underlying bugs: one is about preserving distributed_config/has_ep during loading, the other about expert tensors being overwritten during FSDP2 post-shard handling. Not duplicates.", "confidence": 0.94, "canonical_issue_reason": null, - "canonical_pr_reason": "No canonical PR as a duplicate representative: the PRs address different EP loading failures with different root causes and fixes.", + "canonical_pr_reason": "No single canonical PR for duplication purposes; if one must anchor the cluster, 45662 is the more concrete bug fix because it targets a specific overwrite path in model loading, while 45548 addresses a separate distributed_config serialization/loading issue.", "best_issue_reason": null, - "best_pr_reason": "PR 45662 is the stronger standalone change because it fixes a concrete parameter-overwrite bug in the loading path and touches the relevant sharding hooks directly.", + "best_pr_reason": "45662 is the stronger standalone PR: it identifies a precise failure mode (experts overwritten by rank-0 broadcast) and changes the relevant loading path directly. 45548 is also valid, but it addresses a different code path and bug.", "soft_edge_verdicts": [ { "left": "pull_request:45548", "right": "pull_request:45662", "accept": false, - "reason": "Different underlying bugs: 45548 preserves distributed_config and adds has_ep for accelerate/ZeRO-3 loading, while 45662 prevents expert weights from being overwritten during FSDP2/sharded loading. Shared EP theme is too broad to merge as duplicates." + "reason": "Shared EP terminology and one common file are not enough: 45548 fixes distributed_config serialization/has_ep for ZeRO-3 loading, while 45662 fixes expert parameters being overwritten in FSDP2 post-shard handling. Different bug mechanisms and merge scopes." } ] }, @@ -27019,7 +27021,7 @@ "soft_pairs": [ "pull_request:45639|pull_request:45645" ], - "prepared_review_unit_hash": "3c55040c70b66f655838bbf2e9a9fb4661b200008aaf15766350c19533440977", + "prepared_review_unit_hash": "04c9477aa31673a6543d9a2bf0ec375bff4eba4f0bd1886604a3d222a32f7b30", "component_budget": { "node_count": 2, "item_count": 2, @@ -27041,18 +27043,18 @@ "aggressively_trimmed": false, "split": false, "analyst_result": { - "summary": "Both PRs address the same xdist collision problem around `captured_info`/patched testing debug logs, and PR #45645 appears to subsume #45639 by adding the CI workflow-side log collection fix as well. They look like the same underlying change and could plausibly be merged into one PR.", - "confidence": 0.94, + "summary": "Both PRs target the same xdist-related captured_info logging problem and share the same core testing_utils change; PR 45645 is the broader, more complete variant because it also updates CI log collection to handle multiple per-worker files.", + "confidence": 0.95, "canonical_issue_reason": null, - "canonical_pr_reason": "PR #45645 is the better canonical PR because it includes the core xdist-safe file handling plus the CI workflow change needed to preserve and aggregate the new per-worker logs, making it the more complete fix for the shared issue.", + "canonical_pr_reason": "PR 45645 is the better canonical PR because it contains the same core xdist-safe output-file fix as 45639 and also updates the workflow that consumes those logs, making it the more complete end-to-end change.", "best_issue_reason": null, - "best_pr_reason": "PR #45645 is the best overall representative because it covers the same bug as #45639 and adds the complementary workflow adjustment, so it better captures the full intended change.", + "best_pr_reason": "PR 45645 is the best representative of the cluster: it addresses the same concrete captured_info collision bug and includes the downstream CI handling needed to preserve debug logs.", "soft_edge_verdicts": [ { "left": "pull_request:45639", "right": "pull_request:45645", "accept": true, - "reason": "Same underlying xdist collision fix for `captured_info` artifacts; #45645 includes the same `testing_utils.py` change and extends it with CI log collection, so the PRs are sufficiently overlapping to treat as duplicates." + "reason": "Accept: both PRs fix the same underlying captured_info xdist collision issue in testing_utils; 45645 is a superset that adds workflow support for multiple output files." } ] }, @@ -27148,7 +27150,7 @@ "soft_pairs": [ "pull_request:46126|pull_request:46269" ], - "prepared_review_unit_hash": "2188b0e5a9e626b2d21334771cb3b83124c6ba9c0858fc49304efdb9a9608f7d", + "prepared_review_unit_hash": "77f41d8c43112bb0d3bf62b680beaa88cec91aa146b7c552834bab7174de8b0b", "component_budget": { "node_count": 2, "item_count": 2, @@ -27170,18 +27172,18 @@ "aggressively_trimmed": false, "split": false, "analyst_result": { - "summary": "These two PRs are not duplicates: #46126 is a focused EP/Trainer integration change around tensor-parallel MoE routing, while #46269 is a broad distributed-branch base PR touching CI, FSDP, and many unrelated distributed files.", - "confidence": 0.95, + "summary": "Two PRs overlap in distributed code, but they are not the same change: one is a focused EP/trainer integration, the other is a broad distributed branch base with many unrelated updates.", + "confidence": 0.88, "canonical_issue_reason": null, - "canonical_pr_reason": "#46126 is the better canonical PR because it is the more specific, self-contained change centered on EP routing and Trainer integration; #46269 is a large umbrella branch-base PR with many unrelated distributed/CI edits.", + "canonical_pr_reason": "PR 46126 is the more focused, concrete change set centered on EP + Trainer integration; PR 46269 is a much broader branch-base PR with many additional distributed and CI modifications, so it is not the best canonical representative of this cluster.", "best_issue_reason": null, - "best_pr_reason": "#46126 is the best standalone PR in this cluster: it has a concrete code-path goal and a narrow scope, whereas #46269 is too broad and mixes multiple subsystems.", + "best_pr_reason": "PR 46126 is the best standalone PR here because it has a clear, specific purpose and a tighter code-path scope. PR 46269 is too broad and mixed-purpose to serve as the best representative of the underlying change.", "soft_edge_verdicts": [ { "left": "pull_request:46126", "right": "pull_request:46269", "accept": false, - "reason": "Reject: despite overlapping distributed filenames, the PRs target different work. #46126 implements EP routing/Trainer integration for MoE, while #46269 is a broad distributed branch-base change with FSDP CI and many unrelated updates; they do not look mergeable into one fix." + "reason": "Reject: the shared distributed filenames do not indicate the same underlying fix. 46126 is a targeted EP/trainer integration PR, while 46269 is a broad distributed branch-base PR spanning many unrelated areas, so they would not plausibly be merged as one PR." } ] }, @@ -27249,220 +27251,232 @@ "pull_request:46504|pull_request:46716", "pull_request:46432|pull_request:46716" ], - "prepared_review_unit_hash": "49a84bc60c8cc02dc3d15abaa38a6a206f4a8a3e268bd5dfad4b4a87c0c5f583", + "prepared_review_unit_hash": "d64e746b79370ae12a1d5ab8408da4a6734246e60d004982b8c1e2614acced21", "component_budget": { "node_count": 10, "item_count": 10, "soft_pair_count": 31, - "serialized_chars": 26689, - "estimated_input_tokens": 6673, - "estimated_eval_tokens": 13602 + "serialized_chars": 26678, + "estimated_input_tokens": 6670, + "estimated_eval_tokens": 13596 }, "budget": { "node_count": 10, "item_count": 10, "soft_pair_count": 31, - "serialized_chars": 26689, - "estimated_input_tokens": 6673, - "estimated_eval_tokens": 13602 + "serialized_chars": 26678, + "estimated_input_tokens": 6670, + "estimated_eval_tokens": 13596 }, "overflow_policy": "truncate_then_skip", "trimmed": false, "aggressively_trimmed": false, "split": false, "analyst_result": { - "summary": "This cluster is mostly a set of unrelated workflow-maintenance PRs. The only dense similarity group is the pr-ci-security-gate.yml series, but those PRs make different incremental edits rather than fixing the same concrete bug/change, so none should be merged as duplicates.", - "confidence": 0.96, + "summary": "This cluster is a loose mix of unrelated workflow-maintenance PRs. The only near-cluster is the set of pr-ci-security-gate.yml edits, but they address different security-hardening steps rather than the same concrete change, so there are no true duplicates here.", + "confidence": 0.94, "canonical_issue_reason": null, - "canonical_pr_reason": "PR 46554 is the best canonical representative because it is the most substantive anchor in the largest near-similar group (pr-ci-security-gate.yml) and the later PRs look like separate follow-up edits rather than the same change.", + "canonical_pr_reason": "PR 46554 is the most representative PR in the only semi-cohesive subcluster: it has the broadest and most substantive security-gate workflow change and the strongest overlap with the other pr-ci-security-gate variants.", "best_issue_reason": null, - "best_pr_reason": "PR 46554 is the strongest representative of the cluster\u2019s only meaningful sub-group: it makes the concrete security-gate workflow fix that the later pr-ci-security-gate PRs build on, while the rest are clearly separate workflow chores.", + "best_pr_reason": "PR 46554 is the best single representative for this mixed cluster because it is the most central member of the security-gate subcluster and the clearest example of a concrete workflow hardening change.", "soft_edge_verdicts": [ { "left": "pull_request:46138", "right": "pull_request:46554", "accept": false, - "reason": "Different workflows and different purposes: self-comment-ci stdin handling vs pr-ci-security-gate security checks." + "reason": "Different workflow fixes: self-comment stdin handling vs PR CI security-gate logic." }, { "left": "pull_request:46138", "right": "pull_request:46558", "accept": false, - "reason": "Different workflows and different purposes; not the same underlying change." + "reason": "Different workflow fixes: self-comment stdin handling vs PR CI security-gate permissions/approval behavior." }, { "left": "pull_request:46138", "right": "pull_request:46563", "accept": false, - "reason": "Different workflows and different purposes; not mergeable as one fix." + "reason": "Different workflow fixes: self-comment stdin handling vs reusable-workflow safety note." }, { "left": "pull_request:46138", "right": "pull_request:46716", "accept": false, - "reason": "Different workflows and different purposes; no shared concrete bug." + "reason": "Different workflow fixes: self-comment stdin handling vs secret-scoping change in security gate." }, { "left": "pull_request:46138", "right": "pull_request:46504", "accept": false, - "reason": "Different workflows and different purposes; self-comment-ci vs build-ci-docker-images." + "reason": "Different workflow fixes: self-comment stdin handling vs build-ci-docker-images command quoting." }, { "left": "pull_request:46432", "right": "pull_request:46503", "accept": false, - "reason": "ssh-runner hardening and benchmark trigger changes are unrelated workflow edits." + "reason": "Unrelated workflow changes: SSH runner hardening vs benchmark dispatch trigger update." }, { "left": "pull_request:46432", "right": "pull_request:46501", "accept": false, - "reason": "ssh-runner hardening vs anti-slop permissions change; unrelated." + "reason": "Unrelated workflow changes: SSH runner hardening vs anti-slop permissions cleanup." }, { "left": "pull_request:46432", "right": "pull_request:46502", "accept": false, - "reason": "ssh-runner hardening vs assign-reviewers checkout ref change; unrelated." + "reason": "Unrelated workflow changes: SSH runner hardening vs assign-reviewers checkout ref change." }, { "left": "pull_request:46432", "right": "pull_request:46554", "accept": false, - "reason": "ssh-runner hardening vs pr-ci-security-gate logic; different workflows and fixes." + "reason": "Different workflow/file and different behavior: SSH runner hardening vs PR CI security-gate checks." }, { "left": "pull_request:46432", "right": "pull_request:46558", "accept": false, - "reason": "Different workflows and different concrete changes; not duplicates." + "reason": "Different workflow/file and different behavior: SSH runner hardening vs PR CI security-gate permissions/approval logic." + }, + { + "left": "pull_request:46432", + "right": "pull_request:46563", + "accept": false, + "reason": "Different workflow/file and different behavior: SSH runner hardening vs security-gate documentation/comment update." + }, + { + "left": "pull_request:46432", + "right": "pull_request:46716", + "accept": false, + "reason": "Different workflow/file and different behavior: SSH runner hardening vs secret-scoping change in security gate." }, { "left": "pull_request:46501", "right": "pull_request:46503", "accept": false, - "reason": "Anti-slop permissions change and benchmark trigger change are unrelated." + "reason": "Same broad workflows area, but different concrete changes: anti-slop permissions vs benchmark dispatch trigger." }, { "left": "pull_request:46501", "right": "pull_request:46502", "accept": false, - "reason": "Anti-slop permissions change vs assign-reviewers checkout ref; unrelated." + "reason": "Different workflow behavior: anti-slop permissions cleanup vs reviewer-assignment checkout ref." }, { "left": "pull_request:46501", "right": "pull_request:46504", "accept": false, - "reason": "Anti-slop permissions change vs build-ci-docker-images commit-message handling; unrelated." + "reason": "Different workflow behavior: anti-slop permissions cleanup vs build-ci-docker-images command quoting." }, { "left": "pull_request:46501", "right": "pull_request:46554", "accept": false, - "reason": "Different workflows and different concrete changes; not the same fix." + "reason": "Different concrete change: anti-slop permissions cleanup vs PR CI security-gate hardening." }, { "left": "pull_request:46502", "right": "pull_request:46503", "accept": false, - "reason": "Assign-reviewers checkout ref change and benchmark trigger change are unrelated." + "reason": "Different workflow behavior: reviewer checkout ref vs benchmark dispatch trigger." }, { "left": "pull_request:46502", "right": "pull_request:46504", "accept": false, - "reason": "Assign-reviewers checkout ref change vs build-ci-docker-images change; unrelated." + "reason": "Different workflow behavior: reviewer checkout ref vs build-ci-docker-images command quoting." }, { "left": "pull_request:46502", "right": "pull_request:46554", "accept": false, - "reason": "Different workflows and different concrete changes; not duplicates." + "reason": "Different concrete change: reviewer checkout ref vs PR CI security-gate hardening." }, { "left": "pull_request:46503", "right": "pull_request:46504", "accept": false, - "reason": "Benchmark trigger condition change vs build-ci-docker-images commit-message handling; unrelated." + "reason": "Different workflow behavior: benchmark dispatch trigger vs build-ci-docker-images command quoting." }, { "left": "pull_request:46503", "right": "pull_request:46554", "accept": false, - "reason": "Different workflows and different purposes; no shared underlying change." + "reason": "Different concrete change: benchmark dispatch trigger vs PR CI security-gate hardening." }, { "left": "pull_request:46504", "right": "pull_request:46554", "accept": false, - "reason": "Build-ci-docker-images token-safe grep change vs pr-ci-security-gate workflow security changes; unrelated." + "reason": "Different workflow behavior: build-ci-docker-images command quoting vs PR CI security-gate checks." }, { "left": "pull_request:46504", "right": "pull_request:46558", "accept": false, - "reason": "Build-ci-docker-images change and pr-ci-security-gate change are unrelated." + "reason": "Different workflow behavior: build-ci-docker-images command quoting vs PR CI security-gate permissions/approval logic." }, { "left": "pull_request:46554", "right": "pull_request:46558", "accept": false, - "reason": "Same file, but the diffs are different concrete edits to the workflow, so they are not the same underlying change." + "reason": "Same file, but different concrete hardening steps: checkout-and-diff checks vs permissions/approval gating." }, { "left": "pull_request:46554", "right": "pull_request:46563", "accept": false, - "reason": "Same file, but this is a different incremental edit to pr-ci-security-gate.yml, not the same fix." + "reason": "Same file, but different concrete changes: workflow checks vs reusable-workflow safety commentary." }, { "left": "pull_request:46554", "right": "pull_request:46716", "accept": false, - "reason": "Same file, but the change is a separate follow-up from the security-gate fix; not one duplicate change." + "reason": "Same file, but different concrete changes: workflow checks vs secrets scoping in the reusable workflow call." }, { "left": "pull_request:46558", "right": "pull_request:46563", "accept": false, - "reason": "Same workflow file, but different concrete edits and intents; not duplicates." + "reason": "Same workflow family, but different fixes: approval gating vs safety-note documentation." }, { "left": "pull_request:46558", "right": "pull_request:46716", "accept": false, - "reason": "Same workflow file, but different concrete edits and intents; not the same change." + "reason": "Same workflow family, but different fixes: approval gating vs secret-scoping change." }, { "left": "pull_request:46563", "right": "pull_request:46716", "accept": false, - "reason": "Same workflow file, but these are distinct incremental updates, not one underlying bugfix." + "reason": "Same file, but different changes: documentation note vs secrets inheritance removal." }, { "left": "pull_request:46504", "right": "pull_request:46563", "accept": false, - "reason": "Unrelated workflow changes; build-ci-docker-images vs pr-ci-security-gate." + "reason": "Different workflow behavior: build-ci-docker-images command quoting vs security-gate documentation note." }, { "left": "pull_request:46432", "right": "pull_request:46563", "accept": false, - "reason": "Different workflows and different purposes; unrelated." + "reason": "Different workflow/file and different behavior: SSH runner hardening vs security-gate documentation note." }, { "left": "pull_request:46504", "right": "pull_request:46716", "accept": false, - "reason": "Unrelated workflow changes; build-ci-docker-images vs pr-ci-security-gate." + "reason": "Different workflow behavior: build-ci-docker-images command quoting vs secret-scoping change." }, { "left": "pull_request:46432", "right": "pull_request:46716", "accept": false, - "reason": "Different workflows and different purposes; unrelated." + "reason": "Different workflow/file and different behavior: SSH runner hardening vs security-gate secret-scoping change." } ] }, @@ -27555,12 +27569,12 @@ "soft_pairs": [ "pull_request:46310|pull_request:47162" ], - "prepared_review_unit_hash": "c34e0d8dc5e468cc7b3f18ea450c4be5e1251da703066ca81fa3a19049599e2f", + "prepared_review_unit_hash": "54fa991378c2d1f6bcbcec5c227c7ee260121072e6fe6ac17b5f03d794b291a0", "component_budget": { "node_count": 2, "item_count": 2, "soft_pair_count": 1, - "serialized_chars": 4882, + "serialized_chars": 4883, "estimated_input_tokens": 1221, "estimated_eval_tokens": 2698 }, @@ -27568,7 +27582,7 @@ "node_count": 2, "item_count": 2, "soft_pair_count": 1, - "serialized_chars": 4882, + "serialized_chars": 4883, "estimated_input_tokens": 1221, "estimated_eval_tokens": 2698 }, @@ -27577,18 +27591,18 @@ "aggressively_trimmed": false, "split": false, "analyst_result": { - "summary": "Both PRs address the same off-by-one bug in the sliding-window mask right boundary (`-dist <= right_window_size` vs `<`). PR 47162 is the more complete fix because it also updates Gemma4 and adds tests, while PR 46310 is the narrower MoonshineStreaming-only version.", - "confidence": 0.94, + "summary": "Both PRs fix the same MoonshineStreaming sliding-window right-bound off-by-one, changing `-dist < right_window_size` to `-dist <= right_window_size` in the same mask function. PR 47162 is broader because it also applies the same fix to Gemma4 and adds tests, but the shared MoonshineStreaming change is the same underlying bug.", + "confidence": 0.95, "canonical_issue_reason": null, - "canonical_pr_reason": "PR 47162 is the better canonical PR because it fixes the same off-by-one mask bug in MoonshineStreaming and also applies the same correction to Gemma4 with accompanying tests, making it the more complete representative change.", + "canonical_pr_reason": "PR 46310 is the narrower, more direct fix for the MoonshineStreaming off-by-one bug and best represents the specific issue target shared by both PRs.", "best_issue_reason": null, - "best_pr_reason": "PR 47162 is the strongest standalone PR in the cluster: it covers the same concrete bug, includes the broader affected model path, and adds test coverage.", + "best_pr_reason": "PR 47162 is the strongest representative overall because it includes the same MoonshineStreaming fix plus the related Gemma4 variant and tests, making it the most complete candidate in the cluster.", "soft_edge_verdicts": [ { "left": "pull_request:46310", "right": "pull_request:47162", "accept": true, - "reason": "Same underlying off-by-one right-window masking bug in the sliding_window_mask_function; 47162 is a superset-like follow-up that fixes the same code-path and could plausibly be merged into one PR with 46310." + "reason": "Same concrete off-by-one in the same MoonshineStreaming mask function; 47162 also carries the same fix and is plausibly part of one combined change set." } ] }, @@ -27618,12 +27632,12 @@ "soft_pairs": [ "pull_request:46574|pull_request:46580" ], - "prepared_review_unit_hash": "ca1d1881bbcd3b05cd0c9156cf25222a9f3c300b0ec74df5e86c8992bc6e1db8", + "prepared_review_unit_hash": "f68334f27ac9be91fae0fba9162a991fb2db64a72271925caa9b5e6cf1272af8", "component_budget": { "node_count": 2, "item_count": 2, "soft_pair_count": 1, - "serialized_chars": 4409, + "serialized_chars": 4410, "estimated_input_tokens": 1103, "estimated_eval_tokens": 2462 }, @@ -27631,7 +27645,7 @@ "node_count": 2, "item_count": 2, "soft_pair_count": 1, - "serialized_chars": 4409, + "serialized_chars": 4410, "estimated_input_tokens": 1103, "estimated_eval_tokens": 2462 }, @@ -27640,18 +27654,18 @@ "aggressively_trimmed": false, "split": false, "analyst_result": { - "summary": "These two PRs are unrelated tokenizer fixes: one changes CodeLlama decode behavior to preserve leading spaces, the other changes AutoTokenizer checkpoint-prefix handling for DeepSeek-Coder/DeepSeek-LLM. They touch different code paths and are not candidates for consolidation.", - "confidence": 0.98, + "summary": "These are two different pull requests touching different code paths: one fixes CodeLlama decode behavior by moving tokenizer decoder setup, while the other fixes AutoTokenizer class selection for specific DeepSeek checkpoints. They are not duplicates and should not be merged together.", + "confidence": 0.97, "canonical_issue_reason": null, - "canonical_pr_reason": "No canonical PR: the PRs address different bugs in different tokenizer subsystems and do not represent the same underlying change.", + "canonical_pr_reason": "No canonical PR for duplication: the PRs address separate bugs in separate subsystems (CodeLlama tokenizer decode vs AutoTokenizer checkpoint routing).", "best_issue_reason": null, - "best_pr_reason": "No single best PR for a duplicate cluster, because there is no duplicate relationship here; each PR is independently scoped to a separate issue.", + "best_pr_reason": "PR #46574 is the cleaner representative of a single concrete bug fix, with a focused code change and an explicit issue target; PR #46580 is a different, unrelated compatibility fix.", "soft_edge_verdicts": [ { "left": "pull_request:46574", "right": "pull_request:46580", "accept": false, - "reason": "Different problems and code paths: CodeLlama decoder stripping vs DeepSeek checkpoint-to-tokenizer-class resolution. They are not the same fix and would not plausibly merge into one PR." + "reason": "Reject: they solve different problems on different paths. #46574 fixes CodeLlama decoder setup so real leading spaces are preserved on decode; #46580 changes AutoTokenizer checkpoint-prefix routing for DeepSeek models that load the wrong tokenizer class. They are not the same underlying change and would not be merged as one PR." } ] }, @@ -27681,12 +27695,12 @@ "soft_pairs": [ "pull_request:46773|pull_request:47087" ], - "prepared_review_unit_hash": "3ee25ce94368dc657295d9d2975ae31780dc5200b57b9d8f64609ae7c0f2800c", + "prepared_review_unit_hash": "f2a70d74e39fef43949049812bdd0b003a9eef8dbc11aabce1751176a13d9433", "component_budget": { "node_count": 2, "item_count": 2, "soft_pair_count": 1, - "serialized_chars": 5822, + "serialized_chars": 5821, "estimated_input_tokens": 1456, "estimated_eval_tokens": 3168 }, @@ -27694,7 +27708,7 @@ "node_count": 2, "item_count": 2, "soft_pair_count": 1, - "serialized_chars": 5822, + "serialized_chars": 5821, "estimated_input_tokens": 1456, "estimated_eval_tokens": 3168 }, @@ -27703,18 +27717,18 @@ "aggressively_trimmed": false, "split": false, "analyst_result": { - "summary": "These are related only at a high level (both fix padding-mask handling in linear/recurrent attention models), but they target different code paths and failure modes: one is a model-forward guard for batch-size-1 masking, the other is recurrent mask creation for continued forwards/chunked prefill/cache continuation.", - "confidence": 0.95, + "summary": "The two pull requests address different masking bugs: one fixes padding masks being skipped for batch size 1 in linear-attention models, while the other fixes padding masks being skipped on continued forwards / chunked prefill / cache continuation. They share some masking code but are not the same underlying change.", + "confidence": 0.92, "canonical_issue_reason": null, - "canonical_pr_reason": "PR #47087 is the better canonical PR because it fixes the more general masking utility path, has an explicit issue target, and covers the shared recurrent-forward bug across several models.", + "canonical_pr_reason": "47087 is the better canonical PR because it fixes the broader shared masking utility behavior, has an explicit issue target, and covers the recurrent continued-forward path across several model families. 46773 is narrower and focused on a batch-size-1 edge case in linear-attention models.", "best_issue_reason": null, - "best_pr_reason": "PR #47087 is the stronger representative of the cluster because it addresses the core reusable mask-construction logic rather than a single forward-path condition in specific models.", + "best_pr_reason": "47087 is the strongest overall PR in the cluster: it addresses a concrete, cross-model masking bug in the shared utility layer and is clearly scoped around one code-path problem. 46773 is useful but more specialized and less representative of the cluster.", "soft_edge_verdicts": [ { "left": "pull_request:46773", "right": "pull_request:47087", "accept": false, - "reason": "Do not merge: #46773 fixes skipping padding masks when batch size is 1 in linear-attention model forwards, while #47087 fixes recurrent padding-mask creation being skipped on continued multi-token forwards. They overlap in theme but are different bugs and code paths." + "reason": "Different bugs and code paths: 46773 fixes batch-size-1 masking being skipped in linear-attention models, while 47087 fixes masking being skipped on continued forwards/chunked prefill/cache continuation. Shared masking code is not enough to treat them as duplicates." } ] }, @@ -27746,12 +27760,12 @@ "pull_request:47074|pull_request:47076", "pull_request:47074|pull_request:47195" ], - "prepared_review_unit_hash": "a72e5ecf07f255b351ff3b9f267315a0e06f367d43cba3abce16b77b3ae73b9a", + "prepared_review_unit_hash": "024bbf134c22cbafc8901d61314a057784c87b7ae8c78621da1aeedb3453b378", "component_budget": { "node_count": 3, "item_count": 3, "soft_pair_count": 2, - "serialized_chars": 6434, + "serialized_chars": 6433, "estimated_input_tokens": 1609, "estimated_eval_tokens": 3474 }, @@ -27759,7 +27773,7 @@ "node_count": 3, "item_count": 3, "soft_pair_count": 2, - "serialized_chars": 6434, + "serialized_chars": 6433, "estimated_input_tokens": 1609, "estimated_eval_tokens": 3474 }, @@ -27768,24 +27782,24 @@ "aggressively_trimmed": false, "split": false, "analyst_result": { - "summary": "These three PRs are only superficially similar: they all touch Transformers infrastructure, but they fix different concrete code paths. PR 47074 normalizes AnyToAnyPipeline input keys, PR 47076 validates Gemma4 audio placeholder batching, and PR 47195 corrects a generation candidate length cap. They should not be clustered as duplicates.", + "summary": "These are three unrelated pull requests in different code paths: multimodal pipeline key normalization, Gemma4 audio placeholder validation, and generation length-capping. The soft similarities are superficial, so none should be merged as duplicates.", "confidence": 0.96, "canonical_issue_reason": null, - "canonical_pr_reason": "No pair looks like the same underlying change, so there is no true canonical duplicate PR. If one PR must stand in as the cluster representative, 47074 is the broadest and most central user-facing fix because it addresses the pipeline input-key mismatch directly.", + "canonical_pr_reason": "No canonical PR: the PRs target different bugs in different subsystems and are not mergeable into one change.", "best_issue_reason": null, - "best_pr_reason": "47074 is the best representative of the cluster only in a loose sense: it is a clear, self-contained fix with the broadest impact and strongest evidence of user-facing behavior correction. It is still not a duplicate of the others.", + "best_pr_reason": "47074 is the clearest standalone bug fix with a minimal code change and tests, but it is not a duplicate of the other PRs.", "soft_edge_verdicts": [ { "left": "pull_request:47074", "right": "pull_request:47076", "accept": false, - "reason": "Both relate to multimodal inputs, but the concrete bugs differ: 47074 fixes the AnyToAnyPipeline passing `videos` instead of `video`, while 47076 adds Gemma4 processor validation for audio/image placeholder batching. Different components and different failure modes." + "reason": "Different bugs and code paths: AnyToAnyPipeline input key normalization vs Gemma4 processor placeholder validation." }, { "left": "pull_request:47074", "right": "pull_request:47195", "accept": false, - "reason": "These address unrelated code paths: 47074 is a multimodal pipeline key normalization fix, while 47195 changes the candidate generation length cap in decoding. Shared Transformers context is too broad to treat as the same change." + "reason": "Different areas and fixes: multimodal pipeline input keys vs generation candidate length budgeting." } ] }, @@ -27847,193 +27861,184 @@ "pull_request:47280|pull_request:47282", "pull_request:47310|pull_request:47380" ], - "prepared_review_unit_hash": "70321143d91dfd8dce53711048e6927eda39cd88c02a960cf15b675f3cf1c5e9", + "prepared_review_unit_hash": "3c8f9e6663b4b5d7de0b33474a92cbd249c4c8a644e5b27e376601f3489162f2", "component_budget": { "node_count": 10, "item_count": 10, "soft_pair_count": 25, - "serialized_chars": 28959, - "estimated_input_tokens": 7240, - "estimated_eval_tokens": 14736 + "serialized_chars": 28950, + "estimated_input_tokens": 7238, + "estimated_eval_tokens": 14732 }, "budget": { "node_count": 10, "item_count": 10, "soft_pair_count": 25, - "serialized_chars": 28959, - "estimated_input_tokens": 7240, - "estimated_eval_tokens": 14736 + "serialized_chars": 28950, + "estimated_input_tokens": 7238, + "estimated_eval_tokens": 14732 }, "overflow_policy": "truncate_then_skip", "trimmed": false, "aggressively_trimmed": false, "split": false, "analyst_result": { - "summary": "This cluster is mostly a set of unrelated auto-generated PRs across different models and failure modes. The real duplicate subclusters are MiniMax-M3-VL load_error fixes and Mamba2 OOM fixes; the EdgeTam mixed bundle 47224 is not a conservative duplicate of 47122 and should be rejected.", - "confidence": 0.69, + "summary": "This cluster is heterogeneous and splits into separate duplicate groups rather than one shared fix. The clearest duplicates are the MiniMax-M3-VL load_error/offload-folder pair (47123, 47280) and the Mamba2 OOM pair (47282, 47380); the remaining PRs target different models or different failure modes.", + "confidence": 0.91, "canonical_issue_reason": null, - "canonical_pr_reason": "47280 is the strongest representative of the MiniMax-M3-VL load_error subcluster: it fixes the same loading path as 47123, but with the more complete offload-folder handling and cleanup.", + "canonical_pr_reason": "The cluster is mixed, but 47380 is a strong representative of one real duplicate subgroup: it addresses the same Mamba2 OOM test issue as 47282 with a concrete memory-cleanup fix in the same test file.", "best_issue_reason": null, - "best_pr_reason": "47280 is the cleanest focused PR in the MiniMax-M3-VL duplicate subcluster and best represents the underlying load_error workaround.", + "best_pr_reason": "47380 is the best single PR representative because it fixes a concrete, self-contained Mamba2 OOM path rather than just adjusting expectations, and it clearly matches 47282.", "soft_edge_verdicts": [ { "left": "pull_request:47122", "right": "pull_request:47281", "accept": false, - "reason": "Different models and code paths: EdgeTam kwargs cleanup vs GLM46V modular changes." + "reason": "Different models and unrelated fixes: Edgetam kwargs handling vs GLM46V model changes." }, { "left": "pull_request:47122", "right": "pull_request:47123", "accept": false, - "reason": "Different models and failure modes: EdgeTam import_or_config vs MiniMax-M3-VL load_error." + "reason": "Different models and failure modes: Edgetam import_or_config vs MiniMax-M3-VL load_error." }, { "left": "pull_request:47122", "right": "pull_request:47280", "accept": false, - "reason": "Different models and fixes: EdgeTam kwargs cleanup vs MiniMax-M3-VL offload-folder loading." + "reason": "Different models and unrelated code paths; not the same bug/change." }, { "left": "pull_request:47122", "right": "pull_request:47224", "accept": false, - "reason": "Reject conservatively: 47224 is a mixed multi-file bundle, so it does not cleanly look like a single-PR duplicate of the EdgeTam fix even though it shares that code hunk." + "reason": "47224 bundles Edgetam with other tokenizer/whisper fixes, so it is not the same standalone change as 47122." }, { "left": "pull_request:47122", "right": "pull_request:47310", "accept": false, - "reason": "Unrelated models and problems: EdgeTam import_or_config vs Gemma3 generation-config behavior." + "reason": "Different models and issues: Edgetam kwargs passthrough vs Gemma3 cache-implementation behavior." }, { "left": "pull_request:47123", "right": "pull_request:47280", "accept": true, - "reason": "Same MiniMax-M3-VL load_error workaround in the same test path; both add offload-folder handling for auto device_map loading." + "reason": "Same MiniMax-M3-VL load_error/offload-folder workaround in the same test path." }, { "left": "pull_request:47123", "right": "pull_request:47266", "accept": false, - "reason": "Both are load_error test fixes, but for different models and likely different root causes." + "reason": "Both are load_error-style fixes, but for different models and different root causes." }, { "left": "pull_request:47123", "right": "pull_request:47281", "accept": false, - "reason": "Different models and unrelated fixes: MiniMax-M3-VL load_error vs GLM46V changes." + "reason": "Different model and unrelated fix area." }, { "left": "pull_request:47123", "right": "pull_request:47124", "accept": false, - "reason": "Different models and failure modes: MiniMax-M3-VL load_error vs Falcon-Mamba output mismatch." + "reason": "Different models and symptoms; MiniMax load_error is not the same as Falcon-Mamba output mismatch." }, { "left": "pull_request:47124", "right": "pull_request:47310", "accept": false, - "reason": "Different models and behaviors: Falcon-Mamba output mismatch vs Gemma3 generation-config change." + "reason": "Different models and concrete issues; Falcon-Mamba output mismatch vs Gemma3 generation-config handling." }, { "left": "pull_request:47124", "right": "pull_request:47282", "accept": false, - "reason": "Different models and problems: Falcon-Mamba output mismatch vs Mamba2 OOM test fix." + "reason": "Different model and failure mode; output mismatch is not the same as Mamba2 OOM." }, { "left": "pull_request:47124", "right": "pull_request:47380", "accept": false, - "reason": "Different models and failure causes: Falcon-Mamba output mismatch vs Mamba2 OOM cleanup." + "reason": "Different models and unrelated fixes." }, { "left": "pull_request:47124", "right": "pull_request:47281", "accept": false, - "reason": "Different models and code paths: Falcon-Mamba test outputs vs GLM46V modular changes." + "reason": "Different model and code path." }, { "left": "pull_request:47266", "right": "pull_request:47280", "accept": false, - "reason": "Different models and different load_error fixes: DeepSeekV3.2 vs MiniMax-M3-VL." + "reason": "Different models and different issues; DeepSeekV3.2 load_error is not the MiniMax offload-folder fix." }, { "left": "pull_request:47266", "right": "pull_request:47380", "accept": false, - "reason": "Different models and causes: DeepSeekV3.2 load_error vs Mamba2 OOM." + "reason": "Different models and failure modes." }, { "left": "pull_request:47266", "right": "pull_request:47281", "accept": false, - "reason": "Different models and code paths: DeepSeekV3.2 test loading vs GLM46V model changes." + "reason": "DeepSeek integration-test setup vs GLM46V model rewrite are unrelated." }, { "left": "pull_request:47280", "right": "pull_request:47281", "accept": false, - "reason": "Different models and fixes: MiniMax-M3-VL load_error workaround vs GLM46V modular changes." + "reason": "Different models and unrelated fixes." }, { "left": "pull_request:47280", "right": "pull_request:47310", "accept": false, - "reason": "Different models and failure types: MiniMax-M3-VL load_error vs Gemma3 generation-config behavior." + "reason": "Different models and issues." }, { "left": "pull_request:47280", "right": "pull_request:47380", "accept": false, - "reason": "Different models and unrelated issues: MiniMax-M3-VL load_error vs Mamba2 OOM." + "reason": "Different models; MiniMax-M3-VL load_error is not the same as Mamba2 OOM cleanup." }, { "left": "pull_request:47281", "right": "pull_request:47282", "accept": false, - "reason": "Different models and problems: GLM46V modular refactor vs Mamba2 OOM." + "reason": "GLM46V model changes vs Mamba2 OOM test fix are unrelated." }, { "left": "pull_request:47282", "right": "pull_request:47380", "accept": true, - "reason": "Same Mamba2 OOM/test instability in the same file; both reduce memory pressure and could plausibly be merged into one fix." + "reason": "Same Mamba2 OOM issue in the same integration test file; both are memory-hygiene workarounds for the same bug." }, { "left": "pull_request:47282", "right": "pull_request:47310", "accept": false, - "reason": "Different models and failure modes: Mamba2 OOM vs Gemma3 generation-config change." + "reason": "Different models and failure modes: Mamba2 OOM vs Gemma3 cache-implementation behavior." }, { "left": "pull_request:47280", "right": "pull_request:47282", "accept": false, - "reason": "Different models and unrelated fixes: MiniMax-M3-VL load_error vs Mamba2 OOM." + "reason": "Different models and unrelated issues." }, { "left": "pull_request:47310", "right": "pull_request:47380", "accept": false, - "reason": "Different models and unrelated fixes: Gemma3 generation-config behavior vs Mamba2 OOM cleanup." - }, - { - "left": "pull_request:47124", - "right": "pull_request:47266", - "accept": false, - "reason": "Different models and code paths: Falcon-Mamba output mismatch vs DeepSeekV3.2 load_error." + "reason": "Different models and independent fixes." } ] }, - "evaluator_result": { - "accept": false, - "feedback": "The overall cluster summary is plausible, but one soft-edge acceptance is too aggressive for the packet: 47122\u201347224. PR 47224 is a mixed bundle across several files/models, so it does not cleanly look like a single-PR duplicate of the EdgeTam fix even though it shares that code change. Because the instructions call for conservative verdicts and rejecting pairs that do not clearly look mergeable into a single bugfix PR, this makes the output non-conservative. The other accepted pairs are better grounded." - }, - "evaluator_used": true, - "retried": true, + "evaluator_result": null, + "evaluator_used": false, + "retried": false, "accepted_nontrivial_soft_edge": false, "error_kind": null, "error_message": null @@ -28057,47 +28062,50 @@ "soft_pairs": [ "pull_request:47140|pull_request:47303" ], - "prepared_review_unit_hash": "c8d974fa56edd6af094727fe6084237f11b7166837918d20545a859c78519aab", + "prepared_review_unit_hash": "bb7a39453883a6637e8b07fc394147af51eea240759278f44c01b0248ad7367c", "component_budget": { "node_count": 2, "item_count": 2, "soft_pair_count": 1, - "serialized_chars": 4408, - "estimated_input_tokens": 1102, - "estimated_eval_tokens": 2460 + "serialized_chars": 4409, + "estimated_input_tokens": 1103, + "estimated_eval_tokens": 2462 }, "budget": { "node_count": 2, "item_count": 2, "soft_pair_count": 1, - "serialized_chars": 4408, - "estimated_input_tokens": 1102, - "estimated_eval_tokens": 2460 + "serialized_chars": 4409, + "estimated_input_tokens": 1103, + "estimated_eval_tokens": 2462 }, "overflow_policy": "truncate_then_skip", "trimmed": false, "aggressively_trimmed": false, "split": false, "analyst_result": { - "summary": "These are related chat CLI URL-fix PRs, but they target different fixes/issue tracks and use different approaches: one normalizes a `/v1` base in the interface, the other routes management calls to the service root for model loading. They are similar enough to compare, but not the same duplicate change.", - "confidence": 0.83, + "summary": "Both PRs address the same chat CLI bug: management endpoints were being built against a `/v1` API base instead of the service root. They appear to be alternative fixes for the same code path in `src/transformers/cli/chat.py`.", + "confidence": 0.78, "canonical_issue_reason": null, - "canonical_pr_reason": "PR 47303 is the better representative because it expresses the broader service-root routing fix more clearly and directly updates the management call site, while 47140 is a narrower `/v1` base normalization variant tied to a different issue.", + "canonical_pr_reason": "PR #47140 is the more direct canonical fix: it normalizes the CLI base URL once in `RichInterface`, so all management requests inherit the correct service-root URL rather than patching individual calls.", "best_issue_reason": null, - "best_pr_reason": "47303 is the strongest single PR in the cluster: it has the clearer intent, explicit management-root routing behavior, and a more targeted code path. 47140 is adjacent but not the same concrete fix.", + "best_pr_reason": "PR #47140 is the best overall fit because it cleanly addresses the root cause for the chat CLI\u2019s management URLs and is likely to cover the affected call paths consistently.", "soft_edge_verdicts": [ { "left": "pull_request:47140", "right": "pull_request:47303", - "accept": false, - "reason": "Both address chat CLI management URL handling, but 47140 normalizes the interface base URL for `/v1` generally, while 47303 routes a specific management request to the service root. Different issue targets and different concrete code-path fixes make them too distinct to treat as duplicates." + "accept": true, + "reason": "Both PRs fix the same underlying chat CLI management-URL bug: requests were being sent relative to a `/v1` base instead of the service root. The helpers and URL rewrites are alternative implementations of the same change, so this is a valid duplicate/overlap pair." } ] }, - "evaluator_result": null, - "evaluator_used": false, + "evaluator_result": { + "accept": true, + "feedback": "The pair looks grounded as an overlap: both PRs change `src/transformers/cli/chat.py`/tests to route chat management calls to the service root instead of inheriting a `/v1` base. The implementations differ (central base-URL normalization vs. per-call routing helper), but they appear to target the same bug family and could plausibly be merged into one fix." + }, + "evaluator_used": true, "retried": false, - "accepted_nontrivial_soft_edge": false, + "accepted_nontrivial_soft_edge": true, "error_kind": null, "error_message": null }, @@ -28120,12 +28128,12 @@ "soft_pairs": [ "pull_request:47261|pull_request:47262" ], - "prepared_review_unit_hash": "26064551f9eacd3358862f8555d8fa17c1c122d7822a7dd9054b5b7758879c50", + "prepared_review_unit_hash": "ef65887c2246e335615ec85da7d2368cc6b7b1fb1a75377c86778156799ad9a6", "component_budget": { "node_count": 2, "item_count": 2, "soft_pair_count": 1, - "serialized_chars": 4417, + "serialized_chars": 4418, "estimated_input_tokens": 1105, "estimated_eval_tokens": 2466 }, @@ -28133,7 +28141,7 @@ "node_count": 2, "item_count": 2, "soft_pair_count": 1, - "serialized_chars": 4417, + "serialized_chars": 4418, "estimated_input_tokens": 1105, "estimated_eval_tokens": 2466 }, @@ -28142,18 +28150,18 @@ "aggressively_trimmed": false, "split": false, "analyst_result": { - "summary": "The two PRs are in different parts of the library and fix unrelated problems: one adjusts label smoothing under gradient accumulation in training, the other adds/validates min_tokens_to_keep for TopHLogitsWarper in generation.", + "summary": "Two unrelated pull requests: one fixes label smoothing/gradient accumulation loss scaling in training, the other adds and enforces `min_tokens_to_keep` for TopH logits warping in generation. They touch different code paths and should not be deduplicated.", "confidence": 0.98, "canonical_issue_reason": null, - "canonical_pr_reason": "PR 47261 is the stronger representative of the cluster because it is a focused bug fix with an explicit issue target and a clear code-path change; 47262 is a separate generation feature/bug fix unrelated to training loss scaling.", + "canonical_pr_reason": "PR 47261 is the better canonical PR if one must be chosen: it has an explicit issue target (46731) and a clearly scoped training-loss fix with tests.", "best_issue_reason": null, - "best_pr_reason": "PR 47261 is the best standalone representative, but only because it is the more concrete bug fix; it should not be treated as a duplicate of 47262.", + "best_pr_reason": "PR 47261 is the strongest representative of the cluster because it is issue-linked and has the clearest bug-to-fix correspondence; PR 47262 is a separate generation feature/bugfix.", "soft_edge_verdicts": [ { "left": "pull_request:47261", "right": "pull_request:47262", "accept": false, - "reason": "Reject: they address different subsystems and different failure modes. 47261 fixes label smoothing loss scaling during gradient accumulation in Trainer, while 47262 changes TopHLogitsWarper behavior and validation in generation. They are not the same underlying bug or mergeable into one PR." + "reason": "Different underlying problems in different subsystems: training loss scaling vs generation logits warping. No shared code path or mergeable common fix." } ] }, diff --git a/analysis/current/manifest.json b/analysis/current/manifest.json index dd6916434d4016af6a24b0183b6003596f6613f3..e34484c16cd458f8afb18f02ffccff4ec770f142 100644 --- a/analysis/current/manifest.json +++ b/analysis/current/manifest.json @@ -1,8 +1,8 @@ { - "analysis_id": "hybrid-model-20260718t060036z", + "analysis_id": "hybrid-model-20260718t120032z", "archived_artifacts": { - "hybrid": "snapshots/20260718T060036Z/analysis-runs/hybrid-model-20260718t060036z/analysis-report-hybrid.json", - "hybrid_reviews": "snapshots/20260718T060036Z/analysis-runs/hybrid-model-20260718t060036z/analysis-report-hybrid.llm-reviews.json" + "hybrid": "snapshots/20260718T120032Z/analysis-runs/hybrid-model-20260718t120032z/analysis-report-hybrid.json", + "hybrid_reviews": "snapshots/20260718T120032Z/analysis-runs/hybrid-model-20260718t120032z/analysis-report-hybrid.llm-reviews.json" }, "artifacts": { "hybrid": "analysis/current/analysis-report-hybrid.json", @@ -10,9 +10,9 @@ }, "channel": "canonical", "model": null, - "published_at": "2026-07-18T06:12:23Z", + "published_at": "2026-07-18T12:15:00Z", "repo": "huggingface/transformers", "schema_version": 1, - "snapshot_id": "20260718T060036Z", + "snapshot_id": "20260718T120032Z", "variant": "hybrid" } diff --git a/snapshots/20260718T120032Z/analysis-runs/hybrid-model-20260718t120032z/analysis-report-hybrid.json b/snapshots/20260718T120032Z/analysis-runs/hybrid-model-20260718t120032z/analysis-report-hybrid.json new file mode 100644 index 0000000000000000000000000000000000000000..50161d03e511b3d348b487662470a38a6906ecbb --- /dev/null +++ b/snapshots/20260718T120032Z/analysis-runs/hybrid-model-20260718t120032z/analysis-report-hybrid.json @@ -0,0 +1,3143 @@ +{ + "schema_version": "1.0", + "repo": "huggingface/transformers", + "snapshot_id": "20260718T120032Z", + "generated_at": "2026-07-18T12:14:56Z", + "evidence_quality": "full", + "llm_enrichment": true, + "meta_bugs": [ + { + "cluster_id": "cluster-43979-11", + "summary": "Cluster of 1 issues and 10 PRs centered on issue #43979.", + "status": "open", + "confidence": 0.8, + "canonical_issue_number": 43979, + "canonical_pr_number": 44007, + "issue_numbers": [ + 43979 + ], + "pr_numbers": [ + 43996, + 44007, + 44013, + 44044, + 44066, + 44072, + 44085, + 44129, + 44154, + 44722 + ], + "evidence_types": [ + "closing_reference", + "shared_issue_target", + "soft_similarity" + ], + "pr_comparisons": [ + { + "left_pr_number": 43996, + "right_pr_number": 44007, + "code_similarity": 0.179, + "size_similarity": 0.576, + "file_overlap": 0.0, + "area_overlap": 0.0, + "patch_similarity": 0.429, + "shared_filenames": [], + "shared_file_areas": [] + }, + { + "left_pr_number": 43996, + "right_pr_number": 44013, + "code_similarity": 0.122, + "size_similarity": 0.318, + "file_overlap": 0.0, + "area_overlap": 0.0, + "patch_similarity": 0.392, + "shared_filenames": [], + "shared_file_areas": [] + }, + { + "left_pr_number": 43996, + "right_pr_number": 44044, + "code_similarity": 0.245, + "size_similarity": 0.864, + "file_overlap": 0.0, + "area_overlap": 0.0, + "patch_similarity": 0.479, + "shared_filenames": [], + "shared_file_areas": [] + }, + { + "left_pr_number": 43996, + "right_pr_number": 44066, + "code_similarity": 0.225, + "size_similarity": 0.818, + "file_overlap": 0.0, + "area_overlap": 0.0, + "patch_similarity": 0.408, + "shared_filenames": [], + "shared_file_areas": [] + }, + { + "left_pr_number": 43996, + "right_pr_number": 44072, + "code_similarity": 0.14, + "size_similarity": 0.303, + "file_overlap": 0.0, + "area_overlap": 0.0, + "patch_similarity": 0.528, + "shared_filenames": [], + "shared_file_areas": [] + }, + { + "left_pr_number": 43996, + "right_pr_number": 44085, + "code_similarity": 0.216, + "size_similarity": 0.783, + "file_overlap": 0.0, + "area_overlap": 0.0, + "patch_similarity": 0.398, + "shared_filenames": [], + "shared_file_areas": [] + }, + { + "left_pr_number": 43996, + "right_pr_number": 44129, + "code_similarity": 0.163, + "size_similarity": 0.643, + "file_overlap": 0.0, + "area_overlap": 0.0, + "patch_similarity": 0.229, + "shared_filenames": [], + "shared_file_areas": [] + }, + { + "left_pr_number": 43996, + "right_pr_number": 44154, + "code_similarity": 0.153, + "size_similarity": 0.535, + "file_overlap": 0.0, + "area_overlap": 0.0, + "patch_similarity": 0.31, + "shared_filenames": [], + "shared_file_areas": [] + }, + { + "left_pr_number": 43996, + "right_pr_number": 44722, + "code_similarity": 0.225, + "size_similarity": 0.848, + "file_overlap": 0.0, + "area_overlap": 0.0, + "patch_similarity": 0.368, + "shared_filenames": [], + "shared_file_areas": [] + }, + { + "left_pr_number": 44007, + "right_pr_number": 44013, + "code_similarity": 0.19, + "size_similarity": 0.553, + "file_overlap": 0.0, + "area_overlap": 0.0, + "patch_similarity": 0.531, + "shared_filenames": [], + "shared_file_areas": [] + }, + { + "left_pr_number": 44007, + "right_pr_number": 44044, + "code_similarity": 0.186, + "size_similarity": 0.667, + "file_overlap": 0.0, + "area_overlap": 0.0, + "patch_similarity": 0.354, + "shared_filenames": [], + "shared_file_areas": [] + }, + { + "left_pr_number": 44007, + "right_pr_number": 44066, + "code_similarity": 0.188, + "size_similarity": 0.704, + "file_overlap": 0.0, + "area_overlap": 0.0, + "patch_similarity": 0.315, + "shared_filenames": [], + "shared_file_areas": [] + }, + { + "left_pr_number": 44007, + "right_pr_number": 44072, + "code_similarity": 0.212, + "size_similarity": 0.526, + "file_overlap": 0.0, + "area_overlap": 0.0, + "patch_similarity": 0.708, + "shared_filenames": [], + "shared_file_areas": [] + }, + { + "left_pr_number": 44007, + "right_pr_number": 44085, + "code_similarity": 0.195, + "size_similarity": 0.735, + "file_overlap": 0.0, + "area_overlap": 0.0, + "patch_similarity": 0.318, + "shared_filenames": [], + "shared_file_areas": [] + }, + { + "left_pr_number": 44007, + "right_pr_number": 44129, + "code_similarity": 0.103, + "size_similarity": 0.37, + "file_overlap": 0.0, + "area_overlap": 0.0, + "patch_similarity": 0.191, + "shared_filenames": [], + "shared_file_areas": [] + }, + { + "left_pr_number": 44007, + "right_pr_number": 44154, + "code_similarity": 0.238, + "size_similarity": 0.93, + "file_overlap": 0.0, + "area_overlap": 0.0, + "patch_similarity": 0.344, + "shared_filenames": [], + "shared_file_areas": [] + }, + { + "left_pr_number": 44007, + "right_pr_number": 44722, + "code_similarity": 0.178, + "size_similarity": 0.679, + "file_overlap": 0.0, + "area_overlap": 0.0, + "patch_similarity": 0.28, + "shared_filenames": [], + "shared_file_areas": [] + }, + { + "left_pr_number": 44013, + "right_pr_number": 44044, + "code_similarity": 0.126, + "size_similarity": 0.368, + "file_overlap": 0.0, + "area_overlap": 0.0, + "patch_similarity": 0.351, + "shared_filenames": [], + "shared_file_areas": [] + }, + { + "left_pr_number": 44013, + "right_pr_number": 44066, + "code_similarity": 0.127, + "size_similarity": 0.389, + "file_overlap": 0.0, + "area_overlap": 0.0, + "patch_similarity": 0.325, + "shared_filenames": [], + "shared_file_areas": [] + }, + { + "left_pr_number": 44013, + "right_pr_number": 44072, + "code_similarity": 0.29, + "size_similarity": 0.952, + "file_overlap": 0.0, + "area_overlap": 0.0, + "patch_similarity": 0.667, + "shared_filenames": [], + "shared_file_areas": [] + }, + { + "left_pr_number": 44013, + "right_pr_number": 44085, + "code_similarity": 0.131, + "size_similarity": 0.406, + "file_overlap": 0.0, + "area_overlap": 0.0, + "patch_similarity": 0.329, + "shared_filenames": [], + "shared_file_areas": [] + }, + { + "left_pr_number": 44013, + "right_pr_number": 44129, + "code_similarity": 0.07, + "size_similarity": 0.205, + "file_overlap": 0.0, + "area_overlap": 0.0, + "patch_similarity": 0.192, + "shared_filenames": [], + "shared_file_areas": [] + }, + { + "left_pr_number": 44013, + "right_pr_number": 44154, + "code_similarity": 0.177, + "size_similarity": 0.594, + "file_overlap": 0.0, + "area_overlap": 0.0, + "patch_similarity": 0.389, + "shared_filenames": [], + "shared_file_areas": [] + }, + { + "left_pr_number": 44013, + "right_pr_number": 44722, + "code_similarity": 0.118, + "size_similarity": 0.375, + "file_overlap": 0.0, + "area_overlap": 0.0, + "patch_similarity": 0.287, + "shared_filenames": [], + "shared_file_areas": [] + }, + { + "left_pr_number": 44044, + "right_pr_number": 44066, + "code_similarity": 0.25, + "size_similarity": 0.947, + "file_overlap": 0.0, + "area_overlap": 0.0, + "patch_similarity": 0.404, + "shared_filenames": [], + "shared_file_areas": [] + }, + { + "left_pr_number": 44044, + "right_pr_number": 44072, + "code_similarity": 0.136, + "size_similarity": 0.351, + "file_overlap": 0.0, + "area_overlap": 0.0, + "patch_similarity": 0.442, + "shared_filenames": [], + "shared_file_areas": [] + }, + { + "left_pr_number": 44044, + "right_pr_number": 44085, + "code_similarity": 0.24, + "size_similarity": 0.906, + "file_overlap": 0.0, + "area_overlap": 0.0, + "patch_similarity": 0.394, + "shared_filenames": [], + "shared_file_areas": [] + }, + { + "left_pr_number": 44044, + "right_pr_number": 44129, + "code_similarity": 0.147, + "size_similarity": 0.555, + "file_overlap": 0.0, + "area_overlap": 0.0, + "patch_similarity": 0.243, + "shared_filenames": [], + "shared_file_areas": [] + }, + { + "left_pr_number": 44044, + "right_pr_number": 44154, + "code_similarity": 0.17, + "size_similarity": 0.62, + "file_overlap": 0.0, + "area_overlap": 0.0, + "patch_similarity": 0.306, + "shared_filenames": [], + "shared_file_areas": [] + }, + { + "left_pr_number": 44044, + "right_pr_number": 44722, + "code_similarity": 0.257, + "size_similarity": 0.982, + "file_overlap": 0.0, + "area_overlap": 0.0, + "patch_similarity": 0.402, + "shared_filenames": [], + "shared_file_areas": [] + }, + { + "left_pr_number": 44066, + "right_pr_number": 44072, + "code_similarity": 0.133, + "size_similarity": 0.37, + "file_overlap": 0.0, + "area_overlap": 0.0, + "patch_similarity": 0.393, + "shared_filenames": [], + "shared_file_areas": [] + }, + { + "left_pr_number": 44066, + "right_pr_number": 44085, + "code_similarity": 0.763, + "size_similarity": 0.957, + "file_overlap": 0.5, + "area_overlap": 0.825, + "patch_similarity": 0.887, + "shared_filenames": [ + "src/transformers/models/gptj/modeling_gptj.py" + ], + "shared_file_areas": [ + { + "filename": "src/transformers/models/gptj/modeling_gptj.py", + "left_ranges": [ + [ + 33, + 41 + ], + [ + 174, + 181 + ], + [ + 250, + 257 + ], + [ + 398, + 419 + ], + [ + 425, + 434 + ], + [ + 459, + 466 + ], + [ + 471, + 488 + ], + [ + 518, + 540 + ], + [ + 554, + 560 + ], + [ + 566, + 575 + ], + [ + 580, + 598 + ], + [ + 601, + 606 + ], + [ + 634, + 640 + ], + [ + 646, + 653 + ], + [ + 658, + 674 + ], + [ + 717, + 722 + ], + [ + 738, + 744 + ], + [ + 749, + 772 + ], + [ + 790, + 795 + ] + ], + "right_ranges": [ + [ + 33, + 40 + ], + [ + 173, + 180 + ], + [ + 249, + 256 + ], + [ + 397, + 405 + ], + [ + 408, + 419 + ], + [ + 425, + 434 + ], + [ + 459, + 465 + ], + [ + 470, + 489 + ], + [ + 517, + 539 + ], + [ + 553, + 559 + ], + [ + 565, + 574 + ], + [ + 579, + 597 + ], + [ + 600, + 611 + ], + [ + 633, + 639 + ], + [ + 645, + 652 + ], + [ + 657, + 673 + ], + [ + 716, + 728 + ], + [ + 737, + 743 + ], + [ + 748, + 771 + ], + [ + 789, + 794 + ] + ] + } + ] + }, + { + "left_pr_number": 44066, + "right_pr_number": 44129, + "code_similarity": 0.145, + "size_similarity": 0.526, + "file_overlap": 0.0, + "area_overlap": 0.0, + "patch_similarity": 0.263, + "shared_filenames": [], + "shared_file_areas": [] + }, + { + "left_pr_number": 44066, + "right_pr_number": 44154, + "code_similarity": 0.174, + "size_similarity": 0.654, + "file_overlap": 0.0, + "area_overlap": 0.0, + "patch_similarity": 0.286, + "shared_filenames": [], + "shared_file_areas": [] + }, + { + "left_pr_number": 44066, + "right_pr_number": 44722, + "code_similarity": 0.907, + "size_similarity": 0.964, + "file_overlap": 1.0, + "area_overlap": 0.808, + "patch_similarity": 0.874, + "shared_filenames": [ + "src/transformers/models/codegen/modeling_codegen.py", + "src/transformers/models/gptj/modeling_gptj.py" + ], + "shared_file_areas": [ + { + "filename": "src/transformers/models/codegen/modeling_codegen.py", + "left_ranges": [ + [ + 245, + 266 + ] + ], + "right_ranges": [ + [ + 228, + 234 + ] + ] + }, + { + "filename": "src/transformers/models/gptj/modeling_gptj.py", + "left_ranges": [ + [ + 33, + 41 + ], + [ + 174, + 181 + ], + [ + 250, + 257 + ], + [ + 398, + 419 + ], + [ + 425, + 434 + ], + [ + 459, + 466 + ], + [ + 471, + 488 + ], + [ + 518, + 540 + ], + [ + 554, + 560 + ], + [ + 566, + 575 + ], + [ + 580, + 598 + ], + [ + 601, + 606 + ], + [ + 634, + 640 + ], + [ + 646, + 653 + ], + [ + 658, + 674 + ], + [ + 717, + 722 + ], + [ + 738, + 744 + ], + [ + 749, + 772 + ], + [ + 790, + 795 + ] + ], + "right_ranges": [ + [ + 33, + 42 + ], + [ + 175, + 182 + ], + [ + 251, + 258 + ], + [ + 399, + 420 + ], + [ + 426, + 435 + ], + [ + 460, + 467 + ], + [ + 472, + 489 + ], + [ + 519, + 540 + ], + [ + 554, + 560 + ], + [ + 566, + 575 + ], + [ + 580, + 598 + ], + [ + 601, + 612 + ], + [ + 634, + 640 + ], + [ + 646, + 653 + ], + [ + 658, + 663 + ], + [ + 666, + 674 + ], + [ + 717, + 722 + ], + [ + 738, + 744 + ], + [ + 749, + 772 + ], + [ + 790, + 795 + ] + ] + } + ] + }, + { + "left_pr_number": 44072, + "right_pr_number": 44085, + "code_similarity": 0.137, + "size_similarity": 0.387, + "file_overlap": 0.0, + "area_overlap": 0.0, + "patch_similarity": 0.398, + "shared_filenames": [], + "shared_file_areas": [] + }, + { + "left_pr_number": 44072, + "right_pr_number": 44129, + "code_similarity": 0.074, + "size_similarity": 0.195, + "file_overlap": 0.0, + "area_overlap": 0.0, + "patch_similarity": 0.231, + "shared_filenames": [], + "shared_file_areas": [] + }, + { + "left_pr_number": 44072, + "right_pr_number": 44154, + "code_similarity": 0.175, + "size_similarity": 0.566, + "file_overlap": 0.0, + "area_overlap": 0.0, + "patch_similarity": 0.414, + "shared_filenames": [], + "shared_file_areas": [] + }, + { + "left_pr_number": 44072, + "right_pr_number": 44722, + "code_similarity": 0.124, + "size_similarity": 0.357, + "file_overlap": 0.0, + "area_overlap": 0.0, + "patch_similarity": 0.347, + "shared_filenames": [], + "shared_file_areas": [] + }, + { + "left_pr_number": 44085, + "right_pr_number": 44129, + "code_similarity": 0.141, + "size_similarity": 0.503, + "file_overlap": 0.0, + "area_overlap": 0.0, + "patch_similarity": 0.272, + "shared_filenames": [], + "shared_file_areas": [] + }, + { + "left_pr_number": 44085, + "right_pr_number": 44154, + "code_similarity": 0.18, + "size_similarity": 0.684, + "file_overlap": 0.0, + "area_overlap": 0.0, + "patch_similarity": 0.289, + "shared_filenames": [], + "shared_file_areas": [] + }, + { + "left_pr_number": 44085, + "right_pr_number": 44722, + "code_similarity": 0.728, + "size_similarity": 0.923, + "file_overlap": 0.5, + "area_overlap": 0.791, + "patch_similarity": 0.78, + "shared_filenames": [ + "src/transformers/models/gptj/modeling_gptj.py" + ], + "shared_file_areas": [ + { + "filename": "src/transformers/models/gptj/modeling_gptj.py", + "left_ranges": [ + [ + 33, + 40 + ], + [ + 173, + 180 + ], + [ + 249, + 256 + ], + [ + 397, + 405 + ], + [ + 408, + 419 + ], + [ + 425, + 434 + ], + [ + 459, + 465 + ], + [ + 470, + 489 + ], + [ + 517, + 539 + ], + [ + 553, + 559 + ], + [ + 565, + 574 + ], + [ + 579, + 597 + ], + [ + 600, + 611 + ], + [ + 633, + 639 + ], + [ + 645, + 652 + ], + [ + 657, + 673 + ], + [ + 716, + 728 + ], + [ + 737, + 743 + ], + [ + 748, + 771 + ], + [ + 789, + 794 + ] + ], + "right_ranges": [ + [ + 33, + 42 + ], + [ + 175, + 182 + ], + [ + 251, + 258 + ], + [ + 399, + 420 + ], + [ + 426, + 435 + ], + [ + 460, + 467 + ], + [ + 472, + 489 + ], + [ + 519, + 540 + ], + [ + 554, + 560 + ], + [ + 566, + 575 + ], + [ + 580, + 598 + ], + [ + 601, + 612 + ], + [ + 634, + 640 + ], + [ + 646, + 653 + ], + [ + 658, + 663 + ], + [ + 666, + 674 + ], + [ + 717, + 722 + ], + [ + 738, + 744 + ], + [ + 749, + 772 + ], + [ + 790, + 795 + ] + ] + } + ] + }, + { + "left_pr_number": 44129, + "right_pr_number": 44154, + "code_similarity": 0.099, + "size_similarity": 0.344, + "file_overlap": 0.0, + "area_overlap": 0.0, + "patch_similarity": 0.199, + "shared_filenames": [], + "shared_file_areas": [] + }, + { + "left_pr_number": 44129, + "right_pr_number": 44722, + "code_similarity": 0.146, + "size_similarity": 0.545, + "file_overlap": 0.0, + "area_overlap": 0.0, + "patch_similarity": 0.247, + "shared_filenames": [], + "shared_file_areas": [] + }, + { + "left_pr_number": 44154, + "right_pr_number": 44722, + "code_similarity": 0.164, + "size_similarity": 0.631, + "file_overlap": 0.0, + "area_overlap": 0.0, + "patch_similarity": 0.255, + "shared_filenames": [], + "shared_file_areas": [] + } + ] + }, + { + "cluster_id": "cluster-46894-8", + "summary": "Cluster of 3 issues and 5 PRs centered on issue #46894.", + "status": "open", + "confidence": 0.8, + "canonical_issue_number": 46894, + "canonical_pr_number": 46903, + "issue_numbers": [ + 46894, + 46901, + 47394 + ], + "pr_numbers": [ + 46895, + 46903, + 46958, + 47090, + 47395 + ], + "evidence_types": [ + "closing_reference", + "shared_issue_target", + "soft_similarity" + ], + "pr_comparisons": [ + { + "left_pr_number": 46895, + "right_pr_number": 46903, + "code_similarity": 0.533, + "size_similarity": 0.404, + "file_overlap": 0.333, + "area_overlap": 0.933, + "patch_similarity": 0.17, + "shared_filenames": [ + "src/transformers/models/cohere_asr/modeling_cohere_asr.py", + "src/transformers/models/cohere_asr/modular_cohere_asr.py" + ], + "shared_file_areas": [ + { + "filename": "src/transformers/models/cohere_asr/modeling_cohere_asr.py", + "left_ranges": [ + [ + 637, + 650 + ] + ], + "right_ranges": [ + [ + 637, + 651 + ] + ] + }, + { + "filename": "src/transformers/models/cohere_asr/modular_cohere_asr.py", + "left_ranges": [ + [ + 500, + 513 + ] + ], + "right_ranges": [ + [ + 500, + 514 + ] + ] + } + ] + }, + { + "left_pr_number": 46895, + "right_pr_number": 46958, + "code_similarity": 0.779, + "size_similarity": 0.915, + "file_overlap": 1.0, + "area_overlap": 0.5, + "patch_similarity": 0.81, + "shared_filenames": [ + "src/transformers/models/cohere_asr/modeling_cohere_asr.py", + "src/transformers/models/cohere_asr/modular_cohere_asr.py", + "tests/models/cohere_asr/test_modeling_cohere_asr.py" + ], + "shared_file_areas": [ + { + "filename": "src/transformers/models/cohere_asr/modeling_cohere_asr.py", + "left_ranges": [ + [ + 637, + 650 + ] + ], + "right_ranges": [ + [ + 22, + 28 + ], + [ + 638, + 645 + ] + ] + }, + { + "filename": "src/transformers/models/cohere_asr/modular_cohere_asr.py", + "left_ranges": [ + [ + 500, + 513 + ] + ], + "right_ranges": [ + [ + 16, + 22 + ], + [ + 501, + 508 + ] + ] + }, + { + "filename": "tests/models/cohere_asr/test_modeling_cohere_asr.py", + "left_ranges": [ + [ + 35, + 41 + ], + [ + 176, + 214 + ] + ], + "right_ranges": [ + [ + 15, + 21 + ], + [ + 180, + 221 + ] + ] + } + ] + }, + { + "left_pr_number": 46895, + "right_pr_number": 47090, + "code_similarity": 0.459, + "size_similarity": 0.281, + "file_overlap": 0.273, + "area_overlap": 0.765, + "patch_similarity": 0.355, + "shared_filenames": [ + "src/transformers/models/cohere_asr/modeling_cohere_asr.py", + "src/transformers/models/cohere_asr/modular_cohere_asr.py", + "tests/models/cohere_asr/test_modeling_cohere_asr.py" + ], + "shared_file_areas": [ + { + "filename": "src/transformers/models/cohere_asr/modeling_cohere_asr.py", + "left_ranges": [ + [ + 637, + 650 + ] + ], + "right_ranges": [ + [ + 637, + 649 + ] + ] + }, + { + "filename": "src/transformers/models/cohere_asr/modular_cohere_asr.py", + "left_ranges": [ + [ + 500, + 513 + ] + ], + "right_ranges": [ + [ + 500, + 512 + ] + ] + }, + { + "filename": "tests/models/cohere_asr/test_modeling_cohere_asr.py", + "left_ranges": [ + [ + 35, + 41 + ], + [ + 176, + 214 + ] + ], + "right_ranges": [ + [ + 175, + 224 + ] + ] + } + ] + }, + { + "left_pr_number": 46895, + "right_pr_number": 47395, + "code_similarity": 0.187, + "size_similarity": 0.683, + "file_overlap": 0.0, + "area_overlap": 0.0, + "patch_similarity": 0.333, + "shared_filenames": [], + "shared_file_areas": [] + }, + { + "left_pr_number": 46903, + "right_pr_number": 46958, + "code_similarity": 0.325, + "size_similarity": 0.369, + "file_overlap": 0.333, + "area_overlap": 0.364, + "patch_similarity": 0.156, + "shared_filenames": [ + "src/transformers/models/cohere_asr/modeling_cohere_asr.py", + "src/transformers/models/cohere_asr/modular_cohere_asr.py" + ], + "shared_file_areas": [ + { + "filename": "src/transformers/models/cohere_asr/modeling_cohere_asr.py", + "left_ranges": [ + [ + 637, + 651 + ] + ], + "right_ranges": [ + [ + 22, + 28 + ], + [ + 638, + 645 + ] + ] + }, + { + "filename": "src/transformers/models/cohere_asr/modular_cohere_asr.py", + "left_ranges": [ + [ + 500, + 514 + ] + ], + "right_ranges": [ + [ + 16, + 22 + ], + [ + 501, + 508 + ] + ] + } + ] + }, + { + "left_pr_number": 46903, + "right_pr_number": 47090, + "code_similarity": 0.576, + "size_similarity": 0.695, + "file_overlap": 0.333, + "area_overlap": 0.871, + "patch_similarity": 0.214, + "shared_filenames": [ + "src/transformers/models/cohere_asr/modeling_cohere_asr.py", + "src/transformers/models/cohere_asr/modular_cohere_asr.py", + "src/transformers/models/pp_formulanet/modeling_pp_formulanet.py", + "src/transformers/models/pp_formulanet/modular_pp_formulanet.py" + ], + "shared_file_areas": [ + { + "filename": "src/transformers/models/cohere_asr/modeling_cohere_asr.py", + "left_ranges": [ + [ + 637, + 651 + ] + ], + "right_ranges": [ + [ + 637, + 649 + ] + ] + }, + { + "filename": "src/transformers/models/cohere_asr/modular_cohere_asr.py", + "left_ranges": [ + [ + 500, + 514 + ] + ], + "right_ranges": [ + [ + 500, + 512 + ] + ] + }, + { + "filename": "src/transformers/models/pp_formulanet/modeling_pp_formulanet.py", + "left_ranges": [ + [ + 1103, + 1116 + ] + ], + "right_ranges": [ + [ + 1103, + 1118 + ] + ] + }, + { + "filename": "src/transformers/models/pp_formulanet/modular_pp_formulanet.py", + "left_ranges": [ + [ + 488, + 501 + ] + ], + "right_ranges": [ + [ + 488, + 503 + ] + ] + } + ] + }, + { + "left_pr_number": 46903, + "right_pr_number": 47395, + "code_similarity": 0.083, + "size_similarity": 0.276, + "file_overlap": 0.0, + "area_overlap": 0.0, + "patch_similarity": 0.186, + "shared_filenames": [], + "shared_file_areas": [] + }, + { + "left_pr_number": 46958, + "right_pr_number": 47090, + "code_similarity": 0.397, + "size_similarity": 0.257, + "file_overlap": 0.273, + "area_overlap": 0.598, + "patch_similarity": 0.362, + "shared_filenames": [ + "src/transformers/models/cohere_asr/modeling_cohere_asr.py", + "src/transformers/models/cohere_asr/modular_cohere_asr.py", + "tests/models/cohere_asr/test_modeling_cohere_asr.py" + ], + "shared_file_areas": [ + { + "filename": "src/transformers/models/cohere_asr/modeling_cohere_asr.py", + "left_ranges": [ + [ + 22, + 28 + ], + [ + 638, + 645 + ] + ], + "right_ranges": [ + [ + 637, + 649 + ] + ] + }, + { + "filename": "src/transformers/models/cohere_asr/modular_cohere_asr.py", + "left_ranges": [ + [ + 16, + 22 + ], + [ + 501, + 508 + ] + ], + "right_ranges": [ + [ + 500, + 512 + ] + ] + }, + { + "filename": "tests/models/cohere_asr/test_modeling_cohere_asr.py", + "left_ranges": [ + [ + 15, + 21 + ], + [ + 180, + 221 + ] + ], + "right_ranges": [ + [ + 175, + 224 + ] + ] + } + ] + }, + { + "left_pr_number": 46958, + "right_pr_number": 47395, + "code_similarity": 0.201, + "size_similarity": 0.747, + "file_overlap": 0.0, + "area_overlap": 0.0, + "patch_similarity": 0.342, + "shared_filenames": [], + "shared_file_areas": [] + }, + { + "left_pr_number": 47090, + "right_pr_number": 47395, + "code_similarity": 0.078, + "size_similarity": 0.192, + "file_overlap": 0.0, + "area_overlap": 0.0, + "patch_similarity": 0.262, + "shared_filenames": [], + "shared_file_areas": [] + } + ] + }, + { + "cluster_id": "cluster-46509-6", + "summary": "Cluster of 2 issues and 4 PRs centered on issue #47333.", + "status": "open", + "confidence": 0.6, + "canonical_issue_number": 47333, + "canonical_pr_number": 47336, + "issue_numbers": [ + 47332, + 47333 + ], + "pr_numbers": [ + 46509, + 47143, + 47301, + 47336 + ], + "evidence_types": [ + "closing_reference", + "soft_similarity" + ], + "pr_comparisons": [ + { + "left_pr_number": 46509, + "right_pr_number": 47143, + "code_similarity": 0.254, + "size_similarity": 0.793, + "file_overlap": 0.071, + "area_overlap": 0.143, + "patch_similarity": 0.161, + "shared_filenames": [ + "src/transformers/models/auto/auto_mappings.py", + "src/transformers/models/auto/modeling_auto.py" + ], + "shared_file_areas": [ + { + "filename": "src/transformers/models/auto/auto_mappings.py", + "left_ranges": [ + [ + 412, + 418 + ], + [ + 1077, + 1083 + ], + [ + 1206, + 1212 + ] + ], + "right_ranges": [ + [ + 412, + 418 + ], + [ + 975, + 981 + ] + ] + }, + { + "filename": "src/transformers/models/auto/modeling_auto.py", + "left_ranges": [ + [ + 1097, + 1103 + ] + ], + "right_ranges": [ + [ + 348, + 354 + ], + [ + 1960, + 1966 + ] + ] + } + ] + }, + { + "left_pr_number": 46509, + "right_pr_number": 47301, + "code_similarity": 0.01, + "size_similarity": 0.013, + "file_overlap": 0.0, + "area_overlap": 0.0, + "patch_similarity": 0.047, + "shared_filenames": [], + "shared_file_areas": [] + }, + { + "left_pr_number": 46509, + "right_pr_number": 47336, + "code_similarity": 0.003, + "size_similarity": 0.008, + "file_overlap": 0.0, + "area_overlap": 0.0, + "patch_similarity": 0.007, + "shared_filenames": [], + "shared_file_areas": [] + }, + { + "left_pr_number": 47143, + "right_pr_number": 47301, + "code_similarity": 0.006, + "size_similarity": 0.01, + "file_overlap": 0.0, + "area_overlap": 0.0, + "patch_similarity": 0.026, + "shared_filenames": [], + "shared_file_areas": [] + }, + { + "left_pr_number": 47143, + "right_pr_number": 47336, + "code_similarity": 0.002, + "size_similarity": 0.006, + "file_overlap": 0.0, + "area_overlap": 0.0, + "patch_similarity": 0.004, + "shared_filenames": [], + "shared_file_areas": [] + }, + { + "left_pr_number": 47301, + "right_pr_number": 47336, + "code_similarity": 0.12, + "size_similarity": 0.6, + "file_overlap": 0.0, + "area_overlap": 0.0, + "patch_similarity": 0.0, + "shared_filenames": [], + "shared_file_areas": [] + } + ] + }, + { + "cluster_id": "cluster-41211-3", + "summary": "Cluster of 1 issues and 2 PRs centered on issue #41211.", + "status": "open", + "confidence": 0.55, + "canonical_issue_number": 41211, + "canonical_pr_number": 44339, + "issue_numbers": [ + 41211 + ], + "pr_numbers": [ + 41356, + 44339 + ], + "evidence_types": [ + "closing_reference" + ], + "pr_comparisons": [ + { + "left_pr_number": 41356, + "right_pr_number": 44339, + "code_similarity": 0.155, + "size_similarity": 0.094, + "file_overlap": 0.273, + "area_overlap": 0.078, + "patch_similarity": 0.183, + "shared_filenames": [ + "docs/source/en/model_doc/deimv2.md", + "src/transformers/models/deimv2/__init__.py", + "src/transformers/models/deimv2/configuration_deimv2.py", + "src/transformers/models/deimv2/modeling_deimv2.py", + "tests/models/deimv2/__init__.py", + "tests/models/deimv2/test_modeling_deimv2.py" + ], + "shared_file_areas": [ + { + "filename": "docs/source/en/model_doc/deimv2.md", + "left_ranges": [ + [ + 1, + 132 + ] + ], + "right_ranges": [ + [ + 1, + 65 + ] + ] + }, + { + "filename": "src/transformers/models/deimv2/__init__.py", + "left_ranges": [ + [ + 1, + 15 + ] + ], + "right_ranges": [ + [ + 1, + 29 + ] + ] + }, + { + "filename": "src/transformers/models/deimv2/configuration_deimv2.py", + "left_ranges": [ + [ + 1, + 74 + ] + ], + "right_ranges": [ + [ + 1, + 266 + ] + ] + }, + { + "filename": "tests/models/deimv2/test_modeling_deimv2.py", + "left_ranges": [ + [ + 1, + 15 + ] + ], + "right_ranges": [ + [ + 1, + 1734 + ] + ] + } + ] + } + ] + }, + { + "cluster_id": "cluster-47138-4", + "summary": "Cluster of 2 issues and 2 PRs centered on issue #47138.", + "status": "open", + "confidence": 0.6, + "canonical_issue_number": 47138, + "canonical_pr_number": 47303, + "issue_numbers": [ + 47138, + 47139 + ], + "pr_numbers": [ + 47140, + 47303 + ], + "evidence_types": [ + "closing_reference", + "soft_similarity" + ], + "pr_comparisons": [ + { + "left_pr_number": 47140, + "right_pr_number": 47303, + "code_similarity": 0.656, + "size_similarity": 0.65, + "file_overlap": 1.0, + "area_overlap": 0.447, + "patch_similarity": 0.466, + "shared_filenames": [ + "src/transformers/cli/chat.py", + "tests/cli/test_chat.py" + ], + "shared_file_areas": [ + { + "filename": "src/transformers/cli/chat.py", + "left_ranges": [ + [ + 20, + 26 + ], + [ + 103, + 122 + ], + [ + 346, + 356 + ], + [ + 383, + 389 + ], + [ + 473, + 479 + ] + ], + "right_ranges": [ + [ + 103, + 115 + ], + [ + 216, + 224 + ], + [ + 383, + 389 + ] + ] + }, + { + "filename": "tests/cli/test_chat.py", + "left_ranges": [ + [ + 15, + 28 + ], + [ + 51, + 105 + ] + ], + "right_ranges": [ + [ + 14, + 29 + ], + [ + 52, + 76 + ] + ] + } + ] + } + ] + }, + { + "cluster_id": "cluster-47175-3", + "summary": "Cluster of 3 related pull requests linked by soft_similarity.", + "status": "open", + "confidence": 0.5, + "canonical_issue_number": null, + "canonical_pr_number": 47298, + "issue_numbers": [], + "pr_numbers": [ + 47175, + 47294, + 47298 + ], + "evidence_types": [ + "soft_similarity" + ], + "pr_comparisons": [ + { + "left_pr_number": 47175, + "right_pr_number": 47294, + "code_similarity": 0.008, + "size_similarity": 0.032, + "file_overlap": 0.0, + "area_overlap": 0.0, + "patch_similarity": 0.011, + "shared_filenames": [], + "shared_file_areas": [] + }, + { + "left_pr_number": 47175, + "right_pr_number": 47298, + "code_similarity": 0.019, + "size_similarity": 0.043, + "file_overlap": 0.0, + "area_overlap": 0.0, + "patch_similarity": 0.067, + "shared_filenames": [], + "shared_file_areas": [] + }, + { + "left_pr_number": 47294, + "right_pr_number": 47298, + "code_similarity": 0.153, + "size_similarity": 0.734, + "file_overlap": 0.0, + "area_overlap": 0.0, + "patch_similarity": 0.044, + "shared_filenames": [], + "shared_file_areas": [] + } + ] + }, + { + "cluster_id": "cluster-43240-3", + "summary": "Cluster of 1 issues and 2 PRs centered on issue #43240.", + "status": "open", + "confidence": 0.75, + "canonical_issue_number": 43240, + "canonical_pr_number": 43251, + "issue_numbers": [ + 43240 + ], + "pr_numbers": [ + 43251, + 43254 + ], + "evidence_types": [ + "closing_reference", + "shared_issue_target" + ], + "pr_comparisons": [ + { + "left_pr_number": 43251, + "right_pr_number": 43254, + "code_similarity": 0.794, + "size_similarity": 0.64, + "file_overlap": 1.0, + "area_overlap": 0.667, + "patch_similarity": 0.882, + "shared_filenames": [ + "src/transformers/loss/loss_utils.py" + ], + "shared_file_areas": [ + { + "filename": "src/transformers/loss/loss_utils.py", + "left_ranges": [ + [ + 30, + 50 + ] + ], + "right_ranges": [ + [ + 30, + 43 + ] + ] + } + ] + } + ] + }, + { + "cluster_id": "cluster-45647-3", + "summary": "Cluster of 1 issues and 2 PRs centered on issue #45647.", + "status": "open", + "confidence": 0.75, + "canonical_issue_number": 45647, + "canonical_pr_number": 45738, + "issue_numbers": [ + 45647 + ], + "pr_numbers": [ + 45737, + 45738 + ], + "evidence_types": [ + "closing_reference", + "shared_issue_target" + ], + "pr_comparisons": [ + { + "left_pr_number": 45737, + "right_pr_number": 45738, + "code_similarity": 0.526, + "size_similarity": 0.11, + "file_overlap": 0.5, + "area_overlap": 1.0, + "patch_similarity": 0.024, + "shared_filenames": [ + "src/transformers/models/musicgen_melody/modeling_musicgen_melody.py" + ], + "shared_file_areas": [ + { + "filename": "src/transformers/models/musicgen_melody/modeling_musicgen_melody.py", + "left_ranges": [ + [ + 491, + 497 + ] + ], + "right_ranges": [ + [ + 491, + 497 + ] + ] + } + ] + } + ] + }, + { + "cluster_id": "cluster-46159-3", + "summary": "Cluster of 1 issues and 2 PRs centered on issue #46159.", + "status": "open", + "confidence": 0.55, + "canonical_issue_number": 46159, + "canonical_pr_number": 46174, + "issue_numbers": [ + 46159 + ], + "pr_numbers": [ + 46174, + 46780 + ], + "evidence_types": [ + "closing_reference" + ], + "pr_comparisons": [ + { + "left_pr_number": 46174, + "right_pr_number": 46780, + "code_similarity": 0.841, + "size_similarity": 0.87, + "file_overlap": 1.0, + "area_overlap": 0.767, + "patch_similarity": 0.657, + "shared_filenames": [ + "src/transformers/models/aimv2/modeling_aimv2.py", + "src/transformers/models/d_fine/modeling_d_fine.py", + "src/transformers/models/deimv2/modeling_deimv2.py", + "src/transformers/models/pp_doclayout_v2/modeling_pp_doclayout_v2.py", + "src/transformers/models/pp_doclayout_v3/modeling_pp_doclayout_v3.py", + "src/transformers/models/rt_detr/modeling_rt_detr.py", + "src/transformers/models/rt_detr_v2/modeling_rt_detr_v2.py", + "src/transformers/models/vit_mae/modeling_vit_mae.py", + "src/transformers/models/vit_mae/modular_vit_mae.py" + ], + "shared_file_areas": [ + { + "filename": "src/transformers/models/aimv2/modeling_aimv2.py", + "left_ranges": [ + [ + 134, + 141 + ], + [ + 144, + 154 + ], + [ + 157, + 165 + ] + ], + "right_ranges": [ + [ + 134, + 140 + ], + [ + 143, + 153 + ], + [ + 156, + 162 + ] + ] + }, + { + "filename": "src/transformers/models/d_fine/modeling_d_fine.py", + "left_ranges": [ + [ + 631, + 638 + ], + [ + 641, + 651 + ], + [ + 654, + 662 + ] + ], + "right_ranges": [ + [ + 631, + 637 + ], + [ + 640, + 650 + ], + [ + 653, + 659 + ] + ] + }, + { + "filename": "src/transformers/models/deimv2/modeling_deimv2.py", + "left_ranges": [ + [ + 703, + 710 + ], + [ + 713, + 723 + ], + [ + 726, + 734 + ] + ], + "right_ranges": [ + [ + 703, + 709 + ], + [ + 712, + 722 + ], + [ + 725, + 731 + ] + ] + }, + { + "filename": "src/transformers/models/pp_doclayout_v2/modeling_pp_doclayout_v2.py", + "left_ranges": [ + [ + 1551, + 1558 + ], + [ + 1561, + 1571 + ], + [ + 1574, + 1582 + ] + ], + "right_ranges": [ + [ + 1551, + 1557 + ], + [ + 1560, + 1570 + ], + [ + 1573, + 1579 + ] + ] + }, + { + "filename": "src/transformers/models/pp_doclayout_v3/modeling_pp_doclayout_v3.py", + "left_ranges": [ + [ + 804, + 811 + ], + [ + 814, + 824 + ], + [ + 827, + 835 + ] + ], + "right_ranges": [ + [ + 804, + 810 + ], + [ + 813, + 823 + ], + [ + 826, + 832 + ] + ] + }, + { + "filename": "src/transformers/models/rt_detr/modeling_rt_detr.py", + "left_ranges": [ + [ + 847, + 854 + ], + [ + 857, + 867 + ], + [ + 870, + 878 + ] + ], + "right_ranges": [ + [ + 847, + 853 + ], + [ + 856, + 866 + ], + [ + 869, + 875 + ] + ] + }, + { + "filename": "src/transformers/models/rt_detr_v2/modeling_rt_detr_v2.py", + "left_ranges": [ + [ + 976, + 983 + ], + [ + 986, + 996 + ], + [ + 999, + 1007 + ] + ], + "right_ranges": [ + [ + 976, + 982 + ], + [ + 985, + 995 + ], + [ + 998, + 1004 + ] + ] + }, + { + "filename": "src/transformers/models/vit_mae/modeling_vit_mae.py", + "left_ranges": [ + [ + 152, + 159 + ], + [ + 162, + 172 + ], + [ + 175, + 183 + ] + ], + "right_ranges": [ + [ + 152, + 158 + ], + [ + 161, + 171 + ], + [ + 174, + 180 + ] + ] + }, + { + "filename": "src/transformers/models/vit_mae/modular_vit_mae.py", + "left_ranges": [ + [ + 62, + 69 + ], + [ + 72, + 82 + ], + [ + 85, + 93 + ] + ], + "right_ranges": [ + [ + 62, + 68 + ], + [ + 71, + 81 + ], + [ + 84, + 90 + ] + ] + } + ] + } + ] + }, + { + "cluster_id": "cluster-46869-3", + "summary": "Cluster of 1 issues and 2 PRs centered on issue #46869.", + "status": "open", + "confidence": 0.75, + "canonical_issue_number": 46869, + "canonical_pr_number": 46886, + "issue_numbers": [ + 46869 + ], + "pr_numbers": [ + 46886, + 46922 + ], + "evidence_types": [ + "closing_reference", + "shared_issue_target" + ], + "pr_comparisons": [ + { + "left_pr_number": 46886, + "right_pr_number": 46922, + "code_similarity": 0.711, + "size_similarity": 0.857, + "file_overlap": 1.0, + "area_overlap": 0.5, + "patch_similarity": 0.429, + "shared_filenames": [ + "src/transformers/models/yolos/modeling_yolos.py" + ], + "shared_file_areas": [ + { + "filename": "src/transformers/models/yolos/modeling_yolos.py", + "left_ranges": [ + [ + 446, + 452 + ] + ], + "right_ranges": [ + [ + 434, + 440 + ], + [ + 446, + 452 + ] + ] + } + ] + } + ] + }, + { + "cluster_id": "cluster-45561-3", + "summary": "Cluster of 1 issues and 2 PRs centered on issue #45561.", + "status": "open", + "confidence": 0.8, + "canonical_issue_number": 45561, + "canonical_pr_number": 45645, + "issue_numbers": [ + 45561 + ], + "pr_numbers": [ + 45639, + 45645 + ], + "evidence_types": [ + "closing_reference", + "shared_issue_target", + "soft_similarity" + ], + "pr_comparisons": [ + { + "left_pr_number": 45639, + "right_pr_number": 45645, + "code_similarity": 0.721, + "size_similarity": 0.973, + "file_overlap": 0.75, + "area_overlap": 0.563, + "patch_similarity": 0.696, + "shared_filenames": [ + "src/transformers/testing_utils.py", + "tests/utils/test_testing_utils.py", + "utils/notification_service.py" + ], + "shared_file_areas": [ + { + "filename": "src/transformers/testing_utils.py", + "left_ranges": [ + [ + 3611, + 3618 + ], + [ + 3835, + 3869 + ] + ], + "right_ranges": [ + [ + 3525, + 3558 + ], + [ + 3782, + 3788 + ] + ] + }, + { + "filename": "tests/utils/test_testing_utils.py", + "left_ranges": [ + [ + 1, + 144 + ] + ], + "right_ranges": [ + [ + 1, + 114 + ] + ] + }, + { + "filename": "utils/notification_service.py", + "left_ranges": [ + [ + 935, + 967 + ] + ], + "right_ranges": [ + [ + 935, + 967 + ] + ] + } + ] + } + ] + }, + { + "cluster_id": "cluster-46305-3", + "summary": "Cluster of 1 issues and 2 PRs centered on issue #46305.", + "status": "open", + "confidence": 0.8, + "canonical_issue_number": 46305, + "canonical_pr_number": 47162, + "issue_numbers": [ + 46305 + ], + "pr_numbers": [ + 46310, + 47162 + ], + "evidence_types": [ + "closing_reference", + "shared_issue_target", + "soft_similarity" + ], + "pr_comparisons": [ + { + "left_pr_number": 46310, + "right_pr_number": 47162, + "code_similarity": 0.577, + "size_similarity": 0.338, + "file_overlap": 0.5, + "area_overlap": 1.0, + "patch_similarity": 0.062, + "shared_filenames": [ + "src/transformers/models/moonshine_streaming/modeling_moonshine_streaming.py", + "src/transformers/models/moonshine_streaming/modular_moonshine_streaming.py" + ], + "shared_file_areas": [ + { + "filename": "src/transformers/models/moonshine_streaming/modeling_moonshine_streaming.py", + "left_ranges": [ + [ + 356, + 362 + ] + ], + "right_ranges": [ + [ + 356, + 362 + ] + ] + }, + { + "filename": "src/transformers/models/moonshine_streaming/modular_moonshine_streaming.py", + "left_ranges": [ + [ + 269, + 275 + ] + ], + "right_ranges": [ + [ + 269, + 275 + ] + ] + } + ] + } + ] + }, + { + "cluster_id": "cluster-45081-3", + "summary": "Cluster of 1 issues and 2 PRs centered on issue #45081.", + "status": "open", + "confidence": 0.8, + "canonical_issue_number": 45081, + "canonical_pr_number": 45317, + "issue_numbers": [ + 45081 + ], + "pr_numbers": [ + 45086, + 45317 + ], + "evidence_types": [ + "closing_reference", + "shared_issue_target", + "soft_similarity" + ], + "pr_comparisons": [ + { + "left_pr_number": 45086, + "right_pr_number": 45317, + "code_similarity": 0.589, + "size_similarity": 0.34, + "file_overlap": 0.5, + "area_overlap": 1.0, + "patch_similarity": 0.136, + "shared_filenames": [ + "src/transformers/tokenization_utils_tokenizers.py" + ], + "shared_file_areas": [ + { + "filename": "src/transformers/tokenization_utils_tokenizers.py", + "left_ranges": [ + [ + 1360, + 1370 + ], + [ + 1374, + 1380 + ] + ], + "right_ranges": [ + [ + 1360, + 1370 + ], + [ + 1374, + 1380 + ] + ] + } + ] + } + ] + }, + { + "cluster_id": "cluster-43698-3", + "summary": "Cluster of 1 issues and 2 PRs centered on issue #43698.", + "status": "open", + "confidence": 0.75, + "canonical_issue_number": 43698, + "canonical_pr_number": 43779, + "issue_numbers": [ + 43698 + ], + "pr_numbers": [ + 43779, + 43816 + ], + "evidence_types": [ + "closing_reference", + "shared_issue_target" + ], + "pr_comparisons": [ + { + "left_pr_number": 43779, + "right_pr_number": 43816, + "code_similarity": 0.418, + "size_similarity": 0.538, + "file_overlap": 1.0, + "area_overlap": 0.02, + "patch_similarity": 0.02, + "shared_filenames": [ + "src/transformers/integrations/integration_utils.py" + ], + "shared_file_areas": [ + { + "filename": "src/transformers/integrations/integration_utils.py", + "left_ranges": [ + [ + 2227, + 2241 + ], + [ + 2303, + 2309 + ] + ], + "right_ranges": [ + [ + 2278, + 2291 + ], + [ + 2309, + 2322 + ] + ] + } + ] + } + ] + }, + { + "cluster_id": "cluster-43656-2", + "summary": "Cluster of 2 related pull requests linked by soft_similarity.", + "status": "open", + "confidence": 0.5, + "canonical_issue_number": null, + "canonical_pr_number": 43656, + "issue_numbers": [], + "pr_numbers": [ + 43656, + 43842 + ], + "evidence_types": [ + "soft_similarity" + ], + "pr_comparisons": [ + { + "left_pr_number": 43656, + "right_pr_number": 43842, + "code_similarity": 0.405, + "size_similarity": 0.5, + "file_overlap": 1.0, + "area_overlap": 0.0, + "patch_similarity": 0.036, + "shared_filenames": [ + "src/transformers/cli/serve.py" + ], + "shared_file_areas": [ + { + "filename": "src/transformers/cli/serve.py", + "left_ranges": [ + [ + 11, + 18 + ], + [ + 30, + 36 + ], + [ + 315, + 323 + ], + [ + 665, + 671 + ], + [ + 931, + 937 + ], + [ + 1843, + 1849 + ], + [ + 1868, + 1874 + ] + ], + "right_ranges": [ + [ + 54, + 61 + ], + [ + 587, + 593 + ] + ] + } + ] + } + ] + }, + { + "cluster_id": "cluster-47123-2", + "summary": "Cluster of 2 related pull requests linked by soft_similarity.", + "status": "open", + "confidence": 0.5, + "canonical_issue_number": null, + "canonical_pr_number": 47123, + "issue_numbers": [], + "pr_numbers": [ + 47123, + 47280 + ], + "evidence_types": [ + "soft_similarity" + ], + "pr_comparisons": [ + { + "left_pr_number": 47123, + "right_pr_number": 47280, + "code_similarity": 0.547, + "size_similarity": 0.522, + "file_overlap": 1.0, + "area_overlap": 0.368, + "patch_similarity": 0.093, + "shared_filenames": [ + "tests/models/minimax_m3_vl/test_modeling_minimax_m3_vl.py" + ], + "shared_file_areas": [ + { + "filename": "tests/models/minimax_m3_vl/test_modeling_minimax_m3_vl.py", + "left_ranges": [ + [ + 14, + 20 + ], + [ + 487, + 493 + ] + ], + "right_ranges": [ + [ + 14, + 21 + ], + [ + 481, + 499 + ], + [ + 520, + 530 + ] + ] + } + ] + } + ] + }, + { + "cluster_id": "cluster-45157-2", + "summary": "Cluster of 2 related pull requests linked by soft_similarity.", + "status": "open", + "confidence": 0.5, + "canonical_issue_number": null, + "canonical_pr_number": 47387, + "issue_numbers": [], + "pr_numbers": [ + 45157, + 47387 + ], + "evidence_types": [ + "soft_similarity" + ], + "pr_comparisons": [ + { + "left_pr_number": 45157, + "right_pr_number": 47387, + "code_similarity": 0.176, + "size_similarity": 0.842, + "file_overlap": 0.0, + "area_overlap": 0.0, + "patch_similarity": 0.053, + "shared_filenames": [], + "shared_file_areas": [] + } + ] + }, + { + "cluster_id": "cluster-47282-2", + "summary": "Cluster of 2 related pull requests linked by soft_similarity.", + "status": "open", + "confidence": 0.5, + "canonical_issue_number": null, + "canonical_pr_number": 47380, + "issue_numbers": [], + "pr_numbers": [ + 47282, + 47380 + ], + "evidence_types": [ + "soft_similarity" + ], + "pr_comparisons": [ + { + "left_pr_number": 47282, + "right_pr_number": 47380, + "code_similarity": 0.476, + "size_similarity": 0.5, + "file_overlap": 1.0, + "area_overlap": 0.194, + "patch_similarity": 0.056, + "shared_filenames": [ + "tests/models/mamba2/test_modeling_mamba2.py" + ], + "shared_file_areas": [ + { + "filename": "tests/models/mamba2/test_modeling_mamba2.py", + "left_ranges": [ + [ + 420, + 438 + ], + [ + 450, + 474 + ], + [ + 482, + 524 + ] + ], + "right_ranges": [ + [ + 18, + 24 + ], + [ + 406, + 419 + ], + [ + 426, + 432 + ], + [ + 457, + 463 + ], + [ + 487, + 493 + ] + ] + } + ] + } + ] + }, + { + "cluster_id": "cluster-47163-2", + "summary": "Cluster of 2 related pull requests linked by soft_similarity.", + "status": "open", + "confidence": 0.5, + "canonical_issue_number": null, + "canonical_pr_number": 47163, + "issue_numbers": [], + "pr_numbers": [ + 47163, + 47210 + ], + "evidence_types": [ + "soft_similarity" + ], + "pr_comparisons": [ + { + "left_pr_number": 47163, + "right_pr_number": 47210, + "code_similarity": 0.021, + "size_similarity": 0.076, + "file_overlap": 0.0, + "area_overlap": 0.0, + "patch_similarity": 0.036, + "shared_filenames": [], + "shared_file_areas": [] + } + ] + } + ], + "duplicate_issues": [ + { + "cluster_id": "cluster-46509-6", + "canonical_issue_number": 47333, + "duplicate_issue_numbers": [ + 47332 + ], + "reason": "Issues in cluster-46509-6 are treated as duplicates because they share closing_reference, soft_similarity evidence." + }, + { + "cluster_id": "cluster-46894-8", + "canonical_issue_number": 46894, + "duplicate_issue_numbers": [ + 46901, + 47394 + ], + "reason": "Issues in cluster-46894-8 are treated as duplicates because they share closing_reference, shared_issue_target, soft_similarity evidence." + }, + { + "cluster_id": "cluster-47138-4", + "canonical_issue_number": 47138, + "duplicate_issue_numbers": [ + 47139 + ], + "reason": "Issues in cluster-47138-4 are treated as duplicates because they share closing_reference, soft_similarity evidence." + } + ], + "duplicate_prs": [ + { + "cluster_id": "cluster-41211-3", + "canonical_pr_number": 44339, + "duplicate_pr_numbers": [ + 41356 + ], + "target_issue_number": 41211, + "reason": "PRs in cluster-41211-3 are treated as duplicates because they converge on issue #41211 with closing_reference evidence." + }, + { + "cluster_id": "cluster-43240-3", + "canonical_pr_number": 43251, + "duplicate_pr_numbers": [ + 43254 + ], + "target_issue_number": 43240, + "reason": "PRs in cluster-43240-3 are treated as duplicates because they converge on issue #43240 with closing_reference, shared_issue_target evidence." + }, + { + "cluster_id": "cluster-43656-2", + "canonical_pr_number": 43656, + "duplicate_pr_numbers": [ + 43842 + ], + "target_issue_number": null, + "reason": "PRs in cluster-43656-2 are treated as duplicates because they share soft_similarity evidence." + }, + { + "cluster_id": "cluster-43698-3", + "canonical_pr_number": 43779, + "duplicate_pr_numbers": [ + 43816 + ], + "target_issue_number": 43698, + "reason": "PRs in cluster-43698-3 are treated as duplicates because they converge on issue #43698 with closing_reference, shared_issue_target evidence." + }, + { + "cluster_id": "cluster-43979-11", + "canonical_pr_number": 44007, + "duplicate_pr_numbers": [ + 43996, + 44013, + 44044, + 44066, + 44072, + 44085, + 44129, + 44154, + 44722 + ], + "target_issue_number": 43979, + "reason": "PRs in cluster-43979-11 are treated as duplicates because they converge on issue #43979 with closing_reference, shared_issue_target, soft_similarity evidence." + }, + { + "cluster_id": "cluster-45081-3", + "canonical_pr_number": 45317, + "duplicate_pr_numbers": [ + 45086 + ], + "target_issue_number": 45081, + "reason": "PRs in cluster-45081-3 are treated as duplicates because they converge on issue #45081 with closing_reference, shared_issue_target, soft_similarity evidence." + }, + { + "cluster_id": "cluster-45157-2", + "canonical_pr_number": 47387, + "duplicate_pr_numbers": [ + 45157 + ], + "target_issue_number": null, + "reason": "PRs in cluster-45157-2 are treated as duplicates because they share soft_similarity evidence." + }, + { + "cluster_id": "cluster-45561-3", + "canonical_pr_number": 45645, + "duplicate_pr_numbers": [ + 45639 + ], + "target_issue_number": 45561, + "reason": "PRs in cluster-45561-3 are treated as duplicates because they converge on issue #45561 with closing_reference, shared_issue_target, soft_similarity evidence." + }, + { + "cluster_id": "cluster-45647-3", + "canonical_pr_number": 45738, + "duplicate_pr_numbers": [ + 45737 + ], + "target_issue_number": 45647, + "reason": "PRs in cluster-45647-3 are treated as duplicates because they converge on issue #45647 with closing_reference, shared_issue_target evidence." + }, + { + "cluster_id": "cluster-46159-3", + "canonical_pr_number": 46174, + "duplicate_pr_numbers": [ + 46780 + ], + "target_issue_number": 46159, + "reason": "PRs in cluster-46159-3 are treated as duplicates because they converge on issue #46159 with closing_reference evidence." + }, + { + "cluster_id": "cluster-46305-3", + "canonical_pr_number": 47162, + "duplicate_pr_numbers": [ + 46310 + ], + "target_issue_number": 46305, + "reason": "PRs in cluster-46305-3 are treated as duplicates because they converge on issue #46305 with closing_reference, shared_issue_target, soft_similarity evidence." + }, + { + "cluster_id": "cluster-46509-6", + "canonical_pr_number": 47336, + "duplicate_pr_numbers": [ + 46509, + 47143, + 47301 + ], + "target_issue_number": 47333, + "reason": "PRs in cluster-46509-6 are treated as duplicates because they converge on issue #47333 with closing_reference, soft_similarity evidence." + }, + { + "cluster_id": "cluster-46869-3", + "canonical_pr_number": 46886, + "duplicate_pr_numbers": [ + 46922 + ], + "target_issue_number": 46869, + "reason": "PRs in cluster-46869-3 are treated as duplicates because they converge on issue #46869 with closing_reference, shared_issue_target evidence." + }, + { + "cluster_id": "cluster-46894-8", + "canonical_pr_number": 46903, + "duplicate_pr_numbers": [ + 46895, + 46958, + 47090, + 47395 + ], + "target_issue_number": 46894, + "reason": "PRs in cluster-46894-8 are treated as duplicates because they converge on issue #46894 with closing_reference, shared_issue_target, soft_similarity evidence." + }, + { + "cluster_id": "cluster-47123-2", + "canonical_pr_number": 47123, + "duplicate_pr_numbers": [ + 47280 + ], + "target_issue_number": null, + "reason": "PRs in cluster-47123-2 are treated as duplicates because they share soft_similarity evidence." + }, + { + "cluster_id": "cluster-47138-4", + "canonical_pr_number": 47303, + "duplicate_pr_numbers": [ + 47140 + ], + "target_issue_number": 47138, + "reason": "PRs in cluster-47138-4 are treated as duplicates because they converge on issue #47138 with closing_reference, soft_similarity evidence." + }, + { + "cluster_id": "cluster-47163-2", + "canonical_pr_number": 47163, + "duplicate_pr_numbers": [ + 47210 + ], + "target_issue_number": null, + "reason": "PRs in cluster-47163-2 are treated as duplicates because they share soft_similarity evidence." + }, + { + "cluster_id": "cluster-47175-3", + "canonical_pr_number": 47298, + "duplicate_pr_numbers": [ + 47175, + 47294 + ], + "target_issue_number": null, + "reason": "PRs in cluster-47175-3 are treated as duplicates because they share soft_similarity evidence." + }, + { + "cluster_id": "cluster-47282-2", + "canonical_pr_number": 47380, + "duplicate_pr_numbers": [ + 47282 + ], + "target_issue_number": null, + "reason": "PRs in cluster-47282-2 are treated as duplicates because they share soft_similarity evidence." + } + ], + "best_issue": { + "cluster_id": "cluster-46894-8", + "issue_number": 46894, + "reason": "Issue #46894 is the strongest global issue candidate because it is open, belongs to a cluster with 8 artifacts, and carries 1 discussion comments plus 10 inbound references.", + "score": 192.0 + }, + "best_pr": { + "cluster_id": "cluster-43979-11", + "pr_number": 44007, + "reason": "PR #44007 is the strongest global PR candidate because it is open, belongs to a cluster with 11 artifacts, links to 1 issues, and carries 2 review/discussion events.", + "score": 238.886 + } +} diff --git a/snapshots/20260718T120032Z/analysis-runs/hybrid-model-20260718t120032z/analysis-report-hybrid.llm-reviews.json b/snapshots/20260718T120032Z/analysis-runs/hybrid-model-20260718t120032z/analysis-report-hybrid.llm-reviews.json new file mode 100644 index 0000000000000000000000000000000000000000..e65d32944397fadfed36c438c2455f8a8a7977af --- /dev/null +++ b/snapshots/20260718T120032Z/analysis-runs/hybrid-model-20260718t120032z/analysis-report-hybrid.llm-reviews.json @@ -0,0 +1,28176 @@ +{ + "schema_version": "1.0", + "repo": "huggingface/transformers", + "snapshot_id": "20260718T120032Z", + "generated_at": "2026-07-18T12:14:56Z", + "model": "gpt-5.4-mini?service_tier=flex", + "reviews": [ + { + "label": "issue", + "component_index": 1, + "component_count": 28, + "review_unit_index": 1, + "review_unit_count": 65, + "status": "reviewed", + "reason": null, + "source": "llm", + "cache_hit": false, + "model": "gpt-5.4-mini?service_tier=flex", + "cluster_id": "cluster-10000-530", + "nodes": [ + "issue:24643", + "issue:30277", + "issue:34634", + "issue:34928", + "issue:35308", + "issue:35545", + "issue:36010", + "issue:36331", + "issue:39290", + "issue:41720", + "issue:41762", + "issue:42915", + "issue:43039", + "issue:43425", + "issue:43541", + "issue:43596", + "issue:43638", + "issue:43716", + "issue:43828", + "issue:43975", + "issue:44292", + "issue:44322", + "issue:44457", + "issue:44509", + "issue:44512", + "issue:44560", + "issue:44805", + "issue:44841", + "issue:44918", + "issue:45081", + "issue:45161", + "issue:45237", + "issue:45464", + "issue:45507", + "issue:45685", + "issue:45800", + "issue:45823", + "issue:45859", + "issue:45874", + "issue:45901", + "issue:45902", + "issue:46015", + "issue:46165", + "issue:46294", + "issue:46468", + "issue:46515", + "issue:46661", + "issue:46869" + ], + "soft_pairs": [ + "issue:44322|issue:45464", + "issue:45237|issue:45685", + "issue:43716|issue:45237", + "issue:43828|issue:45237", + "issue:45237|issue:46468", + "issue:45237|issue:45507", + "issue:30277|issue:35545", + "issue:35545|issue:42915", + "issue:35545|issue:41720", + "issue:45901|issue:46015", + "issue:35545|issue:36010", + "issue:44509|issue:44512", + "issue:44918|issue:46165", + "issue:46165|issue:46869", + "issue:44805|issue:45823", + "issue:43638|issue:44805", + "issue:43638|issue:46165", + "issue:43596|issue:44805", + "issue:44805|issue:46294", + "issue:46165|issue:46468", + "issue:44805|issue:45161", + "issue:44560|issue:46165", + "issue:34634|issue:35545", + "issue:30277|issue:43638", + "issue:43596|issue:43638", + "issue:43039|issue:45161", + "issue:43975|issue:44457", + "issue:44292|issue:44457", + "issue:34928|issue:35308", + "issue:44841|issue:46661", + "issue:24643|issue:30277", + "issue:43716|issue:46869", + "issue:34928|issue:44805", + "issue:45685|issue:46869", + "issue:34928|issue:36331", + "issue:45859|issue:45902", + "issue:43638|issue:45161", + "issue:43596|issue:45161", + "issue:43716|issue:45507", + "issue:44292|issue:45237", + "issue:45874|issue:46294", + "issue:39290|issue:45081", + "issue:45800|issue:46515", + "issue:45823|issue:45874", + "issue:34928|issue:41720", + "issue:43425|issue:44292", + "issue:43828|issue:45161", + "issue:43541|issue:43716", + "issue:43828|issue:44292", + "issue:43716|issue:45685", + "issue:45507|issue:45685", + "issue:44292|issue:44918", + "issue:43716|issue:44292", + "issue:44918|issue:45161", + "issue:43039|issue:44918", + "issue:41762|issue:43638" + ], + "prepared_review_unit_hash": "9a400136cc34d904348faa448fd0abd472fb13ee6274a452a4b0d81aad08433e", + "component_budget": { + "node_count": 530, + "item_count": 530, + "soft_pair_count": 1954, + "serialized_chars": 1013004, + "estimated_input_tokens": 253251, + "estimated_eval_tokens": 506758 + }, + "budget": { + "node_count": 48, + "item_count": 48, + "soft_pair_count": 56, + "serialized_chars": 44518, + "estimated_input_tokens": 11130, + "estimated_eval_tokens": 22516 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": true, + "analyst_result": { + "summary": "Mostly heterogeneous cluster. The only clear duplicate subcluster is the DeepSpeed ZeRO-3 BERT IndexError pair (43596/43638); the other issues span different models, hardware paths, docs, and export/inference bugs.", + "confidence": 0.94, + "canonical_issue_reason": "issue:43596 is the best canonical issue because it is one of the only true duplicates in the set and gives a concrete, reproducible DeepSpeed ZeRO-3/BERT IndexError that matches 43638 almost exactly.", + "canonical_pr_reason": null, + "best_issue_reason": "issue:43596 is the best global issue choice because it represents the only confirmed duplicate pair and is the clearest, most self-contained report for that bug.", + "best_pr_reason": null, + "soft_edge_verdicts": [ + { + "left": "issue:44322", + "right": "issue:45464", + "accept": false, + "reason": "Different failures: Qwen3_5Config missing num_attention_heads versus a Qwen3.5-0.8B streaming chat/completions error." + }, + { + "left": "issue:45237", + "right": "issue:45685", + "accept": false, + "reason": "Different model/platform bugs: GPT-OSS on AMD GPUs versus an MPS histogram kernel limitation for Int." + }, + { + "left": "issue:43716", + "right": "issue:45237", + "accept": false, + "reason": "Different root causes and subsystems: image preprocessor/model dtype mismatch versus AMD GPU device issues." + }, + { + "left": "issue:43828", + "right": "issue:45237", + "accept": false, + "reason": "Autocast dtype mismatch in Phi-tiny-MoE is unrelated to GPT-OSS failing on AMD GPUs." + }, + { + "left": "issue:45237", + "right": "issue:46468", + "accept": false, + "reason": "Different models and code paths: GPT-OSS AMD GPU failure versus a broken Nemotron example." + }, + { + "left": "issue:45237", + "right": "issue:45507", + "accept": false, + "reason": "Different bugs: AMD GPU compatibility failure versus GraniteMoEHybrid calling an invalid method." + }, + { + "left": "issue:30277", + "right": "issue:35545", + "accept": false, + "reason": "DeepSpeed collective mismatch and ModernBERT ONNX export are different failures." + }, + { + "left": "issue:35545", + "right": "issue:42915", + "accept": false, + "reason": "ONNX export error and Qwen3Moe FineGrainedFP8 failure are unrelated." + }, + { + "left": "issue:35545", + "right": "issue:41720", + "accept": false, + "reason": "Different code paths: ONNX export failure versus a CUDA assert during auto device mapping." + }, + { + "left": "issue:45901", + "right": "issue:46015", + "accept": false, + "reason": "Task crash and outdated RAG docs are not the same underlying issue." + }, + { + "left": "issue:35545", + "right": "issue:36010", + "accept": false, + "reason": "ModernBERT ONNX export is unrelated to the GenerationMixin import error." + }, + { + "left": "issue:44509", + "right": "issue:44512", + "accept": false, + "reason": "Both are v5 docs issues, but they cover different removed documentation topics and are not the same change." + }, + { + "left": "issue:44918", + "right": "issue:46165", + "accept": false, + "reason": "TRL SFT embedding unpacking and resume-training with trust_remote_code are different bugs." + }, + { + "left": "issue:46165", + "right": "issue:46869", + "accept": false, + "reason": "Resume-training/trust_remote_code bug is unrelated to YOLO multi-GPU device_map auto failure." + }, + { + "left": "issue:44805", + "right": "issue:45823", + "accept": false, + "reason": "Mask/tensor shape mismatch is a different root cause from Gemma4 PLE device mismatch." + }, + { + "left": "issue:43638", + "right": "issue:44805", + "accept": false, + "reason": "Both involve IndexError, but ZeRO-3/BERT init and mask-shape indexing are different bugs." + }, + { + "left": "issue:43596", + "right": "issue:44805", + "accept": false, + "reason": "Different error shapes and code paths; not the same underlying bug." + }, + { + "left": "issue:44805", + "right": "issue:46294", + "accept": false, + "reason": "Tensor mask mismatch is unrelated to Qwen3.5 video token/feature mismatch." + }, + { + "left": "issue:44805", + "right": "issue:45161", + "accept": false, + "reason": "Different failures: mask/tensor indexing versus GPT-OSS tensor-parallel behavior." + }, + { + "left": "issue:44560", + "right": "issue:46165", + "accept": false, + "reason": "Qwen3-vl video StopIteration is unrelated to resume-training with local models and trust_remote_code." + }, + { + "left": "issue:34634", + "right": "issue:35545", + "accept": false, + "reason": "BarkProcessor voice_preset bug is unrelated to ModernBERT ONNX export." + }, + { + "left": "issue:30277", + "right": "issue:43638", + "accept": false, + "reason": "Both involve DeepSpeed, but the concrete failures differ: collective mismatch versus zero-sized tensor IndexError." + }, + { + "left": "issue:43596", + "right": "issue:43638", + "accept": true, + "reason": "Near-identical DeepSpeed ZeRO-3 BERT reports with the same out-of-bounds index-0 failure during initialization." + }, + { + "left": "issue:43039", + "right": "issue:45161", + "accept": false, + "reason": "Liger cross_entropy being called is a different issue from GPT-OSS tensor-parallel failure." + }, + { + "left": "issue:43975", + "right": "issue:44457", + "accept": false, + "reason": "Incorrect detokenization and LoRA merge/reload output mismatch are different bugs." + }, + { + "left": "issue:44292", + "right": "issue:44457", + "accept": false, + "reason": "Qwen-3 NVFP4 runtime error is unrelated to LoRA save/load mismatch." + }, + { + "left": "issue:34928", + "right": "issue:35308", + "accept": false, + "reason": "Activation-checkpoint/FSDP recomputation mismatch and IterableDataset length mismatch are different distributed-training bugs." + }, + { + "left": "issue:44841", + "right": "issue:46661", + "accept": false, + "reason": "Processor failure for Voxtral is unrelated to the remote-code config_class AttributeError." + }, + { + "left": "issue:24643", + "right": "issue:30277", + "accept": false, + "reason": "Different DeepSpeed failures: 2-D weight runtime error versus collective mismatch." + }, + { + "left": "issue:43716", + "right": "issue:46869", + "accept": false, + "reason": "Different models and causes: image preprocessing dtype mismatch versus YOLO multi-GPU device-map failure." + }, + { + "left": "issue:45685", + "right": "issue:46869", + "accept": false, + "reason": "MPS histogram kernel limitation is unrelated to YOLO device-map auto failure." + }, + { + "left": "issue:43425", + "right": "issue:44292", + "accept": false, + "reason": "Torch version incompatibility is not the same as a Qwen NVFP4 model runtime error." + }, + { + "left": "issue:43828", + "right": "issue:45161", + "accept": false, + "reason": "Autocast dtype mismatch in Phi-tiny-MoE is a different bug from GPT-OSS tensor-parallel failure." + }, + { + "left": "issue:43541", + "right": "issue:43716", + "accept": false, + "reason": "Torch dynamo grouped_mm tracing error is unrelated to Mistral-3 image dtype mismatch." + }, + { + "left": "issue:43828", + "right": "issue:44292", + "accept": false, + "reason": "Different model families and failure modes." + }, + { + "left": "issue:43716", + "right": "issue:45685", + "accept": false, + "reason": "Different subsystem and hardware-specific failures." + }, + { + "left": "issue:45507", + "right": "issue:45685", + "accept": false, + "reason": "GraniteMoEHybrid invalid method is unrelated to the MPS histogram Int limitation." + }, + { + "left": "issue:44292", + "right": "issue:44918", + "accept": false, + "reason": "Different Qwen-related failures with different triggers and stack behavior." + }, + { + "left": "issue:43716", + "right": "issue:44292", + "accept": false, + "reason": "Different model and runtime issues." + }, + { + "left": "issue:43596", + "right": "issue:43638", + "accept": true, + "reason": "Same underlying ZeRO-3 BERT initialization bug and essentially the same error report." + } + ] + }, + "evaluator_result": { + "accept": false, + "feedback": "Mostly conservative, but I\u2019d reject for precision: one soft-edge reason misstates the packet (issue:44322 is titled about Qwen3_5Config lacking num_attention_heads, not Qwen3.5_5Moe), which makes the evidence handling less exact than required. The summary is otherwise plausible, but the review needs to stay tighter to the provided titles." + }, + "evaluator_used": true, + "retried": true, + "accepted_nontrivial_soft_edge": true, + "error_kind": null, + "error_message": null + }, + { + "label": "issue", + "component_index": 1, + "component_count": 28, + "review_unit_index": 2, + "review_unit_count": 65, + "status": "reviewed", + "reason": null, + "source": "llm", + "cache_hit": false, + "model": "gpt-5.4-mini?service_tier=flex", + "cluster_id": "cluster-10000-530", + "nodes": [ + "issue:30277", + "issue:34928", + "issue:35308", + "issue:36010", + "issue:36331", + "issue:39290", + "issue:41093", + "issue:41720", + "issue:41762", + "issue:42617", + "issue:43039", + "issue:43296", + "issue:43425", + "issue:43531", + "issue:43596", + "issue:43638", + "issue:43792", + "issue:43827", + "issue:43828", + "issue:43854", + "issue:43866", + "issue:43872", + "issue:43901", + "issue:43975", + "issue:44292", + "issue:44512", + "issue:44560", + "issue:44589", + "issue:44805", + "issue:44841", + "issue:44863", + "issue:45020", + "issue:45070", + "issue:45237", + "issue:45685", + "issue:45800", + "issue:45812", + "issue:45823", + "issue:45901", + "issue:45920", + "issue:46015", + "issue:46294", + "issue:46304", + "issue:46402", + "issue:46772", + "issue:46901", + "issue:47394" + ], + "soft_pairs": [ + "issue:43827|issue:44512", + "issue:43828|issue:45685", + "issue:43039|issue:43828", + "issue:30277|issue:36331", + "issue:34928|issue:36010", + "issue:45237|issue:46015", + "issue:43039|issue:44805", + "issue:45800|issue:46304", + "issue:43296|issue:46304", + "issue:39290|issue:41720", + "issue:43866|issue:46402", + "issue:41762|issue:43596", + "issue:45237|issue:45901", + "issue:44589|issue:46304", + "issue:43901|issue:44512", + "issue:43039|issue:44560", + "issue:44841|issue:45823", + "issue:45823|issue:46294", + "issue:43975|issue:46015", + "issue:43866|issue:46772", + "issue:45020|issue:46304", + "issue:44863|issue:46015", + "issue:35308|issue:42617", + "issue:45812|issue:45920", + "issue:46901|issue:47394", + "issue:44292|issue:46015", + "issue:43872|issue:46304", + "issue:39290|issue:43531", + "issue:43866|issue:44863", + "issue:41093|issue:43638", + "issue:43792|issue:45901", + "issue:43854|issue:43866", + "issue:43425|issue:45070" + ], + "prepared_review_unit_hash": "ca06b3e7403c76a2f7235fb71c4a34eee0e0e1b8ce5f6b6c75b229347550b0ee", + "component_budget": { + "node_count": 530, + "item_count": 530, + "soft_pair_count": 1954, + "serialized_chars": 1013004, + "estimated_input_tokens": 253251, + "estimated_eval_tokens": 506758 + }, + "budget": { + "node_count": 47, + "item_count": 47, + "soft_pair_count": 33, + "serialized_chars": 36962, + "estimated_input_tokens": 9241, + "estimated_eval_tokens": 18738 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": true, + "analyst_result": { + "summary": "Most pairs are unrelated despite surface-level keyword similarity. I only accept a few where the titles point to the same concrete regression/path: Zero-3 init index-out-of-bounds, AutoTokenizer wrong IDs, and double-shifted training loss.", + "confidence": 0.78, + "canonical_issue_reason": "issue:45920 is the strongest canonical issue candidate: it describes the broad v5 AutoTokenizer regression across multiple models, and it cleanly subsumes the narrower Granite-only report (#45812).", + "canonical_pr_reason": null, + "best_issue_reason": "issue:45920 is the best overall issue to keep as the canonical duplicate target because it is broad, specific about the regression, and already aggregates community evidence.", + "best_pr_reason": null, + "soft_edge_verdicts": [ + { + "left": "issue:43827", + "right": "issue:44512", + "accept": false, + "reason": "Both are docs-related, but they refer to different removed surfaces (pipeline() vs run command) and do not describe the same bug." + }, + { + "left": "issue:43828", + "right": "issue:45685", + "accept": false, + "reason": "Different failures and code paths: autocast dtype mismatch vs MPS histogram op not implemented for Int." + }, + { + "left": "issue:43039", + "right": "issue:43828", + "accept": false, + "reason": "One is a Liger Kernel cross-entropy complaint; the other is an autocast dtype mismatch. Too different." + }, + { + "left": "issue:30277", + "right": "issue:36331", + "accept": false, + "reason": "Unrelated training/distributed errors: collective mismatch with ZeRO3 vs compute_loss API breakage." + }, + { + "left": "issue:34928", + "right": "issue:36010", + "accept": false, + "reason": "Activation checkpointing/FSDP tensor-size mismatch is unrelated to the GenerationMixin import error." + }, + { + "left": "issue:45237", + "right": "issue:46015", + "accept": false, + "reason": "AMD GPU runtime failure and outdated RAG docs are unrelated." + }, + { + "left": "issue:43039", + "right": "issue:44805", + "accept": false, + "reason": "Both mention shape/mask problems, but the underlying reports are different and not the same code-path." + }, + { + "left": "issue:45800", + "right": "issue:46304", + "accept": false, + "reason": "These are separate compatibility regressions: torch/transformers versioning vs USE_DISTRIBUTED=0 support." + }, + { + "left": "issue:43296", + "right": "issue:46304", + "accept": false, + "reason": "Different model-loading failure in vLLM vs distributed-build incompatibility." + }, + { + "left": "issue:39290", + "right": "issue:41720", + "accept": false, + "reason": "Different model families and different failures: missing sliding_window_pattern vs cudaErrorAssert in auto device mapping." + }, + { + "left": "issue:43866", + "right": "issue:46402", + "accept": false, + "reason": "Checkpoint corruption/loader failure for Ovis2 is not the same issue as Sam3Model weights not loading." + }, + { + "left": "issue:41762", + "right": "issue:43596", + "accept": true, + "reason": "Both report the same underlying DeepSpeed ZeRO-3 index-out-of-bounds on loading a model, with nearly identical failure mode and code path." + }, + { + "left": "issue:45237", + "right": "issue:45901", + "accept": false, + "reason": "Different tasks and failure modes; no evidence they share the same underlying bug." + }, + { + "left": "issue:44589", + "right": "issue:46304", + "accept": false, + "reason": "Float8 storage lookup error is unrelated to the distributed build compatibility regression." + }, + { + "left": "issue:43901", + "right": "issue:44512", + "accept": false, + "reason": "Both are docs issues, but they concern different documented behaviors and removals." + }, + { + "left": "issue:43039", + "right": "issue:44560", + "accept": false, + "reason": "Cross-entropy being called by Liger Kernel is unrelated to Qwen3-VL video StopIteration." + }, + { + "left": "issue:44841", + "right": "issue:45823", + "accept": false, + "reason": "Different models and failures: processor handling vs device-mapped forward mismatch." + }, + { + "left": "issue:45823", + "right": "issue:46294", + "accept": false, + "reason": "Different subsystems and symptoms: device_map forward mismatch vs chat-template video token mismatch." + }, + { + "left": "issue:43975", + "right": "issue:46015", + "accept": false, + "reason": "Detokenization regression is unrelated to outdated RAG examples." + }, + { + "left": "issue:43866", + "right": "issue:46772", + "accept": false, + "reason": "Checkpoint corruption vs ModelOpt/NVFP4 compatibility are different problems." + }, + { + "left": "issue:45020", + "right": "issue:46304", + "accept": false, + "reason": "Remote-code regressions and distributed-build incompatibility are not the same bug." + }, + { + "left": "issue:44863", + "right": "issue:46015", + "accept": false, + "reason": "NemotronH checkpoint loading and RAG docs are unrelated." + }, + { + "left": "issue:35308", + "right": "issue:42617", + "accept": false, + "reason": "Both involve multi-GPU contexts, but the actual crashes are different and not clearly the same code path." + }, + { + "left": "issue:45812", + "right": "issue:45920", + "accept": true, + "reason": "Both describe the same v5 AutoTokenizer regression producing wrong token IDs; #45920 is the broader report and #45812 is a narrower model-specific instance." + }, + { + "left": "issue:46901", + "right": "issue:47394", + "accept": true, + "reason": "Both point to the same training-loss double-shift regression in causal LM code; the GitForCausalLM report adds an extra symptom, but the core bug matches." + }, + { + "left": "issue:44292", + "right": "issue:46015", + "accept": false, + "reason": "Qwen-3-8B-NVFP4 runtime failure is unrelated to outdated RAG docs." + }, + { + "left": "issue:43872", + "right": "issue:46304", + "accept": false, + "reason": "bitsandbytes Int8Params incompatibility is unrelated to the distributed build issue." + }, + { + "left": "issue:39290", + "right": "issue:43531", + "accept": false, + "reason": "Both mention sliding_window, but they affect different model/config paths and are not clearly the same bug." + }, + { + "left": "issue:43866", + "right": "issue:44863", + "accept": false, + "reason": "Different model-loading failures; no evidence of a shared root cause." + }, + { + "left": "issue:41093", + "right": "issue:43638", + "accept": false, + "reason": "Similar-looking indexing errors, but one is a mask shape mismatch and the other is a ZeRO-3 zero-size index error." + }, + { + "left": "issue:43792", + "right": "issue:45901", + "accept": false, + "reason": "Whisper loading/runtime failure is unrelated to table-question-answering crashes." + }, + { + "left": "issue:43854", + "right": "issue:43866", + "accept": false, + "reason": "Unit-test loading failure for GLM-4.7-Flash is not the same as a corrupted Ovis2 checkpoint." + }, + { + "left": "issue:43425", + "right": "issue:45070", + "accept": false, + "reason": "Torch 2.10 compatibility and a pydantic PretrainedConfig regression are unrelated." + } + ] + }, + "evaluator_result": { + "accept": true, + "feedback": "The summary is conservative and grounded in the packet: it only accepts a few pairs that clearly share the same concrete failure mode (ZeRO-3 index-out-of-bounds, AutoTokenizer wrong IDs, and double-shifted training loss), and the canonical-issue choice of issue:45920 over issue:45812 is well supported by the broader title and the duplicate relationship. I did not see evidence of overstatement in the remaining rejects." + }, + "evaluator_used": true, + "retried": false, + "accepted_nontrivial_soft_edge": true, + "error_kind": null, + "error_message": null + }, + { + "label": "issue", + "component_index": 1, + "component_count": 28, + "review_unit_index": 3, + "review_unit_count": 65, + "status": "reviewed", + "reason": null, + "source": "llm", + "cache_hit": false, + "model": "gpt-5.4-mini?service_tier=flex", + "cluster_id": "cluster-10000-530", + "nodes": [ + "issue:36010", + "issue:39186", + "issue:39290", + "issue:41093", + "issue:41762", + "issue:43054", + "issue:43296", + "issue:43366", + "issue:43404", + "issue:43531", + "issue:43541", + "issue:43572", + "issue:43596", + "issue:43716", + "issue:43792", + "issue:43828", + "issue:43866", + "issue:43872", + "issue:43975", + "issue:43994", + "issue:44291", + "issue:44292", + "issue:44387", + "issue:44479", + "issue:44805", + "issue:44841", + "issue:44918", + "issue:45005", + "issue:45020", + "issue:45070", + "issue:45084", + "issue:45362", + "issue:45399", + "issue:45464", + "issue:45507", + "issue:45685", + "issue:45759", + "issue:45800", + "issue:45823", + "issue:45901", + "issue:45902", + "issue:46165", + "issue:46294", + "issue:46468", + "issue:46640", + "issue:46661", + "issue:46869", + "issue:47228" + ], + "soft_pairs": [ + "issue:43541|issue:47228", + "issue:43975|issue:45901", + "issue:43596|issue:45800", + "issue:43792|issue:45070", + "issue:45362|issue:45464", + "issue:43541|issue:46869", + "issue:44292|issue:45901", + "issue:39290|issue:45823", + "issue:45464|issue:45685", + "issue:36010|issue:39290", + "issue:43541|issue:45685", + "issue:43541|issue:43828", + "issue:44841|issue:45084", + "issue:45507|issue:46869", + "issue:43404|issue:43866", + "issue:41093|issue:41762", + "issue:45399|issue:46661", + "issue:43572|issue:45070", + "issue:45800|issue:45823", + "issue:43872|issue:47228", + "issue:44805|issue:44918", + "issue:43828|issue:45464", + "issue:43716|issue:47228", + "issue:43296|issue:43366", + "issue:43366|issue:45464", + "issue:44291|issue:45070", + "issue:43531|issue:45070", + "issue:44387|issue:45005", + "issue:46869|issue:47228", + "issue:45020|issue:45800", + "issue:45507|issue:47228", + "issue:43716|issue:46294", + "issue:41093|issue:43596", + "issue:44479|issue:46294", + "issue:43054|issue:43994", + "issue:46468|issue:46869", + "issue:43596|issue:46165", + "issue:45759|issue:45902", + "issue:46640|issue:46661", + "issue:39186|issue:44918", + "issue:43716|issue:44918" + ], + "prepared_review_unit_hash": "fdb555bdfbfc411514e38acdf9b4f09251f604e6cb009d2a960a42661a9f5ce0", + "component_budget": { + "node_count": 530, + "item_count": 530, + "soft_pair_count": 1954, + "serialized_chars": 1013004, + "estimated_input_tokens": 253251, + "estimated_eval_tokens": 506758 + }, + "budget": { + "node_count": 48, + "item_count": 48, + "soft_pair_count": 41, + "serialized_chars": 39837, + "estimated_input_tokens": 9960, + "estimated_eval_tokens": 20176 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": true, + "analyst_result": { + "summary": "Mostly unrelated issue set with one strong soft-duplicate pair around Qwen3.5 video-input regressions. I\u2019d use the broad remote-code regression report as the best issue representative.", + "confidence": 0.79, + "canonical_issue_reason": "issue:45020 is the broadest, most representative regression report in the set; it captures a general recent-version compatibility break affecting remote_code models, while most other candidates are narrower, model-specific symptoms.", + "canonical_pr_reason": null, + "best_issue_reason": "issue:45020 is the best standalone representative because it is broad, cross-model, and framed as a general transformers regression rather than a single downstream symptom.", + "best_pr_reason": null, + "soft_edge_verdicts": [ + { + "left": "issue:43541", + "right": "issue:47228", + "accept": false, + "reason": "Different failures: Mixtral grouped_mm tracing error vs Sam3 bf16 compilation regression." + }, + { + "left": "issue:43975", + "right": "issue:45901", + "accept": false, + "reason": "Different subsystems and symptoms: detokenization bug vs table-question-answering crash." + }, + { + "left": "issue:43596", + "right": "issue:45800", + "accept": false, + "reason": "Zero-3/BertModel loading failure is unrelated to torch/transformers version incompatibility." + }, + { + "left": "issue:43792", + "right": "issue:45070", + "accept": false, + "reason": "Whisper runtime failure and a PretrainedConfig pydantic-field regression are not the same bug." + }, + { + "left": "issue:45362", + "right": "issue:45464", + "accept": false, + "reason": "Both involve Qwen3.5, but one is a chat crash and the other is a streaming API failure; not clearly one code path." + }, + { + "left": "issue:43541", + "right": "issue:46869", + "accept": false, + "reason": "Mixtral tracing/grouped_mm issue is unrelated to YOLO multi-GPU device-map behavior." + }, + { + "left": "issue:44292", + "right": "issue:45901", + "accept": false, + "reason": "NVFP4 model load error is unrelated to table-question-answering crashes." + }, + { + "left": "issue:39290", + "right": "issue:45823", + "accept": false, + "reason": "Gemma3 sliding_window_pattern/vLLM issue is different from Gemma4 PLE device mismatch." + }, + { + "left": "issue:45464", + "right": "issue:45685", + "accept": false, + "reason": "Qwen3.5 streaming inference failure is unrelated to the MPS histogram op error." + }, + { + "left": "issue:36010", + "right": "issue:39290", + "accept": false, + "reason": "ImportError for GenerationMixin is unrelated to Gemma3 sliding_window_pattern behavior." + }, + { + "left": "issue:43541", + "right": "issue:45685", + "accept": false, + "reason": "Different models and code paths: Mixtral tracing vs MPS MoE histogram failure." + }, + { + "left": "issue:43541", + "right": "issue:43828", + "accept": false, + "reason": "Both are MoE-related, but grouped_mm tracing and autocast dtype mismatch are distinct failures." + }, + { + "left": "issue:44841", + "right": "issue:45084", + "accept": false, + "reason": "Voxtral processor failure is unrelated to the non-template node compilation error." + }, + { + "left": "issue:45507", + "right": "issue:46869", + "accept": false, + "reason": "GraniteMoEHybrid invalid method call is not the same bug as YOLO device-map auto failure." + }, + { + "left": "issue:43404", + "right": "issue:43866", + "accept": false, + "reason": "Tied lm_head weights and a corrupted Ovis2 checkpoint are different problems." + }, + { + "left": "issue:41093", + "right": "issue:41762", + "accept": false, + "reason": "Both are index/shape errors, but the mask-shape mismatch and ZeRO-3 empty-tensor index-out-of-bounds are distinct." + }, + { + "left": "issue:45399", + "right": "issue:46661", + "accept": false, + "reason": "Flash-Attn fallback gating and remote-code config_class errors do not share the same code path." + }, + { + "left": "issue:43572", + "right": "issue:45070", + "accept": false, + "reason": "Missing StableLmConfig pad_token_idx and a pydantic PretrainedConfig field regression are different config issues." + }, + { + "left": "issue:45800", + "right": "issue:45823", + "accept": false, + "reason": "Torch/transformers version incompatibility is unrelated to Gemma4 device placement mismatch." + }, + { + "left": "issue:43872", + "right": "issue:47228", + "accept": false, + "reason": "bitsandbytes Int8Params API incompatibility is unrelated to Sam3 compilation regression." + }, + { + "left": "issue:44805", + "right": "issue:44918", + "accept": false, + "reason": "Both mention tensor shape problems, but one is a generic mask-shape error and the other is a Qwen3.5/TRl embedding unpacking issue." + }, + { + "left": "issue:43828", + "right": "issue:45464", + "accept": false, + "reason": "Autocast dtype mismatch in Phi-tiny-MoE is unrelated to Qwen3.5 streaming API failure." + }, + { + "left": "issue:43716", + "right": "issue:47228", + "accept": false, + "reason": "Mistral-3 image preprocessor dtype mismatch and Sam3 bf16 compilation regression are unrelated." + }, + { + "left": "issue:43296", + "right": "issue:43366", + "accept": false, + "reason": "PaddleOCR-VL loading in vLLM and GGUF gpt-oss support are different feature areas." + }, + { + "left": "issue:43366", + "right": "issue:45464", + "accept": false, + "reason": "GGUF architecture support and Qwen3.5 streaming inference failure are unrelated." + }, + { + "left": "issue:44291", + "right": "issue:45070", + "accept": false, + "reason": "init_empty_weights/_is_hf_initialized TypeError is a different failure from the pydantic PretrainedConfig field regression." + }, + { + "left": "issue:43531", + "right": "issue:45070", + "accept": false, + "reason": "Qwen3-MoE sliding_window behavior is unrelated to the PretrainedConfig pydantic-model bug." + }, + { + "left": "issue:44387", + "right": "issue:45005", + "accept": false, + "reason": "int4 CUDA memory/OOM regression and tied-weights translation-model issues are different bugs." + }, + { + "left": "issue:46869", + "right": "issue:47228", + "accept": false, + "reason": "YOLO multi-GPU device-map problems are unrelated to Sam3 bf16 compilation regressions." + }, + { + "left": "issue:45020", + "right": "issue:45800", + "accept": false, + "reason": "Both are broad compatibility reports, but one is about remote_code models and the other is torch-version incompatibility." + }, + { + "left": "issue:45507", + "right": "issue:47228", + "accept": false, + "reason": "GraniteMoEHybrid invalid method call is unrelated to Sam3 bf16 compilation." + }, + { + "left": "issue:43716", + "right": "issue:46294", + "accept": false, + "reason": "Different multimodal bugs: image preprocessor dtype mismatch vs video feature/token mismatch." + }, + { + "left": "issue:41093", + "right": "issue:43596", + "accept": false, + "reason": "Mask-length mismatch and zero-size index-out-of-bounds during load are not the same issue." + }, + { + "left": "issue:44479", + "right": "issue:46294", + "accept": true, + "reason": "Both describe the same Qwen3.5-family video input regression, with 46294 looking like a specific manifestation of the broader bug in 44479." + }, + { + "left": "issue:43054", + "right": "issue:43994", + "accept": false, + "reason": "Both involve SigLIP2, but the reported symptoms differ enough that they are not clearly the same concrete bug." + }, + { + "left": "issue:46468", + "right": "issue:46869", + "accept": false, + "reason": "Nemotron example breakage is unrelated to YOLO multi-GPU device-map behavior." + }, + { + "left": "issue:43596", + "right": "issue:46165", + "accept": false, + "reason": "ZeRO-3/BertModel loading failure and remote-code resume-training issue are different problems." + }, + { + "left": "issue:45759", + "right": "issue:45902", + "accept": false, + "reason": "Composite-config unwrapping and uninitialized rotary-embedding memory are separate bugs." + }, + { + "left": "issue:46640", + "right": "issue:46661", + "accept": false, + "reason": "Flash Attention 2 issue and remote-code model_class/config_class error are unrelated." + }, + { + "left": "issue:39186", + "right": "issue:44918", + "accept": false, + "reason": "FSDP 2-D weight error is unrelated to Qwen3.5/TRl embedding unpacking." + }, + { + "left": "issue:43716", + "right": "issue:44918", + "accept": false, + "reason": "Image-preprocessor dtype mismatch is unrelated to Qwen3.5 embedding unpacking under TRL." + } + ] + }, + "evaluator_result": { + "accept": true, + "feedback": "Grounded overall. The representative issue choice is conservative and supported by the packet, and the soft-edge verdicts generally avoid overstating duplicates. The one accepted pair (44479/46294) looks plausibly like the same Qwen3.5 video-input regression at different specificity levels." + }, + "evaluator_used": true, + "retried": false, + "accepted_nontrivial_soft_edge": true, + "error_kind": null, + "error_message": null + }, + { + "label": "issue", + "component_index": 1, + "component_count": 28, + "review_unit_index": 4, + "review_unit_count": 65, + "status": "reviewed", + "reason": null, + "source": "llm", + "cache_hit": false, + "model": "gpt-5.4-mini?service_tier=flex", + "cluster_id": "cluster-10000-530", + "nodes": [ + "issue:33666", + "issue:35308", + "issue:36331", + "issue:38175", + "issue:39692", + "issue:40990", + "issue:41762", + "issue:42915", + "issue:43054", + "issue:43404", + "issue:43644", + "issue:43645", + "issue:43828", + "issue:43872", + "issue:43873", + "issue:43950", + "issue:43975", + "issue:44292", + "issue:44360", + "issue:44492", + "issue:44512", + "issue:44530", + "issue:44534", + "issue:44589", + "issue:44863", + "issue:44918", + "issue:44928", + "issue:44960", + "issue:45005", + "issue:45183", + "issue:45399", + "issue:45507", + "issue:45632", + "issue:45698", + "issue:45701", + "issue:45902", + "issue:45920", + "issue:45987", + "issue:46032", + "issue:46463", + "issue:46489", + "issue:46491", + "issue:46858", + "issue:46899", + "issue:46932", + "issue:47246", + "issue:47332", + "issue:47333" + ], + "soft_pairs": [ + "issue:43645|issue:45399", + "issue:40990|issue:44960", + "issue:42915|issue:45399", + "issue:45632|issue:45698", + "issue:43404|issue:45005", + "issue:43873|issue:45005", + "issue:44360|issue:44512", + "issue:35308|issue:36331", + "issue:44492|issue:44512", + "issue:44918|issue:45507", + "issue:44530|issue:45005", + "issue:43975|issue:44292", + "issue:45183|issue:46463", + "issue:44863|issue:45005", + "issue:35308|issue:42915", + "issue:39692|issue:43054", + "issue:46489|issue:46491", + "issue:43950|issue:44534", + "issue:36331|issue:41762", + "issue:43644|issue:43950", + "issue:33666|issue:35308", + "issue:46032|issue:47246", + "issue:43872|issue:44589", + "issue:46899|issue:46932", + "issue:43054|issue:45701", + "issue:45920|issue:46489", + "issue:38175|issue:43054", + "issue:47332|issue:47333", + "issue:43828|issue:44928", + "issue:45987|issue:46858", + "issue:43950|issue:45902" + ], + "prepared_review_unit_hash": "8fd0cd03b24e8a44694e7bfc8ee18d0bb7a21d7f6d527c14e5f8780ed824a9a6", + "component_budget": { + "node_count": 530, + "item_count": 530, + "soft_pair_count": 1954, + "serialized_chars": 1013004, + "estimated_input_tokens": 253251, + "estimated_eval_tokens": 506758 + }, + "budget": { + "node_count": 48, + "item_count": 48, + "soft_pair_count": 31, + "serialized_chars": 37131, + "estimated_input_tokens": 9283, + "estimated_eval_tokens": 18822 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": true, + "analyst_result": { + "summary": "This cluster is mostly heterogeneous, but it contains a few clear duplicate sets: the v5 non-persistent-buffer regression, the Gemma4 quantized embedder dtype-cast bug, the Qwen3 offloading save failure, and the StaticCache get_seq_length contract issue. A couple of other pairs are thematically related but not the same concrete bug.", + "confidence": 0.71, + "canonical_issue_reason": "Issue #43950 is the clearest and broadest statement of the shared v5 non-persistent-buffer regression, and it best covers the closely related follow-ups.", + "canonical_pr_reason": null, + "best_issue_reason": "#43950 is the strongest hub issue for the cluster\u2019s duplicate set around non-persistent buffer corruption and has the cleanest root-cause framing.", + "best_pr_reason": null, + "soft_edge_verdicts": [ + { + "left": "issue:43645", + "right": "issue:45399", + "accept": false, + "reason": "Unrelated problems: custom model initialization in notebooks vs flash-attn fallback gating." + }, + { + "left": "issue:40990", + "right": "issue:44960", + "accept": false, + "reason": "Different bugs and models: high perplexity on gpt-oss vs GLM5 model issue." + }, + { + "left": "issue:42915", + "right": "issue:45399", + "accept": false, + "reason": "FP8 quantization failure on Qwen3Moe is unrelated to flash-attn fallback checks." + }, + { + "left": "issue:45632", + "right": "issue:45698", + "accept": false, + "reason": "Both involve custom-module loading, but the symptoms and failure modes differ enough to avoid treating as the same bug." + }, + { + "left": "issue:43404", + "right": "issue:45005", + "accept": false, + "reason": "Both mention tied weights, but they target different model families and code paths." + }, + { + "left": "issue:43873", + "right": "issue:45005", + "accept": false, + "reason": "Quantization/offloading behavior is unrelated to tied-weights regressions." + }, + { + "left": "issue:44360", + "right": "issue:44512", + "accept": false, + "reason": "DSA indexer discussion is unrelated to a docs typo about the removed run command." + }, + { + "left": "issue:35308", + "right": "issue:36331", + "accept": false, + "reason": "Different failure classes: iterable-dataset multi-GPU crash vs compute_loss signature change." + }, + { + "left": "issue:44492", + "right": "issue:44512", + "accept": false, + "reason": "A cache-strategy typo is not the same as outdated docs about the removed run command." + }, + { + "left": "issue:44918", + "right": "issue:45507", + "accept": false, + "reason": "Different model-specific runtime failures; no shared concrete bug." + }, + { + "left": "issue:44530", + "right": "issue:45005", + "accept": false, + "reason": "PagedAttentionCache linear_attention crash is unrelated to tied-weight handling." + }, + { + "left": "issue:43975", + "right": "issue:44292", + "accept": false, + "reason": "Tokenizer detokenization issues for different models are not the same bug." + }, + { + "left": "issue:45183", + "right": "issue:46463", + "accept": false, + "reason": "Both are validation/error-message issues, but they affect different validation paths and are not clearly the same underlying defect." + }, + { + "left": "issue:44863", + "right": "issue:45005", + "accept": false, + "reason": "Checkpoint loading for NemotronH is unrelated to tied-weight regressions." + }, + { + "left": "issue:35308", + "right": "issue:42915", + "accept": false, + "reason": "Completely different topics: dataloader/multi-GPU crash vs FP8 quantization failure." + }, + { + "left": "issue:39692", + "right": "issue:43054", + "accept": false, + "reason": "SigLIP2 docs errors and SigLIP2 embedding quality are related to the same model family but not the same concrete bug." + }, + { + "left": "issue:46489", + "right": "issue:46491", + "accept": false, + "reason": "Both are tokenizer regressions, but they hit different models and different observed behavior." + }, + { + "left": "issue:43950", + "right": "issue:44534", + "accept": true, + "reason": "Same non-persistent-buffer regression in v5; 44534 is a near-duplicate phrasing of the corruption/junk-buffer issue." + }, + { + "left": "issue:36331", + "right": "issue:41762", + "accept": false, + "reason": "Unexpected compute_loss kwarg and ZeRO-3 load-time index error are unrelated." + }, + { + "left": "issue:43644", + "right": "issue:43950", + "accept": true, + "reason": "Same v5 non-persistent-buffer corruption bug, with 43644 describing the same junk-filled buffer symptom." + }, + { + "left": "issue:33666", + "right": "issue:35308", + "accept": false, + "reason": "Both mention multi-GPU training, but the model and failure mechanism are different." + }, + { + "left": "issue:46032", + "right": "issue:47246", + "accept": false, + "reason": "Both concern Mamba2-family recurrence, but one is a generic use_cache correctness bug and the other is a NemotronH slow-path propagation issue." + }, + { + "left": "issue:43872", + "right": "issue:44589", + "accept": false, + "reason": "Different quantization/storage errors with different root causes." + }, + { + "left": "issue:46899", + "right": "issue:46932", + "accept": true, + "reason": "Same Gemma4 quantized vision/audio embedder dtype-cast bug; 46932 is the follow-up fix issue." + }, + { + "left": "issue:43054", + "right": "issue:45701", + "accept": false, + "reason": "SigLIP2 embedding quality and tokenization-version changes are different bugs." + }, + { + "left": "issue:45920", + "right": "issue:46489", + "accept": false, + "reason": "Both are tokenizer regressions, but the affected models and failure modes differ." + }, + { + "left": "issue:38175", + "right": "issue:43054", + "accept": false, + "reason": "Unexpected zero probabilities and worse text embeddings are not the same concrete SigLIP2 issue." + }, + { + "left": "issue:47332", + "right": "issue:47333", + "accept": true, + "reason": "Same offloading save failure for disk-offloaded Qwen3 models; only the model name differs." + }, + { + "left": "issue:43828", + "right": "issue:44928", + "accept": false, + "reason": "Autocast dtype mismatch and RLHF NaN explosion from 3D position_ids are different bugs." + }, + { + "left": "issue:45987", + "right": "issue:46858", + "accept": true, + "reason": "Same StaticCache.get_seq_length contract bug: it returns a tensor instead of a plain int, breaking downstream arithmetic." + }, + { + "left": "issue:43950", + "right": "issue:45902", + "accept": true, + "reason": "Same non-persistent-buffer regression family; 45902 is a specific manifestation via uninitialized `inv_freq` after materialization." + } + ] + }, + "evaluator_result": { + "accept": true, + "feedback": "Grounded overall: the accepted duplicate pairs are supported by the titles and the cluster summary stays conservative. The rejected soft edges are mostly appropriately cautious, with no clear overstatement of evidence." + }, + "evaluator_used": true, + "retried": false, + "accepted_nontrivial_soft_edge": true, + "error_kind": null, + "error_message": null + }, + { + "label": "issue", + "component_index": 1, + "component_count": 28, + "review_unit_index": 5, + "review_unit_count": 65, + "status": "reviewed", + "reason": null, + "source": "llm", + "cache_hit": false, + "model": "gpt-5.4-mini?service_tier=flex", + "cluster_id": "cluster-10000-530", + "nodes": [ + "issue:24643", + "issue:28530", + "issue:41553", + "issue:41762", + "issue:42491", + "issue:42915", + "issue:42947", + "issue:43054", + "issue:43404", + "issue:43493", + "issue:43643", + "issue:43782", + "issue:43824", + "issue:43827", + "issue:43828", + "issue:43856", + "issue:43872", + "issue:43881", + "issue:43883", + "issue:43950", + "issue:43975", + "issue:44368", + "issue:44387", + "issue:44403", + "issue:44451", + "issue:44488", + "issue:44509", + "issue:44661", + "issue:44910", + "issue:44960", + "issue:44991", + "issue:45020", + "issue:45081", + "issue:45237", + "issue:45356", + "issue:45399", + "issue:45488", + "issue:45663", + "issue:45685", + "issue:45715", + "issue:45759", + "issue:45823", + "issue:45859", + "issue:45902", + "issue:46315", + "issue:46489", + "issue:46661", + "issue:46869" + ], + "soft_pairs": [ + "issue:43054|issue:43493", + "issue:44910|issue:45859", + "issue:45663|issue:45902", + "issue:44403|issue:44451", + "issue:42947|issue:44387", + "issue:24643|issue:28530", + "issue:43975|issue:45237", + "issue:44403|issue:44488", + "issue:41762|issue:42915", + "issue:43643|issue:45020", + "issue:44991|issue:45399", + "issue:42947|issue:43856", + "issue:45081|issue:45399", + "issue:43824|issue:43872", + "issue:45685|issue:45823", + "issue:41553|issue:43883", + "issue:45020|issue:45823", + "issue:43782|issue:44960", + "issue:44661|issue:45020", + "issue:43828|issue:45823", + "issue:45020|issue:45356", + "issue:41553|issue:43881", + "issue:42491|issue:44368", + "issue:43872|issue:46869", + "issue:45759|issue:46661", + "issue:45715|issue:46315", + "issue:45020|issue:46661", + "issue:45488|issue:46489", + "issue:43404|issue:43950", + "issue:43827|issue:44509", + "issue:43881|issue:43883" + ], + "prepared_review_unit_hash": "47656424ac07705379ab1645ef65bc607edbabd3c8fa8245c715d5190d698238", + "component_budget": { + "node_count": 530, + "item_count": 530, + "soft_pair_count": 1954, + "serialized_chars": 1013004, + "estimated_input_tokens": 253251, + "estimated_eval_tokens": 506758 + }, + "budget": { + "node_count": 48, + "item_count": 48, + "soft_pair_count": 31, + "serialized_chars": 37489, + "estimated_input_tokens": 9373, + "estimated_eval_tokens": 19002 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": true, + "analyst_result": { + "summary": "The cluster is semantically related but not made of true duplicates: most items are separate regressions around model loading, tokenizers, or device/memory behavior. I would use issue 45020 as the best umbrella representative for the broad remote-code loading regression theme, but I would not merge any of the soft-edge pairs.", + "confidence": 0.69, + "canonical_issue_reason": "issue:45020 is the broadest and most central representative here: it describes a general class of recent-version breakages for remote-code models, which matches several of the loading-failure reports better than the narrower, model-specific issues.", + "canonical_pr_reason": null, + "best_issue_reason": "issue:45020 is the best single issue to summarize the dominant cluster theme because it captures the widest shared symptom set without being tied to one specific model or subsystem.", + "best_pr_reason": null, + "soft_edge_verdicts": [ + { + "left": "issue:43054", + "right": "issue:43493", + "accept": false, + "reason": "Both concern SigLIP2 text-embedding quality/implementation, but one is a performance complaint and the other is a HF-vs-JAX discrepancy report; not enough evidence they are the same concrete bug." + }, + { + "left": "issue:44910", + "right": "issue:45859", + "accept": false, + "reason": "Both mention Qwen3.5/MoE and runtime behavior, but one is a FlashAttention crash from 3D position_ids handling while the other is CPU-offload incompatibility in rotary embeddings; different failure paths." + }, + { + "left": "issue:45663", + "right": "issue:45902", + "accept": false, + "reason": "Both involve Gemma4/Qwen-style training edge cases, but the KeyError in FSDP2 attention and the uninitialized-memory bug after meta-to-empty materialization are distinct defects." + }, + { + "left": "issue:44403", + "right": "issue:44451", + "accept": false, + "reason": "Generic loading noise vs inability to load a specific model; the titles do not point to the same underlying bug." + }, + { + "left": "issue:42947", + "right": "issue:44387", + "accept": false, + "reason": "Gradient checkpointing with LoRA and increased CUDA reserved memory under int4 quantization are unrelated training regressions." + }, + { + "left": "issue:24643", + "right": "issue:28530", + "accept": false, + "reason": "DeepSpeed weight-dimension error vs early-stopping metric configuration warning; completely different subsystems." + }, + { + "left": "issue:43975", + "right": "issue:45237", + "accept": false, + "reason": "Tokenizer detokenization regression and AMD GPU model-loading failure are unrelated." + }, + { + "left": "issue:44403", + "right": "issue:44488", + "accept": false, + "reason": "These are both model-loading complaints, but they target different models and likely different root causes." + }, + { + "left": "issue:41762", + "right": "issue:42915", + "accept": false, + "reason": "Gemma3 ZeRO-3 loading crash and Qwen3Moe FineGrainedFP8 failure are separate model-specific issues." + }, + { + "left": "issue:43643", + "right": "issue:45020", + "accept": false, + "reason": "Missing fields from AutoConfig with trust_remote_code=True is related to remote-code loading, but it is a narrower configuration bug, not the same concrete issue as the broader remote-code regression report." + }, + { + "left": "issue:44991", + "right": "issue:45399", + "accept": false, + "reason": "Tokenizer loading failure for one model and flash-attn fallback blocking for another are different code paths." + }, + { + "left": "issue:42947", + "right": "issue:43856", + "accept": false, + "reason": "LoRA checkpointing behavior and Qwen3 MoE memory usage are different problems despite both involving training efficiency." + }, + { + "left": "issue:45081", + "right": "issue:45399", + "accept": false, + "reason": "Mistral regex patch crash in tokenizer loading and flash-attn fallback gating are unrelated." + }, + { + "left": "issue:43824", + "right": "issue:43872", + "accept": false, + "reason": "ImportError for a missing VL class and bitsandbytes Int8Params incompatibility are distinct API/regression issues." + }, + { + "left": "issue:45685", + "right": "issue:45823", + "accept": false, + "reason": "MPS histogram support in MoE and Gemma4 PLE device mismatch are different backend/device bugs." + }, + { + "left": "issue:41553", + "right": "issue:43883", + "accept": false, + "reason": "Bad AutoTokenizer error message for Voxtral and missing all_tied_weights_keys on Molmo are unrelated model-specific issues." + }, + { + "left": "issue:45020", + "right": "issue:45823", + "accept": false, + "reason": "Remote-code loading regressions and Gemma4 device-map forward mismatch are different problem classes." + }, + { + "left": "issue:43782", + "right": "issue:44960", + "accept": false, + "reason": "Qwen3VL weight_only=True load error and the generic GLM5 issue do not describe the same failure." + }, + { + "left": "issue:44661", + "right": "issue:45020", + "accept": false, + "reason": "Failure in add-new-model-like for TOKENIZER_MAPPING_NAMES is a registry-management bug, not the same as the broader remote-code loading regression." + }, + { + "left": "issue:43828", + "right": "issue:45823", + "accept": false, + "reason": "Autocast dtype mismatch in Phi-tiny-MoE-instruct and Gemma4 PLE device mismatch are separate runtime issues." + }, + { + "left": "issue:45020", + "right": "issue:45356", + "accept": false, + "reason": "General remote-code breakage and the Kimi-K2.5 tokenizer regression have different symptoms and likely different fixes." + }, + { + "left": "issue:41553", + "right": "issue:43881", + "accept": false, + "reason": "Bad AutoTokenizer error messaging for Voxtral and glm-4v-9b load failure are not the same bug." + }, + { + "left": "issue:42491", + "right": "issue:44368", + "accept": false, + "reason": "LoRA compatibility across versions and a tie_word_embeddings warning during Qwen3.5 fine-tuning are not duplicates." + }, + { + "left": "issue:43872", + "right": "issue:46869", + "accept": false, + "reason": "bitsandbytes Int8Params incompatibility and YOLO multi-GPU device-map failure are unrelated." + }, + { + "left": "issue:45759", + "right": "issue:46661", + "accept": false, + "reason": "Composite Qwen config unwrapping and remote-code model_class.config_class AttributeError are both AutoModel-related, but they are different code paths and not clearly the same bug." + }, + { + "left": "issue:45715", + "right": "issue:46315", + "accept": false, + "reason": "Slow-tokenizer conversion inefficiency in two different methods is performance-related, but the titles point to separate hot paths and not one shared defect." + }, + { + "left": "issue:45020", + "right": "issue:46661", + "accept": false, + "reason": "Remote-code breakage is broad, but this is a specific AutoModel.from_pretrained AttributeError with remote-code models, not the same concrete issue as the umbrella report." + }, + { + "left": "issue:45488", + "right": "issue:46489", + "accept": false, + "reason": "Both are DeepSeek tokenizer regressions, but one is a ByteLevel-vs-Metaspace override issue and the other is an incorrect detokenization gap from a prior fix; too different to merge." + }, + { + "left": "issue:43404", + "right": "issue:43950", + "accept": false, + "reason": "lm_head tying bug in Mistral3 and non-persistent buffer corruption during from_pretrained are different state-loading problems." + }, + { + "left": "issue:43827", + "right": "issue:44509", + "accept": false, + "reason": "These are both docs cleanups after pipeline removals, but the pair is not explicitly the same artifact set and the titles suggest separate documentation mentions." + }, + { + "left": "issue:43881", + "right": "issue:43883", + "accept": false, + "reason": "glm-4v-9b loading failure and Molmo missing all_tied_weights_keys are different model-specific loading errors." + } + ] + }, + "evaluator_result": null, + "evaluator_used": false, + "retried": false, + "accepted_nontrivial_soft_edge": false, + "error_kind": null, + "error_message": null + }, + { + "label": "issue", + "component_index": 1, + "component_count": 28, + "review_unit_index": 6, + "review_unit_count": 65, + "status": "reviewed", + "reason": null, + "source": "llm", + "cache_hit": false, + "model": "gpt-5.4-mini?service_tier=flex", + "cluster_id": "cluster-10000-530", + "nodes": [ + "issue:24643", + "issue:30064", + "issue:36331", + "issue:41093", + "issue:41553", + "issue:42617", + "issue:42915", + "issue:42947", + "issue:43257", + "issue:43366", + "issue:43531", + "issue:43596", + "issue:43638", + "issue:43716", + "issue:43792", + "issue:43824", + "issue:43828", + "issue:43854", + "issue:43940", + "issue:43950", + "issue:43957", + "issue:44155", + "issue:44220", + "issue:44479", + "issue:44488", + "issue:44530", + "issue:44560", + "issue:44589", + "issue:44805", + "issue:44863", + "issue:44928", + "issue:44991", + "issue:45305", + "issue:45313", + "issue:45325", + "issue:45584", + "issue:45588", + "issue:45663", + "issue:45685", + "issue:45823", + "issue:45834", + "issue:45902", + "issue:45923", + "issue:46402", + "issue:46468", + "issue:46515", + "issue:46894", + "issue:46897" + ], + "soft_pairs": [ + "issue:43792|issue:44220", + "issue:43257|issue:45305", + "issue:36331|issue:41093", + "issue:46894|issue:46897", + "issue:42947|issue:45923", + "issue:43940|issue:45313", + "issue:43854|issue:44863", + "issue:24643|issue:30064", + "issue:44479|issue:44560", + "issue:41093|issue:44805", + "issue:44928|issue:45588", + "issue:43366|issue:43828", + "issue:43716|issue:46468", + "issue:44530|issue:44863", + "issue:44863|issue:46402", + "issue:43366|issue:43531", + "issue:43596|issue:46515", + "issue:41553|issue:44991", + "issue:45325|issue:45902", + "issue:43366|issue:44589", + "issue:43638|issue:46515", + "issue:45663|issue:45823", + "issue:43716|issue:44560", + "issue:42915|issue:43957", + "issue:44155|issue:45584", + "issue:43824|issue:43957", + "issue:42617|issue:43366", + "issue:43950|issue:45834", + "issue:44560|issue:45685", + "issue:41553|issue:44488" + ], + "prepared_review_unit_hash": "324d6ffc40127287feb70477bf0e00c696fcaa086f775b6ef038349d0fad6a57", + "component_budget": { + "node_count": 530, + "item_count": 530, + "soft_pair_count": 1954, + "serialized_chars": 1013004, + "estimated_input_tokens": 253251, + "estimated_eval_tokens": 506758 + }, + "budget": { + "node_count": 48, + "item_count": 48, + "soft_pair_count": 30, + "serialized_chars": 36758, + "estimated_input_tokens": 9190, + "estimated_eval_tokens": 18636 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": true, + "analyst_result": { + "summary": "The cluster is mostly title-level similarity rather than confirmed duplicates. I would not accept any soft-edge merge on the packet alone; issue:44805 is still the best anchor for the mask/tensor shape-mismatch family, but its relationship to issue:41093 remains tentative.", + "confidence": 0.64, + "canonical_issue_reason": "issue:44805 is the best representative for the mask/tensor shape-mismatch family because it has the clearest error signature and reproduction context; issue:41093 is only loosely similar by title.", + "canonical_pr_reason": null, + "best_issue_reason": "issue:44805 is the strongest triage anchor in this set: the report is detailed, the failure signature is explicit, and it is the easiest case to compare against later shape-mismatch reports.", + "best_pr_reason": null, + "soft_edge_verdicts": [ + { + "left": "issue:43792", + "right": "issue:44220", + "accept": false, + "reason": "Both are audio/Whisper-related, but one is a broad run/load failure and the other is a specific fbank helper issue; different code paths." + }, + { + "left": "issue:43257", + "right": "issue:45305", + "accept": false, + "reason": "Both mention DeepSpeed, but the failures differ: weight conversion vs. gradient averaging with GAS." + }, + { + "left": "issue:36331", + "right": "issue:41093", + "accept": false, + "reason": "Unexpected compute_loss kwargs and a mask/tensor shape mismatch are different error classes and likely unrelated bugs." + }, + { + "left": "issue:46894", + "right": "issue:46897", + "accept": false, + "reason": "They share a bug pattern, but the packet does not show the same concrete model/path or a shared fix." + }, + { + "left": "issue:42947", + "right": "issue:45923", + "accept": false, + "reason": "Both touch gradient checkpointing, but one is PEFT LoRA behavior and the other is a missing model flag; different issues." + }, + { + "left": "issue:43940", + "right": "issue:45313", + "accept": false, + "reason": "The titles are similar in the Qwen3/ZeRO-3 area, but there is no explicit linkage or remediation evidence to confirm the same bug." + }, + { + "left": "issue:43854", + "right": "issue:44863", + "accept": false, + "reason": "These are model-loading failures for different models with different symptoms; not the same concrete bug." + }, + { + "left": "issue:24643", + "right": "issue:30064", + "accept": false, + "reason": "DeepSpeed training and void segmentation-map preprocessing are unrelated failure modes." + }, + { + "left": "issue:44479", + "right": "issue:44560", + "accept": false, + "reason": "Both are video-related, but one is a broad regression across several models and the other is a specific StopIteration report; not enough to merge." + }, + { + "left": "issue:41093", + "right": "issue:44805", + "accept": false, + "reason": "Same general shape-mismatch theme, but the packet only gives title-level similarity and no shared repro or fix evidence." + }, + { + "left": "issue:44928", + "right": "issue:45588", + "accept": false, + "reason": "A gradient-explosion report from 3D position_ids and a flash-attention AttributeError are different bugs." + }, + { + "left": "issue:43366", + "right": "issue:43828", + "accept": false, + "reason": "GGUF gpt-oss support and a Phi-tiny-MoE autocast dtype mismatch are unrelated." + }, + { + "left": "issue:43716", + "right": "issue:46468", + "accept": false, + "reason": "Different model families and different failure modes; no shared code path is evident." + }, + { + "left": "issue:44530", + "right": "issue:44863", + "accept": false, + "reason": "PagedAttentionCache linear_attention crash and NemotronH checkpoint loading are unrelated issues." + }, + { + "left": "issue:44863", + "right": "issue:46402", + "accept": false, + "reason": "Both are model-loading problems, but they affect different model families and different loading paths." + }, + { + "left": "issue:43366", + "right": "issue:43531", + "accept": false, + "reason": "GGUF gpt-oss support and a Qwen3-MoE sliding_window issue do not describe the same bug." + }, + { + "left": "issue:43596", + "right": "issue:46515", + "accept": false, + "reason": "A ZeRO-3 Bert index error and a generic torch incompatibility complaint are not the same concrete failure." + }, + { + "left": "issue:41553", + "right": "issue:44991", + "accept": false, + "reason": "Both are tokenizer-loading reports, but they target different models and error conditions." + }, + { + "left": "issue:45325", + "right": "issue:45902", + "accept": false, + "reason": "Qwen2.5-VL rope-index scaling and a Qwen3_5Moe rotary-buffer materialization bug are different code paths." + }, + { + "left": "issue:43366", + "right": "issue:44589", + "accept": false, + "reason": "GGUF support and a Float8 storage lookup failure are unrelated." + }, + { + "left": "issue:43638", + "right": "issue:46515", + "accept": false, + "reason": "The Bert ZeRO-3 index error and a torch version complaint are not the same bug." + }, + { + "left": "issue:45663", + "right": "issue:45823", + "accept": false, + "reason": "Same model family, but FSDP2 KeyError and device-map forward mismatch are different problems." + }, + { + "left": "issue:43716", + "right": "issue:44560", + "accept": false, + "reason": "Mistral-3 dtype mismatch and Qwen3-VL StopIteration are unrelated failures." + }, + { + "left": "issue:42915", + "right": "issue:43957", + "accept": false, + "reason": "Qwen3-MoE FP8 failure and meta-device loading regression are different issues." + }, + { + "left": "issue:44155", + "right": "issue:45584", + "accept": false, + "reason": "AudioFlamingo3 batched inference token leakage and Whisper empty-transcription generation failure are unrelated." + }, + { + "left": "issue:43824", + "right": "issue:43957", + "accept": false, + "reason": "A missing import and a meta-device loading problem are different failure modes." + }, + { + "left": "issue:42617", + "right": "issue:43366", + "accept": false, + "reason": "3d_parallel.py execution failure and GGUF support are unrelated." + }, + { + "left": "issue:43950", + "right": "issue:45834", + "accept": false, + "reason": "Non-persistent buffer corruption and a Kosmos2.5 OCR index error do not share the same code path." + }, + { + "left": "issue:44560", + "right": "issue:45685", + "accept": false, + "reason": "Qwen3-VL video StopIteration and an MPS histogram limitation in MoE are unrelated." + }, + { + "left": "issue:41553", + "right": "issue:44488", + "accept": false, + "reason": "Both mention tokenizer/model loading, but the models and symptoms differ and there is no shared fix evidence." + } + ] + }, + "evaluator_result": { + "accept": false, + "feedback": "The writeup is mostly careful, but it still overstates evidence on a few soft-edge pairs. In particular, the three accepted pairs (43940\u219445313, 44479\u219444560, 41093\u219444805) are plausible matches by title, but the packet provides no explicit linkage or shared remediation evidence, so calling them the same underlying bug / true duplicates is stronger than warranted. The summary also states 'true duplicates' more broadly than the packet supports. Conservative rephrasing or keeping those as tentative would be safer." + }, + "evaluator_used": true, + "retried": true, + "accepted_nontrivial_soft_edge": false, + "error_kind": null, + "error_message": null + }, + { + "label": "issue", + "component_index": 1, + "component_count": 28, + "review_unit_index": 7, + "review_unit_count": 65, + "status": "reviewed", + "reason": null, + "source": "llm", + "cache_hit": false, + "model": "gpt-5.4-mini?service_tier=flex", + "cluster_id": "cluster-10000-530", + "nodes": [ + "issue:36010", + "issue:38175", + "issue:39692", + "issue:41720", + "issue:42491", + "issue:42915", + "issue:42947", + "issue:43317", + "issue:43550", + "issue:43596", + "issue:43632", + "issue:43638", + "issue:43644", + "issue:43645", + "issue:43646", + "issue:43720", + "issue:43994", + "issue:44451", + "issue:44488", + "issue:44534", + "issue:44560", + "issue:44661", + "issue:44683", + "issue:44756", + "issue:44805", + "issue:44863", + "issue:44877", + "issue:44910", + "issue:44918", + "issue:44960", + "issue:44977", + "issue:45030", + "issue:45276", + "issue:45313", + "issue:45357", + "issue:45507", + "issue:45672", + "issue:45812", + "issue:45823", + "issue:45834", + "issue:46468", + "issue:46489", + "issue:46515", + "issue:46846", + "issue:46858", + "issue:46899", + "issue:47254" + ], + "soft_pairs": [ + "issue:43645|issue:43646", + "issue:44560|issue:44918", + "issue:44683|issue:46515", + "issue:36010|issue:42915", + "issue:44910|issue:46858", + "issue:43550|issue:44910", + "issue:42947|issue:47254", + "issue:45823|issue:46515", + "issue:45812|issue:46489", + "issue:44661|issue:45823", + "issue:44560|issue:44805", + "issue:44877|issue:45030", + "issue:38175|issue:43994", + "issue:43638|issue:44661", + "issue:43596|issue:44661", + "issue:44661|issue:44805", + "issue:42491|issue:43720", + "issue:45672|issue:46899", + "issue:43632|issue:45834", + "issue:43317|issue:44756", + "issue:41720|issue:42915", + "issue:38175|issue:39692", + "issue:42915|issue:44756", + "issue:43644|issue:44534", + "issue:42491|issue:45276", + "issue:44960|issue:44977", + "issue:44451|issue:44488", + "issue:45313|issue:45357", + "issue:44863|issue:46846", + "issue:45507|issue:46468" + ], + "prepared_review_unit_hash": "3bd944c90e1a26f579667044dc859dee165294f9c310c7f8c7efd3dba1ddce3a", + "component_budget": { + "node_count": 530, + "item_count": 530, + "soft_pair_count": 1954, + "serialized_chars": 1013004, + "estimated_input_tokens": 253251, + "estimated_eval_tokens": 506758 + }, + "budget": { + "node_count": 47, + "item_count": 47, + "soft_pair_count": 30, + "serialized_chars": 35975, + "estimated_input_tokens": 8994, + "estimated_eval_tokens": 18244 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": true, + "analyst_result": { + "summary": "This packet contains one clearly exact duplicate pair, a couple of plausible same-bug pairs, and several lookalikes that are too weak to merge. The strongest anchor is the non-persistent-buffer regression (43644/44534). SigLIP2 output regressions (38175/43994) and Gemma4 4-bit multimodal dtype issues (45672/46899) look like real duplicates, while the custom-model-init reports (43645/43646) are related but not evidenced well enough to treat as the same root cause.", + "confidence": 0.84, + "canonical_issue_reason": "issue:43644 is the best canonical issue because it is the earlier, broader report of the exact same 'Transformers v5 fills non-persistent buffers with junk' regression later duplicated by 44534, with broader discussion and an inbound reference.", + "canonical_pr_reason": null, + "best_issue_reason": "issue:43644 is the best global representative: it is an exact duplicate target, predates the follow-up report, and its title captures the underlying regression without narrowing it to one reproduction path.", + "best_pr_reason": null, + "soft_edge_verdicts": [ + { + "left": "issue:43645", + "right": "issue:43646", + "accept": false, + "reason": "Closely related custom-model-init failures, but the packet does not show a shared root cause or code path; treat as related reports, not duplicates." + }, + { + "left": "issue:44560", + "right": "issue:44918", + "accept": false, + "reason": "Different model families and different failure modes: video StopIteration versus TRL SFT unpacking." + }, + { + "left": "issue:44683", + "right": "issue:46515", + "accept": false, + "reason": "Both mention torch compatibility, but they describe different versions and different concrete failures." + }, + { + "left": "issue:36010", + "right": "issue:42915", + "accept": false, + "reason": "Unrelated issues: GenerationMixin import error versus Qwen3 MoE LoRA hf4.x/hf5.x incompatibility." + }, + { + "left": "issue:44910", + "right": "issue:46858", + "accept": false, + "reason": "Both are cache/attention-related, but they hit different mechanisms and backends; not enough evidence of one bug." + }, + { + "left": "issue:43550", + "right": "issue:44910", + "accept": false, + "reason": "Different model and problem class: Bamba torch.compile+SDPA versus Qwen3.5 flash-attention position_ids handling." + }, + { + "left": "issue:42947", + "right": "issue:47254", + "accept": false, + "reason": "Related training stack, but one is ineffective gradient checkpointing and the other is a ZeRO-3 checkpoint error." + }, + { + "left": "issue:45823", + "right": "issue:46515", + "accept": false, + "reason": "Gemma4 device-mismatch during forward is not the same as a generic torch version incompatibility report." + }, + { + "left": "issue:45812", + "right": "issue:46489", + "accept": false, + "reason": "Both are tokenizer regressions, but for different model families and different concrete defects." + }, + { + "left": "issue:44661", + "right": "issue:45823", + "accept": false, + "reason": "Tokenizer-mapping failure and Gemma4 device mismatch are unrelated code paths." + }, + { + "left": "issue:44560", + "right": "issue:44805", + "accept": false, + "reason": "Different symptoms and different model paths; no shared underlying defect is shown." + }, + { + "left": "issue:44877", + "right": "issue:45030", + "accept": false, + "reason": "Both involve stricter config validation, but the affected configs and fixes appear different." + }, + { + "left": "issue:38175", + "right": "issue:43994", + "accept": true, + "reason": "Both report the same SigLIP2 model regression producing wrong/zero outputs; the later issue looks like the same underlying bug re-reported." + }, + { + "left": "issue:43638", + "right": "issue:44661", + "accept": false, + "reason": "DeepSpeed ZeRO-3 Bert indexing error is unrelated to add-new-model-like/tokenizer mapping logic." + }, + { + "left": "issue:43596", + "right": "issue:44661", + "accept": false, + "reason": "Different bug classes: Bert zero3 init IndexError versus tokenizer-mapping failure." + }, + { + "left": "issue:44661", + "right": "issue:44805", + "accept": false, + "reason": "Only a broad runtime/load-failure resemblance; no evidence of the same code-path problem." + }, + { + "left": "issue:42491", + "right": "issue:43720", + "accept": false, + "reason": "Different models and failure modes; not the same underlying bug." + }, + { + "left": "issue:45672", + "right": "issue:46899", + "accept": true, + "reason": "Both are Gemma4 multimodal 4-bit quantization failures involving unsafe dtype handling in audio/vision embedders." + }, + { + "left": "issue:43632", + "right": "issue:45834", + "accept": false, + "reason": "_is_hf_initialized flag regression is unrelated to the Kosmos2.5 OCR indexing issue." + }, + { + "left": "issue:43317", + "right": "issue:44756", + "accept": false, + "reason": "Different resource/loading problems: dequantized offload failure versus mmap OOM workaround." + }, + { + "left": "issue:41720", + "right": "issue:42915", + "accept": false, + "reason": "Qwen3 auto device-mapping cuda assert and Qwen3 MoE LoRA hf4.x/hf5.x incompatibility are different bugs." + }, + { + "left": "issue:38175", + "right": "issue:39692", + "accept": false, + "reason": "Same model family, but one is incorrect outputs and the other is a docs-example mismatch plus quantization failure." + }, + { + "left": "issue:42915", + "right": "issue:44756", + "accept": false, + "reason": "No shared underlying bug; one is a model compatibility issue, the other a platform-specific mmap/OOM workaround." + }, + { + "left": "issue:43644", + "right": "issue:44534", + "accept": true, + "reason": "Exact duplicate regression: both report Transformers v5 filling non-persistent buffers with junk." + }, + { + "left": "issue:42491", + "right": "issue:45276", + "accept": false, + "reason": "LoRA hf4.x/hf5.x incompatibility is unrelated to Gemma4 token-embedding resize behavior." + }, + { + "left": "issue:44960", + "right": "issue:44977", + "accept": false, + "reason": "Unrelated GLM5 and Qwen3.5 flash-attention reports." + }, + { + "left": "issue:44451", + "right": "issue:44488", + "accept": false, + "reason": "Both are BERT load failures, but the packet does not show enough evidence that they share the same root cause." + }, + { + "left": "issue:45313", + "right": "issue:45357", + "accept": false, + "reason": "Distinct code paths: ZeRO-3 weight loading versus incorrect save_pretrained visual-encoder keys." + }, + { + "left": "issue:44863", + "right": "issue:46846", + "accept": false, + "reason": "NemotronH checkpoint loading and Qwen3.5 tensor-parallel plan issues are unrelated." + }, + { + "left": "issue:45507", + "right": "issue:46468", + "accept": false, + "reason": "Different models and failures; no duplicate evidence." + } + ] + }, + "evaluator_result": { + "accept": false, + "feedback": "The packet supports several plausible duplicate clusters, but the response is not conservative enough. In particular, issue:43645 vs issue:43646 is overstated as the same underlying regression; the packet only shows closely related custom-model-init failures, not proof of a shared root cause. The summary also overstates that custom model initialization is a clear duplicate subcluster. The 43644 vs 44534 pair looks well grounded, and 38175/43994 and 45672/46899 are plausible, but the overall precision standard is not met." + }, + "evaluator_used": true, + "retried": true, + "accepted_nontrivial_soft_edge": true, + "error_kind": null, + "error_message": null + }, + { + "label": "issue", + "component_index": 1, + "component_count": 28, + "review_unit_index": 8, + "review_unit_count": 65, + "status": "reviewed", + "reason": null, + "source": "llm", + "cache_hit": false, + "model": "gpt-5.4-mini?service_tier=flex", + "cluster_id": "cluster-10000-530", + "nodes": [ + "issue:39692", + "issue:41720", + "issue:42491", + "issue:42947", + "issue:43257", + "issue:43404", + "issue:43425", + "issue:43454", + "issue:43644", + "issue:43645", + "issue:43708", + "issue:43854", + "issue:43856", + "issue:43873", + "issue:43931", + "issue:43957", + "issue:43994", + "issue:44075", + "issue:44155", + "issue:44164", + "issue:44291", + "issue:44292", + "issue:44387", + "issue:44402", + "issue:44488", + "issue:44756", + "issue:44910", + "issue:44912", + "issue:44960", + "issue:44991", + "issue:45276", + "issue:45292", + "issue:45335", + "issue:45362", + "issue:45406", + "issue:45488", + "issue:45588", + "issue:45663", + "issue:45704", + "issue:45759", + "issue:45812", + "issue:45823", + "issue:45859", + "issue:46050", + "issue:46459", + "issue:46512", + "issue:46897", + "issue:46901" + ], + "soft_pairs": [ + "issue:46459|issue:46512", + "issue:44488|issue:44991", + "issue:43644|issue:43645", + "issue:43257|issue:43854", + "issue:39692|issue:43994", + "issue:43856|issue:43873", + "issue:42491|issue:45335", + "issue:43854|issue:43931", + "issue:42491|issue:45292", + "issue:43856|issue:44387", + "issue:43708|issue:44075", + "issue:44164|issue:44291", + "issue:45488|issue:45812", + "issue:44960|issue:46050", + "issue:45276|issue:45335", + "issue:45362|issue:45406", + "issue:43425|issue:43957", + "issue:44291|issue:44756", + "issue:41720|issue:44155", + "issue:44402|issue:44488", + "issue:45663|issue:45759", + "issue:39692|issue:44960", + "issue:44910|issue:45588", + "issue:43404|issue:43454", + "issue:46897|issue:46901", + "issue:43873|issue:44387", + "issue:42947|issue:45704", + "issue:45663|issue:45859", + "issue:44292|issue:44912", + "issue:44155|issue:45823", + "issue:43957|issue:44756" + ], + "prepared_review_unit_hash": "fa46344100f9b5bf8c1da13868d5a1537ab28ad95fc5ea6baae2604c0e637264", + "component_budget": { + "node_count": 530, + "item_count": 530, + "soft_pair_count": 1954, + "serialized_chars": 1013004, + "estimated_input_tokens": 253251, + "estimated_eval_tokens": 506758 + }, + "budget": { + "node_count": 48, + "item_count": 48, + "soft_pair_count": 31, + "serialized_chars": 37206, + "estimated_input_tokens": 9302, + "estimated_eval_tokens": 18860 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": true, + "analyst_result": { + "summary": "This cluster is highly heterogeneous; most pairs share only a broad theme like loading, tokenization, quantization, or model-init regressions, but not the same concrete bug. No soft pair looks safe to merge as duplicates.", + "confidence": 0.82, + "canonical_issue_reason": "issue:45812 is the strongest representative because it describes a broad, concrete, user-visible regression (all Granite token IDs are wrong) with clear impact and a well-scoped fix target.", + "canonical_pr_reason": null, + "best_issue_reason": "issue:45812 is the best anchor for the cluster: it is specific, high-impact, and reads like a clear regression with a single underlying code path.", + "best_pr_reason": null, + "soft_edge_verdicts": [ + { + "left": "issue:46459", + "right": "issue:46512", + "accept": false, + "reason": "Both mention dtype handling, but one is about loading composite checkpoints and the other about from_config initialization/config staleness; different bug surfaces." + }, + { + "left": "issue:44488", + "right": "issue:44991", + "accept": false, + "reason": "Both are tokenizer-loading failures, but they concern different models and likely different tokenizer files/code paths." + }, + { + "left": "issue:43644", + "right": "issue:43645", + "accept": false, + "reason": "One is about junk-filled non-persistent buffers; the other is a notebook/custom-model initialization break. Similar release theme, not the same bug." + }, + { + "left": "issue:43257", + "right": "issue:43854", + "accept": false, + "reason": "Different models and failure modes: Qwen3 MoE weight conversion under accelerate+deepspeed vs GLM-4.7-Flash loading in tests." + }, + { + "left": "issue:39692", + "right": "issue:43994", + "accept": false, + "reason": "Both involve SigLIP2, but one is a docs example with model/processor mismatch and quantization trouble; the other is nonsensical outputs with AutoModel/pipeline." + }, + { + "left": "issue:43856", + "right": "issue:43873", + "accept": false, + "reason": "Both are memory-related, but training-memory inefficiency and quantization offload behavior are distinct issues." + }, + { + "left": "issue:42491", + "right": "issue:45335", + "accept": false, + "reason": "LoRA incompatibility for Qwen3 MoE is unrelated to resize_token_embeddings not updating decoder.embed_tokens." + }, + { + "left": "issue:43854", + "right": "issue:43931", + "accept": false, + "reason": "Different models and symptoms: GLM-4.7-Flash test loading vs Qwen3-VL weight-shape mismatch." + }, + { + "left": "issue:42491", + "right": "issue:45292", + "accept": false, + "reason": "The LoRA/qwen3_moe compatibility regression is not the same as a generic output_embeddings resize propagation bug." + }, + { + "left": "issue:43856", + "right": "issue:44387", + "accept": false, + "reason": "Both mention memory, but one is Qwen3 MoE training efficiency and the other is int4 quantization causing CUDA reserved-memory OOM." + }, + { + "left": "issue:43708", + "right": "issue:44075", + "accept": false, + "reason": "Trainer resume/max_steps calculation is unrelated to SGD argument handling." + }, + { + "left": "issue:44164", + "right": "issue:44291", + "accept": false, + "reason": "Both involve model save/load flows, but extra_state handling and init_empty_weights/_is_hf_initialized are different failure points." + }, + { + "left": "issue:45488", + "right": "issue:45812", + "accept": false, + "reason": "Both are tokenizer regressions, but one changes LlamaTokenizer pre-tokenization for DeepSeek-style models; the other produces wrong token IDs for Granite." + }, + { + "left": "issue:44960", + "right": "issue:46050", + "accept": false, + "reason": "GLM5 loading and sinq quantization save/load breakage are unrelated bugs." + }, + { + "left": "issue:45276", + "right": "issue:45335", + "accept": false, + "reason": "Both are resize_token_embeddings propagation bugs, but they hit different architectures and output paths; not one concrete bug." + }, + { + "left": "issue:45362", + "right": "issue:45406", + "accept": false, + "reason": "Qwen3.5 chat crashes and Gemma4Processor missing _tokenizer are different code paths." + }, + { + "left": "issue:43425", + "right": "issue:43957", + "accept": false, + "reason": "Torch 2.10 compatibility is a dependency/version issue; meta-device loading breakage is a separate model-loading bug." + }, + { + "left": "issue:44291", + "right": "issue:44756", + "accept": false, + "reason": "Unexpected init arg in init_empty_weights is unrelated to Strix Halo mmap OOM behavior." + }, + { + "left": "issue:41720", + "right": "issue:44155", + "accept": false, + "reason": "CUDA assert on Qwen3 auto device mapping is not the same as AudioFlamingo3 batched inference token/embedding leakage." + }, + { + "left": "issue:44402", + "right": "issue:44488", + "accept": false, + "reason": "Both are tokenizer/model-loading problems, but they concern different models and different underlying defects." + }, + { + "left": "issue:45663", + "right": "issue:45759", + "accept": false, + "reason": "Gemma4 FSDP2 shared_kv_states bug and Qwen multimodal text_config unwrapping are separate composite-model issues." + }, + { + "left": "issue:39692", + "right": "issue:44960", + "accept": false, + "reason": "SigLIP2 docs/quantization errors and GLM5 loading are unrelated." + }, + { + "left": "issue:44910", + "right": "issue:45588", + "accept": false, + "reason": "Both are flash-attention crashes, but one is a packed-sequence/3D position_ids bug and the other is an s_aux=None attribute error." + }, + { + "left": "issue:43404", + "right": "issue:43454", + "accept": false, + "reason": "Both are lm_head tying problems, but they affect different model families and likely different implementations." + }, + { + "left": "issue:46897", + "right": "issue:46901", + "accept": false, + "reason": "Both describe a double-shift training-loss symptom, but they are different models without evidence of one shared code path." + }, + { + "left": "issue:43873", + "right": "issue:44387", + "accept": false, + "reason": "Quantization/offload behavior and int4 reserved-memory OOM are related broadly, but not the same bug." + }, + { + "left": "issue:42947", + "right": "issue:45704", + "accept": false, + "reason": "Gradient checkpointing with PEFT LoRA is unrelated to T5 using apex.FusedRMSNorm memory leak." + }, + { + "left": "issue:45663", + "right": "issue:45859", + "accept": false, + "reason": "Gemma4 FSDP2 attention-state rebuild and Qwen3_5MoeTextRotaryEmbedding CPU-offload incompatibility are different bugs." + }, + { + "left": "issue:44292", + "right": "issue:44912", + "accept": false, + "reason": "Qwen-3-8B-NVFP4 runtime error and git-oss-20b MXFP4 fallback are separate quantization issues." + }, + { + "left": "issue:44155", + "right": "issue:45823", + "accept": false, + "reason": "AudioFlamingo3 batched inference corruption and Gemma4 PLE device mismatch are unrelated." + }, + { + "left": "issue:43957", + "right": "issue:44756", + "accept": false, + "reason": "Meta-device loading regressions and mmap OOM on Strix Halo are different failure modes." + } + ] + }, + "evaluator_result": null, + "evaluator_used": false, + "retried": false, + "accepted_nontrivial_soft_edge": false, + "error_kind": null, + "error_message": null + }, + { + "label": "issue", + "component_index": 1, + "component_count": 28, + "review_unit_index": 9, + "review_unit_count": 65, + "status": "reviewed", + "reason": null, + "source": "llm", + "cache_hit": false, + "model": "gpt-5.4-mini?service_tier=flex", + "cluster_id": "cluster-10000-530", + "nodes": [ + "issue:36010", + "issue:39404", + "issue:42796", + "issue:43335", + "issue:43381", + "issue:43425", + "issue:43645", + "issue:43646", + "issue:43708", + "issue:43716", + "issue:43828", + "issue:43854", + "issue:43957", + "issue:44079", + "issue:44164", + "issue:44387", + "issue:44402", + "issue:44451", + "issue:44479", + "issue:44530", + "issue:44589", + "issue:44683", + "issue:44756", + "issue:44863", + "issue:44877", + "issue:44928", + "issue:45084", + "issue:45216", + "issue:45276", + "issue:45292", + "issue:45335", + "issue:45588", + "issue:45663", + "issue:45684", + "issue:45698", + "issue:45701", + "issue:45812", + "issue:45823", + "issue:46315", + "issue:46326", + "issue:46468", + "issue:46491", + "issue:46829", + "issue:46869", + "issue:46894", + "issue:46901", + "issue:47228", + "issue:47332" + ], + "soft_pairs": [ + "issue:44079|issue:46829", + "issue:43335|issue:45663", + "issue:36010|issue:39404", + "issue:44863|issue:46468", + "issue:44479|issue:44877", + "issue:44387|issue:44756", + "issue:45276|issue:45292", + "issue:43381|issue:44928", + "issue:44164|issue:44756", + "issue:43957|issue:47332", + "issue:42796|issue:46326", + "issue:45588|issue:45663", + "issue:43645|issue:44164", + "issue:43854|issue:45216", + "issue:43425|issue:47228", + "issue:45292|issue:45335", + "issue:45684|issue:45698", + "issue:44589|issue:44756", + "issue:43425|issue:43828", + "issue:44402|issue:44451", + "issue:43646|issue:43708", + "issue:43425|issue:46869", + "issue:46894|issue:46901", + "issue:44530|issue:45084", + "issue:46315|issue:46491", + "issue:44683|issue:45823", + "issue:45812|issue:46315", + "issue:43425|issue:43716", + "issue:39404|issue:45701" + ], + "prepared_review_unit_hash": "486d32fb6ae4261496044dabd70e127209272e06ffa6fc9a5ed3aa8bd3c3fc8b", + "component_budget": { + "node_count": 530, + "item_count": 530, + "soft_pair_count": 1954, + "serialized_chars": 1013004, + "estimated_input_tokens": 253251, + "estimated_eval_tokens": 506758 + }, + "budget": { + "node_count": 48, + "item_count": 48, + "soft_pair_count": 29, + "serialized_chars": 36354, + "estimated_input_tokens": 9089, + "estimated_eval_tokens": 18434 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": true, + "analyst_result": { + "summary": "The set is mostly heterogeneous. I see two real duplicate pairs: the resize_token_embeddings/output_embeddings reports, and the training-loss double-shift reports. The rest are broad thematic overlaps, not the same underlying bug.", + "confidence": 0.67, + "canonical_issue_reason": "46901 is the strongest canonical anchor because it states the bug at implementation level (training against labels[..., 1:] instead of labels), while 46894 is the more model-specific duplicate. Most other items are unrelated enough that they should not influence the canonical choice.", + "canonical_pr_reason": null, + "best_issue_reason": "46901 is the clearest overall issue to keep: it is precise, concrete, and directly actionable, whereas the remaining items are mostly broader or model-specific variants of separate problems.", + "best_pr_reason": null, + "soft_edge_verdicts": [ + { + "left": "issue:44079", + "right": "issue:46829", + "accept": false, + "reason": "Both are output-structure bugs, but one is generic ModelOutput key bookkeeping and the other is a VideoPrism classifier field mix-up; not the same concrete failure." + }, + { + "left": "issue:43335", + "right": "issue:45663", + "accept": false, + "reason": "Different bugs in different models: sparse-layer creation in SwitchTransformers vs shared_kv_states rebuild under FSDP2 in Gemma4." + }, + { + "left": "issue:36010", + "right": "issue:39404", + "accept": false, + "reason": "Unrelated failures: an import error for GenerationMixin vs Whisper pipeline return_language behavior." + }, + { + "left": "issue:44863", + "right": "issue:46468", + "accept": false, + "reason": "Same model family, but one is a checkpoint-loading failure and the other is a broken example; no clear same code path." + }, + { + "left": "issue:44479", + "right": "issue:44877", + "accept": false, + "reason": "Both touch multimodal/model loading, but one is a video-input regression and the other is strict config parsing for granite_speech; too broad to merge." + }, + { + "left": "issue:44387", + "right": "issue:44756", + "accept": false, + "reason": "Both mention OOM, but the causes differ: int4 quantization memory growth vs mmap on Strix Halo." + }, + { + "left": "issue:45276", + "right": "issue:45292", + "accept": true, + "reason": "Same resize_token_embeddings propagation bug; 45276 is the Gemma4-specific variant while 45292 is the broader output_embeddings report." + }, + { + "left": "issue:43381", + "right": "issue:44928", + "accept": false, + "reason": "Gradient checkpointing eval-mode restriction is unrelated to the RLHF NaN/SDPA math-fallback issue." + }, + { + "left": "issue:44164", + "right": "issue:44756", + "accept": false, + "reason": "save/from_pretrained extra_state handling is unrelated to mmap-related OOM." + }, + { + "left": "issue:43957", + "right": "issue:47332", + "accept": false, + "reason": "Both involve loading/offloading, but one is meta-device model loading and the other is saving a disk-offloaded model; different problems." + }, + { + "left": "issue:42796", + "right": "issue:46326", + "accept": false, + "reason": "Both are tokenizer-related, but one is BPE marker handling and the other is continue_final_message behavior for mistral_common tokenizers." + }, + { + "left": "issue:45588", + "right": "issue:45663", + "accept": false, + "reason": "Different failures in different subsystems: flash_attention sinkless crash vs Gemma4 FSDP attention state rebuild." + }, + { + "left": "issue:43645", + "right": "issue:44164", + "accept": false, + "reason": "Custom model initialization in notebooks is not the same bug as save/from_pretrained extra_state handling." + }, + { + "left": "issue:43854", + "right": "issue:45216", + "accept": false, + "reason": "Different save/load regressions affecting different models and APIs; no shared concrete failure mode." + }, + { + "left": "issue:43425", + "right": "issue:47228", + "accept": false, + "reason": "Both are torch-version compatibility regressions, but one is a generic Torch 2.10 incompatibility and the other is a Sam3 bf16 compilation regression." + }, + { + "left": "issue:45292", + "right": "issue:45335", + "accept": false, + "reason": "Both are resize_token_embeddings regressions, but they target different model internals (output_embeddings vs decoder.embed_tokens) and are not clearly the same code path." + }, + { + "left": "issue:45684", + "right": "issue:45698", + "accept": false, + "reason": "One is about permission propagation during save_pretrained; the other is about loading the wrong custom module after save_pretrained." + }, + { + "left": "issue:44589", + "right": "issue:44756", + "accept": false, + "reason": "Float8 storage lookup failure is unrelated to the Strix Halo mmap OOM issue." + }, + { + "left": "issue:43425", + "right": "issue:43828", + "accept": false, + "reason": "Torch compatibility and autocast dtype mismatch are separate problems." + }, + { + "left": "issue:44402", + "right": "issue:44451", + "accept": false, + "reason": "One is tokenizer vocab size / special token placement; the other is a model-loading failure for ScandiBERT." + }, + { + "left": "issue:43646", + "right": "issue:43708", + "accept": false, + "reason": "Custom model initialization breakage is unrelated to Trainer resume_from_checkpoint max_steps miscalculation." + }, + { + "left": "issue:43425", + "right": "issue:46869", + "accept": false, + "reason": "Generic Torch 2.10 incompatibility is not the same as YOLO multi-GPU device_map auto failure." + }, + { + "left": "issue:46894", + "right": "issue:46901", + "accept": true, + "reason": "Same double-shift training-loss bug; both say the model is training against labels[..., 1:] instead of the intended labels." + }, + { + "left": "issue:44530", + "right": "issue:45084", + "accept": false, + "reason": "PagedAttentionCache group-type crash and non-template-node compile TypeError are different failures." + }, + { + "left": "issue:46315", + "right": "issue:46491", + "accept": false, + "reason": "Performance regression in convert_tokens_to_ids is unrelated to CodeLlama leading-space round-trip corruption." + }, + { + "left": "issue:44683", + "right": "issue:45823", + "accept": false, + "reason": "Compiled flex_attention on torch >=2.9 and Gemma4 PLE device mismatch are different bugs." + }, + { + "left": "issue:45812", + "right": "issue:46315", + "accept": false, + "reason": "Wrong Granite token IDs and slow convert_tokens_to_ids are separate tokenizer issues." + }, + { + "left": "issue:43425", + "right": "issue:43716", + "accept": false, + "reason": "Torch 2.10 incompatibility is unrelated to Mistral-3 image preprocessor/model dtype mismatch." + }, + { + "left": "issue:39404", + "right": "issue:45701", + "accept": false, + "reason": "Whisper return_language pipeline regression is unrelated to the tokenization behavior change reported in 45701." + } + ] + }, + "evaluator_result": { + "accept": true, + "feedback": "The analysis is grounded in the packet and stays mostly conservative. The two accepted soft pairs are supported by the issue titles, and the rejected pairs are framed as broad thematic similarities rather than the same concrete bug. The canonical and best-issue choices are also justified by the more precise, implementation-level wording of issue:46901." + }, + "evaluator_used": true, + "retried": false, + "accepted_nontrivial_soft_edge": true, + "error_kind": null, + "error_message": null + }, + { + "label": "issue", + "component_index": 1, + "component_count": 28, + "review_unit_index": 10, + "review_unit_count": 65, + "status": "reviewed", + "reason": null, + "source": "llm", + "cache_hit": false, + "model": "gpt-5.4-mini?service_tier=flex", + "cluster_id": "cluster-10000-530", + "nodes": [ + "issue:36032", + "issue:39404", + "issue:40990", + "issue:42796", + "issue:42915", + "issue:43299", + "issue:43404", + "issue:43452", + "issue:43596", + "issue:43611", + "issue:43632", + "issue:43644", + "issue:43645", + "issue:43716", + "issue:43746", + "issue:43856", + "issue:43901", + "issue:43906", + "issue:43931", + "issue:43994", + "issue:44079", + "issue:44164", + "issue:44265", + "issue:44479", + "issue:44488", + "issue:44509", + "issue:44530", + "issue:44589", + "issue:44671", + "issue:44683", + "issue:44756", + "issue:44805", + "issue:44898", + "issue:44912", + "issue:44987", + "issue:44991", + "issue:45081", + "issue:45084", + "issue:45759", + "issue:45823", + "issue:46050", + "issue:46326", + "issue:46665", + "issue:46821", + "issue:46899", + "issue:47050", + "issue:47331", + "issue:47333" + ], + "soft_pairs": [ + "issue:43611|issue:43645", + "issue:44479|issue:44805", + "issue:44589|issue:45084", + "issue:39404|issue:46665", + "issue:44164|issue:45759", + "issue:40990|issue:47331", + "issue:36032|issue:43452", + "issue:43994|issue:44079", + "issue:43901|issue:44509", + "issue:44265|issue:44898", + "issue:44991|issue:45081", + "issue:46050|issue:46326", + "issue:43906|issue:46326", + "issue:44671|issue:44991", + "issue:43906|issue:44530", + "issue:39404|issue:44912", + "issue:45823|issue:46899", + "issue:43596|issue:44683", + "issue:44756|issue:47333", + "issue:44164|issue:46821", + "issue:43856|issue:44756", + "issue:44488|issue:44987", + "issue:44488|issue:45081", + "issue:36032|issue:42915", + "issue:43632|issue:47050", + "issue:43716|issue:45084", + "issue:43299|issue:43931", + "issue:43746|issue:47331", + "issue:43611|issue:43644", + "issue:42796|issue:43404" + ], + "prepared_review_unit_hash": "d548a79f40a64e471e2c645aca3bfc84d6c09548221ae2463697912d08bda8a6", + "component_budget": { + "node_count": 530, + "item_count": 530, + "soft_pair_count": 1954, + "serialized_chars": 1013004, + "estimated_input_tokens": 253251, + "estimated_eval_tokens": 506758 + }, + "budget": { + "node_count": 48, + "item_count": 48, + "soft_pair_count": 30, + "serialized_chars": 36942, + "estimated_input_tokens": 9236, + "estimated_eval_tokens": 18728 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": true, + "analyst_result": { + "summary": "This cluster is a broad mix of mostly unrelated Transformers v5 regression reports, heavily centered on loading/tokenizer/model-init issues. I do not see any soft edge that clearly represents the same concrete bug, so all candidate duplicate links should stay unmerged.", + "confidence": 0.42, + "canonical_issue_reason": "issue:43611 is the best anchor because it is a clear, concrete v5 loading regression and broadly representative of the common \"v5 breaks loading\" theme across the cluster.", + "canonical_pr_reason": null, + "best_issue_reason": "issue:43611 is the strongest representative issue: it is specific enough to describe a real regression while still being broad compared with the more model-specific reports.", + "best_pr_reason": null, + "soft_edge_verdicts": [ + { + "left": "issue:43611", + "right": "issue:43645", + "accept": false, + "reason": "Different failure surfaces: base_model_prefix loading vs. defining/initializing custom models in Jupyter notebooks." + }, + { + "left": "issue:44479", + "right": "issue:44805", + "accept": false, + "reason": "Both are multimodal regressions, but one is a video-input regression across Qwen VL variants and the other is a mask/tensor shape IndexError." + }, + { + "left": "issue:44589", + "right": "issue:45084", + "accept": false, + "reason": "Unrelated bugs: Float8 storage lookup failure vs. a template-compilation error." + }, + { + "left": "issue:39404", + "right": "issue:46665", + "accept": false, + "reason": "Whisper pipeline return_language regression is a different API/path from from_pretrained behavior inside Python classes." + }, + { + "left": "issue:44164", + "right": "issue:45759", + "accept": false, + "reason": "Different code paths: extra_state save/load handling vs. unwrapping text_config for composite Qwen multimodal configs." + }, + { + "left": "issue:40990", + "right": "issue:47331", + "accept": false, + "reason": "Perplexity quality issue on gpt-oss-20b is not the same as Metal quantization batch corruption in generation." + }, + { + "left": "issue:36032", + "right": "issue:43452", + "accept": false, + "reason": "Tokenizer special-token method conflict and gguf_file from_pretrained handling are separate loading bugs." + }, + { + "left": "issue:43994", + "right": "issue:44079", + "accept": false, + "reason": "SigLIP2 nonsensical outputs and ModelOutput key assignment with None are different layers of the stack." + }, + { + "left": "issue:43901", + "right": "issue:44509", + "accept": false, + "reason": "Both touch docs, but they cover different stale documentation problems and different removed/changed API surface." + }, + { + "left": "issue:44265", + "right": "issue:44898", + "accept": false, + "reason": "torch.export failures with torch_compilable_check are not the same as Perceiver non-default-resolution interpolation failures." + }, + { + "left": "issue:44991", + "right": "issue:45081", + "accept": false, + "reason": "Different tokenizer/loading bugs: EMBEDDIA/est-roberta vs. Mistral regex patching backend_tokenizer." + }, + { + "left": "issue:46050", + "right": "issue:46326", + "accept": false, + "reason": "Quantized save/load corruption is unrelated to continue_final_message behavior for mistral_common tokenizers." + }, + { + "left": "issue:43906", + "right": "issue:46326", + "accept": false, + "reason": "An isolated reproduction of an older issue does not match the specific continue_final_message tokenizer bug." + }, + { + "left": "issue:44671", + "right": "issue:44991", + "accept": false, + "reason": "CamemBERT masked LM prediction regression is unrelated to tokenizer loading for EMBEDDIA/est-roberta." + }, + { + "left": "issue:43906", + "right": "issue:44530", + "accept": false, + "reason": "No shared concrete bug: one is a reproduction issue, the other is a Qwen3.5 cache crash on linear_attention group type." + }, + { + "left": "issue:39404", + "right": "issue:44912", + "accept": false, + "reason": "Whisper return_language and MXFP4 quantization fallback are separate regressions." + }, + { + "left": "issue:45823", + "right": "issue:46899", + "accept": false, + "reason": "Both involve Gemma4 multimodal paths, but one is device-map placement and the other is an unsafe dtype cast in 4-bit embedders." + }, + { + "left": "issue:43596", + "right": "issue:44683", + "accept": false, + "reason": "ZeRO3/BertModel init IndexError and flex_attention compile failure on torch>=2.9 are unrelated failures." + }, + { + "left": "issue:44756", + "right": "issue:47333", + "accept": false, + "reason": "mmap OOM on Strix Halo and disk-offloaded Qwen3 save failure are different memory/offload issues." + }, + { + "left": "issue:44164", + "right": "issue:46821", + "accept": false, + "reason": "extra_state save/load handling is unrelated to the TimesFM window_size AttributeError." + }, + { + "left": "issue:43856", + "right": "issue:44756", + "accept": false, + "reason": "Qwen3 MoE training memory inefficiency is not the same as disabling mmap to avoid OOM on Strix Halo." + }, + { + "left": "issue:44488", + "right": "issue:44987", + "accept": false, + "reason": "Both are load failures, but they affect different models and the titles point to different root causes." + }, + { + "left": "issue:44488", + "right": "issue:45081", + "accept": false, + "reason": "Different model-loading problems: cjvt/sleng-bert vs. the Mistral regex/backend_tokenizer crash." + }, + { + "left": "issue:36032", + "right": "issue:42915", + "accept": false, + "reason": "Tokenizer add_special_tokens conflict and Qwen3Moe FineGrainedFP8Config failure are unrelated." + }, + { + "left": "issue:43632", + "right": "issue:47050", + "accept": false, + "reason": "_is_hf_initialized flag regression and ONNX attention dtype mismatch are separate bugs." + }, + { + "left": "issue:43716", + "right": "issue:45084", + "accept": false, + "reason": "Mistral-3 dtype mismatch in preprocessing/modeling is unrelated to the template-node compilation error." + }, + { + "left": "issue:43299", + "right": "issue:43931", + "accept": false, + "reason": "Same model family, but different concrete problems: Qwen3VL Moe loading break vs. weight-shape mismatch on a specific checkpoint." + }, + { + "left": "issue:43746", + "right": "issue:47331", + "accept": false, + "reason": "GraniteSpeech PEFT local checkpoint loading and MetalConfig batched quantization corruption are unrelated." + }, + { + "left": "issue:43611", + "right": "issue:43644", + "accept": false, + "reason": "base_model_prefix loading regression and non-persistent buffers being filled with junk are different code paths." + }, + { + "left": "issue:42796", + "right": "issue:43404", + "accept": false, + "reason": "Tokenizer BPE-marker handling and lm_head weight tying in Mistral3 are different underlying issues." + } + ] + }, + "evaluator_result": null, + "evaluator_used": false, + "retried": false, + "accepted_nontrivial_soft_edge": false, + "error_kind": null, + "error_message": null + }, + { + "label": "issue", + "component_index": 1, + "component_count": 28, + "review_unit_index": 11, + "review_unit_count": 65, + "status": "reviewed", + "reason": null, + "source": "llm", + "cache_hit": false, + "model": "gpt-5.4-mini?service_tier=flex", + "cluster_id": "cluster-10000-530", + "nodes": [ + "issue:34567", + "issue:36032", + "issue:39404", + "issue:40444", + "issue:40990", + "issue:42175", + "issue:42915", + "issue:43262", + "issue:43421", + "issue:43452", + "issue:43482", + "issue:43541", + "issue:43582", + "issue:43632", + "issue:43643", + "issue:43644", + "issue:43645", + "issue:43646", + "issue:43746", + "issue:43873", + "issue:44206", + "issue:44220", + "issue:44403", + "issue:44451", + "issue:44462", + "issue:44530", + "issue:44661", + "issue:44671", + "issue:44683", + "issue:44737", + "issue:44811", + "issue:44987", + "issue:44991", + "issue:45081", + "issue:45084", + "issue:45479", + "issue:45507", + "issue:45584", + "issue:45859", + "issue:46178", + "issue:46285", + "issue:46326", + "issue:46402", + "issue:46489", + "issue:46894", + "issue:46897", + "issue:47312", + "issue:47331" + ], + "soft_pairs": [ + "issue:45479|issue:46894", + "issue:44462|issue:46489", + "issue:44661|issue:44683", + "issue:46285|issue:46402", + "issue:43541|issue:45084", + "issue:45479|issue:46897", + "issue:44451|issue:44987", + "issue:42915|issue:44530", + "issue:43873|issue:47331", + "issue:43643|issue:44530", + "issue:40990|issue:43421", + "issue:36032|issue:43645", + "issue:40444|issue:46326", + "issue:39404|issue:44671", + "issue:34567|issue:40444", + "issue:44737|issue:45859", + "issue:43644|issue:43646", + "issue:42175|issue:46178", + "issue:44987|issue:44991", + "issue:44206|issue:44220", + "issue:45084|issue:45507", + "issue:43582|issue:43632", + "issue:43452|issue:43482", + "issue:43482|issue:47312", + "issue:43582|issue:47331", + "issue:43746|issue:46326", + "issue:44811|issue:45584", + "issue:44403|issue:45081", + "issue:43262|issue:47331" + ], + "prepared_review_unit_hash": "3cc426d2001703671d4aa90f7a34d07d42a68edbf8f86986ae81cbf9670dee5d", + "component_budget": { + "node_count": 530, + "item_count": 530, + "soft_pair_count": 1954, + "serialized_chars": 1013004, + "estimated_input_tokens": 253251, + "estimated_eval_tokens": 506758 + }, + "budget": { + "node_count": 48, + "item_count": 48, + "soft_pair_count": 29, + "serialized_chars": 36482, + "estimated_input_tokens": 9121, + "estimated_eval_tokens": 18498 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": true, + "analyst_result": { + "summary": "This cluster is mostly a noisy collection of unrelated Transformers issues that happen to share broad v5-regression vocabulary; none of the soft pairs look like true duplicates of the same bug/change.", + "confidence": 0.87, + "canonical_issue_reason": "If a single issue must represent the cluster, issue 43646 is the broadest and most central v5 regression report here (custom model initialization). That said, the set is heterogeneous, so it is only a loose representative, not a true duplicate umbrella.", + "canonical_pr_reason": null, + "best_issue_reason": "Issue 43646 is the best single anchor because it is concise, clearly states a concrete regression, and is broad enough to be a representative reference for nearby v5 initialization complaints.", + "best_pr_reason": null, + "soft_edge_verdicts": [ + { + "left": "issue:45479", + "right": "issue:46894", + "accept": false, + "reason": "Different model families and different manifestations of the training-loss double-shift bug; related pattern, not the same underlying issue." + }, + { + "left": "issue:44462", + "right": "issue:46489", + "accept": false, + "reason": "Both concern tokenizer loading, but one is about ignoring tokenizer.json and the other is a DeepSeek-Coder-specific tokenizer regression from a prior fix gap." + }, + { + "left": "issue:44661", + "right": "issue:44683", + "accept": false, + "reason": "One is a tokenizer mapping/add-new-model failure, the other is a torch flex_attention compilation issue; separate code paths." + }, + { + "left": "issue:46285", + "right": "issue:46402", + "accept": false, + "reason": "Different bugs: class structure inconsistency vs text encoder weights not being loaded." + }, + { + "left": "issue:43541", + "right": "issue:45084", + "accept": false, + "reason": "Unrelated failures in torch tracing/grouped_mm versus Jinja template compilation." + }, + { + "left": "issue:45479", + "right": "issue:46897", + "accept": false, + "reason": "Same general double-shift pattern, but different affected model and separate issue reports." + }, + { + "left": "issue:44451", + "right": "issue:44987", + "accept": false, + "reason": "Both are load failures, but on different models with different breakages and no clear shared bug." + }, + { + "left": "issue:42915", + "right": "issue:44530", + "accept": false, + "reason": "Qwen3Moe FP8 config failure and Qwen3.5 PagedAttentionCache invalid group type are distinct runtime problems." + }, + { + "left": "issue:43873", + "right": "issue:47331", + "accept": false, + "reason": "Quantization/offloading difficulty versus a specific MetalConfig batched-generate corruption bug; not the same code path." + }, + { + "left": "issue:43643", + "right": "issue:44530", + "accept": false, + "reason": "AutoConfig trust_remote_code missing fields is unrelated to Qwen3.5 cache/group-type crashes." + }, + { + "left": "issue:40990", + "right": "issue:43421", + "accept": false, + "reason": "Perplexity regression on gpt-oss is not the same as runtime post-processor updates for tokenizer special tokens." + }, + { + "left": "issue:36032", + "right": "issue:43645", + "accept": false, + "reason": "Tokenizer method conflict during load is unrelated to custom model initialization in notebooks." + }, + { + "left": "issue:40444", + "right": "issue:46326", + "accept": false, + "reason": "Multi-image IterableDataset fine-tuning and mistral_common continue_final_message handling are different bugs." + }, + { + "left": "issue:39404", + "right": "issue:44671", + "accept": false, + "reason": "Whisper pipeline return_language regression and CamemBERT masked-LM predictions are unrelated." + }, + { + "left": "issue:34567", + "right": "issue:40444", + "accept": false, + "reason": "TrainerState token counter not updating is unrelated to Qwen2.5-VL IterableDataset multi-image failures." + }, + { + "left": "issue:44737", + "right": "issue:45859", + "accept": false, + "reason": "CPU arange device bug in XLNet and CPU-offload incompatibility in Qwen3_5Moe are different implementation issues." + }, + { + "left": "issue:43644", + "right": "issue:43646", + "accept": false, + "reason": "Both are v5 regressions, but one is about non-persistent buffers and the other about custom model initialization; not the same bug." + }, + { + "left": "issue:42175", + "right": "issue:46178", + "accept": false, + "reason": "Package/backend dependency inclusion and backend mapping validation are distinct installation/backends issues." + }, + { + "left": "issue:44987", + "right": "issue:44991", + "accept": false, + "reason": "Different model-loading failures with different symptoms; no evidence they share one underlying tokenizer bug." + }, + { + "left": "issue:44206", + "right": "issue:44220", + "accept": false, + "reason": "Unsupported center arg in feature extractor versus _torch_extract_fbank_features() behavior; related audio area but not the same failure." + }, + { + "left": "issue:45084", + "right": "issue:45507", + "accept": false, + "reason": "Template compilation error and GraniteMoEHybrid invalid method calls are separate problems." + }, + { + "left": "issue:43582", + "right": "issue:43632", + "accept": false, + "reason": "Apple Silicon allocator warmup TypeError is unrelated to the _is_hf_initialized flag regression." + }, + { + "left": "issue:43452", + "right": "issue:43482", + "accept": false, + "reason": "GGUF loading breakage and Qwen2.5-GGUF loading failure are different model/loading issues without a shared concrete fix." + }, + { + "left": "issue:43482", + "right": "issue:47312", + "accept": false, + "reason": "Qwen2.5-GGUF v5 loading failure is unrelated to kernels 0.16 incompatibility." + }, + { + "left": "issue:43582", + "right": "issue:47331", + "accept": false, + "reason": "Allocator warmup TypeError and MetalConfig batched-generate corruption are different bugs and code paths." + }, + { + "left": "issue:43746", + "right": "issue:46326", + "accept": false, + "reason": "PEFT-adapter local checkpoint loading for GraniteSpeech is unrelated to mistral_common continue_final_message behavior." + }, + { + "left": "issue:44811", + "right": "issue:45584", + "accept": false, + "reason": "Whisper batch_decode skip_special_tokens handling and empty-transcription generation after align_special_tokens are different issues." + }, + { + "left": "issue:44403", + "right": "issue:45081", + "accept": false, + "reason": "Loading noise vs Mistral regex patch backend_tokenizer crash are not the same underlying defect." + }, + { + "left": "issue:43262", + "right": "issue:47331", + "accept": false, + "reason": "Audio processor chat-template sampling-rate default and Metal batched-generate quantization corruption are unrelated." + } + ] + }, + "evaluator_result": null, + "evaluator_used": false, + "retried": false, + "accepted_nontrivial_soft_edge": false, + "error_kind": null, + "error_message": null + }, + { + "label": "issue", + "component_index": 1, + "component_count": 28, + "review_unit_index": 12, + "review_unit_count": 65, + "status": "reviewed", + "reason": null, + "source": "llm", + "cache_hit": false, + "model": "gpt-5.4-mini?service_tier=flex", + "cluster_id": "cluster-10000-530", + "nodes": [ + "issue:28530", + "issue:34567", + "issue:36032", + "issue:41720", + "issue:42915", + "issue:43208", + "issue:43262", + "issue:43317", + "issue:43322", + "issue:43377", + "issue:43404", + "issue:43425", + "issue:43482", + "issue:43531", + "issue:43540", + "issue:43632", + "issue:43645", + "issue:43646", + "issue:43742", + "issue:43792", + "issue:43819", + "issue:43994", + "issue:44008", + "issue:44062", + "issue:44265", + "issue:44291", + "issue:44403", + "issue:44451", + "issue:44464", + "issue:44466", + "issue:44479", + "issue:44534", + "issue:44589", + "issue:44610", + "issue:44849", + "issue:44877", + "issue:44987", + "issue:45081", + "issue:45084", + "issue:45325", + "issue:45923", + "issue:45987", + "issue:46661", + "issue:46821", + "issue:46869", + "issue:46899", + "issue:47013", + "issue:47254" + ], + "soft_pairs": [ + "issue:43317|issue:43482", + "issue:43208|issue:47013", + "issue:43645|issue:46661", + "issue:43792|issue:46869", + "issue:42915|issue:45084", + "issue:43646|issue:44534", + "issue:44987|issue:45081", + "issue:42915|issue:44291", + "issue:43262|issue:43377", + "issue:36032|issue:43742", + "issue:28530|issue:34567", + "issue:43377|issue:43819", + "issue:44062|issue:44987", + "issue:44451|issue:45081", + "issue:43404|issue:44466", + "issue:43994|issue:44610", + "issue:42915|issue:44589", + "issue:44849|issue:45325", + "issue:36032|issue:41720", + "issue:43531|issue:44877", + "issue:44265|issue:44610", + "issue:44008|issue:46899", + "issue:44479|issue:45325", + "issue:43322|issue:43540", + "issue:45923|issue:47254", + "issue:44464|issue:45987", + "issue:43425|issue:44403", + "issue:43632|issue:46821" + ], + "prepared_review_unit_hash": "f363cab83bac1eb1b8153ad2829f08ef154aaa05c3d4be217d30f2ab65ee6607", + "component_budget": { + "node_count": 530, + "item_count": 530, + "soft_pair_count": 1954, + "serialized_chars": 1013004, + "estimated_input_tokens": 253251, + "estimated_eval_tokens": 506758 + }, + "budget": { + "node_count": 48, + "item_count": 48, + "soft_pair_count": 28, + "serialized_chars": 36256, + "estimated_input_tokens": 9064, + "estimated_eval_tokens": 18384 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": true, + "analyst_result": { + "summary": "This is not a duplicate cluster: the items are a broad mix of unrelated Transformers bugs across training, tokenization, multimodal processing, device mapping, quantization, and serialization. All soft pairs look like loose topical similarity rather than the same underlying issue.", + "confidence": 0.97, + "canonical_issue_reason": "No canonical issue stands out because the set is heterogeneous and does not share one concrete code-path failure.", + "canonical_pr_reason": null, + "best_issue_reason": "No issue is globally suitable as a canonical representative; choosing any one would be arbitrary given the lack of a common bug theme.", + "best_pr_reason": null, + "soft_edge_verdicts": [ + { + "left": "issue:43317", + "right": "issue:43482", + "accept": false, + "reason": "Both involve loading/dequantization behavior, but one is device_map offload and the other is GGUF/v5 loading; different failure modes and code paths." + }, + { + "left": "issue:43208", + "right": "issue:47013", + "accept": false, + "reason": "Both mention xLSTM, but one is a trio of training bugs for smaller-than-7B models and the other is a specific return_last_states unpacking crash." + }, + { + "left": "issue:43645", + "right": "issue:46661", + "accept": false, + "reason": "Both concern custom/remote-code model initialization under v5, but one is a Jupyter-specific notebook issue and the other is an AutoModel config_class error for remote-code models." + }, + { + "left": "issue:43792", + "right": "issue:46869", + "accept": false, + "reason": "Both mention model loading/runtime failure, but the first is Whisper not running and the second is YOLO failing with multi-GPU device_map auto; unrelated." + }, + { + "left": "issue:42915", + "right": "issue:45084", + "accept": false, + "reason": "Qwen3MoE FP8 failure and a chat-template compilation TypeError are unrelated subsystems." + }, + { + "left": "issue:43646", + "right": "issue:44534", + "accept": false, + "reason": "One is custom model initialization breaking in v5, the other is non-persistent buffer serialization corruption; different bugs." + }, + { + "left": "issue:44987", + "right": "issue:45081", + "accept": false, + "reason": "Both are loading failures, but one is a model-specific transformers>=5.1 regression and the other is a Mistral tokenizer regex patch crash." + }, + { + "left": "issue:42915", + "right": "issue:44291", + "accept": false, + "reason": "Both reference v5-era model loading, but one is Qwen3Moe FP8 config failure and the other is init_empty_weights/_is_hf_initialized argument handling." + }, + { + "left": "issue:43262", + "right": "issue:43377", + "accept": false, + "reason": "Audio processor sampling-rate defaulting and MIMI encoder padding-mask behavior are distinct multimodal bugs." + }, + { + "left": "issue:36032", + "right": "issue:43742", + "accept": false, + "reason": "Tokenizer special-token conflict and MobileLLM key error are unrelated model-loading problems." + }, + { + "left": "issue:28530", + "right": "issue:34567", + "accept": false, + "reason": "Early stopping metric selection and num_input_tokens_seen tracking are separate Trainer features." + }, + { + "left": "issue:43377", + "right": "issue:43819", + "accept": false, + "reason": "MIMI encoder batching/padding bug and DAC from_latents/forward mismatch are different audio-model correctness issues." + }, + { + "left": "issue:44062", + "right": "issue:44987", + "accept": false, + "reason": "AddedToken keyword collision and a specific model-loading regression are unrelated tokenizer vs model-loading failures." + }, + { + "left": "issue:44451", + "right": "issue:45081", + "accept": false, + "reason": "ScandiBERT loading failure and Mistral regex patch crash are different tokenizer/model issues." + }, + { + "left": "issue:43404", + "right": "issue:44466", + "accept": false, + "reason": "lm_head weight tying in Mistral3 and device-dependent lm_head serialization are related only broadly by tied weights, but they are not the same concrete bug." + }, + { + "left": "issue:43994", + "right": "issue:44610", + "accept": false, + "reason": "SigLIP2 nonsensical outputs and OmDet-Turbo processor size mismatch are separate vision-model bugs." + }, + { + "left": "issue:42915", + "right": "issue:44589", + "accept": false, + "reason": "Qwen3Moe FP8 config failure and Float8 storage lookup failure are different quantization/runtime issues." + }, + { + "left": "issue:44849", + "right": "issue:45325", + "accept": false, + "reason": "Qwen3.5 output_hidden_states behavior and Qwen2.5-VL rope-index scaling are different model-specific logic bugs." + }, + { + "left": "issue:36032", + "right": "issue:41720", + "accept": false, + "reason": "Tokenizer method-name conflict and Qwen3 auto-device-map cuda assert are unrelated." + }, + { + "left": "issue:43531", + "right": "issue:44877", + "accept": false, + "reason": "Qwen3-MoE sliding_window behavior and strict config loading for granite_speech are different configuration bugs." + }, + { + "left": "issue:44265", + "right": "issue:44610", + "accept": false, + "reason": "torch.export/torch_compilable_check failure and OmDet-Turbo image size mismatch are unrelated export vs preprocessing issues." + }, + { + "left": "issue:44008", + "right": "issue:46899", + "accept": false, + "reason": "Gemma 3n variable-name collision and Gemma4 unsafe dtype cast are different multimodal model bugs." + }, + { + "left": "issue:44479", + "right": "issue:45325", + "accept": false, + "reason": "Both touch Qwen VL models, but one is a video-input regression and the other is still-image rope indexing; not the same bug." + }, + { + "left": "issue:43322", + "right": "issue:43540", + "accept": false, + "reason": "Llava Next segfault and Qwen3OmniMoe video-input ValueError are distinct multimodal loading/processing failures." + }, + { + "left": "issue:45923", + "right": "issue:47254", + "accept": false, + "reason": "Missing supports_gradient_checkpointing on a remote-code model is unrelated to the PEFT/DeepSpeed checkpoint error." + }, + { + "left": "issue:44464", + "right": "issue:45987", + "accept": false, + "reason": "Chunked generation inconsistency with compiled forward and StaticCache.get_seq_length() returning the wrong type are separate generation/cache bugs." + }, + { + "left": "issue:43425", + "right": "issue:44403", + "accept": false, + "reason": "Torch 2.10 incompatibility and loading noise are not the same underlying problem." + }, + { + "left": "issue:43632", + "right": "issue:46821", + "accept": false, + "reason": "_is_hf_initialized flag breakage and TimesFM window_size AttributeError are unrelated v5 compatibility issues." + } + ] + }, + "evaluator_result": null, + "evaluator_used": false, + "retried": false, + "accepted_nontrivial_soft_edge": false, + "error_kind": null, + "error_message": null + }, + { + "label": "issue", + "component_index": 1, + "component_count": 28, + "review_unit_index": 13, + "review_unit_count": 65, + "status": "reviewed", + "reason": null, + "source": "llm", + "cache_hit": false, + "model": "gpt-5.4-mini?service_tier=flex", + "cluster_id": "cluster-10000-530", + "nodes": [ + "issue:33666", + "issue:35707", + "issue:40444", + "issue:40990", + "issue:41720", + "issue:42175", + "issue:42796", + "issue:43089", + "issue:43296", + "issue:43381", + "issue:43421", + "issue:43482", + "issue:43525", + "issue:43531", + "issue:43611", + "issue:43632", + "issue:43716", + "issue:43746", + "issue:43792", + "issue:43872", + "issue:43883", + "issue:44079", + "issue:44220", + "issue:44291", + "issue:44403", + "issue:44462", + "issue:44488", + "issue:44534", + "issue:44568", + "issue:44589", + "issue:44683", + "issue:44756", + "issue:44811", + "issue:44991", + "issue:45081", + "issue:45216", + "issue:45325", + "issue:45357", + "issue:45542", + "issue:45663", + "issue:45812", + "issue:45923", + "issue:45987", + "issue:46315", + "issue:46491", + "issue:46710", + "issue:46821", + "issue:46899" + ], + "soft_pairs": [ + "issue:43089|issue:45987", + "issue:43296|issue:43531", + "issue:40444|issue:45325", + "issue:44220|issue:44683", + "issue:43872|issue:44291", + "issue:43883|issue:44488", + "issue:43611|issue:44534", + "issue:44991|issue:46315", + "issue:45216|issue:45357", + "issue:43716|issue:46899", + "issue:44403|issue:44991", + "issue:42796|issue:43746", + "issue:44291|issue:44589", + "issue:45812|issue:46491", + "issue:43381|issue:45923", + "issue:43792|issue:44589", + "issue:33666|issue:40444", + "issue:40990|issue:44811", + "issue:43632|issue:44756", + "issue:42796|issue:46710", + "issue:43531|issue:44291", + "issue:35707|issue:40444", + "issue:44079|issue:45663", + "issue:43883|issue:44991", + "issue:41720|issue:45081", + "issue:43482|issue:43525", + "issue:43421|issue:44568", + "issue:43296|issue:46821", + "issue:42175|issue:43296", + "issue:42175|issue:45542", + "issue:44462|issue:45812" + ], + "prepared_review_unit_hash": "9eabcd0cfefb81fefa10e7a47cebd42e258a5f082d6a7ded1550a14bd5005ab9", + "component_budget": { + "node_count": 530, + "item_count": 530, + "soft_pair_count": 1954, + "serialized_chars": 1013004, + "estimated_input_tokens": 253251, + "estimated_eval_tokens": 506758 + }, + "budget": { + "node_count": 48, + "item_count": 48, + "soft_pair_count": 31, + "serialized_chars": 37128, + "estimated_input_tokens": 9282, + "estimated_eval_tokens": 18820 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": true, + "analyst_result": { + "summary": "Most items are unrelated issue reports from different models/code paths. I found only two soft-duplicate pairs that look like the same underlying bug: the transformers v5 `_is_hf_initialized`/bitsandbytes loading regression, and the Qwen3.5 `save_pretrained` checkpoint regression.", + "confidence": 0.84, + "canonical_issue_reason": "issue:43872 is the strongest representative: it has the clearest, broadly impactful regression title, high discussion/activity, and inbound references, and it matches a concrete v5 compatibility failure that also appears in a narrower follow-up issue.", + "canonical_pr_reason": null, + "best_issue_reason": "issue:43872 is the best overall issue to keep as the canonical anchor because it is concrete, high-signal, and central to one of the accepted duplicate pairs.", + "best_pr_reason": null, + "soft_edge_verdicts": [ + { + "left": "issue:43089", + "right": "issue:45987", + "accept": false, + "reason": "Different problems: GPU sync/dispatch overhead vs a StaticCache return-type contract bug." + }, + { + "left": "issue:43296", + "right": "issue:43531", + "accept": false, + "reason": "Different models and failures: PaddleOCR-VL load failure vs Qwen3-MoE sliding_window behavior." + }, + { + "left": "issue:40444", + "right": "issue:45325", + "accept": false, + "reason": "Both are Qwen2.5-VL related, but one is multi-image finetuning input handling and the other is ROPE/position_id scaling." + }, + { + "left": "issue:44220", + "right": "issue:44683", + "accept": false, + "reason": "Unrelated subsystems: fbank feature extraction vs compiled flex_attention on newer torch." + }, + { + "left": "issue:43872", + "right": "issue:44291", + "accept": true, + "reason": "Same underlying transformers v5 loading regression around `_is_hf_initialized`, producing the same bitsandbytes TypeError path." + }, + { + "left": "issue:43883", + "right": "issue:44488", + "accept": false, + "reason": "Different model-loading failures with different missing/ignored tokenizer/model assets." + }, + { + "left": "issue:43611", + "right": "issue:44534", + "accept": false, + "reason": "Both are v5 regressions, but they affect different loader internals and different symptoms." + }, + { + "left": "issue:44991", + "right": "issue:46315", + "accept": false, + "reason": "Tokenizer loading failure vs a performance regression in `convert_tokens_to_ids`; different bug classes." + }, + { + "left": "issue:45216", + "right": "issue:45357", + "accept": true, + "reason": "Same Qwen3.5 `save_pretrained` regression; 45357 is a narrower follow-up about incorrect visual encoder keys." + }, + { + "left": "issue:43716", + "right": "issue:46899", + "accept": false, + "reason": "Both are dtype-related multimodal issues, but they affect different models, modalities, and code paths." + }, + { + "left": "issue:44403", + "right": "issue:44991", + "accept": false, + "reason": "Noise/logging during load is not the same as tokenizer repository resolution failure." + }, + { + "left": "issue:42796", + "right": "issue:43746", + "accept": false, + "reason": "Tokenizer BPE marker handling is unrelated to PEFT checkpoint loading." + }, + { + "left": "issue:44291", + "right": "issue:44589", + "accept": false, + "reason": "Different regressions: unexpected `_is_hf_initialized` kwarg vs missing Float8 storage type." + }, + { + "left": "issue:45812", + "right": "issue:46491", + "accept": false, + "reason": "Different tokenizer regressions affecting different models and different behaviors." + }, + { + "left": "issue:43381", + "right": "issue:45923", + "accept": false, + "reason": "Eval-mode gradient checkpointing restriction vs a missing `supports_gradient_checkpointing` flag on a trust_remote_code model." + }, + { + "left": "issue:43792", + "right": "issue:44589", + "accept": false, + "reason": "Whisper runtime failure is not the same as a Float8 storage lookup error." + }, + { + "left": "issue:33666", + "right": "issue:40444", + "accept": false, + "reason": "Both involve training, but they are distinct Qwen-VL issues with different failure modes." + }, + { + "left": "issue:40990", + "right": "issue:44811", + "accept": false, + "reason": "Perplexity on a language model dataset is unrelated to Whisper `batch_decode` special-token handling." + }, + { + "left": "issue:43632", + "right": "issue:44756", + "accept": false, + "reason": "`_is_hf_initialized` flag regression and Strix Halo mmap OOM are unrelated." + }, + { + "left": "issue:42796", + "right": "issue:46710", + "accept": false, + "reason": "Tokenizer marker handling vs accuracy regression after version upgrade; different symptoms and likely different causes." + }, + { + "left": "issue:43531", + "right": "issue:44291", + "accept": false, + "reason": "Qwen3-MoE sliding_window behavior is unrelated to the `_is_hf_initialized` loading error." + }, + { + "left": "issue:35707", + "right": "issue:40444", + "accept": false, + "reason": "Progressive generation with cache inputs is unrelated to Qwen2.5-VL multi-image finetuning failure." + }, + { + "left": "issue:44079", + "right": "issue:45663", + "accept": false, + "reason": "ModelOutput key assignment bug vs Gemma4 FSDP2 shared_kv_states issue." + }, + { + "left": "issue:43883", + "right": "issue:44991", + "accept": false, + "reason": "Missing model attribute error is unrelated to tokenizer repository loading." + }, + { + "left": "issue:41720", + "right": "issue:45081", + "accept": false, + "reason": "Auto device mapping CUDA assert is unrelated to the Mistral regex patch crash." + }, + { + "left": "issue:43482", + "right": "issue:43525", + "accept": false, + "reason": "Qwen2.5-GGUF loading and Llama4Config missing `pad_token_id` are different model/config issues." + }, + { + "left": "issue:43421", + "right": "issue:44568", + "accept": false, + "reason": "Runtime special-token post-processing update bug is different from BOS/EOS not being added on encode." + }, + { + "left": "issue:43296", + "right": "issue:46821", + "accept": false, + "reason": "PaddleOCR-VL load failure vs `window_size` AttributeError in TimesFM 2.5 are unrelated." + }, + { + "left": "issue:42175", + "right": "issue:43296", + "accept": false, + "reason": "Packaging/backend dependency issue is unrelated to model loading failure in vLLM." + }, + { + "left": "issue:42175", + "right": "issue:45542", + "accept": false, + "reason": "Both mention TensorFlow backend, but one is an install/extra dependency issue and the other is a runtime backend detection error." + }, + { + "left": "issue:44462", + "right": "issue:45812", + "accept": false, + "reason": "Repository tokenizer.json being ignored is a different tokenizer loading bug than Granite token ID regression." + } + ] + }, + "evaluator_result": { + "accept": true, + "feedback": "The summary is grounded in the packet and stays conservative. The two accepted soft-duplicate pairs are plausibly the same underlying bug based on the titles, and the rest of the verdicts read as careful non-duplicates rather than overclaims. Canonical issue choice is also reasonable from the available metadata." + }, + "evaluator_used": true, + "retried": false, + "accepted_nontrivial_soft_edge": true, + "error_kind": null, + "error_message": null + }, + { + "label": "issue", + "component_index": 1, + "component_count": 28, + "review_unit_index": 14, + "review_unit_count": 65, + "status": "reviewed", + "reason": null, + "source": "llm", + "cache_hit": false, + "model": "gpt-5.4-mini?service_tier=flex", + "cluster_id": "cluster-10000-530", + "nodes": [ + "issue:34634", + "issue:35707", + "issue:40444", + "issue:40990", + "issue:41720", + "issue:42796", + "issue:43262", + "issue:43317", + "issue:43377", + "issue:43454", + "issue:43482", + "issue:43493", + "issue:43606", + "issue:43611", + "issue:43644", + "issue:43646", + "issue:43701", + "issue:43746", + "issue:43818", + "issue:43828", + "issue:43856", + "issue:43873", + "issue:43994", + "issue:44079", + "issue:44186", + "issue:44265", + "issue:44291", + "issue:44466", + "issue:44479", + "issue:44509", + "issue:44534", + "issue:44829", + "issue:44877", + "issue:45072", + "issue:45325", + "issue:45357", + "issue:45381", + "issue:45479", + "issue:45491", + "issue:45538", + "issue:45584", + "issue:45701", + "issue:45823", + "issue:45834", + "issue:46402", + "issue:46821", + "issue:46909", + "issue:47217" + ], + "soft_pairs": [ + "issue:43317|issue:43856", + "issue:44079|issue:44534", + "issue:43482|issue:45538", + "issue:43611|issue:43646", + "issue:44509|issue:46909", + "issue:44265|issue:45072", + "issue:45325|issue:45834", + "issue:43377|issue:45491", + "issue:40444|issue:45381", + "issue:43644|issue:44079", + "issue:44291|issue:44877", + "issue:45325|issue:45381", + "issue:43493|issue:43994", + "issue:43818|issue:46402", + "issue:40444|issue:47217", + "issue:43454|issue:44466", + "issue:43818|issue:45357", + "issue:43873|issue:45701", + "issue:43701|issue:45325", + "issue:44829|issue:45479", + "issue:43606|issue:44265", + "issue:42796|issue:45381", + "issue:44479|issue:45823", + "issue:43828|issue:46821", + "issue:40990|issue:45584", + "issue:43262|issue:43746", + "issue:34634|issue:35707", + "issue:35707|issue:41720", + "issue:41720|issue:45823", + "issue:44186|issue:44265" + ], + "prepared_review_unit_hash": "7ded0d42bc09e44c2b70be4028d3e37df1103dbf1837f7942ebdcc57c66bc826", + "component_budget": { + "node_count": 530, + "item_count": 530, + "soft_pair_count": 1954, + "serialized_chars": 1013004, + "estimated_input_tokens": 253251, + "estimated_eval_tokens": 506758 + }, + "budget": { + "node_count": 48, + "item_count": 48, + "soft_pair_count": 30, + "serialized_chars": 36885, + "estimated_input_tokens": 9222, + "estimated_eval_tokens": 18700 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": true, + "analyst_result": { + "summary": "This set is mostly a loose collection of Transformers regressions and model-specific bugs; none of the soft pairs is close enough to merge from titles alone. The most representative active issue is the broad offloading/quantization tracker.", + "confidence": 0.61, + "canonical_issue_reason": "issue:43873 is the broadest active issue here and best matches the cluster\u2019s recurring offloading/device-map symptoms; it has the most tracker-like shape and inbound references.", + "canonical_pr_reason": null, + "best_issue_reason": "issue:43873 is the best global representative because it is open, actively discussed, and broad enough to serve as the landing issue for related offloading/quantization reports.", + "best_pr_reason": null, + "soft_edge_verdicts": [ + { + "left": "issue:43317", + "right": "issue:43856", + "accept": false, + "reason": "Different failures: CPU/GPU offload loading error versus MoE training memory inefficiency. Same general area, not the same bug." + }, + { + "left": "issue:44079", + "right": "issue:44534", + "accept": false, + "reason": "ModelOutput key assignment and non-persistent buffer initialization are different code paths and symptoms." + }, + { + "left": "issue:43482", + "right": "issue:45538", + "accept": false, + "reason": "GGUF loading failure with v5 is unrelated to CLIPTokenizer max-length handling." + }, + { + "left": "issue:43611", + "right": "issue:43646", + "accept": false, + "reason": "Both are v5 regressions, but one is about base_model_prefix loading and the other about custom model initialization; not the same underlying change." + }, + { + "left": "issue:44509", + "right": "issue:46909", + "accept": false, + "reason": "Both are docs issues, but they concern different documentation problems and examples." + }, + { + "left": "issue:44265", + "right": "issue:45072", + "accept": false, + "reason": "torch.export failure with torch_compilable_check is a different bug from dtype mismatches in bfloat16 inference." + }, + { + "left": "issue:45325", + "right": "issue:45834", + "accept": false, + "reason": "Qwen2.5-VL rope-index scaling and Kosmos2.5 long-OCR index errors are different model-specific failures." + }, + { + "left": "issue:43377", + "right": "issue:45491", + "accept": false, + "reason": "Missing padding-mask support in MIMI is unrelated to Gemma3 sliding-window NaN embeddings." + }, + { + "left": "issue:40444", + "right": "issue:45381", + "accept": false, + "reason": "Both involve Qwen VL inputs, but one is multiple images in an iterable dataset and the other is video vision_position_ids; not the same defect." + }, + { + "left": "issue:43644", + "right": "issue:44079", + "accept": false, + "reason": "Non-persistent buffer serialization junk and ModelOutput key assignment are separate internal issues." + }, + { + "left": "issue:44291", + "right": "issue:44877", + "accept": false, + "reason": "init_empty_weights argument handling and strict config loading are different regressions." + }, + { + "left": "issue:45325", + "right": "issue:45381", + "accept": false, + "reason": "Both are Qwen2.5-VL position-id bugs, but one is still-image temporal scaling and the other is video input positions." + }, + { + "left": "issue:43493", + "right": "issue:43994", + "accept": false, + "reason": "Same model family, but the issues describe different symptoms and no shared concrete code-path evidence." + }, + { + "left": "issue:43818", + "right": "issue:46402", + "accept": false, + "reason": "Video-LLaVA tower/weight structure issue is unrelated to Sam3Model text encoder loading." + }, + { + "left": "issue:40444", + "right": "issue:47217", + "accept": false, + "reason": "Both are VL input-handling problems, but one is multi-image finetuning and the other is placeholder placement in user text." + }, + { + "left": "issue:43454", + "right": "issue:44466", + "accept": false, + "reason": "AyaVision weight-tying failure and inconsistent lm_head serialization are different tie/serialization bugs." + }, + { + "left": "issue:43818", + "right": "issue:45357", + "accept": false, + "reason": "Video-LLaVA architecture mismatch and Qwen3.5 visual-encoder key saving are unrelated." + }, + { + "left": "issue:43873", + "right": "issue:45701", + "accept": false, + "reason": "Quantized offloading behavior and tokenization-version changes are not the same bug." + }, + { + "left": "issue:43701", + "right": "issue:45325", + "accept": false, + "reason": "Checkpoint key mismatch and Qwen2.5-VL rope-index scaling are unrelated." + }, + { + "left": "issue:44829", + "right": "issue:45479", + "accept": false, + "reason": "FlashAttention-3 training degeneration and single-label classification zero-loss are different training-path bugs." + }, + { + "left": "issue:43606", + "right": "issue:44265", + "accept": false, + "reason": "CPU offload device mismatch and torch.export failures are different failure modes." + }, + { + "left": "issue:42796", + "right": "issue:45381", + "accept": false, + "reason": "Tokenizer BPE marker handling is unrelated to Qwen video vision_position_ids." + }, + { + "left": "issue:44479", + "right": "issue:45823", + "accept": false, + "reason": "Qwen video-input regression and Gemma4 PLE device mismatch are distinct model-specific issues." + }, + { + "left": "issue:43828", + "right": "issue:46821", + "accept": false, + "reason": "Autocast dtype mismatch and TimesFM window_size AttributeError do not share a concrete bug." + }, + { + "left": "issue:40990", + "right": "issue:45584", + "accept": false, + "reason": "High perplexity on gpt-oss-20b and Whisper empty-transcription failure are unrelated." + }, + { + "left": "issue:43262", + "right": "issue:43746", + "accept": false, + "reason": "Audio chat-template sampling-rate defaulting and PEFT checkpoint loading are different concerns." + }, + { + "left": "issue:34634", + "right": "issue:35707", + "accept": false, + "reason": "Bark voice_preset handling and progressive generation with inputs_embeds/past_key_values are unrelated." + }, + { + "left": "issue:35707", + "right": "issue:41720", + "accept": false, + "reason": "Different code paths: generation with inputs_embeds/past_key_values versus auto device mapping cuda asserts." + }, + { + "left": "issue:41720", + "right": "issue:45823", + "accept": false, + "reason": "Both mention auto device mapping, but the model-specific device failures and symptoms are different." + }, + { + "left": "issue:44186", + "right": "issue:44265", + "accept": false, + "reason": "Tokenizer crashes on NER/padding and torch.export failures with torch_compilable_check are different bugs." + } + ] + }, + "evaluator_result": null, + "evaluator_used": false, + "retried": false, + "accepted_nontrivial_soft_edge": false, + "error_kind": null, + "error_message": null + }, + { + "label": "issue", + "component_index": 1, + "component_count": 28, + "review_unit_index": 15, + "review_unit_count": 65, + "status": "reviewed", + "reason": null, + "source": "llm", + "cache_hit": false, + "model": "gpt-5.4-mini?service_tier=flex", + "cluster_id": "cluster-10000-530", + "nodes": [ + "issue:26413", + "issue:28530", + "issue:36331", + "issue:40990", + "issue:41553", + "issue:41720", + "issue:42617", + "issue:42796", + "issue:42915", + "issue:42947", + "issue:43257", + "issue:43317", + "issue:43322", + "issue:43377", + "issue:43450", + "issue:43454", + "issue:43550", + "issue:43572", + "issue:43582", + "issue:43606", + "issue:43720", + "issue:43749", + "issue:43819", + "issue:43873", + "issue:44387", + "issue:44466", + "issue:44534", + "issue:44568", + "issue:44610", + "issue:44756", + "issue:44877", + "issue:44898", + "issue:44977", + "issue:44991", + "issue:45072", + "issue:45357", + "issue:45685", + "issue:45701", + "issue:45704", + "issue:45759", + "issue:45823", + "issue:45920", + "issue:45923", + "issue:46050", + "issue:46209", + "issue:46305", + "issue:46315", + "issue:46821" + ], + "soft_pairs": [ + "issue:43606|issue:43819", + "issue:43454|issue:43550", + "issue:42617|issue:42915", + "issue:45823|issue:46821", + "issue:43257|issue:45923", + "issue:42796|issue:46209", + "issue:40990|issue:42947", + "issue:43454|issue:45072", + "issue:44568|issue:45920", + "issue:43377|issue:44610", + "issue:36331|issue:42617", + "issue:41553|issue:43582", + "issue:41720|issue:43322", + "issue:43317|issue:43873", + "issue:43749|issue:44534", + "issue:36331|issue:42915", + "issue:43454|issue:43720", + "issue:43572|issue:45759", + "issue:44877|issue:44991", + "issue:43377|issue:44898", + "issue:46305|issue:46315", + "issue:41720|issue:44977", + "issue:41720|issue:42617", + "issue:43322|issue:46050", + "issue:44387|issue:45704", + "issue:40990|issue:45701", + "issue:43450|issue:45701", + "issue:43873|issue:44756", + "issue:44466|issue:45357", + "issue:45685|issue:46821", + "issue:26413|issue:28530" + ], + "prepared_review_unit_hash": "f1b7766010b9d430701405e7ff173f81b2ee6ea7806992c7ad54ce503ccb2e78", + "component_budget": { + "node_count": 530, + "item_count": 530, + "soft_pair_count": 1954, + "serialized_chars": 1013004, + "estimated_input_tokens": 253251, + "estimated_eval_tokens": 506758 + }, + "budget": { + "node_count": 48, + "item_count": 48, + "soft_pair_count": 31, + "serialized_chars": 36996, + "estimated_input_tokens": 9249, + "estimated_eval_tokens": 18754 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": true, + "analyst_result": { + "summary": "This set looks like a noisy grab-bag of largely unrelated Transformers issues across tokenization, quantization/offloading, configs, model-specific regressions, and training utilities. I don't see any pair that is clearly the same underlying bug or change, so I reject all soft edges. If forced to pick one representative issue, #43873 is the broadest umbrella around offloading/quantization failures.", + "confidence": 0.94, + "canonical_issue_reason": "No true duplicate cluster emerges from the titles alone; #43873 is the broadest issue and the closest umbrella for nearby offloading/quantization failures.", + "canonical_pr_reason": null, + "best_issue_reason": "#43873 is the most general representative of the set, but only weakly so; the rest are mostly distinct model- or subsystem-specific bugs.", + "best_pr_reason": null, + "soft_edge_verdicts": [ + { + "left": "issue:43606", + "right": "issue:43819", + "accept": false, + "reason": "Different bugs: CPU offload device mismatch for Bark vs a DAC latent/STE mismatch. No shared code-path is evident." + }, + { + "left": "issue:43454", + "right": "issue:43550", + "accept": false, + "reason": "AyaVision lm_head tying bug and Bamba torch.compile/SDPA failure are unrelated." + }, + { + "left": "issue:42617", + "right": "issue:42915", + "accept": false, + "reason": "3d_parallel.py runtime failure is unrelated to Qwen3Moe fine-grained FP8 loading." + }, + { + "left": "issue:45823", + "right": "issue:46821", + "accept": false, + "reason": "Gemma4 device-mapping mismatch and TimesFM window_size AttributeError are different subsystems." + }, + { + "left": "issue:43257", + "right": "issue:45923", + "accept": false, + "reason": "Qwen3 MoE weight conversion under accelerate+deepspeed is unrelated to a missing gradient-checkpointing flag." + }, + { + "left": "issue:42796", + "right": "issue:46209", + "accept": false, + "reason": "Tokenizer BPE-marker handling and FP8 inference on B200 are not the same bug." + }, + { + "left": "issue:40990", + "right": "issue:42947", + "accept": false, + "reason": "High perplexity on a model/dataset pair is unrelated to LoRA gradient checkpointing ineffectiveness." + }, + { + "left": "issue:43454", + "right": "issue:45072", + "accept": false, + "reason": "Both mention model failures, but one is weight tying and the other is a dtype mismatch; different code paths." + }, + { + "left": "issue:44568", + "right": "issue:45920", + "accept": false, + "reason": "Both are tokenizer regressions, but one is missing special tokens and the other is wrong token IDs across several models." + }, + { + "left": "issue:43377", + "right": "issue:44610", + "accept": false, + "reason": "Batching-related issues, but one is MIMI padding-mask behavior and the other is a video processor shape bug." + }, + { + "left": "issue:36331", + "right": "issue:42617", + "accept": false, + "reason": "CustomTrainer API change and 3d_parallel.py failure are unrelated." + }, + { + "left": "issue:41553", + "right": "issue:43582", + "accept": false, + "reason": "Tokenizer loading error for Voxtral and AppleSilicon allocator TypeError share no concrete bug." + }, + { + "left": "issue:41720", + "right": "issue:43322", + "accept": false, + "reason": "Auto device-map assert on Qwen3 and Llava Next segfault are different failures with no common concrete cause shown." + }, + { + "left": "issue:43317", + "right": "issue:43873", + "accept": false, + "reason": "These are both offload/quantization-adjacent, but one is a specific dequantized-model loading failure and the other is a broad offloading complaint; not enough evidence they are the same bug." + }, + { + "left": "issue:43749", + "right": "issue:44534", + "accept": false, + "reason": "FSDP CPU RAM efficient loading and non-persistent buffer junk are distinct regressions." + }, + { + "left": "issue:36331", + "right": "issue:42915", + "accept": false, + "reason": "Trainer signature change and Qwen3Moe FP8 loading failure are unrelated." + }, + { + "left": "issue:43454", + "right": "issue:43720", + "accept": false, + "reason": "AyaVision weight tying and BitNet packed-weight unpacking are different model-specific issues." + }, + { + "left": "issue:43572", + "right": "issue:45759", + "accept": false, + "reason": "Missing pad_token_idx in StableLmConfig and text_config unwrapping for composite Qwen configs are separate config-loading problems." + }, + { + "left": "issue:44877", + "right": "issue:44991", + "accept": false, + "reason": "Strict config validation for granite_speech and tokenizer loading failure for est-roberta are unrelated." + }, + { + "left": "issue:43377", + "right": "issue:44898", + "accept": false, + "reason": "Different models and symptoms: MIMI batching/padding vs Perceiver positional interpolation." + }, + { + "left": "issue:46305", + "right": "issue:46315", + "accept": false, + "reason": "MoonshineStreaming sliding-window off-by-one and tokenizer conversion slowdown are unrelated." + }, + { + "left": "issue:41720", + "right": "issue:44977", + "accept": false, + "reason": "Both involve Qwen3-family models, but one is device-map CUDA assert and the other is flash-attention generation quality; not the same concrete bug." + }, + { + "left": "issue:41720", + "right": "issue:42617", + "accept": false, + "reason": "Qwen3 auto device mapping and 3d_parallel.py runtime failure do not share a concrete code-path." + }, + { + "left": "issue:43322", + "right": "issue:46050", + "accept": false, + "reason": "Llava Next segfault and sinq save/load breakage are different serialization/loading failures." + }, + { + "left": "issue:44387", + "right": "issue:45704", + "accept": false, + "reason": "Int4 quantization OOM and T5 apex memory leak are different memory issues in different implementations." + }, + { + "left": "issue:40990", + "right": "issue:45701", + "accept": false, + "reason": "Perplexity regression on one model and tokenization changes across versions are not the same bug." + }, + { + "left": "issue:43450", + "right": "issue:45701", + "accept": false, + "reason": "Video processor batch-shape bug and version-dependent tokenization changes are unrelated." + }, + { + "left": "issue:43873", + "right": "issue:44756", + "accept": false, + "reason": "Quantization/offloading issue and mmap OOM on Strix Halo are different resource problems." + }, + { + "left": "issue:44466", + "right": "issue:45357", + "accept": false, + "reason": "Both concern save_pretrained regressions, but one is lm_head serialization and the other is visual encoder key naming in Qwen3.5." + }, + { + "left": "issue:45685", + "right": "issue:46821", + "accept": false, + "reason": "MPS histogram error for Int and TimesFM window_size AttributeError are unrelated." + }, + { + "left": "issue:26413", + "right": "issue:28530", + "accept": false, + "reason": "resume_from_checkpoint epoch_iterator failure and early-stopping metric lookup failure are distinct trainer features." + } + ] + }, + "evaluator_result": null, + "evaluator_used": false, + "retried": false, + "accepted_nontrivial_soft_edge": false, + "error_kind": null, + "error_message": null + }, + { + "label": "issue", + "component_index": 1, + "component_count": 28, + "review_unit_index": 16, + "review_unit_count": 65, + "status": "reviewed", + "reason": null, + "source": "llm", + "cache_hit": false, + "model": "gpt-5.4-mini?service_tier=flex", + "cluster_id": "cluster-10000-530", + "nodes": [ + "issue:28530", + "issue:30064", + "issue:33666", + "issue:41720", + "issue:42915", + "issue:43065", + "issue:43295", + "issue:43317", + "issue:43404", + "issue:43441", + "issue:43531", + "issue:43550", + "issue:43572", + "issue:43577", + "issue:43606", + "issue:43646", + "issue:43708", + "issue:43716", + "issue:43720", + "issue:43881", + "issue:44008", + "issue:44186", + "issue:44361", + "issue:44423", + "issue:44451", + "issue:44464", + "issue:44488", + "issue:44610", + "issue:44734", + "issue:44898", + "issue:45072", + "issue:45161", + "issue:45357", + "issue:45406", + "issue:45701", + "issue:45823", + "issue:46291", + "issue:46402", + "issue:46612", + "issue:46665", + "issue:46821", + "issue:46858", + "issue:46869", + "issue:46897", + "issue:47020", + "issue:47317", + "issue:47337" + ], + "soft_pairs": [ + "issue:44186|issue:45072", + "issue:43295|issue:45701", + "issue:44451|issue:47020", + "issue:43065|issue:46402", + "issue:43550|issue:43720", + "issue:43646|issue:46665", + "issue:46612|issue:46858", + "issue:44186|issue:44610", + "issue:43550|issue:43606", + "issue:43577|issue:43708", + "issue:46897|issue:47317", + "issue:46821|issue:46869", + "issue:42915|issue:43317", + "issue:41720|issue:45161", + "issue:43404|issue:45357", + "issue:44423|issue:45406", + "issue:44610|issue:44898", + "issue:28530|issue:33666", + "issue:44008|issue:45823", + "issue:43881|issue:44488", + "issue:42915|issue:46291", + "issue:43531|issue:43572", + "issue:43441|issue:47337", + "issue:44186|issue:44361", + "issue:44734|issue:46858", + "issue:28530|issue:30064", + "issue:43716|issue:46821", + "issue:44464|issue:46858" + ], + "prepared_review_unit_hash": "3955e1e0d2c12f2c04856365199ab49c8efa5133779f3cf92f1c474b70bfab7e", + "component_budget": { + "node_count": 530, + "item_count": 530, + "soft_pair_count": 1954, + "serialized_chars": 1013004, + "estimated_input_tokens": 253251, + "estimated_eval_tokens": 506758 + }, + "budget": { + "node_count": 47, + "item_count": 47, + "soft_pair_count": 28, + "serialized_chars": 35509, + "estimated_input_tokens": 8878, + "estimated_eval_tokens": 18012 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": true, + "analyst_result": { + "summary": "This is a heterogeneous cluster of mostly unrelated bug reports spanning tokenization, loading, device-map/offload, dtype, and model-specific regressions. None of the soft pairs look like the same concrete underlying bug, so all should be rejected. If a representative issue is needed, #45701 is the broadest and most referenced report in the set, but the cluster itself is not cohesive.", + "confidence": 0.39, + "canonical_issue_reason": "#45701 is the broadest, most heavily referenced report here and the closest thing to an umbrella representative, but there is no truly canonical issue because the cluster is not a single bug family.", + "canonical_pr_reason": null, + "best_issue_reason": "#45701 is the best standalone representative because it is broad and well-linked, though it does not actually unify the other issues.", + "best_pr_reason": null, + "soft_edge_verdicts": [ + { + "left": "issue:44186", + "right": "issue:45072", + "accept": false, + "reason": "Both are multimodal/vision-related failures, but one is tokenizer padding/truncation and the other is dtype mismatch in bfloat16 inference; not the same bug." + }, + { + "left": "issue:43295", + "right": "issue:45701", + "accept": false, + "reason": "Both mention regressions around custom code/tokenization, but one is a processor/tokenizer API break and the other is a broad version-caused tokenization change." + }, + { + "left": "issue:44451", + "right": "issue:47020", + "accept": false, + "reason": "Both are tokenizer loading issues, but they involve different models and different failure modes; no shared concrete code path." + }, + { + "left": "issue:43065", + "right": "issue:46402", + "accept": false, + "reason": "Same model family (Sam3), but one is a dummy Conv2d placeholder and the other is missing text encoder weight loading; separate defects." + }, + { + "left": "issue:43550", + "right": "issue:43720", + "accept": false, + "reason": "Both involve loading/model execution with accelerate-like paths, but one is torch.compile+SDPA on Bamba and the other is packed-weight unpacking for BitNet; different bugs." + }, + { + "left": "issue:43646", + "right": "issue:46665", + "accept": false, + "reason": "Both concern custom model/loading behavior, but one is custom model initialization in Transformers 5.0 and the other is from_pretrained inside Python classes; not the same issue." + }, + { + "left": "issue:46612", + "right": "issue:46858", + "accept": false, + "reason": "Both touch generation cache behavior, but one is beam-search cache reorder skipped for specific model classes and the other is a static-cache seq-length tensor issue under torch.compile." + }, + { + "left": "issue:44186", + "right": "issue:44610", + "accept": false, + "reason": "Both are vision processor issues, but one is LayoutLMv2 tokenizer/NER padding and the other is OmDet-Turbo image size mismatch; unrelated." + }, + { + "left": "issue:43550", + "right": "issue:43606", + "accept": false, + "reason": "Both are device-map/offload-related, but one is a torch.compile SDPA failure and the other is a CPU-offload device mismatch in bark-small; different code paths." + }, + { + "left": "issue:43577", + "right": "issue:43708", + "accept": false, + "reason": "One is a dtype propagation bug in Blip2 loading and the other is Trainer checkpoint resumption max_steps calculation; unrelated." + }, + { + "left": "issue:46897", + "right": "issue:47317", + "accept": false, + "reason": "Both are training-loss bugs, but they affect different models and likely different loss computation paths; not the same change." + }, + { + "left": "issue:46821", + "right": "issue:46869", + "accept": false, + "reason": "Both mention multi-GPU/device-map style failures, but one is a TimesFM window_size AttributeError and the other is YOLO auto device-map loading; distinct bugs." + }, + { + "left": "issue:42915", + "right": "issue:43317", + "accept": false, + "reason": "Both are model loading failures, but one is Qwen3Moe with FineGrainedFP8Config and the other is dequantized model loading with gpu+cpu offload; different failure modes." + }, + { + "left": "issue:41720", + "right": "issue:45161", + "accept": false, + "reason": "Both concern model-parallel/device placement, but one is Qwen3 auto device mapping on A800 and the other is GPT-OSS MoE tensor-parallel behavior; not the same bug." + }, + { + "left": "issue:43404", + "right": "issue:45357", + "accept": false, + "reason": "Both involve multimodal generation models and weights, but one is tied lm_head in Mistral3 and the other is incorrect visual encoder keys in Qwen3.5 save_pretrained; separate defects." + }, + { + "left": "issue:44423", + "right": "issue:45406", + "accept": false, + "reason": "Both are serve/runtime crashes for multimodal processors, but one is a 'str' object .to() failure in continuous batching and the other is missing _tokenizer on Gemma4Processor; different root causes." + }, + { + "left": "issue:44610", + "right": "issue:44898", + "accept": false, + "reason": "Both are image-size/preprocessing mismatches, but OmDet-Turbo has 640px vs 224px processor/model mismatch while Perceiver fails on non-default resolution despite interpolate_pos_encoding; not the same bug." + }, + { + "left": "issue:28530", + "right": "issue:33666", + "accept": false, + "reason": "Early stopping metric lookup and Qwen2-VL multi-GPU training are unrelated." + }, + { + "left": "issue:44008", + "right": "issue:45823", + "accept": false, + "reason": "Both are Gemma/Mistral-family multimodal device issues, but one is a variable-name collision causing an AttributeError and the other is a device mismatch under auto mapping in forward." + }, + { + "left": "issue:43881", + "right": "issue:44488", + "accept": false, + "reason": "Both are model loading failures, but they involve different models and likely different tokenizer/config issues; no clear shared bug." + }, + { + "left": "issue:42915", + "right": "issue:46291", + "accept": false, + "reason": "One is a FineGrainedFP8Config load failure and the other is a pipelines import/version constraint on kernels; unrelated." + }, + { + "left": "issue:43531", + "right": "issue:43572", + "accept": false, + "reason": "Both are config-related regressions in newer versions, but one is sliding_window behavior for Qwen3-MoE and the other is missing pad_token_idx in StableLmConfig; separate issues." + }, + { + "left": "issue:43441", + "right": "issue:47337", + "accept": false, + "reason": "FlashAttention failure in Ministral-3 and a flaky bbox coordinate test are unrelated." + }, + { + "left": "issue:44186", + "right": "issue:44361", + "accept": false, + "reason": "Both are tokenizer/attribute errors on task inputs, but they affect different tokenizers and different task flows; no shared code path." + }, + { + "left": "issue:44734", + "right": "issue:46858", + "accept": false, + "reason": "Both involve serve/cache behavior, but one is KV cache continuation in /v1/responses and the other is static cache seq-length arithmetic under torch.compile; different bugs." + }, + { + "left": "issue:28530", + "right": "issue:30064", + "accept": false, + "reason": "Early stopping metric selection and void segmentation map processing are unrelated." + }, + { + "left": "issue:43716", + "right": "issue:46821", + "accept": false, + "reason": "Both mention shape/dtype-ish model input issues, but one is Mistral image-preprocessor dtype mismatch and the other is TimesFM window_size AttributeError; not the same bug." + }, + { + "left": "issue:44464", + "right": "issue:46858", + "accept": false, + "reason": "Both concern compiled generation behavior, but one is inconsistent chunked generation outputs and the other is cache length tensor arithmetic on non-CUDA backends; different failures." + } + ] + }, + "evaluator_result": null, + "evaluator_used": false, + "retried": false, + "accepted_nontrivial_soft_edge": false, + "error_kind": null, + "error_message": null + }, + { + "label": "issue", + "component_index": 1, + "component_count": 28, + "review_unit_index": 17, + "review_unit_count": 65, + "status": "reviewed", + "reason": null, + "source": "llm", + "cache_hit": false, + "model": "gpt-5.4-mini?service_tier=flex", + "cluster_id": "cluster-10000-530", + "nodes": [ + "issue:29159", + "issue:30064", + "issue:33357", + "issue:34634", + "issue:35707", + "issue:36032", + "issue:41720", + "issue:42490", + "issue:42915", + "issue:43257", + "issue:43316", + "issue:43322", + "issue:43329", + "issue:43452", + "issue:43526", + "issue:43531", + "issue:43540", + "issue:43550", + "issue:43606", + "issue:43653", + "issue:43720", + "issue:43746", + "issue:43782", + "issue:43873", + "issue:43881", + "issue:43883", + "issue:44492", + "issue:44509", + "issue:44534", + "issue:44589", + "issue:44671", + "issue:44743", + "issue:44913", + "issue:44991", + "issue:45479", + "issue:45632", + "issue:45684", + "issue:45701", + "issue:45704", + "issue:45759", + "issue:46665", + "issue:46821", + "issue:46894", + "issue:47020", + "issue:47145", + "issue:47254", + "issue:47317" + ], + "soft_pairs": [ + "issue:44671|issue:44743", + "issue:29159|issue:36032", + "issue:43316|issue:44913", + "issue:30064|issue:34634", + "issue:45632|issue:45684", + "issue:43452|issue:47020", + "issue:45479|issue:47317", + "issue:45701|issue:47145", + "issue:42915|issue:43322", + "issue:43873|issue:45704", + "issue:33357|issue:35707", + "issue:43540|issue:43782", + "issue:43531|issue:45759", + "issue:43881|issue:44991", + "issue:41720|issue:43746", + "issue:43550|issue:43653", + "issue:44534|issue:46665", + "issue:46894|issue:47317", + "issue:43257|issue:47254", + "issue:43526|issue:43883", + "issue:42490|issue:43329", + "issue:43606|issue:43720", + "issue:43531|issue:46821", + "issue:44492|issue:44509", + "issue:34634|issue:41720", + "issue:43531|issue:44589" + ], + "prepared_review_unit_hash": "7fd7cb8141e7c1850c4520de4244f1d18bc3074726f107c13c3a0c17b4142be5", + "component_budget": { + "node_count": 530, + "item_count": 530, + "soft_pair_count": 1954, + "serialized_chars": 1013004, + "estimated_input_tokens": 253251, + "estimated_eval_tokens": 506758 + }, + "budget": { + "node_count": 47, + "item_count": 47, + "soft_pair_count": 26, + "serialized_chars": 34870, + "estimated_input_tokens": 8718, + "estimated_eval_tokens": 17692 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": true, + "analyst_result": { + "summary": "No soft edge looks like a true duplicate or mergeable same-bug pair; the cluster is mostly a loose similarity set spanning tokenizer, config, loading, quantization, and model-specific runtime bugs. If one issue must represent the cluster, the Qwen3-MoE sliding-window bug is the closest anchor, but it still does not subsume the others.", + "confidence": 0.77, + "canonical_issue_reason": "issue:43531 is the most central item in the sample, appearing in multiple soft candidates and being a concrete code-path bug rather than a broad meta report; it is the closest representative of the cluster, though not a duplicate of the linked issues.", + "canonical_pr_reason": null, + "best_issue_reason": "issue:43531 is the best available issue representative because it is specific, technical, and the most connected of the sampled issues; however, the cluster is too heterogeneous for any issue to be a strong canonical summary.", + "best_pr_reason": null, + "soft_edge_verdicts": [ + { + "left": "issue:44671", + "right": "issue:44743", + "accept": false, + "reason": "Both are model-specific behavior bugs, but one is CamemBERT masked-LM output and the other is Qwen modular cache state reset; different models and code paths." + }, + { + "left": "issue:29159", + "right": "issue:36032", + "accept": false, + "reason": "Both mention tokenizers, but one is slow-vs-fast inconsistency and the other is a T5Tokenizer load-time attribute conflict; not the same bug." + }, + { + "left": "issue:43316", + "right": "issue:44913", + "accept": false, + "reason": "One is a Gemma3TextConfig API discrepancy, the other is GPTNeoX rotary_pct persistence on reload; different configuration fields and failure modes." + }, + { + "left": "issue:30064", + "right": "issue:34634", + "accept": false, + "reason": "Image processor void segmentation maps and BarkProcessor voice_preset are unrelated processor-specific issues." + }, + { + "left": "issue:45632", + "right": "issue:45684", + "accept": false, + "reason": "Cache-path collision for trust_remote_code and read-only permission propagation in save_pretrained are different file-system concerns with different fixes." + }, + { + "left": "issue:43452", + "right": "issue:47020", + "accept": false, + "reason": "Both are tokenizer/loading-related, but gguf_file breaks from_pretrained broadly while the XLMRobertaTokenizer issue is a specific Unigram vocab type mismatch." + }, + { + "left": "issue:45479", + "right": "issue:47317", + "accept": false, + "reason": "Both involve training loss, but they affect different models and likely different loss computations; too broad to be the same bug." + }, + { + "left": "issue:45701", + "right": "issue:47145", + "accept": false, + "reason": "Tokenization changes across versions and FP8 weight_scale loading corruption are unrelated." + }, + { + "left": "issue:42915", + "right": "issue:43322", + "accept": false, + "reason": "FineGrainedFP8Config failure in Qwen3Moe and Llava Next segmentation fault are different model families and distinct failure classes." + }, + { + "left": "issue:43873", + "right": "issue:45704", + "accept": false, + "reason": "Quantization offloading behavior and T5/apex RMSNorm memory leak are not the same underlying issue." + }, + { + "left": "issue:33357", + "right": "issue:35707", + "accept": false, + "reason": "MacOS bus error in a community CLIP model and progressive generation with inputs_embeds/past_key_values are unrelated." + }, + { + "left": "issue:43540", + "right": "issue:43782", + "accept": false, + "reason": "Both are Qwen multimodal-related, but one is a video-input processing error and the other is a from_pretrained weight_only load failure." + }, + { + "left": "issue:43531", + "right": "issue:45759", + "accept": false, + "reason": "Both involve Qwen3, but one is a sliding_window recurrent-state bug and the other is config unwrapping for composite multimodal models; different code paths." + }, + { + "left": "issue:43881", + "right": "issue:44991", + "accept": false, + "reason": "GLM-4V loading failure and tokenizer loading for EMBEDDIA/est-roberta are unrelated." + }, + { + "left": "issue:41720", + "right": "issue:43746", + "accept": false, + "reason": "Auto device mapping CUDA assert on Qwen3 and GraniteSpeech PEFT checkpoint loading are different problems." + }, + { + "left": "issue:43550", + "right": "issue:43653", + "accept": false, + "reason": "torch.compile/SDPA on Bamba-9B-v2 and BigBirdTokenizer special-token registration are unrelated." + }, + { + "left": "issue:44534", + "right": "issue:46665", + "accept": false, + "reason": "Non-persistent buffer junk in v5 and from_pretrained behavior inside Python classes do not describe the same failure." + }, + { + "left": "issue:46894", + "right": "issue:47317", + "accept": false, + "reason": "Both are training-loss bugs, but they concern different models and likely different shifting logic; not enough to treat as one issue." + }, + { + "left": "issue:43257", + "right": "issue:47254", + "accept": false, + "reason": "Qwen3 MoE weight conversion under accelerate+deepspeed and PEFT+ZeRO-3 gradient checkpointing are distinct deep-learning stack interactions." + }, + { + "left": "issue:43526", + "right": "issue:43883", + "accept": false, + "reason": "BEiTImageProcessorFast reduce_labels and Molmo all_tied_weights_keys are unrelated attribute/processing bugs." + }, + { + "left": "issue:42490", + "right": "issue:43329", + "accept": false, + "reason": "Local-path trailing-slash import warning and undefined video-branch variables in multimodal token counting are unrelated." + }, + { + "left": "issue:43606", + "right": "issue:43720", + "accept": false, + "reason": "CPU offload device mismatch and BitNet packed-weight unpacking during accelerate loading are both loading-related but not the same concrete bug." + }, + { + "left": "issue:43531", + "right": "issue:46821", + "accept": false, + "reason": "Qwen3-MoE sliding-window behavior and TimesFM window_size AttributeError are different model implementations and different parameter handling." + }, + { + "left": "issue:44492", + "right": "issue:44509", + "accept": false, + "reason": "A typo in cache strategies and stale docs for removed pipeline tasks are unrelated documentation/API issues." + }, + { + "left": "issue:34634", + "right": "issue:41720", + "accept": false, + "reason": "BarkProcessor voice_preset and Qwen3 auto device mapping CUDA assert are different models and failure modes." + }, + { + "left": "issue:43531", + "right": "issue:44589", + "accept": false, + "reason": "Qwen3-MoE sliding-window logic and Float8 storage lookup failure are unrelated runtime bugs." + } + ] + }, + "evaluator_result": null, + "evaluator_used": false, + "retried": false, + "accepted_nontrivial_soft_edge": false, + "error_kind": null, + "error_message": null + }, + { + "label": "issue", + "component_index": 1, + "component_count": 28, + "review_unit_index": 18, + "review_unit_count": 65, + "status": "reviewed", + "reason": null, + "source": "llm", + "cache_hit": false, + "model": "gpt-5.4-mini?service_tier=flex", + "cluster_id": "cluster-10000-530", + "nodes": [ + "issue:30064", + "issue:34567", + "issue:35707", + "issue:36010", + "issue:39692", + "issue:41720", + "issue:42886", + "issue:43381", + "issue:43388", + "issue:43421", + "issue:43540", + "issue:43606", + "issue:43644", + "issue:43653", + "issue:43746", + "issue:43873", + "issue:43927", + "issue:44062", + "issue:44077", + "issue:44451", + "issue:44514", + "issue:44545", + "issue:44610", + "issue:44625", + "issue:44843", + "issue:44898", + "issue:44918", + "issue:44936", + "issue:44987", + "issue:44991", + "issue:45042", + "issue:45059", + "issue:45072", + "issue:45357", + "issue:45406", + "issue:45464", + "issue:45479", + "issue:45704", + "issue:45759", + "issue:45859", + "issue:45923", + "issue:46050", + "issue:46402", + "issue:46655", + "issue:46665", + "issue:46693", + "issue:46858", + "issue:47333" + ], + "soft_pairs": [ + "issue:45406|issue:45464", + "issue:34567|issue:35707", + "issue:43606|issue:43653", + "issue:45759|issue:45859", + "issue:44451|issue:44991", + "issue:43540|issue:46858", + "issue:43388|issue:45059", + "issue:44077|issue:44625", + "issue:43644|issue:46665", + "issue:44987|issue:45042", + "issue:45923|issue:47333", + "issue:45357|issue:46402", + "issue:44898|issue:45072", + "issue:44062|issue:44451", + "issue:41720|issue:45704", + "issue:44918|issue:44936", + "issue:39692|issue:44610", + "issue:43746|issue:46050", + "issue:43381|issue:45704", + "issue:44625|issue:45479", + "issue:42886|issue:44843", + "issue:30064|issue:36010", + "issue:44514|issue:44545", + "issue:43746|issue:43873", + "issue:43421|issue:43927", + "issue:46655|issue:46693" + ], + "prepared_review_unit_hash": "b0e8fd75f592ce1dc6850b91f57006920fd9c7dbe4341af763916e2c30af9d67", + "component_budget": { + "node_count": 530, + "item_count": 530, + "soft_pair_count": 1954, + "serialized_chars": 1013004, + "estimated_input_tokens": 253251, + "estimated_eval_tokens": 506758 + }, + "budget": { + "node_count": 48, + "item_count": 48, + "soft_pair_count": 26, + "serialized_chars": 35719, + "estimated_input_tokens": 8930, + "estimated_eval_tokens": 18116 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": true, + "analyst_result": { + "summary": "These items are mostly unrelated issues from different code paths; only the Qwen2_5_VLProcessor batched-padding reports look like a true duplicate pair. The rest share broad themes like loading/offloading/tokenization but describe distinct failures.", + "confidence": 0.86, + "canonical_issue_reason": "issue:43873 is the broadest and most actively discussed issue here, with the most inbound references; if a single issue must represent the set, it is the best umbrella candidate, though this is not a true duplicate cluster.", + "canonical_pr_reason": null, + "best_issue_reason": "issue:43873 has the strongest signal as a representative issue because it is high-activity, open, and broadly scoped around offloading/quantization behavior.", + "best_pr_reason": null, + "soft_edge_verdicts": [ + { + "left": "issue:45406", + "right": "issue:45464", + "accept": false, + "reason": "Different products and failures: Gemma4Processor missing _tokenizer vs Qwen3.5-0.8B streaming chat/completions crash. Same broad area, not the same bug." + }, + { + "left": "issue:34567", + "right": "issue:35707", + "accept": false, + "reason": "Both involve trainer-related state/generation behavior, but one is num_input_tokens_seen not updating and the other is progressive generation with inputs_embeds/past_key_values. Not the same underlying bug." + }, + { + "left": "issue:43606", + "right": "issue:43653", + "accept": false, + "reason": "Different tokenizer/model problems: CPU offload device mismatch for bark-small vs BigBirdTokenizer mask token registration and empty decode output." + }, + { + "left": "issue:45759", + "right": "issue:45859", + "accept": false, + "reason": "Related to Qwen 3.5/3.6 multimodal config handling, but one is AutoModelForCausalLM.from_config text_config unwrapping and the other is a rotary embedding CPU-offload compatibility issue." + }, + { + "left": "issue:44451", + "right": "issue:44991", + "accept": false, + "reason": "Both are tokenizer loading regressions, but they affect different models and likely different root causes; not enough evidence of one concrete bug." + }, + { + "left": "issue:43540", + "right": "issue:46858", + "accept": false, + "reason": "Video-input ValueError in Qwen3OmniMoe is unrelated to cache.get_seq_length() returning a device tensor in static cache/torch.compile prefill." + }, + { + "left": "issue:43388", + "right": "issue:45059", + "accept": false, + "reason": "Different issues: gather_for_metrics dropping tuple labels in the last batch versus SAM3 PCS odd behavior with text and bounding boxes." + }, + { + "left": "issue:44077", + "right": "issue:44625", + "accept": false, + "reason": "patchtsmixer post_init allowance and Qwen3.5 num_labels propagation are unrelated model/config bugs." + }, + { + "left": "issue:43644", + "right": "issue:46665", + "accept": false, + "reason": "Non-persistent buffers filled with junk is a serialization/init bug; from_pretrained inside Python classes is a separate behavioral inconsistency." + }, + { + "left": "issue:44987", + "right": "issue:45042", + "accept": false, + "reason": "One is a loading regression for a specific model; the other is PIL image processors incorrectly requiring torchvision. Same general dependency/loading surface, different bugs." + }, + { + "left": "issue:45923", + "right": "issue:47333", + "accept": false, + "reason": "Both concern offloading on Qwen models, but one is a missing supports_gradient_checkpointing flag and the other is inability to save a disk-offloaded model. Not the same code-path problem." + }, + { + "left": "issue:45357", + "right": "issue:46402", + "accept": false, + "reason": "Both involve Qwen/SAM model loading, but one is incorrect visual encoder keys on save_pretrained and the other is text encoder weights not loaded. Distinct defects." + }, + { + "left": "issue:44898", + "right": "issue:45072", + "accept": false, + "reason": "Perceiver image classification resolution handling and bfloat16 dtype mismatches in SwitchTransformers/TimmWrapperModel are unrelated." + }, + { + "left": "issue:44062", + "right": "issue:44451", + "accept": false, + "reason": "AddedToken special-argument duplication and inability to load a specific tokenizer/model are different tokenizer-layer issues, not the same bug." + }, + { + "left": "issue:41720", + "right": "issue:45704", + "accept": false, + "reason": "Qwen3 auto device mapping CUDA assert and T5 apex.FusedRMSNorm memory leak are unrelated inference/runtime problems." + }, + { + "left": "issue:44918", + "right": "issue:44936", + "accept": false, + "reason": "TRL SFT input-embedding unpacking and trainer.evaluate() after trainer.train() are different trainer behaviors." + }, + { + "left": "issue:39692", + "right": "issue:44610", + "accept": false, + "reason": "SigLIP2 doc example errors are documentation/example problems; OmDet-Turbo processor size mismatch is a runtime processor/model config bug. Not the same issue." + }, + { + "left": "issue:43746", + "right": "issue:46050", + "accept": false, + "reason": "PEFT adapters failing to load from local checkpoints and sinq-quantized models breaking after save/load are different persistence/quantization bugs." + }, + { + "left": "issue:43381", + "right": "issue:45704", + "accept": false, + "reason": "Gradient checkpointing in eval mode and T5 apex RMSNorm memory leak are unrelated." + }, + { + "left": "issue:44625", + "right": "issue:45479", + "accept": false, + "reason": "Qwen3.5 num_labels propagation and single-label classification with num_labels=1 causing zero loss are different configuration/classification bugs." + }, + { + "left": "issue:42886", + "right": "issue:44843", + "accept": false, + "reason": "Both involve offline/loading behavior, but one is tokenizer cache loading with HF_HUB_OFFLINE and the other is an unconditional model_info() call in _patch_mistral_regex. Different root causes." + }, + { + "left": "issue:30064", + "right": "issue:36010", + "accept": false, + "reason": "Void segmentation map processing and GenerationMixin import failure are unrelated subsystems." + }, + { + "left": "issue:44514", + "right": "issue:44545", + "accept": true, + "reason": "These titles describe the same exact Qwen2_5_VLProcessor.apply_chat_template batched-input crash with padding=False; they look like duplicate reports of the same bug." + }, + { + "left": "issue:43746", + "right": "issue:43873", + "accept": false, + "reason": "Both mention loading/offloading/quantization, but one is PEFT local checkpoint loading and the other is general offloading with quantization. Not the same bug." + }, + { + "left": "issue:43421", + "right": "issue:43927", + "accept": false, + "reason": "Runtime post-processor update for special tokens and DiaConfig losing custom token IDs after save/load are related to token metadata but are distinct failures." + }, + { + "left": "issue:46655", + "right": "issue:46693", + "accept": false, + "reason": "SmolVLM video padding calculation and FA2 long-context generation performance regression are separate issues." + } + ] + }, + "evaluator_result": { + "accept": true, + "feedback": "The analysis stays grounded in the packet and the soft-edge calls are conservative. The only clearly supported duplicate pair is issue:44514 / issue:44545, and the choice of issue:43873 as an umbrella representative is reasonable based on activity and breadth. No obvious overstatement." + }, + "evaluator_used": true, + "retried": false, + "accepted_nontrivial_soft_edge": true, + "error_kind": null, + "error_message": null + }, + { + "label": "issue", + "component_index": 1, + "component_count": 28, + "review_unit_index": 19, + "review_unit_count": 65, + "status": "reviewed", + "reason": null, + "source": "llm", + "cache_hit": false, + "model": "gpt-5.4-mini?service_tier=flex", + "cluster_id": "cluster-10000-530", + "nodes": [ + "issue:30064", + "issue:36331", + "issue:39692", + "issue:41720", + "issue:43295", + "issue:43381", + "issue:43421", + "issue:43526", + "issue:43531", + "issue:43540", + "issue:43550", + "issue:43596", + "issue:43643", + "issue:43653", + "issue:43673", + "issue:43720", + "issue:43782", + "issue:43819", + "issue:43844", + "issue:43873", + "issue:43881", + "issue:43927", + "issue:43994", + "issue:44062", + "issue:44186", + "issue:44220", + "issue:44368", + "issue:44464", + "issue:44488", + "issue:44671", + "issue:44857", + "issue:44991", + "issue:44995", + "issue:45072", + "issue:45081", + "issue:45356", + "issue:45357", + "issue:45381", + "issue:45834", + "issue:45907", + "issue:46032", + "issue:46294", + "issue:46463", + "issue:46693", + "issue:46858", + "issue:47020", + "issue:47050", + "issue:47337" + ], + "soft_pairs": [ + "issue:43381|issue:43844", + "issue:43421|issue:43653", + "issue:39692|issue:43873", + "issue:44464|issue:47050", + "issue:45081|issue:45356", + "issue:44995|issue:46858", + "issue:44368|issue:45357", + "issue:43994|issue:44671", + "issue:43673|issue:47337", + "issue:43782|issue:45381", + "issue:43421|issue:44186", + "issue:46032|issue:46463", + "issue:44857|issue:47050", + "issue:45356|issue:45907", + "issue:39692|issue:47050", + "issue:43596|issue:44220", + "issue:43720|issue:43819", + "issue:43295|issue:45834", + "issue:43540|issue:46294", + "issue:43526|issue:43881", + "issue:43550|issue:43927", + "issue:43550|issue:43819", + "issue:44488|issue:47020", + "issue:44062|issue:44488", + "issue:43531|issue:43643", + "issue:43653|issue:43720", + "issue:45072|issue:47050", + "issue:41720|issue:44991", + "issue:30064|issue:36331", + "issue:45381|issue:46693" + ], + "prepared_review_unit_hash": "6b6daa3bdd3de88d7b5bfa972d7609886809d4c0cc81f82347a9c8cbb7ebe37e", + "component_budget": { + "node_count": 530, + "item_count": 530, + "soft_pair_count": 1954, + "serialized_chars": 1013004, + "estimated_input_tokens": 253251, + "estimated_eval_tokens": 506758 + }, + "budget": { + "node_count": 48, + "item_count": 48, + "soft_pair_count": 30, + "serialized_chars": 37015, + "estimated_input_tokens": 9254, + "estimated_eval_tokens": 18764 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": true, + "analyst_result": { + "summary": "This is a heterogeneous cluster of tokenizer, model-loading, generation, vision, and training bugs. I do not see any soft pair that is clearly the same underlying defect, so none should be merged as duplicates.", + "confidence": 0.94, + "canonical_issue_reason": "If forced to pick one anchor, issue 43421 is the broadest reusable tokenizer-runtime regression in the set; however the cluster overall is too mixed to treat as a single duplicate group.", + "canonical_pr_reason": null, + "best_issue_reason": "43421 is the best representative only in the weak sense that it is a general tokenizer/backend behavior bug rather than a model-specific crash; it still does not summarize the whole cluster well.", + "best_pr_reason": null, + "soft_edge_verdicts": [ + { + "left": "issue:43381", + "right": "issue:43844", + "accept": false, + "reason": "Both involve training/runtime behavior, but one is gradient checkpointing in eval mode and the other is a DeepSpeed ZeRO-3 gradient anomaly; different failure modes and code paths." + }, + { + "left": "issue:43421", + "right": "issue:43653", + "accept": false, + "reason": "Both touch tokenizers and special tokens, but 43421 is about runtime post-processor updates while 43653 is a BigBirdTokenizer mask-token registration/empty decode bug." + }, + { + "left": "issue:39692", + "right": "issue:43873", + "accept": false, + "reason": "One is a SigLIP2 doc/example problem; the other is an offloading-with-quantization runtime bug. Same broad area, but not the same defect." + }, + { + "left": "issue:44464", + "right": "issue:47050", + "accept": false, + "reason": "Both mention inference/compiled execution, but 44464 is chunked generation inconsistency and 47050 is an ONNX fp16 attention dtype mismatch." + }, + { + "left": "issue:45081", + "right": "issue:45356", + "accept": false, + "reason": "Both involve tokenizer regex handling, but one crashes on missing backend_tokenizer and the other reports a version regression with incorrect codec handling; not the same concrete bug." + }, + { + "left": "issue:44995", + "right": "issue:46858", + "accept": false, + "reason": "Both mention cache/indexing, but one is a stale indexer cache on repeated forwards and the other is a device-tensor seq-length arithmetic issue in static cache/torch.compile." + }, + { + "left": "issue:44368", + "right": "issue:45357", + "accept": false, + "reason": "One is a warning about tie_word_embeddings in fine-tuning, the other is incorrect keys saved for Qwen3.5 visual encoders; unrelated defects." + }, + { + "left": "issue:43994", + "right": "issue:44671", + "accept": false, + "reason": "Both are SigLIP/CamemBERT model behavior issues, but one is generic wrong outputs with AutoModel/pipeline and the other is masked-LM prediction correctness; different bugs." + }, + { + "left": "issue:43673", + "right": "issue:47337", + "accept": false, + "reason": "One is a generation cache missing in chunked prefill, the other is a flaky bbox range assertion in tests; not the same issue." + }, + { + "left": "issue:43782", + "right": "issue:45381", + "accept": false, + "reason": "Both are multimodal Qwen video/model issues, but one is a from_pretrained weight_only load error and the other is video vision_position_ids mismatch; distinct code paths." + }, + { + "left": "issue:43421", + "right": "issue:44186", + "accept": false, + "reason": "Both are tokenizer-related, but 43421 is about special-token post-processing updates and 44186 is LayoutLMv2 batching/padding/truncation crashes." + }, + { + "left": "issue:46032", + "right": "issue:46463", + "accept": false, + "reason": "One is a cache correctness bug in Mamba2Mixer, the other is an error-message validation issue for T5 decoder inputs; not the same problem." + }, + { + "left": "issue:44857", + "right": "issue:47050", + "accept": false, + "reason": "Both are dtype/AMP-adjacent runtime issues, but one is LwDetrImageLoss under float16 AMP and the other is CanineModel ONNX export attention dtype mismatch." + }, + { + "left": "issue:45356", + "right": "issue:45907", + "accept": false, + "reason": "One is a tokenizer regression, the other is a hub API forwarding issue (`list_repo_files` revision); unrelated." + }, + { + "left": "issue:39692", + "right": "issue:47050", + "accept": false, + "reason": "One is a SigLIP2 documentation/example mismatch, the other is an fp16 ONNX export dtype bug in CanineModel; different artifacts and failures." + }, + { + "left": "issue:43596", + "right": "issue:44220", + "accept": false, + "reason": "Both are runtime errors, but one is DeepSpeed ZeRO-3 init with BertModel and the other is `_torch_extract_fbank_features()`; no concrete overlap." + }, + { + "left": "issue:43720", + "right": "issue:43819", + "accept": false, + "reason": "One is about unpacking packed weights during accelerate loading, the other is a DAC latent/STE mismatch; unrelated model internals." + }, + { + "left": "issue:43295", + "right": "issue:45834", + "accept": false, + "reason": "Both are regression-style correctness issues, but one is processor/tokenizer access in custom code and the other is OCR bbox indexing on long inputs." + }, + { + "left": "issue:43540", + "right": "issue:46294", + "accept": false, + "reason": "Both involve Qwen video inputs, but one is a ValueError in Qwen3OmniMoe video processing and the other is a chat-template token/feature count mismatch." + }, + { + "left": "issue:43526", + "right": "issue:43881", + "accept": false, + "reason": "One is a BeitImageProcessorFast label-reduction bug, the other is a glm-4v-9b load failure; different components and symptoms." + }, + { + "left": "issue:43550", + "right": "issue:43927", + "accept": false, + "reason": "One is Bamba-9B-v2 torch.compile + SDPA failure, the other is DiaConfig save/load losing custom token IDs; unrelated." + }, + { + "left": "issue:43550", + "right": "issue:43819", + "accept": false, + "reason": "Different models and failure modes: compile/SDPA attention bug versus DAC latent consistency bug." + }, + { + "left": "issue:44488", + "right": "issue:47020", + "accept": false, + "reason": "Both are tokenizer/loading issues, but one is a model failing to load and the other is XLMRobertaTokenizer passing the wrong type to Unigram." + }, + { + "left": "issue:44062", + "right": "issue:44488", + "accept": false, + "reason": "One is an AddedToken argument duplication error, the other is a tokenizer loading failure for a specific model; not the same defect." + }, + { + "left": "issue:43531", + "right": "issue:43643", + "accept": false, + "reason": "Sliding-window behavior in Qwen3-MoE is unrelated to AutoConfig missing fields with trust_remote_code=True." + }, + { + "left": "issue:43653", + "right": "issue:43720", + "accept": false, + "reason": "Tokenizer special-token registration and packed-weight unloading are entirely different subsystems." + }, + { + "left": "issue:45072", + "right": "issue:47050", + "accept": false, + "reason": "Both mention dtype mismatches, but one is bfloat16 inference in SwitchTransformers/TimmWrapperModel and the other is fp16 ONNX export attention mismatch." + }, + { + "left": "issue:41720", + "right": "issue:44991", + "accept": false, + "reason": "One is a CUDA assert during auto device mapping for Qwen3; the other is a tokenizer loading failure for EMBEDDIA/est-roberta." + }, + { + "left": "issue:30064", + "right": "issue:36331", + "accept": false, + "reason": "An image-processor void segmentation-map bug is unrelated to a custom trainer compute_loss signature regression." + }, + { + "left": "issue:45381", + "right": "issue:46693", + "accept": false, + "reason": "One is a video position-id bug for qwen2.5-vl, the other is a long-context FA2 performance regression; different code paths and outcomes." + } + ] + }, + "evaluator_result": null, + "evaluator_used": false, + "retried": false, + "accepted_nontrivial_soft_edge": false, + "error_kind": null, + "error_message": null + }, + { + "label": "issue", + "component_index": 1, + "component_count": 28, + "review_unit_index": 20, + "review_unit_count": 65, + "status": "reviewed", + "reason": null, + "source": "llm", + "cache_hit": false, + "model": "gpt-5.4-mini?service_tier=flex", + "cluster_id": "cluster-10000-530", + "nodes": [ + "issue:34634", + "issue:36010", + "issue:39692", + "issue:43322", + "issue:43335", + "issue:43421", + "issue:43540", + "issue:43606", + "issue:43611", + "issue:43618", + "issue:43643", + "issue:43653", + "issue:43697", + "issue:43756", + "issue:43819", + "issue:44008", + "issue:44186", + "issue:44220", + "issue:44442", + "issue:44448", + "issue:44561", + "issue:44661", + "issue:44671", + "issue:44805", + "issue:44811", + "issue:44821", + "issue:44843", + "issue:44877", + "issue:44898", + "issue:44936", + "issue:44991", + "issue:45020", + "issue:45072", + "issue:45081", + "issue:45084", + "issue:45584", + "issue:45685", + "issue:45834", + "issue:45987", + "issue:46291", + "issue:46294", + "issue:46463", + "issue:46665", + "issue:46829", + "issue:46832", + "issue:46858", + "issue:46932", + "issue:47020" + ], + "soft_pairs": [ + "issue:44220|issue:44661", + "issue:44821|issue:45084", + "issue:43611|issue:46665", + "issue:46829|issue:46832", + "issue:44186|issue:44898", + "issue:43653|issue:43819", + "issue:43540|issue:44008", + "issue:44991|issue:47020", + "issue:34634|issue:36010", + "issue:44843|issue:45081", + "issue:43643|issue:44877", + "issue:45987|issue:46829", + "issue:43335|issue:43756", + "issue:44008|issue:46858", + "issue:39692|issue:45834", + "issue:44561|issue:45020", + "issue:44186|issue:44442", + "issue:44448|issue:44671", + "issue:45987|issue:46932", + "issue:43618|issue:43697", + "issue:44805|issue:44936", + "issue:44991|issue:46291", + "issue:43322|issue:45685", + "issue:45584|issue:45834", + "issue:43606|issue:45072", + "issue:43421|issue:44811", + "issue:46294|issue:46463" + ], + "prepared_review_unit_hash": "dffc3460cdb5c149203c5b27f085b37385d0bc7ee5afe599af16697ea5fb9dda", + "component_budget": { + "node_count": 530, + "item_count": 530, + "soft_pair_count": 1954, + "serialized_chars": 1013004, + "estimated_input_tokens": 253251, + "estimated_eval_tokens": 506758 + }, + "budget": { + "node_count": 48, + "item_count": 48, + "soft_pair_count": 27, + "serialized_chars": 36051, + "estimated_input_tokens": 9013, + "estimated_eval_tokens": 18282 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": true, + "analyst_result": { + "summary": "This cluster is noisy and mostly heterogeneous: the paired items share broad HF/Transformers vocabulary, but each points to a different concrete bug path. I did not find any soft pair that looks like the same underlying issue.", + "confidence": 0.84, + "canonical_issue_reason": "issue:45020 is the broadest umbrella-style report in the set and best represents the general v5/remote_code regression theme, even though the cluster overall is not a true duplicate group.", + "canonical_pr_reason": null, + "best_issue_reason": "issue:45020 has the widest scope and highest chance of serving as a central tracker for adjacent loading/regression problems.", + "best_pr_reason": null, + "soft_edge_verdicts": [ + { + "left": "issue:44220", + "right": "issue:44661", + "accept": false, + "reason": "Different subsystems and failure modes: audio feature extraction vs tokenizer-mapping/model-addition logic." + }, + { + "left": "issue:44821", + "right": "issue:45084", + "accept": false, + "reason": "URL image processor loading and template-node compilation are unrelated code paths." + }, + { + "left": "issue:43611", + "right": "issue:46665", + "accept": false, + "reason": "Both involve loading, but one is a base_model_prefix v5 breakage and the other is a broader from_pretrained-in-classes inconsistency." + }, + { + "left": "issue:46829", + "right": "issue:46832", + "accept": false, + "reason": "Different model families and bugs: output field mismatch vs RT-DETR forward shape/type error." + }, + { + "left": "issue:44186", + "right": "issue:44898", + "accept": false, + "reason": "Tokenizer padding/NER crash is unrelated to Perceiver image-classification interpolation behavior." + }, + { + "left": "issue:43653", + "right": "issue:43819", + "accept": false, + "reason": "Tokenizer special-token registration and DAC latent conversion are separate components with no shared code path." + }, + { + "left": "issue:43540", + "right": "issue:44008", + "accept": false, + "reason": "Distinct multimodal model bugs: video input validation vs Gemma3n forward variable collision." + }, + { + "left": "issue:44991", + "right": "issue:47020", + "accept": false, + "reason": "Both are tokenizer-loading problems, but they affect different tokenizer classes and likely different constructor paths." + }, + { + "left": "issue:34634", + "right": "issue:36010", + "accept": false, + "reason": "BarkProcessor voice preset behavior is unrelated to the GenerationMixin import error." + }, + { + "left": "issue:44843", + "right": "issue:45081", + "accept": false, + "reason": "Same helper name, but different concrete failures: offline hub access vs backend_tokenizer attribute crash." + }, + { + "left": "issue:43643", + "right": "issue:44877", + "accept": false, + "reason": "Different config-loading failures: missing fields from trust_remote_code vs strict config rejecting granite_speech." + }, + { + "left": "issue:45987", + "right": "issue:46829", + "accept": false, + "reason": "StaticCache sequence-length contract bug is unrelated to VideoPrism output tensor assignment." + }, + { + "left": "issue:43335", + "right": "issue:43756", + "accept": false, + "reason": "One is sparse-layer creation from zero sparse layers; the other is RoPE layer dropping in Smollm3." + }, + { + "left": "issue:44008", + "right": "issue:46858", + "accept": false, + "reason": "Both mention tensor/device issues, but they arise in different models and different forward/cache logic." + }, + { + "left": "issue:39692", + "right": "issue:45834", + "accept": false, + "reason": "Documentation example errors and long-ocr index error are different problems in different models." + }, + { + "left": "issue:44561", + "right": "issue:45020", + "accept": false, + "reason": "Both are v5 regressions, but one is specifically is_torch_fx_available removal while the other is a broader remote_code breakage report." + }, + { + "left": "issue:44186", + "right": "issue:44442", + "accept": false, + "reason": "Different tokenizer/model classes and distinct crash conditions." + }, + { + "left": "issue:44448", + "right": "issue:44671", + "accept": false, + "reason": "Version-v4/v5 output drift in different models; no shared code path is evident." + }, + { + "left": "issue:45987", + "right": "issue:46932", + "accept": false, + "reason": "StaticCache length return type and Gemma4 dtype casting are unrelated issues." + }, + { + "left": "issue:43618", + "right": "issue:43697", + "accept": false, + "reason": "CLIP attentions field assignment and RTDetrV2 output divergence are separate model-output bugs." + }, + { + "left": "issue:44805", + "right": "issue:44936", + "accept": false, + "reason": "Mask shape IndexError and trainer evaluate-after-train failure are unrelated training/runtime problems." + }, + { + "left": "issue:44991", + "right": "issue:46291", + "accept": false, + "reason": "Tokenizer loading failure and pipelines import failure are different entrypoints and failure modes." + }, + { + "left": "issue:43322", + "right": "issue:45685", + "accept": false, + "reason": "Llava Next segfault and MPS histogram/int dtype error are not the same bug." + }, + { + "left": "issue:45584", + "right": "issue:45834", + "accept": false, + "reason": "Both are multimodal generation issues, but empty-transcription Whisper behavior and Kosmos2.5 long OCR indexing are different code paths." + }, + { + "left": "issue:43606", + "right": "issue:45072", + "accept": false, + "reason": "CPU offload device mismatch and bfloat16 dtype mismatch are separate backend-specific bugs." + }, + { + "left": "issue:43421", + "right": "issue:44811", + "accept": false, + "reason": "Runtime special-token/post-processor updates and Whisper batch_decode skip_special_tokens behavior are distinct tokenizer/processor issues." + }, + { + "left": "issue:46294", + "right": "issue:46463", + "accept": false, + "reason": "Qwen video chat-template mismatch and T5 decoder-input validation are unrelated models and validations." + } + ] + }, + "evaluator_result": null, + "evaluator_used": false, + "retried": false, + "accepted_nontrivial_soft_edge": false, + "error_kind": null, + "error_message": null + }, + { + "label": "issue", + "component_index": 1, + "component_count": 28, + "review_unit_index": 21, + "review_unit_count": 65, + "status": "reviewed", + "reason": null, + "source": "llm", + "cache_hit": false, + "model": "gpt-5.4-mini?service_tier=flex", + "cluster_id": "cluster-10000-530", + "nodes": [ + "issue:28218", + "issue:29159", + "issue:34634", + "issue:36331", + "issue:42175", + "issue:42915", + "issue:43388", + "issue:43540", + "issue:43550", + "issue:43577", + "issue:43582", + "issue:43644", + "issue:43645", + "issue:43716", + "issue:43720", + "issue:43742", + "issue:43749", + "issue:43782", + "issue:43818", + "issue:43825", + "issue:43931", + "issue:43994", + "issue:44008", + "issue:44291", + "issue:44479", + "issue:44610", + "issue:44671", + "issue:44977", + "issue:45020", + "issue:45072", + "issue:45081", + "issue:45183", + "issue:45381", + "issue:45479", + "issue:45563", + "issue:45701", + "issue:45834", + "issue:46032", + "issue:46512", + "issue:46665", + "issue:46752", + "issue:46846", + "issue:47254", + "issue:47276", + "issue:47332", + "issue:47333", + "issue:47337" + ], + "soft_pairs": [ + "issue:43550|issue:45072", + "issue:46752|issue:47337", + "issue:44479|issue:45834", + "issue:43716|issue:44008", + "issue:43782|issue:44977", + "issue:43644|issue:43749", + "issue:43577|issue:43818", + "issue:28218|issue:29159", + "issue:45183|issue:45479", + "issue:43582|issue:46512", + "issue:45563|issue:47276", + "issue:43931|issue:46846", + "issue:44610|issue:45072", + "issue:43825|issue:43994", + "issue:43540|issue:45381", + "issue:43749|issue:44291", + "issue:34634|issue:36331", + "issue:43645|issue:46665", + "issue:43720|issue:45072", + "issue:47276|issue:47337", + "issue:47254|issue:47333", + "issue:42175|issue:45081", + "issue:43388|issue:46032", + "issue:44671|issue:45701", + "issue:43742|issue:45020", + "issue:47254|issue:47332", + "issue:43577|issue:43582", + "issue:42915|issue:47333" + ], + "prepared_review_unit_hash": "8f595d4afb45226d037ef6bdb7af21ad21c45628071c7db0a349d4b4cb6ef2fc", + "component_budget": { + "node_count": 530, + "item_count": 530, + "soft_pair_count": 1954, + "serialized_chars": 1013004, + "estimated_input_tokens": 253251, + "estimated_eval_tokens": 506758 + }, + "budget": { + "node_count": 47, + "item_count": 47, + "soft_pair_count": 28, + "serialized_chars": 35750, + "estimated_input_tokens": 8938, + "estimated_eval_tokens": 18132 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": true, + "analyst_result": { + "summary": "The cluster is heterogeneous: the soft pairs are mostly thematically similar only at a broad subsystem level, but they do not describe the same concrete bug or change. I would not merge any of the candidate edges.", + "confidence": 0.92, + "canonical_issue_reason": "No single canonical issue stands out. The set spans unrelated failures across tokenization, model loading, dtype handling, video inputs, training, offloading, and warnings, so the soft matches are too weak for duplicate triage.", + "canonical_pr_reason": null, + "best_issue_reason": "There is no clear global best issue for deduplication because the cluster is not a true duplicate set. If forced to pick one representative, the most general/central wording is still not enough to subsume the rest.", + "best_pr_reason": null, + "soft_edge_verdicts": [ + { + "left": "issue:43550", + "right": "issue:45072", + "accept": false, + "reason": "Different bugs: torch.compile/SDPA failure vs dtype mismatch during inference." + }, + { + "left": "issue:46752", + "right": "issue:47337", + "accept": false, + "reason": "Empty chat template input crash is unrelated to the bbox range flaky test." + }, + { + "left": "issue:44479", + "right": "issue:45834", + "accept": false, + "reason": "Video-input regression across Qwen VL models is unrelated to Kosmos2.5 long OCR indexing." + }, + { + "left": "issue:43716", + "right": "issue:44008", + "accept": false, + "reason": "Different root causes: image-preprocessor dtype mismatch vs variable name collision in Gemma3nModel.forward()." + }, + { + "left": "issue:43782", + "right": "issue:44977", + "accept": false, + "reason": "Qwen3VL weight_only load error is not the same as Qwen3.5 flash-attention generation failure." + }, + { + "left": "issue:43644", + "right": "issue:43749", + "accept": false, + "reason": "Both concern loading, but one is junk-filled buffers and the other is FSDP CPU RAM efficient loading." + }, + { + "left": "issue:43577", + "right": "issue:43818", + "accept": false, + "reason": "Blip2 dtype propagation bug is unrelated to missing temporal attention/shared weights in Video-LLaVA." + }, + { + "left": "issue:28218", + "right": "issue:29159", + "accept": false, + "reason": "Both are tokenizer-related, but one is an added-token spacing bug and the other is slow/fast inconsistency more broadly." + }, + { + "left": "issue:45183", + "right": "issue:45479", + "accept": false, + "reason": "Input validation error messaging is unrelated to the single-label classification zero-loss bug." + }, + { + "left": "issue:43582", + "right": "issue:46512", + "accept": false, + "reason": "Apple Silicon warmup TypeError is unrelated to stale model.config.dtype after from_config(dtype=...)." + }, + { + "left": "issue:45563", + "right": "issue:47276", + "accept": false, + "reason": "Both are warning-message issues, but they refer to different warnings in different generation paths." + }, + { + "left": "issue:43931", + "right": "issue:46846", + "accept": false, + "reason": "Shape-mismatch loading failure is not the same as missing base_model_tp_plan entries causing TP OOM/shape errors." + }, + { + "left": "issue:44610", + "right": "issue:45072", + "accept": false, + "reason": "Processor image-size mismatch is unrelated to bfloat16 dtype mismatches in inference wrappers." + }, + { + "left": "issue:43825", + "right": "issue:43994", + "accept": false, + "reason": "Unsupported translation-task messaging is unrelated to SigLIP2 nonsense outputs." + }, + { + "left": "issue:43540", + "right": "issue:45381", + "accept": false, + "reason": "Both touch video inputs, but the concrete failures and affected models differ." + }, + { + "left": "issue:43749", + "right": "issue:44291", + "accept": false, + "reason": "FSDP CPU RAM efficient loading failure is unrelated to init_empty_weights receiving an unexpected argument." + }, + { + "left": "issue:34634", + "right": "issue:36331", + "accept": false, + "reason": "BarkProcessor voice preset handling and CustomTrainer.compute_loss signature are unrelated." + }, + { + "left": "issue:43645", + "right": "issue:46665", + "accept": false, + "reason": "Notebook custom-model initialization breakage is not the same as from_pretrained behavior inside Python classes." + }, + { + "left": "issue:43720", + "right": "issue:45072", + "accept": false, + "reason": "Packed-weight unpacking during accelerate loading is a different failure mode from dtype mismatch inference bugs." + }, + { + "left": "issue:47276", + "right": "issue:47337", + "accept": false, + "reason": "A stale synced_gpus warning is unrelated to the bbox coordinate flaky test." + }, + { + "left": "issue:47254", + "right": "issue:47333", + "accept": false, + "reason": "Checkpointing/ZeRO-3 failure and disk-offloaded Qwen3 save failure are different code paths and symptoms." + }, + { + "left": "issue:42175", + "right": "issue:45081", + "accept": false, + "reason": "Backend packaging issue is unrelated to the Mistral tokenizer regex patch crash." + }, + { + "left": "issue:43388", + "right": "issue:46032", + "accept": false, + "reason": "Metric-gathering label truncation is unrelated to Mamba2Mixer use_cache correctness." + }, + { + "left": "issue:44671", + "right": "issue:45701", + "accept": false, + "reason": "CamemBERT masked-LM regression is not the same as version-dependent tokenization changes." + }, + { + "left": "issue:43742", + "right": "issue:45020", + "accept": false, + "reason": "MobileLLM key error and remote_code breakage are different loading failures." + }, + { + "left": "issue:47254", + "right": "issue:47332", + "accept": false, + "reason": "Gradient-checkpointing checkpoint errors are unrelated to saving disk-offloaded Qwen3-VL models." + }, + { + "left": "issue:43577", + "right": "issue:43582", + "accept": false, + "reason": "Blip2 dtype issues and Apple Silicon warmup TypeError are unrelated." + }, + { + "left": "issue:42915", + "right": "issue:47333", + "accept": false, + "reason": "Qwen3Moe FP8 config failure is unrelated to disk-offloaded Qwen3 save errors." + } + ] + }, + "evaluator_result": null, + "evaluator_used": false, + "retried": false, + "accepted_nontrivial_soft_edge": false, + "error_kind": null, + "error_message": null + }, + { + "label": "issue", + "component_index": 1, + "component_count": 28, + "review_unit_index": 22, + "review_unit_count": 65, + "status": "reviewed", + "reason": null, + "source": "llm", + "cache_hit": false, + "model": "gpt-5.4-mini?service_tier=flex", + "cluster_id": "cluster-10000-530", + "nodes": [ + "issue:36010", + "issue:36331", + "issue:42175", + "issue:43257", + "issue:43262", + "issue:43329", + "issue:43452", + "issue:43526", + "issue:43531", + "issue:43550", + "issue:43582", + "issue:43638", + "issue:43645", + "issue:43653", + "issue:43697", + "issue:43782", + "issue:43819", + "issue:43825", + "issue:43827", + "issue:43854", + "issue:43906", + "issue:43927", + "issue:44060", + "issue:44368", + "issue:44589", + "issue:44857", + "issue:44912", + "issue:44936", + "issue:44964", + "issue:44977", + "issue:44991", + "issue:45003", + "issue:45042", + "issue:45081", + "issue:45325", + "issue:45372", + "issue:45406", + "issue:46175", + "issue:46209", + "issue:46285", + "issue:46291", + "issue:46326", + "issue:46697", + "issue:46728", + "issue:46932", + "issue:47217", + "issue:47337" + ], + "soft_pairs": [ + "issue:46697|issue:46728", + "issue:44589|issue:45042", + "issue:43262|issue:47217", + "issue:43782|issue:45325", + "issue:36010|issue:43645", + "issue:43653|issue:44912", + "issue:43262|issue:43819", + "issue:46175|issue:46932", + "issue:44964|issue:47337", + "issue:45003|issue:46285", + "issue:43582|issue:44857", + "issue:36010|issue:36331", + "issue:43825|issue:43827", + "issue:43531|issue:45042", + "issue:43638|issue:44936", + "issue:43550|issue:44912", + "issue:45372|issue:46326", + "issue:42175|issue:46291", + "issue:43452|issue:44991", + "issue:43257|issue:43329", + "issue:45081|issue:45406", + "issue:43819|issue:43927", + "issue:43906|issue:46209", + "issue:43526|issue:43697", + "issue:44060|issue:44368", + "issue:43854|issue:44977" + ], + "prepared_review_unit_hash": "7dd8c8d3d2ad771da3e9e30641da6b27f88cc8e3078047e351e81d69442430a9", + "component_budget": { + "node_count": 530, + "item_count": 530, + "soft_pair_count": 1954, + "serialized_chars": 1013004, + "estimated_input_tokens": 253251, + "estimated_eval_tokens": 506758 + }, + "budget": { + "node_count": 47, + "item_count": 47, + "soft_pair_count": 26, + "serialized_chars": 35500, + "estimated_input_tokens": 8875, + "estimated_eval_tokens": 18006 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": true, + "analyst_result": { + "summary": "The cluster is broadly heterogeneous: the soft-similarity pairs mostly share a subsystem name or API surface, but they do not look like the same underlying bug. No PRs are present.", + "confidence": 0.93, + "canonical_issue_reason": "No clear canonical issue: the issues span unrelated failure modes across tokenization, vision, training, quantization, and docs, so there is no single shared root cause.", + "canonical_pr_reason": null, + "best_issue_reason": "None of the issues is a strong cluster anchor; even the more discussed items target different code paths and should not be treated as duplicates.", + "best_pr_reason": null, + "soft_edge_verdicts": [ + { + "left": "issue:46697", + "right": "issue:46728", + "accept": false, + "reason": "Both are image-processor bugs, but one is over-padding in Swin2SR and the other is a wrong patch-count calculation in Aria; different code paths and symptoms." + }, + { + "left": "issue:44589", + "right": "issue:45042", + "accept": false, + "reason": "One is an FP8 storage/type error, the other is an image-processor torchvision dependency issue; unrelated failures." + }, + { + "left": "issue:43262", + "right": "issue:47217", + "accept": false, + "reason": "Audio sampling-rate defaults and VL image-placeholder handling are different subsystems with different bug mechanisms." + }, + { + "left": "issue:43782", + "right": "issue:45325", + "accept": false, + "reason": "Different Qwen2.5-VL/Qwen3VL model issues: one is a from_pretrained weight_only error, the other is a rope index scaling bug." + }, + { + "left": "issue:36010", + "right": "issue:43645", + "accept": false, + "reason": "Both involve Transformers v5 breakage, but one is an import-path regression and the other is notebook custom-model initialization; not the same bug." + }, + { + "left": "issue:43653", + "right": "issue:44912", + "accept": false, + "reason": "Tokenizer special-token registration is unrelated to MXFP4 quantization fallback." + }, + { + "left": "issue:43262", + "right": "issue:43819", + "accept": false, + "reason": "Audio chat-template sampling rate and DAC latent reconstruction are unrelated features." + }, + { + "left": "issue:46175", + "right": "issue:46932", + "accept": false, + "reason": "Both are Gemma4-related, but one is a huge one-hot position-embedding implementation and the other is a dtype-casting fix for quantized embedders." + }, + { + "left": "issue:44964", + "right": "issue:47337", + "accept": false, + "reason": "A multimodal model load failure and a flaky bbox-range test are unrelated." + }, + { + "left": "issue:45003", + "right": "issue:46285", + "accept": false, + "reason": "Unsafe sys.modules access and CLIP text-model structure changes are different issues." + }, + { + "left": "issue:43582", + "right": "issue:44857", + "accept": false, + "reason": "Apple Silicon allocator warmup TypeError and CUDA AMP loss crash are unrelated runtime problems." + }, + { + "left": "issue:36010", + "right": "issue:36331", + "accept": false, + "reason": "ImportError for GenerationMixin and CustomTrainer.compute_loss signature mismatch are separate regressions." + }, + { + "left": "issue:43825", + "right": "issue:43827", + "accept": false, + "reason": "One is a bad runtime error message, the other is stale docs; related topic but not the same concrete bug." + }, + { + "left": "issue:43531", + "right": "issue:45042", + "accept": false, + "reason": "Qwen3-MoE sliding-window behavior and PIL backend torchvision dependency are unrelated." + }, + { + "left": "issue:43638", + "right": "issue:44936", + "accept": false, + "reason": "DeepSpeed ZeRO-3 initialization failure and evaluate-after-train failure are different trainer bugs." + }, + { + "left": "issue:43550", + "right": "issue:44912", + "accept": false, + "reason": "torch.compile with SDPA and MXFP4 quantization fallback are different model/runtime paths." + }, + { + "left": "issue:45372", + "right": "issue:46326", + "accept": false, + "reason": "Gemma4 processor loading import failure and mistral_common continue_final_message behavior are unrelated." + }, + { + "left": "issue:42175", + "right": "issue:46291", + "accept": false, + "reason": "Packaging/backend dependency omission and pipelines import failure with kernels v0.15 are different dependency issues." + }, + { + "left": "issue:43452", + "right": "issue:44991", + "accept": false, + "reason": "gguf_file handling and tokenizer loading for a specific model are separate from different loaders." + }, + { + "left": "issue:43257", + "right": "issue:43329", + "accept": false, + "reason": "Qwen3 MoE weight conversion and undefined video-token helper usage are unrelated bugs." + }, + { + "left": "issue:45081", + "right": "issue:45406", + "accept": false, + "reason": "Both hit tokenizer/processor loading, but one is a Mistral regex patch crash and the other is a Gemma4Processor missing tokenizer attribute." + }, + { + "left": "issue:43819", + "right": "issue:43927", + "accept": false, + "reason": "DAC latent reconstruction mismatch and DiaConfig token-id persistence loss are unrelated model/config bugs." + }, + { + "left": "issue:43906", + "right": "issue:46209", + "accept": false, + "reason": "A reproduction of an older issue and an FP8 inference failure on B200 are not the same root cause." + }, + { + "left": "issue:43526", + "right": "issue:43697", + "accept": false, + "reason": "A BEiT image-processor label reduction bug and an RTDetrV2 output regression are different code paths." + }, + { + "left": "issue:44060", + "right": "issue:44368", + "accept": false, + "reason": "Both mention tied-weights warnings, but they concern different Qwen model families and different warning causes." + }, + { + "left": "issue:43854", + "right": "issue:44977", + "accept": false, + "reason": "GLM-4.7-Flash test loading and Qwen3.5 flash-attention generation issues are unrelated model-specific failures." + } + ] + }, + "evaluator_result": null, + "evaluator_used": false, + "retried": false, + "accepted_nontrivial_soft_edge": false, + "error_kind": null, + "error_message": null + }, + { + "label": "issue", + "component_index": 1, + "component_count": 28, + "review_unit_index": 23, + "review_unit_count": 65, + "status": "reviewed", + "reason": null, + "source": "llm", + "cache_hit": false, + "model": "gpt-5.4-mini?service_tier=flex", + "cluster_id": "cluster-10000-530", + "nodes": [ + "issue:28282", + "issue:29159", + "issue:29870", + "issue:34567", + "issue:34689", + "issue:38175", + "issue:39692", + "issue:40990", + "issue:42490", + "issue:42915", + "issue:43257", + "issue:43262", + "issue:43329", + "issue:43425", + "issue:43450", + "issue:43525", + "issue:43526", + "issue:43643", + "issue:43696", + "issue:43742", + "issue:43746", + "issue:43931", + "issue:44291", + "issue:44387", + "issue:44479", + "issue:44554", + "issue:44589", + "issue:44610", + "issue:44617", + "issue:44811", + "issue:44821", + "issue:44912", + "issue:44987", + "issue:45042", + "issue:45081", + "issue:45084", + "issue:45459", + "issue:45538", + "issue:45600", + "issue:45759", + "issue:45803", + "issue:45859", + "issue:46291", + "issue:47020", + "issue:47050", + "issue:47145", + "issue:47333" + ], + "soft_pairs": [ + "issue:42915|issue:44821", + "issue:43746|issue:44811", + "issue:43257|issue:43696", + "issue:45084|issue:45538", + "issue:45859|issue:47050", + "issue:43643|issue:45759", + "issue:38175|issue:44610", + "issue:42490|issue:43257", + "issue:43262|issue:44811", + "issue:42915|issue:43742", + "issue:43329|issue:45600", + "issue:43329|issue:44479", + "issue:44291|issue:45042", + "issue:43526|issue:44554", + "issue:43329|issue:43450", + "issue:39692|issue:44912", + "issue:34567|issue:40990", + "issue:29159|issue:34567", + "issue:44387|issue:44617", + "issue:44589|issue:47333", + "issue:43257|issue:43931", + "issue:45081|issue:47020", + "issue:44987|issue:45538", + "issue:43425|issue:46291", + "issue:44912|issue:47145", + "issue:43525|issue:46291", + "issue:28282|issue:29870", + "issue:45459|issue:45803", + "issue:28282|issue:34689" + ], + "prepared_review_unit_hash": "6881fd00cbc7e32cc0e7e3a2047c5eb884b303ea908f27c85b8d982a862162ac", + "component_budget": { + "node_count": 530, + "item_count": 530, + "soft_pair_count": 1954, + "serialized_chars": 1013004, + "estimated_input_tokens": 253251, + "estimated_eval_tokens": 506758 + }, + "budget": { + "node_count": 47, + "item_count": 47, + "soft_pair_count": 29, + "serialized_chars": 36095, + "estimated_input_tokens": 9024, + "estimated_eval_tokens": 18304 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": true, + "analyst_result": { + "summary": "All proposed soft edges look like false positives: the items span unrelated bugs across model loading, quantization, tokenizers, processors, video/audio paths, and trainer state. No single canonical issue stands out as a safe duplicate target.", + "confidence": 0.96, + "canonical_issue_reason": null, + "canonical_pr_reason": null, + "best_issue_reason": null, + "best_pr_reason": null, + "soft_edge_verdicts": [ + { + "left": "issue:42915", + "right": "issue:44821", + "accept": false, + "reason": "Qwen3Moe FineGrainedFP8Config failure vs AutoImageProcessor URL loading; different subsystems and symptoms." + }, + { + "left": "issue:43746", + "right": "issue:44811", + "accept": false, + "reason": "PEFT local-checkpoint loading bug vs Whisper batch_decode skip_special_tokens handling; unrelated paths." + }, + { + "left": "issue:43257", + "right": "issue:43696", + "accept": false, + "reason": "MoE weight conversion under accelerate+deepspeed vs CUDA OOM on GPT-oss-20b; different root causes." + }, + { + "left": "issue:45084", + "right": "issue:45538", + "accept": false, + "reason": "Template compilation TypeError vs CLIPTokenizer model_max_length bug; no shared code path." + }, + { + "left": "issue:45859", + "right": "issue:47050", + "accept": false, + "reason": "CPU offload rotary embedding issue vs fp16 ONNX dtype mismatch; unrelated model/runtime problems." + }, + { + "left": "issue:43643", + "right": "issue:45759", + "accept": false, + "reason": "AutoConfig missing fields with trust_remote_code vs AutoModelForCausalLM text_config unwrapping; distinct config bugs." + }, + { + "left": "issue:38175", + "right": "issue:44610", + "accept": false, + "reason": "Zero probabilities in SigLIP2 vs OmDet-Turbo processor/model input-size mismatch; unrelated model families." + }, + { + "left": "issue:42490", + "right": "issue:43257", + "accept": false, + "reason": "Trailing-slash local-path warning vs Qwen3 MoE weight conversion; different loading issues." + }, + { + "left": "issue:43262", + "right": "issue:44811", + "accept": false, + "reason": "Audio chat-template sampling-rate default vs Whisper batch_decode special-token handling; separate processor behavior." + }, + { + "left": "issue:42915", + "right": "issue:43742", + "accept": false, + "reason": "Qwen3Moe FP8 config failure vs MobileLLM key error on load; different models and failures." + }, + { + "left": "issue:43329", + "right": "issue:45600", + "accept": false, + "reason": "Undefined video token helper vars vs stale auto_mappings config references; unrelated maintenance bugs." + }, + { + "left": "issue:43329", + "right": "issue:44479", + "accept": false, + "reason": "Video token counting helper crash vs model-specific video regression; both video-related but not the same bug." + }, + { + "left": "issue:44291", + "right": "issue:45042", + "accept": false, + "reason": "init_empty_weights argument mismatch vs PIL backend requiring torchvision; unrelated compatibility issues." + }, + { + "left": "issue:43526", + "right": "issue:44554", + "accept": false, + "reason": "BeitImageProcessorFast label reduction bug vs MPS attention correctness issue; different areas entirely." + }, + { + "left": "issue:43329", + "right": "issue:43450", + "accept": false, + "reason": "Undefined multimodal video token helper vs batched video processor output shape bug; not the same code path." + }, + { + "left": "issue:39692", + "right": "issue:44912", + "accept": false, + "reason": "SigLIP2 doc example errors vs gpt-oss MXFP4 quantization fallback; documentation/model-loading mismatch." + }, + { + "left": "issue:34567", + "right": "issue:40990", + "accept": false, + "reason": "TrainerState token counter not updating vs high perplexity on gpt-oss-20b; unrelated training/eval issues." + }, + { + "left": "issue:29159", + "right": "issue:34567", + "accept": false, + "reason": "Slow/fast tokenizer inconsistency vs TrainerState token accounting; different components." + }, + { + "left": "issue:44387", + "right": "issue:44617", + "accept": false, + "reason": "Reserved-memory regression under int4 quantization vs Sam3Video OOM; both memory-related but distinct failures." + }, + { + "left": "issue:44589", + "right": "issue:47333", + "accept": false, + "reason": "Missing Float8 storage class vs disk-offloaded Qwen3 save failure; unrelated serialization/offload bugs." + }, + { + "left": "issue:43257", + "right": "issue:43931", + "accept": false, + "reason": "Qwen3 MoE weight conversion under distributed loading vs Qwen3-VL weight-shape mismatch; different load-time problems." + }, + { + "left": "issue:45081", + "right": "issue:47020", + "accept": false, + "reason": "Mistral regex patch backend_tokenizer crash vs XLMRobertaTokenizer Unigram init bug; distinct tokenizer implementations." + }, + { + "left": "issue:44987", + "right": "issue:45538", + "accept": false, + "reason": "Loading failure for physical-intelligence/fast under 5.1.0 vs CLIPTokenizer max-length constant bug; unrelated." + }, + { + "left": "issue:43425", + "right": "issue:46291", + "accept": false, + "reason": "Torch version incompatibility vs kernels package requiring version/revision; different dependency issues." + }, + { + "left": "issue:44912", + "right": "issue:47145", + "accept": false, + "reason": "MXFP4 quantization fallback to bf16 vs FP8 compressed-tensors missing weight_scale; both quantization-related but different formats and bugs." + }, + { + "left": "issue:43525", + "right": "issue:46291", + "accept": false, + "reason": "Llama4Config missing pad_token_id vs pipelines import failing on kernels dependency; unrelated." + }, + { + "left": "issue:28282", + "right": "issue:29870", + "accept": false, + "reason": "PyTorch missing ImportError vs Bart generate decoder_attention_mask UnboundLocalError; unrelated load vs generation bugs." + }, + { + "left": "issue:45459", + "right": "issue:45803", + "accept": false, + "reason": "Tokenizer decode error handling with protobuf missing vs bare except swallowing interrupts; same exception theme but different functions and bugs." + }, + { + "left": "issue:28282", + "right": "issue:34689", + "accept": false, + "reason": "Missing PyTorch ImportError vs Llama 3.2 Vision model-loading regression; different underlying causes." + } + ] + }, + "evaluator_result": null, + "evaluator_used": false, + "retried": false, + "accepted_nontrivial_soft_edge": false, + "error_kind": null, + "error_message": null + }, + { + "label": "issue", + "component_index": 1, + "component_count": 28, + "review_unit_index": 24, + "review_unit_count": 65, + "status": "reviewed", + "reason": null, + "source": "llm", + "cache_hit": false, + "model": "gpt-5.4-mini?service_tier=flex", + "cluster_id": "cluster-10000-530", + "nodes": [ + "issue:24643", + "issue:28282", + "issue:29127", + "issue:29159", + "issue:33357", + "issue:34567", + "issue:39379", + "issue:42175", + "issue:42915", + "issue:43404", + "issue:43454", + "issue:43531", + "issue:43540", + "issue:43577", + "issue:43582", + "issue:43650", + "issue:43697", + "issue:43742", + "issue:43749", + "issue:43824", + "issue:43906", + "issue:44062", + "issue:44291", + "issue:44479", + "issue:44554", + "issue:44568", + "issue:44589", + "issue:44756", + "issue:44821", + "issue:44857", + "issue:44936", + "issue:44991", + "issue:45127", + "issue:45137", + "issue:45356", + "issue:45357", + "issue:45405", + "issue:45488", + "issue:45538", + "issue:46421", + "issue:46424", + "issue:46693", + "issue:46701", + "issue:46821", + "issue:46832", + "issue:46869", + "issue:47254", + "issue:47332" + ], + "soft_pairs": [ + "issue:43650|issue:43906", + "issue:33357|issue:34567", + "issue:43540|issue:46701", + "issue:43749|issue:44756", + "issue:43454|issue:43577", + "issue:43742|issue:44291", + "issue:44568|issue:45356", + "issue:44291|issue:44821", + "issue:46421|issue:46424", + "issue:44936|issue:46821", + "issue:42175|issue:43824", + "issue:28282|issue:29127", + "issue:44589|issue:47332", + "issue:43404|issue:45127", + "issue:24643|issue:29159", + "issue:45137|issue:45405", + "issue:44479|issue:46693", + "issue:42915|issue:45538", + "issue:43697|issue:45357", + "issue:43582|issue:47254", + "issue:44062|issue:44991", + "issue:43531|issue:44821", + "issue:44857|issue:46832", + "issue:44568|issue:45488", + "issue:39379|issue:44554", + "issue:45538|issue:46869", + "issue:42915|issue:43749" + ], + "prepared_review_unit_hash": "a99a5c582806dc8ac989947efe0e76ac643205ea9b8d771737abf96dfe3982ca", + "component_budget": { + "node_count": 530, + "item_count": 530, + "soft_pair_count": 1954, + "serialized_chars": 1013004, + "estimated_input_tokens": 253251, + "estimated_eval_tokens": 506758 + }, + "budget": { + "node_count": 48, + "item_count": 48, + "soft_pair_count": 27, + "serialized_chars": 35705, + "estimated_input_tokens": 8927, + "estimated_eval_tokens": 18110 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": true, + "analyst_result": { + "summary": "This cluster is mostly a grab bag of unrelated issues across tokenizers, model loading, cache behavior, DeepSpeed/FSDP, and multimodal regressions. The soft pairs share at most broad subsystem similarity, not the same concrete bug, so none should be merged as duplicates.", + "confidence": 0.69, + "canonical_issue_reason": "issue:45488 is the strongest representative: it is a concrete, well-described regression with clear user impact, multiple inbound references, and a specific code-path cause in tokenizer handling.", + "canonical_pr_reason": null, + "best_issue_reason": "issue:45488 is the best standalone issue in the set because it is specific, reproducible, and broad enough to represent a meaningful bug in the cluster.", + "best_pr_reason": null, + "soft_edge_verdicts": [ + { + "left": "issue:43650", + "right": "issue:43906", + "accept": false, + "reason": "One is a vague data request, the other is an isolated reproduction of a separate issue; no shared bug." + }, + { + "left": "issue:33357", + "right": "issue:34567", + "accept": false, + "reason": "MacOS bus error in CLIP loading vs TrainerState token counting; unrelated code paths." + }, + { + "left": "issue:43540", + "right": "issue:46701", + "accept": false, + "reason": "Both are multimodal failures, but they affect different models and different pipelines, so not the same bug." + }, + { + "left": "issue:43749", + "right": "issue:44756", + "accept": false, + "reason": "FSDP CPU RAM efficient loading vs mmap/OOM on Strix Halo are different memory-management problems." + }, + { + "left": "issue:43454", + "right": "issue:43577", + "accept": false, + "reason": "One is a weight-tying bug in AyaVision, the other a dtype-loading issue in BLIP2; different failures." + }, + { + "left": "issue:43742", + "right": "issue:44291", + "accept": false, + "reason": "Model loading KeyError vs init_empty_weights/_is_hf_initialized TypeError; distinct failure mechanisms." + }, + { + "left": "issue:44568", + "right": "issue:45356", + "accept": false, + "reason": "Both tokenizer regressions, but for different models and different root causes; not mergeable as one fix." + }, + { + "left": "issue:44291", + "right": "issue:44821", + "accept": false, + "reason": "init_empty_weights argument handling is unrelated to loading AutoImageProcessor from URL." + }, + { + "left": "issue:46421", + "right": "issue:46424", + "accept": false, + "reason": "Both involve cache behavior, but one is lazy-init recompilation and the other is max_cache_length handling." + }, + { + "left": "issue:44936", + "right": "issue:46821", + "accept": false, + "reason": "Trainer evaluate-after-train bug is unrelated to TimesFM window_size attribute errors." + }, + { + "left": "issue:42175", + "right": "issue:43824", + "accept": false, + "reason": "Packaging/backend dependency issue vs missing Qwen2.5-VL import; unrelated." + }, + { + "left": "issue:28282", + "right": "issue:29127", + "accept": false, + "reason": "Missing PyTorch import dependency vs unclear LayoutLMv3 error message; different bugs." + }, + { + "left": "issue:44589", + "right": "issue:47332", + "accept": false, + "reason": "Float8 storage lookup failure vs saving a disk-offloaded model; not the same underlying defect." + }, + { + "left": "issue:43404", + "right": "issue:45127", + "accept": false, + "reason": "Both mention tied weights, but one is a missing tie in a model, the other is LoRA merge collapse after vocab extension." + }, + { + "left": "issue:24643", + "right": "issue:29159", + "accept": false, + "reason": "DeepSpeed training shape error vs slow/fast tokenizer inconsistency; no overlap beyond being bugs." + }, + { + "left": "issue:45137", + "right": "issue:45405", + "accept": false, + "reason": "ZeRO-3 deque underflow vs a dependency version bump; entirely separate." + }, + { + "left": "issue:44479", + "right": "issue:46693", + "accept": false, + "reason": "Video-input regression and FA2 long-context performance regression are different code paths and symptoms." + }, + { + "left": "issue:42915", + "right": "issue:45538", + "accept": false, + "reason": "FineGrainedFP8Config failure in Qwen3Moe is unrelated to CLIPTokenizer model_max_length." + }, + { + "left": "issue:43697", + "right": "issue:45357", + "accept": false, + "reason": "RT-DETR output drift in v5 vs Qwen3.5 save_pretrained key mismatch; unrelated model families and bugs." + }, + { + "left": "issue:43582", + "right": "issue:47254", + "accept": false, + "reason": "Apple Silicon allocator warmup TypeError vs PEFT/ZeRO-3 checkpointing error; different failures." + }, + { + "left": "issue:44062", + "right": "issue:44991", + "accept": false, + "reason": "AddedToken keyword conflict vs tokenizer loading failure for est-roberta; not the same root cause." + }, + { + "left": "issue:43531", + "right": "issue:44821", + "accept": false, + "reason": "Qwen3-MoE sliding_window behavior vs AutoImageProcessor URL loading; unrelated." + }, + { + "left": "issue:44857", + "right": "issue:46832", + "accept": false, + "reason": "Both are RT-DETR-related, but one is an AMP crash in loss and the other is a forward shape/type issue in a different condition." + }, + { + "left": "issue:44568", + "right": "issue:45488", + "accept": false, + "reason": "Both are tokenizer regressions, but they affect different tokenizer implementations and different incorrect behaviors." + }, + { + "left": "issue:39379", + "right": "issue:44554", + "accept": false, + "reason": "Mllama mask handling is unrelated to the MPS attention correctness issue." + }, + { + "left": "issue:45538", + "right": "issue:46869", + "accept": false, + "reason": "Tokenizer max_length bug vs YOLO device-map/multi-GPU failure; unrelated." + }, + { + "left": "issue:42915", + "right": "issue:43749", + "accept": false, + "reason": "Qwen3Moe FP8 config failure and FSDP CPU RAM efficient loading are separate distributed/runtime issues." + } + ] + }, + "evaluator_result": null, + "evaluator_used": false, + "retried": false, + "accepted_nontrivial_soft_edge": false, + "error_kind": null, + "error_message": null + }, + { + "label": "issue", + "component_index": 1, + "component_count": 28, + "review_unit_index": 25, + "review_unit_count": 65, + "status": "reviewed", + "reason": null, + "source": "llm", + "cache_hit": false, + "model": "gpt-5.4-mini?service_tier=flex", + "cluster_id": "cluster-10000-530", + "nodes": [ + "issue:24643", + "issue:27957", + "issue:28282", + "issue:29870", + "issue:34567", + "issue:36331", + "issue:38175", + "issue:41720", + "issue:42175", + "issue:42915", + "issue:43257", + "issue:43452", + "issue:43540", + "issue:43577", + "issue:43582", + "issue:43716", + "issue:43742", + "issue:43756", + "issue:44062", + "issue:44479", + "issue:44488", + "issue:44560", + "issue:44589", + "issue:44737", + "issue:44792", + "issue:44849", + "issue:44898", + "issue:44964", + "issue:44987", + "issue:45042", + "issue:45072", + "issue:45081", + "issue:45325", + "issue:45381", + "issue:45405", + "issue:45600", + "issue:45672", + "issue:45674", + "issue:45701", + "issue:45823", + "issue:45907", + "issue:46399", + "issue:46463", + "issue:46512", + "issue:46701", + "issue:47276", + "issue:47332", + "issue:47333" + ], + "soft_pairs": [ + "issue:27957|issue:28282", + "issue:41720|issue:43540", + "issue:43540|issue:44479", + "issue:43540|issue:44560", + "issue:43452|issue:43582", + "issue:34567|issue:36331", + "issue:42915|issue:47332", + "issue:43577|issue:46512", + "issue:45042|issue:45081", + "issue:44062|issue:45081", + "issue:42175|issue:44488", + "issue:44479|issue:45381", + "issue:44987|issue:45907", + "issue:38175|issue:44898", + "issue:43742|issue:44589", + "issue:44792|issue:44964", + "issue:44849|issue:45701", + "issue:43742|issue:47333", + "issue:43257|issue:43756", + "issue:43716|issue:46701", + "issue:43756|issue:45325", + "issue:45823|issue:46463", + "issue:43756|issue:46399", + "issue:45672|issue:45674", + "issue:46463|issue:47276", + "issue:44737|issue:46463", + "issue:45405|issue:45600", + "issue:43577|issue:45072", + "issue:24643|issue:29870" + ], + "prepared_review_unit_hash": "8ed6bf0eb36e8d1c3aac1c086ff062d39d500bb9289409ffbdb7654619d508d8", + "component_budget": { + "node_count": 530, + "item_count": 530, + "soft_pair_count": 1954, + "serialized_chars": 1013004, + "estimated_input_tokens": 253251, + "estimated_eval_tokens": 506758 + }, + "budget": { + "node_count": 48, + "item_count": 48, + "soft_pair_count": 29, + "serialized_chars": 36515, + "estimated_input_tokens": 9129, + "estimated_eval_tokens": 18514 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": true, + "analyst_result": { + "summary": "Most items are unrelated, broad issue reports across different models/features. The only clear duplicate-like pair is the Qwen2.5-VL video-input regression between issues 44479 and 45381.", + "confidence": 0.86, + "canonical_issue_reason": "issue:44479 is the better canonical issue: it is earlier and frames the bug as the broader v5.3.0 video-input regression affecting multiple Qwen vision-language variants, which subsumes the narrower follow-up report.", + "canonical_pr_reason": null, + "best_issue_reason": "issue:44479 is the strongest representative for the cluster because it captures the shared regression at the highest level of specificity and scope.", + "best_pr_reason": null, + "soft_edge_verdicts": [ + { + "left": "issue:27957", + "right": "issue:28282", + "accept": false, + "reason": "Different problems: Flash Attention 2 behavior for XLM-Roberta vs a missing PyTorch dependency error." + }, + { + "left": "issue:41720", + "right": "issue:43540", + "accept": false, + "reason": "Different models and failures: Qwen3 auto device mapping cuda assert vs Qwen3OmniMoe video-input ValueError." + }, + { + "left": "issue:43540", + "right": "issue:44479", + "accept": false, + "reason": "Both are video-related, but they target different Qwen model families and different failure modes; not the same concrete bug." + }, + { + "left": "issue:43540", + "right": "issue:44560", + "accept": false, + "reason": "Different error paths and model variants: Qwen3OmniMoe video processing ValueError vs Qwen3-VL-embedding StopIteration." + }, + { + "left": "issue:43452", + "right": "issue:43582", + "accept": false, + "reason": "Unrelated: tokenizer/model loading for gguf_file vs an Apple Silicon allocator warmup TypeError." + }, + { + "left": "issue:34567", + "right": "issue:36331", + "accept": false, + "reason": "Both are Trainer-related, but one is token-count state tracking and the other is a compute_loss signature change." + }, + { + "left": "issue:42915", + "right": "issue:47332", + "accept": false, + "reason": "Different issues: Qwen3Moe FP8 loading failure vs saving a disk-offloaded Qwen3-VL model." + }, + { + "left": "issue:43577", + "right": "issue:46512", + "accept": false, + "reason": "Both mention dtype, but one is a Blip2 loading bug and the other is config.dtype being stale after from_config()." + }, + { + "left": "issue:45042", + "right": "issue:45081", + "accept": false, + "reason": "Different subsystems: image processor torchvision dependency vs Mistral tokenizer regex patch crash." + }, + { + "left": "issue:44062", + "right": "issue:45081", + "accept": false, + "reason": "Unrelated tokenizer bugs with different code paths and symptoms." + }, + { + "left": "issue:42175", + "right": "issue:44488", + "accept": false, + "reason": "Dependency packaging issue vs a specific model loading failure; no shared underlying bug." + }, + { + "left": "issue:44479", + "right": "issue:45381", + "accept": true, + "reason": "Same underlying Qwen2.5-VL video-input regression in transformers 5.3.0; 45381 is a narrower symptom of the broader bug in 44479." + }, + { + "left": "issue:44987", + "right": "issue:45907", + "accept": false, + "reason": "Different API failures: loading a model from transformers>=5.1.0 vs list_repo_files missing revision forwarding." + }, + { + "left": "issue:38175", + "right": "issue:44898", + "accept": false, + "reason": "Different models and failures: SIGLIP2 zero probabilities vs Perceiver image classification failing on non-default resolution." + }, + { + "left": "issue:43742", + "right": "issue:44589", + "accept": false, + "reason": "Different loading/storage errors for unrelated models and dtypes." + }, + { + "left": "issue:44792", + "right": "issue:44964", + "accept": false, + "reason": "Both touch multimodal models, but one is a Janus test failure and the other is a Phi-4-multimodal load regression." + }, + { + "left": "issue:44849", + "right": "issue:45701", + "accept": false, + "reason": "Different bugs: Qwen3.5 output_hidden_states behavior vs tokenization changes across versions." + }, + { + "left": "issue:43742", + "right": "issue:47333", + "accept": false, + "reason": "Unrelated: MobileLLM config/key error vs Qwen3 offloading save failure." + }, + { + "left": "issue:43257", + "right": "issue:43756", + "accept": false, + "reason": "Different Qwen3Moe loading/quantization issue vs Smollm3 RoPE-layer count mismatch." + }, + { + "left": "issue:43716", + "right": "issue:46701", + "accept": false, + "reason": "Different multimodal code paths: image preprocessor dtype mismatch vs DiffusionGemma generation pipeline failure." + }, + { + "left": "issue:43756", + "right": "issue:45325", + "accept": false, + "reason": "Both involve RoPE terminology, but one is model architecture coverage and the other is Qwen2.5-VL temporal position scaling." + }, + { + "left": "issue:45823", + "right": "issue:46463", + "accept": false, + "reason": "Unrelated: Gemma4 device placement during forward vs T5 decoder input validation messaging." + }, + { + "left": "issue:43577", + "right": "issue:45072", + "accept": false, + "reason": "Both are dtype-related, but the affected models and root causes differ; not mergeable as one bug fix." + }, + { + "left": "issue:24643", + "right": "issue:29870", + "accept": false, + "reason": "Different runtime errors in different flows: DeepSpeed weight shape issue vs BART decoder_attention_mask assignment bug." + } + ] + }, + "evaluator_result": { + "accept": true, + "feedback": "The summary is grounded in the packet and stays conservative. The only accepted pair, issue:44479 \u2194 issue:45381, is plausibly the same Qwen2.5-VL video-input regression, and choosing 44479 as the canonical issue is supported by its earlier, broader framing. The other soft-edge rejections do not overstate equivalence." + }, + "evaluator_used": true, + "retried": false, + "accepted_nontrivial_soft_edge": true, + "error_kind": null, + "error_message": null + }, + { + "label": "issue", + "component_index": 1, + "component_count": 28, + "review_unit_index": 26, + "review_unit_count": 65, + "status": "reviewed", + "reason": null, + "source": "llm", + "cache_hit": false, + "model": "gpt-5.4-mini?service_tier=flex", + "cluster_id": "cluster-10000-530", + "nodes": [ + "issue:28282", + "issue:29159", + "issue:32090", + "issue:33357", + "issue:38175", + "issue:41762", + "issue:42796", + "issue:43295", + "issue:43452", + "issue:43540", + "issue:43653", + "issue:43688", + "issue:43697", + "issue:43720", + "issue:43742", + "issue:43749", + "issue:43906", + "issue:43994", + "issue:44361", + "issue:44442", + "issue:44466", + "issue:44488", + "issue:44493", + "issue:44568", + "issue:44625", + "issue:44737", + "issue:44821", + "issue:44843", + "issue:44877", + "issue:44898", + "issue:44913", + "issue:44936", + "issue:45005", + "issue:45216", + "issue:45357", + "issue:45588", + "issue:45663", + "issue:45920", + "issue:46050", + "issue:46121", + "issue:46463", + "issue:46491", + "issue:46693", + "issue:46821", + "issue:46829", + "issue:46846", + "issue:46899", + "issue:47332" + ], + "soft_pairs": [ + "issue:43906|issue:46693", + "issue:44625|issue:44877", + "issue:44568|issue:46491", + "issue:43540|issue:46693", + "issue:43742|issue:47332", + "issue:43452|issue:44488", + "issue:28282|issue:32090", + "issue:44466|issue:45005", + "issue:42796|issue:43906", + "issue:46463|issue:46899", + "issue:41762|issue:44936", + "issue:43452|issue:44843", + "issue:44913|issue:46121", + "issue:45663|issue:46846", + "issue:43697|issue:46829", + "issue:43720|issue:44898", + "issue:44737|issue:45588", + "issue:43653|issue:44361", + "issue:44493|issue:46050", + "issue:38175|issue:43295", + "issue:44821|issue:46821", + "issue:44361|issue:44442", + "issue:43697|issue:45216", + "issue:45357|issue:45920", + "issue:43688|issue:43749", + "issue:43994|issue:44898", + "issue:29159|issue:33357" + ], + "prepared_review_unit_hash": "3f0e0e8e285a65f698d6879c601b5f4766651eea51cfee7832015636706f043c", + "component_budget": { + "node_count": 530, + "item_count": 530, + "soft_pair_count": 1954, + "serialized_chars": 1013004, + "estimated_input_tokens": 253251, + "estimated_eval_tokens": 506758 + }, + "budget": { + "node_count": 48, + "item_count": 48, + "soft_pair_count": 27, + "serialized_chars": 36165, + "estimated_input_tokens": 9042, + "estimated_eval_tokens": 18340 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": true, + "analyst_result": { + "summary": "The set is mostly a collection of unrelated issue reports across tokenizers, model loading, serialization, and distributed/runtime bugs. Only one soft pair looks plausibly like the same tied-weights/save_pretrained regression; the rest are too far apart to merge.", + "confidence": 0.62, + "canonical_issue_reason": "Issue 45920 is the broadest and most representative regression report here: it spans many models, has a clear v5 tokenizer regression framing, and is more general than the other single-model bugs.", + "canonical_pr_reason": null, + "best_issue_reason": "Issue 45920 is the best overall anchor because it affects multiple model families and describes a concrete, high-signal regression rather than a narrow edge case.", + "best_pr_reason": null, + "soft_edge_verdicts": [ + { + "left": "issue:43906", + "right": "issue:46693", + "accept": false, + "reason": "Different problems: one is a reproduction of an earlier issue, the other is a performance regression in long-context generation with FA2." + }, + { + "left": "issue:44625", + "right": "issue:44877", + "accept": false, + "reason": "Unrelated config/loading failures in different model families; no shared underlying code path is evident." + }, + { + "left": "issue:44568", + "right": "issue:46491", + "accept": false, + "reason": "Both are tokenizer bugs, but they affect different tokenizers and different behaviors (special tokens vs leading-space round-trip)." + }, + { + "left": "issue:43540", + "right": "issue:46693", + "accept": false, + "reason": "Video-input processing error in Qwen3OmniMoe is unrelated to the FA2 long-context performance regression." + }, + { + "left": "issue:43742", + "right": "issue:47332", + "accept": false, + "reason": "Different loading/saving paths and different model families; the errors do not point to the same bug." + }, + { + "left": "issue:43452", + "right": "issue:44488", + "accept": false, + "reason": "Both mention loading, but one is a gguf_file/from_pretrained issue while the other is a model-specific load failure; not the same underlying defect." + }, + { + "left": "issue:28282", + "right": "issue:32090", + "accept": false, + "reason": "PyTorch missing ImportError vs GPU broadcast NoneType TypeError are entirely separate failures." + }, + { + "left": "issue:44466", + "right": "issue:45005", + "accept": true, + "reason": "Both point to the same tied-weights/save_pretrained regression in v5, with one describing device-dependent lm_head.weight serialization and the other broader translation-model tied-weight breakage." + }, + { + "left": "issue:42796", + "right": "issue:43906", + "accept": false, + "reason": "Different tokenizer-specific bugs with no shared model or code-path evidence." + }, + { + "left": "issue:46463", + "right": "issue:46899", + "accept": false, + "reason": "T5 decoder-input validation messaging is unrelated to Gemma4 multimodal quantization failures." + }, + { + "left": "issue:41762", + "right": "issue:44936", + "accept": false, + "reason": "Gemma3 ZeRO-3 loading IndexError and trainer.evaluate-after-train behavior are different issues." + }, + { + "left": "issue:43452", + "right": "issue:44843", + "accept": false, + "reason": "Both involve loading, but one is gguf_file handling and the other is offline Hub metadata lookup in Mistral regex patching." + }, + { + "left": "issue:44913", + "right": "issue:46121", + "accept": false, + "reason": "Both touch RoPE/config code, but one is a reload/default regression and the other is a type error on JSON-loaded lists." + }, + { + "left": "issue:45663", + "right": "issue:46846", + "accept": false, + "reason": "Different architectures and different failure modes: FSDP2 per-layer cache rebuilding vs missing TP plan entries." + }, + { + "left": "issue:43697", + "right": "issue:46829", + "accept": false, + "reason": "Different output-shape/return-field bugs in different models; not the same code-path problem." + }, + { + "left": "issue:43720", + "right": "issue:44898", + "accept": false, + "reason": "BitNet accelerate-loading unpacking bug is unrelated to Perceiver interpolation failure." + }, + { + "left": "issue:44737", + "right": "issue:45588", + "accept": false, + "reason": "XLNet CPU placement bug and flash_attention s_aux=None crash are unrelated." + }, + { + "left": "issue:43653", + "right": "issue:44361", + "accept": false, + "reason": "Both are tokenizer-related, but they affect different tokenizers with different symptoms and likely different fixes." + }, + { + "left": "issue:44493", + "right": "issue:46050", + "accept": false, + "reason": "Unexpected key warnings in position IDs and sinq quantization save/load breakage are different issues." + }, + { + "left": "issue:38175", + "right": "issue:43295", + "accept": false, + "reason": "Siglip2 zero-probability outputs and processor.tokenizer regression are different bugs." + }, + { + "left": "issue:44821", + "right": "issue:46821", + "accept": false, + "reason": "AutoImageProcessor URL loading and TimesFM window_size AttributeError are unrelated." + }, + { + "left": "issue:44361", + "right": "issue:44442", + "accept": false, + "reason": "Both mention tokenizer failures, but they concern different tokenizer classes and different failure mechanisms." + }, + { + "left": "issue:43697", + "right": "issue:45216", + "accept": false, + "reason": "RTDetrV2 output drift and Qwen3.5 checkpoint saving are unrelated regressions." + }, + { + "left": "issue:45357", + "right": "issue:45920", + "accept": false, + "reason": "Qwen3.5 visual-encoder save_pretrained bug is unrelated to the broad tokenizer token-id regression across many models." + }, + { + "left": "issue:43688", + "right": "issue:43749", + "accept": false, + "reason": "Auxiliary-loss normalization and FSDP CPU RAM-efficient loading are distinct training/runtime problems." + }, + { + "left": "issue:43994", + "right": "issue:44898", + "accept": false, + "reason": "Different model families and different symptoms; no shared underlying defect is evident." + }, + { + "left": "issue:29159", + "right": "issue:33357", + "accept": false, + "reason": "Tokenizer slow/fast inconsistency and MacOS bus error with CLIP are unrelated." + } + ] + }, + "evaluator_result": { + "accept": true, + "feedback": "The summary is grounded in the packet and stays conservative. The anchor choice of issue 45920 is supported by the title and breadth of affected models, and the lone accepted soft pair (44466/45005) is plausibly the same tied-weights/save_pretrained regression. The other rejections do not overstate similarity and are framed cautiously enough given the limited evidence in the packet." + }, + "evaluator_used": true, + "retried": false, + "accepted_nontrivial_soft_edge": true, + "error_kind": null, + "error_message": null + }, + { + "label": "issue", + "component_index": 1, + "component_count": 28, + "review_unit_index": 27, + "review_unit_count": 65, + "status": "reviewed", + "reason": null, + "source": "llm", + "cache_hit": false, + "model": "gpt-5.4-mini?service_tier=flex", + "cluster_id": "cluster-10000-530", + "nodes": [ + "issue:32090", + "issue:33666", + "issue:36010", + "issue:36331", + "issue:40444", + "issue:41720", + "issue:42898", + "issue:43295", + "issue:43452", + "issue:43525", + "issue:43550", + "issue:43577", + "issue:43643", + "issue:43646", + "issue:43653", + "issue:43688", + "issue:43749", + "issue:43756", + "issue:43761", + "issue:43927", + "issue:44361", + "issue:44464", + "issue:44479", + "issue:44568", + "issue:44625", + "issue:44743", + "issue:44756", + "issue:44811", + "issue:44843", + "issue:44857", + "issue:44991", + "issue:45356", + "issue:45372", + "issue:45375", + "issue:45542", + "issue:45701", + "issue:45704", + "issue:46170", + "issue:46178", + "issue:46402", + "issue:46491", + "issue:46701", + "issue:46829", + "issue:46832", + "issue:46932", + "issue:47000", + "issue:47013", + "issue:47332" + ], + "soft_pairs": [ + "issue:44756|issue:47332", + "issue:40444|issue:44811", + "issue:43643|issue:45375", + "issue:44843|issue:44991", + "issue:45372|issue:46170", + "issue:44857|issue:47000", + "issue:43688|issue:45704", + "issue:43688|issue:45701", + "issue:43295|issue:45356", + "issue:43525|issue:46701", + "issue:45542|issue:46178", + "issue:43646|issue:43749", + "issue:33666|issue:36010", + "issue:43653|issue:43927", + "issue:43653|issue:44568", + "issue:43756|issue:46402", + "issue:43761|issue:47013", + "issue:44464|issue:44743", + "issue:44479|issue:44625", + "issue:46832|issue:46932", + "issue:43550|issue:44361", + "issue:43761|issue:46829", + "issue:43452|issue:43577", + "issue:42898|issue:46491", + "issue:41720|issue:44464", + "issue:33666|issue:36331", + "issue:32090|issue:33666" + ], + "prepared_review_unit_hash": "3be9fb1838b07aee07bc73c5d1b16c65459328b8bb3e3cc87913626f8ef8b95f", + "component_budget": { + "node_count": 530, + "item_count": 530, + "soft_pair_count": 1954, + "serialized_chars": 1013004, + "estimated_input_tokens": 253251, + "estimated_eval_tokens": 506758 + }, + "budget": { + "node_count": 48, + "item_count": 48, + "soft_pair_count": 27, + "serialized_chars": 36098, + "estimated_input_tokens": 9025, + "estimated_eval_tokens": 18306 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": true, + "analyst_result": { + "summary": "I don\u2019t see a true duplicate cluster here: most issues are distinct regressions or model-specific bugs across unrelated subsystems. If an umbrella anchor is needed, #43646 is the broadest generic issue in the set. No PRs were provided.", + "confidence": 0.33, + "canonical_issue_reason": "#43646 is the most general, cross-cutting regression report (\u201cTransformers 5.0.0 breaks custom model initialization\u201d) and is the best umbrella anchor among these mostly unrelated issues.", + "canonical_pr_reason": null, + "best_issue_reason": "#43646 is the best representative issue because it captures a broad class of custom-model breakages, whereas the others are narrower model/version-specific failures or unrelated feature bugs.", + "best_pr_reason": null, + "soft_edge_verdicts": [ + { + "left": "issue:44756", + "right": "issue:47332", + "accept": false, + "reason": "Different problems: Strix Halo mmap/OOM vs saving a disk-offloaded Qwen3-VL model." + }, + { + "left": "issue:40444", + "right": "issue:44811", + "accept": false, + "reason": "Different subsystems: iterable-dataset multi-image finetuning vs Whisper batch_decode skip_special_tokens handling." + }, + { + "left": "issue:43643", + "right": "issue:45375", + "accept": false, + "reason": "Both involve missing config fields, but the titles point to different code paths/models and not a clearly shared underlying bug." + }, + { + "left": "issue:44843", + "right": "issue:44991", + "accept": false, + "reason": "Both are tokenizer-loading regressions, but one is an offline/mistral-regex network call issue and the other is a specific tokenizer load failure for est-roberta." + }, + { + "left": "issue:45372", + "right": "issue:46170", + "accept": false, + "reason": "Unrelated failures: mistral_common import breakage for Gemma 4 vs Windows/ROCm torch.distributed dependency issues." + }, + { + "left": "issue:44857", + "right": "issue:47000", + "accept": false, + "reason": "Different loss implementations and failure modes: float16 CUDA crash vs infeasible cost matrix." + }, + { + "left": "issue:43688", + "right": "issue:45704", + "accept": false, + "reason": "Auxiliary-loss normalization is unrelated to the T5 RMSNorm memory leak." + }, + { + "left": "issue:43688", + "right": "issue:45701", + "accept": false, + "reason": "One is a loss normalization bug; the other is a tokenizer behavior regression across versions." + }, + { + "left": "issue:43295", + "right": "issue:45356", + "accept": false, + "reason": "Both are regressions, but one is processor/tokenizer API breakage and the other is codec/warning behavior for a specific tokenizer." + }, + { + "left": "issue:43525", + "right": "issue:46701", + "accept": false, + "reason": "Llama4 pad_token_id config error is unrelated to the DiffusionGemma image-text pipeline failure." + }, + { + "left": "issue:45542", + "right": "issue:46178", + "accept": false, + "reason": "Both touch backend detection, but they are different backends and different installation/import problems." + }, + { + "left": "issue:43646", + "right": "issue:43749", + "accept": false, + "reason": "Custom model initialization breakage is not the same as FSDP CPU RAM-efficient loading being broken." + }, + { + "left": "issue:33666", + "right": "issue:36010", + "accept": false, + "reason": "Qwen2-VL multi-GPU training and GenerationMixin import failure are unrelated." + }, + { + "left": "issue:43653", + "right": "issue:43927", + "accept": false, + "reason": "Tokenizer special-token registration bug vs config token IDs being lost after save/load are different issues." + }, + { + "left": "issue:43653", + "right": "issue:44568", + "accept": false, + "reason": "Both concern special tokens, but they affect different tokenizers and different symptoms." + }, + { + "left": "issue:43756", + "right": "issue:46402", + "accept": false, + "reason": "Smollm3 RoPE-layer mismatch is unrelated to Sam3 text encoder weights not loading." + }, + { + "left": "issue:43761", + "right": "issue:47013", + "accept": false, + "reason": "CLIP hidden_states regression and xLSTM tuple-unpacking crash are different model-specific forward-path bugs." + }, + { + "left": "issue:44464", + "right": "issue:44743", + "accept": false, + "reason": "Both involve generation/state behavior, but they are distinct models and different root symptoms." + }, + { + "left": "issue:44479", + "right": "issue:44625", + "accept": false, + "reason": "Video-input regression across Qwen VL variants is unrelated to num_labels propagation in Qwen3.5 config." + }, + { + "left": "issue:46832", + "right": "issue:46932", + "accept": false, + "reason": "RT-DETR feature-level mismatch and Gemma4 quantized dtype casting are unrelated." + }, + { + "left": "issue:43550", + "right": "issue:44361", + "accept": false, + "reason": "Bamba torch.compile/SDPA failure is unrelated to MLukeTokenizer AttributeError." + }, + { + "left": "issue:43761", + "right": "issue:46829", + "accept": false, + "reason": "Both mention hidden_states, but they are different models and different output-contract bugs." + }, + { + "left": "issue:43452", + "right": "issue:43577", + "accept": false, + "reason": "Tokenizer/model loading with gguf_file is unrelated to BLIP-2 dtype propagation." + }, + { + "left": "issue:42898", + "right": "issue:46491", + "accept": false, + "reason": "Both are tokenization regressions, but one is a global cleanup-space behavior change and the other is a CodeLlama round-trip spacing bug." + }, + { + "left": "issue:41720", + "right": "issue:44464", + "accept": false, + "reason": "Qwen3 auto device mapping cuda assert and chunked-generation inconsistency are different failures." + }, + { + "left": "issue:33666", + "right": "issue:36331", + "accept": false, + "reason": "Multi-GPU Qwen2-VL training and CustomTrainer.compute_loss signature mismatch are unrelated." + }, + { + "left": "issue:32090", + "right": "issue:33666", + "accept": false, + "reason": "Trainer _gpu_broadcast_one NoneType error is not the same as Qwen2-VL multi-GPU training." + } + ] + }, + "evaluator_result": null, + "evaluator_used": false, + "retried": false, + "accepted_nontrivial_soft_edge": false, + "error_kind": null, + "error_message": null + }, + { + "label": "issue", + "component_index": 1, + "component_count": 28, + "review_unit_index": 28, + "review_unit_count": 65, + "status": "reviewed", + "reason": null, + "source": "llm", + "cache_hit": false, + "model": "gpt-5.4-mini?service_tier=flex", + "cluster_id": "cluster-10000-530", + "nodes": [ + "issue:28218", + "issue:33357", + "issue:39692", + "issue:43278", + "issue:43295", + "issue:43334", + "issue:43493", + "issue:43504", + "issue:43525", + "issue:43550", + "issue:43572", + "issue:43577", + "issue:43606", + "issue:43611", + "issue:43643", + "issue:43688", + "issue:43720", + "issue:43782", + "issue:43825", + "issue:43927", + "issue:43994", + "issue:44361", + "issue:44464", + "issue:44466", + "issue:44484", + "issue:44514", + "issue:44554", + "issue:44561", + "issue:44610", + "issue:44625", + "issue:44704", + "issue:44734", + "issue:44964", + "issue:44977", + "issue:45290", + "issue:45405", + "issue:45440", + "issue:45759", + "issue:45823", + "issue:46050", + "issue:46192", + "issue:46209", + "issue:46612", + "issue:46693", + "issue:46846", + "issue:47254", + "issue:47276" + ], + "soft_pairs": [ + "issue:44466|issue:46050", + "issue:44977|issue:46693", + "issue:44464|issue:44977", + "issue:43504|issue:43550", + "issue:43278|issue:43688", + "issue:45405|issue:47254", + "issue:43688|issue:45440", + "issue:43782|issue:46693", + "issue:43504|issue:43720", + "issue:43720|issue:43927", + "issue:43643|issue:44561", + "issue:43606|issue:43927", + "issue:44734|issue:46612", + "issue:44625|issue:47276", + "issue:39692|issue:43493", + "issue:43504|issue:44610", + "issue:43720|issue:44361", + "issue:44484|issue:44704", + "issue:43577|issue:44554", + "issue:43525|issue:43572", + "issue:28218|issue:33357", + "issue:43611|issue:46846", + "issue:44964|issue:45405", + "issue:44514|issue:45290", + "issue:45823|issue:46050", + "issue:44977|issue:46209", + "issue:39692|issue:43295", + "issue:43493|issue:43825", + "issue:43334|issue:45759", + "issue:43994|issue:46192" + ], + "prepared_review_unit_hash": "5378ca3f86722ffd60e71afb3cc1763c9f508a778185f69b15d3993053bd7064", + "component_budget": { + "node_count": 530, + "item_count": 530, + "soft_pair_count": 1954, + "serialized_chars": 1013004, + "estimated_input_tokens": 253251, + "estimated_eval_tokens": 506758 + }, + "budget": { + "node_count": 47, + "item_count": 47, + "soft_pair_count": 30, + "serialized_chars": 36292, + "estimated_input_tokens": 9073, + "estimated_eval_tokens": 18402 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": true, + "analyst_result": { + "summary": "This cluster is mostly a grab bag of unrelated Transformers bugs with only broad overlap around config/loading, processor, and generation regressions. None of the soft pairs look like the same concrete underlying issue with enough confidence to merge.", + "confidence": 0.84, + "canonical_issue_reason": "issue:43525 is the clearest anchor: it is a concrete, reproducible config-attribute regression (`pad_token_id` missing) and is the closest match to the small subset of config-field failures in the cluster. The cluster overall is still too heterogeneous to treat this as a true duplicate hub.", + "canonical_pr_reason": null, + "best_issue_reason": "issue:43525 is the best single representative because it has a precise failure mode, a clear stack-level symptom, and is actionable. It is not a perfect canonical for the whole set, but it is the strongest issue-level anchor here.", + "best_pr_reason": null, + "soft_edge_verdicts": [ + { + "left": "issue:44466", + "right": "issue:46050", + "accept": false, + "reason": "Both involve save/load behavior, but one is tied-weight serialization varying by device and the other is sinq quantization persistence. Different bugs and code paths." + }, + { + "left": "issue:44977", + "right": "issue:46693", + "accept": false, + "reason": "Shared mention of flash-attn, but one is Qwen3.5 generation correctness and the other is a long-context performance regression after 5.2.0." + }, + { + "left": "issue:44464", + "right": "issue:44977", + "accept": false, + "reason": "Chunked generation with compiled forward is a different failure mode from Qwen3.5 flash-attention generation issues." + }, + { + "left": "issue:43504", + "right": "issue:43550", + "accept": false, + "reason": "Beit legacy-field loading and Bamba torch.compile/SDPA are unrelated model-specific problems." + }, + { + "left": "issue:43278", + "right": "issue:43688", + "accept": false, + "reason": "BF16->FP32 dtype drift in eval is not the same as auxiliary-loss normalization." + }, + { + "left": "issue:45405", + "right": "issue:47254", + "accept": false, + "reason": "PEFT version pinning and a PEFT+ZeRO-3 checkpoint error are different issues despite both involving PEFT." + }, + { + "left": "issue:43688", + "right": "issue:45440", + "accept": false, + "reason": "Auxiliary-loss normalization and DeepseekV3 implementation divergence are distinct model-behavior bugs." + }, + { + "left": "issue:43782", + "right": "issue:46693", + "accept": false, + "reason": "Qwen3VL weight_only load failure is unrelated to a flash-attn long-context performance regression." + }, + { + "left": "issue:43504", + "right": "issue:43720", + "accept": false, + "reason": "Legacy preset-field loading and BitNet packed-weight unpacking during accelerate load are separate loading bugs." + }, + { + "left": "issue:43720", + "right": "issue:43927", + "accept": false, + "reason": "Packed-weight loading in BitNet and DiaConfig save/load token-ID loss are different serialization paths." + }, + { + "left": "issue:43643", + "right": "issue:44561", + "accept": false, + "reason": "Both touch trust_remote_code, but one is missing fields on returned configs and the other is removal of `is_torch_fx_available` breaking remote-code models." + }, + { + "left": "issue:43606", + "right": "issue:43927", + "accept": false, + "reason": "CPU offload device mismatch is not the same as config token-ID persistence after save/load." + }, + { + "left": "issue:44734", + "right": "issue:46612", + "accept": false, + "reason": "KV-cache continuation tensor indexing in serve is different from beam-search cache reorder being skipped for specific model families." + }, + { + "left": "issue:44625", + "right": "issue:47276", + "accept": false, + "reason": "Missing `num_labels` propagation and a misleading `synced_gpus` warning are unrelated." + }, + { + "left": "issue:39692", + "right": "issue:43493", + "accept": false, + "reason": "Both are SigLIP2-related, but one is a doc/example mismatch and quantization failure while the other is an HF-vs-JAX implementation discrepancy." + }, + { + "left": "issue:43504", + "right": "issue:44610", + "accept": false, + "reason": "BEiT legacy-field loading and OmDet processor/model size mismatch are different model-specific bugs." + }, + { + "left": "issue:43720", + "right": "issue:44361", + "accept": false, + "reason": "BitNet packed-weight loading and MLukeTokenizer task AttributeError are unrelated." + }, + { + "left": "issue:44484", + "right": "issue:44704", + "accept": false, + "reason": "A question about default shard size is not the same as AutoProcessor failing to forward kwargs to cached_file." + }, + { + "left": "issue:43577", + "right": "issue:44554", + "accept": false, + "reason": "BLIP2 dtype propagation and MPS attention correctness are unrelated." + }, + { + "left": "issue:43525", + "right": "issue:43572", + "accept": false, + "reason": "Both are missing-pad-token config reports after v5, but they affect different model configs and different attribute names; not enough evidence of one shared defect." + }, + { + "left": "issue:28218", + "right": "issue:33357", + "accept": false, + "reason": "Tokenizer spacing after added tokens has nothing to do with a MacOS bus error on a CLIP model." + }, + { + "left": "issue:43611", + "right": "issue:46846", + "accept": false, + "reason": "Base-model-prefix loading regression and missing `linear_attn` entries in the TP plan are different failures." + }, + { + "left": "issue:44964", + "right": "issue:45405", + "accept": false, + "reason": "Phi-4 multimodal loading failure is unrelated to a PEFT minimum-version bump." + }, + { + "left": "issue:44514", + "right": "issue:45290", + "accept": false, + "reason": "Both hit `apply_chat_template`, but one is batched input with `padding=False` and the other is assistant tool-call messages with no content; different edge cases." + }, + { + "left": "issue:45823", + "right": "issue:46050", + "accept": false, + "reason": "Gemma4 device-mapping mismatch and sinq save/load quantization breakage are separate issues." + }, + { + "left": "issue:44977", + "right": "issue:46209", + "accept": false, + "reason": "Qwen3.5 flash-attention generation problems and FP8 inference failure on B200 are different hardware/runtime bugs." + }, + { + "left": "issue:39692", + "right": "issue:43295", + "accept": false, + "reason": "SigLIP2 documentation/example errors do not match the processor.tokenizer regression for custom model code." + }, + { + "left": "issue:43493", + "right": "issue:43825", + "accept": false, + "reason": "A SigLIP2 implementation discrepancy is unrelated to a pipeline warning message about translation tasks." + }, + { + "left": "issue:43334", + "right": "issue:45759", + "accept": false, + "reason": "Both are config-related, but one is a missing `pad_token_id` AttributeError and the other is failure to unwrap `text_config` for composite multimodal configs; not the same bug." + }, + { + "left": "issue:43994", + "right": "issue:46192", + "accept": false, + "reason": "Nonsensical SigLIP2 outputs and `return_tensors` being ignored with explicit `text_kwargs` are different processor issues." + } + ] + }, + "evaluator_result": null, + "evaluator_used": false, + "retried": false, + "accepted_nontrivial_soft_edge": false, + "error_kind": null, + "error_message": null + }, + { + "label": "issue", + "component_index": 1, + "component_count": 28, + "review_unit_index": 29, + "review_unit_count": 65, + "status": "reviewed", + "reason": null, + "source": "llm", + "cache_hit": false, + "model": "gpt-5.4-mini?service_tier=flex", + "cluster_id": "cluster-10000-530", + "nodes": [ + "issue:15354", + "issue:38175", + "issue:42490", + "issue:42796", + "issue:43278", + "issue:43334", + "issue:43493", + "issue:43504", + "issue:43572", + "issue:43606", + "issue:43646", + "issue:43653", + "issue:43696", + "issue:43701", + "issue:43824", + "issue:43844", + "issue:43873", + "issue:43927", + "issue:43994", + "issue:44112", + "issue:44291", + "issue:44361", + "issue:44464", + "issue:44466", + "issue:44488", + "issue:44493", + "issue:44554", + "issue:44568", + "issue:44589", + "issue:44843", + "issue:44977", + "issue:45072", + "issue:45245", + "issue:45405", + "issue:45491", + "issue:45520", + "issue:45584", + "issue:45812", + "issue:46002", + "issue:46050", + "issue:46209", + "issue:46285", + "issue:46365", + "issue:46489", + "issue:46710", + "issue:46821", + "issue:46832", + "issue:47332" + ], + "soft_pairs": [ + "issue:44488|issue:44843", + "issue:45405|issue:47332", + "issue:43334|issue:43572", + "issue:42796|issue:44977", + "issue:44554|issue:46365", + "issue:45491|issue:46002", + "issue:43646|issue:44466", + "issue:43278|issue:43701", + "issue:44568|issue:46489", + "issue:43927|issue:44361", + "issue:44568|issue:45812", + "issue:43572|issue:44291", + "issue:43493|issue:43844", + "issue:43278|issue:43844", + "issue:46209|issue:46710", + "issue:44589|issue:45520", + "issue:43504|issue:43606", + "issue:43824|issue:44589", + "issue:46821|issue:46832", + "issue:44493|issue:46285", + "issue:43504|issue:43653", + "issue:43994|issue:44493", + "issue:43606|issue:44112", + "issue:15354|issue:38175", + "issue:42490|issue:43696", + "issue:45245|issue:46710", + "issue:44464|issue:45584", + "issue:43873|issue:46050", + "issue:45072|issue:46832" + ], + "prepared_review_unit_hash": "6f7936fc6cba6beea5400b130f1d74b8d60fcd9099e2aefafbbdb1b5a4382555", + "component_budget": { + "node_count": 530, + "item_count": 530, + "soft_pair_count": 1954, + "serialized_chars": 1013004, + "estimated_input_tokens": 253251, + "estimated_eval_tokens": 506758 + }, + "budget": { + "node_count": 48, + "item_count": 48, + "soft_pair_count": 29, + "serialized_chars": 36633, + "estimated_input_tokens": 9159, + "estimated_eval_tokens": 18574 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": true, + "analyst_result": { + "summary": "The set is mostly a collection of unrelated model/tokenizer/config regression reports. I found no soft-edge pair that clearly describes the same underlying bug or change, so I would keep them separated. The strongest standalone issue is the Granite AutoTokenizer regression because it is concrete, broad, and clearly framed as a v4\u2192v5 behavior break.", + "confidence": 0.74, + "canonical_issue_reason": "issue:45812 is the most canonical-style report: it describes a broad but concrete tokenizer regression, names the affected model family, and frames the break as a silent v4\u2192v5 change with clear user impact.", + "canonical_pr_reason": null, + "best_issue_reason": "issue:45812 is the best global anchor because it is specific enough to triage and broad enough to catch related Granite tokenizer regressions without being too vague.", + "best_pr_reason": null, + "soft_edge_verdicts": [ + { + "left": "issue:44488", + "right": "issue:44843", + "accept": false, + "reason": "Both are model-loading failures, but one is a specific local-model loading complaint and the other is an offline-regex/model_info bug in AutoTokenizer; not the same underlying defect." + }, + { + "left": "issue:45405", + "right": "issue:47332", + "accept": false, + "reason": "Unrelated problems: a released-version dependency constraint versus an offloading save failure for a specific Qwen3-VL model." + }, + { + "left": "issue:43334", + "right": "issue:43572", + "accept": false, + "reason": "Different config fields, different models, different failure modes (`pad_token_id` vs `pad_token_idx`)." + }, + { + "left": "issue:42796", + "right": "issue:44977", + "accept": false, + "reason": "Tokenizer BPE-marker handling and flash-attention generation are distinct bugs in different areas." + }, + { + "left": "issue:44554", + "right": "issue:46365", + "accept": false, + "reason": "MPS attention correctness and Trainer import resolution are unrelated." + }, + { + "left": "issue:45491", + "right": "issue:46002", + "accept": false, + "reason": "Both mention NaNs, but they occur in different components and contexts; not enough to treat as the same bug." + }, + { + "left": "issue:43646", + "right": "issue:44466", + "accept": false, + "reason": "Custom model initialization breakage is not the same as device-dependent lm_head serialization inconsistency." + }, + { + "left": "issue:43278", + "right": "issue:43701", + "accept": false, + "reason": "Training/eval dtype mismatch and resume checkpoint key mismatch are unrelated regressions." + }, + { + "left": "issue:44568", + "right": "issue:46489", + "accept": false, + "reason": "Both are tokenizer-related, but they target different model families and different tokenization failures." + }, + { + "left": "issue:43927", + "right": "issue:44361", + "accept": false, + "reason": "Save/load loss of custom token IDs is not the same as an AttributeError in MLukeTokenizer task handling." + }, + { + "left": "issue:44568", + "right": "issue:45812", + "accept": false, + "reason": "Both concern tokenization, but one is missing BOS/EOS insertion for a specific tokenizer while the other is a broad Granite token-ID regression; different code paths and symptoms." + }, + { + "left": "issue:43572", + "right": "issue:44291", + "accept": false, + "reason": "A missing config attribute after upgrade is not the same as an `init_empty_weights` TypeError from an unexpected constructor argument." + }, + { + "left": "issue:43493", + "right": "issue:43844", + "accept": false, + "reason": "A SigLIP2 implementation discrepancy and a DeepSpeed ZeRO-3 gradient issue are unrelated." + }, + { + "left": "issue:43278", + "right": "issue:43844", + "accept": false, + "reason": "These are both training-related, but one is a dtype change in embeddings and the other is a ZeRO-3 gradient anomaly; no shared underlying bug." + }, + { + "left": "issue:46209", + "right": "issue:46710", + "accept": false, + "reason": "FP8 inference failure on B200 and an accuracy regression after upgrade are distinct issues." + }, + { + "left": "issue:44589", + "right": "issue:45520", + "accept": false, + "reason": "Different platform/library failures: Float8 storage lookup vs `flash_attn` import handling on Python 3.13." + }, + { + "left": "issue:43504", + "right": "issue:43606", + "accept": false, + "reason": "A legacy config field loading bug is unrelated to a CPU-offload device mismatch." + }, + { + "left": "issue:43824", + "right": "issue:44589", + "accept": false, + "reason": "Importing a missing Qwen class and an FP8 storage-object error are not the same bug." + }, + { + "left": "issue:46821", + "right": "issue:46832", + "accept": false, + "reason": "Two separate model-forward shape/argument issues in different architectures; not mergeable as one fix." + }, + { + "left": "issue:44493", + "right": "issue:46285", + "accept": false, + "reason": "Both are regression-style reports, but one is a generic unexpected-key issue and the other is a CLIP structure change; different concrete problems." + }, + { + "left": "issue:43504", + "right": "issue:43653", + "accept": false, + "reason": "Legacy-field loading and BigBirdTokenizer special-token registration are unrelated." + }, + { + "left": "issue:43994", + "right": "issue:44493", + "accept": false, + "reason": "SigLIP2 nonsensical outputs and unexpected position-id keys are different symptoms and code paths." + }, + { + "left": "issue:43606", + "right": "issue:44112", + "accept": false, + "reason": "A real CPU-offload mismatch and a stale CI test are not the same underlying change." + }, + { + "left": "issue:15354", + "right": "issue:38175", + "accept": false, + "reason": "TorchScript GeneratorExp export failure and zero-probabilities on SigLIP2 are unrelated." + }, + { + "left": "issue:42490", + "right": "issue:43696", + "accept": false, + "reason": "A trailing-slash local path warning/import issue is unrelated to GPU out-of-memory." + }, + { + "left": "issue:45245", + "right": "issue:46710", + "accept": false, + "reason": "Category-count limits and post-upgrade accuracy regression are different failure classes." + }, + { + "left": "issue:44464", + "right": "issue:45584", + "accept": false, + "reason": "Chunked generation with compiled forward and Whisper empty-transcription failure after token alignment are distinct generation bugs." + }, + { + "left": "issue:43873", + "right": "issue:46050", + "accept": false, + "reason": "Quantized offloading behavior and broken sinq save/load are related to memory/quantization broadly, but not the same concrete defect." + }, + { + "left": "issue:45072", + "right": "issue:46832", + "accept": false, + "reason": "dtype mismatch in bfloat16 inference and RTDetr feature-level shape handling are unrelated." + } + ] + }, + "evaluator_result": null, + "evaluator_used": false, + "retried": false, + "accepted_nontrivial_soft_edge": false, + "error_kind": null, + "error_message": null + }, + { + "label": "issue", + "component_index": 1, + "component_count": 28, + "review_unit_index": 30, + "review_unit_count": 65, + "status": "reviewed", + "reason": null, + "source": "llm", + "cache_hit": false, + "model": "gpt-5.4-mini?service_tier=flex", + "cluster_id": "cluster-10000-530", + "nodes": [ + "issue:28218", + "issue:29159", + "issue:29942", + "issue:30333", + "issue:32090", + "issue:34567", + "issue:37428", + "issue:41720", + "issue:42947", + "issue:43065", + "issue:43295", + "issue:43381", + "issue:43404", + "issue:43450", + "issue:43493", + "issue:43701", + "issue:43844", + "issue:43937", + "issue:43957", + "issue:44423", + "issue:44554", + "issue:44734", + "issue:44849", + "issue:44869", + "issue:44964", + "issue:44977", + "issue:45072", + "issue:45084", + "issue:45198", + "issue:45200", + "issue:45229", + "issue:45357", + "issue:45446", + "issue:45520", + "issue:45588", + "issue:45759", + "issue:45823", + "issue:46050", + "issue:46209", + "issue:46369", + "issue:46399", + "issue:46701", + "issue:46710", + "issue:46752", + "issue:46846", + "issue:47013", + "issue:47145", + "issue:47254" + ], + "soft_pairs": [ + "issue:42947|issue:43381", + "issue:44849|issue:46209", + "issue:44849|issue:45357", + "issue:28218|issue:34567", + "issue:29159|issue:32090", + "issue:43844|issue:44964", + "issue:45084|issue:45520", + "issue:47013|issue:47254", + "issue:45229|issue:45823", + "issue:46399|issue:46846", + "issue:45200|issue:45229", + "issue:45588|issue:46050", + "issue:44423|issue:44734", + "issue:43493|issue:43701", + "issue:44849|issue:47145", + "issue:37428|issue:43957", + "issue:41720|issue:46710", + "issue:43937|issue:46701", + "issue:43295|issue:43450", + "issue:44554|issue:45446", + "issue:43065|issue:46369", + "issue:44977|issue:46701", + "issue:44869|issue:46752", + "issue:29942|issue:30333", + "issue:45072|issue:45198", + "issue:43404|issue:46710", + "issue:45588|issue:45759" + ], + "prepared_review_unit_hash": "b990242e4be5d6d916a45082a963d40ccd7180bc1827a476a097953667bb54d0", + "component_budget": { + "node_count": 530, + "item_count": 530, + "soft_pair_count": 1954, + "serialized_chars": 1013004, + "estimated_input_tokens": 253251, + "estimated_eval_tokens": 506758 + }, + "budget": { + "node_count": 48, + "item_count": 48, + "soft_pair_count": 27, + "serialized_chars": 35875, + "estimated_input_tokens": 8969, + "estimated_eval_tokens": 18194 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": true, + "analyst_result": { + "summary": "This cluster is mostly a grab-bag of unrelated issues. None of the proposed soft edges look like true duplicates; the titles overlap only at a broad subsystem level (tokenizers, training, multimodal, serving, etc.) without matching the same concrete failure mode.", + "confidence": 0.93, + "canonical_issue_reason": "If one issue must represent the cluster, issue 45229 is the strongest standalone report: it has a clear, reproducible failure (Gemma4 31B-IT multi-GPU CUDA OOM), substantial discussion, and a concrete user-facing impact. That said, the items are not actually a duplicate set.", + "canonical_pr_reason": null, + "best_issue_reason": "Issue 45229 is the best standalone issue in this set because it is specific, severe, and actionable, with the clearest described failure path and the most discussion. It is a better representative than the many narrower or more ambiguous reports here.", + "best_pr_reason": null, + "soft_edge_verdicts": [ + { + "left": "issue:42947", + "right": "issue:43381", + "accept": false, + "reason": "Both mention gradient checkpointing, but one is about PEFT LoRA being ineffective and the other about checkpointing being disallowed in eval mode. Different bug conditions and code paths." + }, + { + "left": "issue:44849", + "right": "issue:46209", + "accept": false, + "reason": "Qwen3.5 hidden-state output mismatch vs FP8 inference failure on B200 are unrelated failures affecting different runtime paths." + }, + { + "left": "issue:44849", + "right": "issue:45357", + "accept": false, + "reason": "One is an output_hidden_states bug; the other is a save_pretrained regression with visual encoder keys. Same model family, different bugs." + }, + { + "left": "issue:28218", + "right": "issue:34567", + "accept": false, + "reason": "Tokenizer spacing after added tokens is unrelated to TrainerState.num_input_tokens_seen not updating." + }, + { + "left": "issue:29159", + "right": "issue:32090", + "accept": false, + "reason": "Inconsistent slow vs fast tokenizer behavior is a tokenizer bug; NoneType passed to _gpu_broadcast_one is a training/distributed runtime error." + }, + { + "left": "issue:43844", + "right": "issue:44964", + "accept": false, + "reason": "ZeRO-3 gradient growth with HfDeepSpeedConfig is unrelated to Phi-4 multimodal model loading failures." + }, + { + "left": "issue:45084", + "right": "issue:45520", + "accept": false, + "reason": "Template-node compilation error and flash_attn import KeyError on Python 3.13 are different failure modes in different components." + }, + { + "left": "issue:47013", + "right": "issue:47254", + "accept": false, + "reason": "Both involve training/checkpointing, but xLSTM unpacking a bare tensor and PEFT+DeepSpeed CheckpointError are not the same underlying bug." + }, + { + "left": "issue:45229", + "right": "issue:45823", + "accept": false, + "reason": "Gemma4 multi-GPU OOM and Gemma4 PLE device mismatch with device_map=auto are distinct problems: memory pressure vs tensor placement." + }, + { + "left": "issue:46399", + "right": "issue:46846", + "accept": false, + "reason": "Missing apply_rotary_pos_emb registration decorator and missing linear_attn entries in TP plan are separate Qwen3.5 issues with different failure symptoms." + }, + { + "left": "issue:45200", + "right": "issue:45229", + "accept": false, + "reason": "Defaulting mm_token_type_ids to zeros is a configuration/shape fix; the other issue is a multi-GPU CUDA OOM." + }, + { + "left": "issue:45588", + "right": "issue:46050", + "accept": false, + "reason": "An AttributeError on s_aux=None in flash_attention integration is unrelated to sinq quantized model save/load corruption." + }, + { + "left": "issue:44423", + "right": "issue:44734", + "accept": false, + "reason": "Both are serve-related, but one crashes in continuous batching for multimodal models and the other crashes on KV-cache continuation in /v1/responses." + }, + { + "left": "issue:43493", + "right": "issue:43701", + "accept": false, + "reason": "SigLIP2 implementation discrepancy and resume_from_checkpoint key mismatch are unrelated bugs." + }, + { + "left": "issue:44849", + "right": "issue:47145", + "accept": false, + "reason": "Qwen3.5 hidden-state output bug and FP8 compressed-tensors weight_scale being skipped are different model/runtime issues." + }, + { + "left": "issue:37428", + "right": "issue:43957", + "accept": false, + "reason": "Importing _flash_supports_window_size and meta-device model loading failures are separate loading/runtime problems." + }, + { + "left": "issue:41720", + "right": "issue:46710", + "accept": false, + "reason": "cudaErrorAssert with auto device mapping and an accuracy regression after upgrade are not the same concrete bug." + }, + { + "left": "issue:43937", + "right": "issue:46701", + "accept": false, + "reason": "Invalid GenerationConfig for GLM-5 and DiffusionGemmaGenerationOutput pipeline breakage are unrelated." + }, + { + "left": "issue:43295", + "right": "issue:43450", + "accept": false, + "reason": "Both touch multimodal processor paths, but one is a regression around processor.tokenizer and image passthrough, while the other is batched video output shape handling." + }, + { + "left": "issue:44554", + "right": "issue:45446", + "accept": false, + "reason": "MPS attention correctness with mismatched head dims is unrelated to a PyTorch version check bug in flex_attention imports." + }, + { + "left": "issue:43065", + "right": "issue:46369", + "accept": false, + "reason": "A dummy Conv2d in Sam3PixelDecoder and memory overflow in sam3-litetext are different SAM3 issues with different symptoms." + }, + { + "left": "issue:44977", + "right": "issue:46701", + "accept": false, + "reason": "Qwen3.5 flash-attention generation failure and DiffusionGemma pipeline failure are unrelated." + }, + { + "left": "issue:44869", + "right": "issue:46752", + "accept": false, + "reason": "Whisper timestamp decode crash on trailing replacement character is unrelated to apply_chat_template crashing on an empty list." + }, + { + "left": "issue:29942", + "right": "issue:30333", + "accept": false, + "reason": "Failing Flash Attention 2 tests and MLflow not reporting failed jobs are completely different issues." + }, + { + "left": "issue:45072", + "right": "issue:45198", + "accept": false, + "reason": "bf16 dtype mismatches in SwitchTransformers/TimmWrapperModel are unrelated to Wav2Vec2 save_pretrained/tokenization failures." + }, + { + "left": "issue:45588", + "right": "issue:45759", + "accept": false, + "reason": "flash_attention s_aux=None crash and AutoModelForCausalLM.from_config not unwrapping text_config are different code paths and bug classes." + } + ] + }, + "evaluator_result": null, + "evaluator_used": false, + "retried": false, + "accepted_nontrivial_soft_edge": false, + "error_kind": null, + "error_message": null + }, + { + "label": "issue", + "component_index": 1, + "component_count": 28, + "review_unit_index": 31, + "review_unit_count": 65, + "status": "reviewed", + "reason": null, + "source": "llm", + "cache_hit": false, + "model": "gpt-5.4-mini?service_tier=flex", + "cluster_id": "cluster-10000-530", + "nodes": [ + "issue:28218", + "issue:29870", + "issue:29942", + "issue:35141", + "issue:37428", + "issue:39692", + "issue:42490", + "issue:42915", + "issue:43262", + "issue:43278", + "issue:43295", + "issue:43334", + "issue:43450", + "issue:43493", + "issue:43525", + "issue:43540", + "issue:43550", + "issue:43572", + "issue:43575", + "issue:43653", + "issue:43824", + "issue:43844", + "issue:43856", + "issue:43873", + "issue:43950", + "issue:43957", + "issue:43975", + "issue:44112", + "issue:44186", + "issue:44466", + "issue:44521", + "issue:44610", + "issue:44737", + "issue:45127", + "issue:45200", + "issue:45290", + "issue:45381", + "issue:45406", + "issue:45520", + "issue:45823", + "issue:46121", + "issue:46178", + "issue:46192", + "issue:46209", + "issue:46285", + "issue:46821", + "issue:46894" + ], + "soft_pairs": [ + "issue:43295|issue:43873", + "issue:43572|issue:46821", + "issue:45406|issue:46121", + "issue:43540|issue:45200", + "issue:43295|issue:44610", + "issue:43450|issue:43575", + "issue:43572|issue:43957", + "issue:44186|issue:46192", + "issue:45823|issue:46192", + "issue:37428|issue:43824", + "issue:43550|issue:43844", + "issue:45520|issue:46178", + "issue:45200|issue:45823", + "issue:29870|issue:29942", + "issue:42490|issue:43493", + "issue:43295|issue:43493", + "issue:39692|issue:46192", + "issue:43334|issue:43525", + "issue:42915|issue:43824", + "issue:43278|issue:43856", + "issue:44521|issue:45290", + "issue:43824|issue:43975", + "issue:28218|issue:35141", + "issue:43262|issue:43450", + "issue:45381|issue:46209", + "issue:43975|issue:45520", + "issue:44737|issue:46894", + "issue:44466|issue:45127", + "issue:43653|issue:44112", + "issue:43950|issue:46285" + ], + "prepared_review_unit_hash": "10728492b5b091f50ab8e2246f0a1b2ed22b901c0ead951f192a041142016159", + "component_budget": { + "node_count": 530, + "item_count": 530, + "soft_pair_count": 1954, + "serialized_chars": 1013004, + "estimated_input_tokens": 253251, + "estimated_eval_tokens": 506758 + }, + "budget": { + "node_count": 47, + "item_count": 47, + "soft_pair_count": 30, + "serialized_chars": 36490, + "estimated_input_tokens": 9123, + "estimated_eval_tokens": 18502 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": true, + "analyst_result": { + "summary": "The set is a loose collection of mostly unrelated Transformers issues; none of the soft pairs look like true duplicates or the same fix. If forced to anchor the cluster, the most active and reference-heavy issue is the best representative, but the cluster does not form a real duplicate group.", + "confidence": 0.95, + "canonical_issue_reason": "issue:43873 is the most active open issue and has the most inbound references/discussion, so it is the closest thing to a cluster anchor, even though the surrounding items are not the same bug.", + "canonical_pr_reason": null, + "best_issue_reason": "issue:43873 is the best triage target because it is active, open, and well-referenced; however, it should not be treated as semantically equivalent to the other items.", + "best_pr_reason": null, + "soft_edge_verdicts": [ + { + "left": "issue:43295", + "right": "issue:43873", + "accept": false, + "reason": "Different failures: a processor/tokenizer regression versus an offloading/quantization problem. No shared code path or fix." + }, + { + "left": "issue:43572", + "right": "issue:46821", + "accept": false, + "reason": "Both are AttributeError reports, but they concern different config fields, different models, and different code paths." + }, + { + "left": "issue:45406", + "right": "issue:46121", + "accept": false, + "reason": "Unrelated issues: missing `_tokenizer` on Gemma4Processor vs a TypeError in rope validation for a JSON-loaded list." + }, + { + "left": "issue:43540", + "right": "issue:45200", + "accept": false, + "reason": "Different subsystems and symptoms: Qwen3OmniMoe video processing versus Gemma 4 text-only fine-tuning defaults." + }, + { + "left": "issue:43295", + "right": "issue:44610", + "accept": false, + "reason": "Different bugs in different processors; one is a tokenizer access regression, the other is a wrong image size from the OmDet processor." + }, + { + "left": "issue:43450", + "right": "issue:43575", + "accept": false, + "reason": "Batched video shape handling and tensor-parallel OOM are unrelated problems." + }, + { + "left": "issue:43572", + "right": "issue:43957", + "accept": false, + "reason": "Both mention model loading/configuration, but the failure modes and affected mechanisms are different." + }, + { + "left": "issue:44186", + "right": "issue:46192", + "accept": false, + "reason": "Tokenizer crashes on NER/padding versus `return_tensors` being ignored; not the same bug." + }, + { + "left": "issue:45823", + "right": "issue:46192", + "accept": false, + "reason": "Gemma4 device-placement mismatch is unrelated to `return_tensors` handling." + }, + { + "left": "issue:37428", + "right": "issue:43824", + "accept": false, + "reason": "Both are import errors, but they involve different missing symbols and different causes." + }, + { + "left": "issue:43550", + "right": "issue:43844", + "accept": false, + "reason": "Torch.compile + SDPA failure is unrelated to ZeRO-3 gradient growth in DeepSpeed." + }, + { + "left": "issue:45520", + "right": "issue:46178", + "accept": false, + "reason": "Both touch import/backend lookup code, but the offending backends and root causes differ." + }, + { + "left": "issue:45200", + "right": "issue:45823", + "accept": false, + "reason": "Same model family, but one is a missing default tensor field for fine-tuning and the other is a device-map forward mismatch." + }, + { + "left": "issue:29870", + "right": "issue:29942", + "accept": false, + "reason": "Bart generate mask initialization and Flash Attention 2 CI failures are unrelated." + }, + { + "left": "issue:42490", + "right": "issue:43493", + "accept": false, + "reason": "A trailing-slash import warning and a SigLIP2 implementation discrepancy are not the same issue." + }, + { + "left": "issue:43295", + "right": "issue:43493", + "accept": false, + "reason": "Custom code processor access regression versus SigLIP2 HF/JAX mismatch; no shared fix." + }, + { + "left": "issue:39692", + "right": "issue:46192", + "accept": false, + "reason": "Documentation/example errors are unrelated to the `return_tensors` override bug." + }, + { + "left": "issue:43334", + "right": "issue:43525", + "accept": false, + "reason": "Both are missing `pad_token_id` attribute errors, but they affect different model configs and are not shown to share a single root cause." + }, + { + "left": "issue:42915", + "right": "issue:43824", + "accept": false, + "reason": "FineGrainedFP8Config failure and a missing Qwen2.5-VL import are unrelated." + }, + { + "left": "issue:43278", + "right": "issue:43856", + "accept": false, + "reason": "Embedding dtype drift during evaluate is unrelated to Qwen3 MoE memory inefficiency." + }, + { + "left": "issue:44521", + "right": "issue:45290", + "accept": false, + "reason": "Both concern `apply_chat_template`, but one is assistant mask generation for multimodal inputs and the other is a crash on tool-call messages; distinct bugs." + }, + { + "left": "issue:43824", + "right": "issue:43975", + "accept": false, + "reason": "ImportError versus detokenization regression; different symptoms and code paths." + }, + { + "left": "issue:28218", + "right": "issue:35141", + "accept": false, + "reason": "Tokenizer spacing after added tokens is unrelated to embedding reinitialization after resizing." + }, + { + "left": "issue:43262", + "right": "issue:43450", + "accept": false, + "reason": "Audio chat-template sampling-rate default and batched video shape handling are unrelated." + }, + { + "left": "issue:45381", + "right": "issue:46209", + "accept": false, + "reason": "Qwen2.5-VL video position IDs and FP8 inference on B200 are unrelated issues." + }, + { + "left": "issue:43975", + "right": "issue:45520", + "accept": false, + "reason": "Detokenization bug and flash-attn import mapping error do not share an underlying code path." + }, + { + "left": "issue:44737", + "right": "issue:46894", + "accept": false, + "reason": "XLNet CPU placement bug and CohereASR training-loss shift bug are unrelated." + }, + { + "left": "issue:44466", + "right": "issue:45127", + "accept": false, + "reason": "Both involve tied weights, but one is serialization differing by device and the other is LoRA merge collapse with extended vocabulary." + }, + { + "left": "issue:43653", + "right": "issue:44112", + "accept": false, + "reason": "Tokenizer special-token registration and a GraniteSpeech CI device-override test are unrelated." + }, + { + "left": "issue:43950", + "right": "issue:46285", + "accept": false, + "reason": "Non-persistent buffer corruption during from_pretrained and CLIP text-model structure inconsistency are different regression types." + } + ] + }, + "evaluator_result": null, + "evaluator_used": false, + "retried": false, + "accepted_nontrivial_soft_edge": false, + "error_kind": null, + "error_message": null + }, + { + "label": "issue", + "component_index": 1, + "component_count": 28, + "review_unit_index": 32, + "review_unit_count": 65, + "status": "reviewed", + "reason": null, + "source": "llm", + "cache_hit": false, + "model": "gpt-5.4-mini?service_tier=flex", + "cluster_id": "cluster-10000-530", + "nodes": [ + "issue:28282", + "issue:29942", + "issue:32090", + "issue:33357", + "issue:37428", + "issue:38175", + "issue:41720", + "issue:43335", + "issue:43425", + "issue:43452", + "issue:43504", + "issue:43550", + "issue:43606", + "issue:43611", + "issue:43653", + "issue:43819", + "issue:43827", + "issue:43901", + "issue:44112", + "issue:44206", + "issue:44442", + "issue:44466", + "issue:44479", + "issue:44561", + "issue:44610", + "issue:44792", + "issue:44855", + "issue:44857", + "issue:44898", + "issue:45083", + "issue:45198", + "issue:45200", + "issue:45325", + "issue:45412", + "issue:45563", + "issue:45584", + "issue:45593", + "issue:45759", + "issue:46192", + "issue:46209", + "issue:46234", + "issue:46285", + "issue:46431", + "issue:46512", + "issue:46566", + "issue:46821", + "issue:46899", + "issue:46932" + ], + "soft_pairs": [ + "issue:43452|issue:46512", + "issue:44442|issue:45198", + "issue:44561|issue:46234", + "issue:44479|issue:46192", + "issue:45200|issue:45325", + "issue:43819|issue:44112", + "issue:43335|issue:43550", + "issue:45200|issue:46899", + "issue:44898|issue:45412", + "issue:44610|issue:44857", + "issue:41720|issue:46209", + "issue:45200|issue:46932", + "issue:37428|issue:43425", + "issue:43819|issue:45083", + "issue:43335|issue:43653", + "issue:43606|issue:45083", + "issue:44857|issue:45593", + "issue:44792|issue:45563", + "issue:28282|issue:33357", + "issue:46512|issue:46821", + "issue:43827|issue:43901", + "issue:45759|issue:46285", + "issue:38175|issue:45412", + "issue:29942|issue:32090", + "issue:43335|issue:43504", + "issue:44206|issue:45584", + "issue:43550|issue:44112", + "issue:43611|issue:44466", + "issue:44855|issue:45083", + "issue:46431|issue:46566" + ], + "prepared_review_unit_hash": "e0378b09a3fa377c987c7d4f610b7a7465d4797f5b1bdc2f4595ea7e3eae75aa", + "component_budget": { + "node_count": 530, + "item_count": 530, + "soft_pair_count": 1954, + "serialized_chars": 1013004, + "estimated_input_tokens": 253251, + "estimated_eval_tokens": 506758 + }, + "budget": { + "node_count": 48, + "item_count": 48, + "soft_pair_count": 30, + "serialized_chars": 37022, + "estimated_input_tokens": 9256, + "estimated_eval_tokens": 18768 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": true, + "analyst_result": { + "summary": "This cluster is very heterogeneous: mostly unrelated model-loading, config, device, and CI regressions across different components and model families. The soft pairs share broad keywords or subsystem overlap, but not the same concrete bug/change, so none should be merged as duplicates.", + "confidence": 0.88, + "canonical_issue_reason": "No true duplicate center is obvious; if forced to anchor the cluster, #43452 is the broadest user-facing loading/API regression, but it is still only a loose thematic fit.", + "canonical_pr_reason": null, + "best_issue_reason": "#43452 is the most generally representative of the cluster\u2019s recurring loading/config failure theme, though the overall set is not a tight duplicate cluster.", + "best_pr_reason": null, + "soft_edge_verdicts": [ + { + "left": "issue:43452", + "right": "issue:46512", + "accept": false, + "reason": "Different failure modes: gguf/from_pretrained loading vs stale config dtype after from_config." + }, + { + "left": "issue:44442", + "right": "issue:45198", + "accept": false, + "reason": "Different tokenizer/model paths and symptoms; one is AutoTokenizer loading, the other save_pretrained/tokenization failure." + }, + { + "left": "issue:44561", + "right": "issue:46234", + "accept": false, + "reason": "Both are missing-symbol regressions, but for different utilities and unrelated code paths." + }, + { + "left": "issue:44479", + "right": "issue:46192", + "accept": false, + "reason": "Different input-handling bugs: video regression vs text_kwargs/return_tensors precedence." + }, + { + "left": "issue:45200", + "right": "issue:45325", + "accept": false, + "reason": "Different multimodal processing bugs in different models; not the same underlying issue." + }, + { + "left": "issue:43819", + "right": "issue:44112", + "accept": false, + "reason": "One is an algorithmic mismatch in DAC.from_latents; the other is a CI/device-override test failure." + }, + { + "left": "issue:43335", + "right": "issue:43550", + "accept": false, + "reason": "Unrelated model/config bug vs torch.compile/SDPA runtime failure." + }, + { + "left": "issue:45200", + "right": "issue:46899", + "accept": false, + "reason": "Same model family, but distinct bugs: text-only token type defaults vs quantized audio/vision embedder casting." + }, + { + "left": "issue:44898", + "right": "issue:45412", + "accept": false, + "reason": "Different classes of problems: resolution/interpolation failure vs memory leak on deletion." + }, + { + "left": "issue:44610", + "right": "issue:44857", + "accept": false, + "reason": "Different models and code paths: processor size mismatch vs AMP loss crash." + }, + { + "left": "issue:41720", + "right": "issue:46209", + "accept": false, + "reason": "Both are hardware-related, but one is auto device mapping assert and the other is FP8 inference incompatibility." + }, + { + "left": "issue:45200", + "right": "issue:46932", + "accept": false, + "reason": "#46932 points to a separate quantized vision/audio dtype fix, not the mm_token_type_ids default issue." + }, + { + "left": "issue:37428", + "right": "issue:43425", + "accept": false, + "reason": "Import failure in flash attention utils vs broad Torch 2.10 incompatibility; not the same bug." + }, + { + "left": "issue:43819", + "right": "issue:45083", + "accept": false, + "reason": "Different subsystems entirely: DAC latent conversion vs qwen3_omni_moe feature-length helper." + }, + { + "left": "issue:43335", + "right": "issue:43653", + "accept": false, + "reason": "Different components and symptoms: SwitchTransformers sparse-layer config vs BigBirdTokenizer special-token registration." + }, + { + "left": "issue:43606", + "right": "issue:45083", + "accept": false, + "reason": "CPU offload device mismatch is unrelated to the qwen3_omni_moe helper behavior." + }, + { + "left": "issue:44857", + "right": "issue:45593", + "accept": false, + "reason": "Different detection/training bugs; no shared code path." + }, + { + "left": "issue:44792", + "right": "issue:45563", + "accept": false, + "reason": "One is a failing generation test, the other a stale warning in paged generate." + }, + { + "left": "issue:28282", + "right": "issue:33357", + "accept": false, + "reason": "Completely different problems: missing PyTorch import vs MacOS bus error on a CLIP model." + }, + { + "left": "issue:46512", + "right": "issue:46821", + "accept": false, + "reason": "Different API regressions: dtype handling in from_config vs window_size AttributeError." + }, + { + "left": "issue:43827", + "right": "issue:43901", + "accept": false, + "reason": "Both are docs-related, but they cover different APIs and different issues." + }, + { + "left": "issue:45759", + "right": "issue:46285", + "accept": false, + "reason": "Both involve model structure, but the affected models and breakages are different." + }, + { + "left": "issue:38175", + "right": "issue:45412", + "accept": false, + "reason": "Unrelated symptoms: zero probabilities vs memory not released." + }, + { + "left": "issue:29942", + "right": "issue:32090", + "accept": false, + "reason": "Different failures in different areas: Flash Attention tests vs Trainer GPU broadcast NoneType error." + }, + { + "left": "issue:43335", + "right": "issue:43504", + "accept": false, + "reason": "Distinct bugs in unrelated model/config loading paths." + }, + { + "left": "issue:44206", + "right": "issue:45584", + "accept": false, + "reason": "Different audio-model regressions with no shared underlying code path." + }, + { + "left": "issue:43550", + "right": "issue:44112", + "accept": false, + "reason": "A torch.compile/SDPA runtime issue is not the same as a CI stale-device test." + }, + { + "left": "issue:43611", + "right": "issue:44466", + "accept": false, + "reason": "Both are v5 regressions, but one is model loading with base_model_prefix and the other is serialization of tied weights." + }, + { + "left": "issue:44855", + "right": "issue:45083", + "accept": false, + "reason": "Completely different failures: Python 3.13 import parsing vs qwen3_omni_moe helper behavior." + }, + { + "left": "issue:46431", + "right": "issue:46566", + "accept": false, + "reason": "Processor chat-template availability and device-map meta weights are unrelated bugs." + } + ] + }, + "evaluator_result": null, + "evaluator_used": false, + "retried": false, + "accepted_nontrivial_soft_edge": false, + "error_kind": null, + "error_message": null + }, + { + "label": "issue", + "component_index": 1, + "component_count": 28, + "review_unit_index": 33, + "review_unit_count": 65, + "status": "reviewed", + "reason": null, + "source": "llm", + "cache_hit": false, + "model": "gpt-5.4-mini?service_tier=flex", + "cluster_id": "cluster-10000-530", + "nodes": [ + "issue:10000", + "issue:15354", + "issue:28218", + "issue:29127", + "issue:32090", + "issue:33357", + "issue:34689", + "issue:35141", + "issue:37428", + "issue:38175", + "issue:39692", + "issue:42490", + "issue:43278", + "issue:43335", + "issue:43388", + "issue:43526", + "issue:43550", + "issue:43575", + "issue:43606", + "issue:43701", + "issue:43720", + "issue:43819", + "issue:43825", + "issue:43854", + "issue:44112", + "issue:44206", + "issue:44351", + "issue:44448", + "issue:44493", + "issue:44514", + "issue:44521", + "issue:44617", + "issue:44743", + "issue:44779", + "issue:44857", + "issue:45072", + "issue:45083", + "issue:45290", + "issue:45520", + "issue:45588", + "issue:45823", + "issue:45859", + "issue:46170", + "issue:46326", + "issue:46424", + "issue:46697", + "issue:46858" + ], + "soft_pairs": [ + "issue:43388|issue:43526", + "issue:46424|issue:46858", + "issue:43335|issue:44112", + "issue:43335|issue:43606", + "issue:43335|issue:45083", + "issue:42490|issue:43854", + "issue:37428|issue:44351", + "issue:45290|issue:46326", + "issue:45520|issue:46170", + "issue:44617|issue:46697", + "issue:45588|issue:45859", + "issue:44448|issue:44743", + "issue:43819|issue:44493", + "issue:34689|issue:35141", + "issue:43606|issue:44857", + "issue:39692|issue:43526", + "issue:39692|issue:44493", + "issue:44206|issue:45823", + "issue:10000|issue:15354", + "issue:29127|issue:32090", + "issue:44514|issue:44521", + "issue:28218|issue:32090", + "issue:43720|issue:44112", + "issue:43701|issue:45083", + "issue:44448|issue:44779", + "issue:44857|issue:45072", + "issue:43278|issue:43825", + "issue:32090|issue:33357", + "issue:42490|issue:43575", + "issue:38175|issue:44493", + "issue:43550|issue:45083" + ], + "prepared_review_unit_hash": "e1c48ba412937025c82010ce04d16c736ce5842d73c02ca11ac2a78bb43b5b6a", + "component_budget": { + "node_count": 530, + "item_count": 530, + "soft_pair_count": 1954, + "serialized_chars": 1013004, + "estimated_input_tokens": 253251, + "estimated_eval_tokens": 506758 + }, + "budget": { + "node_count": 47, + "item_count": 47, + "soft_pair_count": 31, + "serialized_chars": 36635, + "estimated_input_tokens": 9159, + "estimated_eval_tokens": 18574 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": true, + "analyst_result": { + "summary": "The candidate pairs are mostly unrelated; none looks like the same underlying bug or change, so all soft edges should be rejected. There is no coherent PR cluster here, and the issue set is too heterogeneous to have a strong canonical beyond a loose representative.", + "confidence": 0.95, + "canonical_issue_reason": "No true canonical issue: the set spans unrelated failures across tokenization, loading, caching, CI, and dtype behavior. If forced to pick an anchor, issue 44521 is the most discussed/open, but it is not a duplicate of the others.", + "canonical_pr_reason": null, + "best_issue_reason": "Issue 44521 is the best representative only in the sense that it has the strongest discussion/inbound-reference signal and a clear reproducible bug, but it should not be treated as a duplicate target for the rest of the set.", + "best_pr_reason": null, + "soft_edge_verdicts": [ + { + "left": "issue:43388", + "right": "issue:43526", + "accept": false, + "reason": "Different label bugs in different components: metric gathering truncation vs BeitImageProcessorFast label reduction." + }, + { + "left": "issue:46424", + "right": "issue:46858", + "accept": false, + "reason": "Both are cache-related, but one is max_cache_length handling for quantized cache and the other is device-tensor seq_length arithmetic in static cache/compile." + }, + { + "left": "issue:43335", + "right": "issue:44112", + "accept": false, + "reason": "Config bug about sparse layer creation vs a GraniteSpeech CI device-override test failure." + }, + { + "left": "issue:43335", + "right": "issue:43606", + "accept": false, + "reason": "Sparse-layer config bug vs a CPU-offload device mismatch in bark-small." + }, + { + "left": "issue:43335", + "right": "issue:45083", + "accept": false, + "reason": "Different models and different failure modes; not the same underlying bug." + }, + { + "left": "issue:42490", + "right": "issue:43854", + "accept": false, + "reason": "Both involve loading, but one is a trailing-slash import warning and the other is a GLM-4.7-Flash unit-test load failure." + }, + { + "left": "issue:37428", + "right": "issue:44351", + "accept": false, + "reason": "Separate import errors for different symbols: _flash_supports_window_size vs HybridCache." + }, + { + "left": "issue:45290", + "right": "issue:46326", + "accept": false, + "reason": "Both touch chat templating, but one crashes on tool-call assistant messages and the other is about continue_final_message for mistral_common tokenizers." + }, + { + "left": "issue:45520", + "right": "issue:46170", + "accept": false, + "reason": "Unrelated platform/import failures: Python 3.13 flash-attn lookup vs Windows ROCm/torch.distributed issues." + }, + { + "left": "issue:44617", + "right": "issue:46697", + "accept": false, + "reason": "CUDA OOM in Sam3Video is unrelated to Swin2SR image over-padding." + }, + { + "left": "issue:45588", + "right": "issue:45859", + "accept": false, + "reason": "Different code paths and failures: flash_attention.py AttributeError on s_aux=None vs Qwen3_5MoeTextRotaryEmbedding CPU-offload incompatibility." + }, + { + "left": "issue:44448", + "right": "issue:44743", + "accept": false, + "reason": "Pegasus v4/v5 output drift is unrelated to Qwen3.5 recurrent state reset with cache." + }, + { + "left": "issue:43819", + "right": "issue:44493", + "accept": false, + "reason": "DAC.from_latents/STE mismatch is unrelated to unexpected position-id key warnings." + }, + { + "left": "issue:34689", + "right": "issue:35141", + "accept": false, + "reason": "Model-loading regression for Llama 3.2 Vision is unrelated to post_init reinitializing untied output embeddings." + }, + { + "left": "issue:43606", + "right": "issue:44857", + "accept": false, + "reason": "Different runtime failures in different paths: CPU offload device mismatch vs float16 AMP crash." + }, + { + "left": "issue:39692", + "right": "issue:43526", + "accept": false, + "reason": "SigLIP2 docs/example problems are not the same as BeitImageProcessorFast returning only one label." + }, + { + "left": "issue:39692", + "right": "issue:44493", + "accept": false, + "reason": "Docs/example mismatch and quantization failure are unrelated to position-id key warnings." + }, + { + "left": "issue:44206", + "right": "issue:45823", + "accept": false, + "reason": "Unsupported feature-extractor argument vs Gemma4 device_map forward mismatch." + }, + { + "left": "issue:10000", + "right": "issue:15354", + "accept": false, + "reason": "Different models and unrelated scripting/export failures." + }, + { + "left": "issue:29127", + "right": "issue:32090", + "accept": false, + "reason": "LayoutLMv3 error-message clarity issue is unrelated to Trainer _gpu_broadcast_one TypeError." + }, + { + "left": "issue:44514", + "right": "issue:44521", + "accept": false, + "reason": "Same area, but different bugs: tool-call/no-content crash vs all-zero assistant masks for multimodal inputs." + }, + { + "left": "issue:28218", + "right": "issue:32090", + "accept": false, + "reason": "Tokenizer extra-space bug is unrelated to Trainer broadcast TypeError." + }, + { + "left": "issue:43720", + "right": "issue:44112", + "accept": false, + "reason": "Packed-weight loading bug is unrelated to the GraniteSpeech CI stale device-override test." + }, + { + "left": "issue:43701", + "right": "issue:45083", + "accept": false, + "reason": "Checkpoint key mismatch is unrelated to the qwen3_omni_moe feature-length helper behavior." + }, + { + "left": "issue:44448", + "right": "issue:44779", + "accept": false, + "reason": "Pegasus generation regression is unrelated to Deepseek tokenizer correctness in v5." + }, + { + "left": "issue:44857", + "right": "issue:45072", + "accept": false, + "reason": "Different dtype/runtime failures in different models: LwDetr AMP crash vs SwitchTransformers/TimmWrapper bfloat16 mismatch." + }, + { + "left": "issue:43278", + "right": "issue:43825", + "accept": false, + "reason": "Training/eval dtype drift is unrelated to the pipeline translation-support error message." + }, + { + "left": "issue:32090", + "right": "issue:33357", + "accept": false, + "reason": "Different platform/backend failures: Trainer broadcast NoneType vs MacOS bus error with CLIP/community model." + }, + { + "left": "issue:42490", + "right": "issue:43575", + "accept": false, + "reason": "Trailing-slash path warning is unrelated to Qwen2-57B-A14B-Instruct tensor-parallel OOM." + }, + { + "left": "issue:38175", + "right": "issue:44493", + "accept": false, + "reason": "SigLIP2 zero-probabilities bug is unrelated to unexpected position-id key warnings." + }, + { + "left": "issue:43550", + "right": "issue:45083", + "accept": false, + "reason": "torch.compile/SDPA failure in Bamba is unrelated to the qwen3_omni_moe feature-length helper issue." + } + ] + }, + "evaluator_result": null, + "evaluator_used": false, + "retried": false, + "accepted_nontrivial_soft_edge": false, + "error_kind": null, + "error_message": null + }, + { + "label": "issue", + "component_index": 1, + "component_count": 28, + "review_unit_index": 34, + "review_unit_count": 65, + "status": "reviewed", + "reason": null, + "source": "llm", + "cache_hit": false, + "model": "gpt-5.4-mini?service_tier=flex", + "cluster_id": "cluster-10000-530", + "nodes": [ + "issue:25251", + "issue:26413", + "issue:27957", + "issue:29870", + "issue:29942", + "issue:31501", + "issue:32090", + "issue:33290", + "issue:34689", + "issue:35141", + "issue:37428", + "issue:40444", + "issue:42175", + "issue:42490", + "issue:43257", + "issue:43278", + "issue:43299", + "issue:43329", + "issue:43381", + "issue:43526", + "issue:43873", + "issue:43927", + "issue:44360", + "issue:44448", + "issue:44464", + "issue:44530", + "issue:44561", + "issue:44779", + "issue:44855", + "issue:44977", + "issue:45127", + "issue:45229", + "issue:45245", + "issue:45290", + "issue:45310", + "issue:45381", + "issue:45459", + "issue:45478", + "issue:45593", + "issue:45698", + "issue:45920", + "issue:46192", + "issue:46234", + "issue:46399", + "issue:46424", + "issue:46655", + "issue:47000", + "issue:47276" + ], + "soft_pairs": [ + "issue:45310|issue:45478", + "issue:26413|issue:31501", + "issue:27957|issue:34689", + "issue:45127|issue:45245", + "issue:45459|issue:46234", + "issue:37428|issue:42175", + "issue:29942|issue:35141", + "issue:45593|issue:47000", + "issue:29870|issue:32090", + "issue:43278|issue:43329", + "issue:44360|issue:44561", + "issue:25251|issue:32090", + "issue:45290|issue:45381", + "issue:29870|issue:33290", + "issue:44448|issue:44977", + "issue:32090|issue:33290", + "issue:44464|issue:46424", + "issue:43278|issue:43381", + "issue:40444|issue:43526", + "issue:43257|issue:43299", + "issue:45290|issue:46655", + "issue:44855|issue:46399", + "issue:44530|issue:47276", + "issue:42490|issue:43873", + "issue:43927|issue:45698", + "issue:44977|issue:45229", + "issue:45920|issue:46192", + "issue:44561|issue:44779" + ], + "prepared_review_unit_hash": "2635452cba784e917c5c5be1b73c007b379f1b73010a8a621c3557c8ef6b9124", + "component_budget": { + "node_count": 530, + "item_count": 530, + "soft_pair_count": 1954, + "serialized_chars": 1013004, + "estimated_input_tokens": 253251, + "estimated_eval_tokens": 506758 + }, + "budget": { + "node_count": 48, + "item_count": 48, + "soft_pair_count": 28, + "serialized_chars": 36220, + "estimated_input_tokens": 9055, + "estimated_eval_tokens": 18366 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": true, + "analyst_result": { + "summary": "Most pairs are not true duplicates; they only share broad subsystem or version-regression themes. The only clear duplicate is the Qwen3.5 MoE from_pretrained error pair (45310/45478).", + "confidence": 0.71, + "canonical_issue_reason": "issue:43873 is the most active and broadly referenced item in the set, so it is the best single anchor even though the overall cluster is thematically diverse.", + "canonical_pr_reason": null, + "best_issue_reason": "issue:43873 has the strongest triage signal here: open, high discussion, and multiple inbound references, making it the most useful representative issue.", + "best_pr_reason": null, + "soft_edge_verdicts": [ + { + "left": "issue:45310", + "right": "issue:45478", + "accept": true, + "reason": "Same title, same Qwen3.5 MoE from_pretrained failure, and both describe the same regression on transformers>=5.4.0." + }, + { + "left": "issue:26413", + "right": "issue:31501", + "accept": false, + "reason": "Both involve training startup failures, but one is a resume_from_checkpoint epoch_iterator edge case and the other is FineWeb SLM training not starting; not the same concrete bug." + }, + { + "left": "issue:27957", + "right": "issue:34689", + "accept": false, + "reason": "Both mention model/attention issues, but they are different models and different failure modes; no clear shared code-path bug." + }, + { + "left": "issue:45127", + "right": "issue:45245", + "accept": false, + "reason": "These are unrelated: LoRA/vocabulary merging versus a category cardinality limit error." + }, + { + "left": "issue:45459", + "right": "issue:46234", + "accept": false, + "reason": "Both touch import/util error handling, but they are different missing symbols and different failure mechanisms." + }, + { + "left": "issue:37428", + "right": "issue:42175", + "accept": false, + "reason": "One is a missing flash-attention helper import, the other is packaging/backend selection for TensorFlow; unrelated." + }, + { + "left": "issue:29942", + "right": "issue:35141", + "accept": false, + "reason": "Flash Attention 2 test failures and output embedding reinitialization are different bugs in different paths." + }, + { + "left": "issue:45593", + "right": "issue:47000", + "accept": false, + "reason": "Both are training-loss problems, but D-FINE auxiliary loss handling and RTDetr infeasible cost matrix are distinct issues." + }, + { + "left": "issue:29870", + "right": "issue:32090", + "accept": false, + "reason": "Bart generate() decoder mask bug and Trainer _gpu_broadcast_one NoneType error are unrelated code paths." + }, + { + "left": "issue:43278", + "right": "issue:43329", + "accept": false, + "reason": "Embedding dtype drift during evaluation is unrelated to the multimodal video token counting bug." + }, + { + "left": "issue:44360", + "right": "issue:44561", + "accept": false, + "reason": "A missing ReLU in DSA indexer and removal of is_torch_fx_available are unrelated changes." + }, + { + "left": "issue:25251", + "right": "issue:32090", + "accept": false, + "reason": "Pipeline top_k nesting and Trainer broadcast NoneType are unrelated." + }, + { + "left": "issue:45290", + "right": "issue:45381", + "accept": false, + "reason": "Both affect multimodal/chat behavior, but one crashes on tool-call messages in chat templates while the other has wrong video vision_position_ids." + }, + { + "left": "issue:29870", + "right": "issue:33290", + "accept": false, + "reason": "UnboundLocalError in BART generation and DeepSpeed/Adafactor OOM are different issues." + }, + { + "left": "issue:44448", + "right": "issue:44977", + "accept": false, + "reason": "Pegasus output regression and Qwen3.5 flash-attention generation failure are different models and code paths." + }, + { + "left": "issue:32090", + "right": "issue:33290", + "accept": false, + "reason": "Trainer broadcast NoneType and Adafactor OOM are unrelated failures." + }, + { + "left": "issue:44464", + "right": "issue:46424", + "accept": false, + "reason": "Chunked generation inconsistency under compiled forward is unrelated to cache length handling in quantized cache." + }, + { + "left": "issue:43278", + "right": "issue:43381", + "accept": false, + "reason": "Embedding dtype change in evaluate and gradient checkpointing in eval mode are separate behaviors." + }, + { + "left": "issue:40444", + "right": "issue:43526", + "accept": false, + "reason": "IterableDataset multi-image fine-tuning failure and BeitImageProcessorFast label reduction are unrelated." + }, + { + "left": "issue:43257", + "right": "issue:43299", + "accept": false, + "reason": "Both involve Qwen3 MoE loading, but one is accelerate+deepspeed conversion and the other is a v5.0.0.dev0 load regression; not the same concrete bug." + }, + { + "left": "issue:45290", + "right": "issue:46655", + "accept": false, + "reason": "Assistant tool-call chat-template crash and SmolVLM video padding error are unrelated multimodal bugs." + }, + { + "left": "issue:44855", + "right": "issue:46399", + "accept": false, + "reason": "Both mention decorators/parsing, but one is a Python 3.13 IndentationError in DebertaV2 and the other is a Qwen3.5 kernel registration failure." + }, + { + "left": "issue:44530", + "right": "issue:47276", + "accept": false, + "reason": "PagedAttentionCache group-type crash and a misleading synced_gpus warning are unrelated." + }, + { + "left": "issue:42490", + "right": "issue:43873", + "accept": false, + "reason": "A local-path import warning and quantization offloading failure are different issues." + }, + { + "left": "issue:43927", + "right": "issue:45698", + "accept": false, + "reason": "Both are save/load related, but one is custom token IDs in config and the other is wrong custom module resolution; different underlying problems." + }, + { + "left": "issue:44977", + "right": "issue:45229", + "accept": false, + "reason": "Flash-attention generation failure for Qwen3.5 and Gemma4 multi-GPU OOM are unrelated." + }, + { + "left": "issue:45920", + "right": "issue:46192", + "accept": false, + "reason": "Tokenizer regression across many models and return_tensors being ignored are different API bugs." + }, + { + "left": "issue:44561", + "right": "issue:44779", + "accept": false, + "reason": "Removal of a utility function and Deepseek tokenizer incorrect results are not the same bug." + } + ] + }, + "evaluator_result": { + "accept": true, + "feedback": "Grounded overall. The canonical issue choice is supported by the packet (43873 is open, has the highest discussion/inbound signal, and is a reasonable anchor). The soft-edge decisions are conservative, and the only accepted duplicate pair (45310/45478) is justified by identical titles and matching regression wording. No obvious overstatement beyond the evidence in the packet." + }, + "evaluator_used": true, + "retried": false, + "accepted_nontrivial_soft_edge": true, + "error_kind": null, + "error_message": null + }, + { + "label": "issue", + "component_index": 1, + "component_count": 28, + "review_unit_index": 35, + "review_unit_count": 65, + "status": "reviewed", + "reason": null, + "source": "llm", + "cache_hit": false, + "model": "gpt-5.4-mini?service_tier=flex", + "cluster_id": "cluster-10000-530", + "nodes": [ + "issue:29127", + "issue:29870", + "issue:29942", + "issue:30064", + "issue:33357", + "issue:34567", + "issue:34689", + "issue:39692", + "issue:41720", + "issue:42490", + "issue:42915", + "issue:42994", + "issue:43278", + "issue:43381", + "issue:43412", + "issue:43575", + "issue:43611", + "issue:43653", + "issue:43701", + "issue:43756", + "issue:43873", + "issue:43976", + "issue:44077", + "issue:44186", + "issue:44206", + "issue:44265", + "issue:44361", + "issue:44442", + "issue:44448", + "issue:44466", + "issue:44479", + "issue:44568", + "issue:44743", + "issue:44871", + "issue:44977", + "issue:45198", + "issue:45229", + "issue:45440", + "issue:45491", + "issue:45563", + "issue:45588", + "issue:45803", + "issue:45823", + "issue:46121", + "issue:46421", + "issue:46858", + "issue:47276" + ], + "soft_pairs": [ + "issue:33357|issue:39692", + "issue:29127|issue:34567", + "issue:43873|issue:45229", + "issue:44871|issue:44977", + "issue:43381|issue:43701", + "issue:44448|issue:44466", + "issue:44186|issue:45198", + "issue:29942|issue:30064", + "issue:43278|issue:43575", + "issue:45588|issue:45823", + "issue:44361|issue:45198", + "issue:46421|issue:46858", + "issue:41720|issue:45229", + "issue:43976|issue:45440", + "issue:29870|issue:34689", + "issue:44206|issue:44479", + "issue:43611|issue:46121", + "issue:45198|issue:45803", + "issue:43653|issue:44442", + "issue:44265|issue:45198", + "issue:44077|issue:47276", + "issue:44977|issue:45563", + "issue:43653|issue:44743", + "issue:43412|issue:45588", + "issue:44568|issue:45491", + "issue:42915|issue:42994", + "issue:43756|issue:45440", + "issue:43575|issue:43701", + "issue:29127|issue:33357", + "issue:42490|issue:43381" + ], + "prepared_review_unit_hash": "d2e620a25bc4738a49152e42b0a724421d15f8424eda1a71008c151d91a20626", + "component_budget": { + "node_count": 530, + "item_count": 530, + "soft_pair_count": 1954, + "serialized_chars": 1013004, + "estimated_input_tokens": 253251, + "estimated_eval_tokens": 506758 + }, + "budget": { + "node_count": 47, + "item_count": 47, + "soft_pair_count": 30, + "serialized_chars": 36237, + "estimated_input_tokens": 9060, + "estimated_eval_tokens": 18376 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": true, + "analyst_result": { + "summary": "All soft pairs are false positives: they mostly share broad subsystem keywords (tokenizers, quantization, cache, generation, vision) but describe different concrete bugs and code paths.", + "confidence": 0.97, + "canonical_issue_reason": null, + "canonical_pr_reason": null, + "best_issue_reason": null, + "best_pr_reason": null, + "soft_edge_verdicts": [ + { + "left": "issue:33357", + "right": "issue:39692", + "accept": false, + "reason": "One is a MacOS bus error loading a CLIP model; the other is a SigLIP2 docs/example issue with model/processor mismatch and quantization." + }, + { + "left": "issue:29127", + "right": "issue:34567", + "accept": false, + "reason": "LayoutLMv3 box-input error messaging is unrelated to TrainerState token-count tracking." + }, + { + "left": "issue:43873", + "right": "issue:45229", + "accept": false, + "reason": "Both mention memory/offloading, but one is a quantization offloading behavior bug and the other is a Gemma4 multi-GPU OOM case." + }, + { + "left": "issue:44871", + "right": "issue:44977", + "accept": false, + "reason": "Gemma eos_token_id mismatch is a config inconsistency, while Qwen3.5 flash-attention generation is a runtime generation bug." + }, + { + "left": "issue:43381", + "right": "issue:43701", + "accept": false, + "reason": "Eval-mode gradient checkpointing and resume_from_checkpoint key mismatch are different training-state issues." + }, + { + "left": "issue:44448", + "right": "issue:44466", + "accept": false, + "reason": "One reports differing outputs across v4/v5; the other is inconsistent lm_head.weight serialization due to tied weights/device state." + }, + { + "left": "issue:44186", + "right": "issue:45198", + "accept": false, + "reason": "LayoutLMv2 tokenizer crashes on NER/padding, while Wav2Vec2 has save_pretrained/tokenization failures." + }, + { + "left": "issue:29942", + "right": "issue:30064", + "accept": false, + "reason": "Flash Attention 2 test failures are unrelated to void segmentation-map processing in image preprocessing." + }, + { + "left": "issue:43278", + "right": "issue:43575", + "accept": false, + "reason": "Embedding dtype drift in evaluation is unrelated to a tensor-parallel OOM when loading Qwen2-57B." + }, + { + "left": "issue:45588", + "right": "issue:45823", + "accept": false, + "reason": "Flash-attention s_aux=None crash is a different code path from Gemma4 device_map auto forward-device mismatch." + }, + { + "left": "issue:44361", + "right": "issue:45198", + "accept": false, + "reason": "MLukeTokenizer AttributeError on tasks is unrelated to Wav2Vec2 save/tokenization failures." + }, + { + "left": "issue:46421", + "right": "issue:46858", + "accept": false, + "reason": "Both are cache-related, but one is lazy-init recompilation behavior and the other is a device-tensor seq_length arithmetic bug." + }, + { + "left": "issue:41720", + "right": "issue:45229", + "accept": false, + "reason": "Qwen3 auto device mapping cudaErrorAssert is not the same as Gemma4 multi-GPU OOM." + }, + { + "left": "issue:43976", + "right": "issue:45440", + "accept": false, + "reason": "Python-version compatibility breakage is unrelated to DeepseekV3MoE diverging from the remote implementation." + }, + { + "left": "issue:29870", + "right": "issue:34689", + "accept": false, + "reason": "BART generate decoder_attention_mask initialization and Llama 3.2 Vision model loading failure are different failures." + }, + { + "left": "issue:44206", + "right": "issue:44479", + "accept": false, + "reason": "Unsupported center arg in a feature extractor is not the same as the Qwen2_5/Qwen3 video-input regression." + }, + { + "left": "issue:43611", + "right": "issue:46121", + "accept": false, + "reason": "base_model_prefix loading regression and rope-parameter dict conversion TypeError are separate loading/config issues." + }, + { + "left": "issue:45198", + "right": "issue:45803", + "accept": false, + "reason": "Wav2Vec2 tokenization/save failures do not match Fuyu's bare-except bug swallowing errors." + }, + { + "left": "issue:43653", + "right": "issue:44442", + "accept": false, + "reason": "BigBirdTokenizer special-token registration and AutoTokenizer loading FastSpeech2ConformerTokenizer are different tokenizer bugs." + }, + { + "left": "issue:44265", + "right": "issue:45198", + "accept": false, + "reason": "torch.export failures with torch_compilable_check are unrelated to Wav2Vec2 save/tokenization failures." + }, + { + "left": "issue:44077", + "right": "issue:47276", + "accept": false, + "reason": "patchtsmixer post_init allowance and a misleading synced_gpus warning are unrelated." + }, + { + "left": "issue:44977", + "right": "issue:45563", + "accept": false, + "reason": "Flash-attention generation failure is not the same as the stale num_return_sequences warning in paged generate()." + }, + { + "left": "issue:43653", + "right": "issue:44743", + "accept": false, + "reason": "BigBirdTokenizer special-token registration and Qwen3.5 recurrent-state reset are unrelated bugs." + }, + { + "left": "issue:43412", + "right": "issue:45588", + "accept": false, + "reason": "Executorch export/training-guard/erfinv issues are unrelated to flash-attention s_aux=None crashes." + }, + { + "left": "issue:44568", + "right": "issue:45491", + "accept": false, + "reason": "mdeberta add_special_tokens BOS/EOS behavior is unrelated to Gemma3 NaN embeddings with mixed-length batching." + }, + { + "left": "issue:42915", + "right": "issue:42994", + "accept": false, + "reason": "Qwen3Moe FP8 config failure and quantized-model saving failure are different quantization-related problems." + }, + { + "left": "issue:43756", + "right": "issue:45440", + "accept": false, + "reason": "Smollm3 RoPE-layer mismatch and DeepseekV3MoE divergence are distinct model implementation issues." + }, + { + "left": "issue:43575", + "right": "issue:43701", + "accept": false, + "reason": "Tensor-parallel OOM on model load is unrelated to resume_from_checkpoint key mismatch." + }, + { + "left": "issue:29127", + "right": "issue:33357", + "accept": false, + "reason": "LayoutLMv3 error clarity and CLIP bus error on MacOS are unrelated model/runtime bugs." + }, + { + "left": "issue:42490", + "right": "issue:43381", + "accept": false, + "reason": "Path-with-trailing-slash import warning and gradient checkpointing in eval mode are different issues." + } + ] + }, + "evaluator_result": null, + "evaluator_used": false, + "retried": false, + "accepted_nontrivial_soft_edge": false, + "error_kind": null, + "error_message": null + }, + { + "label": "issue", + "component_index": 1, + "component_count": 28, + "review_unit_index": 36, + "review_unit_count": 65, + "status": "reviewed", + "reason": null, + "source": "llm", + "cache_hit": false, + "model": "gpt-5.4-mini?service_tier=flex", + "cluster_id": "cluster-10000-530", + "nodes": [ + "issue:27957", + "issue:29127", + "issue:33290", + "issue:34689", + "issue:35141", + "issue:42915", + "issue:43065", + "issue:43066", + "issue:43334", + "issue:43381", + "issue:43421", + "issue:43526", + "issue:43531", + "issue:43575", + "issue:43824", + "issue:43873", + "issue:43931", + "issue:43994", + "issue:44279", + "issue:44387", + "issue:44442", + "issue:44462", + "issue:44514", + "issue:44625", + "issue:44857", + "issue:44898", + "issue:45072", + "issue:45245", + "issue:45250", + "issue:45372", + "issue:45381", + "issue:45406", + "issue:45412", + "issue:45440", + "issue:45459", + "issue:45491", + "issue:45563", + "issue:45584", + "issue:45588", + "issue:45593", + "issue:45712", + "issue:45834", + "issue:46369", + "issue:46640", + "issue:46693", + "issue:46701", + "issue:46821", + "issue:47217" + ], + "soft_pairs": [ + "issue:43065|issue:43994", + "issue:43575|issue:44387", + "issue:43334|issue:43531", + "issue:45491|issue:46693", + "issue:29127|issue:35141", + "issue:43065|issue:45412", + "issue:43994|issue:45412", + "issue:45372|issue:45588", + "issue:43526|issue:47217", + "issue:45406|issue:46701", + "issue:44442|issue:45072", + "issue:44625|issue:45584", + "issue:45440|issue:45593", + "issue:43421|issue:45491", + "issue:45459|issue:45712", + "issue:45412|issue:46369", + "issue:43526|issue:45834", + "issue:45372|issue:45406", + "issue:27957|issue:33290", + "issue:43066|issue:44462", + "issue:33290|issue:34689", + "issue:45250|issue:46640", + "issue:45245|issue:45563", + "issue:45372|issue:45459", + "issue:33290|issue:35141", + "issue:27957|issue:29127", + "issue:43526|issue:46693", + "issue:43381|issue:43873", + "issue:43824|issue:44279", + "issue:43994|issue:46369", + "issue:44857|issue:45588", + "issue:42915|issue:43931", + "issue:45588|issue:46821", + "issue:44514|issue:45381", + "issue:44442|issue:44898" + ], + "prepared_review_unit_hash": "4c0ec7684000661f88a854263d276b1f085272082b3829aca68d721a88159af9", + "component_budget": { + "node_count": 530, + "item_count": 530, + "soft_pair_count": 1954, + "serialized_chars": 1013004, + "estimated_input_tokens": 253251, + "estimated_eval_tokens": 506758 + }, + "budget": { + "node_count": 48, + "item_count": 48, + "soft_pair_count": 35, + "serialized_chars": 37927, + "estimated_input_tokens": 9482, + "estimated_eval_tokens": 19220 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": true, + "analyst_result": { + "summary": "The soft links are mostly false positives: they overlap on broad themes like OOM, Flash Attention, Qwen, or tokenizer handling, but each pair points to a different concrete bug, model, or code path. I would not merge any of these pairs as duplicates.", + "confidence": 0.38, + "canonical_issue_reason": "issue:33290 is the most central item in the noisy memory/OOM subset and has the strongest soft-link connectivity, so it is the safest representative if one must be chosen.", + "canonical_pr_reason": null, + "best_issue_reason": "issue:33290 is the strongest standalone bug report and broadest duplicate target among the issues here, but the linked items still do not look like true duplicates.", + "best_pr_reason": null, + "soft_edge_verdicts": [ + { + "left": "issue:43065", + "right": "issue:43994", + "accept": false, + "reason": "Different failures: dummy Conv2d placeholder in Sam3PixelDecoder vs Siglip2 loading/inference producing nonsense." + }, + { + "left": "issue:43575", + "right": "issue:44387", + "accept": false, + "reason": "Both mention OOM, but one is tensor-parallel model loading and the other is int4 quantization reserved-memory growth." + }, + { + "left": "issue:43334", + "right": "issue:43531", + "accept": false, + "reason": "Qwen3-VL load crash from missing pad_token_id is unrelated to Qwen3-MoE sliding_window behavior." + }, + { + "left": "issue:45491", + "right": "issue:46693", + "accept": false, + "reason": "Both involve FA2/long context, but one is a NaN correctness bug and the other is a performance regression." + }, + { + "left": "issue:29127", + "right": "issue:35141", + "accept": false, + "reason": "LayoutLMV3 error-message clarity is unrelated to token-embedding reinit when resizing embeddings." + }, + { + "left": "issue:43065", + "right": "issue:45412", + "accept": false, + "reason": "Dummy module placeholder and RT-DETR memory leak are different code paths and symptoms." + }, + { + "left": "issue:43994", + "right": "issue:45412", + "accept": false, + "reason": "Siglip2 inference quality issue does not match RT-DETR object lifetime / GC memory release." + }, + { + "left": "issue:45372", + "right": "issue:45588", + "accept": false, + "reason": "Gemma 4 processor import failure is unrelated to flash_attention.py crashing on s_aux=None." + }, + { + "left": "issue:43526", + "right": "issue:47217", + "accept": false, + "reason": "BeitImageProcessorFast label reduction bug is unrelated to VL image-placeholder handling in user text." + }, + { + "left": "issue:45406", + "right": "issue:46701", + "accept": false, + "reason": "Gemma4Processor missing _tokenizer is a processor-loading bug, not the DiffusionGemmaGenerationOutput pipeline failure." + }, + { + "left": "issue:44442", + "right": "issue:45072", + "accept": false, + "reason": "Tokenizer loading failure and dtype mismatch in bfloat16 inference are distinct issues." + }, + { + "left": "issue:44625", + "right": "issue:45584", + "accept": false, + "reason": "num_labels propagation in Qwen3.5 is unrelated to Whisper empty-transcription generation after align_special_tokens." + }, + { + "left": "issue:45440", + "right": "issue:45593", + "accept": false, + "reason": "DeepseekV3MoE implementation divergence and D-FINE auxiliary-loss omission are different model bugs." + }, + { + "left": "issue:43421", + "right": "issue:45491", + "accept": false, + "reason": "Tokenizer post-processor update logic does not match Gemma3 NaN embeddings under mixed-length batching." + }, + { + "left": "issue:45459", + "right": "issue:45712", + "accept": false, + "reason": "Masking tokenizer errors without protobuf and cleanup of dummy PT classes are unrelated maintenance bugs." + }, + { + "left": "issue:45412", + "right": "issue:46369", + "accept": false, + "reason": "RT-DETR GC memory leak is not the same as sam3-litetext memory overflow." + }, + { + "left": "issue:43526", + "right": "issue:45834", + "accept": false, + "reason": "Beit reduce_labels bug and Kosmos2.5 OCR index error affect different preprocessing paths." + }, + { + "left": "issue:45372", + "right": "issue:45406", + "accept": false, + "reason": "Both are Gemma 4-related, but one is a mistral_common import dependency failure and the other is a missing processor attribute." + }, + { + "left": "issue:27957", + "right": "issue:33290", + "accept": false, + "reason": "XLMRoberta Flash Attention 2 support is unrelated to Adafactor+DeepSpeed OOM." + }, + { + "left": "issue:43066", + "right": "issue:44462", + "accept": false, + "reason": "Wrong decoder type in Transformers v5 is a different tokenizer problem than ignoring tokenizer.json from the repo." + }, + { + "left": "issue:33290", + "right": "issue:34689", + "accept": false, + "reason": "OOM in deepspeed/Adafactor does not match model-loading breakage for Llama 3.2 Vision." + }, + { + "left": "issue:45250", + "right": "issue:46640", + "accept": false, + "reason": "These are only generic Flash Attention 2 titles; no concrete shared bug path is evident." + }, + { + "left": "issue:45245", + "right": "issue:45563", + "accept": false, + "reason": "Category-limit runtime error and stale generate() warning are unrelated." + }, + { + "left": "issue:45372", + "right": "issue:45459", + "accept": false, + "reason": "ImportError in Gemma 4 loading and protobuf-related tokenizer error masking are different failure modes." + }, + { + "left": "issue:33290", + "right": "issue:35141", + "accept": false, + "reason": "OOM under Adafactor/DeepSpeed is unrelated to embedding reinitialization after resizing." + }, + { + "left": "issue:27957", + "right": "issue:29127", + "accept": false, + "reason": "Flash Attention 2 with XLMRoberta and LayoutLMV3 error messaging have no shared bug." + }, + { + "left": "issue:43526", + "right": "issue:46693", + "accept": false, + "reason": "Beit label handling and FA2 long-context performance regression are unrelated." + }, + { + "left": "issue:43381", + "right": "issue:43873", + "accept": false, + "reason": "Gradient checkpointing in eval mode and quantization offloading are different issues." + }, + { + "left": "issue:43824", + "right": "issue:44279", + "accept": false, + "reason": "Qwen2_5_VL import failure and a generic dependency issue are too broad to be the same bug." + }, + { + "left": "issue:43994", + "right": "issue:46369", + "accept": false, + "reason": "Siglip2 nonsense outputs and SAM3 lite-text memory overflow are unrelated." + }, + { + "left": "issue:44857", + "right": "issue:45588", + "accept": false, + "reason": "LwDetr float16 AMP crash and flash_attention s_aux=None are different code paths." + }, + { + "left": "issue:42915", + "right": "issue:43931", + "accept": false, + "reason": "Qwen3Moe FineGrainedFP8Config failure and Qwen3-VL shape mismatch are different model-loading problems." + }, + { + "left": "issue:45588", + "right": "issue:46821", + "accept": false, + "reason": "flash_attention.py s_aux=None crash is unrelated to TimesFM window_size AttributeError." + }, + { + "left": "issue:44514", + "right": "issue:45381", + "accept": false, + "reason": "Qwen2.5-VL batched chat-template padding bug is distinct from video vision_position_ids being wrong." + }, + { + "left": "issue:44442", + "right": "issue:44898", + "accept": false, + "reason": "FastSpeech2ConformerTokenizer load failure is unrelated to Perceiver non-default-resolution behavior." + } + ] + }, + "evaluator_result": null, + "evaluator_used": false, + "retried": false, + "accepted_nontrivial_soft_edge": false, + "error_kind": null, + "error_message": null + }, + { + "label": "issue", + "component_index": 1, + "component_count": 28, + "review_unit_index": 37, + "review_unit_count": 65, + "status": "reviewed", + "reason": null, + "source": "llm", + "cache_hit": false, + "model": "gpt-5.4-mini?service_tier=flex", + "cluster_id": "cluster-10000-530", + "nodes": [ + "issue:25251", + "issue:27957", + "issue:28282", + "issue:29870", + "issue:42175", + "issue:42490", + "issue:42617", + "issue:43262", + "issue:43316", + "issue:43493", + "issue:43525", + "issue:43550", + "issue:43575", + "issue:43606", + "issue:43643", + "issue:43646", + "issue:43701", + "issue:43720", + "issue:43742", + "issue:43749", + "issue:43761", + "issue:43827", + "issue:43873", + "issue:43931", + "issue:44062", + "issue:44077", + "issue:44206", + "issue:44442", + "issue:44479", + "issue:44743", + "issue:44811", + "issue:44871", + "issue:44913", + "issue:44964", + "issue:44977", + "issue:44987", + "issue:45245", + "issue:45325", + "issue:45440", + "issue:45759", + "issue:45859", + "issue:45907", + "issue:46121", + "issue:46178", + "issue:46234", + "issue:46399", + "issue:46697", + "issue:47312" + ], + "soft_pairs": [ + "issue:44743|issue:44811", + "issue:43316|issue:47312", + "issue:45325|issue:46697", + "issue:46121|issue:46399", + "issue:43646|issue:44077", + "issue:44977|issue:45245", + "issue:43262|issue:44206", + "issue:43761|issue:45859", + "issue:43720|issue:44442", + "issue:25251|issue:28282", + "issue:27957|issue:29870", + "issue:43873|issue:44871", + "issue:44077|issue:44479", + "issue:42175|issue:46234", + "issue:42490|issue:44743", + "issue:43643|issue:44913", + "issue:43493|issue:43827", + "issue:43316|issue:45759", + "issue:44062|issue:45907", + "issue:43749|issue:43931", + "issue:44987|issue:46178", + "issue:42617|issue:43575", + "issue:43316|issue:43525", + "issue:44964|issue:45440", + "issue:43550|issue:43701", + "issue:43575|issue:43742", + "issue:43606|issue:43701" + ], + "prepared_review_unit_hash": "60b59aafa40aebae84e24f92e55a2e488a1a44babcf37c84eaf5c907ede74726", + "component_budget": { + "node_count": 530, + "item_count": 530, + "soft_pair_count": 1954, + "serialized_chars": 1013004, + "estimated_input_tokens": 253251, + "estimated_eval_tokens": 506758 + }, + "budget": { + "node_count": 48, + "item_count": 48, + "soft_pair_count": 27, + "serialized_chars": 35801, + "estimated_input_tokens": 8951, + "estimated_eval_tokens": 18158 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": true, + "analyst_result": { + "summary": "The soft pairs are mostly superficial overlaps by subsystem or wording, but they describe different bugs, configs, models, or docs changes. I would not merge any of them as duplicates.", + "confidence": 0.98, + "canonical_issue_reason": "No single canonical issue fits this set: the items span unrelated reports. If forced to pick a representative anchor, issue 43873 is the broadest and most active runtime bug report, but it does not subsume the others.", + "canonical_pr_reason": null, + "best_issue_reason": "Issue 43873 is the best representative only in the sense that it is broad, active, and cross-cutting; however, it is not a true duplicate candidate for the rest of the cluster.", + "best_pr_reason": null, + "soft_edge_verdicts": [ + { + "left": "issue:44743", + "right": "issue:44811", + "accept": false, + "reason": "Different bugs: recurrent-state reset in Qwen3.5 vs Whisper batch_decode ignoring skip_special_tokens." + }, + { + "left": "issue:43316", + "right": "issue:47312", + "accept": false, + "reason": "Gemma3TextConfig API mismatch is unrelated to a generic kernels 0.16 compatibility complaint." + }, + { + "left": "issue:45325", + "right": "issue:46697", + "accept": false, + "reason": "Qwen2.5-VL rope-index scaling bug is unrelated to Swin2SR over-padding." + }, + { + "left": "issue:46121", + "right": "issue:46399", + "accept": false, + "reason": "One is rope-validation TypeError; the other is a Qwen3.5 kernel registration crash." + }, + { + "left": "issue:43646", + "right": "issue:44077", + "accept": false, + "reason": "Both touch initialization, but one is a broad v5 regression and the other is a model-specific post_init policy issue." + }, + { + "left": "issue:44977", + "right": "issue:45245", + "accept": false, + "reason": "Flash-attention generation failure and category-cardinality limit are unrelated." + }, + { + "left": "issue:43262", + "right": "issue:44206", + "accept": false, + "reason": "Audio chat-template sampling-rate default and feature-extractor center-arg crash are different processor bugs." + }, + { + "left": "issue:43761", + "right": "issue:45859", + "accept": false, + "reason": "CLIPVision hidden_states regression is unrelated to Qwen3_5Moe CPU offload incompatibility." + }, + { + "left": "issue:43720", + "right": "issue:44442", + "accept": false, + "reason": "BitNet packed-weight loading and FastSpeech2Conformer tokenizer loading are unrelated code paths." + }, + { + "left": "issue:25251", + "right": "issue:28282", + "accept": false, + "reason": "Pipeline output-shape bug is unrelated to missing PyTorch import errors." + }, + { + "left": "issue:27957", + "right": "issue:29870", + "accept": false, + "reason": "Flash Attention 2 model support and BART generate() unbound variable are different failures." + }, + { + "left": "issue:43873", + "right": "issue:44871", + "accept": false, + "reason": "Quantization/offloading behavior does not match an eos_token_id config inconsistency." + }, + { + "left": "issue:44077", + "right": "issue:44479", + "accept": false, + "reason": "Patchtsmixer post_init policy and Qwen video-input regression are unrelated." + }, + { + "left": "issue:42175", + "right": "issue:46234", + "accept": false, + "reason": "Backend packaging mismatch and missing import_utils symbol are different issues." + }, + { + "left": "issue:42490", + "right": "issue:44743", + "accept": false, + "reason": "Local-path import warning is unrelated to recurrent-state reset in Qwen3.5." + }, + { + "left": "issue:43643", + "right": "issue:44913", + "accept": false, + "reason": "Missing fields from trust_remote_code config loading is not the same as rotary_pct not persisting on reload." + }, + { + "left": "issue:43493", + "right": "issue:43827", + "accept": false, + "reason": "SigLIP2 implementation mismatch and docs still referencing pipeline() are unrelated." + }, + { + "left": "issue:43316", + "right": "issue:45759", + "accept": false, + "reason": "Gemma3TextConfig API discrepancy and Qwen multimodal text_config unwrapping are different composite-config bugs." + }, + { + "left": "issue:44062", + "right": "issue:45907", + "accept": false, + "reason": "AddedToken keyword collision is unrelated to list_repo_files missing revision forwarding." + }, + { + "left": "issue:43749", + "right": "issue:43931", + "accept": false, + "reason": "FSDP CPU RAM loading breakage and Qwen3-VL shape mismatch are different loading problems." + }, + { + "left": "issue:44987", + "right": "issue:46178", + "accept": false, + "reason": "Transformers version loading failure and tensorflow_text backend registration error are unrelated." + }, + { + "left": "issue:42617", + "right": "issue:43575", + "accept": false, + "reason": "3d_parallel.py execution failure and Qwen2-57B tensor-parallel OOM are different symptoms and code paths." + }, + { + "left": "issue:43316", + "right": "issue:43525", + "accept": false, + "reason": "Both are config-related, but Gemma3TextConfig API differences and missing pad_token_id on Llama4Config are not the same bug." + }, + { + "left": "issue:44964", + "right": "issue:45440", + "accept": false, + "reason": "Phi-4 multimodal loading failure and DeepSeekV3MoE implementation divergence are separate model-specific issues." + }, + { + "left": "issue:43550", + "right": "issue:43701", + "accept": false, + "reason": "torch.compile/SDPA failure in Bamba is unrelated to resume_from_checkpoint key mismatch." + }, + { + "left": "issue:43575", + "right": "issue:43742", + "accept": false, + "reason": "OOM on large Qwen2 model under tensor parallelism is unrelated to a key error while loading MobileLLM-125M." + }, + { + "left": "issue:43606", + "right": "issue:43701", + "accept": false, + "reason": "CPU offload device mismatch and checkpoint resume key mismatch are unrelated." + } + ] + }, + "evaluator_result": null, + "evaluator_used": false, + "retried": false, + "accepted_nontrivial_soft_edge": false, + "error_kind": null, + "error_message": null + }, + { + "label": "issue", + "component_index": 1, + "component_count": 28, + "review_unit_index": 38, + "review_unit_count": 65, + "status": "reviewed", + "reason": null, + "source": "llm", + "cache_hit": false, + "model": "gpt-5.4-mini?service_tier=flex", + "cluster_id": "cluster-10000-530", + "nodes": [ + "issue:16998", + "issue:22355", + "issue:25251", + "issue:27957", + "issue:32090", + "issue:42490", + "issue:42907", + "issue:43065", + "issue:43232", + "issue:43295", + "issue:43316", + "issue:43388", + "issue:43404", + "issue:43441", + "issue:43452", + "issue:43646", + "issue:43701", + "issue:43716", + "issue:43761", + "issue:43825", + "issue:43931", + "issue:44077", + "issue:44079", + "issue:44188", + "issue:44190", + "issue:44442", + "issue:44464", + "issue:44496", + "issue:44521", + "issue:44534", + "issue:44617", + "issue:44855", + "issue:44871", + "issue:44913", + "issue:45137", + "issue:45216", + "issue:45584", + "issue:45698", + "issue:45803", + "issue:46050", + "issue:46234", + "issue:46291", + "issue:46421", + "issue:46506", + "issue:46829", + "issue:46932", + "issue:47217", + "issue:47276" + ], + "soft_pairs": [ + "issue:44855|issue:45137", + "issue:43701|issue:46506", + "issue:43931|issue:45216", + "issue:44190|issue:44496", + "issue:43761|issue:44079", + "issue:27957|issue:32090", + "issue:44442|issue:45803", + "issue:42490|issue:47276", + "issue:46829|issue:46932", + "issue:43295|issue:44871", + "issue:43316|issue:43716", + "issue:44871|issue:45584", + "issue:43295|issue:43825", + "issue:43232|issue:43388", + "issue:16998|issue:22355", + "issue:43761|issue:44188", + "issue:44521|issue:47217", + "issue:43646|issue:44913", + "issue:46234|issue:46291", + "issue:43452|issue:44496", + "issue:44077|issue:44534", + "issue:42907|issue:45698", + "issue:44871|issue:46050", + "issue:44464|issue:46421", + "issue:43404|issue:43441", + "issue:25251|issue:27957", + "issue:43065|issue:44617" + ], + "prepared_review_unit_hash": "eb16d334873a0477eef07c3c63ac22eb91f2cac13bfd9931ca54964467947e53", + "component_budget": { + "node_count": 530, + "item_count": 530, + "soft_pair_count": 1954, + "serialized_chars": 1013004, + "estimated_input_tokens": 253251, + "estimated_eval_tokens": 506758 + }, + "budget": { + "node_count": 48, + "item_count": 48, + "soft_pair_count": 27, + "serialized_chars": 35766, + "estimated_input_tokens": 8942, + "estimated_eval_tokens": 18140 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": true, + "analyst_result": { + "summary": "The cluster is highly heterogeneous: the soft links are mostly superficial similarity across unrelated Transformers issues, so none of the candidate pairs should be merged as duplicates.", + "confidence": 0.96, + "canonical_issue_reason": "No single canonical issue stands out: the issues cover unrelated bugs across generation, loading, configuration, multimodal templates, and backend/runtime behavior.", + "canonical_pr_reason": null, + "best_issue_reason": "There is no good global issue canonical because the items do not share one underlying bug or code path.", + "best_pr_reason": null, + "soft_edge_verdicts": [ + { + "left": "issue:44855", + "right": "issue:45137", + "accept": false, + "reason": "Different failures: Python 3.13 JIT parsing/import error vs DeepSpeed ZeRO3 deque underflow." + }, + { + "left": "issue:43701", + "right": "issue:46506", + "accept": false, + "reason": "Resume checkpoint key mismatch and CPU/CUDA divergence are separate problems with different symptoms and code paths." + }, + { + "left": "issue:43931", + "right": "issue:45216", + "accept": false, + "reason": "Both are loading/checkpoint issues, but one is Qwen3-VL weight-shape mismatch and the other is a Qwen3.5 save_pretrained regression." + }, + { + "left": "issue:44190", + "right": "issue:44496", + "accept": false, + "reason": "Local dataset loading in an example script is unrelated to unrecognized-model/config loading failure." + }, + { + "left": "issue:43761", + "right": "issue:44079", + "accept": false, + "reason": "CLIPVisionModel hidden_states regression and ModelOutput key assignment are different subsystems and bugs." + }, + { + "left": "issue:27957", + "right": "issue:32090", + "accept": false, + "reason": "FlashAttention-2 support for XLMRoberta is unrelated to the Trainer NoneType broadcast error." + }, + { + "left": "issue:44442", + "right": "issue:45803", + "accept": false, + "reason": "Tokenizer loading failure and a bare-except bug in FuyuBatchFeature are distinct issues." + }, + { + "left": "issue:42490", + "right": "issue:47276", + "accept": false, + "reason": "Trailing-slash local model import warning and misleading synced_gpus warning are different user-facing problems." + }, + { + "left": "issue:46829", + "right": "issue:46932", + "accept": false, + "reason": "VideoPrism hidden_states output bug and Gemma4 quantized-embedder dtype casting are unrelated." + }, + { + "left": "issue:43295", + "right": "issue:44871", + "accept": false, + "reason": "Processor/tokenizer regression for custom model code is not the same as Gemma eos_token_id config inconsistency." + }, + { + "left": "issue:43316", + "right": "issue:43716", + "accept": false, + "reason": "Gemma3TextConfig API mismatch and Mistral3 image-preprocessor dtype mismatch are different model-specific bugs." + }, + { + "left": "issue:44871", + "right": "issue:45584", + "accept": false, + "reason": "Gemma eos_token configuration mismatch is unrelated to Whisper empty-transcription generation failure." + }, + { + "left": "issue:43295", + "right": "issue:43825", + "accept": false, + "reason": "Custom model processor regression and pipeline translation error messaging do not share the same root cause." + }, + { + "left": "issue:43232", + "right": "issue:43388", + "accept": false, + "reason": "Generation kwargs sync logic and gather_for_metrics label truncation are separate code paths." + }, + { + "left": "issue:16998", + "right": "issue:22355", + "accept": false, + "reason": "A model_max_length question for DeBERTa-V3 is unrelated to the missing transformers.onnx module." + }, + { + "left": "issue:43761", + "right": "issue:44188", + "accept": false, + "reason": "Hidden-states regression and attention-kernel divergence under torch.compile are different bugs." + }, + { + "left": "issue:44521", + "right": "issue:47217", + "accept": false, + "reason": "Both involve multimodal text/image handling, but one is all-zero assistant masks and the other is image-placeholder placement in user text." + }, + { + "left": "issue:43646", + "right": "issue:44913", + "accept": false, + "reason": "Custom model initialization breakage in v5 is unrelated to GPTNeoXConfig rotary_pct not persisting on reload." + }, + { + "left": "issue:46234", + "right": "issue:46291", + "accept": false, + "reason": "Missing import_utils helper and kernels dependency/version import failure are different issues." + }, + { + "left": "issue:43452", + "right": "issue:44496", + "accept": false, + "reason": "gguf_file loading breakage and unrecognized-model/model_type config errors both affect loading, but they are different failure modes." + }, + { + "left": "issue:44077", + "right": "issue:44534", + "accept": false, + "reason": "patchtsmixer post_init policy and non-persistent buffer initialization are unrelated." + }, + { + "left": "issue:42907", + "right": "issue:45698", + "accept": false, + "reason": "Dequantized-model save_pretrained failure and wrong custom-module loading after save_pretrained are distinct serialization problems." + }, + { + "left": "issue:44871", + "right": "issue:46050", + "accept": false, + "reason": "Gemma eos_token_id inconsistency is unrelated to sinq quantization save/load breakage." + }, + { + "left": "issue:44464", + "right": "issue:46421", + "accept": false, + "reason": "Chunked generation output inconsistency and chunked-prefill recompilation are separate compile/cache issues." + }, + { + "left": "issue:43404", + "right": "issue:43441", + "accept": false, + "reason": "lm_head weight tying in Mistral3 and FlashAttention failures in Ministral-3 are not the same bug." + }, + { + "left": "issue:43065", + "right": "issue:44617", + "accept": false, + "reason": "Dummy Conv2d in Sam3PixelDecoder and Sam3Video CUDA OOM are unrelated." + } + ] + }, + "evaluator_result": null, + "evaluator_used": false, + "retried": false, + "accepted_nontrivial_soft_edge": false, + "error_kind": null, + "error_message": null + }, + { + "label": "issue", + "component_index": 1, + "component_count": 28, + "review_unit_index": 39, + "review_unit_count": 65, + "status": "reviewed", + "reason": null, + "source": "llm", + "cache_hit": false, + "model": "gpt-5.4-mini?service_tier=flex", + "cluster_id": "cluster-10000-530", + "nodes": [ + "issue:29870", + "issue:35141", + "issue:41950", + "issue:42490", + "issue:43352", + "issue:43475", + "issue:43504", + "issue:43525", + "issue:43531", + "issue:43540", + "issue:43572", + "issue:43650", + "issue:43673", + "issue:43931", + "issue:44060", + "issue:44162", + "issue:44242", + "issue:44479", + "issue:44514", + "issue:44534", + "issue:44610", + "issue:44743", + "issue:44792", + "issue:44855", + "issue:44913", + "issue:44977", + "issue:45127", + "issue:45200", + "issue:45381", + "issue:45405", + "issue:45542", + "issue:45588", + "issue:45600", + "issue:45632", + "issue:45698", + "issue:45803", + "issue:46170", + "issue:46178", + "issue:46202", + "issue:46234", + "issue:46365", + "issue:46369", + "issue:46424", + "issue:46640", + "issue:46693", + "issue:46894", + "issue:47217" + ], + "soft_pairs": [ + "issue:43352|issue:46640", + "issue:44162|issue:44792", + "issue:45803|issue:46365", + "issue:43540|issue:44514", + "issue:43572|issue:46178", + "issue:44610|issue:46369", + "issue:44792|issue:44977", + "issue:45381|issue:47217", + "issue:44479|issue:47217", + "issue:44534|issue:44913", + "issue:45542|issue:46234", + "issue:44792|issue:45600", + "issue:43504|issue:45698", + "issue:43525|issue:43531", + "issue:45632|issue:46424", + "issue:35141|issue:42490", + "issue:45803|issue:46170", + "issue:46693|issue:47217", + "issue:45127|issue:45200", + "issue:44242|issue:46894", + "issue:41950|issue:43475", + "issue:43931|issue:44060", + "issue:43673|issue:44743", + "issue:43650|issue:46202", + "issue:44855|issue:45588", + "issue:44792|issue:45405", + "issue:29870|issue:35141" + ], + "prepared_review_unit_hash": "bed2ef1c628fea7e51f5380e38a7116f1feec090765f0cb4851b94bca89bbebc", + "component_budget": { + "node_count": 530, + "item_count": 530, + "soft_pair_count": 1954, + "serialized_chars": 1013004, + "estimated_input_tokens": 253251, + "estimated_eval_tokens": 506758 + }, + "budget": { + "node_count": 47, + "item_count": 47, + "soft_pair_count": 27, + "serialized_chars": 35243, + "estimated_input_tokens": 8811, + "estimated_eval_tokens": 17878 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": true, + "analyst_result": { + "summary": "Mostly unrelated issues across different models/subsystems. The only soft match that looks like the same underlying bug is the Qwen2.5-VL video-input regression pair. No PRs were provided.", + "confidence": 0.86, + "canonical_issue_reason": "issue:46693 is the strongest standalone bug report: it has broad impact, clear regression framing, and substantial discussion/inbound references around FA2 long-context performance.", + "canonical_pr_reason": null, + "best_issue_reason": "issue:46693 is the most actionable and high-signal issue in the set, with a concrete performance regression and strong community engagement.", + "best_pr_reason": null, + "soft_edge_verdicts": [ + { + "left": "issue:43352", + "right": "issue:46640", + "accept": false, + "reason": "Both mention Flash Attention 2, but one is a model-support error and the other is a performance regression; different bugs and code paths." + }, + { + "left": "issue:44162", + "right": "issue:44792", + "accept": false, + "reason": "ESM2 breakage and a Janus image-generation test failure are unrelated models and failures." + }, + { + "left": "issue:45803", + "right": "issue:46365", + "accept": false, + "reason": "A swallowed-exception bug in Fuyu tensor conversion is unrelated to importing Trainer with datasets installed." + }, + { + "left": "issue:43540", + "right": "issue:44514", + "accept": false, + "reason": "Qwen3OmniMoe video-input ValueError and Qwen2_5_VL batched chat-template crash are different failure modes." + }, + { + "left": "issue:43572", + "right": "issue:46178", + "accept": false, + "reason": "StableLm config padding-token regression and tensorflow_text backend mapping are unrelated." + }, + { + "left": "issue:44610", + "right": "issue:46369", + "accept": false, + "reason": "OmDet-Turbo input-size mismatch and SAM3-LiteText memory overflow do not share the same bug." + }, + { + "left": "issue:44792", + "right": "issue:44977", + "accept": false, + "reason": "Janus test failure and Qwen3.5 flash-attention generation issue are different model/code-path problems." + }, + { + "left": "issue:45381", + "right": "issue:47217", + "accept": false, + "reason": "Both are VL-related, but one is a video-position-id regression and the other is a placeholder-in-user-text parsing failure." + }, + { + "left": "issue:44479", + "right": "issue:47217", + "accept": false, + "reason": "Video-input regression and image-placeholder-in-user-text failure are distinct bugs, even though both involve VL models." + }, + { + "left": "issue:44534", + "right": "issue:44913", + "accept": false, + "reason": "Non-persistent buffer junk initialization is unrelated to GPTNeoX rotary_pct reload behavior." + }, + { + "left": "issue:45542", + "right": "issue:46234", + "accept": false, + "reason": "TensorBoard/TensorFlow backend detection and missing falcon mamba availability helper are unrelated." + }, + { + "left": "issue:44792", + "right": "issue:45600", + "accept": false, + "reason": "A Janus generation test failure and Sam3LiteText auto-mapping breakage are separate issues." + }, + { + "left": "issue:43504", + "right": "issue:45698", + "accept": false, + "reason": "BEiT legacy-field loading failure and wrong custom-module loading after save_pretrained are different serialization/loading bugs." + }, + { + "left": "issue:43525", + "right": "issue:43531", + "accept": false, + "reason": "Missing pad_token_id in Llama4Config and sliding_window problems in Qwen3-MoE are unrelated configuration issues." + }, + { + "left": "issue:45632", + "right": "issue:46424", + "accept": false, + "reason": "Remote-code cache-path collisions and max_cache_length being ignored are distinct cache-management bugs." + }, + { + "left": "issue:35141", + "right": "issue:42490", + "accept": false, + "reason": "Embedding reinitialization on resize and import warnings from trailing-slash local paths are unrelated." + }, + { + "left": "issue:45803", + "right": "issue:46170", + "accept": false, + "reason": "Fuyu exception handling and Windows ROCm/torch.distributed import errors do not share a code path." + }, + { + "left": "issue:46693", + "right": "issue:47217", + "accept": false, + "reason": "FA2 long-context performance regression is unrelated to VL placeholder parsing failures." + }, + { + "left": "issue:45127", + "right": "issue:45200", + "accept": false, + "reason": "LoRA merge collapse with tied embeddings and Gemma 4 mm_token_type_ids defaults are different model behaviors." + }, + { + "left": "issue:44242", + "right": "issue:46894", + "accept": false, + "reason": "Load-balancing loss omission and CohereASR double-shift training loss are both loss-related, but not the same bug." + }, + { + "left": "issue:41950", + "right": "issue:43475", + "accept": false, + "reason": "Video-classification processor mismatch and SAM3 VisionEncoderOutput missing field are unrelated vision bugs." + }, + { + "left": "issue:43931", + "right": "issue:44060", + "accept": false, + "reason": "Qwen3-VL weight-shape mismatch and Qwen3-Next tied-weights warning are different loading/training issues." + }, + { + "left": "issue:43673", + "right": "issue:44743", + "accept": false, + "reason": "Chunked-prefill cache missing and recurrent states resetting with cache are separate cache/state bugs." + }, + { + "left": "issue:43650", + "right": "issue:46202", + "accept": false, + "reason": "An empty-data issue and a Gemma pipeline roundtrip failure are unrelated." + }, + { + "left": "issue:44855", + "right": "issue:45588", + "accept": false, + "reason": "Python 3.13 import parsing failure and flash_attention.py crashing on s_aux=None are different problems." + }, + { + "left": "issue:44792", + "right": "issue:45405", + "accept": false, + "reason": "A Janus test failure and a minimum PEFT version mismatch are unrelated." + }, + { + "left": "issue:29870", + "right": "issue:35141", + "accept": false, + "reason": "Bart generate-time decoder_attention_mask error and embedding reinitialization after resize are distinct bugs." + }, + { + "left": "issue:44479", + "right": "issue:45381", + "accept": true, + "reason": "Both describe the same Qwen2.5-VL video-input regression; 45381 is a more specific manifestation of the broader video-input bug in 44479." + } + ] + }, + "evaluator_result": null, + "evaluator_used": false, + "retried": false, + "accepted_nontrivial_soft_edge": false, + "error_kind": null, + "error_message": null + }, + { + "label": "issue", + "component_index": 1, + "component_count": 28, + "review_unit_index": 40, + "review_unit_count": 65, + "status": "reviewed", + "reason": null, + "source": "llm", + "cache_hit": false, + "model": "gpt-5.4-mini?service_tier=flex", + "cluster_id": "cluster-10000-530", + "nodes": [ + "issue:25251", + "issue:29870", + "issue:39692", + "issue:41720", + "issue:43065", + "issue:43295", + "issue:43335", + "issue:43388", + "issue:43550", + "issue:43644", + "issue:43688", + "issue:43827", + "issue:43854", + "issue:43994", + "issue:44190", + "issue:44479", + "issue:44610", + "issue:44617", + "issue:44734", + "issue:44792", + "issue:44855", + "issue:44869", + "issue:44913", + "issue:44933", + "issue:44964", + "issue:45072", + "issue:45245", + "issue:45276", + "issue:45362", + "issue:45381", + "issue:45406", + "issue:45563", + "issue:45584", + "issue:45588", + "issue:45632", + "issue:46082", + "issue:46369", + "issue:46399", + "issue:46617", + "issue:46655", + "issue:46697", + "issue:46832", + "issue:46858", + "issue:46932", + "issue:47013", + "issue:47312", + "issue:47321" + ], + "soft_pairs": [ + "issue:44190|issue:44964", + "issue:43388|issue:43688", + "issue:46082|issue:46697", + "issue:43827|issue:43994", + "issue:45245|issue:46858", + "issue:45072|issue:46932", + "issue:44617|issue:45584", + "issue:43644|issue:44913", + "issue:45588|issue:47013", + "issue:25251|issue:29870", + "issue:43335|issue:44855", + "issue:44734|issue:45406", + "issue:46399|issue:47312", + "issue:44869|issue:46832", + "issue:45381|issue:46369", + "issue:45632|issue:46617", + "issue:41720|issue:45245", + "issue:43827|issue:47321", + "issue:43854|issue:44792", + "issue:43550|issue:44190", + "issue:39692|issue:47321", + "issue:45563|issue:45584", + "issue:45381|issue:46655", + "issue:44617|issue:46369", + "issue:43065|issue:44610", + "issue:44479|issue:44734", + "issue:43335|issue:44190", + "issue:44933|issue:45362", + "issue:43295|issue:45276", + "issue:39692|issue:44190" + ], + "prepared_review_unit_hash": "8721befed9a11c093bb0938f498cc398fbe9647a5df10c08813368d5264be99c", + "component_budget": { + "node_count": 530, + "item_count": 530, + "soft_pair_count": 1954, + "serialized_chars": 1013004, + "estimated_input_tokens": 253251, + "estimated_eval_tokens": 506758 + }, + "budget": { + "node_count": 47, + "item_count": 47, + "soft_pair_count": 30, + "serialized_chars": 36347, + "estimated_input_tokens": 9087, + "estimated_eval_tokens": 18430 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": true, + "analyst_result": { + "summary": "This set does not form a true duplicate cluster: the soft-pair candidates are mostly unrelated bugs/regressions across different models, processors, and subsystems. I would not merge any of the proposed soft edges, and there is no clear canonical issue/PR target here.", + "confidence": 0.91, + "canonical_issue_reason": null, + "canonical_pr_reason": null, + "best_issue_reason": null, + "best_pr_reason": null, + "soft_edge_verdicts": [ + { + "left": "issue:44190", + "right": "issue:44964", + "accept": false, + "reason": "Different problems: local dataset loading script failure vs multimodal model loading regression." + }, + { + "left": "issue:43388", + "right": "issue:43688", + "accept": false, + "reason": "Both touch evaluation/labels, but one is last-batch label handling and the other is auxiliary-loss normalization in MoE training." + }, + { + "left": "issue:46082", + "right": "issue:46697", + "accept": false, + "reason": "Unrelated subsystems: LlamaConfig head_dim validation vs image processor over-padding." + }, + { + "left": "issue:43827", + "right": "issue:43994", + "accept": false, + "reason": "Docs stale pipeline references vs SigLIP2 model/pipeline output corruption are different issues." + }, + { + "left": "issue:45245", + "right": "issue:46858", + "accept": false, + "reason": "Different runtime failures: category-count limit vs cache sequence-length tensor arithmetic." + }, + { + "left": "issue:45072", + "right": "issue:46932", + "accept": false, + "reason": "Both are dtype-related, but they affect different models and different code paths." + }, + { + "left": "issue:44617", + "right": "issue:45584", + "accept": false, + "reason": "Sam3Video OOM and Whisper empty-transcription crash are unrelated bugs." + }, + { + "left": "issue:43644", + "right": "issue:44913", + "accept": false, + "reason": "Non-persistent buffer initialization bug vs GPTNeoX config reload behavior are separate issues." + }, + { + "left": "issue:45588", + "right": "issue:47013", + "accept": false, + "reason": "Both are crashy edge cases, but they involve different APIs and root causes." + }, + { + "left": "issue:25251", + "right": "issue:29870", + "accept": false, + "reason": "Top-k nesting in pipeline output and Bart decoder_attention_mask UnboundLocalError are unrelated generation bugs." + }, + { + "left": "issue:43335", + "right": "issue:44855", + "accept": false, + "reason": "Sparse-layer config bug vs Python 3.13 IndentationError on import are not the same defect." + }, + { + "left": "issue:44734", + "right": "issue:45406", + "accept": false, + "reason": "KV-cache continuation indexing crash vs missing `_tokenizer` on Gemma4Processor are different failures." + }, + { + "left": "issue:46399", + "right": "issue:47312", + "accept": false, + "reason": "Missing decorator during kernel registration vs incompatibility with kernels 0.16 are related only broadly, not the same bug." + }, + { + "left": "issue:44869", + "right": "issue:46832", + "accept": false, + "reason": "Whisper timestamp decode crash and RT-DETR feature-level mismatch are unrelated." + }, + { + "left": "issue:45381", + "right": "issue:46369", + "accept": false, + "reason": "Qwen video position IDs bug vs Sam3-LiteText memory overflow are different code paths." + }, + { + "left": "issue:45632", + "right": "issue:46617", + "accept": false, + "reason": "Both involve trust_remote_code caching, but one is a leaf-dir collision and the other is a FileNotFoundError; too different to merge." + }, + { + "left": "issue:41720", + "right": "issue:45245", + "accept": false, + "reason": "Auto device mapping CUDA assert vs category limit overflow are unrelated." + }, + { + "left": "issue:43827", + "right": "issue:47321", + "accept": false, + "reason": "Docs about removed pipeline APIs vs AMD quark class update issue are unrelated." + }, + { + "left": "issue:43854", + "right": "issue:44792", + "accept": false, + "reason": "GLM-4.7-Flash test loading failure vs Janus image-generation test failure are different models/tests." + }, + { + "left": "issue:43550", + "right": "issue:44190", + "accept": false, + "reason": "Bamba torch.compile SDPA crash and dataset loading script issue are unrelated." + }, + { + "left": "issue:39692", + "right": "issue:47321", + "accept": false, + "reason": "SigLIP2 docs/model example errors vs AMD quark class update issue are unrelated." + }, + { + "left": "issue:45563", + "right": "issue:45584", + "accept": false, + "reason": "Stale num_return_sequences warning vs Whisper empty-transcription failure are different generation behaviors." + }, + { + "left": "issue:45381", + "right": "issue:46655", + "accept": false, + "reason": "Both are video/image preprocessing issues, but they affect different models and different bugs." + }, + { + "left": "issue:44617", + "right": "issue:46369", + "accept": false, + "reason": "Same Sam3 family, but OOM in Sam3Video is not the same as memory overflow in Sam3-litetext." + }, + { + "left": "issue:43065", + "right": "issue:44610", + "accept": false, + "reason": "Dummy Conv2d in Sam3PixelDecoder vs OmDet-Turbo processor/model size mismatch are unrelated." + }, + { + "left": "issue:44479", + "right": "issue:44734", + "accept": false, + "reason": "Qwen video-input regression and serve KV-cache indexing crash are different subsystems." + }, + { + "left": "issue:43335", + "right": "issue:44190", + "accept": false, + "reason": "SwitchTransformers sparse-layer config bug vs local dataset loading bug are unrelated." + }, + { + "left": "issue:44933", + "right": "issue:45362", + "accept": false, + "reason": "Missing image_utils import and Qwen3.5 chat crash are unrelated." + }, + { + "left": "issue:43295", + "right": "issue:45276", + "accept": false, + "reason": "Processor/tokenizer API regression vs Gemma4 embedding-resize behavior are different bugs." + }, + { + "left": "issue:39692", + "right": "issue:44190", + "accept": false, + "reason": "SigLIP2 docs/model example errors and dataset script loading failure are unrelated." + } + ] + }, + "evaluator_result": null, + "evaluator_used": false, + "retried": false, + "accepted_nontrivial_soft_edge": false, + "error_kind": null, + "error_message": null + }, + { + "label": "issue", + "component_index": 1, + "component_count": 28, + "review_unit_index": 41, + "review_unit_count": 65, + "status": "reviewed", + "reason": null, + "source": "llm", + "cache_hit": false, + "model": "gpt-5.4-mini?service_tier=flex", + "cluster_id": "cluster-10000-530", + "nodes": [ + "issue:27957", + "issue:32090", + "issue:35141", + "issue:42175", + "issue:42796", + "issue:43317", + "issue:43404", + "issue:43526", + "issue:43531", + "issue:43540", + "issue:43611", + "issue:43643", + "issue:43673", + "issue:43854", + "issue:43994", + "issue:44112", + "issue:44242", + "issue:44246", + "issue:44279", + "issue:44462", + "issue:44496", + "issue:44521", + "issue:44734", + "issue:44737", + "issue:44779", + "issue:44964", + "issue:45003", + "issue:45127", + "issue:45216", + "issue:45276", + "issue:45290", + "issue:45357", + "issue:45381", + "issue:45584", + "issue:45600", + "issue:45803", + "issue:45907", + "issue:46421", + "issue:46489", + "issue:46697", + "issue:46752", + "issue:46858", + "issue:46897", + "issue:47145", + "issue:47217", + "issue:47312", + "issue:47321" + ], + "soft_pairs": [ + "issue:44779|issue:46489", + "issue:44242|issue:46897", + "issue:43526|issue:46697", + "issue:43540|issue:44734", + "issue:45127|issue:45357", + "issue:43673|issue:46421", + "issue:45276|issue:45584", + "issue:47145|issue:47321", + "issue:43854|issue:44964", + "issue:43643|issue:44496", + "issue:43994|issue:47321", + "issue:44112|issue:45600", + "issue:44521|issue:45381", + "issue:44462|issue:45907", + "issue:43317|issue:47312", + "issue:43611|issue:47321", + "issue:27957|issue:35141", + "issue:44246|issue:45003", + "issue:44112|issue:44964", + "issue:45290|issue:46752", + "issue:42175|issue:44279", + "issue:43673|issue:46858", + "issue:32090|issue:35141", + "issue:44964|issue:45600", + "issue:43404|issue:45216", + "issue:42796|issue:44521", + "issue:44737|issue:45803", + "issue:43531|issue:45600", + "issue:44496|issue:47217" + ], + "prepared_review_unit_hash": "484e33fbd16c78ca6666e63dd9ce1dffa1dde0a7498da794eb277bfcf18f4b4b", + "component_budget": { + "node_count": 530, + "item_count": 530, + "soft_pair_count": 1954, + "serialized_chars": 1013004, + "estimated_input_tokens": 253251, + "estimated_eval_tokens": 506758 + }, + "budget": { + "node_count": 47, + "item_count": 47, + "soft_pair_count": 29, + "serialized_chars": 35959, + "estimated_input_tokens": 8990, + "estimated_eval_tokens": 18236 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": true, + "analyst_result": { + "summary": "This cluster is mostly heterogeneous issue reports with only superficial overlaps (tokenizers, cache/prefill, multimodal chat templates, model loading). I found no true duplicate pairs among the soft edges.", + "confidence": 0.82, + "canonical_issue_reason": "issue:44521 is the strongest anchor here: it is open, has the most discussion/inbound references, and describes a user-facing multimodal `apply_chat_template` bug that fits the cluster\u2019s most active area.", + "canonical_pr_reason": null, + "best_issue_reason": "issue:44521 is the best single issue to represent the cluster because it is active, broadly impactful, and has the clearest evidence of ongoing interest/impact compared with the other isolated reports.", + "best_pr_reason": null, + "soft_edge_verdicts": [ + { + "left": "issue:44779", + "right": "issue:46489", + "accept": false, + "reason": "Both are DeepSeek tokenizer regressions, but they target different model families and likely different fixes; too broad to treat as the same underlying bug." + }, + { + "left": "issue:44242", + "right": "issue:46897", + "accept": false, + "reason": "Both mention loss computation in MoE-like models, but the concrete failures differ (`output_router_logits=False` vs double-shift training loss) and do not look mergeable as one fix." + }, + { + "left": "issue:43526", + "right": "issue:46697", + "accept": false, + "reason": "Different vision/image-processing bugs in different processors; no shared code path or concrete symptom match." + }, + { + "left": "issue:43540", + "right": "issue:44734", + "accept": false, + "reason": "One is a Qwen3OmniMoe video-input validation error, the other is a serve/KV-cache tensor-indexing crash; unrelated subsystems." + }, + { + "left": "issue:45127", + "right": "issue:45357", + "accept": false, + "reason": "Both concern model save/weight handling, but one is LoRA+extended-vocab collapse and the other is incorrect visual encoder keys in Qwen3.5 checkpoints; not the same bug." + }, + { + "left": "issue:43673", + "right": "issue:46421", + "accept": false, + "reason": "Both touch chunked prefill/cache, but one is a missing cache in `GenerationMixin` and the other is a lazy-init recompilation issue; distinct failure modes." + }, + { + "left": "issue:45276", + "right": "issue:45584", + "accept": false, + "reason": "Unrelated bugs: token embedding resizing vs Whisper empty-transcription generation after special-token alignment." + }, + { + "left": "issue:47145", + "right": "issue:47321", + "accept": false, + "reason": "Completely different areas: FP8 checkpoint loading versus an AMD quark class update issue." + }, + { + "left": "issue:43854", + "right": "issue:44964", + "accept": false, + "reason": "Both are model-loading complaints, but for different models and different root causes; not the same underlying defect." + }, + { + "left": "issue:43643", + "right": "issue:44496", + "accept": false, + "reason": "Both involve config/model loading, but one is `trust_remote_code` returning incomplete configs while the other is a missing `model_type` error; distinct code paths." + }, + { + "left": "issue:43994", + "right": "issue:47321", + "accept": false, + "reason": "No meaningful overlap beyond being issues in the repo; unrelated model inference versus AMD class metadata." + }, + { + "left": "issue:44112", + "right": "issue:45600", + "accept": false, + "reason": "Both cause CI failures, but one is a stale test and the other is a bad auto-mapping reference; different defects." + }, + { + "left": "issue:44521", + "right": "issue:45381", + "accept": false, + "reason": "Both are multimodal/chat-template related, but one is zero assistant masks and the other is wrong vision position ids for Qwen2.5-VL video inputs; too different to merge." + }, + { + "left": "issue:44462", + "right": "issue:45907", + "accept": false, + "reason": "Tokenizer repository loading versus Hub file-listing revision forwarding; unrelated functionality." + }, + { + "left": "issue:43317", + "right": "issue:47312", + "accept": false, + "reason": "Different problems: device-map/offload loading versus kernel compatibility." + }, + { + "left": "issue:43611", + "right": "issue:47321", + "accept": false, + "reason": "Unrelated issues; one is a Transformers 5.0 loading regression, the other an AMD quark update problem." + }, + { + "left": "issue:27957", + "right": "issue:35141", + "accept": false, + "reason": "Flash Attention support for XLMRoberta is unrelated to the token-embedding resize/post-init reinitialization bug." + }, + { + "left": "issue:44246", + "right": "issue:45003", + "accept": false, + "reason": "Import slowness and unsafe `sys.modules` access are not the same concrete bug, even if both touch import-time behavior." + }, + { + "left": "issue:44112", + "right": "issue:44964", + "accept": false, + "reason": "A CI test flake and a model-loading regression are different issues." + }, + { + "left": "issue:45290", + "right": "issue:46752", + "accept": false, + "reason": "Both are `apply_chat_template` crashes, but on different edge cases: tool-call assistant messages versus an empty list." + }, + { + "left": "issue:42175", + "right": "issue:44279", + "accept": false, + "reason": "Packaging/backend extras issue versus a general dependency issue; no same underlying code-path." + }, + { + "left": "issue:43673", + "right": "issue:46858", + "accept": false, + "reason": "Both involve cache/prefill, but the actual bugs differ: missing GenerationMixin cache versus device-tensor `get_seq_length()` breaking arithmetic on non-CUDA backends." + }, + { + "left": "issue:32090", + "right": "issue:35141", + "accept": false, + "reason": "Trainer broadcast TypeError and embedding reinitialization are unrelated." + }, + { + "left": "issue:44964", + "right": "issue:45600", + "accept": false, + "reason": "Different model-loading failures with different causes; not duplicates." + }, + { + "left": "issue:43404", + "right": "issue:45216", + "accept": false, + "reason": "Both relate to tied weights/checkpoints, but they affect different models and different save/load paths." + }, + { + "left": "issue:42796", + "right": "issue:44521", + "accept": false, + "reason": "Tokenizer BPE-marker handling and multimodal assistant-mask generation are unrelated." + }, + { + "left": "issue:44737", + "right": "issue:45803", + "accept": false, + "reason": "CPU-side `arange` device bug in XLNet versus a bare-except error handling problem in Fuyu; no shared underlying defect." + }, + { + "left": "issue:43531", + "right": "issue:45600", + "accept": false, + "reason": "Sliding-window behavior in Qwen3-MoE is unrelated to removed-config references in auto-mappings." + }, + { + "left": "issue:44496", + "right": "issue:47217", + "accept": false, + "reason": "Model config recognition failure versus multimodal placeholder handling; different code paths and symptoms." + } + ] + }, + "evaluator_result": null, + "evaluator_used": false, + "retried": false, + "accepted_nontrivial_soft_edge": false, + "error_kind": null, + "error_message": null + }, + { + "label": "issue", + "component_index": 1, + "component_count": 28, + "review_unit_index": 42, + "review_unit_count": 65, + "status": "reviewed", + "reason": null, + "source": "llm", + "cache_hit": false, + "model": "gpt-5.4-mini?service_tier=flex", + "cluster_id": "cluster-10000-530", + "nodes": [ + "issue:36246", + "issue:38617", + "issue:39692", + "issue:41628", + "issue:41950", + "issue:42907", + "issue:42994", + "issue:43295", + "issue:43335", + "issue:43381", + "issue:43441", + "issue:43526", + "issue:43650", + "issue:43673", + "issue:43901", + "issue:43976", + "issue:44016", + "issue:44062", + "issue:44220", + "issue:44279", + "issue:44295", + "issue:44327", + "issue:44336", + "issue:44496", + "issue:44683", + "issue:44792", + "issue:44938", + "issue:44987", + "issue:45042", + "issue:45083", + "issue:45335", + "issue:45446", + "issue:45542", + "issue:45584", + "issue:45588", + "issue:45593", + "issue:45632", + "issue:45759", + "issue:46082", + "issue:46121", + "issue:46234", + "issue:46424", + "issue:46506", + "issue:46515", + "issue:46752", + "issue:46932", + "issue:47332", + "issue:47333" + ], + "soft_pairs": [ + "issue:44938|issue:45542", + "issue:43335|issue:45593", + "issue:45632|issue:46234", + "issue:36246|issue:41628", + "issue:36246|issue:38617", + "issue:44496|issue:45042", + "issue:43295|issue:45335", + "issue:45083|issue:46506", + "issue:43976|issue:44987", + "issue:44295|issue:46506", + "issue:41950|issue:44016", + "issue:44496|issue:45759", + "issue:39692|issue:46932", + "issue:44683|issue:45446", + "issue:43673|issue:46424", + "issue:43526|issue:43901", + "issue:42994|issue:47333", + "issue:43381|issue:45593", + "issue:43673|issue:44336", + "issue:44016|issue:44220", + "issue:44279|issue:46515", + "issue:46082|issue:46121", + "issue:44016|issue:44062", + "issue:43650|issue:44220", + "issue:42994|issue:47332", + "issue:42907|issue:43441", + "issue:44792|issue:46424", + "issue:43441|issue:45588", + "issue:44327|issue:46752", + "issue:45335|issue:45584", + "issue:43976|issue:44279" + ], + "prepared_review_unit_hash": "a9f834ff13a8cd105610e8ca52d6530fcbad81c1188404ce1ff78468bd5e899f", + "component_budget": { + "node_count": 530, + "item_count": 530, + "soft_pair_count": 1954, + "serialized_chars": 1013004, + "estimated_input_tokens": 253251, + "estimated_eval_tokens": 506758 + }, + "budget": { + "node_count": 48, + "item_count": 48, + "soft_pair_count": 31, + "serialized_chars": 36943, + "estimated_input_tokens": 9236, + "estimated_eval_tokens": 18728 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": true, + "analyst_result": { + "summary": "The set is heterogeneous: the soft-similarity pairs mostly share broad themes (imports, caching, saving, flash-attn, docs) but not the same concrete bug or code path, so I would not merge any of them as duplicates.", + "confidence": 0.93, + "canonical_issue_reason": null, + "canonical_pr_reason": null, + "best_issue_reason": "issue:45584 is the clearest standalone bug report in the set, with a specific repro and a well-scoped failure path; it is not a duplicate canonical for the whole cluster, but it is the strongest representative issue.", + "best_pr_reason": null, + "soft_edge_verdicts": [ + { + "left": "issue:44938", + "right": "issue:45542", + "accept": false, + "reason": "Both are environment-related, but one is a Python 3.14 import/load failure and the other is a TensorBoard-without-TensorFlow backend error; different bugs and fixes." + }, + { + "left": "issue:43335", + "right": "issue:45593", + "accept": false, + "reason": "One is a SwitchTransformers config edge case, the other is missing auxiliary losses in D-FINE; unrelated model logic." + }, + { + "left": "issue:45632", + "right": "issue:46234", + "accept": false, + "reason": "Cache-path collisions for local trust_remote_code models and a missing import_utils helper are unrelated failures." + }, + { + "left": "issue:36246", + "right": "issue:41628", + "accept": false, + "reason": "Both are import errors, but they target different symbols and code paths (`Qwen2_5_VLImageProcessor` vs `AutoImageProcessor`)." + }, + { + "left": "issue:36246", + "right": "issue:38617", + "accept": false, + "reason": "Missing image-processor export and missing `layer_type_validation` are unrelated import problems." + }, + { + "left": "issue:44496", + "right": "issue:45042", + "accept": false, + "reason": "Unrecognized model config and PIL image processors wrongly requiring torchvision are different loading/runtime bugs." + }, + { + "left": "issue:43295", + "right": "issue:45335", + "accept": false, + "reason": "Processor/tokenizer regression and `resize_token_embeddings` not updating `decoder.embed_tokens` are distinct multimodal/model issues." + }, + { + "left": "issue:45083", + "right": "issue:46506", + "accept": false, + "reason": "A helper-length bug in `qwen3_omni_moe` and CPU/CUDA divergence in PPDocLayoutV3 are not the same defect." + }, + { + "left": "issue:43976", + "right": "issue:44987", + "accept": false, + "reason": "Python-version incompatibility and a specific model-loading failure are too broad to be the same underlying bug." + }, + { + "left": "issue:44295", + "right": "issue:46506", + "accept": false, + "reason": "`position_ids` buffer access and CPU/CUDA output mismatches are unrelated code paths." + }, + { + "left": "issue:41950", + "right": "issue:44016", + "accept": false, + "reason": "Pipeline image-processor lookup and a notebook syntax error are clearly unrelated." + }, + { + "left": "issue:44496", + "right": "issue:45759", + "accept": false, + "reason": "Both involve model loading, but one is unrecognized model config and the other is `from_config` not unwrapping `text_config`; different fixes." + }, + { + "left": "issue:39692", + "right": "issue:46932", + "accept": false, + "reason": "A broken documentation example and a Gemma4 quantized-embedder dtype bug are not the same issue." + }, + { + "left": "issue:44683", + "right": "issue:45446", + "accept": false, + "reason": "Both mention flex_attention, but one is a torch>=2.9 compiled failure and the other is a version-check/import-guard bug." + }, + { + "left": "issue:43673", + "right": "issue:46424", + "accept": false, + "reason": "Chunked-prefill cache missing and `max_cache_length` only affecting Quantized Cache are different cache behaviors." + }, + { + "left": "issue:43526", + "right": "issue:43901", + "accept": false, + "reason": "A BEiT fast image-processor label bug and stale docs for `return_all_scores` are unrelated." + }, + { + "left": "issue:42994", + "right": "issue:47333", + "accept": false, + "reason": "General quantized-model saving and disk-offloaded Qwen3 saving failures are not the same concrete save path." + }, + { + "left": "issue:43381", + "right": "issue:45593", + "accept": false, + "reason": "Eval-mode gradient checkpointing and missing auxiliary losses are unrelated training issues." + }, + { + "left": "issue:43673", + "right": "issue:44336", + "accept": false, + "reason": "Cache handling during chunked prefill and ANSI output formatting in loading_report have nothing in common beyond being bugs." + }, + { + "left": "issue:44016", + "right": "issue:44220", + "accept": false, + "reason": "A notebook syntax error and `_torch_extract_fbank_features()` failure are different parts of the project." + }, + { + "left": "issue:44279", + "right": "issue:46515", + "accept": false, + "reason": "A generic dependency issue and a torch 2.4.1 incompatibility are too broad to merge." + }, + { + "left": "issue:46082", + "right": "issue:46121", + "accept": false, + "reason": "Llama head-dim validation and RoPE parameter conversion with JSON-loaded lists are separate configuration bugs." + }, + { + "left": "issue:44016", + "right": "issue:44062", + "accept": false, + "reason": "Notebook syntax and `AddedToken(..., special=...)` argument duplication are unrelated." + }, + { + "left": "issue:43650", + "right": "issue:44220", + "accept": false, + "reason": "`ADD THE DATA` is not a meaningful duplicate target for the audio feature extraction bug." + }, + { + "left": "issue:42994", + "right": "issue:47332", + "accept": false, + "reason": "Both concern saving, but one is quantized-model saving and the other is disk-offloaded Qwen3-VL saving; different mechanisms." + }, + { + "left": "issue:42907", + "right": "issue:43441", + "accept": false, + "reason": "Dequantized-model saving and FlashAttention failures in Ministral-3 are different regressions." + }, + { + "left": "issue:44792", + "right": "issue:46424", + "accept": false, + "reason": "Janus image-generation test failure and cache max-length behavior are unrelated." + }, + { + "left": "issue:43441", + "right": "issue:45588", + "accept": false, + "reason": "Both mention flash attention, but they describe different failure modes and code paths, so they are not duplicate fixes." + }, + { + "left": "issue:44327", + "right": "issue:46752", + "accept": false, + "reason": "QA decode-spans top-k bounds error and empty chat-template IndexError are different edge cases." + }, + { + "left": "issue:45335", + "right": "issue:45584", + "accept": false, + "reason": "Embedding resize propagation and empty-transcription Whisper generation failure are unrelated." + }, + { + "left": "issue:43976", + "right": "issue:44279", + "accept": false, + "reason": "A Transformers/Python version compatibility issue and a dependency issue are too broad to be the same bug." + } + ] + }, + "evaluator_result": null, + "evaluator_used": false, + "retried": false, + "accepted_nontrivial_soft_edge": false, + "error_kind": null, + "error_message": null + }, + { + "label": "issue", + "component_index": 1, + "component_count": 28, + "review_unit_index": 43, + "review_unit_count": 65, + "status": "reviewed", + "reason": null, + "source": "llm", + "cache_hit": false, + "model": "gpt-5.4-mini?service_tier=flex", + "cluster_id": "cluster-10000-530", + "nodes": [ + "issue:39692", + "issue:41950", + "issue:42796", + "issue:43295", + "issue:43299", + "issue:43381", + "issue:43479", + "issue:43577", + "issue:43644", + "issue:43650", + "issue:43653", + "issue:43824", + "issue:43854", + "issue:43901", + "issue:43931", + "issue:43976", + "issue:44016", + "issue:44075", + "issue:44279", + "issue:44291", + "issue:44351", + "issue:44368", + "issue:44462", + "issue:44466", + "issue:44496", + "issue:44779", + "issue:44933", + "issue:45020", + "issue:45081", + "issue:45216", + "issue:45276", + "issue:45278", + "issue:45357", + "issue:45375", + "issue:45405", + "issue:45593", + "issue:45812", + "issue:45823", + "issue:46082", + "issue:46123", + "issue:46209", + "issue:46369", + "issue:46399", + "issue:46614", + "issue:47000", + "issue:47085", + "issue:47145", + "issue:47217" + ], + "soft_pairs": [ + "issue:43824|issue:44933", + "issue:43295|issue:44779", + "issue:43901|issue:44496", + "issue:41950|issue:45020", + "issue:43299|issue:45216", + "issue:44279|issue:45405", + "issue:43901|issue:47217", + "issue:39692|issue:43901", + "issue:42796|issue:43650", + "issue:44779|issue:45812", + "issue:43577|issue:45216", + "issue:46614|issue:47085", + "issue:43479|issue:45375", + "issue:41950|issue:45823", + "issue:46123|issue:46369", + "issue:44466|issue:45276", + "issue:43976|issue:45405", + "issue:43653|issue:45593", + "issue:44075|issue:44368", + "issue:43381|issue:47000", + "issue:44462|issue:45081", + "issue:43299|issue:45357", + "issue:43644|issue:47145", + "issue:43299|issue:43854", + "issue:43650|issue:46209", + "issue:43931|issue:47145", + "issue:44351|issue:45278", + "issue:44016|issue:44291", + "issue:46082|issue:46399" + ], + "prepared_review_unit_hash": "1f2d1a88e50184dca60eef4f777a59d26da7aac509fce5fd35ce81c6a292aa97", + "component_budget": { + "node_count": 530, + "item_count": 530, + "soft_pair_count": 1954, + "serialized_chars": 1013004, + "estimated_input_tokens": 253251, + "estimated_eval_tokens": 506758 + }, + "budget": { + "node_count": 48, + "item_count": 48, + "soft_pair_count": 29, + "serialized_chars": 36572, + "estimated_input_tokens": 9143, + "estimated_eval_tokens": 18542 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": true, + "analyst_result": { + "summary": "The cluster is mostly a set of unrelated Transformers issues linked only by broad similarity (tokenizers, import errors, model loading, docs, regressions). I found no convincing duplicate pairs; all soft-edge candidates should stay separate.", + "confidence": 0.88, + "canonical_issue_reason": "If one anchor is needed, issue:45278 is the broadest regression umbrella in the set, but this cluster is heterogeneous and it should not be treated as a true duplicate canonical for the others.", + "canonical_pr_reason": null, + "best_issue_reason": "Issue:45278 is the most globally representative of the cluster\u2019s broad version-regression theme, but only as a loose anchor, not as a duplicate target.", + "best_pr_reason": null, + "soft_edge_verdicts": [ + { + "left": "issue:43824", + "right": "issue:44933", + "accept": false, + "reason": "Both are import-related, but one is a missing model export and the other is an image_utils import path problem; different code paths and fixes." + }, + { + "left": "issue:43295", + "right": "issue:44779", + "accept": false, + "reason": "One is a processor/tokenizer API regression for multimodal inputs; the other is Deepseek token ID corruption. Same broad area, not the same bug." + }, + { + "left": "issue:43901", + "right": "issue:44496", + "accept": false, + "reason": "Documentation mismatch for a pipeline option versus an unrecognized-model/config loading failure; unrelated failures." + }, + { + "left": "issue:41950", + "right": "issue:45020", + "accept": false, + "reason": "Video pipeline processor selection bug versus remote_code compatibility regressions; different concrete problems." + }, + { + "left": "issue:43299", + "right": "issue:45216", + "accept": false, + "reason": "Both involve Qwen-family models, but one is a load-time MoE model issue and the other is incorrect save_pretrained serialization." + }, + { + "left": "issue:44279", + "right": "issue:45405", + "accept": false, + "reason": "Generic dependency issue versus a specific MIN_PEFT_VERSION bump to an unreleased package; not the same underlying bug." + }, + { + "left": "issue:43901", + "right": "issue:47217", + "accept": false, + "reason": "Docs mentioning return_all_scores versus a VL runtime failure with image placeholders; no shared fix." + }, + { + "left": "issue:39692", + "right": "issue:43901", + "accept": false, + "reason": "Both are documentation issues, but they affect different examples and behaviors." + }, + { + "left": "issue:42796", + "right": "issue:43650", + "accept": false, + "reason": "Tokenizer BPE marker handling is unrelated to the 'ADD THE DATA' issue." + }, + { + "left": "issue:44779", + "right": "issue:45812", + "accept": false, + "reason": "Two tokenizer regressions, but for different model families and different symptoms (wrong results vs wrong IDs)." + }, + { + "left": "issue:43577", + "right": "issue:45216", + "accept": false, + "reason": "Blip2 dtype propagation bug versus Qwen3.5 checkpoint serialization bug; different subsystems." + }, + { + "left": "issue:46614", + "right": "issue:47085", + "accept": false, + "reason": "Stale phonemizer language state is unrelated to fast tokenizer thread-safety." + }, + { + "left": "issue:43479", + "right": "issue:45375", + "accept": false, + "reason": "Both are config issues, but they concern different multimodal models and different fields/defaults." + }, + { + "left": "issue:41950", + "right": "issue:45823", + "accept": false, + "reason": "Pipeline processor lookup bug versus Gemma4 device placement mismatch in forward." + }, + { + "left": "issue:46123", + "right": "issue:46369", + "accept": false, + "reason": "Final-batch flag handling in a pipeline versus memory overflow in a specific SAM model path." + }, + { + "left": "issue:44466", + "right": "issue:45276", + "accept": false, + "reason": "Serialization of tied weights depending on device is not the same bug as resize_token_embeddings not updating embeddings." + }, + { + "left": "issue:43976", + "right": "issue:45405", + "accept": false, + "reason": "Python compatibility breakage versus an unreleased dependency version bump; different causes and remedies." + }, + { + "left": "issue:43653", + "right": "issue:45593", + "accept": false, + "reason": "Tokenizer special-token registration versus D-FINE auxiliary-loss logic; unrelated." + }, + { + "left": "issue:44075", + "right": "issue:44368", + "accept": false, + "reason": "Optimizer arguments being ignored has no overlap with the Qwen3.5 tie_word_embeddings warning." + }, + { + "left": "issue:43381", + "right": "issue:47000", + "accept": false, + "reason": "Gradient checkpointing/eval-mode incompatibility versus RTDetrLoss infeasible cost matrix; separate failures." + }, + { + "left": "issue:44462", + "right": "issue:45081", + "accept": false, + "reason": "AutoTokenizer repo-file selection bug versus a Mistral regex patch crash; both tokenizer-related but different concrete code paths." + }, + { + "left": "issue:43299", + "right": "issue:45357", + "accept": false, + "reason": "Both are Qwen3.5/Qwen3VL regressions, but one is model loading and the other is save_pretrained key serialization." + }, + { + "left": "issue:43644", + "right": "issue:47145", + "accept": false, + "reason": "Non-persistent buffer junk initialization versus FP8 weight_scale not being applied; different state-loading bugs." + }, + { + "left": "issue:43299", + "right": "issue:43854", + "accept": false, + "reason": "Different models and different load failures; no shared underlying bug." + }, + { + "left": "issue:43650", + "right": "issue:46209", + "accept": false, + "reason": "The placeholder issue has no relation to FP8 inference on B200." + }, + { + "left": "issue:43931", + "right": "issue:47145", + "accept": false, + "reason": "Qwen3-VL weight-shape mismatch versus FP8 compressed-tensors scaling; different error classes." + }, + { + "left": "issue:44351", + "right": "issue:45278", + "accept": false, + "reason": "A single missing HybridCache export is not enough to merge with a broad 'many import errors' report." + }, + { + "left": "issue:44016", + "right": "issue:44291", + "accept": false, + "reason": "Notebook syntax error versus init_empty_weights argument mismatch; unrelated." + }, + { + "left": "issue:46082", + "right": "issue:46399", + "accept": false, + "reason": "LlamaConfig head_dim validation is unrelated to Qwen3.5 rotary-kernel registration failures." + }, + { + "left": "issue:43381", + "right": "issue:47000", + "accept": false, + "reason": "Re-listed? No; this pair remains unrelated because the failures affect different training components." + }, + { + "left": "issue:44462", + "right": "issue:45081", + "accept": false, + "reason": "Re-listed? No; same verdict: repo tokenizer selection and Mistral regex patching are distinct bugs." + } + ] + }, + "evaluator_result": null, + "evaluator_used": false, + "retried": false, + "accepted_nontrivial_soft_edge": false, + "error_kind": null, + "error_message": null + }, + { + "label": "issue", + "component_index": 1, + "component_count": 28, + "review_unit_index": 44, + "review_unit_count": 65, + "status": "reviewed", + "reason": null, + "source": "llm", + "cache_hit": false, + "model": "gpt-5.4-mini?service_tier=flex", + "cluster_id": "cluster-10000-530", + "nodes": [ + "issue:33290", + "issue:42175", + "issue:42994", + "issue:43317", + "issue:43441", + "issue:43479", + "issue:43525", + "issue:43575", + "issue:43643", + "issue:43673", + "issue:43761", + "issue:43784", + "issue:43873", + "issue:43881", + "issue:43901", + "issue:43976", + "issue:44016", + "issue:44186", + "issue:44242", + "issue:44246", + "issue:44261", + "issue:44263", + "issue:44336", + "issue:44623", + "issue:44671", + "issue:44683", + "issue:44749", + "issue:44779", + "issue:45081", + "issue:45103", + "issue:45200", + "issue:45325", + "issue:45372", + "issue:45397", + "issue:45542", + "issue:45588", + "issue:45656", + "issue:45800", + "issue:45823", + "issue:45834", + "issue:45859", + "issue:45907", + "issue:46082", + "issue:46512", + "issue:46851", + "issue:47000", + "issue:47145", + "issue:47217" + ], + "soft_pairs": [ + "issue:44683|issue:45542", + "issue:46082|issue:46512", + "issue:44242|issue:45200", + "issue:45081|issue:45907", + "issue:45588|issue:47000", + "issue:43873|issue:47145", + "issue:45103|issue:45372", + "issue:42994|issue:43317", + "issue:45859|issue:46851", + "issue:43441|issue:45834", + "issue:44336|issue:46851", + "issue:44671|issue:44779", + "issue:33290|issue:45656", + "issue:44246|issue:47217", + "issue:45542|issue:45800", + "issue:42175|issue:43976", + "issue:44623|issue:44749", + "issue:44016|issue:45823", + "issue:44186|issue:47000", + "issue:44623|issue:45397", + "issue:43525|issue:43881", + "issue:43479|issue:43525", + "issue:43901|issue:44246", + "issue:43784|issue:45542", + "issue:43673|issue:43761", + "issue:43525|issue:43643", + "issue:44261|issue:44263", + "issue:43575|issue:47145", + "issue:45325|issue:46851" + ], + "prepared_review_unit_hash": "90e54d5a6889df894af45e99ef2591b173b768b4dac9198d60cd2a8a02a78469", + "component_budget": { + "node_count": 530, + "item_count": 530, + "soft_pair_count": 1954, + "serialized_chars": 1013004, + "estimated_input_tokens": 253251, + "estimated_eval_tokens": 506758 + }, + "budget": { + "node_count": 48, + "item_count": 48, + "soft_pair_count": 29, + "serialized_chars": 36378, + "estimated_input_tokens": 9095, + "estimated_eval_tokens": 18446 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": true, + "analyst_result": { + "summary": "This set is not a true duplicate cluster: the issues span unrelated bugs across configs, quantization/offload, tokenizers, docs, and model-specific regressions. I found no soft-edge pair that looks like the same underlying bug/change, so all candidate links should be rejected.", + "confidence": 0.46, + "canonical_issue_reason": "issue:43873 is the broadest and most active item here, centered on quantization/offload behavior, which is the closest thing to a shared theme in this otherwise mixed set.", + "canonical_pr_reason": null, + "best_issue_reason": "issue:43873 is the best anchor among weak candidates because it is open, discussion-heavy, and user-facing; however, it is still not a good duplicate representative for most of the other issues.", + "best_pr_reason": null, + "soft_edge_verdicts": [ + { + "left": "issue:44683", + "right": "issue:45542", + "accept": false, + "reason": "Different bugs: compiled flex_attention on newer torch vs missing TensorFlow backend wiring from tensorboard install." + }, + { + "left": "issue:46082", + "right": "issue:46512", + "accept": false, + "reason": "Unrelated config issues: LlamaConfig head_dim validation vs from_config(dtype=...) leaving config.dtype stale." + }, + { + "left": "issue:44242", + "right": "issue:45200", + "accept": false, + "reason": "Different model/code paths: MoE load-balancing loss gating vs Gemma 4 multimodal token type defaults." + }, + { + "left": "issue:45081", + "right": "issue:45907", + "accept": false, + "reason": "Mistral tokenizer regex patch crash is unrelated to list_repo_files not forwarding revision." + }, + { + "left": "issue:45588", + "right": "issue:47000", + "accept": false, + "reason": "Different failures: flash_attention s_aux=None crash vs RTDetrLoss infeasible cost matrix." + }, + { + "left": "issue:43873", + "right": "issue:47145", + "accept": false, + "reason": "Both touch quantization/offload, but one is about offloading not working while the other is silent corruption from missing weight_scale; not the same bug." + }, + { + "left": "issue:45103", + "right": "issue:45372", + "accept": false, + "reason": "Auto-docstring future-annotations crash is unrelated to the mistral_common ReasoningEffort import error." + }, + { + "left": "issue:42994", + "right": "issue:43317", + "accept": false, + "reason": "Related to quantized models, but saving failure and dequantized loading/offload failure are distinct code paths and symptoms." + }, + { + "left": "issue:45859", + "right": "issue:46851", + "accept": false, + "reason": "Same family of position-related Qwen bugs, but one is CPU offload incompatibility and the other is variable-length position_ids behavior." + }, + { + "left": "issue:43441", + "right": "issue:45834", + "accept": false, + "reason": "Different models and failures: Ministral FlashAttention regression vs Kosmos2.5 OCR length index error." + }, + { + "left": "issue:44336", + "right": "issue:46851", + "accept": false, + "reason": "ANSI-code logging issue is unrelated to Qwen position_ids handling." + }, + { + "left": "issue:44671", + "right": "issue:44779", + "accept": false, + "reason": "CamemBERT masked-LM regression and Deepseek tokenizer corruption are different subsystems and bugs." + }, + { + "left": "issue:33290", + "right": "issue:45656", + "accept": false, + "reason": "Both mention deepspeed/optimizer, but OOM in Adafactor and double optimizer.step() are different problems." + }, + { + "left": "issue:44246", + "right": "issue:47217", + "accept": false, + "reason": "Import performance issue is unrelated to VL image-placeholder handling." + }, + { + "left": "issue:45542", + "right": "issue:45800", + "accept": false, + "reason": "TensorBoard/TensorFlow backend packaging bug is unrelated to torch/transformers version incompatibility." + }, + { + "left": "issue:42175", + "right": "issue:43976", + "accept": false, + "reason": "Package extra backend omission and Python version incompatibility are not the same issue." + }, + { + "left": "issue:44623", + "right": "issue:44749", + "accept": false, + "reason": "Processor save_pretrained missing files is unrelated to the Chinese-reported preprocessing slowdown." + }, + { + "left": "issue:44016", + "right": "issue:45823", + "accept": false, + "reason": "Notebook syntax error and Gemma4 device mismatch are different classes of failure." + }, + { + "left": "issue:44186", + "right": "issue:47000", + "accept": false, + "reason": "Tokenizer crash on NER/padding is unrelated to RTDetrLoss infeasible cost matrix." + }, + { + "left": "issue:44623", + "right": "issue:45397", + "accept": false, + "reason": "Both involve model/processor loading, but missing saved files and gemma-4 zero3 from_pretrained are distinct bugs." + }, + { + "left": "issue:43525", + "right": "issue:43881", + "accept": false, + "reason": "Missing pad_token_id in config and glm-4v-9b load failure are not the same concrete bug." + }, + { + "left": "issue:43479", + "right": "issue:43525", + "accept": false, + "reason": "Both are config initialization issues, but they affect different model configs and different missing fields/defaults." + }, + { + "left": "issue:43901", + "right": "issue:44246", + "accept": false, + "reason": "Docs mismatch for return_all_scores is unrelated to slow imports." + }, + { + "left": "issue:43784", + "right": "issue:45542", + "accept": false, + "reason": "Sentence-transformers NameError and undefined TensorFlow backend are separate import failures." + }, + { + "left": "issue:43673", + "right": "issue:43761", + "accept": false, + "reason": "Both are v5 regressions, but one is chunked_prefill cache handling and the other is CLIPVisionModel hidden_states output." + }, + { + "left": "issue:43525", + "right": "issue:43643", + "accept": false, + "reason": "Missing model config field and AutoConfig trust_remote_code field loss are different causes and fixes." + }, + { + "left": "issue:44261", + "right": "issue:44263", + "accept": false, + "reason": "MLA rms_norm_eps precision issue is unrelated to GlmMoeDsaIndexer torch.split return values." + }, + { + "left": "issue:43575", + "right": "issue:47145", + "accept": false, + "reason": "Qwen2 OOM on tensor-parallel load and FP8 weight_scale omission are different quantization/loading bugs." + }, + { + "left": "issue:45325", + "right": "issue:46851", + "accept": false, + "reason": "Both mention position_ids, but the Qwen2.5-VL temporal scaling bug is distinct from Qwen3.5 variable-length input handling." + } + ] + }, + "evaluator_result": null, + "evaluator_used": false, + "retried": false, + "accepted_nontrivial_soft_edge": false, + "error_kind": null, + "error_message": null + }, + { + "label": "issue", + "component_index": 1, + "component_count": 28, + "review_unit_index": 45, + "review_unit_count": 65, + "status": "reviewed", + "reason": null, + "source": "llm", + "cache_hit": false, + "model": "gpt-5.4-mini?service_tier=flex", + "cluster_id": "cluster-10000-530", + "nodes": [ + "issue:42491", + "issue:42617", + "issue:42796", + "issue:43299", + "issue:43441", + "issue:43526", + "issue:43540", + "issue:43550", + "issue:43723", + "issue:43824", + "issue:43976", + "issue:44162", + "issue:44188", + "issue:44246", + "issue:44295", + "issue:44297", + "issue:44336", + "issue:44373", + "issue:44496", + "issue:44589", + "issue:44871", + "issue:45003", + "issue:45072", + "issue:45083", + "issue:45103", + "issue:45362", + "issue:45397", + "issue:45561", + "issue:45563", + "issue:45588", + "issue:45600", + "issue:45712", + "issue:46291", + "issue:46399", + "issue:46506", + "issue:46614", + "issue:46655", + "issue:46665", + "issue:46726", + "issue:46752", + "issue:46821", + "issue:46829", + "issue:46851", + "issue:47145", + "issue:47176", + "issue:47177", + "issue:47217", + "issue:47333" + ], + "soft_pairs": [ + "issue:44589|issue:45362", + "issue:42796|issue:43441", + "issue:45600|issue:45712", + "issue:44188|issue:44336", + "issue:43824|issue:43976", + "issue:45362|issue:47333", + "issue:44246|issue:44496", + "issue:44336|issue:44373", + "issue:43723|issue:44162", + "issue:42617|issue:43976", + "issue:44871|issue:47145", + "issue:47176|issue:47177", + "issue:44297|issue:45397", + "issue:46291|issue:46399", + "issue:45083|issue:46851", + "issue:46655|issue:46726", + "issue:42491|issue:43299", + "issue:46399|issue:46851", + "issue:45003|issue:47217", + "issue:45072|issue:46829", + "issue:43526|issue:45003", + "issue:43540|issue:46752", + "issue:44246|issue:44871", + "issue:45003|issue:46821", + "issue:45103|issue:46399", + "issue:44295|issue:45397", + "issue:45588|issue:46851", + "issue:45561|issue:45563", + "issue:46614|issue:46665", + "issue:43550|issue:46506" + ], + "prepared_review_unit_hash": "106703d18fd7ab25468f0c5acb91a856ffac7444efc15e22b9bbe042497f458c", + "component_budget": { + "node_count": 530, + "item_count": 530, + "soft_pair_count": 1954, + "serialized_chars": 1013004, + "estimated_input_tokens": 253251, + "estimated_eval_tokens": 506758 + }, + "budget": { + "node_count": 48, + "item_count": 48, + "soft_pair_count": 30, + "serialized_chars": 36711, + "estimated_input_tokens": 9178, + "estimated_eval_tokens": 18612 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": true, + "analyst_result": { + "summary": "This cluster is a loose mix of mostly unrelated Transformers issues, many of them model-specific regression reports. None of the proposed soft edges look like the same underlying bug or change, so all should be rejected. No PRs were provided.", + "confidence": 0.96, + "canonical_issue_reason": null, + "canonical_pr_reason": null, + "best_issue_reason": null, + "best_pr_reason": null, + "soft_edge_verdicts": [ + { + "left": "issue:44589", + "right": "issue:45362", + "accept": false, + "reason": "Different failures: Float8 storage lookup vs Qwen3.5 chat crash; no shared code path." + }, + { + "left": "issue:42796", + "right": "issue:43441", + "accept": false, + "reason": "Same model family but different components: tokenizer BPE handling vs FlashAttention runtime failure." + }, + { + "left": "issue:45600", + "right": "issue:45712", + "accept": false, + "reason": "Different cleanup/CI problems: removed Sam3LiteText configs vs leftover dummy classes." + }, + { + "left": "issue:44188", + "right": "issue:44336", + "accept": false, + "reason": "Attention-kernel torch.compile branching vs ANSI codes in loading_report; unrelated subsystems." + }, + { + "left": "issue:43824", + "right": "issue:43976", + "accept": false, + "reason": "Import error for a missing class vs Python version compatibility; distinct root causes." + }, + { + "left": "issue:45362", + "right": "issue:47333", + "accept": false, + "reason": "Qwen3.5 chat crash and disk-offloaded save failure are different features and failures." + }, + { + "left": "issue:44246", + "right": "issue:44496", + "accept": false, + "reason": "Import slowness is unrelated to unrecognized model/config loading." + }, + { + "left": "issue:44336", + "right": "issue:44373", + "accept": false, + "reason": "Terminal ANSI output bug vs docstring typo; not the same bug." + }, + { + "left": "issue:43723", + "right": "issue:44162", + "accept": false, + "reason": "Tokenizer loading/import issue vs ESM2 model breakage; no concrete overlap." + }, + { + "left": "issue:42617", + "right": "issue:43976", + "accept": false, + "reason": "3d_parallel.py runtime issue and Python version support issue are unrelated." + }, + { + "left": "issue:44871", + "right": "issue:47145", + "accept": false, + "reason": "Both involve model loading/config, but one is eos_token_id mismatch and the other is missing FP8 weight_scale application." + }, + { + "left": "issue:47176", + "right": "issue:47177", + "accept": false, + "reason": "Both are security-related, but one is shell-operator allowlist bypass and the other is archive path traversal; different vulnerabilities." + }, + { + "left": "issue:44297", + "right": "issue:45397", + "accept": false, + "reason": "Tokenizer save_pretrained metadata mismatch vs gemma-4 zero3 from_pretrained failure." + }, + { + "left": "issue:46291", + "right": "issue:46399", + "accept": false, + "reason": "Kernels import/version requirement failure vs Qwen3.5 rotary embedding registration crash; unrelated." + }, + { + "left": "issue:45083", + "right": "issue:46851", + "accept": false, + "reason": "Shared Qwen3 naming, but one is a helper length calculation bug and the other is variable-length position_ids handling." + }, + { + "left": "issue:46655", + "right": "issue:46726", + "accept": false, + "reason": "Video padding calculation bug vs missing-vision-backend processor error; different code paths." + }, + { + "left": "issue:42491", + "right": "issue:43299", + "accept": false, + "reason": "Both mention Qwen3 MoE, but the first is LoRA compatibility across hf4.x/hf5.x and the second is a model loading regression in 5.0.0.dev0." + }, + { + "left": "issue:46399", + "right": "issue:46851", + "accept": false, + "reason": "Rotary embedding registration error vs variable-length position_ids issue; not the same defect." + }, + { + "left": "issue:45003", + "right": "issue:47217", + "accept": false, + "reason": "Unsafe sys.modules access is unrelated to image placeholder handling in VL models." + }, + { + "left": "issue:45072", + "right": "issue:46829", + "accept": false, + "reason": "dtype mismatch in bfloat16 inference vs returning the wrong tensor in ImageClassifierOutput; distinct bugs." + }, + { + "left": "issue:43526", + "right": "issue:45003", + "accept": false, + "reason": "BeitImageProcessorFast reduce_labels bug has no relation to sys.modules access." + }, + { + "left": "issue:43540", + "right": "issue:46752", + "accept": false, + "reason": "Qwen3OmniMoe video-processing ValueError and empty-list chat template IndexError are unrelated." + }, + { + "left": "issue:44246", + "right": "issue:44871", + "accept": false, + "reason": "Slow import times vs eos_token_id config inconsistency; no shared underlying cause." + }, + { + "left": "issue:45003", + "right": "issue:46821", + "accept": false, + "reason": "sys.modules access safety issue and TimesFM window_size AttributeError are unrelated." + }, + { + "left": "issue:45103", + "right": "issue:46399", + "accept": false, + "reason": "Auto-docstring future-annotations crash vs rotary embedding registration bug; different modules." + }, + { + "left": "issue:44295", + "right": "issue:45397", + "accept": false, + "reason": "Position_ids buffer read error vs gemma-4 zero3 loading issue; not the same problem." + }, + { + "left": "issue:45588", + "right": "issue:46851", + "accept": false, + "reason": "FlashAttention crash on s_aux=None is unrelated to variable-length position_ids handling." + }, + { + "left": "issue:45561", + "right": "issue:45563", + "accept": false, + "reason": "pytest-xdist captured_info.txt race and stale generate() warning are different testing/generation issues." + }, + { + "left": "issue:46614", + "right": "issue:46665", + "accept": false, + "reason": "Stale phonemizer backend state vs inconsistent from_pretrained inside Python classes; similar statefulness theme, but different behavior and code path." + }, + { + "left": "issue:43550", + "right": "issue:46506", + "accept": false, + "reason": "torch.compile + SDPA failure in Bamba vs CPU/CUDA result divergence in PPDocLayoutV3; different models and bugs." + } + ] + }, + "evaluator_result": null, + "evaluator_used": false, + "retried": false, + "accepted_nontrivial_soft_edge": false, + "error_kind": null, + "error_message": null + }, + { + "label": "issue", + "component_index": 1, + "component_count": 28, + "review_unit_index": 46, + "review_unit_count": 65, + "status": "reviewed", + "reason": null, + "source": "llm", + "cache_hit": false, + "model": "gpt-5.4-mini?service_tier=flex", + "cluster_id": "cluster-10000-530", + "nodes": [ + "issue:36246", + "issue:41950", + "issue:42757", + "issue:42886", + "issue:42915", + "issue:43257", + "issue:43295", + "issue:43352", + "issue:43540", + "issue:43596", + "issue:43901", + "issue:44060", + "issue:44077", + "issue:44162", + "issue:44220", + "issue:44242", + "issue:44295", + "issue:44336", + "issue:44351", + "issue:44451", + "issue:44466", + "issue:44479", + "issue:44749", + "issue:44779", + "issue:44877", + "issue:44995", + "issue:45003", + "issue:45137", + "issue:45200", + "issue:45250", + "issue:45292", + "issue:45335", + "issue:45375", + "issue:45561", + "issue:45593", + "issue:45600", + "issue:45701", + "issue:45759", + "issue:46132", + "issue:46170", + "issue:46424", + "issue:46655", + "issue:46726", + "issue:46752", + "issue:47217", + "issue:47312", + "issue:47321" + ], + "soft_pairs": [ + "issue:44336|issue:46424", + "issue:43540|issue:46655", + "issue:36246|issue:43596", + "issue:42915|issue:47312", + "issue:44779|issue:45292", + "issue:44242|issue:45593", + "issue:43257|issue:45137", + "issue:43901|issue:45003", + "issue:44479|issue:46655", + "issue:41950|issue:44220", + "issue:46726|issue:47217", + "issue:45375|issue:45759", + "issue:45292|issue:45701", + "issue:36246|issue:44351", + "issue:36246|issue:42757", + "issue:43295|issue:45292", + "issue:42886|issue:46132", + "issue:45137|issue:45600", + "issue:44466|issue:45335", + "issue:44451|issue:46752", + "issue:43352|issue:45250", + "issue:44295|issue:44749", + "issue:46170|issue:47321", + "issue:44162|issue:45200", + "issue:44060|issue:44077", + "issue:44995|issue:45200", + "issue:44877|issue:45375", + "issue:45003|issue:45561" + ], + "prepared_review_unit_hash": "94a25f52ccb2203d62ea579f03cc3e9a85c404335913ac0fc2de20243d55d424", + "component_budget": { + "node_count": 530, + "item_count": 530, + "soft_pair_count": 1954, + "serialized_chars": 1013004, + "estimated_input_tokens": 253251, + "estimated_eval_tokens": 506758 + }, + "budget": { + "node_count": 47, + "item_count": 47, + "soft_pair_count": 28, + "serialized_chars": 35359, + "estimated_input_tokens": 8840, + "estimated_eval_tokens": 17936 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": true, + "analyst_result": { + "summary": "All candidate soft edges are false positives: the paired issues usually share a broad subsystem label (e.g. video, Qwen, deepspeed, tokenization) but describe different failures, different code paths, or different models. No PRs are present.", + "confidence": 0.84, + "canonical_issue_reason": "Issue 44479 is the strongest representative issue in this set: it describes a concrete, broad regression in video-input handling across multiple Qwen-family models, making it a plausible canonical duplicate target if one were needed.", + "canonical_pr_reason": null, + "best_issue_reason": "Issue 44479 is the best overall issue candidate because it is specific, user-facing, and spans several related model variants rather than a narrow one-off symptom.", + "best_pr_reason": null, + "soft_edge_verdicts": [ + { + "left": "issue:44336", + "right": "issue:46424", + "accept": false, + "reason": "Unrelated bugs: ANSI color output in loading reports vs cache length behavior in quantized cache." + }, + { + "left": "issue:43540", + "right": "issue:46655", + "accept": false, + "reason": "Both mention video, but one is Qwen3OmniMoe input processing and the other is SmolVLM padding math; different models and failures." + }, + { + "left": "issue:36246", + "right": "issue:43596", + "accept": false, + "reason": "ImportError for a missing Qwen image processor is unrelated to a DeepSpeed/BertModel zero3 IndexError." + }, + { + "left": "issue:42915", + "right": "issue:47312", + "accept": false, + "reason": "Qwen3Moe FP8 loading failure and kernel compatibility are different problems with no shared code-path." + }, + { + "left": "issue:44779", + "right": "issue:45292", + "accept": false, + "reason": "Tokenizer correctness regression is not the same as resize_token_embeddings not updating output embeddings." + }, + { + "left": "issue:44242", + "right": "issue:45593", + "accept": false, + "reason": "Load balancing loss omission and D-FINE auxiliary loss behavior are different training bugs." + }, + { + "left": "issue:43257", + "right": "issue:45137", + "accept": false, + "reason": "MoE weight conversion with accelerate/deepspeed is unrelated to a deque underflow in ZeRO3 init." + }, + { + "left": "issue:43901", + "right": "issue:45003", + "accept": false, + "reason": "Docs mismatch for return_all_scores is unrelated to unsafe sys.modules access in modeling_utils." + }, + { + "left": "issue:44479", + "right": "issue:46655", + "accept": false, + "reason": "Both concern video processors, but the regression is broader Qwen-family input handling while the other is a SmolVLM padding calculation bug." + }, + { + "left": "issue:41950", + "right": "issue:44220", + "accept": false, + "reason": "Video-classification pipeline image-processor lookup and _torch_extract_fbank_features are different pipeline issues." + }, + { + "left": "issue:46726", + "right": "issue:47217", + "accept": false, + "reason": "Missing vision backend / unrecognized image processor is not the same as image placeholders inside user text causing VL model failure." + }, + { + "left": "issue:45375", + "right": "issue:45759", + "accept": false, + "reason": "Missing config field dropped by strict parsing is unrelated to AutoModelForCausalLM not unwrapping text_config." + }, + { + "left": "issue:45292", + "right": "issue:45701", + "accept": false, + "reason": "resize_token_embeddings wiring bug and tokenization changes across versions are different symptoms and code paths." + }, + { + "left": "issue:36246", + "right": "issue:44351", + "accept": false, + "reason": "Two unrelated import errors: missing Qwen2_5_VLImageProcessor vs missing HybridCache export." + }, + { + "left": "issue:36246", + "right": "issue:42757", + "accept": false, + "reason": "Missing Qwen image processor and missing is_offline_mode are separate import surface regressions." + }, + { + "left": "issue:43295", + "right": "issue:45292", + "accept": false, + "reason": "Custom processor.tokenizer regression is not the same as resize_token_embeddings not affecting output embeddings." + }, + { + "left": "issue:42886", + "right": "issue:46132", + "accept": false, + "reason": "Offline cache loading failure and AttentionInterface.register behavior change are unrelated." + }, + { + "left": "issue:45137", + "right": "issue:45600", + "accept": false, + "reason": "ZeRO3 deque/indexing crash is unrelated to auto_mappings still referencing removed configs." + }, + { + "left": "issue:44466", + "right": "issue:45335", + "accept": false, + "reason": "Both involve tied/resized embeddings, but one is serialization inconsistency and the other is decoder.embed_tokens not updating." + }, + { + "left": "issue:44451", + "right": "issue:46752", + "accept": false, + "reason": "Model loading failure for ScandiBERT is unrelated to apply_chat_template on an empty list." + }, + { + "left": "issue:43352", + "right": "issue:45250", + "accept": false, + "reason": "Model-specific unsupported Flash Attention 2 warning is not the same as the general Flash Attention 2.0 issue." + }, + { + "left": "issue:44295", + "right": "issue:44749", + "accept": false, + "reason": "A position_ids buffer access error is different from a performance regression after upgrading Transformers." + }, + { + "left": "issue:46170", + "right": "issue:47321", + "accept": false, + "reason": "Windows/ROCm import failure due to torch.distributed is unrelated to AMD quark class updates." + }, + { + "left": "issue:44162", + "right": "issue:45200", + "accept": false, + "reason": "ESM2 breakage and Gemma 4 mm_token_type_ids defaulting are different model bugs." + }, + { + "left": "issue:44060", + "right": "issue:44077", + "accept": false, + "reason": "Incorrect tied-weights warning in Qwen3-Next is unrelated to patchtsmixer's post_init allowance." + }, + { + "left": "issue:44995", + "right": "issue:45200", + "accept": false, + "reason": "GlmMoeDsa stale indexer cache and Gemma 4 token type IDs are separate issues." + }, + { + "left": "issue:44877", + "right": "issue:45375", + "accept": false, + "reason": "Strict config blocking granite_speech loading is not the same as missing deepstack_visual_indexes in Qwen3_5MoeVisionConfig." + }, + { + "left": "issue:45003", + "right": "issue:45561", + "accept": false, + "reason": "Unsafe sys.modules access is unrelated to the xdist captured_info.txt race." + } + ] + }, + "evaluator_result": null, + "evaluator_used": false, + "retried": false, + "accepted_nontrivial_soft_edge": false, + "error_kind": null, + "error_message": null + }, + { + "label": "issue", + "component_index": 1, + "component_count": 28, + "review_unit_index": 47, + "review_unit_count": 65, + "status": "reviewed", + "reason": null, + "source": "llm", + "cache_hit": false, + "model": "gpt-5.4-mini?service_tier=flex", + "cluster_id": "cluster-10000-530", + "nodes": [ + "issue:29870", + "issue:30333", + "issue:42199", + "issue:42886", + "issue:42994", + "issue:43065", + "issue:43408", + "issue:43475", + "issue:43540", + "issue:43653", + "issue:43723", + "issue:43906", + "issue:43931", + "issue:44038", + "issue:44162", + "issue:44188", + "issue:44242", + "issue:44361", + "issue:44368", + "issue:44479", + "issue:44623", + "issue:44749", + "issue:44857", + "issue:44877", + "issue:44977", + "issue:44987", + "issue:44995", + "issue:45229", + "issue:45397", + "issue:45412", + "issue:45459", + "issue:45563", + "issue:45588", + "issue:45656", + "issue:45685", + "issue:45987", + "issue:46082", + "issue:46123", + "issue:46170", + "issue:46202", + "issue:46369", + "issue:46431", + "issue:46506", + "issue:46710", + "issue:46726", + "issue:46851", + "issue:47312" + ], + "soft_pairs": [ + "issue:43475|issue:46369", + "issue:44479|issue:46170", + "issue:44987|issue:45459", + "issue:44995|issue:45588", + "issue:29870|issue:30333", + "issue:44623|issue:46726", + "issue:44038|issue:46506", + "issue:44877|issue:46170", + "issue:45685|issue:46431", + "issue:44242|issue:44857", + "issue:45397|issue:45656", + "issue:43475|issue:43540", + "issue:44368|issue:44977", + "issue:43653|issue:44162", + "issue:42994|issue:47312", + "issue:43723|issue:46710", + "issue:42199|issue:46851", + "issue:43065|issue:43408", + "issue:45412|issue:46123", + "issue:42886|issue:45229", + "issue:43931|issue:44368", + "issue:43906|issue:46202", + "issue:44188|issue:46202", + "issue:44361|issue:44749", + "issue:44368|issue:45563", + "issue:45987|issue:46082" + ], + "prepared_review_unit_hash": "7ad5713d6f7fbb818ee1aff5dc6508985e5036a6907b031b541ee0a3726926da", + "component_budget": { + "node_count": 530, + "item_count": 530, + "soft_pair_count": 1954, + "serialized_chars": 1013004, + "estimated_input_tokens": 253251, + "estimated_eval_tokens": 506758 + }, + "budget": { + "node_count": 47, + "item_count": 47, + "soft_pair_count": 26, + "serialized_chars": 34834, + "estimated_input_tokens": 8709, + "estimated_eval_tokens": 17674 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": true, + "analyst_result": { + "summary": "This set is not a true duplicate cluster; it contains many unrelated bugs across tokenizers, vision models, MoE, caching, and training/inference paths. No PRs are present. All soft-edge pairs should be rejected.", + "confidence": 0.98, + "canonical_issue_reason": "No canonical issue: the items are heterogeneous and do not share one underlying defect or code path.", + "canonical_pr_reason": null, + "best_issue_reason": "No global best issue: none clearly subsumes the others, so selecting one as the representative would be arbitrary.", + "best_pr_reason": null, + "soft_edge_verdicts": [ + { + "left": "issue:43475", + "right": "issue:46369", + "accept": false, + "reason": "Both involve SAM3, but one is an encoder attribute error and the other is a litetext memory overflow; different failures and code paths." + }, + { + "left": "issue:44479", + "right": "issue:46170", + "accept": false, + "reason": "Qwen video-input regression and a Windows/ROCm distributed import error are unrelated bugs." + }, + { + "left": "issue:44987", + "right": "issue:45459", + "accept": false, + "reason": "One is a model loading failure for a specific repo; the other is tokenizer error masking when protobuf is missing." + }, + { + "left": "issue:44995", + "right": "issue:45588", + "accept": false, + "reason": "Different subsystems and symptoms: GlmMoeDsa stale cache on second forward versus flash-attention crashing on s_aux=None." + }, + { + "left": "issue:29870", + "right": "issue:30333", + "accept": false, + "reason": "A BART generate() UnboundLocalError and MLflow reporting on failed jobs are unrelated." + }, + { + "left": "issue:44623", + "right": "issue:46726", + "accept": false, + "reason": "Both mention processors, but saving missing files and loading with missing vision backend are distinct issues." + }, + { + "left": "issue:44038", + "right": "issue:46506", + "accept": false, + "reason": "A Qwen3-VL-Moe v5 bug and CPU/CUDA divergence in PPDocLayoutV3 do not share a concrete failure path." + }, + { + "left": "issue:44877", + "right": "issue:46170", + "accept": false, + "reason": "Strict Granite Speech config loading and a Windows ROCm torch.distributed import problem are unrelated." + }, + { + "left": "issue:45685", + "right": "issue:46431", + "accept": false, + "reason": "MoE histogram_mps on Int and missing chat template are different errors in different components." + }, + { + "left": "issue:44242", + "right": "issue:44857", + "accept": false, + "reason": "Loss routing omission in MoE and an AMP/CUDA crash in LwDetrImageLoss are not the same bug." + }, + { + "left": "issue:45397", + "right": "issue:45656", + "accept": false, + "reason": "Deepspeed from_pretrained loading and duplicated optimizer stepping are separate problems." + }, + { + "left": "issue:43475", + "right": "issue:43540", + "accept": false, + "reason": "Different SAM3 video/vision error versus Qwen3OmniMoe video input processing error." + }, + { + "left": "issue:44368", + "right": "issue:44977", + "accept": false, + "reason": "A warning about tie_word_embeddings in fine-tuning is not the same as a flash-attention generation failure." + }, + { + "left": "issue:43653", + "right": "issue:44162", + "accept": false, + "reason": "BigBirdTokenizer special-token registration and ESM2 breakage are unrelated." + }, + { + "left": "issue:42994", + "right": "issue:47312", + "accept": false, + "reason": "Quantized model save failure and kernel compatibility mismatch are different topics." + }, + { + "left": "issue:43723", + "right": "issue:46710", + "accept": false, + "reason": "Tokenizer loading in v5 and accuracy regression after an upgrade are not the same underlying bug." + }, + { + "left": "issue:42199", + "right": "issue:46851", + "accept": false, + "reason": "DETR cardinality/background-class logic and Qwen3.5 variable-length position_ids are unrelated." + }, + { + "left": "issue:43065", + "right": "issue:43408", + "accept": false, + "reason": "Dummy Conv2d in Sam3PixelDecoder and a model-type mismatch warning are separate SAM3 issues." + }, + { + "left": "issue:45412", + "right": "issue:46123", + "accept": false, + "reason": "RT-DETR memory retention and MaskGenerationPipeline final-batch dropping are unrelated." + }, + { + "left": "issue:42886", + "right": "issue:45229", + "accept": false, + "reason": "Offline tokenizer cache loading and Gemma4 multi-GPU CUDA OOM are different failures." + }, + { + "left": "issue:43931", + "right": "issue:44368", + "accept": false, + "reason": "Qwen3-VL-30B weight-shape mismatch and a Qwen3.5 warning about tied embeddings are not the same bug." + }, + { + "left": "issue:43906", + "right": "issue:46202", + "accept": false, + "reason": "An isolated reproduction of another issue and a Gemma pipeline roundtrip failure do not match as a concrete duplicate." + }, + { + "left": "issue:44188", + "right": "issue:46202", + "accept": false, + "reason": "Torch.compile attention-kernel divergence and Gemma roundtrip tokenization failure are unrelated." + }, + { + "left": "issue:44361", + "right": "issue:44749", + "accept": false, + "reason": "MLukeTokenizer AttributeError and a transformers performance regression are different problems." + }, + { + "left": "issue:44368", + "right": "issue:45563", + "accept": false, + "reason": "A fine-tuning warning and a stale generate() warning for num_return_sequences are not the same issue." + }, + { + "left": "issue:45987", + "right": "issue:46082", + "accept": false, + "reason": "StaticCache return-type contract violation and LlamaConfig head_dim validation are unrelated." + } + ] + }, + "evaluator_result": null, + "evaluator_used": false, + "retried": false, + "accepted_nontrivial_soft_edge": false, + "error_kind": null, + "error_message": null + }, + { + "label": "issue", + "component_index": 1, + "component_count": 28, + "review_unit_index": 48, + "review_unit_count": 65, + "status": "reviewed", + "reason": null, + "source": "llm", + "cache_hit": false, + "model": "gpt-5.4-mini?service_tier=flex", + "cluster_id": "cluster-10000-530", + "nodes": [ + "issue:28282", + "issue:30990", + "issue:39692", + "issue:43065", + "issue:43475", + "issue:43643", + "issue:43646", + "issue:43701", + "issue:43723", + "issue:43784", + "issue:43994", + "issue:44162", + "issue:44188", + "issue:44297", + "issue:44373", + "issue:44483", + "issue:44623", + "issue:44749", + "issue:44871", + "issue:44964", + "issue:45103", + "issue:45137", + "issue:45230", + "issue:45375", + "issue:45397", + "issue:45478", + "issue:45561", + "issue:45593", + "issue:45656", + "issue:46123", + "issue:46132", + "issue:46202", + "issue:46326", + "issue:46399", + "issue:46421", + "issue:46463", + "issue:46726", + "issue:46752", + "issue:46821", + "issue:46851", + "issue:46858", + "issue:46877", + "issue:47000", + "issue:47020", + "issue:47068", + "issue:47145", + "issue:47217" + ], + "soft_pairs": [ + "issue:46752|issue:47020", + "issue:46132|issue:47145", + "issue:45103|issue:46858", + "issue:44483|issue:46202", + "issue:45593|issue:47068", + "issue:43994|issue:46726", + "issue:44188|issue:46421", + "issue:43701|issue:44749", + "issue:44749|issue:45656", + "issue:44297|issue:44623", + "issue:43065|issue:46123", + "issue:43646|issue:44162", + "issue:44964|issue:45137", + "issue:43475|issue:46821", + "issue:43475|issue:43643", + "issue:44623|issue:45656", + "issue:43784|issue:46399", + "issue:39692|issue:46726", + "issue:46463|issue:46877", + "issue:44373|issue:46851", + "issue:28282|issue:30990", + "issue:44871|issue:46132", + "issue:45561|issue:47217", + "issue:43723|issue:46326", + "issue:46877|issue:47000", + "issue:44188|issue:46851", + "issue:45375|issue:45478", + "issue:45230|issue:45397" + ], + "prepared_review_unit_hash": "422bc0a996841b5031dfb03cb65deaacfdb170ec6e2f1e917711738b0dbf8fad", + "component_budget": { + "node_count": 530, + "item_count": 530, + "soft_pair_count": 1954, + "serialized_chars": 1013004, + "estimated_input_tokens": 253251, + "estimated_eval_tokens": 506758 + }, + "budget": { + "node_count": 47, + "item_count": 47, + "soft_pair_count": 28, + "serialized_chars": 35477, + "estimated_input_tokens": 8870, + "estimated_eval_tokens": 17996 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": true, + "analyst_result": { + "summary": "This cluster is highly heterogeneous: the paired items share only superficial wording or broad subsystems, not the same underlying bug/change. I would not collapse any of the soft-edge pairs.", + "confidence": 0.97, + "canonical_issue_reason": null, + "canonical_pr_reason": null, + "best_issue_reason": null, + "best_pr_reason": null, + "soft_edge_verdicts": [ + { + "left": "issue:46752", + "right": "issue:47020", + "accept": false, + "reason": "Different tokenizer APIs and failures: empty chat template handling vs Unigram init type mismatch." + }, + { + "left": "issue:46132", + "right": "issue:47145", + "accept": false, + "reason": "Unrelated areas: attention registration behavior vs FP8 checkpoint weight scaling on load." + }, + { + "left": "issue:45103", + "right": "issue:46858", + "accept": false, + "reason": "Docstring/introspection crash vs cache length tensor arithmetic; no shared code path." + }, + { + "left": "issue:44483", + "right": "issue:46202", + "accept": false, + "reason": "API request rejection vs Gemma roundtrip tokenization issue; different failure modes." + }, + { + "left": "issue:45593", + "right": "issue:47068", + "accept": false, + "reason": "Both training-related, but one is loss composition and the other is a CUDA conditional sync bug." + }, + { + "left": "issue:43994", + "right": "issue:46726", + "accept": false, + "reason": "Both vision/model loading related, but one is a broken example and the other is missing backend handling." + }, + { + "left": "issue:44188", + "right": "issue:46421", + "accept": false, + "reason": "Compile-time attention kernel branching vs cache lazy-init recompilation; distinct mechanisms." + }, + { + "left": "issue:43701", + "right": "issue:44749", + "accept": false, + "reason": "Checkpoint key mismatch and training slowdown are unrelated." + }, + { + "left": "issue:44749", + "right": "issue:45656", + "accept": false, + "reason": "Performance regression from data filtering vs duplicate optimizer.step with DeepSpeed." + }, + { + "left": "issue:44297", + "right": "issue:44623", + "accept": false, + "reason": "Both save_pretrained-related, but tokenizer metadata mismatch is not the same as missing processor files." + }, + { + "left": "issue:43065", + "right": "issue:46123", + "accept": false, + "reason": "SAM3 pixel decoder stub vs mask pipeline batch-end logic; no common bug." + }, + { + "left": "issue:43646", + "right": "issue:44162", + "accept": false, + "reason": "Custom config field loss vs ESM2 model breakage; too different to merge." + }, + { + "left": "issue:44964", + "right": "issue:45137", + "accept": false, + "reason": "Model loading failure vs DeepSpeed ZeRO3 deque error; unrelated code paths." + }, + { + "left": "issue:43475", + "right": "issue:46821", + "accept": false, + "reason": "Missing vision encoder attribute vs TimesFM window_size attribute error; different models and causes." + }, + { + "left": "issue:43475", + "right": "issue:43643", + "accept": false, + "reason": "Different missing-field symptoms in different objects; not the same defect." + }, + { + "left": "issue:44623", + "right": "issue:45656", + "accept": false, + "reason": "Save-time artifact omission vs optimizer stepping twice; unrelated bugs." + }, + { + "left": "issue:43784", + "right": "issue:46399", + "accept": false, + "reason": "Import NameError in sentence-transformers vs Qwen3.5 kernel registration crash." + }, + { + "left": "issue:39692", + "right": "issue:46726", + "accept": false, + "reason": "Doc example mistakes vs backend-missing processor error; not duplicate." + }, + { + "left": "issue:46463", + "right": "issue:46877", + "accept": false, + "reason": "T5 validation messaging vs auto_docstring stdout logging; unrelated." + }, + { + "left": "issue:44373", + "right": "issue:46851", + "accept": false, + "reason": "Wrong docstring only vs runtime position_ids bug; different kinds of fixes." + }, + { + "left": "issue:28282", + "right": "issue:30990", + "accept": false, + "reason": "PyTorch-missing import error vs hanging load in sentence-transformers; unrelated." + }, + { + "left": "issue:44871", + "right": "issue:46132", + "accept": false, + "reason": "Gemma eos_token_id config mismatch vs attention registration behavior change." + }, + { + "left": "issue:45561", + "right": "issue:47217", + "accept": false, + "reason": "pytest-xdist file race vs multimodal placeholder parsing; no overlap." + }, + { + "left": "issue:43723", + "right": "issue:46326", + "accept": false, + "reason": "Tokenizer loading regression in v5 vs continue_final_message handling for mistral_common." + }, + { + "left": "issue:46877", + "right": "issue:47000", + "accept": false, + "reason": "Import-time doc-lint output vs RTDetr loss infeasibility; unrelated." + }, + { + "left": "issue:44188", + "right": "issue:46851", + "accept": false, + "reason": "Both involve Qwen3.5/positioning, but one is torch.compile kernel divergence and the other is variable-length input handling." + }, + { + "left": "issue:45375", + "right": "issue:45478", + "accept": false, + "reason": "Missing config field silently dropped vs broader Qwen3.5 from_pretrained error; same family but different bug." + }, + { + "left": "issue:45230", + "right": "issue:45397", + "accept": false, + "reason": "Generic bug report placeholder vs gemma-4 ZeRO3 loading issue." + } + ] + }, + "evaluator_result": null, + "evaluator_used": false, + "retried": false, + "accepted_nontrivial_soft_edge": false, + "error_kind": null, + "error_message": null + }, + { + "label": "issue", + "component_index": 1, + "component_count": 28, + "review_unit_index": 49, + "review_unit_count": 65, + "status": "reviewed", + "reason": null, + "source": "llm", + "cache_hit": false, + "model": "gpt-5.4-mini?service_tier=flex", + "cluster_id": "cluster-10000-530", + "nodes": [ + "issue:30333", + "issue:30990", + "issue:33290", + "issue:35141", + "issue:36010", + "issue:41628", + "issue:42175", + "issue:42371", + "issue:42831", + "issue:42898", + "issue:43232", + "issue:43278", + "issue:43673", + "issue:43784", + "issue:43824", + "issue:43873", + "issue:43976", + "issue:44038", + "issue:44222", + "issue:44351", + "issue:44496", + "issue:44568", + "issue:44623", + "issue:44661", + "issue:44743", + "issue:44792", + "issue:44857", + "issue:44869", + "issue:44933", + "issue:44945", + "issue:45081", + "issue:45092", + "issue:45103", + "issue:45137", + "issue:45310", + "issue:45356", + "issue:45375", + "issue:45405", + "issue:45406", + "issue:45561", + "issue:45593", + "issue:45712", + "issue:45753", + "issue:46202", + "issue:46701", + "issue:46877", + "issue:47059", + "issue:47068" + ], + "soft_pairs": [ + "issue:44869|issue:45356", + "issue:44038|issue:44623", + "issue:43232|issue:43673", + "issue:45310|issue:45375", + "issue:41628|issue:44933", + "issue:36010|issue:41628", + "issue:44661|issue:47059", + "issue:43873|issue:44945", + "issue:43278|issue:45137", + "issue:42371|issue:45092", + "issue:45081|issue:45103", + "issue:44857|issue:46877", + "issue:45406|issue:46202", + "issue:46202|issue:46701", + "issue:45593|issue:46877", + "issue:44038|issue:45310", + "issue:43232|issue:44792", + "issue:41628|issue:44351", + "issue:42175|issue:44933", + "issue:44222|issue:45310", + "issue:30333|issue:33290", + "issue:44945|issue:45753", + "issue:45405|issue:46877", + "issue:45561|issue:47059", + "issue:44743|issue:47068", + "issue:42898|issue:44568", + "issue:42831|issue:44222", + "issue:43784|issue:43824", + "issue:42175|issue:45712", + "issue:43976|issue:44933", + "issue:44496|issue:45561", + "issue:30990|issue:35141" + ], + "prepared_review_unit_hash": "9b54c4fc9971f82ce906fd5cdae6c31c6f7b3c8738339ca6743377d345b96e3e", + "component_budget": { + "node_count": 530, + "item_count": 530, + "soft_pair_count": 1954, + "serialized_chars": 1013004, + "estimated_input_tokens": 253251, + "estimated_eval_tokens": 506758 + }, + "budget": { + "node_count": 48, + "item_count": 48, + "soft_pair_count": 32, + "serialized_chars": 37293, + "estimated_input_tokens": 9324, + "estimated_eval_tokens": 18904 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": true, + "analyst_result": { + "summary": "This cluster is not a duplicate set; it contains many unrelated bug reports that were only weakly text-similar. I would not merge any of the proposed soft edges.", + "confidence": 0.93, + "canonical_issue_reason": "If one issue must stand in for the cluster, issue:43873 is the most central and active: it is open, has the highest discussion/references, and is a broad transformer runtime bug rather than a narrowly model-specific report.", + "canonical_pr_reason": null, + "best_issue_reason": "issue:43873 is the best representative issue because it is active, broadly scoped, and has the strongest engagement signal in this set.", + "best_pr_reason": null, + "soft_edge_verdicts": [ + { + "left": "issue:44869", + "right": "issue:45356", + "accept": false, + "reason": "Both involve tokenization/codec behavior, but one is a Whisper timestamp crash and the other is a Kimi tokenizer regression with fix_mistral_regex; different code paths and symptoms." + }, + { + "left": "issue:44038", + "right": "issue:44623", + "accept": false, + "reason": "Qwen3-VL-Moe loading bug vs processor.save_pretrained missing files; same area, but not the same underlying defect." + }, + { + "left": "issue:43232", + "right": "issue:43673", + "accept": false, + "reason": "Both are generation/cache-related, but one concerns _update_model_kwargs_for_generation after sync_gpus and the other cache absence in chunked_prefill; distinct bugs." + }, + { + "left": "issue:45310", + "right": "issue:45375", + "accept": false, + "reason": "Both mention Qwen3.5 MoE, but one is a from_pretrained failure and the other is a missing config field in a vision config; not the same issue." + }, + { + "left": "issue:41628", + "right": "issue:44933", + "accept": false, + "reason": "Both are import errors, but they concern different symbols and modules; no evidence of the same missing export." + }, + { + "left": "issue:36010", + "right": "issue:41628", + "accept": false, + "reason": "GenerationMixin import failure and AutoImageProcessor import failure are separate API breakages." + }, + { + "left": "issue:44661", + "right": "issue:47059", + "accept": false, + "reason": "Model-registration logic in add-new-model-like is unrelated to a CLI completion regression from huggingface_hub." + }, + { + "left": "issue:43873", + "right": "issue:44945", + "accept": false, + "reason": "Offloading/quantization behavior and incorrect pipeline-parallel outputs are different distributed inference problems." + }, + { + "left": "issue:43278", + "right": "issue:45137", + "accept": false, + "reason": "Embedding dtype drift in eval is unrelated to a DeepSpeed ZeRO3 deque IndexError." + }, + { + "left": "issue:42371", + "right": "issue:45092", + "accept": false, + "reason": "TF32 API guidance and InternVL2 meta-init compatibility are unrelated." + }, + { + "left": "issue:45081", + "right": "issue:45103", + "accept": false, + "reason": "Both crash with AttributeError, but one is Mistral tokenizer patching and the other is auto_docstring kwargs parsing; different causes." + }, + { + "left": "issue:44857", + "right": "issue:46877", + "accept": false, + "reason": "Training loss AMP crash and docstring import-time logging are unrelated." + }, + { + "left": "issue:45406", + "right": "issue:46202", + "accept": false, + "reason": "Both involve Gemma, but one is a processor attribute crash in transformers serve and the other is a tokenizer roundtrip failure." + }, + { + "left": "issue:46202", + "right": "issue:46701", + "accept": false, + "reason": "Gemma pipeline roundtrip failure and DiffusionGemma image-text-to-text pipeline failure are different concrete bugs." + }, + { + "left": "issue:45593", + "right": "issue:46877", + "accept": false, + "reason": "D-FINE auxiliary-loss behavior is unrelated to auto_docstring stdout logging." + }, + { + "left": "issue:44038", + "right": "issue:45310", + "accept": false, + "reason": "Related Qwen model family, but different models and failure modes; insufficient evidence they are the same bug." + }, + { + "left": "issue:41628", + "right": "issue:44351", + "accept": false, + "reason": "Both are import-name issues, but they target different classes and likely different missing exports." + }, + { + "left": "issue:44222", + "right": "issue:45310", + "accept": false, + "reason": "FP8 save_pretrained behavior is distinct from a from_pretrained model-loading error." + }, + { + "left": "issue:30333", + "right": "issue:33290", + "accept": false, + "reason": "MLflow reporting on failed training jobs is unrelated to DeepSpeed OOM with Adafactor." + }, + { + "left": "issue:44945", + "right": "issue:45753", + "accept": false, + "reason": "Pipeline-parallel output corruption is unrelated to a Qwen3_5 image infinite loop." + }, + { + "left": "issue:45405", + "right": "issue:46877", + "accept": false, + "reason": "PEFT version bump and docstring import-time diagnostics are unrelated." + }, + { + "left": "issue:45561", + "right": "issue:47059", + "accept": false, + "reason": "Pytest-xdist file race and CLI completion breakage are different testing/packaging issues." + }, + { + "left": "issue:44743", + "right": "issue:47068", + "accept": false, + "reason": "Qwen3_5 recurrent-state reset and DeepSpeed loss synchronization are unrelated runtime bugs." + }, + { + "left": "issue:42898", + "right": "issue:44568", + "accept": false, + "reason": "Both are tokenizer regressions, but one is clean-up spacing and the other is add_special_tokens/BOS/EOS behavior; not the same defect." + }, + { + "left": "issue:42831", + "right": "issue:44222", + "accept": false, + "reason": "Both mention FP8, but one is accuracy loss and the other is save_pretrained behavior for MoE; different problems." + }, + { + "left": "issue:43784", + "right": "issue:43824", + "accept": false, + "reason": "Sentence-transformers import failure and Qwen2.5-VL import failure are unrelated." + }, + { + "left": "issue:42175", + "right": "issue:45712", + "accept": false, + "reason": "One is missing TensorFlow in a torch extra install, the other is dummy PT object cleanup; not the same underlying issue." + }, + { + "left": "issue:43976", + "right": "issue:44933", + "accept": false, + "reason": "Python version compatibility and image_utils import cleanup are unrelated." + }, + { + "left": "issue:44496", + "right": "issue:45561", + "accept": false, + "reason": "Model config recognition and xdist race conditions are unrelated." + }, + { + "left": "issue:30990", + "right": "issue:35141", + "accept": false, + "reason": "Sentence-Transformers loading hang and embedding reinitialization after resize are different bugs." + } + ] + }, + "evaluator_result": null, + "evaluator_used": false, + "retried": false, + "accepted_nontrivial_soft_edge": false, + "error_kind": null, + "error_message": null + }, + { + "label": "issue", + "component_index": 1, + "component_count": 28, + "review_unit_index": 50, + "review_unit_count": 65, + "status": "reviewed", + "reason": null, + "source": "llm", + "cache_hit": false, + "model": "gpt-5.4-mini?service_tier=flex", + "cluster_id": "cluster-10000-530", + "nodes": [ + "issue:22355", + "issue:27957", + "issue:28282", + "issue:30333", + "issue:34689", + "issue:36331", + "issue:39692", + "issue:42222", + "issue:42673", + "issue:42796", + "issue:42898", + "issue:43116", + "issue:43295", + "issue:43479", + "issue:43704", + "issue:43723", + "issue:43874", + "issue:44016", + "issue:44060", + "issue:44075", + "issue:44079", + "issue:44186", + "issue:44190", + "issue:44295", + "issue:44448", + "issue:44568", + "issue:44623", + "issue:44734", + "issue:44779", + "issue:44843", + "issue:44869", + "issue:45092", + "issue:45103", + "issue:45125", + "issue:45200", + "issue:45310", + "issue:45600", + "issue:45656", + "issue:45812", + "issue:46077", + "issue:46123", + "issue:46231", + "issue:46693", + "issue:46829", + "issue:46909", + "issue:47059", + "issue:47085" + ], + "soft_pairs": [ + "issue:44568|issue:47085", + "issue:44843|issue:47085", + "issue:42898|issue:45812", + "issue:42673|issue:43704", + "issue:44186|issue:47085", + "issue:43479|issue:46829", + "issue:42673|issue:46693", + "issue:44623|issue:45310", + "issue:39692|issue:46909", + "issue:43723|issue:44568", + "issue:44060|issue:45125", + "issue:45600|issue:47059", + "issue:30333|issue:34689", + "issue:44016|issue:46231", + "issue:27957|issue:30333", + "issue:43479|issue:44079", + "issue:30333|issue:36331", + "issue:42898|issue:44779", + "issue:42673|issue:47085", + "issue:42898|issue:43295", + "issue:43479|issue:44869", + "issue:42898|issue:44448", + "issue:46077|issue:46123", + "issue:22355|issue:28282", + "issue:44295|issue:45656", + "issue:42222|issue:45200", + "issue:44734|issue:45103", + "issue:43874|issue:45092", + "issue:42796|issue:43723", + "issue:43116|issue:44190", + "issue:44075|issue:46231" + ], + "prepared_review_unit_hash": "92efe59ad41336342aba24b36c48777d641ea9a9b29b97d9ad49c722c9e1847d", + "component_budget": { + "node_count": 530, + "item_count": 530, + "soft_pair_count": 1954, + "serialized_chars": 1013004, + "estimated_input_tokens": 253251, + "estimated_eval_tokens": 506758 + }, + "budget": { + "node_count": 47, + "item_count": 47, + "soft_pair_count": 31, + "serialized_chars": 36613, + "estimated_input_tokens": 9154, + "estimated_eval_tokens": 18564 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": true, + "analyst_result": { + "summary": "Only one soft edge is a true duplicate: the two Qwen3ForCausalLM VRAM-leak reports. The rest are same-subsystem or same-version regressions at a broad level, but they describe different concrete bugs and should stay separate.", + "confidence": 0.96, + "canonical_issue_reason": "issue:42673 is the strongest canonical candidate: it is the earliest, most detailed report of the exact Qwen3ForCausalLM multi-dataloader VRAM leak, and issue:43704 is essentially the same report.", + "canonical_pr_reason": null, + "best_issue_reason": "issue:42673 is the best global representative because it captures the concrete bug most completely and is duplicated by issue:43704; it is the cleanest anchor for triage.", + "best_pr_reason": null, + "soft_edge_verdicts": [ + { + "left": "issue:44568", + "right": "issue:47085", + "accept": false, + "reason": "Both mention tokenizer behavior, but one is a special-token regression for mDeBERTa and the other is a thread-safety regression in fast tokenizers; different bugs." + }, + { + "left": "issue:44843", + "right": "issue:47085", + "accept": false, + "reason": "Offline HF hub patching in AutoTokenizer is unrelated to tokenizer thread safety." + }, + { + "left": "issue:42898", + "right": "issue:45812", + "accept": false, + "reason": "Both are v5 tokenizer regressions, but one is clean_up_tokenization_spaces behavior and the other is Granite token ID generation; not the same bug." + }, + { + "left": "issue:42673", + "right": "issue:43704", + "accept": true, + "reason": "Same title and same concrete bug: Qwen3ForCausalLM VRAM leak when used in multiple dataloader threads." + }, + { + "left": "issue:44186", + "right": "issue:47085", + "accept": false, + "reason": "LayoutLMv2Tokenizer crash on NER/padding is unrelated to tokenizer thread-safety." + }, + { + "left": "issue:43479", + "right": "issue:46829", + "accept": false, + "reason": "Different model families and different failure modes: config initialization vs output field misassignment." + }, + { + "left": "issue:42673", + "right": "issue:46693", + "accept": false, + "reason": "One is a VRAM leak in multithreaded dataloaders; the other is a FA2 long-context performance regression." + }, + { + "left": "issue:44623", + "right": "issue:45310", + "accept": false, + "reason": "Missing processor save files and Qwen3.5 Moe from_pretrained failure are different code paths." + }, + { + "left": "issue:39692", + "right": "issue:46909", + "accept": false, + "reason": "Both are documentation-example bugs, but they involve different models/examples and different error classes." + }, + { + "left": "issue:43723", + "right": "issue:44568", + "accept": false, + "reason": "Both are tokenizer v5 issues, but one is AutoTokenizer loading and the other is add_special_tokens behavior for a specific model." + }, + { + "left": "issue:44060", + "right": "issue:45125", + "accept": false, + "reason": "Different Qwen3-Next/Qwen3.5-MoE model internals: tied-weights warning vs missing tensor-parallel plan." + }, + { + "left": "issue:45600", + "right": "issue:47059", + "accept": false, + "reason": "Broken auto-mapping for removed configs and a CLI completion API break are unrelated." + }, + { + "left": "issue:30333", + "right": "issue:34689", + "accept": false, + "reason": "MLflow reporting on failed training is unrelated to model loading regression for Llama 3.2 Vision Instruct." + }, + { + "left": "issue:44016", + "right": "issue:46231", + "accept": false, + "reason": "Notebook syntax error and a docs parameter-name mismatch are different documentation issues." + }, + { + "left": "issue:27957", + "right": "issue:30333", + "accept": false, + "reason": "Flash Attention 2 behavior for XLM-R is unrelated to MLflow reporting after job failure." + }, + { + "left": "issue:43479", + "right": "issue:44079", + "accept": false, + "reason": "Config default handling and ModelOutput key assignment are different bugs." + }, + { + "left": "issue:30333", + "right": "issue:36331", + "accept": false, + "reason": "Training failure reporting vs CustomTrainer compute_loss signature mismatch are different trainer issues." + }, + { + "left": "issue:42898", + "right": "issue:44779", + "accept": false, + "reason": "Both are tokenizer regressions in v5, but they affect different models and different tokenization behaviors." + }, + { + "left": "issue:42673", + "right": "issue:47085", + "accept": false, + "reason": "The titles are identical in theme, but without concrete shared symptoms or code path evidence, treat as separate reports rather than a confirmed duplicate." + }, + { + "left": "issue:42898", + "right": "issue:43295", + "accept": false, + "reason": "Different regressions: clean_up_tokenization_spaces vs processor.tokenizer/image-passing compatibility." + }, + { + "left": "issue:43479", + "right": "issue:44869", + "accept": false, + "reason": "Phi4 multimodal config initialization and Whisper timestamp decoding are unrelated." + }, + { + "left": "issue:42898", + "right": "issue:44448", + "accept": false, + "reason": "Both are v4-to-v5 output regressions, but on different models and behaviors; not one bug." + }, + { + "left": "issue:46077", + "right": "issue:46123", + "accept": false, + "reason": "Gemma attention-mask construction and last-partial-batch handling are unrelated pipeline/model bugs." + }, + { + "left": "issue:22355", + "right": "issue:28282", + "accept": false, + "reason": "Importing transformers.onnx and missing PyTorch for AutoModel are different dependency/import failures." + }, + { + "left": "issue:44295", + "right": "issue:45656", + "accept": false, + "reason": "Position_ids buffer access and duplicate optimizer stepping with deepspeed are unrelated." + }, + { + "left": "issue:42222", + "right": "issue:45200", + "accept": false, + "reason": "Both involve multimodal models, but one is broken vitpose loading and the other is Gemma 4 text-only mm_token_type_ids behavior." + }, + { + "left": "issue:44734", + "right": "issue:45103", + "accept": false, + "reason": "KV-cache continuation crash in serve and auto_docstring annotation handling are unrelated." + }, + { + "left": "issue:43874", + "right": "issue:45092", + "accept": false, + "reason": "Missing image patch count API and remote-code meta-init incompatibility are different multimodal issues." + }, + { + "left": "issue:42796", + "right": "issue:43723", + "accept": false, + "reason": "Tokenizer BPE-marker handling and AutoTokenizer loading failure are not the same defect." + }, + { + "left": "issue:43116", + "right": "issue:44190", + "accept": false, + "reason": "Two example-script bugs, but in different scripts and with different failure modes." + }, + { + "left": "issue:44075", + "right": "issue:46231", + "accept": false, + "reason": "Optimizer argument handling is unrelated to a documentation parameter-name typo." + } + ] + }, + "evaluator_result": { + "accept": true, + "feedback": "The summary is grounded in the packet, and the one accepted soft edge (issue:42673 \u2194 issue:43704) is supported by identical titles and closely aligned metadata. The remaining soft-edge rejections are conservative and do not overclaim duplication from broad subsystem similarity alone. No obvious overstatement or mergeability problem." + }, + "evaluator_used": true, + "retried": false, + "accepted_nontrivial_soft_edge": true, + "error_kind": null, + "error_message": null + }, + { + "label": "issue", + "component_index": 1, + "component_count": 28, + "review_unit_index": 51, + "review_unit_count": 65, + "status": "reviewed", + "reason": null, + "source": "llm", + "cache_hit": false, + "model": "gpt-5.4-mini?service_tier=flex", + "cluster_id": "cluster-10000-530", + "nodes": [ + "issue:30990", + "issue:38175", + "issue:41628", + "issue:42757", + "issue:42831", + "issue:42907", + "issue:43122", + "issue:43232", + "issue:43475", + "issue:43479", + "issue:43504", + "issue:43644", + "issue:43723", + "issue:44190", + "issue:44222", + "issue:44263", + "issue:44295", + "issue:44297", + "issue:44327", + "issue:44360", + "issue:44373", + "issue:44448", + "issue:44484", + "issue:44485", + "issue:44623", + "issue:44933", + "issue:44977", + "issue:45071", + "issue:45092", + "issue:45245", + "issue:45310", + "issue:45405", + "issue:45656", + "issue:45712", + "issue:45834", + "issue:46050", + "issue:46082", + "issue:46132", + "issue:46209", + "issue:46305", + "issue:46369", + "issue:46431", + "issue:46614", + "issue:46762", + "issue:46909", + "issue:46932", + "issue:47059", + "issue:47276" + ], + "soft_pairs": [ + "issue:45405|issue:45712", + "issue:44222|issue:44297", + "issue:46209|issue:46431", + "issue:44448|issue:46132", + "issue:44297|issue:45656", + "issue:42831|issue:44623", + "issue:43479|issue:43644", + "issue:45245|issue:46132", + "issue:44222|issue:44623", + "issue:44933|issue:45712", + "issue:44373|issue:44485", + "issue:43723|issue:44977", + "issue:44297|issue:45310", + "issue:43122|issue:46209", + "issue:41628|issue:42757", + "issue:43232|issue:47276", + "issue:43504|issue:47059", + "issue:46614|issue:46932", + "issue:45071|issue:46082", + "issue:43475|issue:46305", + "issue:43232|issue:45834", + "issue:30990|issue:38175", + "issue:44222|issue:44484", + "issue:44295|issue:44623", + "issue:44360|issue:46762", + "issue:44327|issue:47059", + "issue:44263|issue:45092", + "issue:46369|issue:46909", + "issue:44190|issue:46082", + "issue:42907|issue:46050" + ], + "prepared_review_unit_hash": "8326779039b4f6a68776d62814f456abbc21c10597d7d681514fae6a8213de2a", + "component_budget": { + "node_count": 530, + "item_count": 530, + "soft_pair_count": 1954, + "serialized_chars": 1013004, + "estimated_input_tokens": 253251, + "estimated_eval_tokens": 506758 + }, + "budget": { + "node_count": 48, + "item_count": 48, + "soft_pair_count": 30, + "serialized_chars": 36297, + "estimated_input_tokens": 9075, + "estimated_eval_tokens": 18406 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": true, + "analyst_result": { + "summary": "This cluster is mostly a grab-bag of unrelated Transformers issues. The soft-similarity pairs share broad keywords like save_pretrained, from_pretrained, config, or generation, but they do not describe the same concrete bug/change, so none should be merged as duplicates.", + "confidence": 0.83, + "canonical_issue_reason": "issue:44222 is the most representative concrete bug in the set, centered on broken FP8 save_pretrained behavior and close to several other serialization-related issues, even though the nearby soft matches are not true duplicates.", + "canonical_pr_reason": null, + "best_issue_reason": "issue:44222 is the best single issue to anchor this cluster because it is specific, reproducible, and sits near the strongest serialization-related similarity edges.", + "best_pr_reason": null, + "soft_edge_verdicts": [ + { + "left": "issue:45405", + "right": "issue:45712", + "accept": false, + "reason": "One is a PEFT version bump complaint; the other is about leftover dummy classes leaking into the API. Different problems." + }, + { + "left": "issue:44222", + "right": "issue:44297", + "accept": false, + "reason": "Both mention save_pretrained, but one is FP8 MoE serialization and the other is tokenizer_class metadata mismatch. Not the same bug." + }, + { + "left": "issue:46209", + "right": "issue:46431", + "accept": false, + "reason": "FP8 inference failure on B200 and missing chat template are unrelated code paths." + }, + { + "left": "issue:44448", + "right": "issue:46132", + "accept": false, + "reason": "Pegasus output drift and AttentionInterface.register behavior are different model/runtime issues." + }, + { + "left": "issue:44297", + "right": "issue:45656", + "accept": false, + "reason": "Tokenizer save metadata mismatch is unrelated to deepspeed causing optimizer.step to run twice." + }, + { + "left": "issue:42831", + "right": "issue:44623", + "accept": false, + "reason": "FP8 accuracy regression and processor.save_pretrained missing files are different serialization/runtime problems." + }, + { + "left": "issue:43479", + "right": "issue:43644", + "accept": false, + "reason": "Config default initialization for multimodal vision/audio and non-persistent buffer junk are not the same defect." + }, + { + "left": "issue:45245", + "right": "issue:46132", + "accept": false, + "reason": "Category cardinality limit and attention registration semantics are unrelated." + }, + { + "left": "issue:44222", + "right": "issue:44623", + "accept": false, + "reason": "Both involve save_pretrained, but one is FP8 MoE model saving and the other is processor artifact omission. Different failure modes." + }, + { + "left": "issue:44933", + "right": "issue:45712", + "accept": false, + "reason": "Missing import from image_utils and leftover dummy classes are separate API cleanup issues." + }, + { + "left": "issue:44373", + "right": "issue:44485", + "accept": false, + "reason": "A docstring mismatch for position_ids is not the same as a RoPE implementation bug." + }, + { + "left": "issue:43723", + "right": "issue:44977", + "accept": false, + "reason": "Tokenizer loading failure and flash-attention generation failure are unrelated." + }, + { + "left": "issue:44297", + "right": "issue:45310", + "accept": false, + "reason": "Tokenizer save metadata and Qwen3.5 MoE from_pretrained error are different problems." + }, + { + "left": "issue:43122", + "right": "issue:46209", + "accept": false, + "reason": "Tokenization differences across versions and FP8 inference on B200 do not share the same underlying bug." + }, + { + "left": "issue:41628", + "right": "issue:42757", + "accept": false, + "reason": "Missing AutoImageProcessor and missing is_offline_mode are separate import/export regressions." + }, + { + "left": "issue:43232", + "right": "issue:47276", + "accept": false, + "reason": "One is a generation state bug after sync_gpus; the other is only a misleading warning message." + }, + { + "left": "issue:43504", + "right": "issue:47059", + "accept": false, + "reason": "Legacy field handling in BEiT loading is unrelated to a huggingface_hub CLI incompatibility." + }, + { + "left": "issue:46614", + "right": "issue:46932", + "accept": false, + "reason": "Stale phonemizer backend state and Gemma4 quantized dtype casting are unrelated." + }, + { + "left": "issue:45071", + "right": "issue:46082", + "accept": false, + "reason": "Both concern config validation, but type-checking breakage and head_dim rejection are not the same concrete issue." + }, + { + "left": "issue:43475", + "right": "issue:46305", + "accept": false, + "reason": "Missing fpn_position_embeddings and a sliding-window off-by-one are different model bugs." + }, + { + "left": "issue:43232", + "right": "issue:45834", + "accept": false, + "reason": "Generation kwargs update after sync_gpus and long-ocr index overflow are unrelated." + }, + { + "left": "issue:30990", + "right": "issue:38175", + "accept": false, + "reason": "Sentence Transformers loading hang and zero-probabilities from SigLIP2 are different symptom sets and code paths." + }, + { + "left": "issue:44222", + "right": "issue:44484", + "accept": false, + "reason": "Both touch saving, but FP8 MoE serialization and the 50GB shard-size question are not the same bug." + }, + { + "left": "issue:44295", + "right": "issue:44623", + "accept": false, + "reason": "Reading a registered buffer and processor save_pretrained missing files are not the same issue." + }, + { + "left": "issue:44360", + "right": "issue:46762", + "accept": false, + "reason": "A DSA indexer activation concern and Minimax attention numerical drift are distinct implementation problems." + }, + { + "left": "issue:44327", + "right": "issue:47059", + "accept": false, + "reason": "Top-k decode crash and CLI completion attribute breakage are unrelated." + }, + { + "left": "issue:44263", + "right": "issue:45092", + "accept": false, + "reason": "torch.split indexing in GlmMoeDsaIndexer and InternVL2 meta-init incompatibility are different code paths." + }, + { + "left": "issue:46369", + "right": "issue:46909", + "accept": false, + "reason": "SAM3 LiteText memory overflow and a broken documentation example do not match." + }, + { + "left": "issue:44190", + "right": "issue:46082", + "accept": false, + "reason": "Local dataset loading in a training script and LlamaConfig head_dim validation are unrelated." + }, + { + "left": "issue:42907", + "right": "issue:46050", + "accept": false, + "reason": "Both involve save/load after quantization, but dequantized Ministrals/Devstrals saving and sinq round-trip breakage are different concrete bugs." + } + ] + }, + "evaluator_result": null, + "evaluator_used": false, + "retried": false, + "accepted_nontrivial_soft_edge": false, + "error_kind": null, + "error_message": null + }, + { + "label": "issue", + "component_index": 1, + "component_count": 28, + "review_unit_index": 52, + "review_unit_count": 65, + "status": "reviewed", + "reason": null, + "source": "llm", + "cache_hit": false, + "model": "gpt-5.4-mini?service_tier=flex", + "cluster_id": "cluster-10000-530", + "nodes": [ + "issue:22355", + "issue:27957", + "issue:28218", + "issue:39401", + "issue:41628", + "issue:42796", + "issue:42831", + "issue:43122", + "issue:43232", + "issue:43408", + "issue:43540", + "issue:43576", + "issue:43644", + "issue:43756", + "issue:43784", + "issue:43824", + "issue:43986", + "issue:44038", + "issue:44060", + "issue:44188", + "issue:44261", + "issue:44263", + "issue:44279", + "issue:44360", + "issue:44373", + "issue:44483", + "issue:44568", + "issue:44829", + "issue:44964", + "issue:45092", + "issue:45310", + "issue:45357", + "issue:45381", + "issue:45406", + "issue:45600", + "issue:45753", + "issue:46002", + "issue:46305", + "issue:46328", + "issue:46369", + "issue:46402", + "issue:46424", + "issue:46655", + "issue:46726", + "issue:46909", + "issue:47059", + "issue:47217" + ], + "soft_pairs": [ + "issue:45092|issue:45357", + "issue:43232|issue:46424", + "issue:22355|issue:27957", + "issue:43986|issue:46726", + "issue:44263|issue:44373", + "issue:44279|issue:46328", + "issue:43408|issue:46402", + "issue:46305|issue:46655", + "issue:45381|issue:46305", + "issue:42831|issue:45310", + "issue:41628|issue:43824", + "issue:22355|issue:28218", + "issue:44483|issue:45406", + "issue:44263|issue:44360", + "issue:44964|issue:47059", + "issue:42831|issue:44038", + "issue:44829|issue:46002", + "issue:44261|issue:45092", + "issue:43784|issue:45600", + "issue:44188|issue:44373", + "issue:39401|issue:44568", + "issue:42796|issue:43122", + "issue:43644|issue:46305", + "issue:43576|issue:43784", + "issue:43986|issue:47217", + "issue:43756|issue:44060", + "issue:42831|issue:45753", + "issue:43408|issue:46369", + "issue:43540|issue:46909", + "issue:45381|issue:46909" + ], + "prepared_review_unit_hash": "f896a617cadf3148fbf65defa3075559eef35a72dab7d65a633f9d2a039fa4c5", + "component_budget": { + "node_count": 530, + "item_count": 530, + "soft_pair_count": 1954, + "serialized_chars": 1013004, + "estimated_input_tokens": 253251, + "estimated_eval_tokens": 506758 + }, + "budget": { + "node_count": 47, + "item_count": 47, + "soft_pair_count": 30, + "serialized_chars": 35929, + "estimated_input_tokens": 8983, + "estimated_eval_tokens": 18222 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": true, + "analyst_result": { + "summary": "This cluster is heterogeneous: it groups many unrelated Transformers issues across tokenization, vision-language loading, generation, CLI, and model-specific regressions. None of the soft pairs appear to be true duplicates of the same underlying bug or change.", + "confidence": 0.97, + "canonical_issue_reason": null, + "canonical_pr_reason": null, + "best_issue_reason": null, + "best_pr_reason": null, + "soft_edge_verdicts": [ + { + "left": "issue:45092", + "right": "issue:45357", + "accept": false, + "reason": "Both concern Qwen-related multimodal models, but one is a v5 meta-init incompatibility and the other is incorrect visual encoder keys on save_pretrained; different code paths and fixes." + }, + { + "left": "issue:43232", + "right": "issue:46424", + "accept": false, + "reason": "One is a generation kwargs sync bug after sync_gpus, the other is cache max length only being honored for quantized cache; same area, different defect." + }, + { + "left": "issue:22355", + "right": "issue:27957", + "accept": false, + "reason": "Unrelated issues: missing transformers.onnx module import vs XLMRoberta Flash Attention 2 support." + }, + { + "left": "issue:43986", + "right": "issue:46726", + "accept": false, + "reason": "Both involve processor/backend handling, but one is a video-model crash without torchvision and the other is an unrecognized image processor when a vision backend is missing." + }, + { + "left": "issue:44263", + "right": "issue:44373", + "accept": false, + "reason": "Different bugs: a torch.split/indexer implementation problem vs a wrong docstring for position_ids." + }, + { + "left": "issue:44279", + "right": "issue:46328", + "accept": false, + "reason": "A generic dependency issue is not the same as a SequenceFeatureExtractor.pad performance regression." + }, + { + "left": "issue:43408", + "right": "issue:46402", + "accept": false, + "reason": "One is a model-type warning during instantiation, the other is missing text-encoder weights in Sam3Model." + }, + { + "left": "issue:46305", + "right": "issue:46655", + "accept": false, + "reason": "Both are video-processing bugs, but one is a sliding-window mask off-by-one and the other is incorrect padding calculation." + }, + { + "left": "issue:45381", + "right": "issue:46305", + "accept": false, + "reason": "Qwen2.5-VL vision_position_ids bug vs MoonshineStreaming mask bug; different models and logic." + }, + { + "left": "issue:42831", + "right": "issue:45310", + "accept": false, + "reason": "FineGrainedFP8 accuracy issue is unrelated to Qwen3.5 MoE from_pretrained loading failure." + }, + { + "left": "issue:41628", + "right": "issue:43824", + "accept": false, + "reason": "Both are import errors, but for different symbols and likely different missing exposure paths." + }, + { + "left": "issue:22355", + "right": "issue:28218", + "accept": false, + "reason": "Different tokenizer/import problems: ONNX module import vs extra space after added token." + }, + { + "left": "issue:44483", + "right": "issue:45406", + "accept": false, + "reason": "A chat-completions request rejection is not the same as a Gemma4Processor crash in transformers serve." + }, + { + "left": "issue:44263", + "right": "issue:44360", + "accept": false, + "reason": "Same subsystem, but distinct bugs: split return values in GlmMoeDsaIndexer vs a missing ReLU in the DSA indexer." + }, + { + "left": "issue:44964", + "right": "issue:47059", + "accept": false, + "reason": "Phi-4 multimodal model loading failure is unrelated to the huggingface_hub CLI completion breakage." + }, + { + "left": "issue:42831", + "right": "issue:44038", + "accept": false, + "reason": "An FP8 accuracy regression is not the same as a Qwen3-VL-Moe bug." + }, + { + "left": "issue:44829", + "right": "issue:46002", + "accept": false, + "reason": "Flash Attention 3 causes degenerate training, while the other is NaNs after checkout/init; different symptoms and likely causes." + }, + { + "left": "issue:44261", + "right": "issue:45092", + "accept": false, + "reason": "Layernorm eps precision mismatch and remote-code/meta-init incompatibility are unrelated defects." + }, + { + "left": "issue:43784", + "right": "issue:45600", + "accept": false, + "reason": "Importing sentence-transformers fails with NameError in one case; the other is stale auto_mappings references breaking CI." + }, + { + "left": "issue:44188", + "right": "issue:44373", + "accept": false, + "reason": "torch.compile attention-kernel divergence is unrelated to a wrong docstring." + }, + { + "left": "issue:39401", + "right": "issue:44568", + "accept": false, + "reason": "Both are tokenizer regressions, but one is offset_mapping and the other is add_special_tokens not inserting BOS/EOS." + }, + { + "left": "issue:42796", + "right": "issue:43122", + "accept": false, + "reason": "The first is a specific BPE-marker handling bug for Ministral3, while the second is a broader version-to-version tokenization regression." + }, + { + "left": "issue:43644", + "right": "issue:46305", + "accept": false, + "reason": "Non-persistent buffer initialization junk is unrelated to a video sliding-window mask off-by-one." + }, + { + "left": "issue:43576", + "right": "issue:43784", + "accept": false, + "reason": "A broken transformers env command and a sentence-transformers import NameError are different failure modes." + }, + { + "left": "issue:43986", + "right": "issue:47217", + "accept": false, + "reason": "Both are VL processor/input issues, but one is missing torchvision handling and the other is failure when the image placeholder appears in user text." + }, + { + "left": "issue:43756", + "right": "issue:44060", + "accept": false, + "reason": "Smollm3 RoPE layer loss and Qwen3-Next tied-weights warning are different model bugs." + }, + { + "left": "issue:43540", + "right": "issue:46909", + "accept": false, + "reason": "A Qwen3OmniMoe video-input ValueError is unrelated to a broken documentation example." + }, + { + "left": "issue:45381", + "right": "issue:46909", + "accept": false, + "reason": "Model video-position-id bug vs documentation bug; no shared underlying change." + }, + { + "left": "issue:42831", + "right": "issue:45753", + "accept": false, + "reason": "FP8 accuracy regression and Qwen3.5 infinite-loop-on-image are different defects." + } + ] + }, + "evaluator_result": null, + "evaluator_used": false, + "retried": false, + "accepted_nontrivial_soft_edge": false, + "error_kind": null, + "error_message": null + }, + { + "label": "issue", + "component_index": 1, + "component_count": 28, + "review_unit_index": 53, + "review_unit_count": 65, + "status": "reviewed", + "reason": null, + "source": "llm", + "cache_hit": false, + "model": "gpt-5.4-mini?service_tier=flex", + "cluster_id": "cluster-10000-530", + "nodes": [ + "issue:22355", + "issue:27957", + "issue:29159", + "issue:30990", + "issue:33357", + "issue:39379", + "issue:39401", + "issue:41628", + "issue:42371", + "issue:42757", + "issue:42886", + "issue:42907", + "issue:42994", + "issue:43066", + "issue:43116", + "issue:43122", + "issue:43329", + "issue:43408", + "issue:43452", + "issue:43475", + "issue:43646", + "issue:43824", + "issue:43937", + "issue:43976", + "issue:44038", + "issue:44117", + "issue:44263", + "issue:44295", + "issue:44297", + "issue:44351", + "issue:44360", + "issue:44485", + "issue:44488", + "issue:44521", + "issue:44625", + "issue:44704", + "issue:44829", + "issue:44945", + "issue:44964", + "issue:45200", + "issue:45216", + "issue:45278", + "issue:45290", + "issue:45478", + "issue:46002", + "issue:46172", + "issue:46326", + "issue:47059" + ], + "soft_pairs": [ + "issue:42886|issue:43066", + "issue:42886|issue:42994", + "issue:43646|issue:46002", + "issue:39379|issue:46172", + "issue:42757|issue:45278", + "issue:27957|issue:30990", + "issue:30990|issue:33357", + "issue:43122|issue:46326", + "issue:43329|issue:47059", + "issue:44295|issue:44297", + "issue:44625|issue:44829", + "issue:39401|issue:44488", + "issue:42907|issue:44964", + "issue:39401|issue:44521", + "issue:43976|issue:44945", + "issue:42907|issue:44945", + "issue:41628|issue:45278", + "issue:44117|issue:44704", + "issue:22355|issue:29159", + "issue:43122|issue:44521", + "issue:44704|issue:45290", + "issue:42371|issue:44263", + "issue:43408|issue:43475", + "issue:42907|issue:43937", + "issue:42907|issue:43452", + "issue:44038|issue:44297", + "issue:43824|issue:44351", + "issue:44038|issue:44295", + "issue:45216|issue:45478", + "issue:43116|issue:45200", + "issue:44360|issue:44485" + ], + "prepared_review_unit_hash": "5ea6f365e6bebd24a10deff74e9fe8c06755093f56cee80aafe7dfb3ac1f634d", + "component_budget": { + "node_count": 530, + "item_count": 530, + "soft_pair_count": 1954, + "serialized_chars": 1013004, + "estimated_input_tokens": 253251, + "estimated_eval_tokens": 506758 + }, + "budget": { + "node_count": 48, + "item_count": 48, + "soft_pair_count": 31, + "serialized_chars": 36522, + "estimated_input_tokens": 9131, + "estimated_eval_tokens": 18518 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": true, + "analyst_result": { + "summary": "These issues are mostly unrelated, with only superficial overlap around imports, tokenizers, and version regressions. None of the soft-edge pairs look like the same underlying bug or change, so I reject them all. The broadest representative issue is the general upgrade-regression report about many import errors after moving to 5.5.0.", + "confidence": 0.91, + "canonical_issue_reason": "issue:45278 is the broadest umbrella report in the set: it describes widespread import breakage after a version upgrade, which makes it the best representative issue if one needs a single anchor for this mixed cluster.", + "canonical_pr_reason": null, + "best_issue_reason": "Same as canonical: it is the most general, cluster-like report rather than a narrow model- or tokenizer-specific bug.", + "best_pr_reason": null, + "soft_edge_verdicts": [ + { + "left": "issue:42886", + "right": "issue:43066", + "accept": false, + "reason": "Both involve tokenization/versioning, but one is offline cache loading and the other is wrong decoder type in v5; different failure modes and code paths." + }, + { + "left": "issue:42886", + "right": "issue:42994", + "accept": false, + "reason": "Offline tokenizer cache loading and quantized model saving are unrelated I/O paths." + }, + { + "left": "issue:43646", + "right": "issue:46002", + "accept": false, + "reason": "Both mention init-time problems, but one is a Transformers 5.0 custom-model init break and the other is NaNs in classification heads; not the same bug." + }, + { + "left": "issue:39379", + "right": "issue:46172", + "accept": false, + "reason": "Different models and different internal logic: mllama mask handling vs deepseek-v4 tensor transposition." + }, + { + "left": "issue:42757", + "right": "issue:45278", + "accept": false, + "reason": "42978-style broad import regressions are not the same as a specific missing huggingface_hub symbol; related symptom class, different underlying issue." + }, + { + "left": "issue:27957", + "right": "issue:30990", + "accept": false, + "reason": "XLMRoberta Flash Attention 2 support and Sentence Transformers hanging on load are unrelated." + }, + { + "left": "issue:30990", + "right": "issue:33357", + "accept": false, + "reason": "Both are runtime failures, but one is a loading hang and the other is a MacOS bus error with a CLIP model." + }, + { + "left": "issue:43122", + "right": "issue:46326", + "accept": false, + "reason": "Both are tokenizer-related, but one is general version-to-version tokenization drift and the other is a specific continue_final_message bug for mistral_common tokenizers." + }, + { + "left": "issue:43329", + "right": "issue:47059", + "accept": false, + "reason": "Multimodal token counting on video and a CLI completion attribute error in huggingface_hub are unrelated." + }, + { + "left": "issue:44295", + "right": "issue:44297", + "accept": false, + "reason": "A position_ids buffer read error and a tokenizer_class mismatch on save_pretrained affect different parts of the stack." + }, + { + "left": "issue:44625", + "right": "issue:44829", + "accept": false, + "reason": "Qwen3.5 num_labels propagation and flash_attention_3 training collapse are different bugs." + }, + { + "left": "issue:39401", + "right": "issue:44488", + "accept": false, + "reason": "Wrong offset_mapping for Qwen3 tokenizer and a load failure for cjvt/sleng-bert are unrelated." + }, + { + "left": "issue:42907", + "right": "issue:44964", + "accept": false, + "reason": "Saving dequantized Ministrals/Devstrals and loading Phi-4 multimodal-instruct are different model and serialization paths." + }, + { + "left": "issue:39401", + "right": "issue:44521", + "accept": false, + "reason": "Tokenizer offset mapping and multimodal assistant mask generation are distinct tokenizer/template behaviors." + }, + { + "left": "issue:43976", + "right": "issue:44945", + "accept": false, + "reason": "Python-version compatibility in 5.1.0 and incorrect pipeline-parallelism output are unrelated." + }, + { + "left": "issue:42907", + "right": "issue:44945", + "accept": false, + "reason": "Quantized save failure and pipeline parallelism output corruption are different problems." + }, + { + "left": "issue:41628", + "right": "issue:45278", + "accept": false, + "reason": "A missing AutoImageProcessor import is a specific export issue, not the same as broad upgrade-wide import breakage." + }, + { + "left": "issue:44117", + "right": "issue:44704", + "accept": false, + "reason": "Tokenizer mapping returning None and AutoProcessor not forwarding kwargs to cached_file are separate logic bugs." + }, + { + "left": "issue:22355", + "right": "issue:29159", + "accept": false, + "reason": "Missing transformers.onnx module import and slow/fast tokenizer inconsistency are unrelated." + }, + { + "left": "issue:43122", + "right": "issue:44521", + "accept": false, + "reason": "General tokenization drift and multimodal assistant-mask zeros are different behaviors with different causes." + }, + { + "left": "issue:44704", + "right": "issue:45290", + "accept": false, + "reason": "AutoProcessor argument forwarding and chat-template crashes on tool-call messages are distinct code paths." + }, + { + "left": "issue:42371", + "right": "issue:44263", + "accept": false, + "reason": "TF32 API warning and a torch.split return-value bug in GlmMoeDsaIndexer are unrelated." + }, + { + "left": "issue:43408", + "right": "issue:43475", + "accept": false, + "reason": "SAM3 tracker/model-type warning and a missing fpn_position_embeddings attribute are different SAM3 issues." + }, + { + "left": "issue:42907", + "right": "issue:43937", + "accept": false, + "reason": "Model saving failure and GLM-5 generation config validation are unrelated." + }, + { + "left": "issue:42907", + "right": "issue:43452", + "accept": false, + "reason": "Dequantized save failure and gguf_file load issues affect different serialization flows." + }, + { + "left": "issue:44038", + "right": "issue:44297", + "accept": false, + "reason": "Qwen3-VL-Moe compatibility and tokenizer_class mismatch on save are not the same bug." + }, + { + "left": "issue:43824", + "right": "issue:44351", + "accept": false, + "reason": "Two missing imports, but for different symbols and different model subsystems; not one underlying defect." + }, + { + "left": "issue:44038", + "right": "issue:44295", + "accept": false, + "reason": "Qwen3-VL-Moe issue and position_ids buffer error are unrelated." + }, + { + "left": "issue:45216", + "right": "issue:45478", + "accept": false, + "reason": "Both concern Qwen3.5 and version 5.4+, but one is save_pretrained checkpoint correctness and the other is from_pretrained failure; strict duplicate criteria not met." + }, + { + "left": "issue:43116", + "right": "issue:45200", + "accept": false, + "reason": "Example-script multi-label classification returning empty results and Gemma 4 mm_token_type_ids defaults are unrelated." + }, + { + "left": "issue:44360", + "right": "issue:44485", + "accept": false, + "reason": "A missing ReLU in the DSA indexer and GLM-5 RoPE implementation are different algorithmic components." + } + ] + }, + "evaluator_result": null, + "evaluator_used": false, + "retried": false, + "accepted_nontrivial_soft_edge": false, + "error_kind": null, + "error_message": null + }, + { + "label": "issue", + "component_index": 1, + "component_count": 28, + "review_unit_index": 54, + "review_unit_count": 65, + "status": "reviewed", + "reason": null, + "source": "llm", + "cache_hit": false, + "model": "gpt-5.4-mini?service_tier=flex", + "cluster_id": "cluster-10000-530", + "nodes": [ + "issue:22355", + "issue:28348", + "issue:29127", + "issue:36296", + "issue:38617", + "issue:39379", + "issue:39401", + "issue:42175", + "issue:42222", + "issue:42673", + "issue:42757", + "issue:42898", + "issue:43116", + "issue:43122", + "issue:43232", + "issue:43278", + "issue:43493", + "issue:43611", + "issue:43976", + "issue:44038", + "issue:44188", + "issue:44222", + "issue:44351", + "issue:44483", + "issue:44625", + "issue:44779", + "issue:44938", + "issue:44945", + "issue:44995", + "issue:45072", + "issue:45083", + "issue:45278", + "issue:45357", + "issue:45446", + "issue:45478", + "issue:45753", + "issue:46002", + "issue:46172", + "issue:46231", + "issue:46328", + "issue:46614", + "issue:46640", + "issue:46697", + "issue:46762", + "issue:46877", + "issue:47217", + "issue:47312" + ], + "soft_pairs": [ + "issue:43122|issue:44779", + "issue:44038|issue:44222", + "issue:46172|issue:46762", + "issue:22355|issue:29127", + "issue:43278|issue:44945", + "issue:42175|issue:44938", + "issue:39401|issue:46614", + "issue:42222|issue:47312", + "issue:39379|issue:43232", + "issue:42757|issue:44351", + "issue:36296|issue:43116", + "issue:43493|issue:45753", + "issue:44995|issue:46697", + "issue:44188|issue:46328", + "issue:28348|issue:46640", + "issue:43976|issue:45446", + "issue:46231|issue:46877", + "issue:42673|issue:45072", + "issue:44483|issue:47217", + "issue:42898|issue:46614", + "issue:43611|issue:46002", + "issue:45357|issue:45478", + "issue:44625|issue:46002", + "issue:38617|issue:45278", + "issue:45083|issue:45446" + ], + "prepared_review_unit_hash": "0dc5e25c8951b07e739aa85d9ae8f49d0682b1092bee37209ec25132767c16a7", + "component_budget": { + "node_count": 530, + "item_count": 530, + "soft_pair_count": 1954, + "serialized_chars": 1013004, + "estimated_input_tokens": 253251, + "estimated_eval_tokens": 506758 + }, + "budget": { + "node_count": 47, + "item_count": 47, + "soft_pair_count": 25, + "serialized_chars": 34107, + "estimated_input_tokens": 8527, + "estimated_eval_tokens": 17310 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": true, + "analyst_result": { + "summary": "Mostly unrelated issue titles; I found one plausible duplicate-style pair around v5 import regressions, but the rest are distinct bugs/features across different models and subsystems.", + "confidence": 0.84, + "canonical_issue_reason": "issue:45278 is the best umbrella issue here: it describes the broad upgrade-related import regression, and is the most representative of the narrower missing-symbol import failure in issue:38617.", + "canonical_pr_reason": null, + "best_issue_reason": "issue:45278 best captures the shared symptom class among the only close match in the set (post-upgrade import errors).", + "best_pr_reason": null, + "soft_edge_verdicts": [ + { + "left": "issue:43122", + "right": "issue:44779", + "accept": false, + "reason": "Both are tokenizer regressions, but one is a broad v5 tokenization difference and the other is Deepseek-specific incorrect results; not clearly the same bug." + }, + { + "left": "issue:44038", + "right": "issue:44222", + "accept": false, + "reason": "Different problems: Qwen3-VL-Moe compatibility vs FP8 save_pretrained behavior for MoE." + }, + { + "left": "issue:46172", + "right": "issue:46762", + "accept": false, + "reason": "Both mention numerical issues, but they concern different models and attention implementations." + }, + { + "left": "issue:22355", + "right": "issue:29127", + "accept": false, + "reason": "Completely unrelated: ONNX import failure vs LayoutLMv3 box-validation error messaging." + }, + { + "left": "issue:43278", + "right": "issue:44945", + "accept": false, + "reason": "Different failure modes: dtype drift between train/eval vs incorrect output under pipeline parallelism." + }, + { + "left": "issue:42175", + "right": "issue:44938", + "accept": false, + "reason": "Installation/backend dependency issue vs Python 3.14 load failure; no shared concrete code path." + }, + { + "left": "issue:39401", + "right": "issue:46614", + "accept": false, + "reason": "Tokenizer offset mapping bug vs stale phonemizer language backend; different subsystems and symptoms." + }, + { + "left": "issue:42222", + "right": "issue:47312", + "accept": false, + "reason": "VitPose model breakage is unrelated to kernel compatibility with 0.16." + }, + { + "left": "issue:39379", + "right": "issue:43232", + "accept": false, + "reason": "Different generation-path bugs with no clear shared underlying defect." + }, + { + "left": "issue:42757", + "right": "issue:44351", + "accept": false, + "reason": "Both are import errors, but for different symbols and different packages; not the same bug." + }, + { + "left": "issue:36296", + "right": "issue:43116", + "accept": false, + "reason": "Tensor parallel training bug vs multi-label classification example returning empty results; unrelated." + }, + { + "left": "issue:43493", + "right": "issue:45753", + "accept": false, + "reason": "Different models and failures: SigLIP2 implementation mismatch vs Qwen3_5 infinite loop on a specific image." + }, + { + "left": "issue:44995", + "right": "issue:46697", + "accept": false, + "reason": "Stale cache crash in GlmMoeDsa vs image over-padding in Swin2SR; unrelated." + }, + { + "left": "issue:44188", + "right": "issue:46328", + "accept": false, + "reason": "Attention-kernel divergence under torch.compile is unrelated to SequenceFeatureExtractor padding inefficiency." + }, + { + "left": "issue:28348", + "right": "issue:46640", + "accept": false, + "reason": "A feature request for GPT2 flash-attention support is not the same as a general Flash Attention 2 issue." + }, + { + "left": "issue:43976", + "right": "issue:45446", + "accept": false, + "reason": "Version-check incompatibility in Transformers vs a PyTorch version check bug for flex_attention; different code paths." + }, + { + "left": "issue:46231", + "right": "issue:46877", + "accept": false, + "reason": "Docs parameter-name mismatch vs import-time docstring/lint output; not the same underlying issue." + }, + { + "left": "issue:42673", + "right": "issue:45072", + "accept": false, + "reason": "VRAM leak in multi-threaded Qwen3ForCausalLM usage vs dtype mismatches in inference." + }, + { + "left": "issue:44483", + "right": "issue:47217", + "accept": false, + "reason": "Both involve chat/VL behavior, but one is a generic v5.3 request rejection and the other is a placeholder-in-text failure; too different." + }, + { + "left": "issue:42898", + "right": "issue:46614", + "accept": false, + "reason": "Tokenization cleanup behavior change is unrelated to phonemizer backend state retention." + }, + { + "left": "issue:43611", + "right": "issue:46002", + "accept": false, + "reason": "Model-loading breakage with base_model_prefix is unrelated to NaNs in classification heads." + }, + { + "left": "issue:45357", + "right": "issue:45478", + "accept": false, + "reason": "Both are Qwen3.5 serialization/load issues, but one is about incorrect visual encoder keys in save_pretrained and the other is a from_pretrained error; not enough to treat as the same bug." + }, + { + "left": "issue:44625", + "right": "issue:46002", + "accept": false, + "reason": "Configuration propagation in Qwen3.5 has no clear relation to NaNs in classification heads." + }, + { + "left": "issue:38617", + "right": "issue:45278", + "accept": true, + "reason": "This looks like the same upgrade regression: a specific missing import from configuration_utils under the broader 'many import errors after update' report." + }, + { + "left": "issue:45083", + "right": "issue:45446", + "accept": false, + "reason": "Unexpected helper output in qwen3_omni_moe is unrelated to a PyTorch version check bug in flex_attention." + } + ] + }, + "evaluator_result": { + "accept": true, + "feedback": "The summary is grounded in the packet and stays conservative. The one accepted pair (issue:38617 \u2194 issue:45278) is a reasonable umbrella-style match based on post-update import regressions, and the rejected soft pairs are generally well-justified as distinct bug classes rather than overstated duplicates." + }, + "evaluator_used": true, + "retried": false, + "accepted_nontrivial_soft_edge": true, + "error_kind": null, + "error_message": null + }, + { + "label": "issue", + "component_index": 1, + "component_count": 28, + "review_unit_index": 55, + "review_unit_count": 65, + "status": "reviewed", + "reason": null, + "source": "llm", + "cache_hit": false, + "model": "gpt-5.4-mini?service_tier=flex", + "cluster_id": "cluster-10000-530", + "nodes": [ + "issue:33453", + "issue:39379", + "issue:39401", + "issue:42371", + "issue:42673", + "issue:42796", + "issue:43064", + "issue:43066", + "issue:43116", + "issue:43493", + "issue:43618", + "issue:43673", + "issue:43688", + "issue:43824", + "issue:43825", + "issue:43827", + "issue:43874", + "issue:43937", + "issue:43992", + "issue:44075", + "issue:44117", + "issue:44291", + "issue:44355", + "issue:44466", + "issue:44485", + "issue:44610", + "issue:44704", + "issue:44779", + "issue:44936", + "issue:45071", + "issue:45081", + "issue:45092", + "issue:45362", + "issue:45405", + "issue:45446", + "issue:45478", + "issue:45563", + "issue:45588", + "issue:46002", + "issue:46231", + "issue:46328", + "issue:46710", + "issue:46726", + "issue:46762", + "issue:46829", + "issue:46909", + "issue:46981", + "issue:47116" + ], + "soft_pairs": [ + "issue:44075|issue:44936", + "issue:43066|issue:44779", + "issue:44610|issue:46909", + "issue:43116|issue:43493", + "issue:43937|issue:44485", + "issue:45362|issue:45478", + "issue:43066|issue:46710", + "issue:44355|issue:46231", + "issue:42673|issue:42796", + "issue:46726|issue:46981", + "issue:43116|issue:43825", + "issue:42371|issue:43874", + "issue:43618|issue:46829", + "issue:43827|issue:45446", + "issue:43992|issue:44704", + "issue:45405|issue:45446", + "issue:46328|issue:47116", + "issue:33453|issue:39401", + "issue:43824|issue:46981", + "issue:44291|issue:45071", + "issue:39379|issue:43688", + "issue:44117|issue:45588", + "issue:43673|issue:46328", + "issue:43064|issue:43673", + "issue:44117|issue:45081", + "issue:45092|issue:46762", + "issue:43937|issue:45563", + "issue:44466|issue:46002" + ], + "prepared_review_unit_hash": "0b8a83fbfe49bfb0083cb7b9a5f948e4643b557ad02a3b2334629ca6fe28bceb", + "component_budget": { + "node_count": 530, + "item_count": 530, + "soft_pair_count": 1954, + "serialized_chars": 1013004, + "estimated_input_tokens": 253251, + "estimated_eval_tokens": 506758 + }, + "budget": { + "node_count": 48, + "item_count": 48, + "soft_pair_count": 28, + "serialized_chars": 36047, + "estimated_input_tokens": 9012, + "estimated_eval_tokens": 18280 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": true, + "analyst_result": { + "summary": "This cluster is mostly heterogeneous. Several items are tokenizer or v5-regression themed, but the soft pairs usually differ in concrete bug, model, or code path, so they should not be deduplicated.", + "confidence": 0.73, + "canonical_issue_reason": "No single canonical issue stands out: the items span tokenizer regressions, model-specific bugs, training/runtime failures, and docs/API issues rather than one underlying defect.", + "canonical_pr_reason": null, + "best_issue_reason": "issue:43066 is the best representative of the tokenizer/v5 regression subset because it is concrete, user-visible, and closest in theme to the other tokenizer-related complaints.", + "best_pr_reason": null, + "soft_edge_verdicts": [ + { + "left": "issue:44075", + "right": "issue:44936", + "accept": false, + "reason": "Different failures: unused SGD args vs evaluate() breaking after train()." + }, + { + "left": "issue:43066", + "right": "issue:44779", + "accept": false, + "reason": "Both are tokenizer regressions, but they describe different concrete problems and code paths." + }, + { + "left": "issue:44610", + "right": "issue:46909", + "accept": false, + "reason": "Model input-size bug vs a broken documentation example; not the same issue." + }, + { + "left": "issue:43116", + "right": "issue:43493", + "accept": false, + "reason": "Multi-label example output bug vs SigLIP2 implementation discrepancy." + }, + { + "left": "issue:43937", + "right": "issue:44485", + "accept": false, + "reason": "Invalid GenerationConfig vs RoPE implementation discussion; same model family but different bugs." + }, + { + "left": "issue:45362", + "right": "issue:45478", + "accept": false, + "reason": "Different Qwen3.5 failures: chat crash vs from_pretrained error." + }, + { + "left": "issue:43066", + "right": "issue:46710", + "accept": false, + "reason": "Tokenizer decoder-type regression is not the same as a model accuracy regression after upgrade." + }, + { + "left": "issue:44355", + "right": "issue:46231", + "accept": false, + "reason": "Compiled Python file errors vs a documentation parameter-name mismatch." + }, + { + "left": "issue:42673", + "right": "issue:42796", + "accept": false, + "reason": "VRAM leak in dataloader threads vs tokenizer BPE-marker handling." + }, + { + "left": "issue:46726", + "right": "issue:46981", + "accept": false, + "reason": "Different AutoProcessor failure modes: missing vision backend vs repeated 429 retries." + }, + { + "left": "issue:43116", + "right": "issue:43825", + "accept": false, + "reason": "Example script multi-label output bug vs pipeline error-message wording." + }, + { + "left": "issue:42371", + "right": "issue:43874", + "accept": false, + "reason": "TF32 configuration guidance vs a missing image-processor method." + }, + { + "left": "issue:43618", + "right": "issue:46829", + "accept": false, + "reason": "Both touch model outputs, but they are different attributes and different models." + }, + { + "left": "issue:43827", + "right": "issue:45446", + "accept": false, + "reason": "Docs referencing removed pipeline APIs vs a PyTorch version check bug." + }, + { + "left": "issue:43992", + "right": "issue:44704", + "accept": false, + "reason": "Missing encoder weights vs AutoProcessor not forwarding kwargs." + }, + { + "left": "issue:45405", + "right": "issue:45446", + "accept": false, + "reason": "Dependency-version bump issue vs an unrelated import/version check bug." + }, + { + "left": "issue:46328", + "right": "issue:47116", + "accept": false, + "reason": "Pad performance issue vs streaming pipeline regression." + }, + { + "left": "issue:33453", + "right": "issue:39401", + "accept": false, + "reason": "Tokenizer loading regression vs wrong offset mapping; related area but not the same bug." + }, + { + "left": "issue:43824", + "right": "issue:46981", + "accept": false, + "reason": "Import error for a specific model class vs HTTP retry behavior in AutoProcessor." + }, + { + "left": "issue:44291", + "right": "issue:45071", + "accept": false, + "reason": "Unexpected meta-init argument vs PretrainedConfig type-checking breakage." + }, + { + "left": "issue:39379", + "right": "issue:43688", + "accept": false, + "reason": "Mllama mask handling vs auxiliary-loss normalization; unrelated model bugs." + }, + { + "left": "issue:44117", + "right": "issue:45588", + "accept": false, + "reason": "Tokenizer mapping assumption vs flash-attention aux-sentinel crash." + }, + { + "left": "issue:43673", + "right": "issue:46328", + "accept": false, + "reason": "Generation cache bug vs SequenceFeatureExtractor padding performance." + }, + { + "left": "issue:43064", + "right": "issue:43673", + "accept": false, + "reason": "Distributed optimizer-state bug vs generation cache missing; different subsystems." + }, + { + "left": "issue:44117", + "right": "issue:45081", + "accept": false, + "reason": "Mapping-name None handling vs Mistral regex patch backend_tokenizer crash." + }, + { + "left": "issue:45092", + "right": "issue:46762", + "accept": false, + "reason": "Meta-initialization incompatibility vs numerical attention differences." + }, + { + "left": "issue:43937", + "right": "issue:45563", + "accept": false, + "reason": "GenerationConfig validation error vs a stale warning in paged generate()." + }, + { + "left": "issue:44466", + "right": "issue:46002", + "accept": false, + "reason": "Serialization inconsistency vs NaNs in classification heads." + } + ] + }, + "evaluator_result": null, + "evaluator_used": false, + "retried": false, + "accepted_nontrivial_soft_edge": false, + "error_kind": null, + "error_message": null + }, + { + "label": "issue", + "component_index": 1, + "component_count": 28, + "review_unit_index": 56, + "review_unit_count": 65, + "status": "reviewed", + "reason": null, + "source": "llm", + "cache_hit": false, + "model": "gpt-5.4-mini?service_tier=flex", + "cluster_id": "cluster-10000-530", + "nodes": [ + "issue:38617", + "issue:39379", + "issue:39401", + "issue:41628", + "issue:42222", + "issue:42617", + "issue:43122", + "issue:43208", + "issue:43408", + "issue:43421", + "issue:43441", + "issue:43450", + "issue:43742", + "issue:43937", + "issue:43986", + "issue:44220", + "issue:44263", + "issue:44380", + "issue:44410", + "issue:44483", + "issue:44496", + "issue:44704", + "issue:44779", + "issue:44829", + "issue:44877", + "issue:44964", + "issue:44977", + "issue:45071", + "issue:45216", + "issue:45278", + "issue:45310", + "issue:45341", + "issue:45464", + "issue:45561", + "issue:45632", + "issue:46077", + "issue:46231", + "issue:46298", + "issue:46328", + "issue:46512", + "issue:46752", + "issue:46762", + "issue:46894", + "issue:46897", + "issue:46981", + "issue:47000", + "issue:47059", + "issue:47110" + ], + "soft_pairs": [ + "issue:44380|issue:46077", + "issue:42222|issue:47000", + "issue:42617|issue:45341", + "issue:43937|issue:44977", + "issue:39379|issue:46897", + "issue:43408|issue:43742", + "issue:43450|issue:43986", + "issue:42222|issue:43441", + "issue:42222|issue:46752", + "issue:44263|issue:46762", + "issue:45071|issue:46512", + "issue:43208|issue:46894", + "issue:39401|issue:44779", + "issue:44829|issue:45216", + "issue:44410|issue:45310", + "issue:44877|issue:46231", + "issue:45071|issue:47059", + "issue:45632|issue:46981", + "issue:45341|issue:45561", + "issue:45341|issue:46328", + "issue:44704|issue:47110", + "issue:39401|issue:43421", + "issue:38617|issue:41628", + "issue:43122|issue:46298", + "issue:43937|issue:44496", + "issue:38617|issue:44220", + "issue:44483|issue:45464", + "issue:44964|issue:45278" + ], + "prepared_review_unit_hash": "700157a6fd4ad7086806122b9213041a6a9187dea7a5ce39c379b908becb9326", + "component_budget": { + "node_count": 530, + "item_count": 530, + "soft_pair_count": 1954, + "serialized_chars": 1013004, + "estimated_input_tokens": 253251, + "estimated_eval_tokens": 506758 + }, + "budget": { + "node_count": 48, + "item_count": 48, + "soft_pair_count": 28, + "serialized_chars": 35830, + "estimated_input_tokens": 8958, + "estimated_eval_tokens": 18172 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": true, + "analyst_result": { + "summary": "This is a loose cluster of mostly unrelated Transformers bugs and regressions. The soft links are usually just superficial similarity (same subsystem, same model family, or both being \u201cbugs\u201d) rather than the same underlying defect.", + "confidence": 0.93, + "canonical_issue_reason": "issue:45278 is the broadest representative in the set (\u201cmany import errors after update from 4.57.0 to 5.5.0\u201d) and is the closest thing to a cluster umbrella; most others are narrow model- or feature-specific bugs.", + "canonical_pr_reason": null, + "best_issue_reason": "issue:45278 is the most general issue here and the only one that plausibly groups a broader class of upgrade/import regressions, unlike the other highly specific reports.", + "best_pr_reason": null, + "soft_edge_verdicts": [ + { + "left": "issue:44380", + "right": "issue:46077", + "accept": false, + "reason": "Both concern attention behavior, but one is GPT-2 backend scaling and the other is Gemma4 bidirectional masking; different models and different code paths." + }, + { + "left": "issue:42222", + "right": "issue:47000", + "accept": false, + "reason": "Completely different problems: vitpose model loading/breakage vs RTDetrLoss infeasible cost matrix during training." + }, + { + "left": "issue:42617", + "right": "issue:45341", + "accept": false, + "reason": "One is a runtime failure for 3d_parallel.py; the other is a small testing_utils bug. No shared underlying defect." + }, + { + "left": "issue:43937", + "right": "issue:44977", + "accept": false, + "reason": "Both are generation-related, but one is a malformed GenerationConfig on GLM-5 and the other is Qwen3.5 flash-attention generation behavior." + }, + { + "left": "issue:39379", + "right": "issue:46897", + "accept": false, + "reason": "Different models and different bugs: incorrect mllama masking vs Florence2 training-loss double-shift." + }, + { + "left": "issue:43408", + "right": "issue:43742", + "accept": false, + "reason": "SAM3 tracker/model-type warning is unrelated to a MobileLLM loading KeyError." + }, + { + "left": "issue:43450", + "right": "issue:43986", + "accept": false, + "reason": "Both touch video processing, but one is batched output shape and the other is a missing torchvision dependency crash during AutoProcessor loading." + }, + { + "left": "issue:42222", + "right": "issue:43441", + "accept": false, + "reason": "Unrelated model families and failure modes: vitpose breakage vs Ministral-3 FlashAttention failure." + }, + { + "left": "issue:42222", + "right": "issue:46752", + "accept": false, + "reason": "Vitpose model breakage and empty-list apply_chat_template crash are unrelated." + }, + { + "left": "issue:44263", + "right": "issue:46762", + "accept": false, + "reason": "Different functionality and symptoms: torch.split return handling in a MoE indexer vs Minimax MSA numerical differences." + }, + { + "left": "issue:45071", + "right": "issue:46512", + "accept": false, + "reason": "Both mention config/dtype, but one is a type-checking regression in PretrainedConfig and the other is stale config.dtype after from_config." + }, + { + "left": "issue:43208", + "right": "issue:46894", + "accept": false, + "reason": "xLSTM training bugs for small models are unrelated to CohereASR's training-loss double-shift issue." + }, + { + "left": "issue:39401", + "right": "issue:44779", + "accept": false, + "reason": "Both are tokenizer regressions, but they affect different models and likely different tokenization logic." + }, + { + "left": "issue:44829", + "right": "issue:45216", + "accept": false, + "reason": "One is a flash-attention training degeneration bug; the other is an incorrect saved checkpoint from save_pretrained." + }, + { + "left": "issue:44410", + "right": "issue:45310", + "accept": false, + "reason": "Both involve Qwen variants, but one is missing layer projections and the other is a from_pretrained error in Qwen3.5 MoE." + }, + { + "left": "issue:44877", + "right": "issue:46231", + "accept": false, + "reason": "Strict granite_speech config loading and a docs parameter-name mismatch are not the same bug." + }, + { + "left": "issue:45071", + "right": "issue:47059", + "accept": false, + "reason": "A PretrainedConfig type-checking regression is unrelated to a CLI completion attribute error from huggingface_hub." + }, + { + "left": "issue:45632", + "right": "issue:46981", + "accept": false, + "reason": "Cache-path collisions for local trust_remote_code models are unrelated to repeated 429 retries in AutoProcessor.from_pretrained." + }, + { + "left": "issue:45341", + "right": "issue:45561", + "accept": false, + "reason": "Both are testing_utils-related, but one is only described as a small bug while the other is a specific xdist race on captured_info.txt; insufficient evidence they are the same defect." + }, + { + "left": "issue:45341", + "right": "issue:46328", + "accept": false, + "reason": "A testing_utils bug and a SequenceFeatureExtractor.pad performance issue are unrelated." + }, + { + "left": "issue:44704", + "right": "issue:47110", + "accept": false, + "reason": "Passing kwargs to cached_file in AutoProcessor is unrelated to special_tokens being a list without keys()." + }, + { + "left": "issue:39401", + "right": "issue:43421", + "accept": false, + "reason": "Tokenizer offset_mapping correctness and runtime post-processor updates are different tokenizer bugs." + }, + { + "left": "issue:38617", + "right": "issue:41628", + "accept": false, + "reason": "Both are import errors, but they concern different missing symbols and likely different packaging/export issues." + }, + { + "left": "issue:43122", + "right": "issue:46298", + "accept": false, + "reason": "Tokenizer behavior drift across versions is unrelated to a noisy logging message." + }, + { + "left": "issue:43937", + "right": "issue:44496", + "accept": false, + "reason": "GenerationConfig invalid vs unrecognized model/config.json model_type are different loading failures." + }, + { + "left": "issue:38617", + "right": "issue:44220", + "accept": false, + "reason": "ImportError for layer_type_validation and an audio feature extraction issue are unrelated." + }, + { + "left": "issue:44483", + "right": "issue:45464", + "accept": false, + "reason": "Both hit chat/completions, but one is a general request rejection and the other is a streaming inference failure for a specific Qwen3.5 model." + }, + { + "left": "issue:44964", + "right": "issue:45278", + "accept": false, + "reason": "A specific Phi-4 multimodal model-load failure is not the same defect as a broad set of import errors after upgrading to 5.5.0." + } + ] + }, + "evaluator_result": null, + "evaluator_used": false, + "retried": false, + "accepted_nontrivial_soft_edge": false, + "error_kind": null, + "error_message": null + }, + { + "label": "issue", + "component_index": 1, + "component_count": 28, + "review_unit_index": 57, + "review_unit_count": 65, + "status": "reviewed", + "reason": null, + "source": "llm", + "cache_hit": false, + "model": "gpt-5.4-mini?service_tier=flex", + "cluster_id": "cluster-10000-530", + "nodes": [ + "issue:27957", + "issue:41628", + "issue:42371", + "issue:42831", + "issue:43012", + "issue:43066", + "issue:43097", + "issue:43122", + "issue:43208", + "issue:43352", + "issue:43441", + "issue:43452", + "issue:43531", + "issue:43577", + "issue:43582", + "issue:43673", + "issue:43874", + "issue:43931", + "issue:43957", + "issue:43986", + "issue:44038", + "issue:44075", + "issue:44077", + "issue:44117", + "issue:44230", + "issue:44261", + "issue:44263", + "issue:44351", + "issue:44360", + "issue:44410", + "issue:44484", + "issue:44485", + "issue:44493", + "issue:44829", + "issue:44987", + "issue:45071", + "issue:45310", + "issue:45341", + "issue:45478", + "issue:45588", + "issue:45656", + "issue:46050", + "issue:46172", + "issue:46752", + "issue:46762", + "issue:46897", + "issue:46981", + "issue:47116" + ], + "soft_pairs": [ + "issue:42371|issue:44261", + "issue:44075|issue:45656", + "issue:43097|issue:44493", + "issue:42371|issue:42831", + "issue:44261|issue:46762", + "issue:44484|issue:45310", + "issue:43066|issue:43122", + "issue:44077|issue:46981", + "issue:44410|issue:45478", + "issue:43931|issue:44410", + "issue:44263|issue:44485", + "issue:43673|issue:45341", + "issue:41628|issue:45071", + "issue:43097|issue:43452", + "issue:43208|issue:46897", + "issue:43931|issue:44038", + "issue:43441|issue:44829", + "issue:43012|issue:43577", + "issue:46752|issue:47116", + "issue:43874|issue:44261", + "issue:27957|issue:43352", + "issue:44117|issue:44987", + "issue:44484|issue:45478", + "issue:44263|issue:46172", + "issue:43531|issue:45341", + "issue:43957|issue:44351", + "issue:44360|issue:46172", + "issue:43097|issue:45588", + "issue:43582|issue:45341", + "issue:44038|issue:44230", + "issue:44261|issue:44485", + "issue:43986|issue:45341", + "issue:43097|issue:46050" + ], + "prepared_review_unit_hash": "9a9c91a1c299c408e7fedbf6a14c970d0c4d1e1d8c658ffdf8a140aeba3e2eb8", + "component_budget": { + "node_count": 530, + "item_count": 530, + "soft_pair_count": 1954, + "serialized_chars": 1013004, + "estimated_input_tokens": 253251, + "estimated_eval_tokens": 506758 + }, + "budget": { + "node_count": 48, + "item_count": 48, + "soft_pair_count": 33, + "serialized_chars": 37182, + "estimated_input_tokens": 9296, + "estimated_eval_tokens": 18848 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": true, + "analyst_result": { + "summary": "The cluster is broadly heterogeneous: most items are separate Transformers regressions across precision, tokenizer, flash-attention, loading, and optimizer behavior. None of the provided soft edges look like the same underlying bug/change, so I would not merge any of them as duplicates. The clearest standalone issue in the set is the Qwen3.5 MoE from_pretrained failure (45310), which also has a near-identical later follow-up (45478).", + "confidence": 0.87, + "canonical_issue_reason": "issue:45310 is the clearest concrete bug report in the set and appears to have a near-identical follow-up in issue:45478; it is specific, reproducible, and earlier than the duplicate-looking follow-up.", + "canonical_pr_reason": null, + "best_issue_reason": "issue:45310 is the best representative issue: it describes a specific model-loading failure with enough detail to anchor a fix, and it is the strongest apparent duplicate target among the listed issues.", + "best_pr_reason": null, + "soft_edge_verdicts": [ + { + "left": "issue:42371", + "right": "issue:44261", + "accept": false, + "reason": "Both are precision-related, but one is TF32 API guidance and the other is a missing rms_norm_eps causing MLA accuracy drift; different code paths and fixes." + }, + { + "left": "issue:44075", + "right": "issue:45656", + "accept": false, + "reason": "Different optimizer bugs: SGD args ignored vs deepspeed causing the optimizer step to run twice." + }, + { + "left": "issue:43097", + "right": "issue:44493", + "accept": false, + "reason": "One is about tie_embeddings_and_encoder_decoder removal in v5; the other is unexpected position-id keys. Both are v5 regressions but not the same bug." + }, + { + "left": "issue:42371", + "right": "issue:42831", + "accept": false, + "reason": "TF32 control API behavior and FineGrainedFP8 accuracy issues are distinct precision topics with unrelated failure modes." + }, + { + "left": "issue:44261", + "right": "issue:46762", + "accept": false, + "reason": "Different model/math issues: MLA q_a_layernorm epsilon handling vs Minimax MSA attention numerical differences." + }, + { + "left": "issue:44484", + "right": "issue:45310", + "accept": false, + "reason": "save_pretrained shard-size behavior is unrelated to a Qwen3.5 MoE from_pretrained loading error." + }, + { + "left": "issue:43066", + "right": "issue:43122", + "accept": false, + "reason": "Both involve tokenizer behavior in v5, but one is wrong decoder type and the other is changed tokenization output; not clearly the same defect." + }, + { + "left": "issue:44077", + "right": "issue:46981", + "accept": false, + "reason": "Optional post_init validation in a model config is unrelated to AutoProcessor retry behavior on 429s." + }, + { + "left": "issue:44410", + "right": "issue:45478", + "accept": false, + "reason": "Different Qwen-family loading problems: missing attn projections in qwen3next vs Qwen3.5 MoE from_pretrained failure." + }, + { + "left": "issue:43931", + "right": "issue:44410", + "accept": false, + "reason": "Weight-shape mismatch for Qwen3-VL-30B-A3B-Instruct is not the same as missing layer-0 projections in qwen3next." + }, + { + "left": "issue:44263", + "right": "issue:44485", + "accept": false, + "reason": "torch.split return values in a GLM MoE indexer is unrelated to the GLM-5 RoPE implementation question." + }, + { + "left": "issue:43673", + "right": "issue:45341", + "accept": false, + "reason": "Generation cache missing during chunked_prefill and a small testing_utils bug are unrelated." + }, + { + "left": "issue:41628", + "right": "issue:45071", + "accept": false, + "reason": "Missing AutoImageProcessor import and PretrainedConfig type-checking regression affect different APIs and code paths." + }, + { + "left": "issue:43097", + "right": "issue:43452", + "accept": false, + "reason": "A removed embedding-tie config and gguf_file loading failures are different from_pretrained/config issues, not one bug." + }, + { + "left": "issue:43208", + "right": "issue:46897", + "accept": false, + "reason": "xLSTM training blockers and Florence2 double-shift loss bug are separate model-specific training problems." + }, + { + "left": "issue:43931", + "right": "issue:44038", + "accept": false, + "reason": "Both are Qwen3-VL-related, but one is a weight mismatch and the other is a transformers v5 + Qwen3-VL-Moe issue; too different to merge." + }, + { + "left": "issue:43441", + "right": "issue:44829", + "accept": false, + "reason": "FlashAttention failures in Ministral-3 and degenerate training with flash_attention_3 are different implementations and symptoms." + }, + { + "left": "issue:43012", + "right": "issue:43577", + "accept": false, + "reason": "A PyTorch warning when compiling bfloat16 models is not the same as Blip2 submodules staying float32 on load." + }, + { + "left": "issue:46752", + "right": "issue:47116", + "accept": false, + "reason": "Empty-list chat template crash and streaming pipeline regression are unrelated API paths." + }, + { + "left": "issue:43874", + "right": "issue:44261", + "accept": false, + "reason": "Image patch counting in GLM46V and MLA epsilon precision drift are different multimodal/model-numerics bugs." + }, + { + "left": "issue:27957", + "right": "issue:43352", + "accept": false, + "reason": "Both mention Flash Attention 2, but they are different models with separate support issues." + }, + { + "left": "issue:44117", + "right": "issue:44987", + "accept": false, + "reason": "Tokenizer mapping None handling and loading physical-intelligence/fast are not the same concrete failure." + }, + { + "left": "issue:44484", + "right": "issue:45478", + "accept": false, + "reason": "Shard-size defaults in save_pretrained are unrelated to the Qwen3.5 MoE loading error." + }, + { + "left": "issue:44263", + "right": "issue:46172", + "accept": false, + "reason": "GLM MoE DSA indexer split semantics and deepseek-v4 dtype/transpose code are distinct bugs." + }, + { + "left": "issue:43531", + "right": "issue:45341", + "accept": false, + "reason": "Qwen3-MoE sliding_window behavior and a testing_utils bug do not share the same code path." + }, + { + "left": "issue:43957", + "right": "issue:44351", + "accept": false, + "reason": "Meta-device loading breakage in some models and cannot import HybridCache are unrelated import/loading regressions." + }, + { + "left": "issue:44360", + "right": "issue:46172", + "accept": false, + "reason": "A missing ReLU in the DSA indexer is not the same as the deepseek-v4 comb.to(dtype).transpose(-1, -2) bug." + }, + { + "left": "issue:43097", + "right": "issue:45588", + "accept": false, + "reason": "Removed tie_embeddings behavior and flash_attention.py crashing on sink-less models are different regressions." + }, + { + "left": "issue:43582", + "right": "issue:45341", + "accept": false, + "reason": "Apple Silicon warmup TypeError and a testing utility bug are unrelated." + }, + { + "left": "issue:44038", + "right": "issue:44230", + "accept": false, + "reason": "Qwen3-VL-Moe inference support and qwen3-vl-fp8/qwen3.5 moe fp8 inference support are different feature requests/failures." + }, + { + "left": "issue:44261", + "right": "issue:44485", + "accept": false, + "reason": "Missing rms_norm_eps in MLA q_a_layernorm and GLM-5 RoPE implementation are separate numerical correctness issues." + }, + { + "left": "issue:43986", + "right": "issue:45341", + "accept": false, + "reason": "AutoProcessor without torchvision and a testing_utils bug are unrelated." + }, + { + "left": "issue:43097", + "right": "issue:46050", + "accept": false, + "reason": "Embedding tie removal in v5 and quantized model save/load corruption are different model persistence issues." + } + ] + }, + "evaluator_result": null, + "evaluator_used": false, + "retried": false, + "accepted_nontrivial_soft_edge": false, + "error_kind": null, + "error_message": null + }, + { + "label": "issue", + "component_index": 1, + "component_count": 28, + "review_unit_index": 58, + "review_unit_count": 65, + "status": "reviewed", + "reason": null, + "source": "llm", + "cache_hit": false, + "model": "gpt-5.4-mini?service_tier=flex", + "cluster_id": "cluster-10000-530", + "nodes": [ + "issue:38617", + "issue:42757", + "issue:42890", + "issue:43012", + "issue:43064", + "issue:43097", + "issue:43208", + "issue:43502", + "issue:43550", + "issue:43673", + "issue:43873", + "issue:43874", + "issue:43986", + "issue:43992", + "issue:44038", + "issue:44062", + "issue:44117", + "issue:44246", + "issue:44263", + "issue:44315", + "issue:44351", + "issue:44380", + "issue:44410", + "issue:44451", + "issue:44484", + "issue:44704", + "issue:44936", + "issue:44938", + "issue:45103", + "issue:45372", + "issue:45375", + "issue:45440", + "issue:45753", + "issue:46123", + "issue:46172", + "issue:46178", + "issue:46298", + "issue:46365", + "issue:46369", + "issue:46421", + "issue:46655", + "issue:46664", + "issue:46701", + "issue:46710", + "issue:46752", + "issue:46821", + "issue:46981", + "issue:47116" + ], + "soft_pairs": [ + "issue:43986|issue:45372", + "issue:45753|issue:46369", + "issue:44246|issue:46981", + "issue:45440|issue:46172", + "issue:43097|issue:43873", + "issue:44410|issue:45375", + "issue:44062|issue:44117", + "issue:43673|issue:46123", + "issue:43012|issue:46821", + "issue:44117|issue:46701", + "issue:43874|issue:44263", + "issue:44038|issue:44410", + "issue:38617|issue:42757", + "issue:46655|issue:46664", + "issue:43208|issue:43550", + "issue:46298|issue:46752", + "issue:44936|issue:46365", + "issue:43064|issue:44315", + "issue:46172|issue:46710", + "issue:43986|issue:44484", + "issue:42890|issue:46298", + "issue:44704|issue:45103", + "issue:38617|issue:44351", + "issue:43502|issue:43992", + "issue:44451|issue:47116", + "issue:44380|issue:46421", + "issue:44938|issue:46178", + "issue:43064|issue:47116", + "issue:43012|issue:43550" + ], + "prepared_review_unit_hash": "7029afebe72103d7d974f9e766b52d9c5d3bd12ccc33c823345efbd3f36ea341", + "component_budget": { + "node_count": 530, + "item_count": 530, + "soft_pair_count": 1954, + "serialized_chars": 1013004, + "estimated_input_tokens": 253251, + "estimated_eval_tokens": 506758 + }, + "budget": { + "node_count": 48, + "item_count": 48, + "soft_pair_count": 29, + "serialized_chars": 36167, + "estimated_input_tokens": 9042, + "estimated_eval_tokens": 18340 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": true, + "analyst_result": { + "summary": "The set is a loose collection of unrelated issues; none of the soft-edge pairs look like true duplicates or mergeable twin PRs, so all soft edges are rejected. I picked a single broad, concrete bug report as the best representative issue.", + "confidence": 0.2, + "canonical_issue_reason": "issue:43502 is a clear, self-contained bug with broad user impact: `local_files_only=True` still causes network/API requests. It has a concrete behavior mismatch and a straightforward reproduction target, making it a strong representative issue.", + "canonical_pr_reason": null, + "best_issue_reason": "Same as canonical: issue:43502 is the strongest high-signal issue here because it describes a specific, impactful regression with well-defined expected behavior.", + "best_pr_reason": null, + "soft_edge_verdicts": [ + { + "left": "issue:43986", + "right": "issue:45372", + "accept": false, + "reason": "Both are ImportErrors, but they involve different missing symbols, different dependencies, and different loading paths; not the same underlying bug." + }, + { + "left": "issue:45753", + "right": "issue:46369", + "accept": false, + "reason": "Different models and different failure modes: an infinite loop on a specific image vs a memory overflow. No shared underlying code-path bug is evident." + }, + { + "left": "issue:44246", + "right": "issue:46981", + "accept": false, + "reason": "One is slow `import transformers`; the other is repeated retry behavior in `AutoProcessor.from_pretrained` with `HF_ENDPOINT`. Different operations and fixes." + }, + { + "left": "issue:45440", + "right": "issue:46172", + "accept": false, + "reason": "Both mention DeepSeek models, but one is a divergence from remote implementation and the other is a tensor dtype/transpose bug. Different concrete issues." + }, + { + "left": "issue:43097", + "right": "issue:43873", + "accept": false, + "reason": "Unrelated problems: a removed config attribute in v5 vs offloading behavior under quantization." + }, + { + "left": "issue:44410", + "right": "issue:45375", + "accept": false, + "reason": "Different Qwen-family model bugs: missing projections in one model and a missing vision config field in another." + }, + { + "left": "issue:44062", + "right": "issue:44117", + "accept": false, + "reason": "Tokenizer construction error vs a mapping-name lookup returning None. They touch tokenizer code but are not the same defect." + }, + { + "left": "issue:43673", + "right": "issue:46123", + "accept": false, + "reason": "Both are generation-related regressions, but one is cache availability in chunked prefill and the other is a pipeline batching flag bug. Different code paths." + }, + { + "left": "issue:43012", + "right": "issue:46821", + "accept": false, + "reason": "PyTorch warning about bfloat16 precision vs an `AttributeError` on `window_size`. No meaningful duplicate relationship." + }, + { + "left": "issue:44117", + "right": "issue:46701", + "accept": false, + "reason": "Tokenizer mapping assumption bug vs an image-text-to-text pipeline failure. Different components and symptoms." + }, + { + "left": "issue:43874", + "right": "issue:44263", + "accept": false, + "reason": "Both are GLM-related, but one is a missing image-processor method and the other is a `torch.split` return-value issue. Not the same bug." + }, + { + "left": "issue:44038", + "right": "issue:44410", + "accept": false, + "reason": "Different Qwen-family regressions with different missing pieces and different failure surfaces." + }, + { + "left": "issue:38617", + "right": "issue:42757", + "accept": false, + "reason": "Both are import errors, but they concern unrelated symbols from different modules; no shared root cause is shown." + }, + { + "left": "issue:46655", + "right": "issue:46664", + "accept": false, + "reason": "Same broad SmolVLM area, but one is padding math and the other is the meaning of `num_frames`. Distinct bugs." + }, + { + "left": "issue:43208", + "right": "issue:43550", + "accept": false, + "reason": "xLSTM training bugs are unrelated to a Bamba torch.compile/SDPA failure." + }, + { + "left": "issue:46298", + "right": "issue:46752", + "accept": false, + "reason": "Noisy logging and an `IndexError` in `apply_chat_template` are unrelated." + }, + { + "left": "issue:44936", + "right": "issue:46365", + "accept": false, + "reason": "A post-train evaluation failure and an import-time Trainer conflict with `datasets` are different failure modes." + }, + { + "left": "issue:43064", + "right": "issue:44315", + "accept": false, + "reason": "One is a rank-sharded optimizer-state bug with FSDP2/PEFT; the other is Liger Kernel not applied under `model_init`. Different code paths." + }, + { + "left": "issue:46172", + "right": "issue:46710", + "accept": false, + "reason": "A specific tensor operation bug in DeepSeek-V4 is not the same as an accuracy regression after upgrading Transformers." + }, + { + "left": "issue:43986", + "right": "issue:44484", + "accept": false, + "reason": "A crash loading a video model without torchvision is unrelated to a question about `max_shard_size` defaults." + }, + { + "left": "issue:42890", + "right": "issue:46298", + "accept": false, + "reason": "Test flakiness from missing `set_seed()` is unrelated to a noisy logging message." + }, + { + "left": "issue:44704", + "right": "issue:45103", + "accept": false, + "reason": "Both mention kwargs handling, but one is `AutoProcessor.from_pretrained` not forwarding kwargs and the other is a docstring parser crash with future annotations." + }, + { + "left": "issue:38617", + "right": "issue:44351", + "accept": false, + "reason": "Both are missing-import complaints, but the missing names and affected APIs differ; these are separate export regressions." + }, + { + "left": "issue:43502", + "right": "issue:43992", + "accept": false, + "reason": "`local_files_only` network leakage and a missing `embed_tokens.weight` in UMT5 are different bugs in different loading paths." + }, + { + "left": "issue:44451", + "right": "issue:47116", + "accept": false, + "reason": "Model load failure for ScandiBERT vs a streaming pipeline regression. No common underlying defect is indicated." + }, + { + "left": "issue:44380", + "right": "issue:46421", + "accept": false, + "reason": "Attention-scaling being ignored under SDPA/FlashAttention is unrelated to cache lazy initialization causing recompilation." + }, + { + "left": "issue:44938", + "right": "issue:46178", + "accept": false, + "reason": "Python 3.14 import failure and a backend mapping error for `tensorflow_text` are not the same issue." + }, + { + "left": "issue:43064", + "right": "issue:47116", + "accept": false, + "reason": "A distributed-training optimizer-state bug is unrelated to a streaming pipeline regression." + }, + { + "left": "issue:43012", + "right": "issue:43550", + "accept": false, + "reason": "A precision warning during bfloat16 compilation is not the same as a model failure under SDPA with torch.compile." + } + ] + }, + "evaluator_result": null, + "evaluator_used": false, + "retried": false, + "accepted_nontrivial_soft_edge": false, + "error_kind": null, + "error_message": null + }, + { + "label": "issue", + "component_index": 1, + "component_count": 28, + "review_unit_index": 59, + "review_unit_count": 65, + "status": "reviewed", + "reason": null, + "source": "llm", + "cache_hit": false, + "model": "gpt-5.4-mini?service_tier=flex", + "cluster_id": "cluster-10000-530", + "nodes": [ + "issue:36010", + "issue:36683", + "issue:42175", + "issue:42757", + "issue:42831", + "issue:42890", + "issue:42907", + "issue:43012", + "issue:43064", + "issue:43066", + "issue:43097", + "issue:43208", + "issue:43352", + "issue:43526", + "issue:43582", + "issue:43618", + "issue:43756", + "issue:43761", + "issue:43824", + "issue:43825", + "issue:43827", + "issue:43906", + "issue:43976", + "issue:43992", + "issue:44077", + "issue:44162", + "issue:44222", + "issue:44261", + "issue:44368", + "issue:44380", + "issue:44483", + "issue:44484", + "issue:44485", + "issue:44514", + "issue:44655", + "issue:44734", + "issue:45478", + "issue:45479", + "issue:45584", + "issue:45712", + "issue:45753", + "issue:46132", + "issue:46172", + "issue:46298", + "issue:46365", + "issue:47059", + "issue:47116", + "issue:47276" + ], + "soft_pairs": [ + "issue:42907|issue:45753", + "issue:45712|issue:46298", + "issue:43066|issue:43906", + "issue:42907|issue:44655", + "issue:43208|issue:43526", + "issue:45584|issue:46365", + "issue:43825|issue:44655", + "issue:43012|issue:44368", + "issue:43352|issue:43976", + "issue:43992|issue:44484", + "issue:43827|issue:47116", + "issue:44734|issue:47116", + "issue:44077|issue:46365", + "issue:44380|issue:46132", + "issue:44162|issue:46172", + "issue:36683|issue:43582", + "issue:43352|issue:44380", + "issue:42757|issue:43824", + "issue:44380|issue:45479", + "issue:43097|issue:46298", + "issue:43012|issue:47276", + "issue:42175|issue:46365", + "issue:43756|issue:44485", + "issue:44222|issue:45478", + "issue:36010|issue:42757", + "issue:43064|issue:44380", + "issue:42890|issue:47059", + "issue:44483|issue:44514", + "issue:42831|issue:44261", + "issue:43618|issue:43761" + ], + "prepared_review_unit_hash": "dcdf7d39a23f37282dcc2194566ed2eba35db592d5a5faae030089afe37bd44b", + "component_budget": { + "node_count": 530, + "item_count": 530, + "soft_pair_count": 1954, + "serialized_chars": 1013004, + "estimated_input_tokens": 253251, + "estimated_eval_tokens": 506758 + }, + "budget": { + "node_count": 48, + "item_count": 48, + "soft_pair_count": 30, + "serialized_chars": 36590, + "estimated_input_tokens": 9148, + "estimated_eval_tokens": 18552 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": true, + "analyst_result": { + "summary": "This cluster is mostly a loose theme of Transformers v5 regressions and model-specific bugs, not a set of true duplicates. I did not find any soft pair that clearly shares the same underlying bug/change.", + "confidence": 0.87, + "canonical_issue_reason": "issue:43824 is the clearest, most self-contained user-facing regression in the set, with strong discussion and inbound references, making it the best representative issue for this mixed cluster.", + "canonical_pr_reason": null, + "best_issue_reason": "issue:43824 has the strongest signal for triage: a concrete import failure, clear repro surface, and enough activity to serve as the cluster anchor.", + "best_pr_reason": null, + "soft_edge_verdicts": [ + { + "left": "issue:42907", + "right": "issue:45753", + "accept": false, + "reason": "Different bugs and models: failing to save dequantized Ministrals/Devstrals versus Qwen3_5 image-induced infinite loop." + }, + { + "left": "issue:45712", + "right": "issue:46298", + "accept": false, + "reason": "Unrelated maintenance issues: leftover dummy classes versus a noisy logging message." + }, + { + "left": "issue:43066", + "right": "issue:43906", + "accept": false, + "reason": "Both are vague v5-related issues, but they do not describe the same concrete bug or code path." + }, + { + "left": "issue:42907", + "right": "issue:44655", + "accept": false, + "reason": "Both involve saving, but one is about dequantized model saving and the other about saving Pipeline objects; different code paths." + }, + { + "left": "issue:43208", + "right": "issue:43526", + "accept": false, + "reason": "Completely different areas: xLSTM training bugs versus BEiT image processor label reduction." + }, + { + "left": "issue:45584", + "right": "issue:46365", + "accept": false, + "reason": "Whisper generation failure and Trainer import dependency conflict are unrelated." + }, + { + "left": "issue:43825", + "right": "issue:44655", + "accept": false, + "reason": "One is a pipeline error message regression, the other is pipeline serialization; not the same issue." + }, + { + "left": "issue:43012", + "right": "issue:44368", + "accept": false, + "reason": "Different warnings from different mechanisms: bf16 compilation precision warning versus tie_word_embeddings advice." + }, + { + "left": "issue:43352", + "right": "issue:43976", + "accept": false, + "reason": "One is a model Flash Attention support limitation, the other is a Python version compatibility problem." + }, + { + "left": "issue:43992", + "right": "issue:44484", + "accept": false, + "reason": "Missing embed_tokens weights on load and max_shard_size behavior in save_pretrained are different problems." + }, + { + "left": "issue:43827", + "right": "issue:47116", + "accept": false, + "reason": "Docs still referencing pipeline() and a streaming pipeline runtime regression are not the same bug." + }, + { + "left": "issue:44734", + "right": "issue:47116", + "accept": false, + "reason": "Different pipeline-related failures: server KV-cache indexing versus streaming pipeline regression." + }, + { + "left": "issue:44077", + "right": "issue:46365", + "accept": false, + "reason": "PatchTSMixer post_init validation and Trainer import dependency issues are unrelated." + }, + { + "left": "issue:44380", + "right": "issue:46132", + "accept": false, + "reason": "Both touch attention code, but one is a backend-specific config being ignored while the other is register() behavior changing." + }, + { + "left": "issue:44162", + "right": "issue:46172", + "accept": false, + "reason": "ESM2 breakage and deepseek-v4 tensor/transposition behavior are unrelated model-specific bugs." + }, + { + "left": "issue:36683", + "right": "issue:43582", + "accept": false, + "reason": "Gemma3Config missing vocab_size and Apple Silicon allocator TypeError have no shared root cause." + }, + { + "left": "issue:43352", + "right": "issue:44380", + "accept": false, + "reason": "Flash Attention support for Nemotron and GPT2 attention scaling being ignored are different issues." + }, + { + "left": "issue:42757", + "right": "issue:43824", + "accept": false, + "reason": "Both are import errors, but they involve different packages, symbols, and failure modes." + }, + { + "left": "issue:44380", + "right": "issue:45479", + "accept": false, + "reason": "Attention scaling being ignored is unrelated to single-label classification producing zero loss." + }, + { + "left": "issue:43097", + "right": "issue:46298", + "accept": false, + "reason": "A removed config flag and a noisy logging message are unrelated." + }, + { + "left": "issue:43012", + "right": "issue:47276", + "accept": false, + "reason": "Both mention warnings, but they concern different features and code paths." + }, + { + "left": "issue:42175", + "right": "issue:46365", + "accept": false, + "reason": "Torch extra backend selection and Trainer import failure via datasets are different dependency problems." + }, + { + "left": "issue:43756", + "right": "issue:44485", + "accept": false, + "reason": "Both mention RoPE, but they are distinct model implementations with different symptoms." + }, + { + "left": "issue:44222", + "right": "issue:45478", + "accept": false, + "reason": "Both are MoE-related, but one is FP8 save_pretrained and the other is from_pretrained loading failure; not the same concrete bug." + }, + { + "left": "issue:36010", + "right": "issue:42757", + "accept": false, + "reason": "Different import failures from different modules and symbols." + }, + { + "left": "issue:43064", + "right": "issue:44380", + "accept": false, + "reason": "Trainer/FSDP2 optimizer-state corruption is unrelated to GPT2 attention scaling being ignored." + }, + { + "left": "issue:42890", + "right": "issue:47059", + "accept": false, + "reason": "Test flakiness from missing set_seed and a huggingface_hub CLI incompatibility are unrelated." + }, + { + "left": "issue:44483", + "right": "issue:44514", + "accept": false, + "reason": "Chat completions API rejection and Qwen2_5_VLProcessor batched-template crash are different endpoints and bugs." + }, + { + "left": "issue:42831", + "right": "issue:44261", + "accept": false, + "reason": "Both concern precision, but one is a broad FineGrainedFP8 accuracy issue and the other is a missing rms_norm_eps in MLA q_a_layernorm." + }, + { + "left": "issue:43618", + "right": "issue:43761", + "accept": false, + "reason": "Both are CLIP regressions, but hidden_states not returned and attentions no longer assigned are separate outputs/code paths." + } + ] + }, + "evaluator_result": null, + "evaluator_used": false, + "retried": false, + "accepted_nontrivial_soft_edge": false, + "error_kind": null, + "error_message": null + }, + { + "label": "issue", + "component_index": 1, + "component_count": 28, + "review_unit_index": 60, + "review_unit_count": 65, + "status": "reviewed", + "reason": null, + "source": "llm", + "cache_hit": false, + "model": "gpt-5.4-mini?service_tier=flex", + "cluster_id": "cluster-10000-530", + "nodes": [ + "issue:29870", + "issue:33453", + "issue:36296", + "issue:36683", + "issue:38617", + "issue:43010", + "issue:43064", + "issue:43299", + "issue:43352", + "issue:43475", + "issue:43502", + "issue:43611", + "issue:43618", + "issue:43867", + "issue:43976", + "issue:43992", + "issue:44075", + "issue:44077", + "issue:44263", + "issue:44355", + "issue:44380", + "issue:44393", + "issue:44492", + "issue:44625", + "issue:44655", + "issue:44855", + "issue:44908", + "issue:44938", + "issue:45216", + "issue:45276", + "issue:45278", + "issue:45310", + "issue:45440", + "issue:45593", + "issue:46077", + "issue:46202", + "issue:46365", + "issue:46512", + "issue:46851", + "issue:46877", + "issue:47000", + "issue:47059", + "issue:47110", + "issue:47317", + "issue:47321", + "issue:47332", + "issue:47337" + ], + "soft_pairs": [ + "issue:38617|issue:43502", + "issue:45276|issue:46077", + "issue:44625|issue:46365", + "issue:43299|issue:43867", + "issue:44655|issue:47321", + "issue:44655|issue:47332", + "issue:43618|issue:44380", + "issue:43611|issue:44655", + "issue:44625|issue:46077", + "issue:36683|issue:43475", + "issue:43352|issue:45593", + "issue:36683|issue:44077", + "issue:44355|issue:46877", + "issue:43976|issue:44938", + "issue:46077|issue:46512", + "issue:43867|issue:45216", + "issue:44263|issue:45440", + "issue:44938|issue:47059", + "issue:43618|issue:45216", + "issue:36683|issue:47110", + "issue:36683|issue:44625", + "issue:29870|issue:36296", + "issue:44355|issue:44855", + "issue:44908|issue:46851", + "issue:43010|issue:45278", + "issue:43867|issue:46202", + "issue:43992|issue:47317", + "issue:44075|issue:46365", + "issue:29870|issue:33453", + "issue:43992|issue:45310", + "issue:43064|issue:44492", + "issue:44393|issue:47000", + "issue:44393|issue:47337" + ], + "prepared_review_unit_hash": "7359977461f904586429252bd3e3ac95692d23e33d10935df91b58870a39c40b", + "component_budget": { + "node_count": 530, + "item_count": 530, + "soft_pair_count": 1954, + "serialized_chars": 1013004, + "estimated_input_tokens": 253251, + "estimated_eval_tokens": 506758 + }, + "budget": { + "node_count": 47, + "item_count": 47, + "soft_pair_count": 33, + "serialized_chars": 36591, + "estimated_input_tokens": 9148, + "estimated_eval_tokens": 18552 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": true, + "analyst_result": { + "summary": "This cluster is heterogeneous: the items mostly share superficial themes like loading, config, or version regressions, but not the same underlying bug. I would not merge any of the soft pairs as duplicates; issue 36683 is only the most central item in the similarity graph, not a true canonical duplicate.", + "confidence": 0.18, + "canonical_issue_reason": "Issue 36683 is the most connected and discussed item among the soft-similarity links, so it is the best single anchor for the cluster, even though the nearby issues are not the same bug.", + "canonical_pr_reason": null, + "best_issue_reason": "36683 has the strongest cluster connectivity and is a concrete runtime failure, making it the best representative issue despite the lack of true duplicates.", + "best_pr_reason": null, + "soft_edge_verdicts": [ + { + "left": "issue:38617", + "right": "issue:43502", + "accept": false, + "reason": "Different bugs: import-time API breakage vs offline-loading network requests." + }, + { + "left": "issue:45276", + "right": "issue:46077", + "accept": false, + "reason": "Same model family, but different code paths and symptoms: embedding resize vs attention-mask construction." + }, + { + "left": "issue:44625", + "right": "issue:46365", + "accept": false, + "reason": "Unrelated failures: Qwen3.5 config propagation vs Trainer import conflict." + }, + { + "left": "issue:43299", + "right": "issue:43867", + "accept": false, + "reason": "Both are loading issues, but for different models and different failure modes." + }, + { + "left": "issue:44655", + "right": "issue:47321", + "accept": false, + "reason": "Different subsystems: pipeline serialization vs AMD quark class update state." + }, + { + "left": "issue:44655", + "right": "issue:47332", + "accept": false, + "reason": "Different save paths and root causes; offloading save is not the same as pipeline save_pretrained." + }, + { + "left": "issue:43618", + "right": "issue:44380", + "accept": false, + "reason": "Attention-related but not the same bug: CLIP output fields vs GPT-2 backend scaling." + }, + { + "left": "issue:43611", + "right": "issue:44655", + "accept": false, + "reason": "One is model loading with base_model_prefix; the other is pipeline saving." + }, + { + "left": "issue:44625", + "right": "issue:46077", + "accept": false, + "reason": "Different model configs and different behavior: label propagation vs attention masking." + }, + { + "left": "issue:36683", + "right": "issue:43475", + "accept": false, + "reason": "Both are attribute errors, but on different model outputs and different missing fields." + }, + { + "left": "issue:43352", + "right": "issue:45593", + "accept": false, + "reason": "Unrelated: flash-attention support check vs denoising auxiliary-loss logic." + }, + { + "left": "issue:36683", + "right": "issue:44077", + "accept": false, + "reason": "Different configuration bugs in different models; no shared code path." + }, + { + "left": "issue:44355", + "right": "issue:46877", + "accept": false, + "reason": "Both touch import-time behavior, but one is compiled-file errors and the other is docstring lint output." + }, + { + "left": "issue:43976", + "right": "issue:44938", + "accept": false, + "reason": "Both are Python-version compatibility reports, but for different versions and likely different causes." + }, + { + "left": "issue:46077", + "right": "issue:46512", + "accept": false, + "reason": "Different regression types: attention-mask construction vs stale config dtype after from_config." + }, + { + "left": "issue:43867", + "right": "issue:45216", + "accept": false, + "reason": "Both involve model checkpoint/loading, but the reported failures and affected models differ." + }, + { + "left": "issue:44263", + "right": "issue:45440", + "accept": false, + "reason": "Different MoE/model-implementation bugs with no concrete shared code path." + }, + { + "left": "issue:44938", + "right": "issue:47059", + "accept": false, + "reason": "Python 3.14 import failure vs CLI breakage from huggingface_hub API change." + }, + { + "left": "issue:43618", + "right": "issue:45216", + "accept": false, + "reason": "Different models and different regressions: missing attentions vs incorrect saved checkpoint." + }, + { + "left": "issue:36683", + "right": "issue:47110", + "accept": false, + "reason": "Both are attribute errors, but on unrelated objects and different APIs." + }, + { + "left": "issue:36683", + "right": "issue:44625", + "accept": false, + "reason": "Related only at a very high level; the missing/propagated field is different and model-specific." + }, + { + "left": "issue:29870", + "right": "issue:36296", + "accept": false, + "reason": "Completely different failures: generate() mask bug vs tensor-parallel training issue." + }, + { + "left": "issue:44355", + "right": "issue:44855", + "accept": false, + "reason": "Both are Python/import related, but one is compiled-file handling and the other is torch.jit parsing." + }, + { + "left": "issue:44908", + "right": "issue:46851", + "accept": false, + "reason": "Different features: scheduler kwargs handling vs variable-length position_ids in Qwen3.5." + }, + { + "left": "issue:43010", + "right": "issue:45278", + "accept": false, + "reason": "One is a cache no_grad proposal; the other is a broad import regression report." + }, + { + "left": "issue:43867", + "right": "issue:46202", + "accept": false, + "reason": "Different issue domains: state_dict loading vs tokenizer roundtrip failure." + }, + { + "left": "issue:43992", + "right": "issue:47317", + "accept": false, + "reason": "Unrelated: missing encoder weight in from_pretrained vs a training-loss bug." + }, + { + "left": "issue:44075", + "right": "issue:46365", + "accept": false, + "reason": "Optimizer argument handling is unrelated to Trainer import failure." + }, + { + "left": "issue:29870", + "right": "issue:33453", + "accept": false, + "reason": "Different bugs in different subsystems: generation mask error vs tokenizer loading regression." + }, + { + "left": "issue:43992", + "right": "issue:45310", + "accept": false, + "reason": "Both are model-loading regressions, but for different models and likely different missing-key causes." + }, + { + "left": "issue:43064", + "right": "issue:44492", + "accept": false, + "reason": "Training/optimizer state bug vs a cache-strategy typo; no shared root cause." + }, + { + "left": "issue:44393", + "right": "issue:47000", + "accept": false, + "reason": "Different vision tasks and failure modes: bbox hallucination/output vs RTDetr loss infeasibility." + }, + { + "left": "issue:44393", + "right": "issue:47337", + "accept": false, + "reason": "Both are vision-related, but one is model output error and the other is a flaky test." + } + ] + }, + "evaluator_result": null, + "evaluator_used": false, + "retried": false, + "accepted_nontrivial_soft_edge": false, + "error_kind": null, + "error_message": null + }, + { + "label": "issue", + "component_index": 1, + "component_count": 28, + "review_unit_index": 61, + "review_unit_count": 65, + "status": "reviewed", + "reason": null, + "source": "llm", + "cache_hit": false, + "model": "gpt-5.4-mini?service_tier=flex", + "cluster_id": "cluster-10000-530", + "nodes": [ + "issue:28282", + "issue:30333", + "issue:33453", + "issue:36296", + "issue:41669", + "issue:42548", + "issue:42890", + "issue:42898", + "issue:43010", + "issue:43065", + "issue:43502", + "issue:43525", + "issue:43575", + "issue:43576", + "issue:43618", + "issue:43723", + "issue:43824", + "issue:43867", + "issue:43874", + "issue:43976", + "issue:44062", + "issue:44077", + "issue:44112", + "issue:44222", + "issue:44246", + "issue:44351", + "issue:44355", + "issue:44360", + "issue:44387", + "issue:44485", + "issue:44556", + "issue:44734", + "issue:44749", + "issue:44964", + "issue:44977", + "issue:45071", + "issue:45125", + "issue:45357", + "issue:45375", + "issue:45440", + "issue:45478", + "issue:45561", + "issue:45850", + "issue:46132", + "issue:46617", + "issue:46664", + "issue:47110", + "issue:47317" + ], + "soft_pairs": [ + "issue:44222|issue:45125", + "issue:36296|issue:43575", + "issue:41669|issue:44246", + "issue:43618|issue:44077", + "issue:43867|issue:45375", + "issue:43010|issue:43976", + "issue:42890|issue:45561", + "issue:43867|issue:45357", + "issue:28282|issue:33453", + "issue:43010|issue:45850", + "issue:42548|issue:44351", + "issue:45125|issue:45478", + "issue:43618|issue:44977", + "issue:45125|issue:45375", + "issue:44360|issue:45440", + "issue:43502|issue:44062", + "issue:43525|issue:47317", + "issue:42890|issue:44112", + "issue:42890|issue:44964", + "issue:44734|issue:46617", + "issue:44387|issue:44556", + "issue:43723|issue:44556", + "issue:42898|issue:46664", + "issue:43824|issue:46617", + "issue:43010|issue:43065", + "issue:45071|issue:47317", + "issue:43576|issue:46132", + "issue:43502|issue:44485", + "issue:44355|issue:44749", + "issue:30333|issue:33453", + "issue:45850|issue:47110", + "issue:43874|issue:46617" + ], + "prepared_review_unit_hash": "723900a7f123346185b2cbd9f77f932e7cd4397581cf015b2f219f0db7269d7c", + "component_budget": { + "node_count": 530, + "item_count": 530, + "soft_pair_count": 1954, + "serialized_chars": 1013004, + "estimated_input_tokens": 253251, + "estimated_eval_tokens": 506758 + }, + "budget": { + "node_count": 48, + "item_count": 48, + "soft_pair_count": 32, + "serialized_chars": 36789, + "estimated_input_tokens": 9198, + "estimated_eval_tokens": 18652 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": true, + "analyst_result": { + "summary": "This is a noisy, heterogeneous issue cluster: a few items are broadly import/loading/performance themed, but most soft pairs point to different failures and different code paths. I would not merge any of the proposed soft edges; the best representative issue is a broad import-slowdown report, but only weakly.", + "confidence": 0.82, + "canonical_issue_reason": "Issue 44246 is the most generic user-facing symptom in the cluster and best matches the broad import/performance theme, though the overall cluster is not a clean duplicate set.", + "canonical_pr_reason": null, + "best_issue_reason": "44246 is the most searchable, broadly phrased issue here and can serve as the least-bad anchor for this mixed set, even if it does not truly subsume the others.", + "best_pr_reason": null, + "soft_edge_verdicts": [ + { + "left": "issue:44222", + "right": "issue:45125", + "accept": false, + "reason": "Different problems: FP8 save_pretrained behavior vs missing tensor-parallel plan for Qwen3.5 MoE." + }, + { + "left": "issue:36296", + "right": "issue:43575", + "accept": false, + "reason": "Both mention tensor parallelism, but one is a generic training bug and the other is an OOM while loading a specific model." + }, + { + "left": "issue:41669", + "right": "issue:44246", + "accept": false, + "reason": "Related to import cost, but one is a refactor removing import * while the other is an intermittent import-slowdown report." + }, + { + "left": "issue:43618", + "right": "issue:44077", + "accept": false, + "reason": "Unrelated: CLIP attentions assignment vs a patchtsmixer post_init policy issue." + }, + { + "left": "issue:43867", + "right": "issue:45375", + "accept": false, + "reason": "Different serialization/config issues: sorted state_dict load failure vs missing vision config field under strict parsing." + }, + { + "left": "issue:43010", + "right": "issue:43976", + "accept": false, + "reason": "No shared bug path: no_grad decoration for cache updates vs Python version compatibility." + }, + { + "left": "issue:42890", + "right": "issue:45561", + "accept": false, + "reason": "Both are test-related, but one is seed-related flakiness and the other is an xdist file race." + }, + { + "left": "issue:43867", + "right": "issue:45357", + "accept": false, + "reason": "Load-time state_dict ordering bug vs save_pretrained writing wrong visual encoder keys." + }, + { + "left": "issue:28282", + "right": "issue:33453", + "accept": false, + "reason": "Different import/model loading failures: missing PyTorch backend vs tokenizer-loading regression." + }, + { + "left": "issue:43010", + "right": "issue:45850", + "accept": false, + "reason": "A specific cache-update correctness request is not the same as a broad 'many models break' complaint." + }, + { + "left": "issue:42548", + "right": "issue:44351", + "accept": false, + "reason": "Both are import errors, but for different symbols and likely different underlying causes." + }, + { + "left": "issue:45125", + "right": "issue:45478", + "accept": false, + "reason": "The latter is a general Qwen3.5 MoE from_pretrained regression; no clear evidence it is the same as the missing _tp_plan issue." + }, + { + "left": "issue:43618", + "right": "issue:44977", + "accept": false, + "reason": "Different models and failures: CLIPOutput metadata vs Qwen3.5 flash-attention generation." + }, + { + "left": "issue:45125", + "right": "issue:45375", + "accept": false, + "reason": "Both are Qwen3.5-related, but one is tensor-parallel support and the other is a strict vision-config field omission." + }, + { + "left": "issue:44360", + "right": "issue:45440", + "accept": false, + "reason": "Completely different subsystems: DSA indexing vs DeepSeekV3MoE implementation divergence." + }, + { + "left": "issue:43502", + "right": "issue:44062", + "accept": false, + "reason": "Unrelated bugs: offline loading still makes requests vs AddedToken keyword duplication." + }, + { + "left": "issue:43525", + "right": "issue:47317", + "accept": false, + "reason": "Different model families and symptoms: missing pad_token_id vs a training-loss bug." + }, + { + "left": "issue:42890", + "right": "issue:44112", + "accept": false, + "reason": "Both are CI/test problems, but the causes are different and not the same failing test path." + }, + { + "left": "issue:42890", + "right": "issue:44964", + "accept": false, + "reason": "Seed-related test instability is not the same as a model-loading failure for Phi-4 multimodal." + }, + { + "left": "issue:44734", + "right": "issue:46617", + "accept": false, + "reason": "Serve-time KV-cache indexing crash vs trust_remote_code cache FileNotFoundError are different loading/runtime paths." + }, + { + "left": "issue:44387", + "right": "issue:44556", + "accept": false, + "reason": "Memory regression under int4 quantization is not the same as checkpoint incompatibility after upgrade." + }, + { + "left": "issue:43723", + "right": "issue:44556", + "accept": false, + "reason": "Tokenizer-loading regression and checkpoint reload incompatibility are different failures." + }, + { + "left": "issue:45850", + "right": "issue:47110", + "accept": false, + "reason": "A broad complaint that many models break is not the same as a specific special_tokens type error." + }, + { + "left": "issue:43874", + "right": "issue:46617", + "accept": false, + "reason": "Different failures: multimodal token counting helper missing vs trust_remote_code cache file lookup." + }, + { + "left": "issue:43824", + "right": "issue:46617", + "accept": false, + "reason": "ImportError for a model class and a cache FileNotFoundError are not the same bug." + }, + { + "left": "issue:43010", + "right": "issue:43065", + "accept": false, + "reason": "No shared code path: no_grad on cache updates vs a dummy Conv2d in Sam3PixelDecoder." + }, + { + "left": "issue:45071", + "right": "issue:47317", + "accept": false, + "reason": "Type-checking regression in PretrainedConfig is unrelated to a model training-loss bug." + }, + { + "left": "issue:43576", + "right": "issue:46132", + "accept": false, + "reason": "Both are API-behavior complaints, but one is a CLI env command and the other is attention registration semantics." + }, + { + "left": "issue:43502", + "right": "issue:44485", + "accept": false, + "reason": "Offline-loading network requests and GLM-5 RoPE implementation are unrelated." + }, + { + "left": "issue:44355", + "right": "issue:44749", + "accept": false, + "reason": "Compiled-Python-file errors and a reported upgrade slowdown are not the same bug." + }, + { + "left": "issue:30333", + "right": "issue:33453", + "accept": false, + "reason": "MLFlow reporting on failed training jobs is unrelated to tokenizer-loading regression." + } + ] + }, + "evaluator_result": null, + "evaluator_used": false, + "retried": false, + "accepted_nontrivial_soft_edge": false, + "error_kind": null, + "error_message": null + }, + { + "label": "issue", + "component_index": 1, + "component_count": 28, + "review_unit_index": 62, + "review_unit_count": 65, + "status": "reviewed", + "reason": null, + "source": "llm", + "cache_hit": false, + "model": "gpt-5.4-mini?service_tier=flex", + "cluster_id": "cluster-10000-530", + "nodes": [ + "issue:29870", + "issue:30333", + "issue:30990", + "issue:31795", + "issue:33357", + "issue:33453", + "issue:36296", + "issue:38617", + "issue:41628", + "issue:42548", + "issue:42757", + "issue:43010", + "issue:43232", + "issue:43299", + "issue:43352", + "issue:43475", + "issue:43502", + "issue:43525", + "issue:43576", + "issue:43673", + "issue:43824", + "issue:43827", + "issue:43901", + "issue:44038", + "issue:44222", + "issue:44230", + "issue:44315", + "issue:44351", + "issue:44355", + "issue:44360", + "issue:44485", + "issue:44492", + "issue:44556", + "issue:44961", + "issue:45071", + "issue:45125", + "issue:45362", + "issue:45375", + "issue:45850", + "issue:46095", + "issue:46143", + "issue:46144", + "issue:46209", + "issue:46617", + "issue:46664", + "issue:46726", + "issue:47059", + "issue:47110" + ], + "soft_pairs": [ + "issue:30990|issue:31795", + "issue:41628|issue:42548", + "issue:44222|issue:44230", + "issue:43010|issue:43232", + "issue:43299|issue:44556", + "issue:46143|issue:46144", + "issue:46664|issue:46726", + "issue:43502|issue:44222", + "issue:30333|issue:36296", + "issue:43827|issue:44961", + "issue:29870|issue:31795", + "issue:33357|issue:33453", + "issue:44038|issue:45125", + "issue:44355|issue:45071", + "issue:43502|issue:44961", + "issue:43576|issue:43824", + "issue:43576|issue:45071", + "issue:44230|issue:46095", + "issue:43576|issue:47059", + "issue:42757|issue:43576", + "issue:31795|issue:38617", + "issue:36296|issue:43901", + "issue:43475|issue:46617", + "issue:44351|issue:46617", + "issue:44315|issue:44360", + "issue:45362|issue:45850", + "issue:45375|issue:47110", + "issue:43525|issue:47110", + "issue:44485|issue:44961", + "issue:43673|issue:44492", + "issue:44315|issue:46209", + "issue:43352|issue:44961" + ], + "prepared_review_unit_hash": "303366f2afeb1755352f787840cec1775ec7a4b24594fbee01f4d15a4f8dd3fb", + "component_budget": { + "node_count": 530, + "item_count": 530, + "soft_pair_count": 1954, + "serialized_chars": 1013004, + "estimated_input_tokens": 253251, + "estimated_eval_tokens": 506758 + }, + "budget": { + "node_count": 48, + "item_count": 48, + "soft_pair_count": 32, + "serialized_chars": 36693, + "estimated_input_tokens": 9174, + "estimated_eval_tokens": 18604 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": true, + "analyst_result": { + "summary": "This cluster is mostly a loose mix of unrelated Transformers v5 regressions, import errors, model-specific bugs, and docs/tooling issues. None of the soft pairs look like true duplicates or mergeable into one fix.", + "confidence": 0.86, + "canonical_issue_reason": "issue:45071 is the most concrete representative of the cluster\u2019s broad v5 compatibility-regression theme: it names a specific breakage in PretrainedConfig type checking, unlike the more generic, model-specific, or docs-only reports.", + "canonical_pr_reason": null, + "best_issue_reason": "issue:45071 is the strongest standalone issue in the set because it is specific, version-scoped, and broadly indicative of the kind of v5 breakages seen across the cluster; the other issues are either narrower, less actionable, or clearly unrelated.", + "best_pr_reason": null, + "soft_edge_verdicts": [ + { + "left": "issue:30990", + "right": "issue:31795", + "accept": false, + "reason": "Loading hang vs documentation confusion; different failure mode and no shared code-path evidence." + }, + { + "left": "issue:41628", + "right": "issue:42548", + "accept": false, + "reason": "Both are import errors, but for different symbols and likely different causes." + }, + { + "left": "issue:44222", + "right": "issue:44230", + "accept": false, + "reason": "FP8 save/load bug vs FP8 inference support request; related area, but not the same concrete issue." + }, + { + "left": "issue:43010", + "right": "issue:43232", + "accept": false, + "reason": "Different generation/cache bugs with different triggers and code paths." + }, + { + "left": "issue:43299", + "right": "issue:44556", + "accept": false, + "reason": "Both are upgrade regressions, but one is Qwen3VL loading and the other is checkpoint reload compatibility." + }, + { + "left": "issue:46143", + "right": "issue:46144", + "accept": false, + "reason": "`**kwargs` forwarding bug vs attention-interface implementation issue; too different to be duplicates." + }, + { + "left": "issue:46664", + "right": "issue:46726", + "accept": false, + "reason": "Video processor semantics vs missing-vision-backend error; unrelated causes." + }, + { + "left": "issue:43502", + "right": "issue:44222", + "accept": false, + "reason": "Offline/local-files-only network leak vs FP8 moe save_pretrained support; different subsystems." + }, + { + "left": "issue:30333", + "right": "issue:36296", + "accept": false, + "reason": "MLFlow reporting on failed jobs vs tensor-parallel training bug; no overlap beyond training context." + }, + { + "left": "issue:43827", + "right": "issue:44961", + "accept": false, + "reason": "Docs referencing removed pipeline APIs vs an unrelated placeholder issue." + }, + { + "left": "issue:29870", + "right": "issue:31795", + "accept": false, + "reason": "Runtime UnboundLocalError in generate vs documentation ambiguity; not the same bug." + }, + { + "left": "issue:33357", + "right": "issue:33453", + "accept": false, + "reason": "MacOS CLIP bus error vs tokenizer loading regression; different model and symptom." + }, + { + "left": "issue:44038", + "right": "issue:45125", + "accept": false, + "reason": "General Qwen3-VL-Moe breakage vs missing tensor-parallel plan for Qwen3_5Moe; related family, but not one fix." + }, + { + "left": "issue:44355", + "right": "issue:45071", + "accept": false, + "reason": "Compiled Python file errors vs PretrainedConfig type checking regression; unrelated." + }, + { + "left": "issue:43502", + "right": "issue:44961", + "accept": false, + "reason": "Local-files-only network leak vs unrelated placeholder issue." + }, + { + "left": "issue:43576", + "right": "issue:43824", + "accept": false, + "reason": "Broken `transformers env` CLI vs Qwen2.5-VL import error; different tools and failures." + }, + { + "left": "issue:43576", + "right": "issue:45071", + "accept": false, + "reason": "CLI breakage and config type-checking regression are distinct problems." + }, + { + "left": "issue:44230", + "right": "issue:46095", + "accept": false, + "reason": "FP8 inference support request vs weight-conversion question; not the same underlying bug." + }, + { + "left": "issue:43576", + "right": "issue:47059", + "accept": false, + "reason": "Different CLI failures: `transformers env` vs hub completion attribute error." + }, + { + "left": "issue:42757", + "right": "issue:43576", + "accept": false, + "reason": "Importing `is_offline_mode` from hub vs broken transformers env command; unrelated." + }, + { + "left": "issue:31795", + "right": "issue:38617", + "accept": false, + "reason": "Docs confusion vs missing `layer_type_validation` import; different scope." + }, + { + "left": "issue:36296", + "right": "issue:43901", + "accept": false, + "reason": "Tensor-parallel training bug vs docs mismatch about `return_all_scores`; not duplicates." + }, + { + "left": "issue:43475", + "right": "issue:46617", + "accept": false, + "reason": "Missing SAM3 attribute vs cache file lookup error; different models and code paths." + }, + { + "left": "issue:44351", + "right": "issue:46617", + "accept": false, + "reason": "Import error for `HybridCache` vs cache FileNotFoundError; separate failures." + }, + { + "left": "issue:44315", + "right": "issue:44360", + "accept": false, + "reason": "Liger Kernel/model_init issue vs DSA indexer discussion; no shared bug." + }, + { + "left": "issue:45362", + "right": "issue:45850", + "accept": false, + "reason": "Specific Qwen3.5-35B chat crash vs generic 'v5 breaks too many models' complaint; not one concrete issue." + }, + { + "left": "issue:45375", + "right": "issue:47110", + "accept": false, + "reason": "Missing config field in Qwen3_5MoeVisionConfig vs special_tokens list-type error; different object and failure." + }, + { + "left": "issue:43525", + "right": "issue:47110", + "accept": false, + "reason": "Missing pad_token_id on Llama4Config vs special_tokens list attribute error; unrelated attribute mismatches." + }, + { + "left": "issue:44485", + "right": "issue:44961", + "accept": false, + "reason": "GLM-5 RoPE implementation discussion vs unrelated placeholder issue." + }, + { + "left": "issue:43673", + "right": "issue:44492", + "accept": false, + "reason": "Generation cache missing in chunked prefill vs cache-strategy typo; different severity and topic." + }, + { + "left": "issue:44315", + "right": "issue:46209", + "accept": false, + "reason": "Model_init/Liger Kernel application bug vs FP8 inference on B200; different code paths." + }, + { + "left": "issue:43352", + "right": "issue:44961", + "accept": false, + "reason": "Flash Attention support limitation vs unrelated placeholder issue." + } + ] + }, + "evaluator_result": null, + "evaluator_used": false, + "retried": false, + "accepted_nontrivial_soft_edge": false, + "error_kind": null, + "error_message": null + }, + { + "label": "issue", + "component_index": 1, + "component_count": 28, + "review_unit_index": 63, + "review_unit_count": 65, + "status": "reviewed", + "reason": null, + "source": "llm", + "cache_hit": false, + "model": "gpt-5.4-mini?service_tier=flex", + "cluster_id": "cluster-10000-530", + "nodes": [ + "issue:41669", + "issue:42371", + "issue:42890", + "issue:43298", + "issue:43408", + "issue:43525", + "issue:43572", + "issue:43576", + "issue:43673", + "issue:43867", + "issue:43976", + "issue:44077", + "issue:44246", + "issue:44273", + "issue:44295", + "issue:44297", + "issue:44315", + "issue:44351", + "issue:44393", + "issue:44466", + "issue:44485", + "issue:44492", + "issue:44556", + "issue:44623", + "issue:44655", + "issue:44704", + "issue:44861", + "issue:44908", + "issue:44961", + "issue:45003", + "issue:45071", + "issue:45092", + "issue:45230", + "issue:45341", + "issue:45381", + "issue:45468", + "issue:45542", + "issue:45753", + "issue:45850", + "issue:46018", + "issue:46095", + "issue:46172", + "issue:46218", + "issue:46328", + "issue:46664", + "issue:47059", + "issue:47110", + "issue:47176" + ], + "soft_pairs": [ + "issue:44961|issue:45468", + "issue:44273|issue:45542", + "issue:44077|issue:44861", + "issue:43867|issue:47110", + "issue:45381|issue:46664", + "issue:44623|issue:45230", + "issue:44273|issue:45003", + "issue:44485|issue:46664", + "issue:41669|issue:43673", + "issue:44466|issue:44556", + "issue:43525|issue:45850", + "issue:43572|issue:45850", + "issue:44393|issue:45753", + "issue:44393|issue:45071", + "issue:44556|issue:45092", + "issue:42371|issue:44492", + "issue:45850|issue:47059", + "issue:44246|issue:44273", + "issue:44655|issue:44861", + "issue:43408|issue:44315", + "issue:44492|issue:45850", + "issue:41669|issue:46218", + "issue:44492|issue:46328", + "issue:43976|issue:44393", + "issue:44295|issue:45230", + "issue:42890|issue:44908", + "issue:44297|issue:45230", + "issue:44704|issue:44908", + "issue:44908|issue:45071", + "issue:41669|issue:45341", + "issue:43576|issue:44908", + "issue:43576|issue:47176", + "issue:41669|issue:44351", + "issue:46018|issue:46172", + "issue:43976|issue:46218", + "issue:43976|issue:47176", + "issue:43298|issue:46095" + ], + "prepared_review_unit_hash": "1c048d3c78bd4e729cd55b83baa4d1c7488c02f60c6691b975f1a7e85930712c", + "component_budget": { + "node_count": 530, + "item_count": 530, + "soft_pair_count": 1954, + "serialized_chars": 1013004, + "estimated_input_tokens": 253251, + "estimated_eval_tokens": 506758 + }, + "budget": { + "node_count": 48, + "item_count": 48, + "soft_pair_count": 37, + "serialized_chars": 38147, + "estimated_input_tokens": 9537, + "estimated_eval_tokens": 19330 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": true, + "analyst_result": { + "summary": "This batch is mostly heterogeneous. The closest recurring theme is import/lazy-loading behavior, but the soft pairs are still too weakly connected to treat as duplicates; I would not merge any of them.", + "confidence": 0.34, + "canonical_issue_reason": "issue:44246 is the clearest and broadest actionable report in the set: a concrete import-time performance symptom that best represents the import/lazy-loading thread.", + "canonical_pr_reason": null, + "best_issue_reason": "issue:44246 is the best canonical issue candidate because it states a direct user-facing regression clearly and broadly enough to subsume related import-performance complaints.", + "best_pr_reason": null, + "soft_edge_verdicts": [ + { + "left": "issue:44961", + "right": "issue:45468", + "accept": false, + "reason": "Unrelated model-specific issues; no shared code path or symptom." + }, + { + "left": "issue:44273", + "right": "issue:45542", + "accept": false, + "reason": "Both are v5-era regressions, but one is lazy loading and the other is a TensorBoard/TF backend error." + }, + { + "left": "issue:44077", + "right": "issue:44861", + "accept": false, + "reason": "Different problems in config validation vs tied-weight key handling." + }, + { + "left": "issue:43867", + "right": "issue:47110", + "accept": false, + "reason": "Both are attribute errors, but on different objects and code paths." + }, + { + "left": "issue:45381", + "right": "issue:46664", + "accept": false, + "reason": "Both concern vision/video inputs, but the concrete bugs are different." + }, + { + "left": "issue:44623", + "right": "issue:45230", + "accept": false, + "reason": "One is a specific save_pretrained bug; the other is too vague to establish sameness." + }, + { + "left": "issue:44273", + "right": "issue:45003", + "accept": false, + "reason": "Related to loading/lazy import behavior, but the reports describe different failures." + }, + { + "left": "issue:44485", + "right": "issue:46664", + "accept": false, + "reason": "Different model features and different bug descriptions." + }, + { + "left": "issue:41669", + "right": "issue:43673", + "accept": false, + "reason": "Import-performance issue versus missing cache during chunked prefill; not the same bug." + }, + { + "left": "issue:44466", + "right": "issue:44556", + "accept": false, + "reason": "Both involve checkpoint serialization/loading, but the symptoms and failure modes differ." + }, + { + "left": "issue:43525", + "right": "issue:45850", + "accept": false, + "reason": "A single config attribute error is not the same as a broad complaint that many models break." + }, + { + "left": "issue:43572", + "right": "issue:45850", + "accept": false, + "reason": "StableLmConfig pad-token mismatch is a specific model bug, not the same as a general breakage report." + }, + { + "left": "issue:44393", + "right": "issue:45753", + "accept": false, + "reason": "Same general model family, but distinct symptoms and inputs." + }, + { + "left": "issue:44393", + "right": "issue:45071", + "accept": false, + "reason": "Model output bug versus a framework-wide type-checking regression." + }, + { + "left": "issue:44556", + "right": "issue:45092", + "accept": false, + "reason": "Both are upgrade/loading issues, but they affect different checkpoint and initialization paths." + }, + { + "left": "issue:42371", + "right": "issue:44492", + "accept": false, + "reason": "Completely different topics: TF32 settings versus cache-strategy typo." + }, + { + "left": "issue:45850", + "right": "issue:47059", + "accept": false, + "reason": "Generic v5 breakage versus a CLI compatibility failure with huggingface_hub." + }, + { + "left": "issue:44246", + "right": "issue:44273", + "accept": false, + "reason": "Potentially related to import/lazy loading, but one is performance and the other is a broader lazy-loading bug report." + }, + { + "left": "issue:44655", + "right": "issue:44861", + "accept": false, + "reason": "Pipeline save_pretrained failure is not the same as tied-weight key attribute handling." + }, + { + "left": "issue:43408", + "right": "issue:44315", + "accept": false, + "reason": "Different model subsystems and unrelated failure modes." + }, + { + "left": "issue:44492", + "right": "issue:45850", + "accept": false, + "reason": "A typo in cache-strategy naming is unrelated to broad v5 regressions." + }, + { + "left": "issue:41669", + "right": "issue:46218", + "accept": false, + "reason": "Both mention import-related behavior, but one is performance and the other is a missing import symbol." + }, + { + "left": "issue:44492", + "right": "issue:46328", + "accept": false, + "reason": "Typo in cache strategies versus a slow padding implementation; no shared bug." + }, + { + "left": "issue:43976", + "right": "issue:44393", + "accept": false, + "reason": "Version/platform compatibility issue versus a specific Qwen3-VL output bug." + }, + { + "left": "issue:44295", + "right": "issue:45230", + "accept": false, + "reason": "Buffer access bug is concrete; the other report is too generic to match it." + }, + { + "left": "issue:42890", + "right": "issue:44908", + "accept": false, + "reason": "Test seeding flakiness and scheduler kwargs handling are unrelated." + }, + { + "left": "issue:44297", + "right": "issue:45230", + "accept": false, + "reason": "Tokenizer save metadata mismatch is a specific bug, not a generic report." + }, + { + "left": "issue:44704", + "right": "issue:44908", + "accept": false, + "reason": "AutoProcessor kwargs forwarding and scheduler kwargs forwarding are unrelated code paths." + }, + { + "left": "issue:44908", + "right": "issue:45071", + "accept": false, + "reason": "Scheduler configuration bug versus framework type-checking regression." + }, + { + "left": "issue:41669", + "right": "issue:45341", + "accept": false, + "reason": "Import-time slowdown is unrelated to the testing_utils bug." + }, + { + "left": "issue:43576", + "right": "issue:44908", + "accept": false, + "reason": "CLI env command issue and scheduler kwargs issue are unrelated." + }, + { + "left": "issue:43576", + "right": "issue:47176", + "accept": false, + "reason": "Environment command breakage is not the same as a shell-command allowlist bypass." + }, + { + "left": "issue:41669", + "right": "issue:44351", + "accept": false, + "reason": "Import-time slowdown versus a missing HybridCache symbol; different symptoms and fixes." + }, + { + "left": "issue:46018", + "right": "issue:46172", + "accept": false, + "reason": "DeepSeek-V4 gating/architecture issue versus a tensor transpose bug." + }, + { + "left": "issue:43976", + "right": "issue:46218", + "accept": false, + "reason": "Python-version compatibility problem versus missing import name; not duplicates." + }, + { + "left": "issue:43976", + "right": "issue:47176", + "accept": false, + "reason": "Unrelated: interpreter/version support versus shell operator security behavior." + }, + { + "left": "issue:43298", + "right": "issue:46095", + "accept": false, + "reason": "Continuous batching support for audio models is not the same as a DeepSeek-V4 weight-conversion question." + } + ] + }, + "evaluator_result": null, + "evaluator_used": false, + "retried": false, + "accepted_nontrivial_soft_edge": false, + "error_kind": null, + "error_message": null + }, + { + "label": "issue", + "component_index": 1, + "component_count": 28, + "review_unit_index": 64, + "review_unit_count": 65, + "status": "reviewed", + "reason": null, + "source": "llm", + "cache_hit": false, + "model": "gpt-5.4-mini?service_tier=flex", + "cluster_id": "cluster-10000-530", + "nodes": [ + "issue:28348", + "issue:29870", + "issue:30333", + "issue:31515", + "issue:31795", + "issue:33453", + "issue:34689", + "issue:38617", + "issue:41669", + "issue:42371", + "issue:42913", + "issue:43066", + "issue:43352", + "issue:43502", + "issue:43519", + "issue:43576", + "issue:43606", + "issue:43618", + "issue:43704", + "issue:43976", + "issue:44230", + "issue:44263", + "issue:44273", + "issue:44315", + "issue:44351", + "issue:44360", + "issue:44485", + "issue:44492", + "issue:44623", + "issue:44829", + "issue:44861", + "issue:44908", + "issue:44998", + "issue:45230", + "issue:45250", + "issue:45290", + "issue:45310", + "issue:45341", + "issue:45468", + "issue:45656", + "issue:45850", + "issue:46218", + "issue:46655", + "issue:46894", + "issue:47059", + "issue:47110", + "issue:47176", + "issue:47177" + ], + "soft_pairs": [ + "issue:31795|issue:33453", + "issue:43519|issue:46655", + "issue:44315|issue:44829", + "issue:44485|issue:47176", + "issue:45230|issue:45656", + "issue:45468|issue:46894", + "issue:44360|issue:45468", + "issue:43976|issue:47177", + "issue:43576|issue:44273", + "issue:43519|issue:44485", + "issue:44485|issue:47177", + "issue:44861|issue:47110", + "issue:44273|issue:44908", + "issue:31795|issue:34689", + "issue:43519|issue:44908", + "issue:41669|issue:43704", + "issue:43576|issue:46218", + "issue:42913|issue:43066", + "issue:44485|issue:45468", + "issue:43704|issue:44485", + "issue:42371|issue:43704", + "issue:45850|issue:47177", + "issue:45850|issue:47176", + "issue:43352|issue:47177", + "issue:43606|issue:45468", + "issue:43502|issue:43519", + "issue:44861|issue:47059", + "issue:43519|issue:44623", + "issue:45850|issue:46218", + "issue:44263|issue:45468", + "issue:43704|issue:45341", + "issue:28348|issue:45250", + "issue:43352|issue:46218", + "issue:44351|issue:46218", + "issue:44861|issue:45290", + "issue:44230|issue:44623", + "issue:44908|issue:44998", + "issue:44998|issue:47177", + "issue:44998|issue:47176", + "issue:28348|issue:43352", + "issue:44492|issue:44998", + "issue:43618|issue:44998", + "issue:44230|issue:45310", + "issue:31515|issue:38617", + "issue:29870|issue:31515", + "issue:30333|issue:31515" + ], + "prepared_review_unit_hash": "43b8eb55a9400a7325edad12264dff861bc08163b673489dab97ae291391d965", + "component_budget": { + "node_count": 530, + "item_count": 530, + "soft_pair_count": 1954, + "serialized_chars": 1013004, + "estimated_input_tokens": 253251, + "estimated_eval_tokens": 506758 + }, + "budget": { + "node_count": 48, + "item_count": 48, + "soft_pair_count": 46, + "serialized_chars": 40661, + "estimated_input_tokens": 10166, + "estimated_eval_tokens": 20588 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": true, + "analyst_result": { + "summary": "Mostly unrelated Transformers issues; the only near-match is the tokenizer v4/v5 regression pair, but it is still too broad/specifc to merge safely. All soft edges should be rejected.", + "confidence": 0.74, + "canonical_issue_reason": "issue:42913 is the clearest broad issue in the set and the best representative of the tokenizer-v5 regression theme, though it is still not a safe duplicate match with issue:43066.", + "canonical_pr_reason": null, + "best_issue_reason": "issue:42913 has the most general, well-framed problem statement among the items and is the best issue to anchor this mini-cluster around.", + "best_pr_reason": null, + "soft_edge_verdicts": [ + { + "left": "issue:31795", + "right": "issue:33453", + "accept": false, + "reason": "Docs/forward-arg confusion vs tokenizer loading regression; different bugs." + }, + { + "left": "issue:43519", + "right": "issue:46655", + "accept": false, + "reason": "Both are processor math issues, but timestamp calculation and video padding are different code paths." + }, + { + "left": "issue:44315", + "right": "issue:44829", + "accept": false, + "reason": "model_init/Liger-kernel application bug vs flash_attention_3 training degeneration; unrelated." + }, + { + "left": "issue:44485", + "right": "issue:47176", + "accept": false, + "reason": "RoPE implementation discussion vs shell-command allowlist bypass; different domains." + }, + { + "left": "issue:45230", + "right": "issue:45656", + "accept": false, + "reason": "Generic bug report vs deepspeed optimizer-step double-call; no concrete overlap." + }, + { + "left": "issue:45468", + "right": "issue:46894", + "accept": false, + "reason": "Gemma audio positional encoding vs CohereASR training-loss double-shift; different model/code paths." + }, + { + "left": "issue:44360", + "right": "issue:45468", + "accept": false, + "reason": "DSA indexer/ReLU discussion vs audio positional encoding bug; unrelated." + }, + { + "left": "issue:43976", + "right": "issue:47177", + "accept": false, + "reason": "Python version compatibility vs tar/zip-slip RCE; unrelated." + }, + { + "left": "issue:43576", + "right": "issue:44273", + "accept": false, + "reason": "Broken env command vs lazy loading malfunction; different features." + }, + { + "left": "issue:43519", + "right": "issue:44485", + "accept": false, + "reason": "Timestamp math in Qwen3VL processor vs GLM-5 RoPE implementation; not the same bug." + }, + { + "left": "issue:44485", + "right": "issue:47177", + "accept": false, + "reason": "RoPE implementation vs archive-slip security issue; unrelated." + }, + { + "left": "issue:44861", + "right": "issue:47110", + "accept": false, + "reason": "Tied-weights helper crash vs special_tokens dict/list error; different failure modes." + }, + { + "left": "issue:44273", + "right": "issue:44908", + "accept": false, + "reason": "Lazy loading issue vs inverse_sqrt scheduler kwargs bug; unrelated." + }, + { + "left": "issue:31795", + "right": "issue:34689", + "accept": false, + "reason": "Forward signature documentation vs Llama 3.2 model loading regression; different issues." + }, + { + "left": "issue:43519", + "right": "issue:44908", + "accept": false, + "reason": "Processor timestamp bug vs scheduler kwargs omission; unrelated." + }, + { + "left": "issue:41669", + "right": "issue:43704", + "accept": false, + "reason": "import * refactor/perf issue vs VRAM leak in multi-threaded dataloading; unrelated." + }, + { + "left": "issue:43576", + "right": "issue:46218", + "accept": false, + "reason": "Env CLI breakage vs module import failure for Qwen3_5; different code paths." + }, + { + "left": "issue:42913", + "right": "issue:43066", + "accept": false, + "reason": "Both concern tokenizer-v5 behavior, but one is broad regression reporting and the other is a specific decoder-type bug; not the same concrete issue." + }, + { + "left": "issue:44485", + "right": "issue:45468", + "accept": false, + "reason": "GLM-5 RoPE implementation vs Gemma audio positional encoding; different models and bugs." + }, + { + "left": "issue:43704", + "right": "issue:44485", + "accept": false, + "reason": "VRAM leak in threaded dataloading vs RoPE implementation; unrelated." + }, + { + "left": "issue:42371", + "right": "issue:43704", + "accept": false, + "reason": "TF32 API/settings issue vs VRAM leak issue; unrelated." + }, + { + "left": "issue:45850", + "right": "issue:47177", + "accept": false, + "reason": "Broad model-breakage complaint vs archive-slip RCE; unrelated." + }, + { + "left": "issue:45850", + "right": "issue:47176", + "accept": false, + "reason": "Broad model-breakage complaint vs shell-operator allowlist bypass; unrelated." + }, + { + "left": "issue:43352", + "right": "issue:47177", + "accept": false, + "reason": "Flash Attention 2.0 support gap vs archive-slip RCE; unrelated." + }, + { + "left": "issue:43606", + "right": "issue:45468", + "accept": false, + "reason": "CPU offload device mismatch vs Gemma audio positional encoding; unrelated." + }, + { + "left": "issue:43502", + "right": "issue:43519", + "accept": false, + "reason": "local_files_only network leak vs timestamp calculation bug; different behaviors." + }, + { + "left": "issue:44861", + "right": "issue:47059", + "accept": false, + "reason": "_tied_weights_keys crash vs CLI completion API breakage; unrelated." + }, + { + "left": "issue:43519", + "right": "issue:44623", + "accept": false, + "reason": "Timestamp calculation vs processor.save_pretrained missing files; different processor bugs." + }, + { + "left": "issue:45850", + "right": "issue:46218", + "accept": false, + "reason": "Generic model-breakage complaint vs Qwen3_5 import failure; not the same bug." + }, + { + "left": "issue:44230", + "right": "issue:45468", + "accept": false, + "reason": "fp8 inference support request vs audio positional encoding bug; unrelated." + }, + { + "left": "issue:44263", + "right": "issue:45468", + "accept": false, + "reason": "torch.split return-value issue in indexer vs audio positional encoding; unrelated." + }, + { + "left": "issue:43704", + "right": "issue:45341", + "accept": false, + "reason": "VRAM leak vs a small testing_utils bug; unrelated." + }, + { + "left": "issue:28348", + "right": "issue:45250", + "accept": false, + "reason": "GPT2-specific Flash Attention 2.0 support request vs generic Flash Attention 2.0 issue; too vague to merge." + }, + { + "left": "issue:43352", + "right": "issue:46218", + "accept": false, + "reason": "Flash Attention support gap vs module import failure; unrelated." + }, + { + "left": "issue:44351", + "right": "issue:46218", + "accept": false, + "reason": "HybridCache import failure vs Qwen3_5 module import failure; different missing symbols." + }, + { + "left": "issue:44861", + "right": "issue:45290", + "accept": false, + "reason": "Tied-weight helper crash vs chat-template/tool-call crash; unrelated." + }, + { + "left": "issue:44230", + "right": "issue:44623", + "accept": false, + "reason": "fp8 support request vs missing save_pretrained files; unrelated." + }, + { + "left": "issue:44908", + "right": "issue:44998", + "accept": false, + "reason": "Scheduler kwargs bug vs unrelated low-information issue title." + }, + { + "left": "issue:44998", + "right": "issue:47177", + "accept": false, + "reason": "Unemployment vs archive-slip RCE; unrelated." + }, + { + "left": "issue:44998", + "right": "issue:47176", + "accept": false, + "reason": "Unemployment vs shell-command bypass; unrelated." + }, + { + "left": "issue:28348", + "right": "issue:43352", + "accept": false, + "reason": "Both mention Flash Attention 2.0, but one is GPT2-specific support and the other is NemotronH support; not the same concrete change." + }, + { + "left": "issue:44492", + "right": "issue:44998", + "accept": false, + "reason": "Cache-strategy typo vs unrelated low-information issue." + }, + { + "left": "issue:43618", + "right": "issue:44998", + "accept": false, + "reason": "CLIPOutput attentions regression vs unrelated low-information issue." + }, + { + "left": "issue:44230", + "right": "issue:45310", + "accept": false, + "reason": "fp8 support request vs Qwen3.5 from_pretrained error; different problems." + }, + { + "left": "issue:31515", + "right": "issue:38617", + "accept": false, + "reason": "Slow checkpoint loading vs missing layer_type_validation import; unrelated." + }, + { + "left": "issue:29870", + "right": "issue:31515", + "accept": false, + "reason": "BART generate decoder_attention_mask error vs slow from_pretrained loading; unrelated." + }, + { + "left": "issue:30333", + "right": "issue:31515", + "accept": false, + "reason": "MLFlow run-finalization bug vs slow checkpoint loading; unrelated." + } + ] + }, + "evaluator_result": null, + "evaluator_used": false, + "retried": false, + "accepted_nontrivial_soft_edge": false, + "error_kind": null, + "error_message": null + }, + { + "label": "issue", + "component_index": 1, + "component_count": 28, + "review_unit_index": 65, + "review_unit_count": 65, + "status": "reviewed", + "reason": null, + "source": "llm", + "cache_hit": false, + "model": "gpt-5.4-mini?service_tier=flex", + "cluster_id": "cluster-10000-530", + "nodes": [ + "issue:31515", + "issue:31795", + "issue:36296", + "issue:37934", + "issue:42936", + "issue:43240", + "issue:43298", + "issue:43668", + "issue:43717", + "issue:43813", + "issue:44031", + "issue:44230", + "issue:46018", + "issue:46144" + ], + "soft_pairs": [ + "issue:31515|issue:36296", + "issue:43298|issue:44230", + "issue:31515|issue:31795", + "issue:43717|issue:46144", + "issue:42936|issue:46018", + "issue:43668|issue:44031", + "issue:42936|issue:43240", + "issue:43240|issue:46144", + "issue:42936|issue:44031", + "issue:42936|issue:46144", + "issue:42936|issue:43813", + "issue:43668|issue:43813", + "issue:43240|issue:43668", + "issue:37934|issue:43668", + "issue:42936|issue:43668" + ], + "prepared_review_unit_hash": "ff5ba67badca7ed56f5b5c877be2abca170aec350ceb534f9d66ab5d51192e85", + "component_budget": { + "node_count": 530, + "item_count": 530, + "soft_pair_count": 1954, + "serialized_chars": 1013004, + "estimated_input_tokens": 253251, + "estimated_eval_tokens": 506758 + }, + "budget": { + "node_count": 14, + "item_count": 14, + "soft_pair_count": 15, + "serialized_chars": 12258, + "estimated_input_tokens": 3065, + "estimated_eval_tokens": 6386 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": true, + "analyst_result": { + "summary": "This set is highly heterogeneous: loading/performance, docs, config typing, tokenizer warnings, architecture quirks, and unrelated model bugs. None of the soft pairs look like the same underlying issue/change, so no merges are justified.", + "confidence": 0.9, + "canonical_issue_reason": "If one issue must represent the set, #46144 is the clearest actionable code-path bug and the only open issue; the rest are largely unrelated docs/typo/config/model-specific reports.", + "canonical_pr_reason": null, + "best_issue_reason": "#46144 is the most concrete and merge-worthy issue in the list: it points to an implementation problem in attention wiring rather than a doc typo or narrow configuration nit.", + "best_pr_reason": null, + "soft_edge_verdicts": [ + { + "left": "issue:31515", + "right": "issue:36296", + "accept": false, + "reason": "Unrelated symptoms: checkpoint loading speed vs tensor-parallel training correctness." + }, + { + "left": "issue:43298", + "right": "issue:44230", + "accept": false, + "reason": "Audio continuous batching support and fp8 inference support are different feature gaps." + }, + { + "left": "issue:31515", + "right": "issue:31795", + "accept": false, + "reason": "Performance/loading bug vs documentation confusion; not the same underlying problem." + }, + { + "left": "issue:43717", + "right": "issue:46144", + "accept": false, + "reason": "Mamba init-weight behavior is unrelated to RoFormer attention interface usage." + }, + { + "left": "issue:42936", + "right": "issue:46018", + "accept": false, + "reason": "Mask2Former ignore_value handling and DeepSeek shared-expert gating are different model bugs." + }, + { + "left": "issue:43668", + "right": "issue:44031", + "accept": false, + "reason": "Incorrect type hint in config vs tokenizer regex warning after a version change." + }, + { + "left": "issue:42936", + "right": "issue:43240", + "accept": false, + "reason": "Unused ignore_value and kwargs not forwarded to loss calculation are separate code paths." + }, + { + "left": "issue:43240", + "right": "issue:46144", + "accept": false, + "reason": "Loss-function kwargs forwarding is unrelated to RoFormer attention implementation details." + }, + { + "left": "issue:42936", + "right": "issue:44031", + "accept": false, + "reason": "Model loss handling and tokenizer warning behavior do not share the same bug." + }, + { + "left": "issue:42936", + "right": "issue:46144", + "accept": false, + "reason": "Different subsystems: segmentation loss config vs attention interface wiring." + }, + { + "left": "issue:42936", + "right": "issue:43813", + "accept": false, + "reason": "Mask2Former loss handling is unrelated to the quantization-operation typo." + }, + { + "left": "issue:43668", + "right": "issue:43813", + "accept": false, + "reason": "Both are small metadata/doc-style issues, but they concern different symbols and fixes." + }, + { + "left": "issue:43240", + "right": "issue:43668", + "accept": false, + "reason": "Loss kwargs propagation and config type-hint correctness are unrelated." + }, + { + "left": "issue:37934", + "right": "issue:43668", + "accept": false, + "reason": "Llama4 norm-implementation question is unrelated to a ModernBERT type hint." + }, + { + "left": "issue:42936", + "right": "issue:43668", + "accept": false, + "reason": "Mask2Former loss usage bug and ModernBERT config typing bug are distinct." + } + ] + }, + "evaluator_result": null, + "evaluator_used": false, + "retried": false, + "accepted_nontrivial_soft_edge": false, + "error_kind": null, + "error_message": null + }, + { + "label": "issue", + "component_index": 2, + "component_count": 28, + "review_unit_index": 1, + "review_unit_count": 1, + "status": "reviewed", + "reason": null, + "source": "llm", + "cache_hit": false, + "model": "gpt-5.4-mini?service_tier=flex", + "cluster_id": "cluster-10519-3", + "nodes": [ + "issue:10519", + "issue:11005", + "issue:11307" + ], + "soft_pairs": [ + "issue:10519|issue:11307", + "issue:10519|issue:11005", + "issue:11005|issue:11307" + ], + "prepared_review_unit_hash": "eec0657430f517550f3c0d745bce124452098fa7893dfcf361ab59b0b982370b", + "component_budget": { + "node_count": 3, + "item_count": 3, + "soft_pair_count": 3, + "serialized_chars": 2874, + "estimated_input_tokens": 719, + "estimated_eval_tokens": 1694 + }, + "budget": { + "node_count": 3, + "item_count": 3, + "soft_pair_count": 3, + "serialized_chars": 2874, + "estimated_input_tokens": 719, + "estimated_eval_tokens": 1694 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": false, + "analyst_result": { + "summary": "These three issues are unrelated: tokenizer truncation behavior, a ReduceLROnPlateau-like training feature, and Wav2Vec2 word boundary timestamps. None of the soft pairs describe the same underlying bug or change.", + "confidence": 0.98, + "canonical_issue_reason": "No single canonical issue fits this cluster because the issues cover three different feature requests in different parts of the library.", + "canonical_pr_reason": "No pull requests were provided.", + "best_issue_reason": "issue:10519 is the clearest, most concrete feature request and is the best representative issue if one must be chosen.", + "best_pr_reason": "No pull requests were provided.", + "soft_edge_verdicts": [ + { + "left": "issue:10519", + "right": "issue:11307", + "accept": false, + "reason": "Different domains and requests: tokenizer truncation control vs. Wav2Vec2 word timestamp extraction." + }, + { + "left": "issue:10519", + "right": "issue:11005", + "accept": false, + "reason": "Truncation direction for tokenization is unrelated to a ReduceLROnPlateau-like learning-rate scheduling feature." + }, + { + "left": "issue:11005", + "right": "issue:11307", + "accept": false, + "reason": "A learning-rate scheduler request and Wav2Vec2 word offset timestamps are unrelated issues." + } + ] + }, + "evaluator_result": null, + "evaluator_used": false, + "retried": false, + "accepted_nontrivial_soft_edge": false, + "error_kind": null, + "error_message": null + }, + { + "label": "issue", + "component_index": 3, + "component_count": 28, + "review_unit_index": 1, + "review_unit_count": 1, + "status": "reviewed", + "reason": null, + "source": "llm", + "cache_hit": false, + "model": "gpt-5.4-mini?service_tier=flex", + "cluster_id": "cluster-13244-3", + "nodes": [ + "issue:13244", + "issue:15980", + "issue:16618" + ], + "soft_pairs": [ + "issue:13244|issue:16618", + "issue:15980|issue:16618" + ], + "prepared_review_unit_hash": "0669f6e8c060e1b60675c739899f0111098ee18772844d6344f7c9e66866810f", + "component_budget": { + "node_count": 3, + "item_count": 3, + "soft_pair_count": 2, + "serialized_chars": 2478, + "estimated_input_tokens": 620, + "estimated_eval_tokens": 1496 + }, + "budget": { + "node_count": 3, + "item_count": 3, + "soft_pair_count": 2, + "serialized_chars": 2478, + "estimated_input_tokens": 620, + "estimated_eval_tokens": 1496 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": false, + "analyst_result": { + "summary": "The cluster is mostly a false-positive mix: one tokenization issue and two model-loading issues. The two model-loading reports are related at a broad subsystem level, but they do not clearly describe the same underlying bug.", + "confidence": 0.58, + "canonical_issue_reason": "Issue 16618 is the broadest and most central of the set, with the clearest model-loading failure wording and the highest discussion activity.", + "canonical_pr_reason": null, + "best_issue_reason": "16618 is the best representative issue for this cluster because it anchors the only apparent theme shared by multiple items: model loading failures.", + "best_pr_reason": null, + "soft_edge_verdicts": [ + { + "left": "issue:13244", + "right": "issue:16618", + "accept": false, + "reason": "Tapas tokenization versus a model-loading failure are different bugs and different code paths; the similarity is only superficial." + }, + { + "left": "issue:15980", + "right": "issue:16618", + "accept": false, + "reason": "Both concern model loading, but 15980 is specifically about the error message when downloading a private model without login, while 16618 is a generic 'can't load the model' report; not the same underlying issue." + } + ] + }, + "evaluator_result": null, + "evaluator_used": false, + "retried": false, + "accepted_nontrivial_soft_edge": false, + "error_kind": null, + "error_message": null + }, + { + "label": "issue", + "component_index": 4, + "component_count": 28, + "review_unit_index": 1, + "review_unit_count": 1, + "status": "reviewed", + "reason": null, + "source": "llm", + "cache_hit": false, + "model": "gpt-5.4-mini?service_tier=flex", + "cluster_id": "cluster-19631-3", + "nodes": [ + "issue:19631", + "issue:20362", + "issue:22372" + ], + "soft_pairs": [ + "issue:19631|issue:22372", + "issue:19631|issue:20362" + ], + "prepared_review_unit_hash": "0727ee12fa81236531394837f8f78dc03ae7f509fc72d716e83c161750b4bcf7", + "component_budget": { + "node_count": 3, + "item_count": 3, + "soft_pair_count": 2, + "serialized_chars": 2351, + "estimated_input_tokens": 588, + "estimated_eval_tokens": 1432 + }, + "budget": { + "node_count": 3, + "item_count": 3, + "soft_pair_count": 2, + "serialized_chars": 2351, + "estimated_input_tokens": 588, + "estimated_eval_tokens": 1432 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": false, + "analyst_result": { + "summary": "These are three distinct feature requests to add different models (EDSR/MDSR, FlexiBERT, Restormer). They share only a broad \u201cadd model\u201d theme and are not the same underlying issue.", + "confidence": 0.98, + "canonical_issue_reason": null, + "canonical_pr_reason": null, + "best_issue_reason": null, + "best_pr_reason": null, + "soft_edge_verdicts": [ + { + "left": "issue:19631", + "right": "issue:22372", + "accept": false, + "reason": "Different requested models in different domains: EDSR/MDSR are super-resolution networks, while Restormer is a separate transformer-based restoration model. No evidence they refer to the same implementation or code-path change." + }, + { + "left": "issue:19631", + "right": "issue:20362", + "accept": false, + "reason": "EDSR/MDSR and FlexiBERT are unrelated model additions. They do not describe the same bug or feature, only the same generic request type ('add a model')." + } + ] + }, + "evaluator_result": null, + "evaluator_used": false, + "retried": false, + "accepted_nontrivial_soft_edge": false, + "error_kind": null, + "error_message": null + }, + { + "label": "issue", + "component_index": 5, + "component_count": 28, + "review_unit_index": 1, + "review_unit_count": 1, + "status": "reviewed", + "reason": null, + "source": "llm", + "cache_hit": false, + "model": "gpt-5.4-mini?service_tier=flex", + "cluster_id": "cluster-26350-2", + "nodes": [ + "issue:26350", + "issue:33260" + ], + "soft_pairs": [ + "issue:26350|issue:33260" + ], + "prepared_review_unit_hash": "c8f851321ff55e74104f6d809151ae6b40021f75fb707228398384fb254f214d", + "component_budget": { + "node_count": 2, + "item_count": 2, + "soft_pair_count": 1, + "serialized_chars": 1589, + "estimated_input_tokens": 398, + "estimated_eval_tokens": 1052 + }, + "budget": { + "node_count": 2, + "item_count": 2, + "soft_pair_count": 1, + "serialized_chars": 1589, + "estimated_input_tokens": 398, + "estimated_eval_tokens": 1052 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": false, + "analyst_result": { + "summary": "These are unrelated feature-request issues: one for Flash Attention 2 support, the other for GGUF support. They share only a generic \"more architectures\" theme and should not be clustered as duplicates.", + "confidence": 0.98, + "canonical_issue_reason": "Issue #26350 is a reasonable representative only because it is the earlier and more heavily discussed request, but it is not a duplicate of #33260.", + "canonical_pr_reason": null, + "best_issue_reason": "#26350 is the better standalone issue to keep as the cluster representative due to earlier creation and higher discussion activity.", + "best_pr_reason": null, + "soft_edge_verdicts": [ + { + "left": "issue:26350", + "right": "issue:33260", + "accept": false, + "reason": "They request different capabilities in different areas (Flash Attention 2 vs GGUF) and do not describe the same underlying bug or change." + } + ] + }, + "evaluator_result": null, + "evaluator_used": false, + "retried": false, + "accepted_nontrivial_soft_edge": false, + "error_kind": null, + "error_message": null + }, + { + "label": "issue", + "component_index": 6, + "component_count": 28, + "review_unit_index": 1, + "review_unit_count": 1, + "status": "reviewed", + "reason": null, + "source": "llm", + "cache_hit": false, + "model": "gpt-5.4-mini?service_tier=flex", + "cluster_id": "cluster-27088-12", + "nodes": [ + "issue:27088", + "issue:28936", + "issue:30191", + "issue:30240", + "issue:38435", + "issue:43834", + "issue:44365", + "issue:44370", + "issue:46222", + "issue:46226", + "issue:46931", + "issue:47299" + ], + "soft_pairs": [ + "issue:27088|issue:44370", + "issue:27088|issue:44365", + "issue:27088|issue:43834", + "issue:27088|issue:30191", + "issue:27088|issue:47299", + "issue:28936|issue:43834", + "issue:27088|issue:28936", + "issue:28936|issue:44365", + "issue:28936|issue:30240", + "issue:28936|issue:38435", + "issue:30191|issue:30240", + "issue:28936|issue:30191", + "issue:30240|issue:44370", + "issue:30240|issue:44365", + "issue:30240|issue:43834", + "issue:30240|issue:47299", + "issue:27088|issue:30240", + "issue:38435|issue:44370", + "issue:38435|issue:44365", + "issue:38435|issue:43834", + "issue:38435|issue:47299", + "issue:38435|issue:46226", + "issue:43834|issue:44370", + "issue:43834|issue:44365", + "issue:43834|issue:47299", + "issue:43834|issue:46226", + "issue:43834|issue:46931", + "issue:44365|issue:44370", + "issue:44365|issue:47299", + "issue:44365|issue:46226", + "issue:44365|issue:46931", + "issue:44370|issue:47299", + "issue:44370|issue:46226", + "issue:44370|issue:46931", + "issue:46222|issue:47299", + "issue:44370|issue:46222", + "issue:44365|issue:46222", + "issue:43834|issue:46222", + "issue:46222|issue:46226", + "issue:46226|issue:46931", + "issue:46226|issue:47299", + "issue:46931|issue:47299" + ], + "prepared_review_unit_hash": "827aa168085b5677e0f0f29712f81a2a6ff346fff2758e8bb153bea9174737c8", + "component_budget": { + "node_count": 12, + "item_count": 12, + "soft_pair_count": 42, + "serialized_chars": 22276, + "estimated_input_tokens": 5569, + "estimated_eval_tokens": 11394 + }, + "budget": { + "node_count": 12, + "item_count": 12, + "soft_pair_count": 42, + "serialized_chars": 22276, + "estimated_input_tokens": 5569, + "estimated_eval_tokens": 11394 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": false, + "analyst_result": { + "summary": "This cluster is mostly composed of language-specific documentation translation requests, plus a few garbled/spam-like entries. The titles/body patterns are highly similar, but they do not indicate the same underlying request, so the soft links should not be merged as duplicates.", + "confidence": 0.88, + "canonical_issue_reason": "Issue 43834 is the best cluster exemplar because it is the generic translation-request template that most closely matches the repeated pattern across the set, even though the language-specific requests still appear distinct.", + "canonical_pr_reason": null, + "best_issue_reason": "Issue 43834 is the clearest representative of the cluster's repeated template and is the most central example of the translation-request pattern.", + "best_pr_reason": null, + "soft_edge_verdicts": [ + { + "left": "issue:27088", + "right": "issue:44370", + "accept": false, + "reason": "Different translation requests; no evidence they target the same language or same underlying change." + }, + { + "left": "issue:27088", + "right": "issue:44365", + "accept": false, + "reason": "Different translation requests; same template style is not enough to treat them as duplicates." + }, + { + "left": "issue:27088", + "right": "issue:43834", + "accept": false, + "reason": "Concrete Turkish request vs generic template issue; not the same underlying request." + }, + { + "left": "issue:27088", + "right": "issue:30191", + "accept": false, + "reason": "Turkish vs Persian translation request; distinct language targets." + }, + { + "left": "issue:27088", + "right": "issue:47299", + "accept": false, + "reason": "The latter looks garbled/spam-like, not the same docs translation request." + }, + { + "left": "issue:28936", + "right": "issue:43834", + "accept": false, + "reason": "Spanish request vs generic template placeholder; likely separate language-specific issue." + }, + { + "left": "issue:27088", + "right": "issue:28936", + "accept": false, + "reason": "Turkish vs Spanish translation request; distinct targets, not duplicates." + }, + { + "left": "issue:28936", + "right": "issue:44365", + "accept": false, + "reason": "Different language requests; similar wording is just the shared template." + }, + { + "left": "issue:28936", + "right": "issue:30240", + "accept": false, + "reason": "Spanish vs Polish translation request; different underlying requests." + }, + { + "left": "issue:28936", + "right": "issue:38435", + "accept": false, + "reason": "Spanish vs Romanian translation request; distinct language targets." + }, + { + "left": "issue:28936", + "right": "issue:30191", + "accept": false, + "reason": "Spanish vs Persian translation request; not the same issue." + }, + { + "left": "issue:30191", + "right": "issue:30240", + "accept": false, + "reason": "Persian vs Polish translation request; same template, different concrete request." + }, + { + "left": "issue:30240", + "right": "issue:44370", + "accept": false, + "reason": "Polish request vs generic placeholder issue; no concrete shared target." + }, + { + "left": "issue:30240", + "right": "issue:44365", + "accept": false, + "reason": "Same translation template, but not enough evidence of the same language/request." + }, + { + "left": "issue:30240", + "right": "issue:43834", + "accept": false, + "reason": "Polish request vs generic template issue; not the same concrete change." + }, + { + "left": "issue:30240", + "right": "issue:47299", + "accept": false, + "reason": "The latter is garbled/spam-like and does not match the Polish translation request." + }, + { + "left": "issue:27088", + "right": "issue:30240", + "accept": false, + "reason": "Turkish vs Polish translation request; separate language targets." + }, + { + "left": "issue:38435", + "right": "issue:44370", + "accept": false, + "reason": "Romanian request vs generic template placeholder; no concrete duplicate evidence." + }, + { + "left": "issue:38435", + "right": "issue:44365", + "accept": false, + "reason": "Romanian request vs generic template placeholder; likely separate issues." + }, + { + "left": "issue:38435", + "right": "issue:43834", + "accept": false, + "reason": "Romanian request vs generic template issue; same theme, different concrete request." + }, + { + "left": "issue:38435", + "right": "issue:47299", + "accept": false, + "reason": "Garbled/spam-like issue does not look like the Romanian translation request." + }, + { + "left": "issue:38435", + "right": "issue:46226", + "accept": false, + "reason": "Romanian translation request vs garbled Thai text; unrelated underlying content." + }, + { + "left": "issue:43834", + "right": "issue:44370", + "accept": false, + "reason": "Both are template-like translation requests, but there is no evidence they are the same language-specific issue." + }, + { + "left": "issue:43834", + "right": "issue:44365", + "accept": false, + "reason": "Near-identical template text, but no concrete shared target language or request." + }, + { + "left": "issue:43834", + "right": "issue:47299", + "accept": false, + "reason": "Generic translation template vs spam/garbled title; not the same request." + }, + { + "left": "issue:43834", + "right": "issue:46226", + "accept": false, + "reason": "Template-like translation issue vs garbled Thai content; no shared concrete bug/change." + }, + { + "left": "issue:43834", + "right": "issue:46931", + "accept": false, + "reason": "Template-like translation issue vs unrelated Thai text; not duplicates." + }, + { + "left": "issue:44365", + "right": "issue:44370", + "accept": false, + "reason": "Titles/body look templated, but no explicit target language or concrete identical request." + }, + { + "left": "issue:44365", + "right": "issue:47299", + "accept": false, + "reason": "Generic translation template vs spam/garbled content; different underlying issue." + }, + { + "left": "issue:44365", + "right": "issue:46226", + "accept": false, + "reason": "Template-like translation issue vs garbled Thai text; no shared concrete request." + }, + { + "left": "issue:44365", + "right": "issue:46931", + "accept": false, + "reason": "Same template family, but the Thai text is unrelated to docs translation." + }, + { + "left": "issue:44370", + "right": "issue:47299", + "accept": false, + "reason": "Generic translation template vs spam/garbled content; not the same issue." + }, + { + "left": "issue:44370", + "right": "issue:46226", + "accept": false, + "reason": "Template-like translation issue vs garbled Thai content; no concrete overlap." + }, + { + "left": "issue:44370", + "right": "issue:46931", + "accept": false, + "reason": "Template-like translation issue vs unrelated Thai text; not duplicates." + }, + { + "left": "issue:46222", + "right": "issue:47299", + "accept": false, + "reason": "Both are garbled/spam-like, but there is no evidence they are the same concrete request." + }, + { + "left": "issue:44370", + "right": "issue:46222", + "accept": false, + "reason": "Generic translation template vs garbled Thai content; different underlying issues." + }, + { + "left": "issue:44365", + "right": "issue:46222", + "accept": false, + "reason": "Generic translation template vs garbled Thai content; not the same request." + }, + { + "left": "issue:43834", + "right": "issue:46222", + "accept": false, + "reason": "Template-like translation issue vs garbled Thai content; no shared concrete target." + }, + { + "left": "issue:46222", + "right": "issue:46226", + "accept": false, + "reason": "Both are garbled Thai text, but there is no concrete evidence they describe the same change or bug." + }, + { + "left": "issue:46226", + "right": "issue:46931", + "accept": false, + "reason": "Garbled Thai content in both, but not enough evidence of the same underlying request." + }, + { + "left": "issue:46226", + "right": "issue:47299", + "accept": false, + "reason": "Garbled/spam-like entries only; no clear same concrete issue." + }, + { + "left": "issue:46931", + "right": "issue:47299", + "accept": false, + "reason": "Unrelated garbled titles; no evidence of the same underlying bug or change." + } + ] + }, + "evaluator_result": null, + "evaluator_used": false, + "retried": false, + "accepted_nontrivial_soft_edge": false, + "error_kind": null, + "error_message": null + }, + { + "label": "issue", + "component_index": 7, + "component_count": 28, + "review_unit_index": 1, + "review_unit_count": 1, + "status": "reviewed", + "reason": null, + "source": "llm", + "cache_hit": false, + "model": "gpt-5.4-mini?service_tier=flex", + "cluster_id": "cluster-36718-22", + "nodes": [ + "issue:36718", + "issue:36817", + "issue:36879", + "issue:37015", + "issue:37334", + "issue:41115", + "issue:41211", + "issue:42503", + "issue:42738", + "issue:42954", + "issue:43671", + "issue:43837", + "issue:43909", + "issue:44121", + "issue:44205", + "issue:45027", + "issue:45177", + "issue:45322", + "issue:46133", + "issue:46645", + "issue:46653", + "issue:47004" + ], + "soft_pairs": [ + "issue:36718|issue:47004", + "issue:36718|issue:41115", + "issue:36817|issue:41211", + "issue:36817|issue:43671", + "issue:36817|issue:36879", + "issue:36817|issue:42954", + "issue:36879|issue:41211", + "issue:36879|issue:43909", + "issue:36879|issue:44121", + "issue:36879|issue:43837", + "issue:36879|issue:43671", + "issue:37015|issue:43837", + "issue:37015|issue:44121", + "issue:37015|issue:43909", + "issue:37015|issue:37334", + "issue:36879|issue:37015", + "issue:37334|issue:41115", + "issue:37334|issue:43837", + "issue:37334|issue:43671", + "issue:37334|issue:42503", + "issue:41115|issue:43837", + "issue:41115|issue:44121", + "issue:41115|issue:42503", + "issue:41211|issue:43671", + "issue:37015|issue:42503", + "issue:42503|issue:43837", + "issue:42503|issue:43671", + "issue:42738|issue:45027", + "issue:42503|issue:42738", + "issue:42738|issue:42954", + "issue:42738|issue:46653", + "issue:42954|issue:44205", + "issue:37334|issue:42954", + "issue:42954|issue:43671", + "issue:42503|issue:42954", + "issue:42954|issue:45177", + "issue:43671|issue:43837", + "issue:43671|issue:45027", + "issue:43671|issue:45177", + "issue:43671|issue:44121", + "issue:43837|issue:43909", + "issue:43837|issue:45322", + "issue:43837|issue:44121", + "issue:43909|issue:44205", + "issue:43909|issue:46645", + "issue:44121|issue:45322", + "issue:41115|issue:45027", + "issue:43837|issue:45027", + "issue:37015|issue:45027", + "issue:37334|issue:45027", + "issue:43837|issue:45177", + "issue:45322|issue:46133", + "issue:42954|issue:46653", + "issue:42503|issue:46653", + "issue:45027|issue:46653" + ], + "prepared_review_unit_hash": "8db3fe801a57badc5ad401c5cd878422fea1d89e64e367bb4065b03b2de8c90f", + "component_budget": { + "node_count": 22, + "item_count": 22, + "soft_pair_count": 55, + "serialized_chars": 31689, + "estimated_input_tokens": 7923, + "estimated_eval_tokens": 16102 + }, + "budget": { + "node_count": 22, + "item_count": 22, + "soft_pair_count": 55, + "serialized_chars": 31689, + "estimated_input_tokens": 7923, + "estimated_eval_tokens": 16102 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": false, + "analyst_result": { + "summary": "This is a loose topic cluster of unrelated model-request issues (vision, audio, language, detectors, matchers, and config additions). The similarities are mostly generic \u201cAdd model\u201d wording, so the soft edges should all be rejected.", + "confidence": 0.96, + "canonical_issue_reason": "No single canonical issue: these are different requested models and features across unrelated code paths and modalities, not duplicates of one underlying bug/change.", + "canonical_pr_reason": null, + "best_issue_reason": "If a representative issue is needed, 42503 is the closest to a generic model-add request, but it still does not cover the rest of the cluster well.", + "best_pr_reason": null, + "soft_edge_verdicts": [ + { + "left": "issue:36718", + "right": "issue:47004", + "accept": false, + "reason": "Different keypoint projects (RoMa matcher vs LoMa detector+matcher); same domain, but not the same requested change." + }, + { + "left": "issue:36718", + "right": "issue:41115", + "accept": false, + "reason": "RoMa matcher is a vision matching request; MiniCPM3 is a model architecture request in a different area." + }, + { + "left": "issue:36817", + "right": "issue:41211", + "accept": false, + "reason": "EuroBert config support and DEIMv2 are different model additions." + }, + { + "left": "issue:36817", + "right": "issue:43671", + "accept": false, + "reason": "EuroBert config support is unrelated to Qwen3-TTS speech support." + }, + { + "left": "issue:36817", + "right": "issue:36879", + "accept": false, + "reason": "EuroBert and RF-DETR are distinct models with different integration work." + }, + { + "left": "issue:36817", + "right": "issue:42954", + "accept": false, + "reason": "EuroBert config work is unrelated to XiaoMi MiMo." + }, + { + "left": "issue:36879", + "right": "issue:41211", + "accept": false, + "reason": "RF-DETR vision detection and DEIMv2 are separate model requests." + }, + { + "left": "issue:36879", + "right": "issue:43909", + "accept": false, + "reason": "RF-DETR and LFM2.5 Audio target different modalities and code paths." + }, + { + "left": "issue:36879", + "right": "issue:44121", + "accept": false, + "reason": "RF-DETR detection support is unrelated to the weight-sparse transformer request." + }, + { + "left": "issue:36879", + "right": "issue:43837", + "accept": false, + "reason": "RF-DETR vision detection and Qwen3-ASR speech recognition are different changes." + }, + { + "left": "issue:36879", + "right": "issue:43671", + "accept": false, + "reason": "RF-DETR and Qwen3-TTS are different tasks and model families." + }, + { + "left": "issue:37015", + "right": "issue:43837", + "accept": false, + "reason": "NeoBERT is a language model request; Qwen3-ASR is an audio transcription request." + }, + { + "left": "issue:37015", + "right": "issue:44121", + "accept": false, + "reason": "NeoBERT and the OpenAI weight-sparse transformer are unrelated model additions." + }, + { + "left": "issue:37015", + "right": "issue:43909", + "accept": false, + "reason": "NeoBERT and LFM2.5 Audio 1.5B are different model families in different modalities." + }, + { + "left": "issue:37015", + "right": "issue:37334", + "accept": false, + "reason": "NeoBERT and LLMDet are distinct requests; one is a BERT-like LM, the other a detector." + }, + { + "left": "issue:36879", + "right": "issue:37015", + "accept": false, + "reason": "RF-DETR detection support is not the same change as NeoBERT support." + }, + { + "left": "issue:37334", + "right": "issue:41115", + "accept": false, + "reason": "LLMDet support and MiniCPM3 architecture are separate model requests." + }, + { + "left": "issue:37334", + "right": "issue:43837", + "accept": false, + "reason": "LLMDet detection support is unrelated to Qwen3-ASR speech recognition." + }, + { + "left": "issue:37334", + "right": "issue:43671", + "accept": false, + "reason": "LLMDet and Qwen3-TTS are different tasks and model code paths." + }, + { + "left": "issue:37334", + "right": "issue:42503", + "accept": false, + "reason": "LLMDet and ModernVBERT are different model additions, not one duplicated request." + }, + { + "left": "issue:41115", + "right": "issue:43837", + "accept": false, + "reason": "MiniCPM3 architecture work and Qwen3-ASR are unrelated requests." + }, + { + "left": "issue:41115", + "right": "issue:44121", + "accept": false, + "reason": "MiniCPM3 and the weight-sparse transformer are different models and changes." + }, + { + "left": "issue:41115", + "right": "issue:42503", + "accept": false, + "reason": "MiniCPM3 support is not the same as adding ModernVBERT models." + }, + { + "left": "issue:41211", + "right": "issue:43671", + "accept": false, + "reason": "DEIMv2 detection support and Qwen3-TTS are unrelated." + }, + { + "left": "issue:37015", + "right": "issue:42503", + "accept": false, + "reason": "NeoBERT and ModernVBERT are different BERT-like model requests, not duplicates." + }, + { + "left": "issue:42503", + "right": "issue:43837", + "accept": false, + "reason": "ModernVBERT model addition and Qwen3-ASR speech recognition are unrelated." + }, + { + "left": "issue:42503", + "right": "issue:43671", + "accept": false, + "reason": "ModernVBERT and Qwen3-TTS are different model families and tasks." + }, + { + "left": "issue:42503", + "right": "issue:46653", + "accept": false, + "reason": "ModernVBERT model support has nothing to do with the Mirko Privitera issue." + }, + { + "left": "issue:42738", + "right": "issue:45027", + "accept": false, + "reason": "BERT-like models with RoPE and Voxtral TTS are different feature requests." + }, + { + "left": "issue:42503", + "right": "issue:42738", + "accept": false, + "reason": "ModernVBERT models and BERT-like models with RoPE overlap only in broad language-model wording, not the same change." + }, + { + "left": "issue:42738", + "right": "issue:42954", + "accept": false, + "reason": "BERT-like RoPE support and XiaoMi MiMo are separate model requests." + }, + { + "left": "issue:42738", + "right": "issue:46653", + "accept": false, + "reason": "BERT-like models with RoPE is unrelated to the Mirko Privitera issue." + }, + { + "left": "issue:42954", + "right": "issue:44205", + "accept": false, + "reason": "XiaoMi MiMo and SAM3-LiteText are different model additions." + }, + { + "left": "issue:37334", + "right": "issue:42954", + "accept": false, + "reason": "LLMDet and XiaoMi MiMo are different requested models and tasks." + }, + { + "left": "issue:42954", + "right": "issue:43671", + "accept": false, + "reason": "XiaoMi MiMo and Qwen3-TTS are unrelated model requests." + }, + { + "left": "issue:42503", + "right": "issue:42954", + "accept": false, + "reason": "ModernVBERT and XiaoMi MiMo are different models; only the request style is similar." + }, + { + "left": "issue:42954", + "right": "issue:45177", + "accept": false, + "reason": "XiaoMi MiMo and DeepSeek-OCR2 are different requests across different model types." + }, + { + "left": "issue:43671", + "right": "issue:43837", + "accept": false, + "reason": "Qwen3-TTS and Qwen3-ASR are related brands but different concrete tasks and implementations." + }, + { + "left": "issue:43671", + "right": "issue:45027", + "accept": false, + "reason": "Qwen3-TTS and Voxtral-4B-TTS-2603 are different TTS model requests." + }, + { + "left": "issue:43671", + "right": "issue:44121", + "accept": false, + "reason": "Qwen3-TTS support is unrelated to the weight-sparse transformer request." + }, + { + "left": "issue:43671", + "right": "issue:45322", + "accept": false, + "reason": "Qwen3-TTS and EUPE are different model families and integration work." + }, + { + "left": "issue:43837", + "right": "issue:43909", + "accept": false, + "reason": "Qwen3-ASR and LFM2.5 Audio are different audio model requests, not the same change." + }, + { + "left": "issue:43837", + "right": "issue:45322", + "accept": false, + "reason": "Qwen3-ASR and EUPE are unrelated." + }, + { + "left": "issue:43837", + "right": "issue:44121", + "accept": false, + "reason": "Qwen3-ASR support is not the same as the weight-sparse transformer change." + }, + { + "left": "issue:43909", + "right": "issue:44205", + "accept": false, + "reason": "LFM2.5 Audio and SAM3-LiteText are different modalities and code paths." + }, + { + "left": "issue:43909", + "right": "issue:46645", + "accept": false, + "reason": "LFM2.5 Audio support is unrelated to the Granite SWA+sink option." + }, + { + "left": "issue:44121", + "right": "issue:45322", + "accept": false, + "reason": "Weight-sparse transformer support and EUPE are distinct model requests." + }, + { + "left": "issue:41115", + "right": "issue:45027", + "accept": false, + "reason": "MiniCPM3 architecture work and Voxtral-4B-TTS support are unrelated." + }, + { + "left": "issue:43837", + "right": "issue:45027", + "accept": false, + "reason": "Qwen3-ASR and Voxtral TTS are different speech models and tasks." + }, + { + "left": "issue:37015", + "right": "issue:45027", + "accept": false, + "reason": "NeoBERT and Voxtral TTS are unrelated model requests." + }, + { + "left": "issue:37334", + "right": "issue:45027", + "accept": false, + "reason": "LLMDet detection support is unrelated to Voxtral TTS." + }, + { + "left": "issue:43837", + "right": "issue:45177", + "accept": false, + "reason": "Qwen3-ASR and DeepSeek-OCR2 are different modalities and changes." + }, + { + "left": "issue:45322", + "right": "issue:46133", + "accept": false, + "reason": "EUPE and TIPSv2 are different model additions from different authors and tasks." + }, + { + "left": "issue:42954", + "right": "issue:46653", + "accept": false, + "reason": "XiaoMi MiMo and Mirko Privitera are entirely different issues." + }, + { + "left": "issue:42503", + "right": "issue:46653", + "accept": false, + "reason": "ModernVBERT and Mirko Privitera are unrelated." + }, + { + "left": "issue:45027", + "right": "issue:46653", + "accept": false, + "reason": "Voxtral TTS support and Mirko Privitera are unrelated." + } + ] + }, + "evaluator_result": null, + "evaluator_used": false, + "retried": false, + "accepted_nontrivial_soft_edge": false, + "error_kind": null, + "error_message": null + }, + { + "label": "issue", + "component_index": 8, + "component_count": 28, + "review_unit_index": 1, + "review_unit_count": 1, + "status": "reviewed", + "reason": null, + "source": "llm", + "cache_hit": false, + "model": "gpt-5.4-mini?service_tier=flex", + "cluster_id": "cluster-38740-2", + "nodes": [ + "issue:38740", + "issue:46443" + ], + "soft_pairs": [ + "issue:38740|issue:46443" + ], + "prepared_review_unit_hash": "6de9d99d6b1eb5931d8959cfbffa4643ebf387d7140e30e1509507fe76f4fba2", + "component_budget": { + "node_count": 2, + "item_count": 2, + "soft_pair_count": 1, + "serialized_chars": 1505, + "estimated_input_tokens": 377, + "estimated_eval_tokens": 1010 + }, + "budget": { + "node_count": 2, + "item_count": 2, + "soft_pair_count": 1, + "serialized_chars": 1505, + "estimated_input_tokens": 377, + "estimated_eval_tokens": 1010 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": false, + "analyst_result": { + "summary": "The only pair is a weak text-similarity match, but the issues are unrelated: one is a docs request to add Pruna, the other is a Vision RoPE refactor. This is not a duplicate cluster.", + "confidence": 0.98, + "canonical_issue_reason": "No canonical issue: the docs request and the Vision RoPE refactor are different topics and do not describe the same bug or change.", + "canonical_pr_reason": null, + "best_issue_reason": "No global best issue: neither item can represent the cluster because they are unrelated; 46443 is the more technical task, but it is not a duplicate of 38740.", + "best_pr_reason": null, + "soft_edge_verdicts": [ + { + "left": "issue:38740", + "right": "issue:46443", + "accept": false, + "reason": "Different subjects and scopes: documentation about adding Pruna vs a Vision RoPE refactor. The similarity is only superficial, not the same underlying issue." + } + ] + }, + "evaluator_result": null, + "evaluator_used": false, + "retried": false, + "accepted_nontrivial_soft_edge": false, + "error_kind": null, + "error_message": null + }, + { + "label": "issue", + "component_index": 9, + "component_count": 28, + "review_unit_index": 1, + "review_unit_count": 1, + "status": "reviewed", + "reason": null, + "source": "llm", + "cache_hit": false, + "model": "gpt-5.4-mini?service_tier=flex", + "cluster_id": "cluster-41084-2", + "nodes": [ + "issue:41084", + "issue:43037" + ], + "soft_pairs": [ + "issue:41084|issue:43037" + ], + "prepared_review_unit_hash": "3708a8e6d459fd56de454109be97c96cb965650c75575db531343fbfb7f96193", + "component_budget": { + "node_count": 2, + "item_count": 2, + "soft_pair_count": 1, + "serialized_chars": 1479, + "estimated_input_tokens": 370, + "estimated_eval_tokens": 996 + }, + "budget": { + "node_count": 2, + "item_count": 2, + "soft_pair_count": 1, + "serialized_chars": 1479, + "estimated_input_tokens": 370, + "estimated_eval_tokens": 996 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": false, + "analyst_result": { + "summary": "The two issues are not duplicates: one is a generic request about block decoding, while the other is a model-specific support request for DeepSeek v3.2. They do not describe the same underlying bug or change.", + "confidence": 0.97, + "canonical_issue_reason": null, + "canonical_pr_reason": null, + "best_issue_reason": "Issue #43037 is the more specific and actionable request, but it is not a duplicate of #41084.", + "best_pr_reason": null, + "soft_edge_verdicts": [ + { + "left": "issue:41084", + "right": "issue:43037", + "accept": false, + "reason": "Different scope and intent: 'Set Block Decoding' is a generic decoding-related request, while 'DeepSeek v3.2 support' is about adding support for a specific model. No clear shared underlying bug/change." + } + ] + }, + "evaluator_result": null, + "evaluator_used": false, + "retried": false, + "accepted_nontrivial_soft_edge": false, + "error_kind": null, + "error_message": null + }, + { + "label": "issue", + "component_index": 10, + "component_count": 28, + "review_unit_index": 1, + "review_unit_count": 1, + "status": "reviewed", + "reason": null, + "source": "llm", + "cache_hit": false, + "model": "gpt-5.4-mini?service_tier=flex", + "cluster_id": "cluster-41306-2", + "nodes": [ + "issue:41306", + "issue:41826" + ], + "soft_pairs": [ + "issue:41306|issue:41826" + ], + "prepared_review_unit_hash": "6a7d7147d9e00f448b8b0906e3837f80a4cf2f892eb031271aec73f3bfcdfff0", + "component_budget": { + "node_count": 2, + "item_count": 2, + "soft_pair_count": 1, + "serialized_chars": 1585, + "estimated_input_tokens": 397, + "estimated_eval_tokens": 1050 + }, + "budget": { + "node_count": 2, + "item_count": 2, + "soft_pair_count": 1, + "serialized_chars": 1585, + "estimated_input_tokens": 397, + "estimated_eval_tokens": 1050 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": false, + "analyst_result": { + "summary": "These are not duplicates: one is a specific `from_pretrained()` flag request for TiledFusedLogitsLoss/FLCE, the other is a broader TiledMLP memory-footprint feature request. The shared \"Tiled\" theme is superficial.", + "confidence": 0.94, + "canonical_issue_reason": "Issue #41306 is the more concrete, narrowly scoped request tied to a specific code path (`from_pretrained()` behavior for TiledFusedLogitsLoss/FLCE).", + "canonical_pr_reason": null, + "best_issue_reason": "Issue #41306 is the best representative because it is more actionable and specific; #41826 is a separate broader enhancement about TiledMLP memory usage.", + "best_pr_reason": null, + "soft_edge_verdicts": [ + { + "left": "issue:41306", + "right": "issue:41826", + "accept": false, + "reason": "Different underlying requests: one is about a `from_pretrained()` flag for FLCE/TiledFusedLogitsLoss, the other is about integrating TiledMLP for lower memory use. Similar terminology is not enough to treat them as the same bug or change." + } + ] + }, + "evaluator_result": null, + "evaluator_used": false, + "retried": false, + "accepted_nontrivial_soft_edge": false, + "error_kind": null, + "error_message": null + }, + { + "label": "issue", + "component_index": 11, + "component_count": 28, + "review_unit_index": 1, + "review_unit_count": 1, + "status": "reviewed", + "reason": null, + "source": "llm", + "cache_hit": false, + "model": "gpt-5.4-mini?service_tier=flex", + "cluster_id": "cluster-43598-2", + "nodes": [ + "issue:43598", + "issue:43600" + ], + "soft_pairs": [ + "issue:43598|issue:43600" + ], + "prepared_review_unit_hash": "8cd29eef3765601e22e27df15865302114c194c5f674fcc96bf8ae483ad16d6e", + "component_budget": { + "node_count": 2, + "item_count": 2, + "soft_pair_count": 1, + "serialized_chars": 1584, + "estimated_input_tokens": 396, + "estimated_eval_tokens": 1048 + }, + "budget": { + "node_count": 2, + "item_count": 2, + "soft_pair_count": 1, + "serialized_chars": 1584, + "estimated_input_tokens": 396, + "estimated_eval_tokens": 1048 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": false, + "analyst_result": { + "summary": "These two issues are related to Trainer extensibility, but they describe different problems: one is about the `remove_unused_column` behavior, the other about FLOPS estimation for custom trainers. They should not be treated as duplicates.", + "confidence": 0.89, + "canonical_issue_reason": "Issue 43598 is more specific about a concrete Trainer customization pain point (`remove_unused_column`) and has a clearer actionable scope.", + "canonical_pr_reason": null, + "best_issue_reason": "Issue 43598 is the better canonical issue because it targets a specific Trainer option and likely maps to a narrower, more directly fixable code path.", + "best_pr_reason": null, + "soft_edge_verdicts": [ + { + "left": "issue:43598", + "right": "issue:43600", + "accept": false, + "reason": "They share the broad theme of custom trainers, but the underlying bugs are different: one concerns unused-column removal, the other concerns FLOPS estimation logic. Not the same change or defect." + } + ] + }, + "evaluator_result": null, + "evaluator_used": false, + "retried": false, + "accepted_nontrivial_soft_edge": false, + "error_kind": null, + "error_message": null + }, + { + "label": "issue", + "component_index": 12, + "component_count": 28, + "review_unit_index": 1, + "review_unit_count": 1, + "status": "reviewed", + "reason": null, + "source": "llm", + "cache_hit": false, + "model": "gpt-5.4-mini?service_tier=flex", + "cluster_id": "cluster-43684-2", + "nodes": [ + "issue:43684", + "issue:44405" + ], + "soft_pairs": [ + "issue:43684|issue:44405" + ], + "prepared_review_unit_hash": "32f42af16420db0b029771312aaac1fb67234bb94899285d4a641a235bbd695a", + "component_budget": { + "node_count": 2, + "item_count": 2, + "soft_pair_count": 1, + "serialized_chars": 1612, + "estimated_input_tokens": 403, + "estimated_eval_tokens": 1062 + }, + "budget": { + "node_count": 2, + "item_count": 2, + "soft_pair_count": 1, + "serialized_chars": 1612, + "estimated_input_tokens": 403, + "estimated_eval_tokens": 1062 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": false, + "analyst_result": { + "summary": "These are two different feature requests in the same broad Transformers area, but they target different model families and different task heads. They should not be clustered as duplicates.", + "confidence": 0.96, + "canonical_issue_reason": "Issue #43684 is the broader and more central request: registering Qwen3-Omni with AutoModel and AutoModelForConditionalGeneration. It describes a concrete missing model registration path rather than a narrow task-head variant.", + "canonical_pr_reason": null, + "best_issue_reason": "#43684 is the best representative issue because it covers the main registration gap and is more general than the sequence-classification-only request in #44405.", + "best_pr_reason": null, + "soft_edge_verdicts": [ + { + "left": "issue:43684", + "right": "issue:44405", + "accept": false, + "reason": "Different underlying requests: Qwen3-Omni model registration for AutoModel/AutoModelForConditionalGeneration vs Qwen3.5 sequence-classification support for AutoModelForSequenceClassification. Same general ecosystem, but not the same bug/change." + } + ] + }, + "evaluator_result": null, + "evaluator_used": false, + "retried": false, + "accepted_nontrivial_soft_edge": false, + "error_kind": null, + "error_message": null + }, + { + "label": "issue", + "component_index": 13, + "component_count": 28, + "review_unit_index": 1, + "review_unit_count": 1, + "status": "reviewed", + "reason": null, + "source": "llm", + "cache_hit": false, + "model": "gpt-5.4-mini?service_tier=flex", + "cluster_id": "cluster-43835-4", + "nodes": [ + "issue:43835", + "issue:43845", + "issue:43846", + "issue:43859" + ], + "soft_pairs": [ + "issue:43835|issue:43845", + "issue:43835|issue:43846", + "issue:43835|issue:43859", + "issue:43845|issue:43846", + "issue:43845|issue:43859", + "issue:43846|issue:43859" + ], + "prepared_review_unit_hash": "9342bb68cacc26fb4c306ff37d4710a1b4c2346d1feddd72197ba3e43d82a68b", + "component_budget": { + "node_count": 4, + "item_count": 4, + "soft_pair_count": 6, + "serialized_chars": 4262, + "estimated_input_tokens": 1066, + "estimated_eval_tokens": 2388 + }, + "budget": { + "node_count": 4, + "item_count": 4, + "soft_pair_count": 6, + "serialized_chars": 4262, + "estimated_input_tokens": 1066, + "estimated_eval_tokens": 2388 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": false, + "analyst_result": { + "summary": "Issue 43845 remains the best canonical placeholder for the repeated low-information \"huggingface \" reports, but the pairwise evidence is too thin to confidently mark those items as true duplicates; they should be treated as likely similar/unresolved. \"Fett\" appears unrelated. No PRs are present.", + "confidence": 0.76, + "canonical_issue_reason": "issue:43845 is the earliest of the identically titled \"huggingface \" issues and is the most reasonable placeholder canonical among that cluster.", + "canonical_pr_reason": null, + "best_issue_reason": "issue:43845 is the best issue placeholder because it is one of the repeated identical-title reports and the earliest available member of that cluster.", + "best_pr_reason": null, + "soft_edge_verdicts": [ + { + "left": "issue:43835", + "right": "issue:43845", + "accept": false, + "reason": "Different titles and no body-level evidence tying \"Fett\" to the low-information \"huggingface \" report." + }, + { + "left": "issue:43835", + "right": "issue:43846", + "accept": false, + "reason": "Different titles and no concrete shared problem details; only weak similarity signals are present." + }, + { + "left": "issue:43835", + "right": "issue:43859", + "accept": false, + "reason": "Different titles and no concrete shared problem details; only weak similarity signals are present." + }, + { + "left": "issue:43845", + "right": "issue:43846", + "accept": false, + "reason": "Same title fragment, but the bodies are unavailable and deterministic_accept is false, so this is not enough to confirm a duplicate; treat as likely similar/unresolved." + }, + { + "left": "issue:43845", + "right": "issue:43859", + "accept": false, + "reason": "Same title fragment, but there is no body evidence and deterministic_accept is false; keep this unresolved rather than accepting a duplicate." + }, + { + "left": "issue:43846", + "right": "issue:43859", + "accept": false, + "reason": "Same title fragment, but the bodies are unavailable and deterministic_accept is false, so duplicate status is unproven." + } + ] + }, + "evaluator_result": { + "accept": false, + "feedback": "The canonical placeholder choice is reasonable, but the soft-edge verdicts are too strong. In the packet, the three \"huggingface \" items only share title-level similarity and soft_similarity signals; the bodies are not available, and every pair is explicitly marked deterministic_accept=false. That is not enough to confidently accept them as duplicates. Please keep those verdicts conservative (e.g., likely similar / unresolved) unless you have stronger evidence from the issue bodies." + }, + "evaluator_used": true, + "retried": true, + "accepted_nontrivial_soft_edge": false, + "error_kind": null, + "error_message": null + }, + { + "label": "issue", + "component_index": 14, + "component_count": 28, + "review_unit_index": 1, + "review_unit_count": 1, + "status": "reviewed", + "reason": null, + "source": "llm", + "cache_hit": false, + "model": "gpt-5.4-mini?service_tier=flex", + "cluster_id": "cluster-44371-6", + "nodes": [ + "issue:44371", + "issue:44418", + "issue:44572", + "issue:44573", + "issue:44937", + "issue:47036" + ], + "soft_pairs": [ + "issue:44371|issue:44572", + "issue:44371|issue:44573", + "issue:44371|issue:44418", + "issue:44371|issue:47036", + "issue:44371|issue:44937", + "issue:44418|issue:44572", + "issue:44418|issue:44573", + "issue:44572|issue:44573", + "issue:44572|issue:47036", + "issue:44572|issue:44937", + "issue:44573|issue:47036", + "issue:44573|issue:44937" + ], + "prepared_review_unit_hash": "5889c1884ae448147daae06cab5c976c92301c55c6f4c76446c21eca3c72c4d7", + "component_budget": { + "node_count": 6, + "item_count": 6, + "soft_pair_count": 12, + "serialized_chars": 7355, + "estimated_input_tokens": 1839, + "estimated_eval_tokens": 3934 + }, + "budget": { + "node_count": 6, + "item_count": 6, + "soft_pair_count": 12, + "serialized_chars": 7355, + "estimated_input_tokens": 1839, + "estimated_eval_tokens": 3934 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": false, + "analyst_result": { + "summary": "These look like unrelated low-content/spam or vague issues, not duplicates of one underlying bug. No soft edge has enough concrete overlap to merge.", + "confidence": 0.93, + "canonical_issue_reason": null, + "canonical_pr_reason": null, + "best_issue_reason": "issue:47036 is the most specific/legible item in the set, but it is still too vague to serve as a true canonical duplicate target.", + "best_pr_reason": null, + "soft_edge_verdicts": [ + { + "left": "issue:44371", + "right": "issue:44572", + "accept": false, + "reason": "Both are spam-like titles with no concrete bug details; no shared underlying issue is visible." + }, + { + "left": "issue:44371", + "right": "issue:44573", + "accept": false, + "reason": "Both are spam-like titles with no actionable content, so they do not appear to describe the same bug." + }, + { + "left": "issue:44371", + "right": "issue:44418", + "accept": false, + "reason": "One is a generic documentation suggestion, the other is spam; no duplicate bug/change can be inferred." + }, + { + "left": "issue:44371", + "right": "issue:47036", + "accept": false, + "reason": "These titles are unrelated and too vague to indicate the same underlying issue." + }, + { + "left": "issue:44371", + "right": "issue:44937", + "accept": false, + "reason": "Unrelated spam/generic titles with no shared concrete problem statement." + }, + { + "left": "issue:44418", + "right": "issue:44572", + "accept": false, + "reason": "Generic documentation enhancement versus spam; not the same issue." + }, + { + "left": "issue:44418", + "right": "issue:44573", + "accept": false, + "reason": "Generic documentation enhancement versus spam; no evidence of duplication." + }, + { + "left": "issue:44572", + "right": "issue:44573", + "accept": false, + "reason": "Both are spam-like placeholders, but there is no concrete bug or change to match." + }, + { + "left": "issue:44572", + "right": "issue:47036", + "accept": false, + "reason": "No shared concrete subject matter beyond vague, low-information titles." + }, + { + "left": "issue:44572", + "right": "issue:44937", + "accept": false, + "reason": "Unrelated vague/spam titles; no common underlying bug or request is visible." + }, + { + "left": "issue:44573", + "right": "issue:47036", + "accept": false, + "reason": "No concrete overlap; the titles do not indicate the same problem." + }, + { + "left": "issue:44573", + "right": "issue:44937", + "accept": false, + "reason": "The items are unrelated and too sparse to support a duplicate merge." + } + ] + }, + "evaluator_result": null, + "evaluator_used": false, + "retried": false, + "accepted_nontrivial_soft_edge": false, + "error_kind": null, + "error_message": null + }, + { + "label": "issue", + "component_index": 15, + "component_count": 28, + "review_unit_index": 1, + "review_unit_count": 1, + "status": "reviewed", + "reason": null, + "source": "llm", + "cache_hit": false, + "model": "gpt-5.4-mini?service_tier=flex", + "cluster_id": "cluster-45145-2", + "nodes": [ + "issue:45145", + "issue:45151" + ], + "soft_pairs": [ + "issue:45145|issue:45151" + ], + "prepared_review_unit_hash": "efd101d63df9118bcc7b607f7948a293cda0771bc959c269bdd3530a184b5470", + "component_budget": { + "node_count": 2, + "item_count": 2, + "soft_pair_count": 1, + "serialized_chars": 1646, + "estimated_input_tokens": 412, + "estimated_eval_tokens": 1080 + }, + "budget": { + "node_count": 2, + "item_count": 2, + "soft_pair_count": 1, + "serialized_chars": 1646, + "estimated_input_tokens": 412, + "estimated_eval_tokens": 1080 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": false, + "analyst_result": { + "summary": "The two closed issues are highly similar and likely related, but the evidence is limited to matching titles, similar body lengths, and weak soft-similarity signals. #45145 is still the best canonical issue because it is earlier and has the only inbound reference, but the duplicate call should be treated as tentative rather than certain.", + "confidence": 0.82, + "canonical_issue_reason": "#45145 is the earlier issue and has the only inbound reference, so it is the best canonical anchor if these are deduplicated.", + "canonical_pr_reason": null, + "best_issue_reason": "#45145 is the better issue to keep as the representative record because it predates #45151 and has stronger linkage, though the duplication evidence is not definitive.", + "best_pr_reason": null, + "soft_edge_verdicts": [ + { + "left": "issue:45145", + "right": "issue:45151", + "accept": false, + "reason": "The titles match and the body lengths are close, but we do not have body content, explicit links, or targets. That is enough to suspect a duplicate, not enough to confidently assert the same underlying report." + } + ] + }, + "evaluator_result": { + "accept": false, + "feedback": "The canonical choice of issue #45145 is grounded: it is earlier and has the only inbound reference. However, the duplicate verdict is a bit too strong for the packet. We only know the titles match, body lengths are similar, and there are no explicit links/targets; we do not have body content, so claims like \"same report content\" or \"no meaningful differences\" overstate the evidence. The soft match should be treated as tentative rather than asserted as near-certain." + }, + "evaluator_used": true, + "retried": true, + "accepted_nontrivial_soft_edge": false, + "error_kind": null, + "error_message": null + }, + { + "label": "issue", + "component_index": 16, + "component_count": 28, + "review_unit_index": 1, + "review_unit_count": 1, + "status": "reviewed", + "reason": null, + "source": "cache", + "cache_hit": true, + "model": "gpt-5.4-mini?service_tier=flex", + "cluster_id": "cluster-45295-2", + "nodes": [ + "issue:45295", + "issue:45373" + ], + "soft_pairs": [], + "prepared_review_unit_hash": "f5fd95fada2f530ab0999e343c17c16266cde0c7182bb22509972ca44827cd49", + "component_budget": { + "node_count": 2, + "item_count": 2, + "soft_pair_count": 0, + "serialized_chars": 1286, + "estimated_input_tokens": 322, + "estimated_eval_tokens": 900 + }, + "budget": { + "node_count": 2, + "item_count": 2, + "soft_pair_count": 0, + "serialized_chars": 1286, + "estimated_input_tokens": 322, + "estimated_eval_tokens": 900 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": false, + "analyst_result": { + "summary": "These two issues describe the same missing Gemma 4 sequence-classification support; the later, more specific open issue is the better cluster representative.", + "confidence": 0.94, + "canonical_issue_reason": "issue #45373 is the clearest canonical tracker: it is open, explicitly names the missing class `Gemma4ForSequenceClassification`, and states the module-level gap relative to Gemma 2/3.", + "canonical_pr_reason": null, + "best_issue_reason": "issue #45373 best captures the concrete request and likely remaining work, while #45295 is a shorter earlier duplicate with less detail and is already closed.", + "best_pr_reason": null, + "soft_edge_verdicts": [] + }, + "evaluator_result": null, + "evaluator_used": false, + "retried": false, + "accepted_nontrivial_soft_edge": false, + "error_kind": null, + "error_message": null + }, + { + "label": "issue", + "component_index": 17, + "component_count": 28, + "review_unit_index": 1, + "review_unit_count": 1, + "status": "reviewed", + "reason": null, + "source": "llm", + "cache_hit": false, + "model": "gpt-5.4-mini?service_tier=flex", + "cluster_id": "cluster-45762-2", + "nodes": [ + "issue:45762", + "issue:45763" + ], + "soft_pairs": [ + "issue:45762|issue:45763" + ], + "prepared_review_unit_hash": "561b1f6512395a8c741218878d48fe58ebb19510de4e0588e26e74ac22aacb50", + "component_budget": { + "node_count": 2, + "item_count": 2, + "soft_pair_count": 1, + "serialized_chars": 1477, + "estimated_input_tokens": 370, + "estimated_eval_tokens": 996 + }, + "budget": { + "node_count": 2, + "item_count": 2, + "soft_pair_count": 1, + "serialized_chars": 1477, + "estimated_input_tokens": 370, + "estimated_eval_tokens": 996 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": false, + "analyst_result": { + "summary": "The cluster contains two unrelated closed issues with personal-name/date titles and no technical content; they do not appear to be duplicates of the same underlying bug or change.", + "confidence": 0.97, + "canonical_issue_reason": "Issue 45762 is the earliest-created item and, with no substantive issue details on either side, it is the cleanest representative if a single issue must be chosen.", + "canonical_pr_reason": null, + "best_issue_reason": "Issue 45762 is the best issue representative only by chronology; neither issue has enough technical signal to prefer one on content.", + "best_pr_reason": null, + "soft_edge_verdicts": [ + { + "left": "issue:45762", + "right": "issue:45763", + "accept": false, + "reason": "Different names and dates with no shared problem description, code path, or issue target; these do not look like the same underlying bug/change." + } + ] + }, + "evaluator_result": null, + "evaluator_used": false, + "retried": false, + "accepted_nontrivial_soft_edge": false, + "error_kind": null, + "error_message": null + }, + { + "label": "issue", + "component_index": 18, + "component_count": 28, + "review_unit_index": 1, + "review_unit_count": 1, + "status": "reviewed", + "reason": null, + "source": "llm", + "cache_hit": false, + "model": "gpt-5.4-mini?service_tier=flex", + "cluster_id": "cluster-45764-2", + "nodes": [ + "issue:45764", + "issue:45765" + ], + "soft_pairs": [ + "issue:45764|issue:45765" + ], + "prepared_review_unit_hash": "15cf1939f3a92a71d710301805684c1ab8b708314a9cd820a26226a57ef5b7ea", + "component_budget": { + "node_count": 2, + "item_count": 2, + "soft_pair_count": 1, + "serialized_chars": 1481, + "estimated_input_tokens": 371, + "estimated_eval_tokens": 998 + }, + "budget": { + "node_count": 2, + "item_count": 2, + "soft_pair_count": 1, + "serialized_chars": 1481, + "estimated_input_tokens": 371, + "estimated_eval_tokens": 998 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": false, + "analyst_result": { + "summary": "The two issues are unrelated low-content items with different personal-name/date titles and no shared bug or feature context, so they should not be clustered as duplicates.", + "confidence": 0.97, + "canonical_issue_reason": null, + "canonical_pr_reason": null, + "best_issue_reason": "Neither issue is a good canonical representative because neither contains actionable problem detail; if forced, no single issue stands out as the same underlying report.", + "best_pr_reason": null, + "soft_edge_verdicts": [ + { + "left": "issue:45764", + "right": "issue:45765", + "accept": false, + "reason": "Different names and dates, no body text, no shared targets or evidence of the same underlying bug/change." + } + ] + }, + "evaluator_result": null, + "evaluator_used": false, + "retried": false, + "accepted_nontrivial_soft_edge": false, + "error_kind": null, + "error_message": null + }, + { + "label": "issue", + "component_index": 19, + "component_count": 28, + "review_unit_index": 1, + "review_unit_count": 1, + "status": "reviewed", + "reason": null, + "source": "llm", + "cache_hit": false, + "model": "gpt-5.4-mini?service_tier=flex", + "cluster_id": "cluster-45810-2", + "nodes": [ + "issue:45810", + "issue:46295" + ], + "soft_pairs": [ + "issue:45810|issue:46295" + ], + "prepared_review_unit_hash": "c533ef251bea22f1d7ba288209c105a5f7c92c05cc859d0aac8a331bcebf65e7", + "component_budget": { + "node_count": 2, + "item_count": 2, + "soft_pair_count": 1, + "serialized_chars": 1540, + "estimated_input_tokens": 385, + "estimated_eval_tokens": 1026 + }, + "budget": { + "node_count": 2, + "item_count": 2, + "soft_pair_count": 1, + "serialized_chars": 1540, + "estimated_input_tokens": 385, + "estimated_eval_tokens": 1026 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": false, + "analyst_result": { + "summary": "These two issues are not duplicates: one requests a new Qwen3_5 token-classification model class, while the other asks for FP8 training support. They concern different features and code paths.", + "confidence": 0.96, + "canonical_issue_reason": null, + "canonical_pr_reason": null, + "best_issue_reason": "Issue 45810 is the more concrete, self-contained feature request, but it is not a substitute for 46295 and should not be treated as the same bug/change.", + "best_pr_reason": null, + "soft_edge_verdicts": [ + { + "left": "issue:45810", + "right": "issue:46295", + "accept": false, + "reason": "No same underlying bug or change: model-class addition for token classification vs. FP8 training support are separate capabilities with different implementation targets." + } + ] + }, + "evaluator_result": null, + "evaluator_used": false, + "retried": false, + "accepted_nontrivial_soft_edge": false, + "error_kind": null, + "error_message": null + }, + { + "label": "issue", + "component_index": 20, + "component_count": 28, + "review_unit_index": 1, + "review_unit_count": 1, + "status": "reviewed", + "reason": null, + "source": "llm", + "cache_hit": false, + "model": "gpt-5.4-mini?service_tier=flex", + "cluster_id": "cluster-45869-5", + "nodes": [ + "issue:45869", + "issue:45870", + "issue:45871", + "issue:45872", + "issue:45873" + ], + "soft_pairs": [ + "issue:45869|issue:45870", + "issue:45869|issue:45871", + "issue:45870|issue:45871", + "issue:45870|issue:45872", + "issue:45870|issue:45873", + "issue:45871|issue:45872", + "issue:45871|issue:45873", + "issue:45872|issue:45873" + ], + "prepared_review_unit_hash": "9a006a6e12fce010532cb05154220f1e676b846f494b263e69043f5ffa211e3c", + "component_budget": { + "node_count": 5, + "item_count": 5, + "soft_pair_count": 8, + "serialized_chars": 5726, + "estimated_input_tokens": 1432, + "estimated_eval_tokens": 3120 + }, + "budget": { + "node_count": 5, + "item_count": 5, + "soft_pair_count": 8, + "serialized_chars": 5726, + "estimated_input_tokens": 1432, + "estimated_eval_tokens": 3120 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": false, + "analyst_result": { + "summary": "There are two likely duplicate issue pairs: 45870/45871 and 45872/45873. The evidence supports duplicate triage, but not a claim of proven exact body matches. Issue 45869 appears unrelated to those pairs.", + "confidence": 0.88, + "canonical_issue_reason": "45870 is the earliest member of the strongest likely-duplicate pair with 45871, and it is a reasonable representative to keep for that request.", + "canonical_pr_reason": null, + "best_issue_reason": "45870 is the strongest overall issue candidate to retain because it is part of a likely duplicate pair with the earliest timestamp, while 45869 is unrelated and the 45872/45873 pair is later.", + "best_pr_reason": null, + "soft_edge_verdicts": [ + { + "left": "issue:45869", + "right": "issue:45870", + "accept": false, + "reason": "Different requests: a dataset bundle versus a financial prediction example." + }, + { + "left": "issue:45869", + "right": "issue:45871", + "accept": false, + "reason": "Different requests: a dataset bundle versus a financial prediction example." + }, + { + "left": "issue:45870", + "right": "issue:45871", + "accept": true, + "reason": "Same title and matching metadata strongly suggest the same underlying issue; likely duplicates." + }, + { + "left": "issue:45870", + "right": "issue:45872", + "accept": false, + "reason": "Different request area: financial prediction example versus scraping-service integration." + }, + { + "left": "issue:45870", + "right": "issue:45873", + "accept": false, + "reason": "Different request area: financial prediction example versus scraping-service integration." + }, + { + "left": "issue:45871", + "right": "issue:45872", + "accept": false, + "reason": "Only broad domain overlap; not the same underlying issue." + }, + { + "left": "issue:45871", + "right": "issue:45873", + "accept": false, + "reason": "Only broad domain overlap; not the same underlying issue." + }, + { + "left": "issue:45872", + "right": "issue:45873", + "accept": true, + "reason": "Same title and matching metadata strongly suggest the same underlying issue; likely duplicates." + } + ] + }, + "evaluator_result": { + "accept": false, + "feedback": "The grouping is directionally plausible, but the analyst overstates the evidence. The packet only shows identical titles and matching body lengths/other metadata, not exact body text, so calling 45870/45871 and 45872/45873 \u201cexact duplicates\u201d or \u201cexact title/body twins\u201d is too strong. The conservative conclusion is that they are likely duplicate issues, not proven exact matches." + }, + "evaluator_used": true, + "retried": true, + "accepted_nontrivial_soft_edge": true, + "error_kind": null, + "error_message": null + }, + { + "label": "issue", + "component_index": 21, + "component_count": 28, + "review_unit_index": 1, + "review_unit_count": 1, + "status": "reviewed", + "reason": null, + "source": "llm", + "cache_hit": false, + "model": "gpt-5.4-mini?service_tier=flex", + "cluster_id": "cluster-46439-2", + "nodes": [ + "issue:46439", + "issue:46441" + ], + "soft_pairs": [ + "issue:46439|issue:46441" + ], + "prepared_review_unit_hash": "70a9df9077e3bc51497e9ebbbf7e43343aa9c03708ec078aaa1012b2f29c1d74", + "component_budget": { + "node_count": 2, + "item_count": 2, + "soft_pair_count": 1, + "serialized_chars": 1641, + "estimated_input_tokens": 411, + "estimated_eval_tokens": 1078 + }, + "budget": { + "node_count": 2, + "item_count": 2, + "soft_pair_count": 1, + "serialized_chars": 1641, + "estimated_input_tokens": 411, + "estimated_eval_tokens": 1078 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": false, + "analyst_result": { + "summary": "These two issues are related to static cache handling in linear-attention models, but they describe different failure modes and likely different root causes: one is a shape corruption/RuntimeError during early initialization, the other is a Qwen3.5-specific KeyError for `linear_attention`.", + "confidence": 0.89, + "canonical_issue_reason": "Issue 46439 is the better canonical issue because it describes the broader underlying static-cache/linear-attention initialization problem in a model-agnostic way.", + "canonical_pr_reason": null, + "best_issue_reason": "Issue 46441 is the most actionable report for reproduction because it names the affected model and gives a concrete failure signature (`KeyError: 'linear_attention'`).", + "best_pr_reason": null, + "soft_edge_verdicts": [ + { + "left": "issue:46439", + "right": "issue:46441", + "accept": false, + "reason": "Reject: both involve static cache and linear attention, but one fails with shape corruption/RuntimeError during early initialization while the other is a Qwen3.5 `KeyError`; they do not clearly look like the same underlying bug." + } + ] + }, + "evaluator_result": null, + "evaluator_used": false, + "retried": false, + "accepted_nontrivial_soft_edge": false, + "error_kind": null, + "error_message": null + }, + { + "label": "issue", + "component_index": 22, + "component_count": 28, + "review_unit_index": 1, + "review_unit_count": 1, + "status": "reviewed", + "reason": null, + "source": "llm", + "cache_hit": false, + "model": "gpt-5.4-mini?service_tier=flex", + "cluster_id": "cluster-46551-2", + "nodes": [ + "issue:46551", + "issue:46552" + ], + "soft_pairs": [ + "issue:46551|issue:46552" + ], + "prepared_review_unit_hash": "cc81fe7f9d0835648e5f39e7387802d0e262e564e091507324fe06d1959ecf5d", + "component_budget": { + "node_count": 2, + "item_count": 2, + "soft_pair_count": 1, + "serialized_chars": 1707, + "estimated_input_tokens": 427, + "estimated_eval_tokens": 1110 + }, + "budget": { + "node_count": 2, + "item_count": 2, + "soft_pair_count": 1, + "serialized_chars": 1707, + "estimated_input_tokens": 427, + "estimated_eval_tokens": 1110 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": false, + "analyst_result": { + "summary": "The cluster contains two identical reports of the same Wav2Vec2CTCTokenizer v5 bug. #46552 is the better living issue because it is open and already has inbound references; #46551 is an older closed instance of the same report.", + "confidence": 0.95, + "canonical_issue_reason": "#46552 is the open, referenced tracker for the same report, so it is the better canonical issue.", + "canonical_pr_reason": null, + "best_issue_reason": "#46552 is the strongest issue to keep because it is active and has inbound references while matching #46551\u2019s title/body exactly.", + "best_pr_reason": null, + "soft_edge_verdicts": [ + { + "left": "issue:46551", + "right": "issue:46552", + "accept": true, + "reason": "Same underlying bug report with identical title/body; #46552 is the better living tracker." + } + ] + }, + "evaluator_result": { + "accept": false, + "feedback": "The packet supports that these two issues have identical titles/bodies and that #46552 is the better living tracker because it is open and has inbound references. However, the analysis overstates the evidence by calling #46551 a \u201cclosed duplicate\u201d and by presenting the pair as definitively duplicate rather than clearly same/near-identical reports. The soft-edge verdict should stay more conservative." + }, + "evaluator_used": true, + "retried": true, + "accepted_nontrivial_soft_edge": true, + "error_kind": null, + "error_message": null + }, + { + "label": "issue", + "component_index": 23, + "component_count": 28, + "review_unit_index": 1, + "review_unit_count": 1, + "status": "reviewed", + "reason": null, + "source": "llm", + "cache_hit": false, + "model": "gpt-5.4-mini?service_tier=flex", + "cluster_id": "cluster-46657-2", + "nodes": [ + "issue:46657", + "issue:46692" + ], + "soft_pairs": [ + "issue:46657|issue:46692" + ], + "prepared_review_unit_hash": "28da07af1dfbb9866f97121fb30e5855b06bc1f9055822b4c13b56d961a197cf", + "component_budget": { + "node_count": 2, + "item_count": 2, + "soft_pair_count": 1, + "serialized_chars": 1569, + "estimated_input_tokens": 393, + "estimated_eval_tokens": 1042 + }, + "budget": { + "node_count": 2, + "item_count": 2, + "soft_pair_count": 1, + "serialized_chars": 1569, + "estimated_input_tokens": 393, + "estimated_eval_tokens": 1042 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": false, + "analyst_result": { + "summary": "These two issues both concern model training, but they describe different failures: a tensor shape mismatch versus out-of-memory after several epochs. The shared wording is too broad to treat them as duplicates.", + "confidence": 0.96, + "canonical_issue_reason": "Issue 46657 is the more concrete, technically specific failure report and is the better representative of the cluster\u2019s training-bug theme, though it is not actually a duplicate of 46692.", + "canonical_pr_reason": null, + "best_issue_reason": "Issue 46657 is the best single issue to surface here because it describes a clearer code-path problem (tensor size mismatch) than the more generic OOM report.", + "best_pr_reason": null, + "soft_edge_verdicts": [ + { + "left": "issue:46657", + "right": "issue:46692", + "accept": false, + "reason": "Reject: one report is a tensor-size mismatch during object detection training, the other is an out-of-memory condition after some epochs in RT-DETR v2 training. These are different failure modes and would not plausibly be fixed by the same change." + } + ] + }, + "evaluator_result": null, + "evaluator_used": false, + "retried": false, + "accepted_nontrivial_soft_edge": false, + "error_kind": null, + "error_message": null + }, + { + "label": "issue", + "component_index": 24, + "component_count": 28, + "review_unit_index": 1, + "review_unit_count": 1, + "status": "reviewed", + "reason": null, + "source": "llm", + "cache_hit": false, + "model": "gpt-5.4-mini?service_tier=flex", + "cluster_id": "cluster-46822-2", + "nodes": [ + "issue:46822", + "issue:46823" + ], + "soft_pairs": [ + "issue:46822|issue:46823" + ], + "prepared_review_unit_hash": "d2bc4fbda5de0d0c77845f41e51cc2d7ef5c91c453cc2ef516f68880f5350897", + "component_budget": { + "node_count": 2, + "item_count": 2, + "soft_pair_count": 1, + "serialized_chars": 1673, + "estimated_input_tokens": 419, + "estimated_eval_tokens": 1094 + }, + "budget": { + "node_count": 2, + "item_count": 2, + "soft_pair_count": 1, + "serialized_chars": 1673, + "estimated_input_tokens": 419, + "estimated_eval_tokens": 1094 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": false, + "analyst_result": { + "summary": "These are different loading bugs: one is ZeRO-3 zero.Init failing to partition a composite submodule and causing OOM, the other is 4-bit auto device mapping with max_memory incorrectly sending modules to CPU and raising ValueError. They should not be treated as duplicates.", + "confidence": 0.97, + "canonical_issue_reason": "No single canonical duplicate here: 46822 is about ZeRO-3 composite submodule partitioning/OOM, while 46823 is about 4-bit load_in_4bit + device_map='auto' + max_memory mis-dispatching to CPU.", + "canonical_pr_reason": null, + "best_issue_reason": "46822 is the better standalone issue to keep as a representative of its specific bug report because it is a clear, concrete ZeRO-3 failure mode and is still open.", + "best_pr_reason": null, + "soft_edge_verdicts": [ + { + "left": "issue:46822", + "right": "issue:46823", + "accept": false, + "reason": "They describe different underlying bugs, different code paths, and different failure signatures; the shared loading context is too broad to merge them." + } + ] + }, + "evaluator_result": null, + "evaluator_used": false, + "retried": false, + "accepted_nontrivial_soft_edge": false, + "error_kind": null, + "error_message": null + }, + { + "label": "issue", + "component_index": 25, + "component_count": 28, + "review_unit_index": 1, + "review_unit_count": 1, + "status": "reviewed", + "reason": null, + "source": "llm", + "cache_hit": false, + "model": "gpt-5.4-mini?service_tier=flex", + "cluster_id": "cluster-46906-20", + "nodes": [ + "issue:46906", + "issue:46940", + "issue:46948", + "issue:46969", + "issue:46985", + "issue:46998", + "issue:47037", + "issue:47049", + "issue:47053", + "issue:47091", + "issue:47102", + "issue:47120", + "issue:47182", + "issue:47227", + "issue:47256", + "issue:47265", + "issue:47279", + "issue:47309", + "issue:47377", + "issue:47390" + ], + "soft_pairs": [ + "issue:46906|issue:46940", + "issue:46906|issue:46948", + "issue:46906|issue:46969", + "issue:46906|issue:46985", + "issue:46906|issue:47037", + "issue:46940|issue:46948", + "issue:46940|issue:46969", + "issue:46940|issue:46985", + "issue:46940|issue:46998", + "issue:46948|issue:46969", + "issue:46948|issue:46985", + "issue:46948|issue:46998", + "issue:46969|issue:46985", + "issue:46969|issue:47037", + "issue:46969|issue:47091", + "issue:46985|issue:47037", + "issue:46985|issue:47049", + "issue:46998|issue:47053", + "issue:46998|issue:47037", + "issue:46998|issue:47049", + "issue:46985|issue:46998", + "issue:46998|issue:47091", + "issue:47037|issue:47049", + "issue:47037|issue:47091", + "issue:47037|issue:47053", + "issue:47049|issue:47053", + "issue:47049|issue:47091", + "issue:47053|issue:47091", + "issue:47053|issue:47102", + "issue:47091|issue:47102", + "issue:46985|issue:47091", + "issue:47037|issue:47102", + "issue:46985|issue:47102", + "issue:47049|issue:47102", + "issue:47120|issue:47182", + "issue:47120|issue:47256", + "issue:47120|issue:47279", + "issue:47120|issue:47265", + "issue:47120|issue:47377", + "issue:47182|issue:47377", + "issue:47182|issue:47256", + "issue:47182|issue:47265", + "issue:47182|issue:47279", + "issue:47227|issue:47377", + "issue:47227|issue:47256", + "issue:47227|issue:47279", + "issue:47182|issue:47227", + "issue:47227|issue:47265", + "issue:47256|issue:47265", + "issue:47256|issue:47279", + "issue:47102|issue:47256", + "issue:47265|issue:47279", + "issue:47265|issue:47309", + "issue:47265|issue:47377", + "issue:47279|issue:47309", + "issue:47279|issue:47377", + "issue:47309|issue:47377", + "issue:47182|issue:47309", + "issue:47227|issue:47309", + "issue:47377|issue:47390", + "issue:47279|issue:47390", + "issue:47227|issue:47390", + "issue:47256|issue:47390", + "issue:47182|issue:47390" + ], + "prepared_review_unit_hash": "84dd606a34c8c1b451acfe7dd1e04f9dfdec8980f2252a1e945d771daf9a3e37", + "component_budget": { + "node_count": 20, + "item_count": 20, + "soft_pair_count": 64, + "serialized_chars": 34212, + "estimated_input_tokens": 8553, + "estimated_eval_tokens": 17362 + }, + "budget": { + "node_count": 20, + "item_count": 20, + "soft_pair_count": 64, + "serialized_chars": 34212, + "estimated_input_tokens": 8553, + "estimated_eval_tokens": 17362 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": false, + "analyst_result": { + "summary": "These are dated daily triage issues for the same Serde/\"serge\" integration-failure workflow, not duplicate reports of one underlying bug. They should stay separate; at most, one can be treated as the best representative thread for the series.", + "confidence": 0.95, + "canonical_issue_reason": "Issue 47279 is the strongest representative of the series: it has the highest engagement/inbound references among the dated triage issues and sits near the middle of the run.", + "canonical_pr_reason": null, + "best_issue_reason": "Issue 47279 is the best umbrella issue if one must be chosen, because it is the most central and active triage thread in this dated sequence.", + "best_pr_reason": null, + "soft_edge_verdicts": [ + { + "left": "issue:46906", + "right": "issue:46940", + "accept": false, + "reason": "Reject: same triage template, but these are date-specific integration-failure reports; no evidence they describe one identical underlying bug." + }, + { + "left": "issue:46906", + "right": "issue:46948", + "accept": false, + "reason": "Reject: same triage template, but these are date-specific integration-failure reports; no evidence they describe one identical underlying bug." + }, + { + "left": "issue:46906", + "right": "issue:46969", + "accept": false, + "reason": "Reject: same triage template, but these are date-specific integration-failure reports; no evidence they describe one identical underlying bug." + }, + { + "left": "issue:46906", + "right": "issue:46985", + "accept": false, + "reason": "Reject: same triage template, but these are date-specific integration-failure reports; no evidence they describe one identical underlying bug." + }, + { + "left": "issue:46906", + "right": "issue:47037", + "accept": false, + "reason": "Reject: same triage template, but these are date-specific integration-failure reports; no evidence they describe one identical underlying bug." + }, + { + "left": "issue:46940", + "right": "issue:46948", + "accept": false, + "reason": "Reject: same triage template, but these are date-specific integration-failure reports; no evidence they describe one identical underlying bug." + }, + { + "left": "issue:46940", + "right": "issue:46969", + "accept": false, + "reason": "Reject: same triage template, but these are date-specific integration-failure reports; no evidence they describe one identical underlying bug." + }, + { + "left": "issue:46940", + "right": "issue:46985", + "accept": false, + "reason": "Reject: same triage template, but these are date-specific integration-failure reports; no evidence they describe one identical underlying bug." + }, + { + "left": "issue:46940", + "right": "issue:46998", + "accept": false, + "reason": "Reject: same triage template, but these are date-specific integration-failure reports; no evidence they describe one identical underlying bug." + }, + { + "left": "issue:46948", + "right": "issue:46969", + "accept": false, + "reason": "Reject: same triage template, but these are date-specific integration-failure reports; no evidence they describe one identical underlying bug." + }, + { + "left": "issue:46948", + "right": "issue:46985", + "accept": false, + "reason": "Reject: same triage template, but these are date-specific integration-failure reports; no evidence they describe one identical underlying bug." + }, + { + "left": "issue:46948", + "right": "issue:46998", + "accept": false, + "reason": "Reject: same triage template, but these are date-specific integration-failure reports; no evidence they describe one identical underlying bug." + }, + { + "left": "issue:46969", + "right": "issue:46985", + "accept": false, + "reason": "Reject: same triage template, but these are date-specific integration-failure reports; no evidence they describe one identical underlying bug." + }, + { + "left": "issue:46969", + "right": "issue:47037", + "accept": false, + "reason": "Reject: same triage template, but these are date-specific integration-failure reports; no evidence they describe one identical underlying bug." + }, + { + "left": "issue:46969", + "right": "issue:47091", + "accept": false, + "reason": "Reject: same triage template, but these are date-specific integration-failure reports; no evidence they describe one identical underlying bug." + }, + { + "left": "issue:46985", + "right": "issue:47037", + "accept": false, + "reason": "Reject: same triage template, but these are date-specific integration-failure reports; no evidence they describe one identical underlying bug." + }, + { + "left": "issue:46985", + "right": "issue:47049", + "accept": false, + "reason": "Reject: same triage template, but these are date-specific integration-failure reports; no evidence they describe one identical underlying bug." + }, + { + "left": "issue:46998", + "right": "issue:47053", + "accept": false, + "reason": "Reject: same triage template, but these are date-specific integration-failure reports; no evidence they describe one identical underlying bug." + }, + { + "left": "issue:46998", + "right": "issue:47037", + "accept": false, + "reason": "Reject: same triage template, but these are date-specific integration-failure reports; no evidence they describe one identical underlying bug." + }, + { + "left": "issue:46998", + "right": "issue:47049", + "accept": false, + "reason": "Reject: same triage template, but these are date-specific integration-failure reports; no evidence they describe one identical underlying bug." + }, + { + "left": "issue:46985", + "right": "issue:46998", + "accept": false, + "reason": "Reject: same triage template, but these are date-specific integration-failure reports; no evidence they describe one identical underlying bug." + }, + { + "left": "issue:46998", + "right": "issue:47091", + "accept": false, + "reason": "Reject: same triage template, but these are date-specific integration-failure reports; no evidence they describe one identical underlying bug." + }, + { + "left": "issue:47037", + "right": "issue:47049", + "accept": false, + "reason": "Reject: same triage template, but these are date-specific integration-failure reports; no evidence they describe one identical underlying bug." + }, + { + "left": "issue:47037", + "right": "issue:47091", + "accept": false, + "reason": "Reject: same triage template, but these are date-specific integration-failure reports; no evidence they describe one identical underlying bug." + }, + { + "left": "issue:47037", + "right": "issue:47053", + "accept": false, + "reason": "Reject: same triage template, but these are date-specific integration-failure reports; no evidence they describe one identical underlying bug." + }, + { + "left": "issue:47049", + "right": "issue:47053", + "accept": false, + "reason": "Reject: same triage template, but these are date-specific integration-failure reports; no evidence they describe one identical underlying bug." + }, + { + "left": "issue:47049", + "right": "issue:47091", + "accept": false, + "reason": "Reject: same triage template, but these are date-specific integration-failure reports; no evidence they describe one identical underlying bug." + }, + { + "left": "issue:47053", + "right": "issue:47091", + "accept": false, + "reason": "Reject: same triage template, but these are date-specific integration-failure reports; no evidence they describe one identical underlying bug." + }, + { + "left": "issue:47053", + "right": "issue:47102", + "accept": false, + "reason": "Reject: same triage template, but these are date-specific integration-failure reports; no evidence they describe one identical underlying bug." + }, + { + "left": "issue:47091", + "right": "issue:47102", + "accept": false, + "reason": "Reject: same triage template, but these are date-specific integration-failure reports; no evidence they describe one identical underlying bug." + }, + { + "left": "issue:46985", + "right": "issue:47091", + "accept": false, + "reason": "Reject: same triage template, but these are date-specific integration-failure reports; no evidence they describe one identical underlying bug." + }, + { + "left": "issue:47037", + "right": "issue:47102", + "accept": false, + "reason": "Reject: same triage template, but these are date-specific integration-failure reports; no evidence they describe one identical underlying bug." + }, + { + "left": "issue:46985", + "right": "issue:47102", + "accept": false, + "reason": "Reject: same triage template, but these are date-specific integration-failure reports; no evidence they describe one identical underlying bug." + }, + { + "left": "issue:47049", + "right": "issue:47102", + "accept": false, + "reason": "Reject: same triage template, but these are date-specific integration-failure reports; no evidence they describe one identical underlying bug." + }, + { + "left": "issue:47120", + "right": "issue:47182", + "accept": false, + "reason": "Reject: same triage template, but these are date-specific integration-failure reports; no evidence they describe one identical underlying bug." + }, + { + "left": "issue:47120", + "right": "issue:47256", + "accept": false, + "reason": "Reject: same triage template, but these are date-specific integration-failure reports; no evidence they describe one identical underlying bug." + }, + { + "left": "issue:47120", + "right": "issue:47279", + "accept": false, + "reason": "Reject: same triage template, but these are date-specific integration-failure reports; no evidence they describe one identical underlying bug." + }, + { + "left": "issue:47120", + "right": "issue:47265", + "accept": false, + "reason": "Reject: same triage template, but these are date-specific integration-failure reports; no evidence they describe one identical underlying bug." + }, + { + "left": "issue:47120", + "right": "issue:47377", + "accept": false, + "reason": "Reject: same triage template, but these are date-specific integration-failure reports; no evidence they describe one identical underlying bug." + }, + { + "left": "issue:47182", + "right": "issue:47227", + "accept": false, + "reason": "Reject: same triage template, but these are date-specific integration-failure reports; no evidence they describe one identical underlying bug." + }, + { + "left": "issue:47182", + "right": "issue:47256", + "accept": false, + "reason": "Reject: same triage template, but these are date-specific integration-failure reports; no evidence they describe one identical underlying bug." + }, + { + "left": "issue:47182", + "right": "issue:47265", + "accept": false, + "reason": "Reject: same triage template, but these are date-specific integration-failure reports; no evidence they describe one identical underlying bug." + }, + { + "left": "issue:47182", + "right": "issue:47279", + "accept": false, + "reason": "Reject: same triage template, but these are date-specific integration-failure reports; no evidence they describe one identical underlying bug." + }, + { + "left": "issue:47182", + "right": "issue:47309", + "accept": false, + "reason": "Reject: same triage template, but these are date-specific integration-failure reports; no evidence they describe one identical underlying bug." + }, + { + "left": "issue:47182", + "right": "issue:47377", + "accept": false, + "reason": "Reject: same triage template, but these are date-specific integration-failure reports; no evidence they describe one identical underlying bug." + }, + { + "left": "issue:47182", + "right": "issue:47390", + "accept": false, + "reason": "Reject: same triage template, but these are date-specific integration-failure reports; no evidence they describe one identical underlying bug." + }, + { + "left": "issue:47227", + "right": "issue:47256", + "accept": false, + "reason": "Reject: same triage template, but these are date-specific integration-failure reports; no evidence they describe one identical underlying bug." + }, + { + "left": "issue:47227", + "right": "issue:47265", + "accept": false, + "reason": "Reject: same triage template, but these are date-specific integration-failure reports; no evidence they describe one identical underlying bug." + }, + { + "left": "issue:47227", + "right": "issue:47279", + "accept": false, + "reason": "Reject: same triage template, but these are date-specific integration-failure reports; no evidence they describe one identical underlying bug." + }, + { + "left": "issue:47227", + "right": "issue:47309", + "accept": false, + "reason": "Reject: same triage template, but these are date-specific integration-failure reports; no evidence they describe one identical underlying bug." + }, + { + "left": "issue:47227", + "right": "issue:47377", + "accept": false, + "reason": "Reject: same triage template, but these are date-specific integration-failure reports; no evidence they describe one identical underlying bug." + }, + { + "left": "issue:47227", + "right": "issue:47390", + "accept": false, + "reason": "Reject: same triage template, but these are date-specific integration-failure reports; no evidence they describe one identical underlying bug." + }, + { + "left": "issue:47256", + "right": "issue:47265", + "accept": false, + "reason": "Reject: same triage template, but these are date-specific integration-failure reports; no evidence they describe one identical underlying bug." + }, + { + "left": "issue:47256", + "right": "issue:47279", + "accept": false, + "reason": "Reject: same triage template, but these are date-specific integration-failure reports; no evidence they describe one identical underlying bug." + }, + { + "left": "issue:47256", + "right": "issue:47390", + "accept": false, + "reason": "Reject: same triage template, but these are date-specific integration-failure reports; no evidence they describe one identical underlying bug." + }, + { + "left": "issue:47265", + "right": "issue:47279", + "accept": false, + "reason": "Reject: same triage template, but these are date-specific integration-failure reports; no evidence they describe one identical underlying bug." + }, + { + "left": "issue:47265", + "right": "issue:47309", + "accept": false, + "reason": "Reject: same triage template, but these are date-specific integration-failure reports; no evidence they describe one identical underlying bug." + }, + { + "left": "issue:47265", + "right": "issue:47377", + "accept": false, + "reason": "Reject: same triage template, but these are date-specific integration-failure reports; no evidence they describe one identical underlying bug." + }, + { + "left": "issue:47279", + "right": "issue:47309", + "accept": false, + "reason": "Reject: same triage template, but these are date-specific integration-failure reports; no evidence they describe one identical underlying bug." + }, + { + "left": "issue:47279", + "right": "issue:47377", + "accept": false, + "reason": "Reject: same triage template, but these are date-specific integration-failure reports; no evidence they describe one identical underlying bug." + }, + { + "left": "issue:47279", + "right": "issue:47390", + "accept": false, + "reason": "Reject: same triage template, but these are date-specific integration-failure reports; no evidence they describe one identical underlying bug." + }, + { + "left": "issue:47309", + "right": "issue:47377", + "accept": false, + "reason": "Reject: same triage template, but these are date-specific integration-failure reports; no evidence they describe one identical underlying bug." + }, + { + "left": "issue:47377", + "right": "issue:47390", + "accept": false, + "reason": "Reject: same triage template, but these are date-specific integration-failure reports; no evidence they describe one identical underlying bug." + }, + { + "left": "issue:47102", + "right": "issue:47256", + "accept": false, + "reason": "Reject: same triage template, but these are date-specific integration-failure reports; no evidence they describe one identical underlying bug." + } + ] + }, + "evaluator_result": null, + "evaluator_used": false, + "retried": false, + "accepted_nontrivial_soft_edge": false, + "error_kind": null, + "error_message": null + }, + { + "label": "issue", + "component_index": 26, + "component_count": 28, + "review_unit_index": 1, + "review_unit_count": 1, + "status": "reviewed", + "reason": null, + "source": "llm", + "cache_hit": false, + "model": "gpt-5.4-mini?service_tier=flex", + "cluster_id": "cluster-46983-2", + "nodes": [ + "issue:46983", + "issue:46984" + ], + "soft_pairs": [ + "issue:46983|issue:46984" + ], + "prepared_review_unit_hash": "19c4008274f3f0e2d68fc9a1ecd50fe3e4f901546c7d54b9ce67fe9111d155ba", + "component_budget": { + "node_count": 2, + "item_count": 2, + "soft_pair_count": 1, + "serialized_chars": 1492, + "estimated_input_tokens": 373, + "estimated_eval_tokens": 1002 + }, + "budget": { + "node_count": 2, + "item_count": 2, + "soft_pair_count": 1, + "serialized_chars": 1492, + "estimated_input_tokens": 373, + "estimated_eval_tokens": 1002 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": false, + "analyst_result": { + "summary": "These two issues are duplicates: same title, similar body size/timing, and strong text similarity with no conflicting scope signals. Keep one canonical issue and close the other as a duplicate.", + "confidence": 0.99, + "canonical_issue_reason": "issue #46983 is the better canonical anchor by lower number and identical title; there is no evidence it differs in scope from #46984.", + "canonical_pr_reason": null, + "best_issue_reason": "issue #46983 is suitable as the global best issue because it is the earliest and equally represents the shared request for RWKV7 support.", + "best_pr_reason": null, + "soft_edge_verdicts": [ + { + "left": "issue:46983", + "right": "issue:46984", + "accept": true, + "reason": "Same underlying request: identical title and highly similar content indicate a duplicate issue rather than two distinct feature requests." + } + ] + }, + "evaluator_result": { + "accept": true, + "feedback": "The conclusion is grounded in the packet: both issues have the same title, similar creation timing/body size, and strong soft-similarity evidence (Jaccard 0.837). The duplicate/soft-edge verdict is conservative and does not overclaim beyond the available evidence." + }, + "evaluator_used": true, + "retried": false, + "accepted_nontrivial_soft_edge": true, + "error_kind": null, + "error_message": null + }, + { + "label": "issue", + "component_index": 27, + "component_count": 28, + "review_unit_index": 1, + "review_unit_count": 1, + "status": "reviewed", + "reason": null, + "source": "llm", + "cache_hit": false, + "model": "gpt-5.4-mini?service_tier=flex", + "cluster_id": "cluster-47007-2", + "nodes": [ + "issue:47007", + "issue:47015" + ], + "soft_pairs": [ + "issue:47007|issue:47015" + ], + "prepared_review_unit_hash": "41c5c5e4bd01db13723e16623e72916bae6d2bfd2e64705e74d39dd84e9bf4fe", + "component_budget": { + "node_count": 2, + "item_count": 2, + "soft_pair_count": 1, + "serialized_chars": 1626, + "estimated_input_tokens": 407, + "estimated_eval_tokens": 1070 + }, + "budget": { + "node_count": 2, + "item_count": 2, + "soft_pair_count": 1, + "serialized_chars": 1626, + "estimated_input_tokens": 407, + "estimated_eval_tokens": 1070 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": false, + "analyst_result": { + "summary": "The two closed issues are thematically similar AI/cybersecurity writeups, but they do not clearly describe the same actionable bug, request, or code change. They look more like overlapping topic proposals than duplicates.", + "confidence": 0.84, + "canonical_issue_reason": null, + "canonical_pr_reason": null, + "best_issue_reason": "Issue 47015 is the broader, more generic framing and is the better representative if one item must stand in for the cluster, but it is not a clear duplicate of 47007.", + "best_pr_reason": null, + "soft_edge_verdicts": [ + { + "left": "issue:47007", + "right": "issue:47015", + "accept": false, + "reason": "High topical overlap, but the titles and bodies suggest different writeups/themes rather than the same underlying issue or change; not enough evidence to merge as duplicates." + } + ] + }, + "evaluator_result": null, + "evaluator_used": false, + "retried": false, + "accepted_nontrivial_soft_edge": false, + "error_kind": null, + "error_message": null + }, + { + "label": "issue", + "component_index": 28, + "component_count": 28, + "review_unit_index": 1, + "review_unit_count": 1, + "status": "reviewed", + "reason": null, + "source": "llm", + "cache_hit": false, + "model": "gpt-5.4-mini?service_tier=flex", + "cluster_id": "cluster-47138-2", + "nodes": [ + "issue:47138", + "issue:47139" + ], + "soft_pairs": [ + "issue:47138|issue:47139" + ], + "prepared_review_unit_hash": "bc88424995408e9eb9d79b6a17ca4a3ab2df8721fa0a27bc2265cf253d046adf", + "component_budget": { + "node_count": 2, + "item_count": 2, + "soft_pair_count": 1, + "serialized_chars": 1590, + "estimated_input_tokens": 398, + "estimated_eval_tokens": 1052 + }, + "budget": { + "node_count": 2, + "item_count": 2, + "soft_pair_count": 1, + "serialized_chars": 1590, + "estimated_input_tokens": 398, + "estimated_eval_tokens": 1052 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": false, + "analyst_result": { + "summary": "#47138 and #47139 look like the same issue report, but the evidence is limited to identical titles, matching body length, and very similar metadata. #47138 is the better canonical anchor because it was created first and has slightly more inbound references.", + "confidence": 0.88, + "canonical_issue_reason": "#47138 is the earliest-created report and has marginally more inbound references, so it is the best anchor for consolidating discussion.", + "canonical_pr_reason": null, + "best_issue_reason": "#47138 is the clearest representative because it appears first and has the stronger reference signal.", + "best_pr_reason": null, + "soft_edge_verdicts": [ + { + "left": "issue:47138", + "right": "issue:47139", + "accept": true, + "reason": "The two issues have the same title and matching body length, with creation times only seconds apart, which is enough to treat them as the same underlying report, though the evidence is not exhaustive." + } + ] + }, + "evaluator_result": { + "accept": false, + "feedback": "The pair is plausibly the same issue, but the analyst overstates the evidence by calling them \u201cexact duplicates\u201d and \u201cnot less complete\u201d when the packet only shows matching titles/body lengths and similar metadata, not body text. The canonical choice of #47138 is grounded by earlier creation time and slightly higher inbound references, but the soft-edge verdict should be more conservative." + }, + "evaluator_used": true, + "retried": true, + "accepted_nontrivial_soft_edge": true, + "error_kind": null, + "error_message": null + }, + { + "label": "pull_request", + "component_index": 1, + "component_count": 31, + "review_unit_index": 1, + "review_unit_count": 4, + "status": "reviewed", + "reason": null, + "source": "llm", + "cache_hit": false, + "model": "gpt-5.4-mini?service_tier=flex", + "cluster_id": "cluster-40546-79", + "nodes": [ + "pull_request:43251", + "pull_request:43254", + "pull_request:44775", + "pull_request:45157", + "pull_request:46196", + "pull_request:46266", + "pull_request:46395", + "pull_request:46836", + "pull_request:46903", + "pull_request:46912", + "pull_request:46954", + "pull_request:46957", + "pull_request:46958", + "pull_request:46974", + "pull_request:46995", + "pull_request:47003", + "pull_request:47077", + "pull_request:47078", + "pull_request:47082", + "pull_request:47090", + "pull_request:47093", + "pull_request:47105", + "pull_request:47108", + "pull_request:47129", + "pull_request:47163", + "pull_request:47166", + "pull_request:47173", + "pull_request:47179", + "pull_request:47184", + "pull_request:47202", + "pull_request:47204", + "pull_request:47210", + "pull_request:47212", + "pull_request:47216", + "pull_request:47243", + "pull_request:47268", + "pull_request:47275", + "pull_request:47284", + "pull_request:47291", + "pull_request:47292", + "pull_request:47295", + "pull_request:47301", + "pull_request:47336", + "pull_request:47360", + "pull_request:47374", + "pull_request:47383", + "pull_request:47384", + "pull_request:47395" + ], + "soft_pairs": [ + "pull_request:43251|pull_request:43254", + "pull_request:46903|pull_request:47090", + "pull_request:47090|pull_request:47395", + "pull_request:46903|pull_request:46958", + "pull_request:46903|pull_request:47395", + "pull_request:47268|pull_request:47284", + "pull_request:46958|pull_request:46974", + "pull_request:46954|pull_request:46958", + "pull_request:47173|pull_request:47284", + "pull_request:47202|pull_request:47204", + "pull_request:47204|pull_request:47374", + "pull_request:47284|pull_request:47292", + "pull_request:47108|pull_request:47202", + "pull_request:47077|pull_request:47078", + "pull_request:47204|pull_request:47243", + "pull_request:47202|pull_request:47360", + "pull_request:47108|pull_request:47360", + "pull_request:47105|pull_request:47129", + "pull_request:47301|pull_request:47360", + "pull_request:47077|pull_request:47212", + "pull_request:46266|pull_request:46957", + "pull_request:47360|pull_request:47383", + "pull_request:47173|pull_request:47184", + "pull_request:47275|pull_request:47360", + "pull_request:47291|pull_request:47384", + "pull_request:47166|pull_request:47173", + "pull_request:47077|pull_request:47275", + "pull_request:44775|pull_request:47275", + "pull_request:47275|pull_request:47301", + "pull_request:46912|pull_request:47179", + "pull_request:47216|pull_request:47275", + "pull_request:46196|pull_request:47202", + "pull_request:46912|pull_request:47284", + "pull_request:47163|pull_request:47284", + "pull_request:46836|pull_request:47202", + "pull_request:46196|pull_request:47210", + "pull_request:47003|pull_request:47204", + "pull_request:47093|pull_request:47275", + "pull_request:45157|pull_request:47093", + "pull_request:46995|pull_request:47163", + "pull_request:47163|pull_request:47173", + "pull_request:46836|pull_request:47275", + "pull_request:45157|pull_request:46395", + "pull_request:46912|pull_request:47082", + "pull_request:47163|pull_request:47384", + "pull_request:47204|pull_request:47295", + "pull_request:44775|pull_request:47301", + "pull_request:46196|pull_request:46995", + "pull_request:47301|pull_request:47384", + "pull_request:47216|pull_request:47384", + "pull_request:47077|pull_request:47301", + "pull_request:47336|pull_request:47384", + "pull_request:44775|pull_request:47077", + "pull_request:47166|pull_request:47292" + ], + "prepared_review_unit_hash": "4afcec023b6365c247e70cfb348d23603532deedb6d51c2c8630c719ae809ba4", + "component_budget": { + "node_count": 79, + "item_count": 79, + "soft_pair_count": 214, + "serialized_chars": 270269, + "estimated_input_tokens": 67568, + "estimated_eval_tokens": 135392 + }, + "budget": { + "node_count": 48, + "item_count": 48, + "soft_pair_count": 54, + "serialized_chars": 126518, + "estimated_input_tokens": 31630, + "estimated_eval_tokens": 63516 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": true, + "analyst_result": { + "summary": "Most candidate pairs are clearly unrelated. The only strong duplicate cluster is the cross-entropy / double-shift loss fixes around CohereASR/PPFormulaNet; I treated PR 47090 as the best representative because it uses the shared loss path cleanly and covers the broader affected models.", + "confidence": 0.81, + "canonical_issue_reason": null, + "canonical_pr_reason": "PR 47090 is the best representative of the duplicate loss-fix cluster: it fixes the double-shift bug through the shared loss helper, keeps gradient-accumulation behavior intact, and covers the broader set of affected models.", + "best_issue_reason": null, + "best_pr_reason": "PR 47090 is the strongest overall PR candidate among the near-duplicates because it is the most complete and least ad hoc fix for the training-loss double-shift problem.", + "soft_edge_verdicts": [ + { + "left": "pull_request:43251", + "right": "pull_request:43254", + "accept": true, + "reason": "Same underlying cross_entropy kwargs bug in the same utility function; both patch the same code path." + }, + { + "left": "pull_request:46903", + "right": "pull_request:47090", + "accept": true, + "reason": "Same double-shift training-loss bug in the CohereASR/PPFormulaNet loss path; 47090 is a broader refinement of the same fix." + }, + { + "left": "pull_request:47090", + "right": "pull_request:47395", + "accept": false, + "reason": "Both are loss-shift bugs, but they affect different models/code paths and are not the same concrete change." + }, + { + "left": "pull_request:46903", + "right": "pull_request:46958", + "accept": true, + "reason": "Same CohereASR/PPFormulaNet double-shift issue; the implementation differs, but the bug is the same." + }, + { + "left": "pull_request:46903", + "right": "pull_request:47395", + "accept": false, + "reason": "Different model and different forward/loss path, so not the same underlying fix." + }, + { + "left": "pull_request:47268", + "right": "pull_request:47284", + "accept": false, + "reason": "Unrelated test failures in different models and subsystems." + }, + { + "left": "pull_request:46958", + "right": "pull_request:46974", + "accept": false, + "reason": "CohereASR loss fix vs a Musicgen generate typo; unrelated." + }, + { + "left": "pull_request:46954", + "right": "pull_request:46958", + "accept": false, + "reason": "Pytest-v9 compatibility change vs a training-loss bug fix; not the same change." + }, + { + "left": "pull_request:47173", + "right": "pull_request:47284", + "accept": false, + "reason": "Different model test updates with no shared underlying bug." + }, + { + "left": "pull_request:47202", + "right": "pull_request:47204", + "accept": false, + "reason": "Both touch docker/export infra, but one removes a builder and the other adds exporter deps; not the same change." + }, + { + "left": "pull_request:47204", + "right": "pull_request:47374", + "accept": false, + "reason": "Exporter test dependencies vs exporter docs; related area, but not the same bug/change." + }, + { + "left": "pull_request:47284", + "right": "pull_request:47292", + "accept": false, + "reason": "Different model test fixes; no shared concrete issue." + }, + { + "left": "pull_request:47108", + "right": "pull_request:47202", + "accept": false, + "reason": "Doc-builder pin comment tweak vs CI docker-image builder removal; unrelated." + }, + { + "left": "pull_request:47077", + "right": "pull_request:47078", + "accept": false, + "reason": "Both are Japanese docs edits, but they are different documentation changes, not one underlying bug." + }, + { + "left": "pull_request:47204", + "right": "pull_request:47243", + "accept": false, + "reason": "CPU exporter test deps vs Executorch exporter code fixes; same broad area, different changes." + }, + { + "left": "pull_request:47202", + "right": "pull_request:47360", + "accept": false, + "reason": "Different CI workflow changes with no shared concrete bug." + }, + { + "left": "pull_request:47105", + "right": "pull_request:47129", + "accept": false, + "reason": "Test assertion fix vs RT-DETR v2 converter update; unrelated." + }, + { + "left": "pull_request:47301", + "right": "pull_request:47360", + "accept": false, + "reason": "MTP docs vs OTEL CI workflow; unrelated." + }, + { + "left": "pull_request:46266", + "right": "pull_request:46957", + "accept": false, + "reason": "DINO model/conversion work vs finegrained FP8 decode optimization; unrelated." + }, + { + "left": "pull_request:47360", + "right": "pull_request:47383", + "accept": false, + "reason": "CI workflow change vs Kimi integration tests; unrelated." + }, + { + "left": "pull_request:47173", + "right": "pull_request:47184", + "accept": false, + "reason": "Both are test-related, but for different models and different platform coverage." + }, + { + "left": "pull_request:47275", + "right": "pull_request:47360", + "accept": false, + "reason": "Model support PR vs CI workflow PR; unrelated." + }, + { + "left": "pull_request:47291", + "right": "pull_request:47384", + "accept": false, + "reason": "Generation-config deprecation vs Gemma 4 heterogeneous attention config; different issues." + }, + { + "left": "pull_request:47166", + "right": "pull_request:47173", + "accept": false, + "reason": "Different test fixes in different areas; not the same bug." + }, + { + "left": "pull_request:47077", + "right": "pull_request:47275", + "accept": false, + "reason": "Japanese docs translation vs new model support; unrelated." + }, + { + "left": "pull_request:44775", + "right": "pull_request:47275", + "accept": false, + "reason": "Docs navigation update vs model addition; unrelated." + }, + { + "left": "pull_request:47275", + "right": "pull_request:47301", + "accept": false, + "reason": "EfficientViTSam model support vs assisted-decoding docs; unrelated." + }, + { + "left": "pull_request:46912", + "right": "pull_request:47179", + "accept": false, + "reason": "RoPE test expansion vs new Granite model support; not the same change." + }, + { + "left": "pull_request:47216", + "right": "pull_request:47275", + "accept": false, + "reason": "Compressed-tensors quantization work vs EfficientViTSam model support; unrelated." + }, + { + "left": "pull_request:46196", + "right": "pull_request:47202", + "accept": false, + "reason": "SDPA model dispatch support vs docker builder cleanup; unrelated." + }, + { + "left": "pull_request:47093", + "right": "pull_request:47275", + "accept": false, + "reason": "OpenPangu-V2 model addition vs EfficientViTSam model addition; different model families." + }, + { + "left": "pull_request:45157", + "right": "pull_request:47093", + "accept": false, + "reason": "PrismML GGUF support vs OpenPangu-V2 model support; unrelated." + }, + { + "left": "pull_request:46995", + "right": "pull_request:47163", + "accept": false, + "reason": "Flash linear attention availability check vs XPU paged-attention tests; different concerns." + }, + { + "left": "pull_request:47163", + "right": "pull_request:47173", + "accept": false, + "reason": "Both target XPU test coverage, but they fix different tests and failure modes." + }, + { + "left": "pull_request:46836", + "right": "pull_request:47202", + "accept": false, + "reason": "Unlimited OCR model addition vs docker builder removal; unrelated." + }, + { + "left": "pull_request:46196", + "right": "pull_request:47210", + "accept": false, + "reason": "General SDPA dispatch support vs packed-row varlen routing; related area but different concrete changes." + }, + { + "left": "pull_request:47003", + "right": "pull_request:47204", + "accept": false, + "reason": "OpenVINO exporter implementation vs CPU test image dependency changes; not the same fix." + }, + { + "left": "pull_request:47093", + "right": "pull_request:47275", + "accept": false, + "reason": "Different model additions with no shared underlying bug." + }, + { + "left": "pull_request:45157", + "right": "pull_request:47093", + "accept": false, + "reason": "Different model support/conversion efforts." + }, + { + "left": "pull_request:46912", + "right": "pull_request:47082", + "accept": false, + "reason": "RoPE test work vs RoMa model support; unrelated." + }, + { + "left": "pull_request:47163", + "right": "pull_request:47384", + "accept": false, + "reason": "XPU test adjustments vs Gemma 4 config refactor; unrelated." + }, + { + "left": "pull_request:47204", + "right": "pull_request:47295", + "accept": false, + "reason": "Exporter dependency changes vs DiffusionGemma mask casting; unrelated." + }, + { + "left": "pull_request:44775", + "right": "pull_request:47301", + "accept": false, + "reason": "Docs for parallelism vs docs for assisted decoding; different topics." + }, + { + "left": "pull_request:46196", + "right": "pull_request:46995", + "accept": false, + "reason": "Different attention-backend availability and model dispatch changes." + }, + { + "left": "pull_request:47301", + "right": "pull_request:47384", + "accept": false, + "reason": "Assisted decoding docs vs Gemma 4 config refactor; unrelated." + }, + { + "left": "pull_request:47216", + "right": "pull_request:47384", + "accept": false, + "reason": "Quantization docs/code vs Gemma 4 config refactor; unrelated." + }, + { + "left": "pull_request:47077", + "right": "pull_request:47275", + "accept": false, + "reason": "Docs translation vs model support; unrelated." + }, + { + "left": "pull_request:47275", + "right": "pull_request:47301", + "accept": false, + "reason": "Model addition vs docs for MTP; unrelated." + }, + { + "left": "pull_request:46912", + "right": "pull_request:47275", + "accept": false, + "reason": "RoPE tests vs EfficientViTSam model support; unrelated." + }, + { + "left": "pull_request:47291", + "right": "pull_request:47384", + "accept": false, + "reason": "Continuous batching deprecation vs Gemma 4 heterogeneous attention config; different changes." + } + ] + }, + "evaluator_result": null, + "evaluator_used": false, + "retried": false, + "accepted_nontrivial_soft_edge": true, + "error_kind": null, + "error_message": null + }, + { + "label": "pull_request", + "component_index": 1, + "component_count": 31, + "review_unit_index": 2, + "review_unit_count": 4, + "status": "reviewed", + "reason": null, + "source": "llm", + "cache_hit": false, + "model": "gpt-5.4-mini?service_tier=flex", + "cluster_id": "cluster-40546-79", + "nodes": [ + "pull_request:41855", + "pull_request:43424", + "pull_request:43451", + "pull_request:44517", + "pull_request:44775", + "pull_request:45157", + "pull_request:46196", + "pull_request:46266", + "pull_request:46267", + "pull_request:46381", + "pull_request:46509", + "pull_request:46658", + "pull_request:46836", + "pull_request:46895", + "pull_request:46912", + "pull_request:46957", + "pull_request:46995", + "pull_request:47077", + "pull_request:47078", + "pull_request:47082", + "pull_request:47093", + "pull_request:47096", + "pull_request:47105", + "pull_request:47108", + "pull_request:47113", + "pull_request:47143", + "pull_request:47163", + "pull_request:47166", + "pull_request:47175", + "pull_request:47179", + "pull_request:47184", + "pull_request:47210", + "pull_request:47216", + "pull_request:47243", + "pull_request:47275", + "pull_request:47291", + "pull_request:47294", + "pull_request:47298", + "pull_request:47301", + "pull_request:47304", + "pull_request:47306", + "pull_request:47320", + "pull_request:47322", + "pull_request:47336", + "pull_request:47352", + "pull_request:47360", + "pull_request:47367", + "pull_request:47383" + ], + "soft_pairs": [ + "pull_request:46267|pull_request:47383", + "pull_request:47294|pull_request:47320", + "pull_request:46509|pull_request:46995", + "pull_request:46995|pull_request:47143", + "pull_request:47093|pull_request:47143", + "pull_request:46266|pull_request:47301", + "pull_request:44775|pull_request:47304", + "pull_request:46266|pull_request:47383", + "pull_request:47082|pull_request:47210", + "pull_request:46196|pull_request:47179", + "pull_request:47093|pull_request:47163", + "pull_request:47179|pull_request:47210", + "pull_request:47082|pull_request:47108", + "pull_request:46266|pull_request:47216", + "pull_request:47108|pull_request:47301", + "pull_request:47078|pull_request:47275", + "pull_request:47082|pull_request:47275", + "pull_request:46266|pull_request:47336", + "pull_request:46995|pull_request:47301", + "pull_request:47175|pull_request:47179", + "pull_request:44775|pull_request:46836", + "pull_request:47082|pull_request:47216", + "pull_request:47163|pull_request:47179", + "pull_request:44517|pull_request:45157", + "pull_request:47275|pull_request:47336", + "pull_request:47210|pull_request:47304", + "pull_request:47108|pull_request:47383", + "pull_request:46957|pull_request:47216", + "pull_request:47216|pull_request:47336", + "pull_request:46912|pull_request:47078", + "pull_request:47179|pull_request:47301", + "pull_request:47166|pull_request:47352", + "pull_request:46912|pull_request:47294", + "pull_request:47184|pull_request:47352", + "pull_request:47143|pull_request:47320", + "pull_request:47175|pull_request:47294", + "pull_request:46895|pull_request:47306", + "pull_request:46196|pull_request:47301", + "pull_request:47105|pull_request:47298", + "pull_request:45157|pull_request:46658", + "pull_request:47275|pull_request:47304", + "pull_request:47078|pull_request:47322", + "pull_request:44517|pull_request:47301", + "pull_request:47078|pull_request:47243", + "pull_request:47294|pull_request:47360", + "pull_request:46912|pull_request:47291", + "pull_request:47113|pull_request:47304", + "pull_request:47163|pull_request:47352", + "pull_request:43451|pull_request:44517", + "pull_request:47352|pull_request:47383", + "pull_request:44517|pull_request:47383", + "pull_request:43451|pull_request:47352", + "pull_request:44517|pull_request:47143", + "pull_request:47243|pull_request:47294", + "pull_request:46957|pull_request:47383", + "pull_request:47210|pull_request:47291", + "pull_request:46957|pull_request:47360", + "pull_request:46895|pull_request:47298", + "pull_request:46381|pull_request:47322", + "pull_request:47113|pull_request:47175", + "pull_request:46957|pull_request:47301", + "pull_request:47113|pull_request:47243", + "pull_request:47113|pull_request:47275", + "pull_request:46912|pull_request:47113", + "pull_request:47077|pull_request:47367", + "pull_request:47294|pull_request:47306", + "pull_request:46658|pull_request:46895", + "pull_request:47275|pull_request:47367", + "pull_request:43451|pull_request:46381", + "pull_request:46381|pull_request:47096", + "pull_request:47210|pull_request:47367", + "pull_request:41855|pull_request:43424" + ], + "prepared_review_unit_hash": "1b5522e50d918c29d1dca5be6a32f4306fbc577e3006e3463ace6669b23ecadb", + "component_budget": { + "node_count": 79, + "item_count": 79, + "soft_pair_count": 214, + "serialized_chars": 270269, + "estimated_input_tokens": 67568, + "estimated_eval_tokens": 135392 + }, + "budget": { + "node_count": 48, + "item_count": 48, + "soft_pair_count": 72, + "serialized_chars": 138861, + "estimated_input_tokens": 34716, + "estimated_eval_tokens": 69688 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": true, + "analyst_result": { + "summary": "This cluster is mostly broad topical similarity rather than true duplication: model-support PRs, docs updates, tests, export/quantization fixes, and CI/workflow changes are all mixed together. None of the soft edges look like the same underlying change.", + "confidence": 0.93, + "canonical_issue_reason": null, + "canonical_pr_reason": "No single canonical PR: the cluster is a grab-bag of unrelated model additions, docs, tests, and infra changes, and the soft links are only broad similarity.", + "best_issue_reason": null, + "best_pr_reason": "No global best PR for deduping this cluster; none of the items is a clear superset or duplicate target for the others.", + "soft_edge_verdicts": [ + { + "left": "pull_request:46267", + "right": "pull_request:47383", + "accept": false, + "reason": "Chat parsing/serve changes vs Kimi integration tests; different subsystems and no shared code path." + }, + { + "left": "pull_request:47294", + "right": "pull_request:47320", + "accept": false, + "reason": "Kernel/native causal_conv1d work vs response-schema deprecation; unrelated functionality." + }, + { + "left": "pull_request:46509", + "right": "pull_request:46995", + "accept": false, + "reason": "Nemotron Omni model support vs MLU availability check; model addition vs backend capability." + }, + { + "left": "pull_request:46995", + "right": "pull_request:47143", + "accept": false, + "reason": "Flash-linear-attention availability on MLU vs NeuCodec model support; unrelated changes." + }, + { + "left": "pull_request:47093", + "right": "pull_request:47143", + "accept": false, + "reason": "OpenPangu-V2 model addition vs NeuCodec model addition; different models and codebases." + }, + { + "left": "pull_request:46266", + "right": "pull_request:47301", + "accept": false, + "reason": "DINO modernization/model conversion work vs assisted-decoding documentation; not the same bug/change." + }, + { + "left": "pull_request:44775", + "right": "pull_request:47304", + "accept": false, + "reason": "Docs navigation for N-D parallelism vs MPS graph-cache docs cleanup; unrelated documentation changes." + }, + { + "left": "pull_request:46266", + "right": "pull_request:47383", + "accept": false, + "reason": "DINO-related model modernization vs Kimi integration tests; different model families and purposes." + }, + { + "left": "pull_request:47082", + "right": "pull_request:47210", + "accept": false, + "reason": "RoMa keypoint matcher addition vs packed SDPA varlen routing; different domains and code paths." + }, + { + "left": "pull_request:46196", + "right": "pull_request:47179", + "accept": false, + "reason": "SDPA dispatch support for wav2vec2-family audio models vs Granite SWA model support; unrelated." + }, + { + "left": "pull_request:47093", + "right": "pull_request:47163", + "accept": false, + "reason": "OpenPangu-V2 model support vs XPU paged-attention test coverage; not the same change." + }, + { + "left": "pull_request:47179", + "right": "pull_request:47210", + "accept": false, + "reason": "Granite SWA model support vs packed SDPA varlen attention; different areas and no shared fix." + }, + { + "left": "pull_request:47082", + "right": "pull_request:47108", + "accept": false, + "reason": "RoMa model support vs doc-builder pin comment tweak; unrelated." + }, + { + "left": "pull_request:46266", + "right": "pull_request:47275", + "accept": false, + "reason": "DINO/DINOV2 modernization vs EfficientViTSAM model support; broad vision-model overlap only." + }, + { + "left": "pull_request:47108", + "right": "pull_request:47301", + "accept": false, + "reason": "Doc-builder pin comment vs MTP docs; both docs-related but clearly unrelated changes." + }, + { + "left": "pull_request:47078", + "right": "pull_request:47275", + "accept": false, + "reason": "Japanese doc wording improvements vs EfficientViTSAM model support; different scopes." + }, + { + "left": "pull_request:47082", + "right": "pull_request:47336", + "accept": false, + "reason": "RoMa keypoint matcher support vs fully offloaded model saving; different code paths." + }, + { + "left": "pull_request:46995", + "right": "pull_request:47301", + "accept": false, + "reason": "MLU availability helper vs multi-token prediction docs; no shared implementation." + }, + { + "left": "pull_request:47175", + "right": "pull_request:47179", + "accept": false, + "reason": "Olmo3 CI expectation updates vs Granite SWA model support; test data vs model implementation." + }, + { + "left": "pull_request:44775", + "right": "pull_request:46836", + "accept": false, + "reason": "Both touch docs trees, but one is parallelism docs and the other is Unlimited OCR model support; not a duplicate." + }, + { + "left": "pull_request:47082", + "right": "pull_request:47216", + "accept": false, + "reason": "RoMa keypoint matcher vs compressed-tensors fp8 changes; unrelated vision/model vs quantization work." + }, + { + "left": "pull_request:47163", + "right": "pull_request:47179", + "accept": false, + "reason": "Paged-attention XPU tests vs Granite SWA model support; test coverage vs new model code." + }, + { + "left": "pull_request:44517", + "right": "pull_request:45157", + "accept": false, + "reason": "Qwen3-TTS model support vs PrismML Bonsai support; different models and implementation paths." + }, + { + "left": "pull_request:47275", + "right": "pull_request:47336", + "accept": false, + "reason": "EfficientViTSAM model addition vs offloading save fix; unrelated." + }, + { + "left": "pull_request:47210", + "right": "pull_request:47304", + "accept": false, + "reason": "Packed SDPA varlen routing vs MPS graph-cache docs cleanup; unrelated." + }, + { + "left": "pull_request:47108", + "right": "pull_request:47383", + "accept": false, + "reason": "Doc-builder pin comment vs Kimi integration tests; no overlap beyond being repository maintenance." + }, + { + "left": "pull_request:46957", + "right": "pull_request:47216", + "accept": false, + "reason": "Finegrained FP8 decode changes vs compressed-tensors fp8 docs/runtime work; both quantization-adjacent but not the same concrete fix." + }, + { + "left": "pull_request:46912", + "right": "pull_request:47078", + "accept": false, + "reason": "RoPE test updates vs Japanese wording/doc translation; unrelated." + }, + { + "left": "pull_request:47179", + "right": "pull_request:47301", + "accept": false, + "reason": "Granite SWA model support vs MTP docs; model implementation vs generation documentation." + }, + { + "left": "pull_request:47166", + "right": "pull_request:47352", + "accept": false, + "reason": "Quantization test tweaks vs distributed runtime utilities; separate test areas and different subsystems." + }, + { + "left": "pull_request:46196", + "right": "pull_request:47301", + "accept": false, + "reason": "SDPA dispatch support vs assisted-decoding/MTP docs; unrelated." + }, + { + "left": "pull_request:45157", + "right": "pull_request:46658", + "accept": false, + "reason": "PrismML support vs Step-3.7 Flash checkpoint builder; different model families and tasks." + }, + { + "left": "pull_request:47275", + "right": "pull_request:47304", + "accept": false, + "reason": "EfficientViTSAM model support vs MPS graph-cache docs cleanup; unrelated." + }, + { + "left": "pull_request:47078", + "right": "pull_request:47322", + "accept": false, + "reason": "Japanese documentation wording vs AMD Quark config class changes; different domains." + }, + { + "left": "pull_request:44517", + "right": "pull_request:47301", + "accept": false, + "reason": "Qwen3-TTS model support vs MTP documentation; unrelated." + }, + { + "left": "pull_request:47078", + "right": "pull_request:47243", + "accept": false, + "reason": "Japanese docs wording vs Executorch exporter fixes; unrelated changes." + }, + { + "left": "pull_request:47294", + "right": "pull_request:47291", + "accept": false, + "reason": "Kernel-native causal_conv1d functions vs generation config deprecation; separate features." + }, + { + "left": "pull_request:46957", + "right": "pull_request:47336", + "accept": false, + "reason": "Finegrained FP8 routing changes vs fully offloaded save-pretrained fix; unrelated." + }, + { + "left": "pull_request:47216", + "right": "pull_request:47336", + "accept": false, + "reason": "Compressed-tensors fp8 support vs offloading save fix; different subsystems." + }, + { + "left": "pull_request:46912", + "right": "pull_request:47113", + "accept": false, + "reason": "RoPE testing updates vs audio requirements fixes; unrelated." + }, + { + "left": "pull_request:47113", + "right": "pull_request:47304", + "accept": false, + "reason": "Audio feature/proc requirement fixes vs MPS graph-cache docs cleanup; unrelated." + }, + { + "left": "pull_request:43451", + "right": "pull_request:44517", + "accept": false, + "reason": "Molmo2 model support vs Qwen3-TTS model support; both add models, but they are different model families and code paths." + }, + { + "left": "pull_request:47275", + "right": "pull_request:47294", + "accept": false, + "reason": "EfficientViTSAM model addition vs native causal_conv1d kernel functions; unrelated." + }, + { + "left": "pull_request:47077", + "right": "pull_request:47367", + "accept": false, + "reason": "Japanese DistilBERT translation vs pipeline device-placement clarification; different documentation topics and code impact." + }, + { + "left": "pull_request:46957", + "right": "pull_request:47383", + "accept": false, + "reason": "Finegrained FP8 generation optimizations vs Kimi integration tests; unrelated." + }, + { + "left": "pull_request:45157", + "right": "pull_request:46658", + "accept": false, + "reason": "PrismML Bonsai support vs Step-3.7 Flash utilities/model builder; different models and mechanisms." + }, + { + "left": "pull_request:47275", + "right": "pull_request:47304", + "accept": false, + "reason": "EfficientViTSAM model support vs MPS graph-cache docs cleanup; not the same change." + }, + { + "left": "pull_request:47078", + "right": "pull_request:47322", + "accept": false, + "reason": "Japanese doc wording changes vs Quark config-class update; unrelated." + }, + { + "left": "pull_request:47113", + "right": "pull_request:47352", + "accept": false, + "reason": "Audio feature/proc requirement fixes vs distributed runtime utils; different subsystems." + }, + { + "left": "pull_request:43451", + "right": "pull_request:44517", + "accept": false, + "reason": "Molmo2 and Qwen3-TTS are distinct model additions; no shared fix or mergeable overlap." + }, + { + "left": "pull_request:47243", + "right": "pull_request:47294", + "accept": false, + "reason": "Executorch exporter fixes vs kernel-native causal_conv1d functions; unrelated exporter vs model kernel changes." + }, + { + "left": "pull_request:46957", + "right": "pull_request:47301", + "accept": false, + "reason": "Quantization/generation optimization work vs MTP docs; no shared concrete issue." + }, + { + "left": "pull_request:47108", + "right": "pull_request:47360", + "accept": false, + "reason": "Doc-builder pin-comment change vs OTEL CI workflow pilot; different maintenance areas." + }, + { + "left": "pull_request:47210", + "right": "pull_request:47367", + "accept": false, + "reason": "Packed SDPA varlen attention routing vs pipeline device-placement docs; unrelated." + }, + { + "left": "pull_request:46912", + "right": "pull_request:47291", + "accept": false, + "reason": "RoPE test updates vs generation-config deprecation; different code paths." + }, + { + "left": "pull_request:47184", + "right": "pull_request:47352", + "accept": false, + "reason": "Qwen3 Next test adjustments vs distributed runtime utils; test coverage for a model vs runtime infrastructure." + }, + { + "left": "pull_request:47143", + "right": "pull_request:47383", + "accept": false, + "reason": "NeuCodec model support vs Kimi integration tests; different models and purposes." + }, + { + "left": "pull_request:44517", + "right": "pull_request:47383", + "accept": false, + "reason": "Qwen3-TTS model support vs Kimi integration tests; unrelated model families." + }, + { + "left": "pull_request:43451", + "right": "pull_request:47143", + "accept": false, + "reason": "Molmo2 and NeuCodec are separate model-support PRs with different architectures and code paths." + }, + { + "left": "pull_request:47077", + "right": "pull_request:47367", + "accept": false, + "reason": "Japanese translation for DistilBERT vs pipeline device clarification; unrelated documentation work." + }, + { + "left": "pull_request:41855", + "right": "pull_request:43424", + "accept": false, + "reason": "Mistral tokenizer methods vs executorch dynamic-shape export tests; unrelated tokenizer API vs export validation." + } + ] + }, + "evaluator_result": null, + "evaluator_used": false, + "retried": false, + "accepted_nontrivial_soft_edge": false, + "error_kind": null, + "error_message": null + }, + { + "label": "pull_request", + "component_index": 1, + "component_count": 31, + "review_unit_index": 3, + "review_unit_count": 4, + "status": "reviewed", + "reason": null, + "source": "llm", + "cache_hit": false, + "model": "gpt-5.4-mini?service_tier=flex", + "cluster_id": "cluster-40546-79", + "nodes": [ + "pull_request:40546", + "pull_request:41855", + "pull_request:42668", + "pull_request:43251", + "pull_request:43424", + "pull_request:43451", + "pull_request:43823", + "pull_request:43995", + "pull_request:44827", + "pull_request:45157", + "pull_request:45170", + "pull_request:45213", + "pull_request:45655", + "pull_request:46048", + "pull_request:46196", + "pull_request:46509", + "pull_request:46658", + "pull_request:46836", + "pull_request:46895", + "pull_request:46912", + "pull_request:46958", + "pull_request:47003", + "pull_request:47077", + "pull_request:47090", + "pull_request:47096", + "pull_request:47143", + "pull_request:47163", + "pull_request:47166", + "pull_request:47173", + "pull_request:47175", + "pull_request:47184", + "pull_request:47210", + "pull_request:47216", + "pull_request:47243", + "pull_request:47291", + "pull_request:47292", + "pull_request:47294", + "pull_request:47298", + "pull_request:47301", + "pull_request:47304", + "pull_request:47306", + "pull_request:47322", + "pull_request:47336", + "pull_request:47367", + "pull_request:47383", + "pull_request:47387" + ], + "soft_pairs": [ + "pull_request:41855|pull_request:43823", + "pull_request:42668|pull_request:43823", + "pull_request:43424|pull_request:47301", + "pull_request:43424|pull_request:47383", + "pull_request:43823|pull_request:45655", + "pull_request:44827|pull_request:47383", + "pull_request:44827|pull_request:45170", + "pull_request:43424|pull_request:44827", + "pull_request:43823|pull_request:43995", + "pull_request:43251|pull_request:43823", + "pull_request:44827|pull_request:46048", + "pull_request:43424|pull_request:43823", + "pull_request:43251|pull_request:44827", + "pull_request:44827|pull_request:45655", + "pull_request:43995|pull_request:45213", + "pull_request:44827|pull_request:45213", + "pull_request:43995|pull_request:44827", + "pull_request:43823|pull_request:45213", + "pull_request:46895|pull_request:47090", + "pull_request:46958|pull_request:47090", + "pull_request:46895|pull_request:46958", + "pull_request:47173|pull_request:47292", + "pull_request:47166|pull_request:47184", + "pull_request:46912|pull_request:47175", + "pull_request:47096|pull_request:47383", + "pull_request:47184|pull_request:47292", + "pull_request:47301|pull_request:47304", + "pull_request:47077|pull_request:47304", + "pull_request:46658|pull_request:46836", + "pull_request:47210|pull_request:47301", + "pull_request:47003|pull_request:47243", + "pull_request:46509|pull_request:47143", + "pull_request:47163|pull_request:47166", + "pull_request:47163|pull_request:47184", + "pull_request:46509|pull_request:47163", + "pull_request:45157|pull_request:46836", + "pull_request:46836|pull_request:47387", + "pull_request:47163|pull_request:47292", + "pull_request:47291|pull_request:47322", + "pull_request:45157|pull_request:46509", + "pull_request:47301|pull_request:47336", + "pull_request:47216|pull_request:47383", + "pull_request:47336|pull_request:47383", + "pull_request:46509|pull_request:47387", + "pull_request:47304|pull_request:47367", + "pull_request:46912|pull_request:47298", + "pull_request:47301|pull_request:47383", + "pull_request:47216|pull_request:47301", + "pull_request:43451|pull_request:46836", + "pull_request:46509|pull_request:47301", + "pull_request:47175|pull_request:47298", + "pull_request:47163|pull_request:47210", + "pull_request:47294|pull_request:47298", + "pull_request:45157|pull_request:47387", + "pull_request:46658|pull_request:47387", + "pull_request:47143|pull_request:47163", + "pull_request:46196|pull_request:47143", + "pull_request:46658|pull_request:47143", + "pull_request:46658|pull_request:47322", + "pull_request:47143|pull_request:47387", + "pull_request:47143|pull_request:47322", + "pull_request:45157|pull_request:47143", + "pull_request:47301|pull_request:47367", + "pull_request:43451|pull_request:47243", + "pull_request:47096|pull_request:47243", + "pull_request:43451|pull_request:47291", + "pull_request:43451|pull_request:47175", + "pull_request:47166|pull_request:47383", + "pull_request:47184|pull_request:47383", + "pull_request:40546|pull_request:43451", + "pull_request:47306|pull_request:47367", + "pull_request:43451|pull_request:47143" + ], + "prepared_review_unit_hash": "e6f5cf15cf4e5b81fdccf3c653383a4b12ff9b88830755b6b332be6fa05a1cde", + "component_budget": { + "node_count": 79, + "item_count": 79, + "soft_pair_count": 214, + "serialized_chars": 270269, + "estimated_input_tokens": 67568, + "estimated_eval_tokens": 135392 + }, + "budget": { + "node_count": 46, + "item_count": 46, + "soft_pair_count": 72, + "serialized_chars": 129459, + "estimated_input_tokens": 32365, + "estimated_eval_tokens": 64986 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": true, + "analyst_result": { + "summary": "There is one real duplicate cluster around CohereASR training-loss double-shift handling; most other soft pairs are only superficially similar (same subsystem, docs, tests, or model-add PRs) and should stay separate.", + "confidence": 0.91, + "canonical_issue_reason": null, + "canonical_pr_reason": "47090 is the best canonical PR: it explicitly targets the CohereASR double-shift bug (#46894) and is the clearest, most complete version of that fix, superseding the earlier variants in 46895 and 46958.", + "best_issue_reason": null, + "best_pr_reason": "47090 is the strongest overall canonical PR candidate because it is the only clearly duplicated fix cluster here with an explicit issue link and a concrete, mergeable code-path change.", + "soft_edge_verdicts": [ + { + "left": "pull_request:41855", + "right": "pull_request:43823", + "accept": false, + "reason": "Different features: Mistral tokenizer methods vs adding MobileLLM support. Same broad area of Transformers, but not the same bug or change." + }, + { + "left": "pull_request:42668", + "right": "pull_request:43823", + "accept": false, + "reason": "Both touch model/processor plumbing, but one is processor auto-loading robustness and the other is a new model addition. Not the same underlying change." + }, + { + "left": "pull_request:43424", + "right": "pull_request:47301", + "accept": false, + "reason": "ExecuTorch dynamic-shape export test vs docs for multi-token prediction. No shared concrete code-path problem." + }, + { + "left": "pull_request:43424", + "right": "pull_request:47383", + "accept": false, + "reason": "Dynamic-shape export test and Kimi integration tests are unrelated test additions for different concerns." + }, + { + "left": "pull_request:43823", + "right": "pull_request:45655", + "accept": false, + "reason": "New MobileLLM support vs config-resolution ordering fix. Different code paths and unrelated changes." + }, + { + "left": "pull_request:44827", + "right": "pull_request:47383", + "accept": false, + "reason": "Mistral4 test fixes vs Kimi integration tests. Same general testing domain only." + }, + { + "left": "pull_request:44827", + "right": "pull_request:45170", + "accept": false, + "reason": "Mistral4 test fixes and layrnorm->layernorm renaming are different issues." + }, + { + "left": "pull_request:43424", + "right": "pull_request:44827", + "accept": false, + "reason": "One is an export test, the other is Mistral4 test/runtime cleanup. Not the same bug." + }, + { + "left": "pull_request:43823", + "right": "pull_request:43995", + "accept": false, + "reason": "New MobileLLM model support vs Falcon output-interface refactor. Different functionality." + }, + { + "left": "pull_request:43251", + "right": "pull_request:43823", + "accept": false, + "reason": "Cross-entropy kwargs passthrough fix vs adding MobileLLM support. No shared concrete fix." + }, + { + "left": "pull_request:44827", + "right": "pull_request:46048", + "accept": false, + "reason": "Mistral4 tests and TDT loss kernel support are unrelated." + }, + { + "left": "pull_request:43424", + "right": "pull_request:43823", + "accept": false, + "reason": "Dynamic export testing vs model addition. Same broad area only." + }, + { + "left": "pull_request:43251", + "right": "pull_request:44827", + "accept": false, + "reason": "Loss utility kwargs handling vs Mistral4 tests. Not the same code-path problem." + }, + { + "left": "pull_request:44827", + "right": "pull_request:45655", + "accept": false, + "reason": "Mistral4 test cleanup and config-resolution fix are unrelated." + }, + { + "left": "pull_request:43995", + "right": "pull_request:45213", + "accept": false, + "reason": "Falcon refactor vs model-creation skill. No overlap beyond generic tooling." + }, + { + "left": "pull_request:44827", + "right": "pull_request:45213", + "accept": false, + "reason": "Mistral4 tests vs a model-creation skill PR. Completely different changes." + }, + { + "left": "pull_request:43995", + "right": "pull_request:44827", + "accept": false, + "reason": "Falcon output collection refactor vs Mistral4 test fixes. Different code paths." + }, + { + "left": "pull_request:43823", + "right": "pull_request:45213", + "accept": false, + "reason": "MobileLLM model support vs model-creation skill docs/scripts. Not a duplicate." + }, + { + "left": "pull_request:46895", + "right": "pull_request:47090", + "accept": true, + "reason": "Same CohereASR training-loss double-shift bug; 47090 is just the refined version of the same fix." + }, + { + "left": "pull_request:46958", + "right": "pull_request:47090", + "accept": true, + "reason": "Both patch CohereASR's double-shift loss handling on the same files and issue target." + }, + { + "left": "pull_request:46895", + "right": "pull_request:46958", + "accept": true, + "reason": "They implement the same CohereASR loss fix with only implementation details differing." + }, + { + "left": "pull_request:47173", + "right": "pull_request:47292", + "accept": false, + "reason": "Both are XPU-related test expectation updates, but they target different models and likely different failing cases, so they are not one concrete bug." + }, + { + "left": "pull_request:47166", + "right": "pull_request:47184", + "accept": false, + "reason": "Quantization test updates and multi-accelerator test adjustments are unrelated." + }, + { + "left": "pull_request:46912", + "right": "pull_request:47175", + "accept": false, + "reason": "Both are test-value updates, but they cover different models and different CI issues; not mergeable as one change." + }, + { + "left": "pull_request:47096", + "right": "pull_request:47383", + "accept": false, + "reason": "Exportability fixes for Kimi vs Kimi integration tests; related model name, but different code-path problems." + }, + { + "left": "pull_request:47184", + "right": "pull_request:47292", + "accept": false, + "reason": "Both tweak accelerator-specific tests, but for different models and failures." + }, + { + "left": "pull_request:47301", + "right": "pull_request:47304", + "accept": false, + "reason": "Docs for multi-token prediction vs a docs tweak about pipeline device placement. Different topics." + }, + { + "left": "pull_request:47077", + "right": "pull_request:47304", + "accept": false, + "reason": "Japanese DistilBERT docs translation and MPS graph-cache docs are unrelated docs changes." + }, + { + "left": "pull_request:46658", + "right": "pull_request:46836", + "accept": false, + "reason": "Step3.7 Flash model work vs Unlimited OCR model work. Both are model additions, not duplicates." + }, + { + "left": "pull_request:47210", + "right": "pull_request:47301", + "accept": false, + "reason": "Packed-SDPA varlen attention routing vs MTP docs. Completely different areas." + }, + { + "left": "pull_request:47003", + "right": "pull_request:47243", + "accept": false, + "reason": "OpenVINO exporter support vs ExecuTorch exporter fixes. Same exporter family, but different concrete exporter paths." + }, + { + "left": "pull_request:46509", + "right": "pull_request:46836", + "accept": false, + "reason": "Nemotron Omni model support and Unlimited OCR model support are separate model integrations." + }, + { + "left": "pull_request:47163", + "right": "pull_request:47166", + "accept": false, + "reason": "Paged-attention XPU tests vs quantization test extension. Different subsystems." + }, + { + "left": "pull_request:47163", + "right": "pull_request:47184", + "accept": false, + "reason": "Paged-attention XPU tests and qwen3_next accelerator test changes are unrelated." + }, + { + "left": "pull_request:46509", + "right": "pull_request:47163", + "accept": false, + "reason": "Nemotron Omni model support vs paged-attention XPU test coverage. Different changes." + }, + { + "left": "pull_request:45157", + "right": "pull_request:46836", + "accept": false, + "reason": "PrismML Bonsai support and Unlimited OCR support are different model additions." + }, + { + "left": "pull_request:46836", + "right": "pull_request:47387", + "accept": false, + "reason": "Unlimited OCR support vs one-to-many conversion infrastructure. Not the same bug or feature." + }, + { + "left": "pull_request:47163", + "right": "pull_request:47292", + "accept": false, + "reason": "Both are XPU-related test fixes, but for different models and failure modes." + }, + { + "left": "pull_request:47291", + "right": "pull_request:47322", + "accept": false, + "reason": "Continuous batching deprecation/config handling and AMD Quark config-class changes are unrelated quantization/config PRs." + }, + { + "left": "pull_request:45157", + "right": "pull_request:46509", + "accept": false, + "reason": "PrismML Bonsai support and Nemotron Omni support are separate model integrations." + }, + { + "left": "pull_request:47301", + "right": "pull_request:47367", + "accept": false, + "reason": "MTP docs and pipeline device-placement docs are unrelated documentation updates." + }, + { + "left": "pull_request:47216", + "right": "pull_request:47383", + "accept": false, + "reason": "Compressed-tensors fp8 docs/code vs Kimi integration tests. Different domains." + }, + { + "left": "pull_request:47336", + "right": "pull_request:47383", + "accept": false, + "reason": "Offloading save_pretrained fix vs Kimi integration tests. Different problems." + }, + { + "left": "pull_request:46509", + "right": "pull_request:47387", + "accept": false, + "reason": "Nemotron Omni support vs one-to-many conversion support. Different features." + }, + { + "left": "pull_request:47304", + "right": "pull_request:47367", + "accept": false, + "reason": "MPS graph-cache docs and pipeline device-placement docs are separate doc changes." + }, + { + "left": "pull_request:46912", + "right": "pull_request:47298", + "accept": false, + "reason": "Both adjust expectations for hardware-specific test failures, but they touch different models and are not one concrete bug." + }, + { + "left": "pull_request:47301", + "right": "pull_request:47383", + "accept": false, + "reason": "MTP docs vs Kimi integration tests. No overlap." + }, + { + "left": "pull_request:47216", + "right": "pull_request:47301", + "accept": false, + "reason": "Compressed-tensors fp8 work vs MTP docs. Not the same change." + }, + { + "left": "pull_request:43451", + "right": "pull_request:46836", + "accept": false, + "reason": "Molmo2 model addition vs Unlimited OCR model addition. Same broad type only." + }, + { + "left": "pull_request:47143", + "right": "pull_request:47163", + "accept": false, + "reason": "NeuCodec model support vs paged-attention XPU tests. Unrelated." + }, + { + "left": "pull_request:46196", + "right": "pull_request:47143", + "accept": false, + "reason": "wav2vec2 SDPA dispatch support vs NeuCodec model support. Different code paths." + }, + { + "left": "pull_request:46658", + "right": "pull_request:47143", + "accept": false, + "reason": "Step3.7 Flash model work and NeuCodec model support are separate model additions." + }, + { + "left": "pull_request:46658", + "right": "pull_request:47322", + "accept": false, + "reason": "Step3.7 Flash model work vs AMD Quark config changes. Different areas." + }, + { + "left": "pull_request:46658", + "right": "pull_request:47387", + "accept": false, + "reason": "Step3.7 Flash model work vs one-to-many conversion support. Different problems." + }, + { + "left": "pull_request:47143", + "right": "pull_request:47387", + "accept": false, + "reason": "NeuCodec model support vs one-to-many conversion infrastructure. Not a duplicate." + }, + { + "left": "pull_request:47143", + "right": "pull_request:47322", + "accept": false, + "reason": "NeuCodec model support and Quark config-class changes are unrelated." + }, + { + "left": "pull_request:45157", + "right": "pull_request:47143", + "accept": false, + "reason": "PrismML Bonsai support vs NeuCodec support. Different model additions." + }, + { + "left": "pull_request:47301", + "right": "pull_request:47367", + "accept": false, + "reason": "Docs on MTP vs pipeline device placement. Separate documentation topics." + }, + { + "left": "pull_request:43451", + "right": "pull_request:47243", + "accept": false, + "reason": "Molmo2 model addition vs ExecuTorch exporter fixes. Different features." + }, + { + "left": "pull_request:47096", + "right": "pull_request:47243", + "accept": false, + "reason": "Kimi exportability changes vs exporter framework fixes for a different model path. Not the same concrete issue." + }, + { + "left": "pull_request:43451", + "right": "pull_request:47291", + "accept": false, + "reason": "Molmo2 model support and continuous-batching deprecation/config changes are unrelated." + }, + { + "left": "pull_request:43451", + "right": "pull_request:47175", + "accept": false, + "reason": "Molmo2 model support vs OLMo3 CI test updates. Different purposes." + }, + { + "left": "pull_request:47166", + "right": "pull_request:47383", + "accept": false, + "reason": "Quantization test updates vs Kimi integration tests. Different subsystems." + }, + { + "left": "pull_request:47184", + "right": "pull_request:47383", + "accept": false, + "reason": "Accelerator-specific test tweaks vs Kimi integration tests. Not the same bug." + }, + { + "left": "pull_request:40546", + "right": "pull_request:43451", + "accept": false, + "reason": "VibeVoice model support and Molmo2 model support are distinct additions." + }, + { + "left": "pull_request:47306", + "right": "pull_request:47367", + "accept": false, + "reason": "Neuron experiment scripts vs pipeline documentation. Different artifacts and goals." + }, + { + "left": "pull_request:43451", + "right": "pull_request:47143", + "accept": false, + "reason": "Molmo2 and NeuCodec are different model integrations, even though both add support." + } + ] + }, + "evaluator_result": null, + "evaluator_used": false, + "retried": false, + "accepted_nontrivial_soft_edge": false, + "error_kind": null, + "error_message": null + }, + { + "label": "pull_request", + "component_index": 1, + "component_count": 31, + "review_unit_index": 4, + "review_unit_count": 4, + "status": "reviewed", + "reason": null, + "source": "llm", + "cache_hit": false, + "model": "gpt-5.4-mini?service_tier=flex", + "cluster_id": "cluster-40546-79", + "nodes": [ + "pull_request:43451", + "pull_request:46509", + "pull_request:47003", + "pull_request:47096", + "pull_request:47143", + "pull_request:47163", + "pull_request:47298", + "pull_request:47306", + "pull_request:47322", + "pull_request:47383", + "pull_request:47387" + ], + "soft_pairs": [ + "pull_request:43451|pull_request:47163", + "pull_request:43451|pull_request:47383", + "pull_request:43451|pull_request:46509", + "pull_request:43451|pull_request:47096", + "pull_request:43451|pull_request:47003", + "pull_request:47003|pull_request:47096", + "pull_request:43451|pull_request:47322", + "pull_request:47096|pull_request:47322", + "pull_request:43451|pull_request:47298", + "pull_request:47096|pull_request:47298", + "pull_request:43451|pull_request:47306", + "pull_request:47096|pull_request:47306", + "pull_request:47003|pull_request:47306", + "pull_request:47096|pull_request:47387", + "pull_request:47003|pull_request:47387", + "pull_request:47096|pull_request:47143" + ], + "prepared_review_unit_hash": "5d02b0a2fe74b181627c8045d296fa982cf15760c6148dde5ce05215a9002a36", + "component_budget": { + "node_count": 79, + "item_count": 79, + "soft_pair_count": 214, + "serialized_chars": 270269, + "estimated_input_tokens": 67568, + "estimated_eval_tokens": 135392 + }, + "budget": { + "node_count": 11, + "item_count": 11, + "soft_pair_count": 16, + "serialized_chars": 31523, + "estimated_input_tokens": 7881, + "estimated_eval_tokens": 16018 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": true, + "analyst_result": { + "summary": "This cluster is heterogeneous: standalone model additions (Molmo2, Nemotron Omni, NeuCodec), exporter/infra work (OpenVINO exporter, Kimi exportability, one-to-many conversion support), a hardware-specific image-loading fix, and test/experiment-only PRs. They are not duplicate changes and should be split for triage.", + "confidence": 0.98, + "canonical_issue_reason": null, + "canonical_pr_reason": "No single PR canonically represents this cluster; the items span unrelated features and fixes rather than one underlying change.", + "best_issue_reason": null, + "best_pr_reason": "No PR is a good global representative for duplicate triage here; 47096 is the closest exporter-focused item, but it does not subsume the rest of the cluster.", + "soft_edge_verdicts": [ + { + "left": "pull_request:43451", + "right": "pull_request:47163", + "accept": false, + "reason": "Molmo2 model onboarding vs XPU paged-attention test additions; different change types and code paths." + }, + { + "left": "pull_request:43451", + "right": "pull_request:47383", + "accept": false, + "reason": "Molmo2 model onboarding vs Kimi integration tests; unrelated PRs." + }, + { + "left": "pull_request:43451", + "right": "pull_request:46509", + "accept": false, + "reason": "Different model additions (Molmo2 vs Nemotron Omni) with separate configs, docs, and integrations." + }, + { + "left": "pull_request:43451", + "right": "pull_request:47096", + "accept": false, + "reason": "Model addition vs exporter/Kimi exportability work; not the same underlying bug or feature." + }, + { + "left": "pull_request:43451", + "right": "pull_request:47003", + "accept": false, + "reason": "Model onboarding vs OpenVINO exporter implementation; distinct code areas and goals." + }, + { + "left": "pull_request:47003", + "right": "pull_request:47096", + "accept": false, + "reason": "Both touch export-related code, but one adds OpenVINO exporter support and the other changes Kimi exportability/attention handling; not one mergeable fix." + }, + { + "left": "pull_request:43451", + "right": "pull_request:47322", + "accept": false, + "reason": "Molmo2 model onboarding vs AMD Quark config-class rename; unrelated features." + }, + { + "left": "pull_request:47096", + "right": "pull_request:47322", + "accept": false, + "reason": "Kimi exportability changes vs Quark config-class changes; different exporter stacks and problem statements." + }, + { + "left": "pull_request:43451", + "right": "pull_request:47298", + "accept": false, + "reason": "Model onboarding vs MI450 image-decoding fallback fix; unrelated." + }, + { + "left": "pull_request:47096", + "right": "pull_request:47298", + "accept": false, + "reason": "Kimi exporter work vs MI450 image decoder compatibility fixes; different code paths and targets." + }, + { + "left": "pull_request:43451", + "right": "pull_request:47306", + "accept": false, + "reason": "Model onboarding vs Trainium/neuron experiment scripts and tensor-parallel integration; unrelated." + }, + { + "left": "pull_request:47096", + "right": "pull_request:47306", + "accept": false, + "reason": "Exporter/Kimi work vs Neuron training experiments; not the same concrete change." + }, + { + "left": "pull_request:47003", + "right": "pull_request:47306", + "accept": false, + "reason": "OpenVINO exporter work vs Neuron training experiments; different subsystems and objectives." + }, + { + "left": "pull_request:47096", + "right": "pull_request:47387", + "accept": false, + "reason": "Kimi exportability changes vs one-to-many conversion support in model loading; separate core mechanisms." + }, + { + "left": "pull_request:47003", + "right": "pull_request:47387", + "accept": false, + "reason": "OpenVINO exporter additions vs one-to-many model-loading conversions; not the same underlying change." + }, + { + "left": "pull_request:47096", + "right": "pull_request:47143", + "accept": false, + "reason": "Kimi exportability work vs NeuCodec model onboarding; unrelated PRs." + } + ] + }, + "evaluator_result": null, + "evaluator_used": false, + "retried": false, + "accepted_nontrivial_soft_edge": false, + "error_kind": null, + "error_message": null + }, + { + "label": "pull_request", + "component_index": 2, + "component_count": 31, + "review_unit_index": 1, + "review_unit_count": 1, + "status": "reviewed", + "reason": null, + "source": "llm", + "cache_hit": false, + "model": "gpt-5.4-mini?service_tier=flex", + "cluster_id": "cluster-40755-2", + "nodes": [ + "pull_request:40755", + "pull_request:41224" + ], + "soft_pairs": [ + "pull_request:40755|pull_request:41224" + ], + "prepared_review_unit_hash": "48e1dead2cc93d13a679f6ecc395b132f04b5ab40205a9e86568d822b0f9aeb4", + "component_budget": { + "node_count": 2, + "item_count": 2, + "soft_pair_count": 1, + "serialized_chars": 4675, + "estimated_input_tokens": 1169, + "estimated_eval_tokens": 2594 + }, + "budget": { + "node_count": 2, + "item_count": 2, + "soft_pair_count": 1, + "serialized_chars": 4675, + "estimated_input_tokens": 1169, + "estimated_eval_tokens": 2594 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": false, + "analyst_result": { + "summary": "The two PRs are both feature additions in different model families and do not look like the same underlying change. One adds covariate forecasting support to TimesFM; the other adds DINOv3 image classification support.", + "confidence": 0.97, + "canonical_issue_reason": null, + "canonical_pr_reason": "No single canonical PR: these are unrelated feature PRs in different code paths and model families.", + "best_issue_reason": null, + "best_pr_reason": "Neither PR is a duplicate of the other; each addresses a distinct model capability and should be triaged independently.", + "soft_edge_verdicts": [ + { + "left": "pull_request:40755", + "right": "pull_request:41224", + "accept": false, + "reason": "Different model features and code paths: TimesFM covariate forecasting vs. DINOv3 image classification. They are not the same underlying bug/change and could not plausibly be merged into one PR." + } + ] + }, + "evaluator_result": null, + "evaluator_used": false, + "retried": false, + "accepted_nontrivial_soft_edge": false, + "error_kind": null, + "error_message": null + }, + { + "label": "pull_request", + "component_index": 3, + "component_count": 31, + "review_unit_index": 1, + "review_unit_count": 1, + "status": "reviewed", + "reason": null, + "source": "cache", + "cache_hit": true, + "model": "gpt-5.4-mini?service_tier=flex", + "cluster_id": "cluster-41356-2", + "nodes": [ + "pull_request:41356", + "pull_request:44339" + ], + "soft_pairs": [ + "pull_request:41356|pull_request:44339" + ], + "prepared_review_unit_hash": "aead1308c167b0bae6ff11bf9358340baa5e76f816618383418b505a10a9cd01", + "component_budget": { + "node_count": 2, + "item_count": 2, + "soft_pair_count": 1, + "serialized_chars": 6050, + "estimated_input_tokens": 1513, + "estimated_eval_tokens": 3282 + }, + "budget": { + "node_count": 2, + "item_count": 2, + "soft_pair_count": 1, + "serialized_chars": 6050, + "estimated_input_tokens": 1513, + "estimated_eval_tokens": 3282 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": false, + "analyst_result": { + "summary": "Two DEIMv2 pull requests target the same model family, but they are not the same change: PR 41356 is an early/basic model + image processor add, while PR 44339 is a much broader follow-up that adds loss code, auto-mappings, conversion tooling, modular files, and more integration work.", + "confidence": 0.91, + "canonical_issue_reason": null, + "canonical_pr_reason": "PR 44339 is the more complete and representative DEIMv2 implementation; PR 41356 looks like an earlier partial version of the same model bring-up.", + "best_issue_reason": null, + "best_pr_reason": "PR 44339 is the strongest single artifact for the cluster because it subsumes the basic model work and includes the broader integration needed for a full Transformers addition.", + "soft_edge_verdicts": [ + { + "left": "pull_request:41356", + "right": "pull_request:44339", + "accept": false, + "reason": "Same model area, but not the same concrete change: 44339 expands the implementation substantially and adds new code paths beyond the basic PR 41356. They look like related iterations, not mergeable duplicates." + } + ] + }, + "evaluator_result": null, + "evaluator_used": false, + "retried": false, + "accepted_nontrivial_soft_edge": false, + "error_kind": null, + "error_message": null + }, + { + "label": "pull_request", + "component_index": 4, + "component_count": 31, + "review_unit_index": 1, + "review_unit_count": 1, + "status": "reviewed", + "reason": null, + "source": "llm", + "cache_hit": false, + "model": "gpt-5.4-mini?service_tier=flex", + "cluster_id": "cluster-42975-2", + "nodes": [ + "pull_request:42975", + "pull_request:42976" + ], + "soft_pairs": [ + "pull_request:42975|pull_request:42976" + ], + "prepared_review_unit_hash": "b95ae04487e9562c25141b65361f567dc16d74439fd8e2a5bb9b72c129c8c2c6", + "component_budget": { + "node_count": 2, + "item_count": 2, + "soft_pair_count": 1, + "serialized_chars": 5934, + "estimated_input_tokens": 1484, + "estimated_eval_tokens": 3224 + }, + "budget": { + "node_count": 2, + "item_count": 2, + "soft_pair_count": 1, + "serialized_chars": 5934, + "estimated_input_tokens": 1484, + "estimated_eval_tokens": 3224 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": false, + "analyst_result": { + "summary": "Both items are maintenance PRs updating GitHub Actions versions in CI workflows, but they are not the same underlying change: one is a broad Node 24 compatibility sweep across many workflow actions, while the other is a separate upgrade focused on latest action versions such as docker/build-push-action v6.", + "confidence": 0.84, + "canonical_issue_reason": null, + "canonical_pr_reason": "PR 42975 is the better canonical PR because it is the broader, more explicit workflow-actions compatibility upgrade and touches more of the shared GitHub Actions surface.", + "best_issue_reason": null, + "best_pr_reason": "PR 42975 is the strongest representative of the cluster since it has the clearest stated goal and wider scope; PR 42976 is a narrower, separate dependency bump.", + "soft_edge_verdicts": [ + { + "left": "pull_request:42975", + "right": "pull_request:42976", + "accept": false, + "reason": "Do not merge: they overlap on workflow maintenance but update different action families and have different objectives/scope, so they are not the same concrete code-path change." + } + ] + }, + "evaluator_result": null, + "evaluator_used": false, + "retried": false, + "accepted_nontrivial_soft_edge": false, + "error_kind": null, + "error_message": null + }, + { + "label": "pull_request", + "component_index": 5, + "component_count": 31, + "review_unit_index": 1, + "review_unit_count": 1, + "status": "reviewed", + "reason": null, + "source": "llm", + "cache_hit": false, + "model": "gpt-5.4-mini?service_tier=flex", + "cluster_id": "cluster-42978-4", + "nodes": [ + "pull_request:42978", + "pull_request:43448", + "pull_request:45490", + "pull_request:46562" + ], + "soft_pairs": [ + "pull_request:42978|pull_request:43448", + "pull_request:42978|pull_request:45490", + "pull_request:43448|pull_request:45490", + "pull_request:45490|pull_request:46562" + ], + "prepared_review_unit_hash": "f6d69f404ce7f2ea7b4d5612a560132fabba559e136fb83de9e6ef0afc6fd393", + "component_budget": { + "node_count": 4, + "item_count": 4, + "soft_pair_count": 4, + "serialized_chars": 13046, + "estimated_input_tokens": 3262, + "estimated_eval_tokens": 6780 + }, + "budget": { + "node_count": 4, + "item_count": 4, + "soft_pair_count": 4, + "serialized_chars": 13046, + "estimated_input_tokens": 3262, + "estimated_eval_tokens": 6780 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": false, + "analyst_result": { + "summary": "These pull requests all add different model integrations and documentation, with only boilerplate overlap in shared auto-mapping/docs files. They are related by subsystem, not by the same underlying change, so none should be merged as duplicates.", + "confidence": 0.98, + "canonical_issue_reason": null, + "canonical_pr_reason": "No single canonical PR: ViT NEPA, Molmo, CTSM, and Step3p7 are distinct model additions with different implementations, tests, and model-specific files.", + "best_issue_reason": null, + "best_pr_reason": "No global best duplicate candidate: each PR targets a separate model integration rather than the same bug or feature, so there is no meaningful representative duplicate PR.", + "soft_edge_verdicts": [ + { + "left": "pull_request:42978", + "right": "pull_request:43448", + "accept": false, + "reason": "Different model additions (ViT NEPA vs Molmo). Shared docs and auto-registry edits are boilerplate for adding a model, not the same code-path fix." + }, + { + "left": "pull_request:42978", + "right": "pull_request:45490", + "accept": false, + "reason": "ViT NEPA and CTSM are unrelated model implementations. Overlap is limited to generic docs/auto-mapping files, so they are not duplicates." + }, + { + "left": "pull_request:43448", + "right": "pull_request:45490", + "accept": false, + "reason": "Molmo and CTSM add different architectures and model-specific files. Same subsystem, but not the same concrete change." + }, + { + "left": "pull_request:45490", + "right": "pull_request:46562", + "accept": false, + "reason": "CTSM and Step3p7 are separate model launches with different modeling/processing code. Shared auto-mapping updates do not make them the same PR." + } + ] + }, + "evaluator_result": null, + "evaluator_used": false, + "retried": false, + "accepted_nontrivial_soft_edge": false, + "error_kind": null, + "error_message": null + }, + { + "label": "pull_request", + "component_index": 6, + "component_count": 31, + "review_unit_index": 1, + "review_unit_count": 1, + "status": "reviewed", + "reason": null, + "source": "llm", + "cache_hit": false, + "model": "gpt-5.4-mini?service_tier=flex", + "cluster_id": "cluster-43291-8", + "nodes": [ + "pull_request:43291", + "pull_request:43665", + "pull_request:44815", + "pull_request:46380", + "pull_request:46539", + "pull_request:46886", + "pull_request:46922", + "pull_request:47393" + ], + "soft_pairs": [ + "pull_request:43291|pull_request:43665", + "pull_request:43291|pull_request:44815", + "pull_request:43291|pull_request:46380", + "pull_request:43291|pull_request:46922", + "pull_request:43665|pull_request:46539", + "pull_request:43665|pull_request:47393", + "pull_request:43665|pull_request:44815", + "pull_request:43665|pull_request:46380", + "pull_request:43665|pull_request:46922", + "pull_request:44815|pull_request:46539", + "pull_request:44815|pull_request:46380", + "pull_request:44815|pull_request:46922", + "pull_request:46380|pull_request:46539", + "pull_request:46380|pull_request:46922", + "pull_request:46539|pull_request:46922", + "pull_request:46886|pull_request:46922" + ], + "prepared_review_unit_hash": "a880f5f84182b7dde2638f57d301fa9201a41ce117e4f013fa2d76c1d57c801a", + "component_budget": { + "node_count": 8, + "item_count": 8, + "soft_pair_count": 16, + "serialized_chars": 21115, + "estimated_input_tokens": 5279, + "estimated_eval_tokens": 10814 + }, + "budget": { + "node_count": 8, + "item_count": 8, + "soft_pair_count": 16, + "serialized_chars": 21115, + "estimated_input_tokens": 5279, + "estimated_eval_tokens": 10814 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": false, + "analyst_result": { + "summary": "This cluster is mostly unrelated PRs with weak title-based similarity; the only clear duplicate-like pair is the two YOLOS fixes, where one PR is a more complete version of the same device-mismatch change.", + "confidence": 0.95, + "canonical_issue_reason": null, + "canonical_pr_reason": "pull_request:46922 is the best canonical PR because it directly targets the YOLOS device-mismatch bug, has an explicit linked issue, and subsumes the narrower no-split-module change in pull_request:46886.", + "best_issue_reason": null, + "best_pr_reason": "pull_request:46922 is the strongest representative of the cluster\u2019s only real duplicate subtopic: it fixes the same YOLOS code path with a concrete forward-pass device placement change plus the supporting split-module adjustment.", + "soft_edge_verdicts": [ + { + "left": "pull_request:43291", + "right": "pull_request:43665", + "accept": false, + "reason": "Both are test-focused, but they touch different models and different test files; no shared bug or mergeable single change." + }, + { + "left": "pull_request:43291", + "right": "pull_request:44815", + "accept": false, + "reason": "Whisper tokenizer tests vs FP8 dequant/loading code are unrelated changes despite superficial \"fix\" wording." + }, + { + "left": "pull_request:43291", + "right": "pull_request:46380", + "accept": false, + "reason": "Whisper test cleanup and generation/CIS validation fixes are different problems in different subsystems." + }, + { + "left": "pull_request:43291", + "right": "pull_request:46922", + "accept": false, + "reason": "Whisper tokenizer/test changes are unrelated to YOLOS device placement and split-module handling." + }, + { + "left": "pull_request:43665", + "right": "pull_request:46539", + "accept": false, + "reason": "Both modify tests/infra, but one is model test assertions and the other is CI memory/device-map handling; not the same underlying change." + }, + { + "left": "pull_request:43665", + "right": "pull_request:47393", + "accept": false, + "reason": "Model tests and a README URL edit are clearly unrelated." + }, + { + "left": "pull_request:43665", + "right": "pull_request:44815", + "accept": false, + "reason": "CLIP/ViT test edits are not the same as the FP8 dequant/loading fix." + }, + { + "left": "pull_request:43665", + "right": "pull_request:46380", + "accept": false, + "reason": "Different models and different failure modes; no concrete shared code path." + }, + { + "left": "pull_request:43665", + "right": "pull_request:46922", + "accept": false, + "reason": "Model test files for CLIP/ViT do not match the YOLOS forward-pass/device-mismatch fix." + }, + { + "left": "pull_request:44815", + "right": "pull_request:46539", + "accept": false, + "reason": "FP8 dequant/loading logic and CI/device-map integration are unrelated subsystems." + }, + { + "left": "pull_request:44815", + "right": "pull_request:46380", + "accept": false, + "reason": "These address different generation/loading issues and should not be merged as one PR." + }, + { + "left": "pull_request:44815", + "right": "pull_request:46922", + "accept": false, + "reason": "No shared code path beyond both touching model internals; the bugs are different." + }, + { + "left": "pull_request:46380", + "right": "pull_request:46539", + "accept": false, + "reason": "Generation watermarking validation and Phimoe CI memory setup are unrelated." + }, + { + "left": "pull_request:46380", + "right": "pull_request:46922", + "accept": false, + "reason": "Generation config validation and YOLOS device handling are different features and failures." + }, + { + "left": "pull_request:46539", + "right": "pull_request:46922", + "accept": false, + "reason": "CI runner memory limits and YOLOS device mismatch are separate problems." + }, + { + "left": "pull_request:46886", + "right": "pull_request:46922", + "accept": true, + "reason": "Both are YOLOS fixes for the same device-mismatch issue in the same file; 46922 extends the same fix with the forward-pass device move and the split-module adjustment, so they could plausibly be one PR." + } + ] + }, + "evaluator_result": { + "accept": true, + "feedback": "The analyst summary is grounded in the packet and stays conservative. The non-YOLOS pairs are reasonably judged unrelated, and the 46886/46922 pair is plausibly mergeable as one YOLOS device-mismatch bugfix, with 46922 looking like the more complete version." + }, + "evaluator_used": true, + "retried": false, + "accepted_nontrivial_soft_edge": true, + "error_kind": null, + "error_message": null + }, + { + "label": "pull_request", + "component_index": 7, + "component_count": 31, + "review_unit_index": 1, + "review_unit_count": 1, + "status": "reviewed", + "reason": null, + "source": "llm", + "cache_hit": false, + "model": "gpt-5.4-mini?service_tier=flex", + "cluster_id": "cluster-43488-3", + "nodes": [ + "pull_request:43488", + "pull_request:43532", + "pull_request:46747" + ], + "soft_pairs": [ + "pull_request:43488|pull_request:46747", + "pull_request:43488|pull_request:43532", + "pull_request:43532|pull_request:46747" + ], + "prepared_review_unit_hash": "582e7ad64dad1fbdde342cd39cea6374d6eceab571eae652513411c98ede047c", + "component_budget": { + "node_count": 3, + "item_count": 3, + "soft_pair_count": 3, + "serialized_chars": 6349, + "estimated_input_tokens": 1588, + "estimated_eval_tokens": 3432 + }, + "budget": { + "node_count": 3, + "item_count": 3, + "soft_pair_count": 3, + "serialized_chars": 6349, + "estimated_input_tokens": 1588, + "estimated_eval_tokens": 3432 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": false, + "analyst_result": { + "summary": "These PRs are not duplicates: one is a repo-bot/version-format test change, one is a workflow/diff-display change, and one is a security-gate workflow tweak. The similarity is mostly superficial (shared testing/workflow context), not the same underlying fix or feature.", + "confidence": 0.94, + "canonical_issue_reason": null, + "canonical_pr_reason": "No single canonical PR: the three changes target different code paths and CI concerns, so they should not be collapsed into one duplicate cluster.", + "best_issue_reason": null, + "best_pr_reason": "If forced to pick the most self-contained reference, PR #43488 is the cleanest standalone change, but it is not a canonical representative for the others because they address unrelated workflow behavior.", + "soft_edge_verdicts": [ + { + "left": "pull_request:43488", + "right": "pull_request:46747", + "accept": false, + "reason": "Different changes: #43488 edits packaging/test files for a repo-bot/version-format check, while #46747 changes the security-gate workflow trigger from pull_request_target to pull_request. Same broad CI area, but not the same bug or mergeable change set." + }, + { + "left": "pull_request:43488", + "right": "pull_request:43532", + "accept": false, + "reason": "#43488 is about dependency/version-format and a Vit test, whereas #43532 restructures multiple workflows and testing utilities to show diffs. Shared test file is incidental; these are not the same underlying change." + }, + { + "left": "pull_request:43532", + "right": "pull_request:46747", + "accept": false, + "reason": "#43532 updates comment/diff-display CI behavior across several files, but #46747 only changes the security gate workflow trigger. They touch the same repo area, not the same concrete problem." + } + ] + }, + "evaluator_result": null, + "evaluator_used": false, + "retried": false, + "accepted_nontrivial_soft_edge": false, + "error_kind": null, + "error_message": null + }, + { + "label": "pull_request", + "component_index": 8, + "component_count": 31, + "review_unit_index": 1, + "review_unit_count": 1, + "status": "reviewed", + "reason": null, + "source": "llm", + "cache_hit": false, + "model": "gpt-5.4-mini?service_tier=flex", + "cluster_id": "cluster-43612-2", + "nodes": [ + "pull_request:43612", + "pull_request:43613" + ], + "soft_pairs": [ + "pull_request:43612|pull_request:43613" + ], + "prepared_review_unit_hash": "de7832969d52e56240bb780ce903833a1e3964d95d53dad6b9c810843aa893ca", + "component_budget": { + "node_count": 2, + "item_count": 2, + "soft_pair_count": 1, + "serialized_chars": 6082, + "estimated_input_tokens": 1521, + "estimated_eval_tokens": 3298 + }, + "budget": { + "node_count": 2, + "item_count": 2, + "soft_pair_count": 1, + "serialized_chars": 6082, + "estimated_input_tokens": 1521, + "estimated_eval_tokens": 3298 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": false, + "analyst_result": { + "summary": "Both PRs add new, similarly structured segmentation pipelines, but they target different concrete features: promptable concept segmentation vs promptable visual segmentation. The overlap is mostly scaffolding/docs, not the same underlying bug or change.", + "confidence": 0.95, + "canonical_issue_reason": null, + "canonical_pr_reason": "PR 43613 is the better representative of the cluster because it is the broader, more complete pipeline addition with wider model support and tests.", + "best_issue_reason": null, + "best_pr_reason": "PR 43613 is the best single PR to reference: it has the larger scope and more integrated support, so it is the most representative artifact in this cluster.", + "soft_edge_verdicts": [ + { + "left": "pull_request:43612", + "right": "pull_request:43613", + "accept": false, + "reason": "Reject: these are separate pipeline features for different tasks and code paths. They share documentation and registration scaffolding, but they do not appear to fix or implement the same concrete change." + } + ] + }, + "evaluator_result": null, + "evaluator_used": false, + "retried": false, + "accepted_nontrivial_soft_edge": false, + "error_kind": null, + "error_message": null + }, + { + "label": "pull_request", + "component_index": 9, + "component_count": 31, + "review_unit_index": 1, + "review_unit_count": 1, + "status": "reviewed", + "reason": null, + "source": "llm", + "cache_hit": false, + "model": "gpt-5.4-mini?service_tier=flex", + "cluster_id": "cluster-43651-2", + "nodes": [ + "pull_request:43651", + "pull_request:43663" + ], + "soft_pairs": [ + "pull_request:43651|pull_request:43663" + ], + "prepared_review_unit_hash": "02ace308d0e604b2930a8b915a88709693a38b84a309412845254fb8908ff0f4", + "component_budget": { + "node_count": 2, + "item_count": 2, + "soft_pair_count": 1, + "serialized_chars": 4293, + "estimated_input_tokens": 1074, + "estimated_eval_tokens": 2404 + }, + "budget": { + "node_count": 2, + "item_count": 2, + "soft_pair_count": 1, + "serialized_chars": 4293, + "estimated_input_tokens": 1074, + "estimated_eval_tokens": 2404 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": false, + "analyst_result": { + "summary": "Both PRs are trainer extensibility tweaks in the same file, but they address different code paths: one changes gradient-accumulation loss scaling, the other refactors signature-column filtering. They are not the same underlying change.", + "confidence": 0.91, + "canonical_issue_reason": null, + "canonical_pr_reason": "PR #43651 is the more concrete code-path change: it alters training_step loss scaling behavior and adds a new hook for custom trainers.", + "best_issue_reason": null, + "best_pr_reason": "PR #43651 is the best representative of the cluster because it is the more directly behavior-changing trainer fix; PR #43663 is a separate refactor around input-column filtering.", + "soft_edge_verdicts": [ + { + "left": "pull_request:43651", + "right": "pull_request:43663", + "accept": false, + "reason": "Shared trainer.py location and extensibility theme, but they fix different mechanisms: gradient-accumulation loss scaling vs signature-based column filtering. Not mergeable as one PR." + } + ] + }, + "evaluator_result": null, + "evaluator_used": false, + "retried": false, + "accepted_nontrivial_soft_edge": false, + "error_kind": null, + "error_message": null + }, + { + "label": "pull_request", + "component_index": 10, + "component_count": 31, + "review_unit_index": 1, + "review_unit_count": 1, + "status": "reviewed", + "reason": null, + "source": "llm", + "cache_hit": false, + "model": "gpt-5.4-mini?service_tier=flex", + "cluster_id": "cluster-43656-2", + "nodes": [ + "pull_request:43656", + "pull_request:43842" + ], + "soft_pairs": [ + "pull_request:43656|pull_request:43842" + ], + "prepared_review_unit_hash": "f681f7a59774fd774a20bcf60ccb2160fd48aa679445fbe805fc24ca96978ea3", + "component_budget": { + "node_count": 2, + "item_count": 2, + "soft_pair_count": 1, + "serialized_chars": 3663, + "estimated_input_tokens": 916, + "estimated_eval_tokens": 2088 + }, + "budget": { + "node_count": 2, + "item_count": 2, + "soft_pair_count": 1, + "serialized_chars": 3663, + "estimated_input_tokens": 916, + "estimated_eval_tokens": 2088 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": false, + "analyst_result": { + "summary": "Both PRs patch `src/transformers/cli/serve.py` to eliminate a `TypeAdapter`-related runtime NameError in the CLI server path; they differ in implementation style, but they target the same underlying failure mode.", + "confidence": 0.65, + "canonical_issue_reason": null, + "canonical_pr_reason": "PR 43842 is the more direct and narrowly scoped fix: it keeps `TypeAdapter` out of runtime imports with `TYPE_CHECKING` and quotes the annotation, matching the reported NameError more precisely.", + "best_issue_reason": null, + "best_pr_reason": "PR 43842 is the best PR candidate here because it directly addresses the `TypeAdapter` NameError without broader annotation churn.", + "soft_edge_verdicts": [ + { + "left": "pull_request:43656", + "right": "pull_request:43842", + "accept": true, + "reason": "Both PRs fix the same `TypeAdapter` NameError in the CLI server code path in `serve.py`; the differences are implementation detail, so they could plausibly be merged into one change." + } + ] + }, + "evaluator_result": { + "accept": true, + "feedback": "The summary is grounded in the packet: both PRs touch `src/transformers/cli/serve.py` and are plausibly aimed at the same `TypeAdapter`-related NameError. The mergeability verdict is suitably conservative given the shared file and similar bugfix scope." + }, + "evaluator_used": true, + "retried": false, + "accepted_nontrivial_soft_edge": true, + "error_kind": null, + "error_message": null + }, + { + "label": "pull_request", + "component_index": 11, + "component_count": 31, + "review_unit_index": 1, + "review_unit_count": 1, + "status": "reviewed", + "reason": null, + "source": "cache", + "cache_hit": true, + "model": "gpt-5.4-mini?service_tier=flex", + "cluster_id": "cluster-43779-2", + "nodes": [ + "pull_request:43779", + "pull_request:43816" + ], + "soft_pairs": [ + "pull_request:43779|pull_request:43816" + ], + "prepared_review_unit_hash": "e7c674adcc031aa8883109a5dfc678b4407fb20b10764302092e384d031ba83a", + "component_budget": { + "node_count": 2, + "item_count": 2, + "soft_pair_count": 1, + "serialized_chars": 4362, + "estimated_input_tokens": 1091, + "estimated_eval_tokens": 2438 + }, + "budget": { + "node_count": 2, + "item_count": 2, + "soft_pair_count": 1, + "serialized_chars": 4362, + "estimated_input_tokens": 1091, + "estimated_eval_tokens": 2438 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": false, + "analyst_result": { + "summary": "Both pull requests appear to implement the same SwanLab integration change: adding run id/resume support in `integration_utils.py` for issue #43698, with one variant also expanding the docs. They are strong duplicate PR candidates.", + "confidence": 0.94, + "canonical_issue_reason": null, + "canonical_pr_reason": "PR #43816 is the cleaner representative: it carries the same SwanLab id/resume support plus the accompanying documentation updates, making it the more complete version of the shared change.", + "best_issue_reason": null, + "best_pr_reason": "PR #43816 is the best global representative because it appears to include the full feature change and docs for SwanLab run id/resume support, while #43779 is the narrower implementation-focused variant.", + "soft_edge_verdicts": [ + { + "left": "pull_request:43779", + "right": "pull_request:43816", + "accept": true, + "reason": "Same underlying SwanLab callback enhancement for id/resume support in the same file and same issue target; the doc differences do not change the core code-path change, so these are duplicate PRs." + } + ] + }, + "evaluator_result": { + "accept": true, + "feedback": "The summary is grounded in the packet: both PRs target issue #43698, touch the same file, and the previews indicate the same SwanLab id/resume support with #43816 adding docs. The duplicate/mergeable verdict is conservative enough given the shared target and overlapping code change." + }, + "evaluator_used": true, + "retried": false, + "accepted_nontrivial_soft_edge": true, + "error_kind": null, + "error_message": null + }, + { + "label": "pull_request", + "component_index": 12, + "component_count": 31, + "review_unit_index": 1, + "review_unit_count": 2, + "status": "reviewed", + "reason": null, + "source": "llm", + "cache_hit": false, + "model": "gpt-5.4-mini?service_tier=flex", + "cluster_id": "cluster-43996-28", + "nodes": [ + "pull_request:43996", + "pull_request:43997", + "pull_request:44007", + "pull_request:44010", + "pull_request:44013", + "pull_request:44017", + "pull_request:44018", + "pull_request:44019", + "pull_request:44024", + "pull_request:44044", + "pull_request:44056", + "pull_request:44059", + "pull_request:44066", + "pull_request:44068", + "pull_request:44071", + "pull_request:44072", + "pull_request:44073", + "pull_request:44074", + "pull_request:44076", + "pull_request:44085", + "pull_request:44086", + "pull_request:44098", + "pull_request:44101", + "pull_request:44116", + "pull_request:44129", + "pull_request:44154", + "pull_request:44161", + "pull_request:44722" + ], + "soft_pairs": [ + "pull_request:43996|pull_request:44085", + "pull_request:43996|pull_request:44044", + "pull_request:44066|pull_request:44085", + "pull_request:44007|pull_request:44072", + "pull_request:44072|pull_request:44722", + "pull_request:44066|pull_request:44072", + "pull_request:44013|pull_request:44044", + "pull_request:44066|pull_request:44086", + "pull_request:44066|pull_request:44071", + "pull_request:44018|pull_request:44068", + "pull_request:44066|pull_request:44068", + "pull_request:44018|pull_request:44066", + "pull_request:44019|pull_request:44722", + "pull_request:44086|pull_request:44722", + "pull_request:44019|pull_request:44085", + "pull_request:44018|pull_request:44019", + "pull_request:44019|pull_request:44071", + "pull_request:44068|pull_request:44116", + "pull_request:44068|pull_request:44722", + "pull_request:44018|pull_request:44116", + "pull_request:44071|pull_request:44722", + "pull_request:44018|pull_request:44085", + "pull_request:44018|pull_request:44086", + "pull_request:44068|pull_request:44086", + "pull_request:44071|pull_request:44085", + "pull_request:44085|pull_request:44086", + "pull_request:44059|pull_request:44086", + "pull_request:44098|pull_request:44101", + "pull_request:44059|pull_request:44722", + "pull_request:44018|pull_request:44071", + "pull_request:44017|pull_request:44161", + "pull_request:44017|pull_request:44073", + "pull_request:44017|pull_request:44722", + "pull_request:44068|pull_request:44071", + "pull_request:43997|pull_request:44073", + "pull_request:44017|pull_request:44071", + "pull_request:44071|pull_request:44086", + "pull_request:44073|pull_request:44101", + "pull_request:44073|pull_request:44161", + "pull_request:44056|pull_request:44101", + "pull_request:44059|pull_request:44071", + "pull_request:44018|pull_request:44059", + "pull_request:44059|pull_request:44066", + "pull_request:43997|pull_request:44017", + "pull_request:44101|pull_request:44161", + "pull_request:43997|pull_request:44056", + "pull_request:44071|pull_request:44076", + "pull_request:44074|pull_request:44101", + "pull_request:44056|pull_request:44161", + "pull_request:43997|pull_request:44018", + "pull_request:43997|pull_request:44161", + "pull_request:44074|pull_request:44161", + "pull_request:44018|pull_request:44161", + "pull_request:44010|pull_request:44071", + "pull_request:44018|pull_request:44129", + "pull_request:44010|pull_request:44073", + "pull_request:43997|pull_request:44074", + "pull_request:44010|pull_request:44018", + "pull_request:43997|pull_request:44129", + "pull_request:44071|pull_request:44129", + "pull_request:43996|pull_request:44161", + "pull_request:43996|pull_request:44017", + "pull_request:44010|pull_request:44044", + "pull_request:44044|pull_request:44074", + "pull_request:44066|pull_request:44722", + "pull_request:44085|pull_request:44722", + "pull_request:44085|pull_request:44129", + "pull_request:43996|pull_request:44007", + "pull_request:43996|pull_request:44129", + "pull_request:44044|pull_request:44129", + "pull_request:44044|pull_request:44154", + "pull_request:44024|pull_request:44076" + ], + "prepared_review_unit_hash": "c3c5b83c8d3742ad5fdc4ddd5822f946305511daf3f7ce554bb731f8edb7a614", + "component_budget": { + "node_count": 28, + "item_count": 28, + "soft_pair_count": 77, + "serialized_chars": 84502, + "estimated_input_tokens": 21126, + "estimated_eval_tokens": 42508 + }, + "budget": { + "node_count": 28, + "item_count": 28, + "soft_pair_count": 72, + "serialized_chars": 76933, + "estimated_input_tokens": 19234, + "estimated_eval_tokens": 38724 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": true, + "analyst_result": { + "summary": "This cluster is a broad wave of output-tracing refactors across many unrelated model implementations. Most soft edges should be rejected as broad pattern matches rather than duplicate fixes. The only tight duplicate-like subgroup in the packet is the GPT-J/CodeGen family around 44066/44085/44722; the GPT-Neo pair 44018/44068 is too weak to call the same concrete change under a conservative standard.", + "confidence": 0.79, + "canonical_issue_reason": null, + "canonical_pr_reason": "PR 44722 is the best canonical representative of the only tight duplicate subgroup: it is issue-linked, touches the same GPT-J/CodeGen code paths as the nearby GPT-J variants, and looks like the most consolidated version of that refactor.", + "best_issue_reason": null, + "best_pr_reason": "PR 44722 is the strongest anchor for the cluster because it is explicit about the standardized-decorator refactor and most clearly subsumes the closely related GPT-J/CodeGen variants.", + "soft_edge_verdicts": [ + { + "left": "pull_request:43996", + "right": "pull_request:44085", + "accept": false, + "reason": "Different model families and code paths; same output-tracing theme is too broad." + }, + { + "left": "pull_request:43996", + "right": "pull_request:44044", + "accept": false, + "reason": "Different model families; not the same concrete fix." + }, + { + "left": "pull_request:44066", + "right": "pull_request:44085", + "accept": false, + "reason": "Same general GPT-J area, but the packet only shows weak similarity here and explicitly marks this pair non-deterministic; not enough to call it one duplicate PR." + }, + { + "left": "pull_request:44007", + "right": "pull_request:44072", + "accept": false, + "reason": "RegNet/ResNet vs EfficientNet are separate model implementations; shared refactor pattern is not enough." + }, + { + "left": "pull_request:44072", + "right": "pull_request:44722", + "accept": false, + "reason": "EfficientNet and GPT-J are unrelated code paths." + }, + { + "left": "pull_request:44066", + "right": "pull_request:44072", + "accept": false, + "reason": "GPT-J/CodeGen vs EfficientNet are unrelated model changes." + }, + { + "left": "pull_request:44013", + "right": "pull_request:44044", + "accept": false, + "reason": "MobileNetV2 vs DeBERTa are different subsystems and different fixes." + }, + { + "left": "pull_request:44066", + "right": "pull_request:44086", + "accept": false, + "reason": "GPT-J/CodeGen and MGP-STR are different model-specific changes." + }, + { + "left": "pull_request:44066", + "right": "pull_request:44071", + "accept": false, + "reason": "GPT-J/CodeGen vs MPT are separate model code paths." + }, + { + "left": "pull_request:44018", + "right": "pull_request:44068", + "accept": false, + "reason": "Same broad GPT-Neo theme, but the packet does not show enough evidence that these are the same concrete refactor." + }, + { + "left": "pull_request:44066", + "right": "pull_request:44068", + "accept": false, + "reason": "GPT-J/CodeGen vs GPT-Neo are distinct model refactors." + }, + { + "left": "pull_request:44018", + "right": "pull_request:44019", + "accept": false, + "reason": "GPT-Neo vs ResNet are unrelated model code paths." + }, + { + "left": "pull_request:44019", + "right": "pull_request:44722", + "accept": false, + "reason": "ResNet vs GPT-J are unrelated model changes." + }, + { + "left": "pull_request:44086", + "right": "pull_request:44722", + "accept": false, + "reason": "MGP-STR and GPT-J are different model families and fixes." + }, + { + "left": "pull_request:44019", + "right": "pull_request:44085", + "accept": false, + "reason": "ResNet vs GPT-J are unrelated." + }, + { + "left": "pull_request:44068", + "right": "pull_request:44116", + "accept": false, + "reason": "GPT-Neo vs Flaubert are different model-specific refactors." + }, + { + "left": "pull_request:44068", + "right": "pull_request:44722", + "accept": false, + "reason": "GPT-Neo and GPT-J/CodeGen are different concrete changes." + }, + { + "left": "pull_request:44018", + "right": "pull_request:44116", + "accept": false, + "reason": "GPT-Neo and Flaubert are unrelated model changes." + }, + { + "left": "pull_request:44071", + "right": "pull_request:44722", + "accept": false, + "reason": "MPT vs GPT-J are unrelated model code paths." + }, + { + "left": "pull_request:44018", + "right": "pull_request:44085", + "accept": false, + "reason": "GPT-Neo vs GPT-J are related only at the broad pattern level, not the same PR." + }, + { + "left": "pull_request:44018", + "right": "pull_request:44086", + "accept": false, + "reason": "GPT-Neo vs MGP-STR are unrelated changes." + }, + { + "left": "pull_request:44068", + "right": "pull_request:44086", + "accept": false, + "reason": "GPT-Neo and MGP-STR are different models." + }, + { + "left": "pull_request:44071", + "right": "pull_request:44086", + "accept": false, + "reason": "MPT vs MGP-STR are separate model implementations." + }, + { + "left": "pull_request:44059", + "right": "pull_request:44086", + "accept": false, + "reason": "GPT-2 vs MGP-STR are unrelated refactors." + }, + { + "left": "pull_request:44098", + "right": "pull_request:44101", + "accept": false, + "reason": "ViLT and Flaubert are different model families." + }, + { + "left": "pull_request:44059", + "right": "pull_request:44722", + "accept": false, + "reason": "GPT-2 and GPT-J are different code paths; same theme only." + }, + { + "left": "pull_request:44017", + "right": "pull_request:44161", + "accept": false, + "reason": "SegFormer vs LongT5 are unrelated model changes." + }, + { + "left": "pull_request:44017", + "right": "pull_request:44073", + "accept": false, + "reason": "SegFormer vs VisualBert are different model-specific refactors." + }, + { + "left": "pull_request:44017", + "right": "pull_request:44722", + "accept": false, + "reason": "SegFormer and GPT-J do not share the same concrete change." + }, + { + "left": "pull_request:43997", + "right": "pull_request:44073", + "accept": false, + "reason": "RegNet vs VisualBert are unrelated model implementations." + }, + { + "left": "pull_request:44056", + "right": "pull_request:44101", + "accept": false, + "reason": "MPNet vs Flaubert are unrelated model-specific refactors." + }, + { + "left": "pull_request:44056", + "right": "pull_request:44161", + "accept": false, + "reason": "MPNet vs LongT5 are separate model-specific changes." + }, + { + "left": "pull_request:43997", + "right": "pull_request:44056", + "accept": false, + "reason": "RegNet and MPNet are different model families." + }, + { + "left": "pull_request:44071", + "right": "pull_request:44076", + "accept": false, + "reason": "MPT and ImageGPT are unrelated model changes." + }, + { + "left": "pull_request:44074", + "right": "pull_request:44101", + "accept": false, + "reason": "TextNet vs Flaubert are different model implementations." + }, + { + "left": "pull_request:44056", + "right": "pull_request:44073", + "accept": false, + "reason": "MPNet and VisualBert are unrelated model-specific refactors." + }, + { + "left": "pull_request:43996", + "right": "pull_request:44161", + "accept": false, + "reason": "FNet/CVT vs LongT5 are unrelated model families." + }, + { + "left": "pull_request:43996", + "right": "pull_request:44017", + "accept": false, + "reason": "FNet/CVT vs SegFormer are different model implementations." + }, + { + "left": "pull_request:44010", + "right": "pull_request:44044", + "accept": false, + "reason": "SqueezeBert and DeBERTa are unrelated changes." + }, + { + "left": "pull_request:44044", + "right": "pull_request:44074", + "accept": false, + "reason": "DeBERTa vs TextNet are different model-specific output changes." + }, + { + "left": "pull_request:44024", + "right": "pull_request:44076", + "accept": false, + "reason": "FocalNet vs ImageGPT are unrelated models; same standardized-output pattern is too broad." + }, + { + "left": "pull_request:44085", + "right": "pull_request:44129", + "accept": false, + "reason": "GPT-J vs SpeechT5 are different concrete code paths." + }, + { + "left": "pull_request:44044", + "right": "pull_request:44129", + "accept": false, + "reason": "DeBERTa and SpeechT5 are unrelated." + }, + { + "left": "pull_request:44044", + "right": "pull_request:44154", + "accept": false, + "reason": "DeBERTa vs ViLT are separate model changes." + } + ] + }, + "evaluator_result": { + "accept": false, + "feedback": "Reject: the cluster summary is mostly conservative, but two soft-edge acceptances overstate the evidence. In particular, 44066\u219444085 and 44018\u219444068 are treated as true duplicates / same concrete refactors, while the packet only supports broad same-model similarity and, for 44066\u219444085, explicitly lists the pair as non-deterministic. That is too strong for this review standard. The rest of the reasoning is broadly grounded, but these acceptances are not conservative enough." + }, + "evaluator_used": true, + "retried": true, + "accepted_nontrivial_soft_edge": false, + "error_kind": null, + "error_message": null + }, + { + "label": "pull_request", + "component_index": 12, + "component_count": 31, + "review_unit_index": 2, + "review_unit_count": 2, + "status": "reviewed", + "reason": null, + "source": "llm", + "cache_hit": false, + "model": "gpt-5.4-mini?service_tier=flex", + "cluster_id": "cluster-43996-28", + "nodes": [ + "pull_request:43997", + "pull_request:44010", + "pull_request:44056", + "pull_request:44073", + "pull_request:44074", + "pull_request:44129" + ], + "soft_pairs": [ + "pull_request:44073|pull_request:44074", + "pull_request:44056|pull_request:44073", + "pull_request:44056|pull_request:44074", + "pull_request:43997|pull_request:44010", + "pull_request:44010|pull_request:44129" + ], + "prepared_review_unit_hash": "870236ba72b7387db711dfc03cd048ffb9c2d39b3596c7fb8009c3835dfc3dd7", + "component_budget": { + "node_count": 28, + "item_count": 28, + "soft_pair_count": 77, + "serialized_chars": 84502, + "estimated_input_tokens": 21126, + "estimated_eval_tokens": 42508 + }, + "budget": { + "node_count": 6, + "item_count": 6, + "soft_pair_count": 5, + "serialized_chars": 13069, + "estimated_input_tokens": 3268, + "estimated_eval_tokens": 6792 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": true, + "analyst_result": { + "summary": "These are all model-specific refactors to standardized output capture/tracing, but they touch different architectures and code paths, so they are not duplicates of one another.", + "confidence": 0.93, + "canonical_issue_reason": null, + "canonical_pr_reason": "No single PR is a true duplicate anchor for the set; 44129 is the strongest representative because it has an explicit issue target and a concrete, model-specific output-tracing refactor.", + "best_issue_reason": null, + "best_pr_reason": "44129 is the best representative PR: it targets a specific issue, changes a single model implementation, and shows a substantive output-tracing cleanup rather than just a title-level similarity.", + "soft_edge_verdicts": [ + { + "left": "pull_request:44073", + "right": "pull_request:44074", + "accept": false, + "reason": "Same refactor theme, but VisualBert and TextNet are different models with different forward paths; not the same concrete change." + }, + { + "left": "pull_request:44056", + "right": "pull_request:44073", + "accept": false, + "reason": "Both are output-tracing migrations, but MPNet and VisualBert have distinct implementations and no shared code-path bug." + }, + { + "left": "pull_request:44056", + "right": "pull_request:44074", + "accept": false, + "reason": "Different model families and different model files; this is parallel refactoring, not one mergeable fix." + }, + { + "left": "pull_request:43997", + "right": "pull_request:44010", + "accept": false, + "reason": "RegNet and SqueezeBert are unrelated code paths; same high-level decorator migration is not enough to treat them as duplicates." + }, + { + "left": "pull_request:44010", + "right": "pull_request:44129", + "accept": false, + "reason": "Both modernize output handling, but SqueezeBert and SpeechT5 fix different model-specific implementations and could not plausibly be one PR." + } + ] + }, + "evaluator_result": null, + "evaluator_used": false, + "retried": false, + "accepted_nontrivial_soft_edge": false, + "error_kind": null, + "error_message": null + }, + { + "label": "pull_request", + "component_index": 13, + "component_count": 31, + "review_unit_index": 1, + "review_unit_count": 1, + "status": "reviewed", + "reason": null, + "source": "llm", + "cache_hit": false, + "model": "gpt-5.4-mini?service_tier=flex", + "cluster_id": "cluster-43998-13", + "nodes": [ + "pull_request:43998", + "pull_request:43999", + "pull_request:44000", + "pull_request:44001", + "pull_request:44002", + "pull_request:44003", + "pull_request:44004", + "pull_request:44025", + "pull_request:44026", + "pull_request:44027", + "pull_request:44028", + "pull_request:44029", + "pull_request:44030" + ], + "soft_pairs": [ + "pull_request:43998|pull_request:44002", + "pull_request:43998|pull_request:44000", + "pull_request:43998|pull_request:44001", + "pull_request:43998|pull_request:44003", + "pull_request:43998|pull_request:44029", + "pull_request:43999|pull_request:44001", + "pull_request:43999|pull_request:44002", + "pull_request:43999|pull_request:44025", + "pull_request:43999|pull_request:44026", + "pull_request:43999|pull_request:44027", + "pull_request:44000|pull_request:44001", + "pull_request:44000|pull_request:44002", + "pull_request:44000|pull_request:44003", + "pull_request:44000|pull_request:44025", + "pull_request:44001|pull_request:44002", + "pull_request:44001|pull_request:44003", + "pull_request:44001|pull_request:44025", + "pull_request:44002|pull_request:44029", + "pull_request:44002|pull_request:44003", + "pull_request:44003|pull_request:44004", + "pull_request:43998|pull_request:44004", + "pull_request:44001|pull_request:44004", + "pull_request:44002|pull_request:44004", + "pull_request:44000|pull_request:44004", + "pull_request:44025|pull_request:44026", + "pull_request:44025|pull_request:44027", + "pull_request:44025|pull_request:44028", + "pull_request:44025|pull_request:44029", + "pull_request:44025|pull_request:44030", + "pull_request:44026|pull_request:44029", + "pull_request:44026|pull_request:44027", + "pull_request:44026|pull_request:44028", + "pull_request:44026|pull_request:44030", + "pull_request:44027|pull_request:44028", + "pull_request:44027|pull_request:44029", + "pull_request:44027|pull_request:44030", + "pull_request:44028|pull_request:44029", + "pull_request:44028|pull_request:44030", + "pull_request:44029|pull_request:44030" + ], + "prepared_review_unit_hash": "e73142a3e2ca188db96f5f72a4c2d7e2cca36b8af82ec275f54438f9ea7e3a17", + "component_budget": { + "node_count": 13, + "item_count": 13, + "soft_pair_count": 39, + "serialized_chars": 39790, + "estimated_input_tokens": 9948, + "estimated_eval_tokens": 20152 + }, + "budget": { + "node_count": 13, + "item_count": 13, + "soft_pair_count": 39, + "serialized_chars": 39790, + "estimated_input_tokens": 9948, + "estimated_eval_tokens": 20152 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": false, + "analyst_result": { + "summary": "These are all model-specific output-tracing refactors, not duplicate fixes. They share a broad implementation theme, but each PR targets different model code paths and should be kept separate.", + "confidence": 0.94, + "canonical_issue_reason": null, + "canonical_pr_reason": "If a representative PR is needed, 44003 is the broadest of the set because it spans two related model implementations and adds shared output-capture plumbing; however, it is not a duplicate of the others.", + "best_issue_reason": null, + "best_pr_reason": "44003 is the strongest representative of the cluster because it has the widest model coverage and the clearest core refactor, but the set still consists of separate PRs rather than duplicates.", + "soft_edge_verdicts": [ + { + "left": "pull_request:43998", + "right": "pull_request:44002", + "accept": false, + "reason": "Different models and code paths; same output-tracing pattern only." + }, + { + "left": "pull_request:43998", + "right": "pull_request:44000", + "accept": false, + "reason": "Different model families; no shared concrete bug or change." + }, + { + "left": "pull_request:43998", + "right": "pull_request:44001", + "accept": false, + "reason": "Separate model-specific refactors, not one mergeable fix." + }, + { + "left": "pull_request:43998", + "right": "pull_request:44003", + "accept": false, + "reason": "Both are tracing refactors, but they change different model implementations and underlying paths." + }, + { + "left": "pull_request:43998", + "right": "pull_request:44029", + "accept": false, + "reason": "Unrelated architectures; same refactor theme is too broad to duplicate." + }, + { + "left": "pull_request:43999", + "right": "pull_request:44001", + "accept": false, + "reason": "Different models; only the tracing helper usage overlaps." + }, + { + "left": "pull_request:43999", + "right": "pull_request:44002", + "accept": false, + "reason": "Distinct model code paths with similar output-tracing cleanup." + }, + { + "left": "pull_request:43999", + "right": "pull_request:44025", + "accept": false, + "reason": "Different model families and files; not the same underlying fix." + }, + { + "left": "pull_request:43999", + "right": "pull_request:44026", + "accept": false, + "reason": "Same high-level refactor, but separate encoder-decoder model targets." + }, + { + "left": "pull_request:43999", + "right": "pull_request:44027", + "accept": false, + "reason": "Different model implementations; no duplicate code-path problem." + }, + { + "left": "pull_request:44000", + "right": "pull_request:44001", + "accept": false, + "reason": "Separate model-specific output-tracing refactors." + }, + { + "left": "pull_request:44000", + "right": "pull_request:44002", + "accept": false, + "reason": "Different architectures and forward paths; not one fix." + }, + { + "left": "pull_request:44000", + "right": "pull_request:44003", + "accept": false, + "reason": "Shared refactor style, but the affected models and logic are different." + }, + { + "left": "pull_request:44000", + "right": "pull_request:44025", + "accept": false, + "reason": "Different model stacks; same abstraction pattern only." + }, + { + "left": "pull_request:44001", + "right": "pull_request:44002", + "accept": false, + "reason": "These touch different models and do not address the same bug." + }, + { + "left": "pull_request:44001", + "right": "pull_request:44003", + "accept": false, + "reason": "Not the same code path; only the refactor mechanism matches." + }, + { + "left": "pull_request:44001", + "right": "pull_request:44025", + "accept": false, + "reason": "Distinct model implementations; no plausible single PR merge." + }, + { + "left": "pull_request:44002", + "right": "pull_request:44029", + "accept": false, + "reason": "Different model families; same output-tracing theme is insufficient." + }, + { + "left": "pull_request:44002", + "right": "pull_request:44003", + "accept": false, + "reason": "Separate model-specific changes, not a duplicate fix." + }, + { + "left": "pull_request:44003", + "right": "pull_request:44004", + "accept": false, + "reason": "Different models and forward implementations; not the same change." + }, + { + "left": "pull_request:43998", + "right": "pull_request:44004", + "accept": false, + "reason": "Model-specific refactors with no shared concrete bug." + }, + { + "left": "pull_request:44001", + "right": "pull_request:44004", + "accept": false, + "reason": "Different code paths; similarity is only at the refactor pattern level." + }, + { + "left": "pull_request:44002", + "right": "pull_request:44004", + "accept": false, + "reason": "Not the same underlying change; different model logic." + }, + { + "left": "pull_request:44000", + "right": "pull_request:44004", + "accept": false, + "reason": "Same broad subsystem, but distinct model-specific refactors." + }, + { + "left": "pull_request:44025", + "right": "pull_request:44026", + "accept": false, + "reason": "Different model families; same output-tracing modernization only." + }, + { + "left": "pull_request:44025", + "right": "pull_request:44027", + "accept": false, + "reason": "Separate encoder-decoder model refactors, not one bug fix." + }, + { + "left": "pull_request:44025", + "right": "pull_request:44028", + "accept": false, + "reason": "Different model and code path; no duplicate change." + }, + { + "left": "pull_request:44025", + "right": "pull_request:44029", + "accept": false, + "reason": "Different architectures; too broad to be a duplicate." + }, + { + "left": "pull_request:44025", + "right": "pull_request:44030", + "accept": false, + "reason": "Distinct model targets and fixes." + }, + { + "left": "pull_request:44026", + "right": "pull_request:44029", + "accept": false, + "reason": "Different model code paths; only the refactor theme overlaps." + }, + { + "left": "pull_request:44026", + "right": "pull_request:44027", + "accept": false, + "reason": "Separate model-specific changes, not a single mergeable PR." + }, + { + "left": "pull_request:44026", + "right": "pull_request:44028", + "accept": false, + "reason": "Different models and outputs; not the same underlying bug." + }, + { + "left": "pull_request:44026", + "right": "pull_request:44030", + "accept": false, + "reason": "Different code paths with a shared output-tracing pattern." + }, + { + "left": "pull_request:44027", + "right": "pull_request:44028", + "accept": false, + "reason": "Distinct model implementations; similarity is too generic." + }, + { + "left": "pull_request:44027", + "right": "pull_request:44029", + "accept": false, + "reason": "Different families and files; not duplicate changes." + }, + { + "left": "pull_request:44027", + "right": "pull_request:44030", + "accept": false, + "reason": "Same refactor motif, but unrelated model code paths." + }, + { + "left": "pull_request:44028", + "right": "pull_request:44029", + "accept": false, + "reason": "Different models and logic; no shared concrete fix." + }, + { + "left": "pull_request:44028", + "right": "pull_request:44030", + "accept": false, + "reason": "Separate model-specific refactors, not one PR." + }, + { + "left": "pull_request:44029", + "right": "pull_request:44030", + "accept": false, + "reason": "Different architectures and forward paths; not duplicates." + } + ] + }, + "evaluator_result": null, + "evaluator_used": false, + "retried": false, + "accepted_nontrivial_soft_edge": false, + "error_kind": null, + "error_message": null + }, + { + "label": "pull_request", + "component_index": 14, + "component_count": 31, + "review_unit_index": 1, + "review_unit_count": 1, + "status": "reviewed", + "reason": null, + "source": "llm", + "cache_hit": false, + "model": "gpt-5.4-mini?service_tier=flex", + "cluster_id": "cluster-44662-2", + "nodes": [ + "pull_request:44662", + "pull_request:45401" + ], + "soft_pairs": [ + "pull_request:44662|pull_request:45401" + ], + "prepared_review_unit_hash": "9ac8370b3f72014b9925d555cbe3bde2ad9e5729646ff4a6c8b2928bf44f4b42", + "component_budget": { + "node_count": 2, + "item_count": 2, + "soft_pair_count": 1, + "serialized_chars": 6551, + "estimated_input_tokens": 1638, + "estimated_eval_tokens": 3532 + }, + "budget": { + "node_count": 2, + "item_count": 2, + "soft_pair_count": 1, + "serialized_chars": 6551, + "estimated_input_tokens": 1638, + "estimated_eval_tokens": 3532 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": false, + "analyst_result": { + "summary": "These are two unrelated model-implementation PRs that happen to touch shared auto-registry and docs files; they are not the same underlying change.", + "confidence": 0.98, + "canonical_issue_reason": null, + "canonical_pr_reason": "PR #45401 is the more canonical representative because it has an explicit linked issue, a focused scope, and a model-specific implementation for Voxtral TTS.", + "best_issue_reason": null, + "best_pr_reason": "PR #45401 is the better standalone PR to keep in a cluster view because it is narrowly scoped and tied to a specific tracking issue, whereas PR #44662 is a separate PenguinVL implementation.", + "soft_edge_verdicts": [ + { + "left": "pull_request:44662", + "right": "pull_request:45401", + "accept": false, + "reason": "Different model additions with different code paths, filenames, and goals (PenguinVL vs Voxtral TTS); shared auto/doc registry edits are not enough to treat them as the same change." + } + ] + }, + "evaluator_result": null, + "evaluator_used": false, + "retried": false, + "accepted_nontrivial_soft_edge": false, + "error_kind": null, + "error_message": null + }, + { + "label": "pull_request", + "component_index": 15, + "component_count": 31, + "review_unit_index": 1, + "review_unit_count": 1, + "status": "reviewed", + "reason": null, + "source": "llm", + "cache_hit": false, + "model": "gpt-5.4-mini?service_tier=flex", + "cluster_id": "cluster-44979-2", + "nodes": [ + "pull_request:44979", + "pull_request:45363" + ], + "soft_pairs": [ + "pull_request:44979|pull_request:45363" + ], + "prepared_review_unit_hash": "5cef09eaf93cfb83956df7a7d0d498c81614fe51fd597ae3a4975de0f0550499", + "component_budget": { + "node_count": 2, + "item_count": 2, + "soft_pair_count": 1, + "serialized_chars": 4407, + "estimated_input_tokens": 1102, + "estimated_eval_tokens": 2460 + }, + "budget": { + "node_count": 2, + "item_count": 2, + "soft_pair_count": 1, + "serialized_chars": 4407, + "estimated_input_tokens": 1102, + "estimated_eval_tokens": 2460 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": false, + "analyst_result": { + "summary": "Both PRs concern fusion infrastructure in Transformers, but they target different layers: PR #44979 introduces a general Module Fusion API and tests, while PR #45363 implements a specific n-to-1 kernel fusion path via `KernelConfig` plus integration changes. Related theme, not the same underlying change.", + "confidence": 0.86, + "canonical_issue_reason": null, + "canonical_pr_reason": "PR #44979 is the broader foundational API change and is the better cluster anchor for the general fusion concept.", + "best_issue_reason": null, + "best_pr_reason": "PR #44979 best represents the cluster because it defines the reusable module-fusion abstraction rather than one concrete kernel-fusion application.", + "soft_edge_verdicts": [ + { + "left": "pull_request:44979", + "right": "pull_request:45363", + "accept": false, + "reason": "They share the fusion theme, but #44979 adds a generic module-fusion API, whereas #45363 wires up a specific n-to-1 kernel-fusion implementation through `KernelConfig` and model loading. Different concrete code paths and not plausibly one merged PR." + } + ] + }, + "evaluator_result": null, + "evaluator_used": false, + "retried": false, + "accepted_nontrivial_soft_edge": false, + "error_kind": null, + "error_message": null + }, + { + "label": "pull_request", + "component_index": 16, + "component_count": 31, + "review_unit_index": 1, + "review_unit_count": 1, + "status": "reviewed", + "reason": null, + "source": "llm", + "cache_hit": false, + "model": "gpt-5.4-mini?service_tier=flex", + "cluster_id": "cluster-45086-2", + "nodes": [ + "pull_request:45086", + "pull_request:45317" + ], + "soft_pairs": [ + "pull_request:45086|pull_request:45317" + ], + "prepared_review_unit_hash": "c13cf43b3eed59e27175ade45b8a1591a88abf5db2cde09117109fed16ad1449", + "component_budget": { + "node_count": 2, + "item_count": 2, + "soft_pair_count": 1, + "serialized_chars": 4414, + "estimated_input_tokens": 1104, + "estimated_eval_tokens": 2464 + }, + "budget": { + "node_count": 2, + "item_count": 2, + "soft_pair_count": 1, + "serialized_chars": 4414, + "estimated_input_tokens": 1104, + "estimated_eval_tokens": 2464 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": false, + "analyst_result": { + "summary": "Both pull requests address the same AttributeError in `_patch_mistral_regex` by switching from `backend_tokenizer.pre_tokenizer` to `tokenizer.pre_tokenizer`; PR 45317 is the more complete version because it also adds tests and clearer framing.", + "confidence": 0.98, + "canonical_issue_reason": null, + "canonical_pr_reason": "PR 45317 is the strongest canonical candidate: it fixes the same code path as 45086, adds test coverage, and is more explicit about the `fix_mistral_regex=True` case.", + "best_issue_reason": null, + "best_pr_reason": "PR 45317 is the best overall PR because it contains the same fix as 45086 plus regression tests, making it the more complete and merge-ready change.", + "soft_edge_verdicts": [ + { + "left": "pull_request:45086", + "right": "pull_request:45317", + "accept": true, + "reason": "Same underlying bug in `_patch_mistral_regex` on the same file and same line-level fix; the second PR is essentially the same patch with added tests, so they are duplicates rather than separate changes." + } + ] + }, + "evaluator_result": null, + "evaluator_used": false, + "retried": false, + "accepted_nontrivial_soft_edge": false, + "error_kind": null, + "error_message": null + }, + { + "label": "pull_request", + "component_index": 17, + "component_count": 31, + "review_unit_index": 1, + "review_unit_count": 1, + "status": "reviewed", + "reason": null, + "source": "llm", + "cache_hit": false, + "model": "gpt-5.4-mini?service_tier=flex", + "cluster_id": "cluster-45115-2", + "nodes": [ + "pull_request:45115", + "pull_request:45569" + ], + "soft_pairs": [ + "pull_request:45115|pull_request:45569" + ], + "prepared_review_unit_hash": "966439515184f94a5283a7d6f37fbd8e7a5f2f1550bbbacf131fa407b1857aa4", + "component_budget": { + "node_count": 2, + "item_count": 2, + "soft_pair_count": 1, + "serialized_chars": 6066, + "estimated_input_tokens": 1517, + "estimated_eval_tokens": 3290 + }, + "budget": { + "node_count": 2, + "item_count": 2, + "soft_pair_count": 1, + "serialized_chars": 6066, + "estimated_input_tokens": 1517, + "estimated_eval_tokens": 3290 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": false, + "analyst_result": { + "summary": "Both PRs work on NemotronH model support, but they target different changes: #45115 is a narrower refactor to inherit GraniteMoeHybrid-style internals and update conversion mappings, while #45569 is a broader split into separate NemotronH Dense/Sparse model families plus docs and auto-mapping updates. They overlap in shared core files, but not the same concrete fix/change.", + "confidence": 0.87, + "canonical_issue_reason": null, + "canonical_pr_reason": "PR #45569 is the better cluster representative because it is broader, more complete, and more user-facing: it introduces the dense/sparse model split, updates docs and auto mappings, and appears to subsume the architectural direction of NemotronH support.", + "best_issue_reason": null, + "best_pr_reason": "PR #45569 is the strongest standalone candidate since it covers the larger end-to-end NemotronH rework; #45115 is narrower and draft-only, focused mainly on inheritance/conversion plumbing.", + "soft_edge_verdicts": [ + { + "left": "pull_request:45115", + "right": "pull_request:45569", + "accept": false, + "reason": "Related NemotronH work, but not the same underlying change. #45115 is a refactor/conversion-mapping PR; #45569 adds separate NemotronH Dense/Sparse models, docs, and auto-mapping updates. Shared files alone are not enough to treat them as duplicates." + } + ] + }, + "evaluator_result": null, + "evaluator_used": false, + "retried": false, + "accepted_nontrivial_soft_edge": false, + "error_kind": null, + "error_message": null + }, + { + "label": "pull_request", + "component_index": 18, + "component_count": 31, + "review_unit_index": 1, + "review_unit_count": 1, + "status": "reviewed", + "reason": null, + "source": "llm", + "cache_hit": false, + "model": "gpt-5.4-mini?service_tier=flex", + "cluster_id": "cluster-45415-2", + "nodes": [ + "pull_request:45415", + "pull_request:45425" + ], + "soft_pairs": [ + "pull_request:45415|pull_request:45425" + ], + "prepared_review_unit_hash": "b980b2a2789630312891dee629839c19f0f121b265ba038553ea3aeff229f7d4", + "component_budget": { + "node_count": 2, + "item_count": 2, + "soft_pair_count": 1, + "serialized_chars": 5514, + "estimated_input_tokens": 1379, + "estimated_eval_tokens": 3014 + }, + "budget": { + "node_count": 2, + "item_count": 2, + "soft_pair_count": 1, + "serialized_chars": 5514, + "estimated_input_tokens": 1379, + "estimated_eval_tokens": 3014 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": false, + "analyst_result": { + "summary": "Both items are typing-focused pull requests, but they address different scopes: one is a broad repository-wide type-checking pass, the other is a narrower set of typing protocol/annotation fixes in core utilities. They are related, but not the same underlying change set, so I would not deduplicate them.", + "confidence": 0.82, + "canonical_issue_reason": null, + "canonical_pr_reason": "PR 45425 is the better canonical PR for this cluster because it is the more concrete, code-path-specific typing change set in core modules (`modeling_utils`, `_typing.py`, `peft.py`, flash-attention utilities) and has substantial review activity. PR 45415 is a much broader sweep across many unrelated files plus type-check config.", + "best_issue_reason": null, + "best_pr_reason": "PR 45425 is the strongest representative: it introduces specific typing protocols/aliases and targeted fixes that can be reasoned about as one change. PR 45415 is too broad and looks like a separate large-scale typing cleanup rather than the same patch.", + "soft_edge_verdicts": [ + { + "left": "pull_request:45415", + "right": "pull_request:45425", + "accept": false, + "reason": "Both are typing-related, but 45415 is a wide repo-wide type-checking cleanup while 45425 adds targeted protocols and annotations in core modules; they are not the same concrete fix and would not plausibly be merged as one PR." + } + ] + }, + "evaluator_result": null, + "evaluator_used": false, + "retried": false, + "accepted_nontrivial_soft_edge": false, + "error_kind": null, + "error_message": null + }, + { + "label": "pull_request", + "component_index": 19, + "component_count": 31, + "review_unit_index": 1, + "review_unit_count": 1, + "status": "reviewed", + "reason": null, + "source": "llm", + "cache_hit": false, + "model": "gpt-5.4-mini?service_tier=flex", + "cluster_id": "cluster-45548-2", + "nodes": [ + "pull_request:45548", + "pull_request:45662" + ], + "soft_pairs": [ + "pull_request:45548|pull_request:45662" + ], + "prepared_review_unit_hash": "df6630ff440056c3dd319a8ea7fe10b121f51d54c528c9b18a99588829d0791f", + "component_budget": { + "node_count": 2, + "item_count": 2, + "soft_pair_count": 1, + "serialized_chars": 4508, + "estimated_input_tokens": 1127, + "estimated_eval_tokens": 2510 + }, + "budget": { + "node_count": 2, + "item_count": 2, + "soft_pair_count": 1, + "serialized_chars": 4508, + "estimated_input_tokens": 1127, + "estimated_eval_tokens": 2510 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": false, + "analyst_result": { + "summary": "These are both EP-related PRs, but they fix different underlying bugs: one is about preserving distributed_config/has_ep during loading, the other about expert tensors being overwritten during FSDP2 post-shard handling. Not duplicates.", + "confidence": 0.94, + "canonical_issue_reason": null, + "canonical_pr_reason": "No single canonical PR for duplication purposes; if one must anchor the cluster, 45662 is the more concrete bug fix because it targets a specific overwrite path in model loading, while 45548 addresses a separate distributed_config serialization/loading issue.", + "best_issue_reason": null, + "best_pr_reason": "45662 is the stronger standalone PR: it identifies a precise failure mode (experts overwritten by rank-0 broadcast) and changes the relevant loading path directly. 45548 is also valid, but it addresses a different code path and bug.", + "soft_edge_verdicts": [ + { + "left": "pull_request:45548", + "right": "pull_request:45662", + "accept": false, + "reason": "Shared EP terminology and one common file are not enough: 45548 fixes distributed_config serialization/has_ep for ZeRO-3 loading, while 45662 fixes expert parameters being overwritten in FSDP2 post-shard handling. Different bug mechanisms and merge scopes." + } + ] + }, + "evaluator_result": null, + "evaluator_used": false, + "retried": false, + "accepted_nontrivial_soft_edge": false, + "error_kind": null, + "error_message": null + }, + { + "label": "pull_request", + "component_index": 20, + "component_count": 31, + "review_unit_index": 1, + "review_unit_count": 1, + "status": "reviewed", + "reason": null, + "source": "llm", + "cache_hit": false, + "model": "gpt-5.4-mini?service_tier=flex", + "cluster_id": "cluster-45639-2", + "nodes": [ + "pull_request:45639", + "pull_request:45645" + ], + "soft_pairs": [ + "pull_request:45639|pull_request:45645" + ], + "prepared_review_unit_hash": "04c9477aa31673a6543d9a2bf0ec375bff4eba4f0bd1886604a3d222a32f7b30", + "component_budget": { + "node_count": 2, + "item_count": 2, + "soft_pair_count": 1, + "serialized_chars": 4672, + "estimated_input_tokens": 1168, + "estimated_eval_tokens": 2592 + }, + "budget": { + "node_count": 2, + "item_count": 2, + "soft_pair_count": 1, + "serialized_chars": 4672, + "estimated_input_tokens": 1168, + "estimated_eval_tokens": 2592 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": false, + "analyst_result": { + "summary": "Both PRs target the same xdist-related captured_info logging problem and share the same core testing_utils change; PR 45645 is the broader, more complete variant because it also updates CI log collection to handle multiple per-worker files.", + "confidence": 0.95, + "canonical_issue_reason": null, + "canonical_pr_reason": "PR 45645 is the better canonical PR because it contains the same core xdist-safe output-file fix as 45639 and also updates the workflow that consumes those logs, making it the more complete end-to-end change.", + "best_issue_reason": null, + "best_pr_reason": "PR 45645 is the best representative of the cluster: it addresses the same concrete captured_info collision bug and includes the downstream CI handling needed to preserve debug logs.", + "soft_edge_verdicts": [ + { + "left": "pull_request:45639", + "right": "pull_request:45645", + "accept": true, + "reason": "Accept: both PRs fix the same underlying captured_info xdist collision issue in testing_utils; 45645 is a superset that adds workflow support for multiple output files." + } + ] + }, + "evaluator_result": null, + "evaluator_used": false, + "retried": false, + "accepted_nontrivial_soft_edge": false, + "error_kind": null, + "error_message": null + }, + { + "label": "pull_request", + "component_index": 21, + "component_count": 31, + "review_unit_index": 1, + "review_unit_count": 1, + "status": "reviewed", + "reason": null, + "source": "cache", + "cache_hit": true, + "model": "gpt-5.4-mini?service_tier=flex", + "cluster_id": "cluster-45737-2", + "nodes": [ + "pull_request:45737", + "pull_request:45738" + ], + "soft_pairs": [ + "pull_request:45737|pull_request:45738" + ], + "prepared_review_unit_hash": "cc2ade73a49ec6a8883a7a24c433988645588bcd049e8aad2bce56b1b24daa1c", + "component_budget": { + "node_count": 2, + "item_count": 2, + "soft_pair_count": 1, + "serialized_chars": 4131, + "estimated_input_tokens": 1033, + "estimated_eval_tokens": 2322 + }, + "budget": { + "node_count": 2, + "item_count": 2, + "soft_pair_count": 1, + "serialized_chars": 4131, + "estimated_input_tokens": 1033, + "estimated_eval_tokens": 2322 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": false, + "analyst_result": { + "summary": "Both PRs address the same MusicgenMelody cache initialization bug in `forward()`: replacing `EncoderDecoderCache(DynamicCache, DynamicCache)` with a single `DynamicCache`. PR #45737 is the more complete duplicate because it bundles the fix with a regression test, while #45738 is the fix-only variant.", + "confidence": 0.97, + "canonical_issue_reason": null, + "canonical_pr_reason": "PR #45737 is the best canonical PR because it contains the same code fix as #45738 plus a regression test, making it the most complete representation of the bug fix.", + "best_issue_reason": null, + "best_pr_reason": "PR #45737 is the strongest overall candidate: it fixes the concrete cache bug and adds a targeted regression test, whereas #45738 only changes the implementation.", + "soft_edge_verdicts": [ + { + "left": "pull_request:45737", + "right": "pull_request:45738", + "accept": true, + "reason": "Same underlying code-path bug in `musicgen_melody` cache setup; the diffs match on the core fix and the test-only addition in #45737 could plausibly live in the same PR as #45738." + } + ] + }, + "evaluator_result": { + "accept": true, + "feedback": "The assessment is grounded in the packet: both PRs make the same `musicgen_melody` cache initialization change, and #45737 only adds a regression test on top of that fix. The soft-edge accept is conservative enough because the diffs are mergeable into one PR for the same bugfix." + }, + "evaluator_used": true, + "retried": false, + "accepted_nontrivial_soft_edge": true, + "error_kind": null, + "error_message": null + }, + { + "label": "pull_request", + "component_index": 22, + "component_count": 31, + "review_unit_index": 1, + "review_unit_count": 1, + "status": "reviewed", + "reason": null, + "source": "llm", + "cache_hit": false, + "model": "gpt-5.4-mini?service_tier=flex", + "cluster_id": "cluster-46126-2", + "nodes": [ + "pull_request:46126", + "pull_request:46269" + ], + "soft_pairs": [ + "pull_request:46126|pull_request:46269" + ], + "prepared_review_unit_hash": "77f41d8c43112bb0d3bf62b680beaa88cec91aa146b7c552834bab7174de8b0b", + "component_budget": { + "node_count": 2, + "item_count": 2, + "soft_pair_count": 1, + "serialized_chars": 5525, + "estimated_input_tokens": 1382, + "estimated_eval_tokens": 3020 + }, + "budget": { + "node_count": 2, + "item_count": 2, + "soft_pair_count": 1, + "serialized_chars": 5525, + "estimated_input_tokens": 1382, + "estimated_eval_tokens": 3020 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": false, + "analyst_result": { + "summary": "Two PRs overlap in distributed code, but they are not the same change: one is a focused EP/trainer integration, the other is a broad distributed branch base with many unrelated updates.", + "confidence": 0.88, + "canonical_issue_reason": null, + "canonical_pr_reason": "PR 46126 is the more focused, concrete change set centered on EP + Trainer integration; PR 46269 is a much broader branch-base PR with many additional distributed and CI modifications, so it is not the best canonical representative of this cluster.", + "best_issue_reason": null, + "best_pr_reason": "PR 46126 is the best standalone PR here because it has a clear, specific purpose and a tighter code-path scope. PR 46269 is too broad and mixed-purpose to serve as the best representative of the underlying change.", + "soft_edge_verdicts": [ + { + "left": "pull_request:46126", + "right": "pull_request:46269", + "accept": false, + "reason": "Reject: the shared distributed filenames do not indicate the same underlying fix. 46126 is a targeted EP/trainer integration PR, while 46269 is a broad distributed branch-base PR spanning many unrelated areas, so they would not plausibly be merged as one PR." + } + ] + }, + "evaluator_result": null, + "evaluator_used": false, + "retried": false, + "accepted_nontrivial_soft_edge": false, + "error_kind": null, + "error_message": null + }, + { + "label": "pull_request", + "component_index": 23, + "component_count": 31, + "review_unit_index": 1, + "review_unit_count": 1, + "status": "reviewed", + "reason": null, + "source": "llm", + "cache_hit": false, + "model": "gpt-5.4-mini?service_tier=flex", + "cluster_id": "cluster-46138-10", + "nodes": [ + "pull_request:46138", + "pull_request:46432", + "pull_request:46501", + "pull_request:46502", + "pull_request:46503", + "pull_request:46504", + "pull_request:46554", + "pull_request:46558", + "pull_request:46563", + "pull_request:46716" + ], + "soft_pairs": [ + "pull_request:46138|pull_request:46554", + "pull_request:46138|pull_request:46558", + "pull_request:46138|pull_request:46563", + "pull_request:46138|pull_request:46716", + "pull_request:46138|pull_request:46504", + "pull_request:46432|pull_request:46503", + "pull_request:46432|pull_request:46501", + "pull_request:46432|pull_request:46502", + "pull_request:46432|pull_request:46554", + "pull_request:46432|pull_request:46558", + "pull_request:46501|pull_request:46503", + "pull_request:46501|pull_request:46502", + "pull_request:46501|pull_request:46504", + "pull_request:46501|pull_request:46554", + "pull_request:46502|pull_request:46503", + "pull_request:46502|pull_request:46504", + "pull_request:46502|pull_request:46554", + "pull_request:46503|pull_request:46504", + "pull_request:46503|pull_request:46554", + "pull_request:46504|pull_request:46554", + "pull_request:46504|pull_request:46558", + "pull_request:46554|pull_request:46558", + "pull_request:46554|pull_request:46563", + "pull_request:46554|pull_request:46716", + "pull_request:46558|pull_request:46563", + "pull_request:46558|pull_request:46716", + "pull_request:46563|pull_request:46716", + "pull_request:46504|pull_request:46563", + "pull_request:46432|pull_request:46563", + "pull_request:46504|pull_request:46716", + "pull_request:46432|pull_request:46716" + ], + "prepared_review_unit_hash": "d64e746b79370ae12a1d5ab8408da4a6734246e60d004982b8c1e2614acced21", + "component_budget": { + "node_count": 10, + "item_count": 10, + "soft_pair_count": 31, + "serialized_chars": 26678, + "estimated_input_tokens": 6670, + "estimated_eval_tokens": 13596 + }, + "budget": { + "node_count": 10, + "item_count": 10, + "soft_pair_count": 31, + "serialized_chars": 26678, + "estimated_input_tokens": 6670, + "estimated_eval_tokens": 13596 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": false, + "analyst_result": { + "summary": "This cluster is a loose mix of unrelated workflow-maintenance PRs. The only near-cluster is the set of pr-ci-security-gate.yml edits, but they address different security-hardening steps rather than the same concrete change, so there are no true duplicates here.", + "confidence": 0.94, + "canonical_issue_reason": null, + "canonical_pr_reason": "PR 46554 is the most representative PR in the only semi-cohesive subcluster: it has the broadest and most substantive security-gate workflow change and the strongest overlap with the other pr-ci-security-gate variants.", + "best_issue_reason": null, + "best_pr_reason": "PR 46554 is the best single representative for this mixed cluster because it is the most central member of the security-gate subcluster and the clearest example of a concrete workflow hardening change.", + "soft_edge_verdicts": [ + { + "left": "pull_request:46138", + "right": "pull_request:46554", + "accept": false, + "reason": "Different workflow fixes: self-comment stdin handling vs PR CI security-gate logic." + }, + { + "left": "pull_request:46138", + "right": "pull_request:46558", + "accept": false, + "reason": "Different workflow fixes: self-comment stdin handling vs PR CI security-gate permissions/approval behavior." + }, + { + "left": "pull_request:46138", + "right": "pull_request:46563", + "accept": false, + "reason": "Different workflow fixes: self-comment stdin handling vs reusable-workflow safety note." + }, + { + "left": "pull_request:46138", + "right": "pull_request:46716", + "accept": false, + "reason": "Different workflow fixes: self-comment stdin handling vs secret-scoping change in security gate." + }, + { + "left": "pull_request:46138", + "right": "pull_request:46504", + "accept": false, + "reason": "Different workflow fixes: self-comment stdin handling vs build-ci-docker-images command quoting." + }, + { + "left": "pull_request:46432", + "right": "pull_request:46503", + "accept": false, + "reason": "Unrelated workflow changes: SSH runner hardening vs benchmark dispatch trigger update." + }, + { + "left": "pull_request:46432", + "right": "pull_request:46501", + "accept": false, + "reason": "Unrelated workflow changes: SSH runner hardening vs anti-slop permissions cleanup." + }, + { + "left": "pull_request:46432", + "right": "pull_request:46502", + "accept": false, + "reason": "Unrelated workflow changes: SSH runner hardening vs assign-reviewers checkout ref change." + }, + { + "left": "pull_request:46432", + "right": "pull_request:46554", + "accept": false, + "reason": "Different workflow/file and different behavior: SSH runner hardening vs PR CI security-gate checks." + }, + { + "left": "pull_request:46432", + "right": "pull_request:46558", + "accept": false, + "reason": "Different workflow/file and different behavior: SSH runner hardening vs PR CI security-gate permissions/approval logic." + }, + { + "left": "pull_request:46432", + "right": "pull_request:46563", + "accept": false, + "reason": "Different workflow/file and different behavior: SSH runner hardening vs security-gate documentation/comment update." + }, + { + "left": "pull_request:46432", + "right": "pull_request:46716", + "accept": false, + "reason": "Different workflow/file and different behavior: SSH runner hardening vs secret-scoping change in security gate." + }, + { + "left": "pull_request:46501", + "right": "pull_request:46503", + "accept": false, + "reason": "Same broad workflows area, but different concrete changes: anti-slop permissions vs benchmark dispatch trigger." + }, + { + "left": "pull_request:46501", + "right": "pull_request:46502", + "accept": false, + "reason": "Different workflow behavior: anti-slop permissions cleanup vs reviewer-assignment checkout ref." + }, + { + "left": "pull_request:46501", + "right": "pull_request:46504", + "accept": false, + "reason": "Different workflow behavior: anti-slop permissions cleanup vs build-ci-docker-images command quoting." + }, + { + "left": "pull_request:46501", + "right": "pull_request:46554", + "accept": false, + "reason": "Different concrete change: anti-slop permissions cleanup vs PR CI security-gate hardening." + }, + { + "left": "pull_request:46502", + "right": "pull_request:46503", + "accept": false, + "reason": "Different workflow behavior: reviewer checkout ref vs benchmark dispatch trigger." + }, + { + "left": "pull_request:46502", + "right": "pull_request:46504", + "accept": false, + "reason": "Different workflow behavior: reviewer checkout ref vs build-ci-docker-images command quoting." + }, + { + "left": "pull_request:46502", + "right": "pull_request:46554", + "accept": false, + "reason": "Different concrete change: reviewer checkout ref vs PR CI security-gate hardening." + }, + { + "left": "pull_request:46503", + "right": "pull_request:46504", + "accept": false, + "reason": "Different workflow behavior: benchmark dispatch trigger vs build-ci-docker-images command quoting." + }, + { + "left": "pull_request:46503", + "right": "pull_request:46554", + "accept": false, + "reason": "Different concrete change: benchmark dispatch trigger vs PR CI security-gate hardening." + }, + { + "left": "pull_request:46504", + "right": "pull_request:46554", + "accept": false, + "reason": "Different workflow behavior: build-ci-docker-images command quoting vs PR CI security-gate checks." + }, + { + "left": "pull_request:46504", + "right": "pull_request:46558", + "accept": false, + "reason": "Different workflow behavior: build-ci-docker-images command quoting vs PR CI security-gate permissions/approval logic." + }, + { + "left": "pull_request:46554", + "right": "pull_request:46558", + "accept": false, + "reason": "Same file, but different concrete hardening steps: checkout-and-diff checks vs permissions/approval gating." + }, + { + "left": "pull_request:46554", + "right": "pull_request:46563", + "accept": false, + "reason": "Same file, but different concrete changes: workflow checks vs reusable-workflow safety commentary." + }, + { + "left": "pull_request:46554", + "right": "pull_request:46716", + "accept": false, + "reason": "Same file, but different concrete changes: workflow checks vs secrets scoping in the reusable workflow call." + }, + { + "left": "pull_request:46558", + "right": "pull_request:46563", + "accept": false, + "reason": "Same workflow family, but different fixes: approval gating vs safety-note documentation." + }, + { + "left": "pull_request:46558", + "right": "pull_request:46716", + "accept": false, + "reason": "Same workflow family, but different fixes: approval gating vs secret-scoping change." + }, + { + "left": "pull_request:46563", + "right": "pull_request:46716", + "accept": false, + "reason": "Same file, but different changes: documentation note vs secrets inheritance removal." + }, + { + "left": "pull_request:46504", + "right": "pull_request:46563", + "accept": false, + "reason": "Different workflow behavior: build-ci-docker-images command quoting vs security-gate documentation note." + }, + { + "left": "pull_request:46432", + "right": "pull_request:46563", + "accept": false, + "reason": "Different workflow/file and different behavior: SSH runner hardening vs security-gate documentation note." + }, + { + "left": "pull_request:46504", + "right": "pull_request:46716", + "accept": false, + "reason": "Different workflow behavior: build-ci-docker-images command quoting vs secret-scoping change." + }, + { + "left": "pull_request:46432", + "right": "pull_request:46716", + "accept": false, + "reason": "Different workflow/file and different behavior: SSH runner hardening vs security-gate secret-scoping change." + } + ] + }, + "evaluator_result": null, + "evaluator_used": false, + "retried": false, + "accepted_nontrivial_soft_edge": false, + "error_kind": null, + "error_message": null + }, + { + "label": "pull_request", + "component_index": 24, + "component_count": 31, + "review_unit_index": 1, + "review_unit_count": 1, + "status": "reviewed", + "reason": null, + "source": "cache", + "cache_hit": true, + "model": "gpt-5.4-mini?service_tier=flex", + "cluster_id": "cluster-46174-2", + "nodes": [ + "pull_request:46174", + "pull_request:46780" + ], + "soft_pairs": [ + "pull_request:46174|pull_request:46780" + ], + "prepared_review_unit_hash": "a48ce4ca162f555682f79411d7229ac1ccc08d7822dfab654bb881639fccddbc", + "component_budget": { + "node_count": 2, + "item_count": 2, + "soft_pair_count": 1, + "serialized_chars": 6030, + "estimated_input_tokens": 1508, + "estimated_eval_tokens": 3272 + }, + "budget": { + "node_count": 2, + "item_count": 2, + "soft_pair_count": 1, + "serialized_chars": 6030, + "estimated_input_tokens": 1508, + "estimated_eval_tokens": 3272 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": false, + "analyst_result": { + "summary": "Two PRs target the same reported MPS crash in the sinusoidal position embedding helper, but they implement different fixes and are not safe to treat as duplicates.", + "confidence": 0.89, + "canonical_issue_reason": null, + "canonical_pr_reason": "PR 46174 is the broader, more complete fix for the float64-on-MPS crash: it updates the helper across all affected model files and keeps the internal math on CPU float64 before casting back.", + "best_issue_reason": null, + "best_pr_reason": "PR 46174 is the best representative of the cluster because it is the original, more comprehensive fix and has more discussion/review context than the narrower follow-up.", + "soft_edge_verdicts": [ + { + "left": "pull_request:46174", + "right": "pull_request:46780", + "accept": false, + "reason": "They share the same issue target and helper, but the patches are different solutions: 46174 avoids float64 on MPS by doing the arithmetic on CPU first, while 46780 switches the helper to use dtype for internal arithmetic. Those are alternative implementations, not one mergeable duplicate PR." + } + ] + }, + "evaluator_result": null, + "evaluator_used": false, + "retried": false, + "accepted_nontrivial_soft_edge": false, + "error_kind": null, + "error_message": null + }, + { + "label": "pull_request", + "component_index": 25, + "component_count": 31, + "review_unit_index": 1, + "review_unit_count": 1, + "status": "reviewed", + "reason": null, + "source": "llm", + "cache_hit": false, + "model": "gpt-5.4-mini?service_tier=flex", + "cluster_id": "cluster-46310-2", + "nodes": [ + "pull_request:46310", + "pull_request:47162" + ], + "soft_pairs": [ + "pull_request:46310|pull_request:47162" + ], + "prepared_review_unit_hash": "54fa991378c2d1f6bcbcec5c227c7ee260121072e6fe6ac17b5f03d794b291a0", + "component_budget": { + "node_count": 2, + "item_count": 2, + "soft_pair_count": 1, + "serialized_chars": 4883, + "estimated_input_tokens": 1221, + "estimated_eval_tokens": 2698 + }, + "budget": { + "node_count": 2, + "item_count": 2, + "soft_pair_count": 1, + "serialized_chars": 4883, + "estimated_input_tokens": 1221, + "estimated_eval_tokens": 2698 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": false, + "analyst_result": { + "summary": "Both PRs fix the same MoonshineStreaming sliding-window right-bound off-by-one, changing `-dist < right_window_size` to `-dist <= right_window_size` in the same mask function. PR 47162 is broader because it also applies the same fix to Gemma4 and adds tests, but the shared MoonshineStreaming change is the same underlying bug.", + "confidence": 0.95, + "canonical_issue_reason": null, + "canonical_pr_reason": "PR 46310 is the narrower, more direct fix for the MoonshineStreaming off-by-one bug and best represents the specific issue target shared by both PRs.", + "best_issue_reason": null, + "best_pr_reason": "PR 47162 is the strongest representative overall because it includes the same MoonshineStreaming fix plus the related Gemma4 variant and tests, making it the most complete candidate in the cluster.", + "soft_edge_verdicts": [ + { + "left": "pull_request:46310", + "right": "pull_request:47162", + "accept": true, + "reason": "Same concrete off-by-one in the same MoonshineStreaming mask function; 47162 also carries the same fix and is plausibly part of one combined change set." + } + ] + }, + "evaluator_result": null, + "evaluator_used": false, + "retried": false, + "accepted_nontrivial_soft_edge": false, + "error_kind": null, + "error_message": null + }, + { + "label": "pull_request", + "component_index": 26, + "component_count": 31, + "review_unit_index": 1, + "review_unit_count": 1, + "status": "reviewed", + "reason": null, + "source": "llm", + "cache_hit": false, + "model": "gpt-5.4-mini?service_tier=flex", + "cluster_id": "cluster-46574-2", + "nodes": [ + "pull_request:46574", + "pull_request:46580" + ], + "soft_pairs": [ + "pull_request:46574|pull_request:46580" + ], + "prepared_review_unit_hash": "f68334f27ac9be91fae0fba9162a991fb2db64a72271925caa9b5e6cf1272af8", + "component_budget": { + "node_count": 2, + "item_count": 2, + "soft_pair_count": 1, + "serialized_chars": 4410, + "estimated_input_tokens": 1103, + "estimated_eval_tokens": 2462 + }, + "budget": { + "node_count": 2, + "item_count": 2, + "soft_pair_count": 1, + "serialized_chars": 4410, + "estimated_input_tokens": 1103, + "estimated_eval_tokens": 2462 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": false, + "analyst_result": { + "summary": "These are two different pull requests touching different code paths: one fixes CodeLlama decode behavior by moving tokenizer decoder setup, while the other fixes AutoTokenizer class selection for specific DeepSeek checkpoints. They are not duplicates and should not be merged together.", + "confidence": 0.97, + "canonical_issue_reason": null, + "canonical_pr_reason": "No canonical PR for duplication: the PRs address separate bugs in separate subsystems (CodeLlama tokenizer decode vs AutoTokenizer checkpoint routing).", + "best_issue_reason": null, + "best_pr_reason": "PR #46574 is the cleaner representative of a single concrete bug fix, with a focused code change and an explicit issue target; PR #46580 is a different, unrelated compatibility fix.", + "soft_edge_verdicts": [ + { + "left": "pull_request:46574", + "right": "pull_request:46580", + "accept": false, + "reason": "Reject: they solve different problems on different paths. #46574 fixes CodeLlama decoder setup so real leading spaces are preserved on decode; #46580 changes AutoTokenizer checkpoint-prefix routing for DeepSeek models that load the wrong tokenizer class. They are not the same underlying change and would not be merged as one PR." + } + ] + }, + "evaluator_result": null, + "evaluator_used": false, + "retried": false, + "accepted_nontrivial_soft_edge": false, + "error_kind": null, + "error_message": null + }, + { + "label": "pull_request", + "component_index": 27, + "component_count": 31, + "review_unit_index": 1, + "review_unit_count": 1, + "status": "reviewed", + "reason": null, + "source": "llm", + "cache_hit": false, + "model": "gpt-5.4-mini?service_tier=flex", + "cluster_id": "cluster-46773-2", + "nodes": [ + "pull_request:46773", + "pull_request:47087" + ], + "soft_pairs": [ + "pull_request:46773|pull_request:47087" + ], + "prepared_review_unit_hash": "f2a70d74e39fef43949049812bdd0b003a9eef8dbc11aabce1751176a13d9433", + "component_budget": { + "node_count": 2, + "item_count": 2, + "soft_pair_count": 1, + "serialized_chars": 5821, + "estimated_input_tokens": 1456, + "estimated_eval_tokens": 3168 + }, + "budget": { + "node_count": 2, + "item_count": 2, + "soft_pair_count": 1, + "serialized_chars": 5821, + "estimated_input_tokens": 1456, + "estimated_eval_tokens": 3168 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": false, + "analyst_result": { + "summary": "The two pull requests address different masking bugs: one fixes padding masks being skipped for batch size 1 in linear-attention models, while the other fixes padding masks being skipped on continued forwards / chunked prefill / cache continuation. They share some masking code but are not the same underlying change.", + "confidence": 0.92, + "canonical_issue_reason": null, + "canonical_pr_reason": "47087 is the better canonical PR because it fixes the broader shared masking utility behavior, has an explicit issue target, and covers the recurrent continued-forward path across several model families. 46773 is narrower and focused on a batch-size-1 edge case in linear-attention models.", + "best_issue_reason": null, + "best_pr_reason": "47087 is the strongest overall PR in the cluster: it addresses a concrete, cross-model masking bug in the shared utility layer and is clearly scoped around one code-path problem. 46773 is useful but more specialized and less representative of the cluster.", + "soft_edge_verdicts": [ + { + "left": "pull_request:46773", + "right": "pull_request:47087", + "accept": false, + "reason": "Different bugs and code paths: 46773 fixes batch-size-1 masking being skipped in linear-attention models, while 47087 fixes masking being skipped on continued forwards/chunked prefill/cache continuation. Shared masking code is not enough to treat them as duplicates." + } + ] + }, + "evaluator_result": null, + "evaluator_used": false, + "retried": false, + "accepted_nontrivial_soft_edge": false, + "error_kind": null, + "error_message": null + }, + { + "label": "pull_request", + "component_index": 28, + "component_count": 31, + "review_unit_index": 1, + "review_unit_count": 1, + "status": "reviewed", + "reason": null, + "source": "llm", + "cache_hit": false, + "model": "gpt-5.4-mini?service_tier=flex", + "cluster_id": "cluster-47074-3", + "nodes": [ + "pull_request:47074", + "pull_request:47076", + "pull_request:47195" + ], + "soft_pairs": [ + "pull_request:47074|pull_request:47076", + "pull_request:47074|pull_request:47195" + ], + "prepared_review_unit_hash": "024bbf134c22cbafc8901d61314a057784c87b7ae8c78621da1aeedb3453b378", + "component_budget": { + "node_count": 3, + "item_count": 3, + "soft_pair_count": 2, + "serialized_chars": 6433, + "estimated_input_tokens": 1609, + "estimated_eval_tokens": 3474 + }, + "budget": { + "node_count": 3, + "item_count": 3, + "soft_pair_count": 2, + "serialized_chars": 6433, + "estimated_input_tokens": 1609, + "estimated_eval_tokens": 3474 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": false, + "analyst_result": { + "summary": "These are three unrelated pull requests in different code paths: multimodal pipeline key normalization, Gemma4 audio placeholder validation, and generation length-capping. The soft similarities are superficial, so none should be merged as duplicates.", + "confidence": 0.96, + "canonical_issue_reason": null, + "canonical_pr_reason": "No canonical PR: the PRs target different bugs in different subsystems and are not mergeable into one change.", + "best_issue_reason": null, + "best_pr_reason": "47074 is the clearest standalone bug fix with a minimal code change and tests, but it is not a duplicate of the other PRs.", + "soft_edge_verdicts": [ + { + "left": "pull_request:47074", + "right": "pull_request:47076", + "accept": false, + "reason": "Different bugs and code paths: AnyToAnyPipeline input key normalization vs Gemma4 processor placeholder validation." + }, + { + "left": "pull_request:47074", + "right": "pull_request:47195", + "accept": false, + "reason": "Different areas and fixes: multimodal pipeline input keys vs generation candidate length budgeting." + } + ] + }, + "evaluator_result": null, + "evaluator_used": false, + "retried": false, + "accepted_nontrivial_soft_edge": false, + "error_kind": null, + "error_message": null + }, + { + "label": "pull_request", + "component_index": 29, + "component_count": 31, + "review_unit_index": 1, + "review_unit_count": 1, + "status": "reviewed", + "reason": null, + "source": "llm", + "cache_hit": false, + "model": "gpt-5.4-mini?service_tier=flex", + "cluster_id": "cluster-47122-10", + "nodes": [ + "pull_request:47122", + "pull_request:47123", + "pull_request:47124", + "pull_request:47224", + "pull_request:47266", + "pull_request:47280", + "pull_request:47281", + "pull_request:47282", + "pull_request:47310", + "pull_request:47380" + ], + "soft_pairs": [ + "pull_request:47122|pull_request:47281", + "pull_request:47122|pull_request:47123", + "pull_request:47122|pull_request:47280", + "pull_request:47122|pull_request:47224", + "pull_request:47122|pull_request:47310", + "pull_request:47123|pull_request:47280", + "pull_request:47123|pull_request:47266", + "pull_request:47123|pull_request:47281", + "pull_request:47123|pull_request:47124", + "pull_request:47124|pull_request:47310", + "pull_request:47124|pull_request:47282", + "pull_request:47124|pull_request:47380", + "pull_request:47124|pull_request:47281", + "pull_request:47266|pull_request:47280", + "pull_request:47266|pull_request:47380", + "pull_request:47266|pull_request:47281", + "pull_request:47124|pull_request:47266", + "pull_request:47280|pull_request:47281", + "pull_request:47280|pull_request:47310", + "pull_request:47280|pull_request:47380", + "pull_request:47281|pull_request:47282", + "pull_request:47282|pull_request:47380", + "pull_request:47282|pull_request:47310", + "pull_request:47280|pull_request:47282", + "pull_request:47310|pull_request:47380" + ], + "prepared_review_unit_hash": "3c8f9e6663b4b5d7de0b33474a92cbd249c4c8a644e5b27e376601f3489162f2", + "component_budget": { + "node_count": 10, + "item_count": 10, + "soft_pair_count": 25, + "serialized_chars": 28950, + "estimated_input_tokens": 7238, + "estimated_eval_tokens": 14732 + }, + "budget": { + "node_count": 10, + "item_count": 10, + "soft_pair_count": 25, + "serialized_chars": 28950, + "estimated_input_tokens": 7238, + "estimated_eval_tokens": 14732 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": false, + "analyst_result": { + "summary": "This cluster is heterogeneous and splits into separate duplicate groups rather than one shared fix. The clearest duplicates are the MiniMax-M3-VL load_error/offload-folder pair (47123, 47280) and the Mamba2 OOM pair (47282, 47380); the remaining PRs target different models or different failure modes.", + "confidence": 0.91, + "canonical_issue_reason": null, + "canonical_pr_reason": "The cluster is mixed, but 47380 is a strong representative of one real duplicate subgroup: it addresses the same Mamba2 OOM test issue as 47282 with a concrete memory-cleanup fix in the same test file.", + "best_issue_reason": null, + "best_pr_reason": "47380 is the best single PR representative because it fixes a concrete, self-contained Mamba2 OOM path rather than just adjusting expectations, and it clearly matches 47282.", + "soft_edge_verdicts": [ + { + "left": "pull_request:47122", + "right": "pull_request:47281", + "accept": false, + "reason": "Different models and unrelated fixes: Edgetam kwargs handling vs GLM46V model changes." + }, + { + "left": "pull_request:47122", + "right": "pull_request:47123", + "accept": false, + "reason": "Different models and failure modes: Edgetam import_or_config vs MiniMax-M3-VL load_error." + }, + { + "left": "pull_request:47122", + "right": "pull_request:47280", + "accept": false, + "reason": "Different models and unrelated code paths; not the same bug/change." + }, + { + "left": "pull_request:47122", + "right": "pull_request:47224", + "accept": false, + "reason": "47224 bundles Edgetam with other tokenizer/whisper fixes, so it is not the same standalone change as 47122." + }, + { + "left": "pull_request:47122", + "right": "pull_request:47310", + "accept": false, + "reason": "Different models and issues: Edgetam kwargs passthrough vs Gemma3 cache-implementation behavior." + }, + { + "left": "pull_request:47123", + "right": "pull_request:47280", + "accept": true, + "reason": "Same MiniMax-M3-VL load_error/offload-folder workaround in the same test path." + }, + { + "left": "pull_request:47123", + "right": "pull_request:47266", + "accept": false, + "reason": "Both are load_error-style fixes, but for different models and different root causes." + }, + { + "left": "pull_request:47123", + "right": "pull_request:47281", + "accept": false, + "reason": "Different model and unrelated fix area." + }, + { + "left": "pull_request:47123", + "right": "pull_request:47124", + "accept": false, + "reason": "Different models and symptoms; MiniMax load_error is not the same as Falcon-Mamba output mismatch." + }, + { + "left": "pull_request:47124", + "right": "pull_request:47310", + "accept": false, + "reason": "Different models and concrete issues; Falcon-Mamba output mismatch vs Gemma3 generation-config handling." + }, + { + "left": "pull_request:47124", + "right": "pull_request:47282", + "accept": false, + "reason": "Different model and failure mode; output mismatch is not the same as Mamba2 OOM." + }, + { + "left": "pull_request:47124", + "right": "pull_request:47380", + "accept": false, + "reason": "Different models and unrelated fixes." + }, + { + "left": "pull_request:47124", + "right": "pull_request:47281", + "accept": false, + "reason": "Different model and code path." + }, + { + "left": "pull_request:47266", + "right": "pull_request:47280", + "accept": false, + "reason": "Different models and different issues; DeepSeekV3.2 load_error is not the MiniMax offload-folder fix." + }, + { + "left": "pull_request:47266", + "right": "pull_request:47380", + "accept": false, + "reason": "Different models and failure modes." + }, + { + "left": "pull_request:47266", + "right": "pull_request:47281", + "accept": false, + "reason": "DeepSeek integration-test setup vs GLM46V model rewrite are unrelated." + }, + { + "left": "pull_request:47280", + "right": "pull_request:47281", + "accept": false, + "reason": "Different models and unrelated fixes." + }, + { + "left": "pull_request:47280", + "right": "pull_request:47310", + "accept": false, + "reason": "Different models and issues." + }, + { + "left": "pull_request:47280", + "right": "pull_request:47380", + "accept": false, + "reason": "Different models; MiniMax-M3-VL load_error is not the same as Mamba2 OOM cleanup." + }, + { + "left": "pull_request:47281", + "right": "pull_request:47282", + "accept": false, + "reason": "GLM46V model changes vs Mamba2 OOM test fix are unrelated." + }, + { + "left": "pull_request:47282", + "right": "pull_request:47380", + "accept": true, + "reason": "Same Mamba2 OOM issue in the same integration test file; both are memory-hygiene workarounds for the same bug." + }, + { + "left": "pull_request:47282", + "right": "pull_request:47310", + "accept": false, + "reason": "Different models and failure modes: Mamba2 OOM vs Gemma3 cache-implementation behavior." + }, + { + "left": "pull_request:47280", + "right": "pull_request:47282", + "accept": false, + "reason": "Different models and unrelated issues." + }, + { + "left": "pull_request:47310", + "right": "pull_request:47380", + "accept": false, + "reason": "Different models and independent fixes." + } + ] + }, + "evaluator_result": null, + "evaluator_used": false, + "retried": false, + "accepted_nontrivial_soft_edge": false, + "error_kind": null, + "error_message": null + }, + { + "label": "pull_request", + "component_index": 30, + "component_count": 31, + "review_unit_index": 1, + "review_unit_count": 1, + "status": "reviewed", + "reason": null, + "source": "llm", + "cache_hit": false, + "model": "gpt-5.4-mini?service_tier=flex", + "cluster_id": "cluster-47140-2", + "nodes": [ + "pull_request:47140", + "pull_request:47303" + ], + "soft_pairs": [ + "pull_request:47140|pull_request:47303" + ], + "prepared_review_unit_hash": "bb7a39453883a6637e8b07fc394147af51eea240759278f44c01b0248ad7367c", + "component_budget": { + "node_count": 2, + "item_count": 2, + "soft_pair_count": 1, + "serialized_chars": 4409, + "estimated_input_tokens": 1103, + "estimated_eval_tokens": 2462 + }, + "budget": { + "node_count": 2, + "item_count": 2, + "soft_pair_count": 1, + "serialized_chars": 4409, + "estimated_input_tokens": 1103, + "estimated_eval_tokens": 2462 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": false, + "analyst_result": { + "summary": "Both PRs address the same chat CLI bug: management endpoints were being built against a `/v1` API base instead of the service root. They appear to be alternative fixes for the same code path in `src/transformers/cli/chat.py`.", + "confidence": 0.78, + "canonical_issue_reason": null, + "canonical_pr_reason": "PR #47140 is the more direct canonical fix: it normalizes the CLI base URL once in `RichInterface`, so all management requests inherit the correct service-root URL rather than patching individual calls.", + "best_issue_reason": null, + "best_pr_reason": "PR #47140 is the best overall fit because it cleanly addresses the root cause for the chat CLI\u2019s management URLs and is likely to cover the affected call paths consistently.", + "soft_edge_verdicts": [ + { + "left": "pull_request:47140", + "right": "pull_request:47303", + "accept": true, + "reason": "Both PRs fix the same underlying chat CLI management-URL bug: requests were being sent relative to a `/v1` base instead of the service root. The helpers and URL rewrites are alternative implementations of the same change, so this is a valid duplicate/overlap pair." + } + ] + }, + "evaluator_result": { + "accept": true, + "feedback": "The pair looks grounded as an overlap: both PRs change `src/transformers/cli/chat.py`/tests to route chat management calls to the service root instead of inheriting a `/v1` base. The implementations differ (central base-URL normalization vs. per-call routing helper), but they appear to target the same bug family and could plausibly be merged into one fix." + }, + "evaluator_used": true, + "retried": false, + "accepted_nontrivial_soft_edge": true, + "error_kind": null, + "error_message": null + }, + { + "label": "pull_request", + "component_index": 31, + "component_count": 31, + "review_unit_index": 1, + "review_unit_count": 1, + "status": "reviewed", + "reason": null, + "source": "llm", + "cache_hit": false, + "model": "gpt-5.4-mini?service_tier=flex", + "cluster_id": "cluster-47261-2", + "nodes": [ + "pull_request:47261", + "pull_request:47262" + ], + "soft_pairs": [ + "pull_request:47261|pull_request:47262" + ], + "prepared_review_unit_hash": "ef65887c2246e335615ec85da7d2368cc6b7b1fb1a75377c86778156799ad9a6", + "component_budget": { + "node_count": 2, + "item_count": 2, + "soft_pair_count": 1, + "serialized_chars": 4418, + "estimated_input_tokens": 1105, + "estimated_eval_tokens": 2466 + }, + "budget": { + "node_count": 2, + "item_count": 2, + "soft_pair_count": 1, + "serialized_chars": 4418, + "estimated_input_tokens": 1105, + "estimated_eval_tokens": 2466 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": false, + "analyst_result": { + "summary": "Two unrelated pull requests: one fixes label smoothing/gradient accumulation loss scaling in training, the other adds and enforces `min_tokens_to_keep` for TopH logits warping in generation. They touch different code paths and should not be deduplicated.", + "confidence": 0.98, + "canonical_issue_reason": null, + "canonical_pr_reason": "PR 47261 is the better canonical PR if one must be chosen: it has an explicit issue target (46731) and a clearly scoped training-loss fix with tests.", + "best_issue_reason": null, + "best_pr_reason": "PR 47261 is the strongest representative of the cluster because it is issue-linked and has the clearest bug-to-fix correspondence; PR 47262 is a separate generation feature/bugfix.", + "soft_edge_verdicts": [ + { + "left": "pull_request:47261", + "right": "pull_request:47262", + "accept": false, + "reason": "Different underlying problems in different subsystems: training loss scaling vs generation logits warping. No shared code path or mergeable common fix." + } + ] + }, + "evaluator_result": null, + "evaluator_used": false, + "retried": false, + "accepted_nontrivial_soft_edge": false, + "error_kind": null, + "error_message": null + } + ] +} diff --git a/snapshots/20260718T120032Z/analysis-runs/hybrid-model-20260718t120032z/manifest.json b/snapshots/20260718T120032Z/analysis-runs/hybrid-model-20260718t120032z/manifest.json new file mode 100644 index 0000000000000000000000000000000000000000..68955d7fc12e15eb4b0ac8f4dcae740fa286f130 --- /dev/null +++ b/snapshots/20260718T120032Z/analysis-runs/hybrid-model-20260718t120032z/manifest.json @@ -0,0 +1,14 @@ +{ + "analysis_id": "hybrid-model-20260718t120032z", + "artifacts": { + "hybrid": "snapshots/20260718T120032Z/analysis-runs/hybrid-model-20260718t120032z/analysis-report-hybrid.json", + "hybrid_reviews": "snapshots/20260718T120032Z/analysis-runs/hybrid-model-20260718t120032z/analysis-report-hybrid.llm-reviews.json" + }, + "channel": "canonical", + "model": null, + "published_at": "2026-07-18T12:15:00Z", + "repo": "huggingface/transformers", + "schema_version": 1, + "snapshot_id": "20260718T120032Z", + "variant": "hybrid" +} diff --git a/snapshots/20260718T120032Z/manifest.json b/snapshots/20260718T120032Z/manifest.json index 1f3cc331107b1bb70ae89c4b413750be68f0288b..cbda00b9bcc1154a2d83d729a46bd08be6e38895 100644 --- a/snapshots/20260718T120032Z/manifest.json +++ b/snapshots/20260718T120032Z/manifest.json @@ -32,6 +32,24 @@ "timeline_events": 97 }, "extracted_at": "2026-07-18T12:00:32Z", + "published_analysis": { + "canonical_analysis_id": "hybrid-model-20260718t120032z", + "runs": { + "hybrid-model-20260718t120032z": { + "analysis_id": "hybrid-model-20260718t120032z", + "artifacts": { + "hybrid": "snapshots/20260718T120032Z/analysis-runs/hybrid-model-20260718t120032z/analysis-report-hybrid.json", + "hybrid_reviews": "snapshots/20260718T120032Z/analysis-runs/hybrid-model-20260718t120032z/analysis-report-hybrid.llm-reviews.json" + }, + "channel": "canonical", + "manifest_path": "snapshots/20260718T120032Z/analysis-runs/hybrid-model-20260718t120032z/manifest.json", + "model": null, + "published_at": "2026-07-18T12:15:00Z", + "variant": "hybrid" + } + }, + "schema_version": 1 + }, "repo": "huggingface/transformers", "snapshot_id": "20260718T120032Z", "watermark": {