diff --git a/analysis/current/analysis-report-hybrid.json b/analysis/current/analysis-report-hybrid.json index 3cd08a39fd09e6f2a5b7e3c7241aa917508d33c4..4262aa64c0da6c0ed019af9382047837cfeba124 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": "20260615T120125Z", - "generated_at": "2026-06-15T12:12:43Z", + "snapshot_id": "20260615T180022Z", + "generated_at": "2026-06-15T18:12:03Z", "evidence_quality": "full", "llm_enrichment": true, "meta_bugs": [ @@ -1148,56 +1148,6 @@ } ] }, - { - "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-41115-3", "summary": "Cluster of 1 issues and 2 PRs centered on issue #41115.", @@ -1372,6 +1322,56 @@ } ] }, + { + "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.", @@ -1444,10 +1444,10 @@ { "left_pr_number": 46173, "right_pr_number": 46474, - "code_similarity": 1.0, + "code_similarity": 0.65, "size_similarity": 1.0, "file_overlap": 1.0, - "area_overlap": 1.0, + "area_overlap": 0.0, "patch_similarity": 1.0, "shared_filenames": [ "src/transformers/cache_utils.py" @@ -1467,12 +1467,12 @@ ], "right_ranges": [ [ - 69, - 75 + 79, + 85 ], [ - 375, - 381 + 461, + 467 ] ] } @@ -1559,7 +1559,7 @@ "cluster_id": "cluster-46305-3", "summary": "Cluster of 1 issues and 2 PRs centered on issue #46305.", "status": "open", - "confidence": 0.55, + "confidence": 0.8, "canonical_issue_number": 46305, "canonical_pr_number": 46546, "issue_numbers": [ @@ -1570,7 +1570,9 @@ 46546 ], "evidence_types": [ - "closing_reference" + "closing_reference", + "shared_issue_target", + "soft_similarity" ], "pr_comparisons": [ { @@ -1620,72 +1622,6 @@ } ] }, - { - "cluster_id": "cluster-46614-3", - "summary": "Cluster of 1 issues and 2 PRs centered on issue #46614.", - "status": "open", - "confidence": 0.75, - "canonical_issue_number": 46614, - "canonical_pr_number": 46616, - "issue_numbers": [ - 46614 - ], - "pr_numbers": [ - 46616, - 46638 - ], - "evidence_types": [ - "closing_reference", - "shared_issue_target" - ], - "pr_comparisons": [ - { - "left_pr_number": 46616, - "right_pr_number": 46638, - "code_similarity": 0.83, - "size_similarity": 0.917, - "file_overlap": 1.0, - "area_overlap": 0.828, - "patch_similarity": 0.377, - "shared_filenames": [ - "src/transformers/models/wav2vec2_phoneme/tokenization_wav2vec2_phoneme.py", - "tests/models/wav2vec2_phoneme/test_tokenization_wav2vec2_phoneme.py" - ], - "shared_file_areas": [ - { - "filename": "src/transformers/models/wav2vec2_phoneme/tokenization_wav2vec2_phoneme.py", - "left_ranges": [ - [ - 255, - 261 - ] - ], - "right_ranges": [ - [ - 256, - 262 - ] - ] - }, - { - "filename": "tests/models/wav2vec2_phoneme/test_tokenization_wav2vec2_phoneme.py", - "left_ranges": [ - [ - 229, - 249 - ] - ], - "right_ranges": [ - [ - 229, - 246 - ] - ] - } - ] - } - ] - }, { "cluster_id": "cluster-45081-3", "summary": "Cluster of 1 issues and 2 PRs centered on issue #45081.", @@ -1745,60 +1681,6 @@ } ] }, - { - "cluster_id": "cluster-46612-3", - "summary": "Cluster of 1 issues and 2 PRs centered on issue #46612.", - "status": "open", - "confidence": 0.75, - "canonical_issue_number": 46612, - "canonical_pr_number": 46639, - "issue_numbers": [ - 46612 - ], - "pr_numbers": [ - 46615, - 46639 - ], - "evidence_types": [ - "closing_reference", - "shared_issue_target" - ], - "pr_comparisons": [ - { - "left_pr_number": 46615, - "right_pr_number": 46639, - "code_similarity": 0.204, - "size_similarity": 0.144, - "file_overlap": 0.5, - "area_overlap": 0.0, - "patch_similarity": 0.169, - "shared_filenames": [ - "src/transformers/generation/utils.py" - ], - "shared_file_areas": [ - { - "filename": "src/transformers/generation/utils.py", - "left_ranges": [ - [ - 955, - 989 - ], - [ - 3421, - 3429 - ] - ], - "right_ranges": [ - [ - 3392, - 3404 - ] - ] - } - ] - } - ] - }, { "cluster_id": "cluster-43698-3", "summary": "Cluster of 1 issues and 2 PRs centered on issue #43698.", @@ -1858,102 +1740,273 @@ ] }, { - "cluster_id": "cluster-46562-2", - "summary": "Cluster of 2 related pull requests linked by soft_similarity.", - "status": "open", - "confidence": 0.5, - "canonical_issue_number": null, - "canonical_pr_number": 46562, - "issue_numbers": [], - "pr_numbers": [ - 46562, - 46658 - ], - "evidence_types": [ - "soft_similarity" - ], - "pr_comparisons": [ - { - "left_pr_number": 46562, - "right_pr_number": 46658, - "code_similarity": 0.127, - "size_similarity": 0.556, - "file_overlap": 0.0, - "area_overlap": 0.0, - "patch_similarity": 0.107, - "shared_filenames": [], - "shared_file_areas": [] - } - ] - }, - { - "cluster_id": "cluster-43656-2", + "cluster_id": "cluster-44018-2", "summary": "Cluster of 2 related pull requests linked by soft_similarity.", "status": "open", "confidence": 0.5, "canonical_issue_number": null, - "canonical_pr_number": 43656, + "canonical_pr_number": 44068, "issue_numbers": [], "pr_numbers": [ - 43656, - 43842 + 44018, + 44068 ], "evidence_types": [ "soft_similarity" ], "pr_comparisons": [ { - "left_pr_number": 43656, - "right_pr_number": 43842, - "code_similarity": 0.405, - "size_similarity": 0.5, + "left_pr_number": 44018, + "right_pr_number": 44068, + "code_similarity": 0.766, + "size_similarity": 0.939, "file_overlap": 1.0, - "area_overlap": 0.0, - "patch_similarity": 0.036, + "area_overlap": 0.425, + "patch_similarity": 0.866, "shared_filenames": [ - "src/transformers/cli/serve.py" + "src/transformers/models/gpt_neo/modeling_gpt_neo.py", + "tests/models/gpt_neo/test_modeling_gpt_neo.py" ], "shared_file_areas": [ { - "filename": "src/transformers/cli/serve.py", + "filename": "src/transformers/models/gpt_neo/modeling_gpt_neo.py", "left_ranges": [ [ - 11, - 18 + 26, + 46 ], [ - 30, - 36 + 138, + 143 + ], + [ + 182, + 187 ], [ - 315, - 323 + 283, + 295 ], [ - 665, - 671 + 327, + 341 ], [ - 931, - 937 + 348, + 354 ], [ - 1843, - 1849 + 360, + 369 ], [ - 1868, - 1874 + 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": [ [ - 54, - 61 + 26, + 31 + ], + [ + 34, + 44 ], [ - 587, - 593 + 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 ] ] } @@ -1991,15 +2044,6 @@ "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, @@ -2026,6 +2070,15 @@ "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, @@ -2069,34 +2122,7 @@ 46310 ], "target_issue_number": 46305, - "reason": "PRs in cluster-46305-3 are treated as duplicates because they converge on issue #46305 with closing_reference evidence." - }, - { - "cluster_id": "cluster-46562-2", - "canonical_pr_number": 46562, - "duplicate_pr_numbers": [ - 46658 - ], - "target_issue_number": null, - "reason": "PRs in cluster-46562-2 are treated as duplicates because they share soft_similarity evidence." - }, - { - "cluster_id": "cluster-46612-3", - "canonical_pr_number": 46639, - "duplicate_pr_numbers": [ - 46615 - ], - "target_issue_number": 46612, - "reason": "PRs in cluster-46612-3 are treated as duplicates because they converge on issue #46612 with closing_reference, shared_issue_target evidence." - }, - { - "cluster_id": "cluster-46614-3", - "canonical_pr_number": 46616, - "duplicate_pr_numbers": [ - 46638 - ], - "target_issue_number": 46614, - "reason": "PRs in cluster-46614-3 are treated as duplicates because they converge on issue #46614 with closing_reference, shared_issue_target evidence." + "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." } ], "best_issue": { diff --git a/analysis/current/analysis-report-hybrid.llm-reviews.json b/analysis/current/analysis-report-hybrid.llm-reviews.json index 8545764d8c245cca9635643e7b4406916a59785e..abc8a3be0b81ff7da77775d2ebde954c9cc0753b 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": "20260615T120125Z", - "generated_at": "2026-06-15T12:12:43Z", + "snapshot_id": "20260615T180022Z", + "generated_at": "2026-06-15T18:12:03Z", "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-10000-472", + "cluster_id": "cluster-10000-473", "nodes": [ "issue:24643", "issue:30277", @@ -130,412 +130,412 @@ "issue:45237|issue:46015", "issue:43039|issue:44805" ], - "prepared_review_unit_hash": "d4e0d6b5933c8afd529deca374796fdea8dfd8ebb4741a4a6b85617042f5f462", + "prepared_review_unit_hash": "506a732315b3892b975a77fe12f29e851a066544837e3d2b6130d677329016d0", "component_budget": { - "node_count": 472, - "item_count": 472, - "soft_pair_count": 1732, - "serialized_chars": 898654, - "estimated_input_tokens": 224664, - "estimated_eval_tokens": 449584 + "node_count": 473, + "item_count": 473, + "soft_pair_count": 1738, + "serialized_chars": 901437, + "estimated_input_tokens": 225360, + "estimated_eval_tokens": 450976 }, "budget": { "node_count": 47, "item_count": 47, "soft_pair_count": 62, - "serialized_chars": 45726, - "estimated_input_tokens": 11432, - "estimated_eval_tokens": 23120 + "serialized_chars": 45730, + "estimated_input_tokens": 11433, + "estimated_eval_tokens": 23122 }, "overflow_policy": "truncate_then_skip", "trimmed": false, "aggressively_trimmed": false, "split": true, "analyst_result": { - "summary": "This cluster is mostly a grab bag of unrelated closed issues across many models and subsystems. Only a couple of soft pairs look like true duplicates: the Bert/DeepSpeed ZeRO-3 IndexError reports, and the torch 2.4.1 compatibility reports.", - "confidence": 0.66, - "canonical_issue_reason": "issue:43596 is the clearest representative of the real duplicate set: it states the concrete Bert + ZeRO-3 IndexError with the failing path more cleanly than 43638.", + "summary": "This cluster is mostly a mix of unrelated bug reports. Only a couple of near-exact duplicate pairs stand out: the Bert/DeepSpeed ZeRO-3 IndexError reports and the torch 2.4.1 / transformers 5.8.0 incompatibility reports. No pull requests are present.", + "confidence": 0.71, + "canonical_issue_reason": "issue:45800 is the clearest and most complete standalone report of the torch 2.4.1 / transformers 5.8.0 incompatibility, making it the best representative issue in the cluster.", "canonical_pr_reason": null, - "best_issue_reason": "issue:43596 is the best cluster representative because it sits in the strongest duplicate pair and names the exact failure mode succinctly.", + "best_issue_reason": "issue:45800 is the best global representative: it is concrete, version-specific, and cleaner than the follow-up report it duplicates.", "best_pr_reason": null, "soft_edge_verdicts": [ { "left": "issue:44322", "right": "issue:45464", "accept": false, - "reason": "Different Qwen3.5 failures: missing config field vs streaming chat/completions error." + "reason": "Different failures: missing config field in Qwen3.5 vs streaming chat/completions API error." }, { "left": "issue:45237", "right": "issue:45685", "accept": false, - "reason": "Unrelated model/runtime problems: AMD GPU failure vs MPS histogram op unsupported." + "reason": "Different models and hardware paths; GPT-OSS on AMD GPUs is unrelated to an MPS int histogram error." }, { "left": "issue:43716", "right": "issue:45237", "accept": false, - "reason": "Different subsystems and symptoms; image preprocessor dtype mismatch is not the AMD GPU issue." + "reason": "Different subsystems: multimodal dtype mismatch vs AMD GPU runtime failure." }, { "left": "issue:43828", "right": "issue:45237", "accept": false, - "reason": "Both are dtype/runtime failures, but on different models and code paths." + "reason": "Autocast dtype mismatch in Phi-tiny-MoE is not the same bug as GPT-OSS on AMD GPUs." }, { "left": "issue:45237", "right": "issue:46468", "accept": false, - "reason": "Different models and failure modes; not the same bug." + "reason": "Unrelated model/runtime issues; Nemotron example breakage is not the same as GPT-OSS AMD GPU failure." }, { "left": "issue:45237", "right": "issue:45507", "accept": false, - "reason": "Different MoE/model-specific errors; no shared concrete code path." + "reason": "Different concrete bugs: AMD GPU incompatibility vs invalid method call in GraniteMoEHybrid." }, { "left": "issue:30277", "right": "issue:35545", "accept": false, - "reason": "DeepSpeed collective mismatch and ONNX export error are unrelated." + "reason": "Collective mismatch under ZeRO-3 is unrelated to ModernBERT ONNX export." }, { "left": "issue:35545", "right": "issue:42915", "accept": false, - "reason": "ModernBERT ONNX export and Qwen3MoE FP8 failure are different bugs." + "reason": "ONNX export error and FineGrainedFP8 Qwen3Moe failure are different code paths." }, { "left": "issue:35545", "right": "issue:41720", "accept": false, - "reason": "Different model/runtime failures; no shared fix path." + "reason": "Export-time ONNX failure is not the same as Qwen3 auto device map cudaErrorAssert." }, { "left": "issue:45901", "right": "issue:46015", "accept": false, - "reason": "Table QA crash and outdated RAG docs are unrelated." + "reason": "Table QA crash vs outdated RAG examples are different issues." }, { "left": "issue:44509", "right": "issue:44512", "accept": false, - "reason": "Both are v5 doc cleanup, but they cover different removed docs topics and are not the same change." + "reason": "Both are docs cleanup, but they refer to different removed docs/topics and are not the same defect." }, { "left": "issue:35545", "right": "issue:36010", "accept": false, - "reason": "ONNX export error vs missing GenerationMixin import are unrelated." + "reason": "Model export error is unrelated to the GenerationMixin import breakage." }, { "left": "issue:44918", "right": "issue:46165", "accept": false, - "reason": "TRL SFT embedding unpacking and resume/trust_remote_code training are different problems." + "reason": "Different training problems: TRL SFT embedding unpacking vs resume-with-local-model remote code issue." }, { "left": "issue:44805", "right": "issue:45823", "accept": false, - "reason": "Different Gemma4 issues: mask shape mismatch vs device-map forward mismatch." + "reason": "Generic mask shape error and Gemma4 PLE device mismatch are different forward failures." }, { "left": "issue:43638", "right": "issue:44805", "accept": false, - "reason": "Both mention size/index errors, but the Bert ZeRO-3 init bug and the mask-shape bug are different." + "reason": "Both are tensor/index errors, but the Bert ZeRO-3 initialization bug is a different code path from the mask-index mismatch." }, { "left": "issue:43638", "right": "issue:46165", "accept": false, - "reason": "Unrelated training resume problem; no shared code path." + "reason": "DeepSpeed ZeRO-3/Bert index error is unrelated to resuming training with remote code." }, { "left": "issue:43596", "right": "issue:44805", "accept": false, - "reason": "Different error classes and model paths; not the same underlying bug." + "reason": "Same broad error class, but not the same concrete bug or model path." }, { "left": "issue:44805", "right": "issue:46294", "accept": false, - "reason": "Mask/index tensor mismatch is unrelated to video token/template mismatch." + "reason": "Mask indexing failure is unrelated to Qwen3.5 video chat-template token mismatch." }, { "left": "issue:44805", "right": "issue:45161", "accept": false, - "reason": "Different model/runtime paths; mask indexing vs TP MoE behavior." + "reason": "Different failures: mask shape mismatch vs GPT-OSS MoE tensor-parallel issue." }, { "left": "issue:46165", "right": "issue:46468", "accept": false, - "reason": "Resume/trust_remote_code and Nemotron example breakage are unrelated." + "reason": "Resume-training remote-code issue is unrelated to the broken Nemotron example." }, { "left": "issue:44560", "right": "issue:46165", "accept": false, - "reason": "Video StopIteration and training resume/trust_remote_code are unrelated." + "reason": "Video StopIteration bug is unrelated to resume-with-local-model remote code." }, { "left": "issue:43596", "right": "issue:46165", "accept": false, - "reason": "Different failures; no shared Bert ZeRO-3 index-error path." + "reason": "Bert/ZeRO-3 IndexError and remote-code resume failure are different issues." }, { "left": "issue:34634", "right": "issue:35545", "accept": false, - "reason": "BarkProcessor voice preset and ModernBERT ONNX export are unrelated." + "reason": "BarkProcessor voice preset issue is unrelated to ModernBERT ONNX export." }, { "left": "issue:30277", "right": "issue:43638", "accept": false, - "reason": "Both involve DeepSpeed, but different errors and model paths." + "reason": "Both involve distributed training, but the collective mismatch and Bert ZeRO-3 IndexError are not the same bug." }, { "left": "issue:43596", "right": "issue:43638", "accept": true, - "reason": "Same Bert/DeepSpeed ZeRO-3 IndexError; titles differ only in model/pretrained wording." + "reason": "Very likely duplicate reports of the same DeepSpeed ZeRO-3/Bert IndexError on empty tensors during initialization." }, { "left": "issue:43039", "right": "issue:45161", "accept": false, - "reason": "Liger Kernel cross-entropy call and GPT-OSS tensor-parallel failure are unrelated." + "reason": "Liger Kernel cross_entropy being called is not the same as GPT-OSS MoE tensor-parallel failure." }, { "left": "issue:43975", "right": "issue:44457", "accept": false, - "reason": "DeepSeek detokenization and LoRA merge/save-load mismatch are different bugs." + "reason": "Detokenization bug and LoRA merge/save-load mismatch are unrelated." }, { "left": "issue:44292", "right": "issue:44457", "accept": false, - "reason": "Qwen3 NVFP4 runtime error and LoRA output mismatch are unrelated." + "reason": "Qwen-3 NVFP4 runtime error and LoRA save/reload mismatch are different problems." }, { "left": "issue:44841", "right": "issue:46661", "accept": false, - "reason": "Voxtral processor failure and remote-code config_class AttributeError are different issues." + "reason": "Processor failure for Voxtral is unrelated to remote-code AutoModel config_class errors." }, { "left": "issue:34928", "right": "issue:44805", "accept": false, - "reason": "Activation checkpointing/FSDP recomputation mismatch is not the same as the mask-shape IndexError." + "reason": "FSDP activation-checkpointing recomputation mismatch is different from the mask-indexing error." }, { "left": "issue:24643", "right": "issue:30277", "accept": false, - "reason": "Different DeepSpeed failures with different symptoms and code paths." + "reason": "Both mention DeepSpeed, but the weight-dimension training error and collective mismatch are different bugs." }, { "left": "issue:45859", "right": "issue:45902", "accept": false, - "reason": "Same model component, but one is CPU-offload incompatibility and the other is meta/to_empty memory initialization; not clearly the same bug." + "reason": "Related model/component, but one is a CPU-offload forward incompatibility and the other is a meta/to_empty uninitialized-buffer bug; not the same defect." }, { "left": "issue:43638", "right": "issue:45161", "accept": false, - "reason": "Bert ZeRO-3 index error and GPT-OSS tensor-parallel issue are unrelated." + "reason": "DeepSpeed ZeRO-3/Bert index error is unrelated to GPT-OSS tensor-parallel behavior." }, { "left": "issue:34928", "right": "issue:36331", "accept": false, - "reason": "FSDP activation-checkpointing mismatch and custom trainer loss signature error are unrelated." + "reason": "Activation checkpointing/FSDP recomputation issue is unrelated to the custom trainer signature breakage." }, { "left": "issue:43596", "right": "issue:45161", "accept": false, - "reason": "Bert ZeRO-3 index error and GPT-OSS tensor-parallel issue are unrelated." + "reason": "Bert ZeRO-3 index error is unrelated to GPT-OSS MoE tensor-parallel failure." }, { "left": "issue:43716", "right": "issue:45507", "accept": false, - "reason": "Different model-specific errors; no shared concrete fix." + "reason": "Different model families and symptoms: multimodal dtype mismatch vs invalid method in GraniteMoEHybrid." }, { "left": "issue:45874", "right": "issue:46294", "accept": false, - "reason": "Gemma4 reverse-embedding lookup and Qwen3.5 video token mismatch are different features and bugs." + "reason": "Gemma4 reverse embedding lookup and Qwen3.5 video token mismatch are different model-specific bugs." }, { "left": "issue:44292", "right": "issue:45237", "accept": false, - "reason": "Different models and failure modes; no shared code path." + "reason": "Qwen NVFP4 runtime error and GPT-OSS AMD GPU failure are unrelated." }, { "left": "issue:39290", "right": "issue:45081", "accept": false, - "reason": "Gemma3TextConfig/vLLM issue and Mistral tokenizer regex patch crash are unrelated." + "reason": "Gemma3/vLLM missing config field and Mistral regex patch crash are different code paths." }, { "left": "issue:45823", "right": "issue:45874", "accept": false, - "reason": "Both are Gemma4, but one is device-map mismatch and the other is expensive reverse embedding lookup." + "reason": "Device mismatch in Gemma4 PLE is unrelated to expensive reverse embedding lookup in Gemma4 E2B/E4B." }, { "left": "issue:45800", "right": "issue:46515", "accept": true, - "reason": "Same torch 2.4.1 compatibility report; titles describe the same version-incompatibility bug." + "reason": "Clear duplicate torch 2.4.1 incompatibility reports; same version-level issue with near-identical scope." }, { "left": "issue:43425", "right": "issue:44292", "accept": false, - "reason": "Torch 2.10 incompatibility and Qwen3 NVFP4 runtime error are unrelated." + "reason": "Torch 2.10 compatibility issue and Qwen-3 NVFP4 runtime failure are different bugs." }, { "left": "issue:34928", "right": "issue:41720", "accept": false, - "reason": "FSDP checkpointing mismatch and Qwen3 auto device mapping assert are unrelated." + "reason": "FSDP checkpointing mismatch is unrelated to Qwen3 auto device mapping cudaErrorAssert." }, { "left": "issue:43828", "right": "issue:45161", "accept": false, - "reason": "Autocast dtype mismatch and GPT-OSS tensor-parallel failure are different bugs." + "reason": "Autocast dtype mismatch in Phi-tiny-MoE is not the same as GPT-OSS tensor-parallel failure." }, { "left": "issue:43541", "right": "issue:43716", "accept": false, - "reason": "Mixtral grouped_mm tracing error and image-preprocessor dtype mismatch are unrelated." + "reason": "Mixtral grouped_mm tracing failure and Mistral-3 image-preprocessor dtype mismatch are different issues." }, { "left": "issue:43828", "right": "issue:44292", "accept": false, - "reason": "Different models and failure modes; not the same underlying bug." + "reason": "Different model families and failure modes; not the same bug." }, { "left": "issue:43716", "right": "issue:46468", "accept": false, - "reason": "Dtype mismatch in Mistral-3 vs broken Nemotron example are unrelated." + "reason": "Mistral-3 image dtype mismatch is unrelated to the Nemotron example bug." }, { "left": "issue:43716", "right": "issue:45685", "accept": false, - "reason": "Different model/runtime problems; no shared fix." + "reason": "Different models and hardware paths; not the same underlying defect." }, { "left": "issue:44292", "right": "issue:44918", "accept": false, - "reason": "Qwen3 NVFP4 runtime error and Qwen3.5 embedding unpacking issue are different." + "reason": "Qwen NVFP4 runtime error and TRL embedding unpacking failure are unrelated." }, { "left": "issue:43716", "right": "issue:44292", "accept": false, - "reason": "Different models and error mechanisms." + "reason": "Different model-specific runtime errors with no concrete overlap." }, { "left": "issue:45507", "right": "issue:45685", "accept": false, - "reason": "GraniteMoE invalid method and MPS histogram op unsupported are unrelated." + "reason": "GraniteMoEHybrid invalid method and MPS histogram_int issue are unrelated." }, { "left": "issue:44918", "right": "issue:45161", "accept": false, - "reason": "TRL SFT embedding unpacking and GPT-OSS TP issue are unrelated." + "reason": "TRL SFT embedding unpacking issue is unrelated to GPT-OSS tensor-parallel failure." }, { "left": "issue:43039", "right": "issue:44918", "accept": false, - "reason": "Liger Kernel cross-entropy call and TRL embedding unpacking are unrelated." + "reason": "Liger Kernel cross_entropy call and Qwen3.5 embedding unpacking are different bugs." }, { "left": "issue:41762", "right": "issue:43638", "accept": false, - "reason": "Gemma3 ZeRO-3 index error and Bert ZeRO-3 index error are similar-looking but distinct model paths." + "reason": "Gemma3 ZeRO-3 load failure and Bert ZeRO-3 index error are similar only superficially; not the same concrete issue." }, { "left": "issue:43827", "right": "issue:44512", "accept": false, - "reason": "Both are docs regressions, but they concern different removed v5 references." + "reason": "Both are docs regressions, but they cover different removed items and would not be one duplicate issue." }, { "left": "issue:43828", "right": "issue:45685", "accept": false, - "reason": "Autocast dtype mismatch and MPS histogram op unsupported are unrelated." + "reason": "Autocast dtype mismatch and MPS int histogram error are unrelated." }, { "left": "issue:44918", "right": "issue:45685", "accept": false, - "reason": "Qwen3.5 embedding unpacking and MPS histogram op unsupported are unrelated." + "reason": "TRL trainer embedding issue is unrelated to the MPS histogram bug." }, { "left": "issue:43039", "right": "issue:43828", "accept": false, - "reason": "Different model/runtime bugs with no shared concrete code path." + "reason": "Different model/runtime problems; no shared concrete code-path." }, { "left": "issue:30277", "right": "issue:36331", "accept": false, - "reason": "DeepSpeed collective mismatch and trainer loss signature mismatch are unrelated." + "reason": "DeepSpeed collective mismatch and custom trainer argument error are unrelated." }, { "left": "issue:34928", "right": "issue:36010", "accept": false, - "reason": "FSDP activation-checkpointing issue and GenerationMixin import failure are unrelated." + "reason": "Activation checkpointing/FSDP issue is unrelated to the GenerationMixin import error." }, { "left": "issue:45237", "right": "issue:46015", "accept": false, - "reason": "AMD GPU runtime failure and outdated RAG examples are unrelated." + "reason": "GPT-OSS AMD GPU failure and RAG docs example issue are unrelated." }, { "left": "issue:43039", "right": "issue:44805", "accept": false, - "reason": "Liger Kernel cross-entropy issue and mask-shape IndexError are unrelated." + "reason": "Liger Kernel cross_entropy call and mask shape mismatch are different problems." } ] }, "evaluator_result": { "accept": true, - "feedback": "The summary is grounded in the packet and stays conservative overall. The accepted soft pairs are well-supported by the issue titles and the strongest similarity signals, and the rejected pairs are phrased cautiously without overclaiming. The representative-issue choice is also defensible based on the close Bert/ZeRO-3 duplicate pair." + "feedback": "Grounded overall. The summary matches the packet: the cluster is mostly unrelated issues, with two conservative duplicate pairs accepted (43596/43638 and 45800/46515), and there are no PRs. The representative/canonical choice of issue:45800 is well supported by the titles and similarity signal." }, "evaluator_used": true, "retried": false, @@ -554,7 +554,7 @@ "source": "llm", "cache_hit": false, "model": "gpt-5.4-mini?service_tier=flex", - "cluster_id": "cluster-10000-472", + "cluster_id": "cluster-10000-473", "nodes": [ "issue:36010", "issue:39290", @@ -611,13 +611,13 @@ "issue:43296|issue:46304", "issue:41762|issue:43596", "issue:39290|issue:41720", - "issue:44589|issue:46304", "issue:43901|issue:44512", + "issue:44589|issue:46304", "issue:45237|issue:45901", "issue:43039|issue:44560", "issue:45823|issue:46294", - "issue:44841|issue:45823", "issue:43975|issue:46015", + "issue:44841|issue:45823", "issue:44863|issue:46015", "issue:45020|issue:46304", "issue:45812|issue:45920", @@ -629,8 +629,8 @@ "issue:43792|issue:45901", "issue:43854|issue:43866", "issue:43425|issue:45070", - "issue:43596|issue:45800", "issue:43975|issue:45901", + "issue:43596|issue:45800", "issue:43792|issue:45070", "issue:45362|issue:45464", "issue:44292|issue:45901", @@ -648,277 +648,283 @@ "issue:45800|issue:45823", "issue:44387|issue:45005" ], - "prepared_review_unit_hash": "694e892dcb305b3764d5b95f0aa14c19ac8430c8ef62ef8d5282998c25a041ef", + "prepared_review_unit_hash": "217bb804c7e222b525afc7cbc17e697297d5b8a90f3b39d2614735db2149c3f7", "component_budget": { - "node_count": 472, - "item_count": 472, - "soft_pair_count": 1732, - "serialized_chars": 898654, - "estimated_input_tokens": 224664, - "estimated_eval_tokens": 449584 + "node_count": 473, + "item_count": 473, + "soft_pair_count": 1738, + "serialized_chars": 901437, + "estimated_input_tokens": 225360, + "estimated_eval_tokens": 450976 }, "budget": { "node_count": 48, "item_count": 48, "soft_pair_count": 41, - "serialized_chars": 39789, - "estimated_input_tokens": 9948, - "estimated_eval_tokens": 20152 + "serialized_chars": 39787, + "estimated_input_tokens": 9947, + "estimated_eval_tokens": 20150 }, "overflow_policy": "truncate_then_skip", "trimmed": false, "aggressively_trimmed": false, "split": true, "analyst_result": { - "summary": "Conservative reading: this cluster is mostly heterogeneous. The only clearly supported duplicate pair is 45812 \u2194 45920, both describing the same v5 AutoTokenizer wrong-token-ID regression. The 45399 \u2194 46661 pair is only a loose similarity and should not be treated as a duplicate family; their titles point to different code paths and bugs.", - "confidence": 0.94, - "canonical_issue_reason": "Issue 45920 is the best canonical issue because it is the broadest live report of the v5 AutoTokenizer wrong-token-ID regression, while 45812 is a narrower Granite-specific instance of the same bug.", + "summary": "Mostly a grab bag of distinct Transformers issues; only the tokenizer regression reports look like the same underlying bug. A few loading/ZeRO-3 and sliding-window items are superficially similar, but not safe to dedupe strictly.", + "confidence": 0.79, + "canonical_issue_reason": "issue:45920 is the best umbrella: it is open, broad, and explicitly covers the same AutoTokenizer wrong-ID regression across multiple model families.", "canonical_pr_reason": null, - "best_issue_reason": "45920 is the best representative issue for the cluster because it is the broad umbrella report for the one confirmed duplicate family and remains open.", + "best_issue_reason": "issue:45920 has the clearest multi-model repro set and ongoing discussion, making it the strongest representative issue for the cluster.", "best_pr_reason": null, "soft_edge_verdicts": [ { "left": "issue:43866", "right": "issue:46402", "accept": false, - "reason": "Both are loading failures, but for different models and different likely root causes." + "reason": "Different models and failure modes: corrupted Ovis2 checkpoint vs missing text-encoder weight loading in Sam3Model." }, { "left": "issue:45800", "right": "issue:46304", "accept": false, - "reason": "Torch/version compatibility and USE_DISTRIBUTED=0 support are different regressions." + "reason": "Both are version-compatibility complaints, but one is torch 2.4.1 vs transformers 5.8.0 and the other is a specific USE_DISTRIBUTED=0 breakage in v5.9.0." }, { "left": "issue:43296", "right": "issue:46304", "accept": false, - "reason": "PaddleOCR-VL load failure and distributed-off support breakage are unrelated." + "reason": "Unrelated regressions: PaddleOCR-VL loading in vLLM vs PyTorch distributed support breakage." }, { "left": "issue:41762", "right": "issue:43596", "accept": false, - "reason": "Both are index errors under ZeRO-3, but on different models and flows." + "reason": "Both hit ZeRO-3 index errors, but they involve different models and different loading/training paths; too broad to treat as the same bug." }, { "left": "issue:39290", "right": "issue:41720", "accept": false, - "reason": "Gemma3 config attribute missing and Qwen3 auto device-map CUDA assert are different bugs." + "reason": "Gemma3TextConfig missing a field for vLLM is unrelated to Qwen3 auto device mapping causing a CUDA assert." }, { - "left": "issue:44589", - "right": "issue:46304", + "left": "issue:43901", + "right": "issue:44512", "accept": false, - "reason": "Float8 storage lookup failure is unrelated to USE_DISTRIBUTED=0 support." + "reason": "Both are docs issues, but they cover different doc text and different removed/changed behaviors." }, { - "left": "issue:43901", - "right": "issue:44512", + "left": "issue:44589", + "right": "issue:46304", "accept": false, - "reason": "Both are docs-ish, but they concern different APIs and regressions." + "reason": "Float8 storage lookup failure is unrelated to distributed support regressions." }, { "left": "issue:45237", "right": "issue:45901", "accept": false, - "reason": "AMD GPU GPT-OSS support and table QA crashes are unrelated." + "reason": "AMD GPU support for GPT-OSS is unrelated to table-question-answering crashes." }, { "left": "issue:43039", "right": "issue:44560", "accept": false, - "reason": "Liger kernel cross-entropy behavior and Qwen3-vl video StopIteration are unrelated code paths." + "reason": "Liger Kernel cross-entropy routing is unrelated to a Qwen3-VL StopIteration video issue." }, { "left": "issue:45823", "right": "issue:46294", "accept": false, - "reason": "Both are multimodal, but device mismatch and video token/feature mismatch are different bugs." - }, - { - "left": "issue:44841", - "right": "issue:45823", - "accept": false, - "reason": "Voxtral processor failure and Gemma4 device-mapping mismatch are different issues." + "reason": "Both touch multimodal/Qwen3.5 areas, but one is a device-map mismatch and the other is a chat-template video token mismatch." }, { "left": "issue:43975", "right": "issue:46015", "accept": false, - "reason": "Detokenization regression and outdated RAG examples are unrelated." + "reason": "Model detokenization bug vs outdated RAG documentation are not the same underlying change." }, { - "left": "issue:44863", - "right": "issue:46015", + "left": "issue:44841", + "right": "issue:45823", "accept": false, - "reason": "NemotronH loading and RAG docs drift are unrelated." + "reason": "Processor failure for Voxtral-Mini and Gemma4 PLE device mismatch are different code paths and symptoms." }, { "left": "issue:45020", "right": "issue:46304", "accept": false, - "reason": "Remote_code breakage and distributed-off support are separate problems." + "reason": "Remote-code model breakage is not the same issue as PyTorch distributed support being disabled." }, { "left": "issue:45812", "right": "issue:45920", "accept": true, - "reason": "Same underlying v5 AutoTokenizer wrong-token-ID regression; 45920 is the broader umbrella and 45812 is a specific Granite example." + "reason": "Same tokenizer regression: both report wrong token IDs across multiple models after the v4\u2192v5 transition, with 45920 being the broader umbrella." }, { "left": "issue:43866", "right": "issue:44863", "accept": false, - "reason": "Different model-loading failures with no evidence of the same root cause." + "reason": "Corrupted Ovis2 checkpoint and NemotronH checkpoint-loading failure are different model-specific problems." }, { "left": "issue:43872", "right": "issue:46304", "accept": false, - "reason": "bitsandbytes constructor incompatibility is unrelated to distributed build support." + "reason": "bitsandbytes Int8Params constructor incompatibility is unrelated to distributed build support." }, { "left": "issue:44292", "right": "issue:46015", "accept": false, - "reason": "Qwen-3-8B-NVFP4 runtime failure and RAG documentation issues are unrelated." + "reason": "Qwen-3 NVFP4 runtime error is unrelated to RAG example docs." }, { "left": "issue:39290", "right": "issue:43531", "accept": false, - "reason": "They both mention sliding_window, but the model/config bugs are different." + "reason": "Both mention sliding_window, but they affect different models/configs and are not clearly the same defect." }, { "left": "issue:41093", "right": "issue:43638", "accept": false, - "reason": "Both are index errors, but one is a mask-shape mismatch and the other is a ZeRO-3 Bert init failure." + "reason": "Both are IndexErrors, but the reported shapes and the zero-sized-tensor symptom point to different bugs." }, { "left": "issue:43792", "right": "issue:45901", "accept": false, - "reason": "Whisper load failure and table-question-answering crashes are unrelated." + "reason": "Whisper-large-v2 loading failure is unrelated to table-question-answering crashes." }, { "left": "issue:43854", "right": "issue:43866", "accept": false, - "reason": "GLM-4.7-Flash test-loading issue and Ovis2 checkpoint corruption are different failures." + "reason": "GLM-4.7-Flash load failure and Ovis2 checkpoint corruption are separate model-specific issues." }, { "left": "issue:43425", "right": "issue:45070", "accept": false, - "reason": "Torch 2.10 incompatibility and PretrainedConfig pydantic breakage are unrelated version regressions." + "reason": "Torch 2.10 compatibility and a pydantic PretrainedConfig field regression are unrelated." }, { - "left": "issue:43596", - "right": "issue:45800", + "left": "issue:43975", + "right": "issue:45901", "accept": false, - "reason": "ZeRO-3 Bert index error and torch/transformers compatibility are not the same bug." + "reason": "Detokenization regression and table QA crashes do not share a concrete bug or fix path." }, { - "left": "issue:43975", - "right": "issue:45901", + "left": "issue:43596", + "right": "issue:45800", "accept": false, - "reason": "Different subsystems: detokenization versus table QA pipeline crash." + "reason": "ZeRO-3 BertModel loading error is unrelated to a torch/transformers version incompatibility report." }, { "left": "issue:43792", "right": "issue:45070", "accept": false, - "reason": "Whisper loading and config-field breakage are unrelated." + "reason": "Whisper model loading and pydantic config-field breakage are unrelated." }, { "left": "issue:45362", "right": "issue:45464", "accept": false, - "reason": "Both are Qwen3.5-related, but one is a crash and the other is a streaming inference API failure." + "reason": "Both involve Qwen3.5 chat behavior, but one is a crash in transformers chat and the other is a streaming chat/completions API failure; not the same bug." + }, + { + "left": "issue:44292", + "right": "issue:45901", + "accept": false, + "reason": "Qwen-3 NVFP4 runtime error does not match table-question-answering task crashes." + }, + { + "left": "issue:39290", + "right": "issue:45823", + "accept": false, + "reason": "Gemma3/vLLM config error and Gemma4 device-map mismatch are different issues." }, { "left": "issue:45464", "right": "issue:45685", "accept": false, - "reason": "Streaming chat/completions failure and MPS histogram op unsupported for int are unrelated." + "reason": "Streaming inference failure for Qwen3.5-0.8B is unrelated to the MPS histogram int-type error." }, { "left": "issue:43541", "right": "issue:45685", "accept": false, - "reason": "Grouped_mm tracing failure and MPS histogram type error are different bugs." + "reason": "Different runtime backends and different ops: torch dynamo grouped_mm tracing vs MPS histogram implementation." }, { "left": "issue:43541", "right": "issue:43828", "accept": false, - "reason": "Different models and different failures: tracing error versus autocast dtype mismatch." + "reason": "Grouped_mm tracing failure in Mixtral is unrelated to an autocast dtype mismatch in Phi-tiny-MoE." }, { "left": "issue:36010", "right": "issue:39290", "accept": false, - "reason": "ImportError for GenerationMixin and Gemma3 missing sliding_window_pattern are unrelated." + "reason": "ImportError for GenerationMixin is unrelated to the Gemma3TextConfig sliding-window regression." }, { "left": "issue:44841", "right": "issue:45084", "accept": false, - "reason": "Processor failure and non-template-node compile error are unrelated." + "reason": "Voxtral processor failure and a template-compilation TypeError are distinct problems." }, { "left": "issue:43404", "right": "issue:43866", "accept": false, - "reason": "Weight tying bug in Mistral3 and corrupted Ovis2 checkpoint are unrelated." + "reason": "Mistral3 tied-weights bug is unrelated to a corrupted Ovis2 checkpoint." }, { "left": "issue:41093", "right": "issue:41762", "accept": false, - "reason": "Both are indexing errors, but one is a mask-shape mismatch and the other is a ZeRO-3 loading failure." + "reason": "Different IndexError signatures and different model/setup contexts; not a reliable duplicate." }, { "left": "issue:43296", "right": "issue:43366", "accept": false, - "reason": "PaddleOCR-VL/vLLM load failure and GGUF gpt-oss support are unrelated." + "reason": "PaddleOCR-VL loading in vLLM and GGUF gpt-oss support are different model-support issues." }, { "left": "issue:45399", "right": "issue:46661", "accept": false, - "reason": "Flash-attn fallback gating and remote-code config_class failures are different code paths and different bugs." + "reason": "flash-attn2 fallback gating and remote-code config_class errors are separate from each other." }, { "left": "issue:43572", "right": "issue:45070", "accept": false, - "reason": "Missing pad_token_idx in StableLmConfig and PretrainedConfig field breakage are separate config regressions." + "reason": "StableLmConfig pad_token_idx regression is unrelated to the pydantic PretrainedConfig field breakage." }, { "left": "issue:45800", "right": "issue:45823", "accept": false, - "reason": "Torch/transformers compatibility and Gemma4 device-map mismatch are unrelated." + "reason": "Torch/transformers version incompatibility and Gemma4 device-map forward mismatch are different bugs." }, { "left": "issue:44387", "right": "issue:45005", "accept": false, - "reason": "Int4 OOM and tied-weights translation-model issues do not share the same bug." + "reason": "Int4 OOM from higher CUDA reserved memory is not the same as tied-weights issues in translation models." } ] }, "evaluator_result": { - "accept": false, - "feedback": "The canonical issue choice for 45920/45812 is well grounded, but the summary overstates the evidence by claiming a second true duplicate family around remote_code loading. In the packet, the only clearly supported duplicate pair is 45812 \u2194 45920; the 45399 \u2194 46661 pair is only soft-similarity and the titles point to different bugs. Because the cluster summary is not fully conservative, I can\u2019t accept it." + "accept": true, + "feedback": "The summary is grounded in the packet and stays conservative overall. The only accepted soft-edge pair, issue:45812 vs issue:45920, is reasonably supported by the titles as the same AutoTokenizer wrong-token-ID regression across different model families. The remaining pair rejections are appropriately cautious given the evidence available." }, "evaluator_used": true, - "retried": true, + "retried": false, "accepted_nontrivial_soft_edge": true, "error_kind": null, "error_message": null @@ -934,7 +940,7 @@ "source": "llm", "cache_hit": false, "model": "gpt-5.4-mini?service_tier=flex", - "cluster_id": "cluster-10000-472", + "cluster_id": "cluster-10000-473", "nodes": [ "issue:39186", "issue:40990", @@ -1017,20 +1023,20 @@ "issue:43975|issue:44292", "issue:44292|issue:45507" ], - "prepared_review_unit_hash": "2321f5619cd8bc71602f0f49633e4fa19f3e5afc307aeb9f5cadb9eeab92e2b0", + "prepared_review_unit_hash": "51b4819a636ad67399c8a1e195e1c425f6a90aa09eef5918c9101f5784fb7e69", "component_budget": { - "node_count": 472, - "item_count": 472, - "soft_pair_count": 1732, - "serialized_chars": 898654, - "estimated_input_tokens": 224664, - "estimated_eval_tokens": 449584 + "node_count": 473, + "item_count": 473, + "soft_pair_count": 1738, + "serialized_chars": 901437, + "estimated_input_tokens": 225360, + "estimated_eval_tokens": 450976 }, "budget": { "node_count": 47, "item_count": 47, "soft_pair_count": 31, - "serialized_chars": 36258, + "serialized_chars": 36257, "estimated_input_tokens": 9065, "estimated_eval_tokens": 18386 }, @@ -1039,168 +1045,162 @@ "aggressively_trimmed": false, "split": true, "analyst_result": { - "summary": "This set is mostly a grab-bag of unrelated Transformers issues. A few pairs share broad themes (validation, tokenizer regressions, remote-code/custom-module loading), but none are clearly the same underlying bug with enough specificity to merge as duplicates.", - "confidence": 0.47, - "canonical_issue_reason": "issue:45183 is the strongest umbrella-style issue in the set: it targets a concrete validation-path bug, is phrased broadly enough to cover related model-specific reports, and is more reusable as a triage anchor than the narrow, model-specific issues.", + "summary": "This cluster is mostly heterogeneous: many unrelated model-specific regressions, tokenizer issues, config/loading bugs, and docs/runtime complaints. Only a couple of soft links look like true duplicates, notably the Qwen video-input regression and the input-validation error-message bug.", + "confidence": 0.88, + "canonical_issue_reason": "No single issue cleanly represents the whole cluster because the items span multiple distinct bugs; if an umbrella representative is needed, issue 45020 is the broadest v5/remote-code regression, but it is still not a true canonical duplicate for most items.", "canonical_pr_reason": null, - "best_issue_reason": "issue:45183 best fits global issue triage because it describes a clear framework-level bug affecting multiple models, making it the most likely duplicate target for related validation complaints.", + "best_issue_reason": "Issue 45020 is the closest high-level umbrella for the cluster\u2019s repeated v5/remote-code breakage theme, but the cluster remains too mixed for a strong canonical choice.", "best_pr_reason": null, "soft_edge_verdicts": [ { "left": "issue:44805", "right": "issue:44918", "accept": false, - "reason": "Both mention indexing/shape failures, but one is a mask-vs-tensor mismatch and the other is about Qwen3.5 embedding unpacking with TRL; different code paths and symptoms." + "reason": "Both are indexing/shape-related, but they describe different failures in different code paths and models; not the same underlying bug." }, { "left": "issue:43541", "right": "issue:45161", "accept": false, - "reason": "Both involve MoE/GPT-OSS/Mixtral-style models, but one is a torch-dynamo grouped_mm tracing crash and the other is a tensor-parallel-only failure; not the same bug." + "reason": "Different MoE/TP failures with different triggers; too broad a similarity to treat as one bug." }, { "left": "issue:43828", "right": "issue:45464", "accept": false, - "reason": "dtype mismatch under autocast is unrelated to a streaming chat/completions failure on Qwen3.5." + "reason": "Different models and symptoms: autocast dtype mismatch vs streaming chat/completions failure." }, { "left": "issue:43366", "right": "issue:45464", "accept": false, - "reason": "GGUF architecture support for gpt-oss is a model-format support request, not the same issue as Qwen3.5 streaming API failure." + "reason": "GGUF gpt-oss support is unrelated to the Qwen3.5 streaming inference failure." }, { "left": "issue:44291", "right": "issue:45070", "accept": false, - "reason": "One is an init_empty_weights/_is_hf_initialized constructor breakage; the other is a pydantic PretrainedConfig-field regression. Different failure modes." + "reason": "Both involve v5 breaking changes, but one is init_empty_weights arg handling and the other is PretrainedConfig/pydantic behavior; not the same bug." }, { "left": "issue:43531", "right": "issue:45070", "accept": false, - "reason": "Qwen3-MoE sliding_window behavior is unrelated to a pydantic model-field regression." + "reason": "Sliding-window Qwen3-MoE behavior is unrelated to the PretrainedConfig pydantic regression." }, { "left": "issue:43541", "right": "issue:44589", "accept": false, - "reason": "Grouped_mm tracing on Mixtral and missing Float8 storage are different runtime errors with different causes." + "reason": "Grouped_mm tracing failure and missing Float8 storage are distinct runtime issues." }, { "left": "issue:43716", "right": "issue:46294", "accept": false, - "reason": "Both are multimodal, but one is an image-preprocessor dtype mismatch and the other is a video-token/template mismatch." + "reason": "Both are multimodal dtype/input issues, but one is image preprocessor dtype mismatch and the other is video token/feature mismatch; not clearly the same bug." }, { "left": "issue:45020", "right": "issue:45800", "accept": false, - "reason": "Remote_code model breakage and torch/transformers incompatibility are too broad and not the same concrete bug." + "reason": "General remote_code breakage and torch/transformers version incompatibility are related only at a high level, not the same concrete defect." }, { "left": "issue:41093", "right": "issue:43596", "accept": false, - "reason": "Both are IndexErrors, but one is a mask-length mismatch and the other is a DeepSpeed ZeRO-3 zero-sized-index failure." + "reason": "Both are IndexErrors, but they arise from different tensor/mask paths and different models." }, { "left": "issue:44479", "right": "issue:46294", - "accept": false, - "reason": "A video-input regression across several Qwen VL models is not the same as a video-feature/video-token mismatch in a specific Qwen3.5 model." + "accept": true, + "reason": "These look like the same Qwen video-input regression: both describe video feature/token mismatch failures in the Qwen3.5/Qwen video pipeline." }, { "left": "issue:43054", "right": "issue:43994", "accept": false, - "reason": "Siglip2 text-embedding degradation and a Siglip2 pipeline/AutoModel nonsensical-output report are related at the model family level, but not clearly the same defect." + "reason": "Both involve Siglip2, but one is a quality/regression report and the other is an AutoModel/pipeline output bug; not enough evidence they are the same defect." }, { "left": "issue:45759", "right": "issue:45902", "accept": false, - "reason": "One is about AutoModel.from_config on composite multimodal configs; the other is an uninitialized-memory bug in Qwen3_5MoeTextRotaryEmbedding." + "reason": "Composite multimodal config handling and Qwen3_5Moe rotary embedding initialization are different code paths." }, { "left": "issue:46640", "right": "issue:46661", "accept": false, - "reason": "Flash Attention 2 issue and remote-code model_class/config_class AttributeError are unrelated." + "reason": "Flash Attention 2 availability/fallback behavior is unrelated to remote-code model_class/config_class loading." }, { "left": "issue:39186", "right": "issue:44918", "accept": false, - "reason": "FSDP weight-rank error and Qwen3.5 embedding unpacking are different subsystems and different bugs." + "reason": "FSDP 2-D weight error and Qwen3.5 embedding unpacking in TRL are distinct failures." }, { "left": "issue:43716", "right": "issue:44918", "accept": false, - "reason": "Image-preprocessor dtype mismatch is not the same as TRL SFT embedding unpacking failure." + "reason": "Image-preprocessor dtype mismatch and TRL input-embedding unpacking are different bugs." }, { "left": "issue:40990", "right": "issue:44960", "accept": false, - "reason": "High perplexity on gpt-oss and a GLM5 issue are not the same underlying bug." + "reason": "These are unrelated model-quality/model-family issues." }, { "left": "issue:45632", "right": "issue:45698", "accept": false, - "reason": "Both concern custom-code loading/caching, but one is a leaf-directory cache collision and the other is loading the wrong module after save_pretrained; similar area, different bug." + "reason": "Both touch remote-code loading, but one is a leaf-dir cache collision and the other is wrong custom module selection after save_pretrained; not the same concrete bug." }, { "left": "issue:43645", "right": "issue:45399", "accept": false, - "reason": "Jupyter custom-model initialization and flash-attn2 fallback gating are unrelated." - }, - { - "left": "issue:43404", - "right": "issue:45005", - "accept": false, - "reason": "Mistral3 tied lm_head weights and translation-model tied-weights regressions are about different model families and different tied-weight behavior." + "reason": "Jupyter custom-model initialization and Flash Attention fallback gating are unrelated." }, { "left": "issue:42915", "right": "issue:45399", "accept": false, - "reason": "Qwen3Moe FP8 config failure and flash-attn2 fallback blocking are unrelated." + "reason": "Qwen3-MoE FP8 failure and Flash Attention fallback checks are different issues." }, { "left": "issue:44530", "right": "issue:45005", "accept": false, - "reason": "PagedAttentionCache group-type crash and tied-weights issues do not share the same code-path problem." + "reason": "PagedAttentionCache linear_attention crash and tied-weights translation-model issues are unrelated." }, { "left": "issue:43873", "right": "issue:45005", "accept": false, - "reason": "Offloading with quantization and tied weights in v5 are distinct concerns." + "reason": "Quantization/offloading behavior and tied-weights translation-model issues do not match." }, { "left": "issue:44360", "right": "issue:44512", "accept": false, - "reason": "A DSA indexer discussion and a docs typo about cache strategies are not the same change or bug." + "reason": "A DSA indexer ReLU discussion and a removed-docs-command typo are unrelated." }, { "left": "issue:44863", "right": "issue:45005", "accept": false, - "reason": "NemotronH checkpoint loading failure is not the same as tied-weights regressions." + "reason": "NemotronH checkpoint loading and translation tied-weights are different problems." }, { "left": "issue:45183", "right": "issue:46463", - "accept": false, - "reason": "Both involve input-validation messaging, but the T5 decoder-specific report is not clearly the same underlying bug as the broader XOR validation issue." + "accept": true, + "reason": "These are the same validation bug family: incorrect input_ids/inputs_embeds XOR/error-message handling, with 46463 as a T5-specific manifestation." }, { "left": "issue:44492", @@ -1212,32 +1212,35 @@ "left": "issue:44918", "right": "issue:45507", "accept": false, - "reason": "Qwen3.5 embedding unpacking with TRL is unrelated to GraniteMoEHybrid calling an invalid method." + "reason": "TRL embedding unpacking and GraniteMoEHybrid invalid method calls are different code paths." }, { "left": "issue:46489", "right": "issue:46491", "accept": false, - "reason": "Both are tokenizer regressions, but they affect different model families and describe different round-trip/output failures." + "reason": "Both are tokenizer regressions, but for different models and different symptoms; not safe to merge as one bug." }, { "left": "issue:43975", "right": "issue:44292", "accept": false, - "reason": "DeepSeek-Coder detokenization and Qwen-3-8B-NVFP4 runtime failure are unrelated." + "reason": "Tokenizer detokenization regression and init_empty_weights/_is_hf_initialized loading error are unrelated." }, { "left": "issue:44292", "right": "issue:45507", "accept": false, - "reason": "NVFP4 runtime failure and GraniteMoEHybrid invalid-method calls are different bugs." + "reason": "Different model-loading/runtime failures with no clear shared underlying defect." } ] }, - "evaluator_result": null, - "evaluator_used": false, + "evaluator_result": { + "accept": true, + "feedback": "The summary is grounded in the packet and stays appropriately conservative about the cluster\u2019s heterogeneity. The two accepted soft links are defensible as same-bug families, and the rejected pairs are phrased cautiously without overclaiming duplicates." + }, + "evaluator_used": true, "retried": false, - "accepted_nontrivial_soft_edge": false, + "accepted_nontrivial_soft_edge": true, "error_kind": null, "error_message": null }, @@ -1252,7 +1255,7 @@ "source": "llm", "cache_hit": false, "model": "gpt-5.4-mini?service_tier=flex", - "cluster_id": "cluster-10000-472", + "cluster_id": "cluster-10000-473", "nodes": [ "issue:24643", "issue:28530", @@ -1311,8 +1314,8 @@ "issue:43828|issue:44928", "issue:43872|issue:44589", "issue:45920|issue:46489", - "issue:43054|issue:45701", "issue:43866|issue:44403", + "issue:43054|issue:45701", "issue:43950|issue:45902", "issue:38175|issue:43054", "issue:41553|issue:43883", @@ -1336,23 +1339,23 @@ "issue:45020|issue:45823", "issue:44661|issue:45020", "issue:45081|issue:45399", - "issue:43828|issue:45823", - "issue:42491|issue:44368" + "issue:42491|issue:44368", + "issue:43828|issue:45823" ], - "prepared_review_unit_hash": "9eebf8ad35bf2e7c30365aeb232de5ffe9679c4028f0c952739a6c49d6ca7978", + "prepared_review_unit_hash": "d7e4b42f388b7e6377b9a13e610a99b8af9bbae7e32bd31aff9ad06d23e67087", "component_budget": { - "node_count": 472, - "item_count": 472, - "soft_pair_count": 1732, - "serialized_chars": 898654, - "estimated_input_tokens": 224664, - "estimated_eval_tokens": 449584 + "node_count": 473, + "item_count": 473, + "soft_pair_count": 1738, + "serialized_chars": 901437, + "estimated_input_tokens": 225360, + "estimated_eval_tokens": 450976 }, "budget": { "node_count": 48, "item_count": 48, "soft_pair_count": 34, - "serialized_chars": 38113, + "serialized_chars": 38114, "estimated_input_tokens": 9529, "estimated_eval_tokens": 19314 }, @@ -1361,222 +1364,222 @@ "aggressively_trimmed": false, "split": true, "analyst_result": { - "summary": "This cluster is mostly a mix of unrelated Transformers v5 regressions across loading, tokenization, precision, and model-specific behavior. The only clear duplicate-like pair from titles alone is the non-persistent-buffer corruption report: issue:43950 and issue:44534 are very similar and likely refer to the same regression.", - "confidence": 0.67, - "canonical_issue_reason": "issue:43950 is the best canonical issue because it gives the most concrete description of the non-persistent-buffer corruption regression, and issue:44534 is a very close textual match. Based on the available metadata, this is a likely duplicate pair rather than something that can be proven exactly identical.", + "summary": "Mostly unrelated loading/training regressions, with one clear non-persistent-buffer corruption cluster and one plausible SigLIP2 implementation-discrepancy pair. The 43950/45902 edge should be rejected: 45902 looks like a related materialization/uninitialized-memory symptom, but not the same regression as the broader from_pretrained() buffer corruption bug.", + "confidence": 0.77, + "canonical_issue_reason": "issue:43950 is the broadest and clearest representative of the transformers 5.x non-persistent-buffer corruption regression; it captures the shared root symptom better than the narrower follow-up reports.", "canonical_pr_reason": null, - "best_issue_reason": "issue:43950 is the strongest representative issue overall: it names a specific v5 regression, has clear problem framing, and is the closest match to another report in the cluster.", + "best_issue_reason": "issue:43950 best represents the shared buffer-corruption defect and subsumes the closely matching junk-filled buffer reports without overextending to adjacent materialization bugs.", "best_pr_reason": null, "soft_edge_verdicts": [ { "left": "issue:43950", "right": "issue:44534", "accept": true, - "reason": "The titles are nearly the same and point to the same concrete regression: v5 corrupting/filling non-persistent buffers. This is a likely duplicate pair." + "reason": "Same underlying transformers 5.x non-persistent-buffer corruption bug." }, { "left": "issue:39692", "right": "issue:43054", "accept": false, - "reason": "Both mention SigLIP2, but one is about a broken documentation example and the other is about poor text embeddings. Related area, different bug." + "reason": "Both mention SigLIP2, but one is a docs/example mismatch/quantization issue and the other is a model-quality complaint; not the same bug." }, { "left": "issue:43644", "right": "issue:43950", - "accept": false, - "reason": "One concerns missing fields from AutoConfig with trust_remote_code; the other concerns non-persistent buffer corruption. Same general release area, not the same issue." + "accept": true, + "reason": "Direct duplicate report of the non-persistent-buffer junk regression." }, { "left": "issue:36331", "right": "issue:41762", "accept": false, - "reason": "These are different failure modes and code paths: a custom compute_loss signature error versus a DeepSpeed ZeRO-3 model loading failure." + "reason": "Different failure modes and code paths: custom compute_loss arg handling vs Gemma3 loading under DeepSpeed ZeRO-3." }, { "left": "issue:43828", "right": "issue:44928", "accept": false, - "reason": "Autocast dtype mismatch for Phi-tiny-MoE is unrelated to the Qwen3.5 flash-attention crash caused by 3D position_ids handling." + "reason": "Different execution bugs: autocast dtype mismatch vs 3D position_ids / flash-attn handling." }, { "left": "issue:43872", "right": "issue:44589", "accept": false, - "reason": "bitsandbytes Int8Params argument mismatch is a different problem from the Float8 storage lookup error." + "reason": "Different low-level compatibility problems: bitsandbytes kwarg mismatch vs missing Float8 storage object." }, { "left": "issue:45920", "right": "issue:46489", "accept": false, - "reason": "Both are tokenizer regressions, but one is broad and multi-model while the other is a specific DeepSeek-Coder v1 output bug. Not enough to merge as duplicates." + "reason": "Related tokenizer-regression area, but not the same concrete defect." }, { - "left": "issue:43054", - "right": "issue:45701", + "left": "issue:43866", + "right": "issue:44403", "accept": false, - "reason": "These both involve tokenization-related behavior changes, but the reported symptoms and scope are different enough that they should stay separate." + "reason": "Unrelated issues: checkpoint corruption vs generic loading noise." }, { - "left": "issue:43866", - "right": "issue:44403", + "left": "issue:43054", + "right": "issue:45701", "accept": false, - "reason": "Checkpoint corruption and generic loading noise are not the same bug." + "reason": "SigLIP2 embedding-quality complaint is not the same as a version-wide tokenization change." }, { "left": "issue:43950", "right": "issue:45902", "accept": false, - "reason": "Both mention non-persistent or materialized state, but 45902 is specifically about uninitialized reads in Qwen3_5MoeTextRotaryEmbedding after to_empty(cuda), which is a different concrete issue." + "reason": "Related symptom area, but 45902 is a more specific meta\u2192to_empty(cuda) materialization/uninitialized-memory problem, not clearly the same non-persistent-buffer corruption regression." }, { "left": "issue:38175", "right": "issue:43054", "accept": false, - "reason": "Unexpected zero probabilities and degraded text embeddings are different SigLIP2 behavior reports." + "reason": "Both concern SigLIP2, but zero probabilities and worse text embeddings are insufficiently specific to call the same bug." }, { "left": "issue:41553", "right": "issue:43883", "accept": false, - "reason": "Bad AutoTokenizer loading for Voxtral and a missing all_tied_weights_keys attribute on Molmo are separate model-specific bugs." + "reason": "Different loading problems for different models; no shared concrete root cause is evident." }, { "left": "issue:44910", "right": "issue:45859", "accept": false, - "reason": "The flash-attention crash from 3D position_ids and the CPU offload incompatibility in Qwen3_5MoeTextRotaryEmbedding are different failure modes." + "reason": "Different Qwen3.5/MoE bugs: flash-attn illegal access with 3D position_ids vs CPU offload incompatibility in rotary embedding." }, { "left": "issue:42947", "right": "issue:44387", "accept": false, - "reason": "LoRA gradient checkpointing ineffectiveness and increased CUDA reserved memory under int4 quantization are distinct problems." + "reason": "Both are memory-related, but the mechanisms differ: gradient checkpointing ineffectiveness vs increased reserved memory under int4 quantization." }, { "left": "issue:43054", "right": "issue:43493", - "accept": false, - "reason": "Both are SigLIP2-related, but one is about worse text embeddings and the other is an HF-vs-JAX implementation discrepancy. Related, not the same bug." + "accept": true, + "reason": "Plausibly the same SigLIP2 implementation discrepancy: degraded text embeddings fit an HF-vs-JAX mismatch." }, { "left": "issue:45663", "right": "issue:45902", "accept": false, - "reason": "Gemma4 FSDP2 KeyError and Qwen3_5Moe uninitialized-memory reads are different code paths." + "reason": "Different bugs in different models/code paths: Gemma4 FSDP2 KeyError vs Qwen3_5Moe buffer materialization." }, { "left": "issue:41553", "right": "issue:43881", "accept": false, - "reason": "Voxtral tokenizer loading and glm-4v-9b loading failure are unrelated." + "reason": "Different loading failures for different models and causes." }, { "left": "issue:44403", "right": "issue:44451", "accept": false, - "reason": "Generic loading noise and failure to load a specific model are not the same underlying issue." + "reason": "Generic loading noise vs inability to load a specific model; not the same defect." }, { "left": "issue:43975", "right": "issue:45520", "accept": false, - "reason": "Incorrect detokenization for DeepSeek-Coder v1 and a Python 3.13 flash_attn import error are unrelated." + "reason": "Tokenizer detokenization bug is unrelated to a Python 3.13 flash_attn import error." }, { "left": "issue:24643", "right": "issue:28530", "accept": false, - "reason": "DeepSpeed training shape error and early stopping metric handling are different issues." + "reason": "DeepSpeed training weight-shape runtime error is unrelated to early-stopping metric selection." }, { "left": "issue:44403", "right": "issue:44488", "accept": false, - "reason": "Generic transformer loading noise and a specific tokenizer load failure do not look like the same bug." + "reason": "Generic loading noise vs a specific model load failure; not the same issue." }, { "left": "issue:42947", "right": "issue:43856", "accept": false, - "reason": "These are both memory/training-related, but one is about gradient checkpointing with LoRA and the other is about Qwen3 MoE memory usage. Too broad to merge." + "reason": "Both involve memory usage, but one is LoRA checkpointing behavior and the other is Qwen3 MoE training efficiency." }, { "left": "issue:43643", "right": "issue:45020", "accept": false, - "reason": "AutoConfig missing fields with trust_remote_code and broad remote_code breakage are related themes, but not the same concrete regression." + "reason": "Related to remote_code, but one is missing fields in AutoConfig and the other is a broader regression breaking remote_code models." }, { "left": "issue:41762", "right": "issue:42915", "accept": false, - "reason": "Gemma3 ZeRO-3 loading failure and Qwen3Moe FP8 failure are different model-specific problems." + "reason": "Different model/load failures: Gemma3 with ZeRO-3 vs Qwen3 MoE with FineGrainedFP8Config." }, { "left": "issue:44991", "right": "issue:45399", "accept": false, - "reason": "Tokenizer loading for a specific model and flash-attn fallback gating are unrelated." + "reason": "Tokenizer-loading failure for a specific model is unrelated to flash-attn fallback gating." }, { "left": "issue:43824", "right": "issue:43872", "accept": false, - "reason": "Missing Qwen2_5_VLForConditionalGeneration import and bitsandbytes Int8Params incompatibility are separate issues." + "reason": "Different compatibility issues: missing Qwen2_5_VL export vs bitsandbytes Int8Params kwarg mismatch." }, { "left": "issue:45685", "right": "issue:45823", "accept": false, - "reason": "An MPS histogram op limitation in MoE and a Gemma4 device mismatch are unrelated." + "reason": "MPS histogram int support error is unrelated to Gemma4 device placement mismatch." }, { "left": "issue:43782", "right": "issue:44960", "accept": false, - "reason": "Qwen3VL weight_only loading error and GLM5 are different model-loading bugs." + "reason": "Different models and different error classes: Qwen3VL weight_only loading error vs GLM5 loading." }, { "left": "issue:43881", "right": "issue:43883", "accept": false, - "reason": "glm-4v-9b loading failure and a missing all_tied_weights_keys attribute on Molmo are not the same issue." + "reason": "Unrelated loading failures for different models and root causes." }, { "left": "issue:45020", "right": "issue:45823", "accept": false, - "reason": "Broad remote_code regressions and Gemma4 PLE device mismatch are too different to treat as duplicates." + "reason": "Broad remote_code regression and Gemma4 device_map forward mismatch are separate issues." }, { "left": "issue:44661", "right": "issue:45020", "accept": false, - "reason": "Tokenizer mapping-name failure and a general remote_code regression are separate problems." + "reason": "Tokenizer-mapping edge case is not the same bug as the remote_code regression." }, { "left": "issue:45081", "right": "issue:45399", "accept": false, - "reason": "Mistral regex patch crash and flash-attn fallback blocking are unrelated." + "reason": "Mistral regex patch crash is unrelated to flash-attn fallback gating." }, { - "left": "issue:43828", - "right": "issue:45823", + "left": "issue:42491", + "right": "issue:44368", "accept": false, - "reason": "Autocast dtype mismatch and Gemma4 device mismatch are different failure modes." + "reason": "Different Qwen3/Qwen3.5 LoRA-related problems: cross-version incompatibility vs tied-embeddings warning." }, { - "left": "issue:42491", - "right": "issue:44368", + "left": "issue:43828", + "right": "issue:45823", "accept": false, - "reason": "LoRA model incompatibility across versions and a tie_word_embeddings warning are not the same bug." + "reason": "Different model execution bugs: autocast dtype mismatch vs device-map placement mismatch." } ] }, "evaluator_result": { "accept": false, - "feedback": "The cluster-level summary is mostly grounded, but the writeup overstates the evidence in a few places. In particular, the packet only gives titles/body lengths, so claims like \"strongest root-cause framing,\" \"same underlying bug,\" and \"near-exact duplicate\" are stronger than the available evidence supports. The 43950/44534 pair looks highly similar, but the reasoning should be phrased more conservatively (e.g. \"likely duplicate\" or \"very similar regression\") unless you can cite body text or explicit links." + "feedback": "The cluster summary is mostly grounded, but one soft-edge verdict is too strong for the evidence in the packet: issue:43950 vs issue:45902. The titles support a possible related loading/materialization bug, but not clearly the same non-persistent-buffer corruption regression, so calling it a duplicate overstates the case. The SigLIP2 pair is more plausible, but the overall set is not conservative enough to accept as-is." }, "evaluator_used": true, "retried": true, @@ -1595,7 +1598,7 @@ "source": "llm", "cache_hit": false, "model": "gpt-5.4-mini?service_tier=flex", - "cluster_id": "cluster-10000-472", + "cluster_id": "cluster-10000-473", "nodes": [ "issue:24643", "issue:30064", @@ -1609,7 +1612,6 @@ "issue:43404", "issue:43531", "issue:43596", - "issue:43638", "issue:43792", "issue:43827", "issue:43828", @@ -1635,8 +1637,10 @@ "issue:45356", "issue:45488", "issue:45588", + "issue:45663", "issue:45715", "issue:45759", + "issue:45823", "issue:45902", "issue:45923", "issue:46315", @@ -1652,8 +1656,8 @@ "issue:43827|issue:44509", "issue:43404|issue:43950", "issue:45759|issue:46661", - "issue:43872|issue:44291", "issue:42915|issue:43872", + "issue:43872|issue:44291", "issue:45020|issue:46661", "issue:43792|issue:44220", "issue:43257|issue:45305", @@ -1663,223 +1667,223 @@ "issue:43940|issue:45313", "issue:43854|issue:44863", "issue:44928|issue:45588", - "issue:41093|issue:44805", "issue:44479|issue:44560", + "issue:41093|issue:44805", "issue:41553|issue:44488", "issue:24643|issue:30064", "issue:44863|issue:46402", "issue:44530|issue:44863", "issue:43366|issue:43828", "issue:43596|issue:46515", - "issue:45325|issue:45902", "issue:43366|issue:43531", - "issue:43638|issue:46515" + "issue:45325|issue:45902", + "issue:45663|issue:45823" ], - "prepared_review_unit_hash": "c50d08d4c9bf3d5692788e3881927573c37770c79a3262d7b68bb300f671edd9", + "prepared_review_unit_hash": "afac19b70ede5dd449c1a6a671a66e4c4a75f6143c2267567392667a65173197", "component_budget": { - "node_count": 472, - "item_count": 472, - "soft_pair_count": 1732, - "serialized_chars": 898654, - "estimated_input_tokens": 224664, - "estimated_eval_tokens": 449584 + "node_count": 473, + "item_count": 473, + "soft_pair_count": 1738, + "serialized_chars": 901437, + "estimated_input_tokens": 225360, + "estimated_eval_tokens": 450976 }, "budget": { - "node_count": 47, - "item_count": 47, + "node_count": 48, + "item_count": 48, "soft_pair_count": 28, - "serialized_chars": 36058, - "estimated_input_tokens": 9015, - "estimated_eval_tokens": 18286 + "serialized_chars": 36623, + "estimated_input_tokens": 9156, + "estimated_eval_tokens": 18568 }, "overflow_policy": "truncate_then_skip", "trimmed": false, "aggressively_trimmed": false, "split": true, "analyst_result": { - "summary": "This cluster is mostly heterogeneous, but it does contain a few clear duplicate subgroups: removed pipeline docs, mask-shape mismatch reports, the `_is_hf_initialized` / empty-weights loading incompatibility, Qwen3-family ZeRO-3 weight-loading failures, and a Qwen video regression in 5.3.0. The remaining pairs look only superficially similar or share a broad subsystem, not the same concrete bug.", - "confidence": 0.76, - "canonical_issue_reason": "No single issue cleanly represents the cluster; it splits into several unrelated duplicate subgroups.", + "summary": "Mostly heterogeneous regressions and model-specific bugs. The strongest duplicate clusters are the v5 docs cleanup, the `_is_hf_initialized` loading regression, and the generic mask-shape IndexError pair. The DeepSeek/Llama tokenizer pair is related but not the same underlying bug from the packet alone.", + "confidence": 0.67, + "canonical_issue_reason": "Issue 44291 is the clearest representative because it states the concrete loading failure and root cause (`_is_hf_initialized` on init-empty loading), while 43872 is a narrower symptom of the same path.", "canonical_pr_reason": null, - "best_issue_reason": "No global best issue: the items are not one coherent bug cluster, so a single representative issue would be misleading.", + "best_issue_reason": "44291 is the best global issue anchor because it most precisely identifies a concrete code-path regression with a plausible single fix target; the rest of the cluster is mostly separate model-specific or docs-only problems.", "best_pr_reason": null, "soft_edge_verdicts": [ { "left": "issue:45020", "right": "issue:45356", "accept": false, - "reason": "Both are version-related tokenizer/model regressions, but they describe different failures and model-specific causes." + "reason": "Both are recent-version regressions, but they affect different areas: remote-code model loading vs tokenizer handling for Kimi-K2.5." }, { "left": "issue:45715", "right": "issue:46315", "accept": false, - "reason": "Same tokenizer API area, but one is convert_ids_to_tokens and the other is convert_tokens_to_ids; different regressions." + "reason": "Both are tokenizer performance issues, but they hit different APIs (`convert_ids_to_tokens` vs `convert_tokens_to_ids`) and do not read as the same bug." }, { "left": "issue:45488", "right": "issue:46489", "accept": false, - "reason": "Both involve DeepSeek tokenization, but the affected model families and failure modes are different enough to keep separate." + "reason": "Both are tokenizer regressions, but the packet only gives titles and they point to different model families/failure modes; related, not clearly the same bug/root cause." }, { "left": "issue:43827", "right": "issue:44509", "accept": true, - "reason": "Both report the same v5 docs regression: summarization/translation pipeline tasks still referenced after removal." + "reason": "Both are the same docs cleanup problem: stale v5 documentation still references removed pipeline tasks like summarization and translation." }, { "left": "issue:43404", "right": "issue:43950", "accept": false, - "reason": "Different problems: tied lm_head weights in Mistral3 vs silent corruption of non-persistent buffers." + "reason": "One is a weight-tying bug in Mistral3ForConditionalGeneration; the other is a from_pretrained buffer-corruption regression." }, { "left": "issue:45759", "right": "issue:46661", "accept": false, - "reason": "Both touch model construction, but one is from_config text_config unwrapping and the other is a remote-code config_class attribute error." - }, - { - "left": "issue:43872", - "right": "issue:44291", - "accept": true, - "reason": "Same underlying `_is_hf_initialized` incompatibility during model loading / empty-weights construction." + "reason": "These are different model-loading failures: composite config unwrapping vs remote-code `config_class` lookup." }, { "left": "issue:42915", "right": "issue:43872", "accept": false, - "reason": "Different quantization/loading issues: FineGrainedFP8Config on Qwen3Moe vs bitsandbytes Int8Params kwarg incompatibility." + "reason": "Qwen3Moe FP8 failure and the bitsandbytes `_is_hf_initialized` incompatibility are different issues affecting different paths." + }, + { + "left": "issue:43872", + "right": "issue:44291", + "accept": true, + "reason": "Same root cause: an unexpected `_is_hf_initialized` argument breaks loading/initialization, with 44291 describing the regression directly." }, { "left": "issue:45020", "right": "issue:46661", "accept": false, - "reason": "Both mention remote-code models, but the concrete bugs are different (remote_code breakage vs missing config_class)." + "reason": "Both involve remote-code models, but one is a broad regression claim and the other is a specific `config_class` attribute error; too broad to merge." }, { "left": "issue:43792", "right": "issue:44220", "accept": false, - "reason": "The Whisper model load failure is too generic to prove it is the same bug as the specific fbank-feature helper issue." + "reason": "The Whisper failure and `_torch_extract_fbank_features()` issue may be related, but the packet does not establish the same concrete bug path." }, { "left": "issue:43257", "right": "issue:45305", "accept": false, - "reason": "Both are DeepSpeed-related, but one is weight conversion/loading and the other is gradient averaging with GAS." + "reason": "One is a Qwen3 MoE weight-conversion/loading problem; the other is DeepSpeed gradient averaging with GAS." }, { "left": "issue:42947", "right": "issue:45923", "accept": false, - "reason": "Gradient checkpointing not working with LoRA is a different issue from a missing supports_gradient_checkpointing flag." + "reason": "Gradient checkpointing being ineffective in PEFT LoRA is not the same as a missing `supports_gradient_checkpointing` flag on a remote-code variant." }, { "left": "issue:41553", "right": "issue:44991", "accept": false, - "reason": "Both are tokenizer-loading failures, but on different models and with different causes." + "reason": "Both are tokenizer-loading complaints, but they involve different models and different failure modes." }, { "left": "issue:36331", "right": "issue:41093", "accept": false, - "reason": "Unexpected compute_loss kwarg and mask/tensor shape mismatch are unrelated errors." + "reason": "Unexpected `num_items_in_batch` in custom trainer loss and a tensor-mask shape mismatch are unrelated errors." }, { "left": "issue:43940", "right": "issue:45313", - "accept": true, - "reason": "Both are Qwen-family DeepSpeed ZeRO-3 weight-loading failures with missing parameters, likely the same code-path bug." + "accept": false, + "reason": "Both are DeepSpeed weight-loading bugs, but one is Qwen3-Next missing params and the other is Qwen3.5 `language_model` loading." }, { "left": "issue:43854", "right": "issue:44863", "accept": false, - "reason": "Different model-loading failures for unrelated model families." + "reason": "Unit-test model loading failure and NemotronH checkpoint loading failure are different model-specific problems." }, { "left": "issue:44928", "right": "issue:45588", "accept": false, - "reason": "One is a NaN training regression from position_ids/SDPA behavior; the other is an AttributeError in flash_attention handling." - }, - { - "left": "issue:41093", - "right": "issue:44805", - "accept": true, - "reason": "Same mask/index shape mismatch error, with the second issue appearing to be the same bug reported with different sizes." + "reason": "RLHF gradient explosion in Qwen3.5 and a flash-attention `s_aux=None` crash are unrelated." }, { "left": "issue:44479", "right": "issue:44560", + "accept": false, + "reason": "Both are Qwen/VL regressions, but one is a video-input regression and the other is a StopIteration on Qwen3-vl-embedding." + }, + { + "left": "issue:41093", + "right": "issue:44805", "accept": true, - "reason": "Both are 5.3.0 Qwen video-input regressions hitting the same video-processing path." + "reason": "Both describe the same generic index/mask length mismatch error with an off-by-one sized indexed tensor." }, { "left": "issue:41553", "right": "issue:44488", "accept": false, - "reason": "Both concern tokenizer/model loading, but they are different models and the Voxtral error-message issue is not the same bug." + "reason": "Different tokenizer-loading issues for different models; one is an error message problem, the other is a load failure." }, { "left": "issue:24643", "right": "issue:30064", "accept": false, - "reason": "Training with DeepSpeed and void segmentation-map preprocessing are unrelated." + "reason": "Training with DeepSpeed and void segmentation-map processing are unrelated." }, { "left": "issue:44863", "right": "issue:46402", "accept": false, - "reason": "Different model-loading failures for NemotronH and Sam3Model." + "reason": "NemotronH checkpoint loading and Sam3 text-encoder weight loading are different model-specific failures." }, { "left": "issue:44530", "right": "issue:44863", "accept": false, - "reason": "PagedAttentionCache linear_attention crash is unrelated to NemotronH checkpoint loading." + "reason": "PagedAttentionCache linear_attention crash and NemotronH checkpoint loading are unrelated." }, { "left": "issue:43366", "right": "issue:43828", "accept": false, - "reason": "GGUF gpt-oss support and Phi-tiny-MoE autocast dtype mismatch are different issues." + "reason": "GGUF gpt-oss support and a torch autocast dtype mismatch in Phi-tiny-MoE are different issues." }, { "left": "issue:43596", "right": "issue:46515", "accept": false, - "reason": "DeepSpeed ZeRO-3 init failure with BertModel is unrelated to torch 2.4.1 incompatibility." + "reason": "Zero-3/BertModel index-out-of-bounds and torch 2.4.1 incompatibility are not the same bug." }, { - "left": "issue:45325", - "right": "issue:45902", + "left": "issue:43366", + "right": "issue:43531", "accept": false, - "reason": "Both involve rotary/position handling, but they affect different models and different bugs." + "reason": "GGUF architecture support and Qwen3-MoE sliding_window behavior are unrelated." }, { - "left": "issue:43366", - "right": "issue:43531", + "left": "issue:45325", + "right": "issue:45902", "accept": false, - "reason": "GGUF architecture support and Qwen3-MoE sliding_window behavior are different problems." + "reason": "Both touch rotary/position handling, but one is a rope-index scaling bug and the other is uninitialized `inv_freq` memory after materialization." }, { - "left": "issue:43638", - "right": "issue:46515", + "left": "issue:45663", + "right": "issue:45823", "accept": false, - "reason": "Non-pretrained Bert ZeRO-3 init failure is unrelated to the torch version incompatibility report." + "reason": "Gemma4 FSDP2 shared_kv_states KeyError and Gemma4 device mismatch are separate forward-pass bugs." } ] }, "evaluator_result": { - "accept": true, - "feedback": "The summary is grounded in the packet and stays mostly conservative. The accepted soft pairs are supported by close title-level alignment, especially the exact/near-exact duplicates around the docs regression, mask-shape mismatch, and `_is_hf_initialized` loading issue. The cluster-level conclusion that this set is heterogeneous with a few duplicate subgroups is reasonable." + "accept": false, + "feedback": "The cluster-level summary is mostly grounded, but one soft-edge accept overstates the evidence: `issue:45488` vs `issue:46489` are both tokenizer regressions, yet the packet only gives titles and they point to different model families/failure modes. From the packet alone, that is better described as related rather than the same bug/root cause. The other accepted pairs look reasonably conservative." }, "evaluator_used": true, - "retried": false, + "retried": true, "accepted_nontrivial_soft_edge": true, "error_kind": null, "error_message": null @@ -1895,7 +1899,7 @@ "source": "llm", "cache_hit": false, "model": "gpt-5.4-mini?service_tier=flex", - "cluster_id": "cluster-10000-472", + "cluster_id": "cluster-10000-473", "nodes": [ "issue:36010", "issue:38175", @@ -1932,10 +1936,11 @@ "issue:44877", "issue:44910", "issue:44918", + "issue:44960", + "issue:44977", "issue:45030", "issue:45276", "issue:45584", - "issue:45663", "issue:45685", "issue:45812", "issue:45823", @@ -1946,7 +1951,7 @@ "issue:46515" ], "soft_pairs": [ - "issue:45663|issue:45823", + "issue:43638|issue:46515", "issue:43716|issue:44560", "issue:44155|issue:45584", "issue:43366|issue:44589", @@ -1956,21 +1961,21 @@ "issue:44560|issue:45685", "issue:43645|issue:43646", "issue:43550|issue:44910", - "issue:44683|issue:46515", "issue:44560|issue:44918", + "issue:44683|issue:46515", "issue:42617|issue:43366", "issue:36010|issue:42915", "issue:45812|issue:46489", "issue:45823|issue:46515", - "issue:44661|issue:45823", "issue:44560|issue:44805", + "issue:44661|issue:45823", "issue:38175|issue:43994", - "issue:44877|issue:45030", "issue:43632|issue:45834", + "issue:44877|issue:45030", "issue:43638|issue:44661", "issue:43596|issue:44661", - "issue:44661|issue:44805", "issue:43317|issue:44756", + "issue:44661|issue:44805", "issue:38175|issue:39692", "issue:42491|issue:43720", "issue:43644|issue:44534", @@ -1978,120 +1983,121 @@ "issue:42915|issue:44756", "issue:42491|issue:45276", "issue:44451|issue:44488", - "issue:46459|issue:46512" + "issue:46459|issue:46512", + "issue:44960|issue:44977" ], - "prepared_review_unit_hash": "4bdc9b81ee6cde5140d5c82f9e8414f050c87431a3f2de9745aa233df6a73813", + "prepared_review_unit_hash": "864186513e4275ee288d5726dc92703705a6154b26d368d7b98ea126d8303a92", "component_budget": { - "node_count": 472, - "item_count": 472, - "soft_pair_count": 1732, - "serialized_chars": 898654, - "estimated_input_tokens": 224664, - "estimated_eval_tokens": 449584 + "node_count": 473, + "item_count": 473, + "soft_pair_count": 1738, + "serialized_chars": 901437, + "estimated_input_tokens": 225360, + "estimated_eval_tokens": 450976 }, "budget": { - "node_count": 47, - "item_count": 47, - "soft_pair_count": 33, - "serialized_chars": 37046, - "estimated_input_tokens": 9262, - "estimated_eval_tokens": 18780 + "node_count": 48, + "item_count": 48, + "soft_pair_count": 34, + "serialized_chars": 37774, + "estimated_input_tokens": 9444, + "estimated_eval_tokens": 19144 }, "overflow_policy": "truncate_then_skip", "trimmed": false, "aggressively_trimmed": false, "split": true, "analyst_result": { - "summary": "Mostly a loose set of unrelated Transformers issues with a few true near-duplicates. The strongest duplicate pairs are the v5 custom-model init breakage, the non-persistent-buffer corruption regression, and the SigLIP2 bad-output reports; the rest are too model-/path-specific to merge.", - "confidence": 0.64, - "canonical_issue_reason": "Issue 43644 is the clearest broad report of the v5 non-persistent-buffer corruption regression and is the best duplicate target among the accepted issue pairs; 44534 is essentially the same bug with a later wording.", + "summary": "Mostly a loose collection of Transformers v5 regressions and model-specific inference/load bugs. Only a few soft edges look like true duplicates; most pairs are different code paths or different models. No PRs are present in this cluster.", + "confidence": 0.82, + "canonical_issue_reason": "issue:43646 is the broadest representative of the main theme here: Transformers v5 breaking custom model initialization/loading. It also subsumes the notebook-specific variant in issue:43645 better than any narrower report.", "canonical_pr_reason": null, - "best_issue_reason": "43644 is the most generally reusable issue in this set: it describes a concrete, cross-model v5 regression with a clear failure mode and is broad enough to absorb the duplicate report in 44534.", + "best_issue_reason": "issue:43646 is the best single anchor for the cluster because it captures the broadest and most reusable regression label among the mostly heterogeneous issues, even though the cluster is not a tight duplicate set.", "best_pr_reason": null, "soft_edge_verdicts": [ { - "left": "issue:45663", - "right": "issue:45823", + "left": "issue:43638", + "right": "issue:46515", "accept": false, - "reason": "Same broad Gemma4 area, but one is an FSDP2 KeyError in attention and the other is a device_map mismatch in PLE; different code paths and fixes." + "reason": "Different bugs: zero3/BertModel init failure vs a generic torch 2.4.1 incompatibility." }, { "left": "issue:43716", "right": "issue:44560", "accept": false, - "reason": "Different models and symptoms: image preprocessor/model dtype mismatch versus a Qwen3-VL video StopIteration error." + "reason": "Different model paths: image preprocessor dtype mismatch vs Qwen3-VL video StopIteration." }, { "left": "issue:44155", "right": "issue:45584", "accept": false, - "reason": "Both are generation bugs, but they affect different models and failure modes; no clear shared underlying code-path." + "reason": "AudioFlamingo3 batch leakage and Whisper empty-transcription failure are unrelated." }, { "left": "issue:43366", "right": "issue:44589", "accept": false, - "reason": "Unrelated functionality: GGUF gpt-oss support versus a Float8 storage deserialization/type error." + "reason": "GGUF gpt-oss support request vs Float8 storage lookup error; not the same bug." }, { "left": "issue:43950", "right": "issue:45834", "accept": false, - "reason": "Both are loading-related, but one is a non-persistent buffer corruption regression and the other is a long-input OCR index error in Kosmos2.5." + "reason": "Both are regressions, but non-persistent buffer corruption and Kosmos2.5 OCR index error are different failures." }, { "left": "issue:43824", "right": "issue:43957", "accept": false, - "reason": "Import/export of a specific Qwen2.5 VL class is a different issue from meta-device loading failures." + "reason": "Import-path regression for Qwen2_5_VL vs generic torch.device(\"meta\") loading breakage." }, { "left": "issue:42915", "right": "issue:43957", "accept": false, - "reason": "Different root causes: Qwen3-MoE FP8 config failure versus meta-device loading breakage." + "reason": "Qwen3MoE FP8 failure and meta-device loading failures are different code paths." }, { "left": "issue:44560", "right": "issue:45685", "accept": false, - "reason": "Separate bugs in different subsystems/models; only a superficial similarity around MPS/runtime errors." + "reason": "Unrelated bugs: Qwen3-VL video StopIteration vs an MPS histogram/int error in MoE." }, { "left": "issue:43645", "right": "issue:43646", "accept": true, - "reason": "These read like the same Transformers v5 custom-model initialization regression, with 43645 being the notebook-specific presentation of the broader 43646 report." + "reason": "Same underlying regression: custom model initialization breaks in Transformers v5; the notebook context is just a reproduction detail." }, { "left": "issue:43550", "right": "issue:44910", "accept": false, - "reason": "Different attention/compile failures on different models; no evidence they share the same concrete bug." + "reason": "Different model/runtime issues: Bamba torch.compile+SDPA vs Qwen3.5 Flash Attention packed-sequence misread." }, { - "left": "issue:44683", - "right": "issue:46515", + "left": "issue:44560", + "right": "issue:44918", "accept": false, - "reason": "Both mention torch version compatibility, but they concern different breakages and different torch releases." + "reason": "Different Qwen3-family failures with different symptoms and code paths." }, { - "left": "issue:44560", - "right": "issue:44918", + "left": "issue:44683", + "right": "issue:46515", "accept": false, - "reason": "Different Qwen3-VL embedding/video issues; one is a StopIteration on video input, the other is an SFT trainer unpacking failure." + "reason": "Compiled flex_attention on torch>=2.9 is unrelated to a torch 2.4.1 compatibility issue." }, { "left": "issue:42617", "right": "issue:43366", "accept": false, - "reason": "3d_parallel.py execution failure is unrelated to GGUF gpt-oss support." + "reason": "3d_parallel.py runtime failure and GGUF gpt-oss support are unrelated." }, { "left": "issue:36010", "right": "issue:42915", "accept": false, - "reason": "ImportError for GenerationMixin and Qwen3-MoE FP8 failure are unrelated bugs." + "reason": "GenerationMixin import error is unrelated to Qwen3MoE FP8 loading failure." }, { "left": "issue:45812", @@ -2103,115 +2109,115 @@ "left": "issue:45823", "right": "issue:46515", "accept": false, - "reason": "Gemma4 PLE device mismatch is not the same as a general torch 2.4.1 incompatibility report." + "reason": "Gemma4 device mismatch during forward is not the same as a generic torch version incompatibility." }, { - "left": "issue:44661", - "right": "issue:45823", + "left": "issue:44560", + "right": "issue:44805", "accept": false, - "reason": "Tokenizer-mapping failure versus Gemma4 PLE device mismatch; different subsystems." + "reason": "Video StopIteration and mask/tensor shape IndexError are different bugs." }, { - "left": "issue:44560", - "right": "issue:44805", + "left": "issue:44661", + "right": "issue:45823", "accept": false, - "reason": "Different runtime errors and different bug surfaces; no shared underlying change." + "reason": "Tokenizer-mapping model-registration bug vs Gemma4 PLE device mismatch; not the same issue." }, { "left": "issue:38175", "right": "issue:43994", "accept": true, - "reason": "Both target SigLIP2 base model bad outputs on the same checkpoint family; 43994 is the broader repro of the earlier zero-probability complaint." + "reason": "Both report broken SigLIP2 base-patch16-224 outputs on the same model family, with the same underlying inference regression." }, { - "left": "issue:44877", - "right": "issue:45030", + "left": "issue:43632", + "right": "issue:45834", "accept": false, - "reason": "Config validation regressions, but for different model families and likely different validation rules." + "reason": "_is_hf_initialized flag regression and Kosmos2.5 long-OCR index error are unrelated." }, { - "left": "issue:43632", - "right": "issue:45834", + "left": "issue:44877", + "right": "issue:45030", "accept": false, - "reason": "_is_hf_initialized behavior under v5 is unrelated to a long-input OCR index error." + "reason": "Both mention config validation, but they are different model configs and not clearly the same concrete bug." }, { "left": "issue:43638", "right": "issue:44661", "accept": false, - "reason": "Both involve initialization or mapping, but one is DeepSpeed zero3/BertModel indexing and the other is add-new-model-like tokenizer mapping behavior." + "reason": "Zero3/BertModel init failure is unrelated to the TOKENIZER_MAPPING_NAMES add-new-model-like bug." }, { "left": "issue:43596", "right": "issue:44661", "accept": false, - "reason": "Different failures: zero3/BertModel index error versus tokenizer mapping limitation." - }, - { - "left": "issue:44661", - "right": "issue:44805", - "accept": false, - "reason": "Tokenizer mapping constraint and mask/tensor shape mismatch are unrelated." + "reason": "Zero3/BertModel init failure and tokenizer-mapping logic failure are different code paths." }, { "left": "issue:43317", "right": "issue:44756", "accept": false, - "reason": "Device-map offload loading bug versus mmap OOM on Strix Halo; different fixes." + "reason": "gpu+cpu offload loading failure and Strix Halo mmap OOM are different problems." }, { "left": "issue:38175", "right": "issue:39692", "accept": false, - "reason": "Same model family, but the second issue is a docs example with model/processor mismatch and quantization failure, not the same output-correction bug." + "reason": "Same model family, but the SigLIP2 output regression and the doc-example model/processor mismatch are not clearly the same bug." }, { "left": "issue:42491", "right": "issue:43720", "accept": false, - "reason": "Different model stacks and different accelerated-loading failures; too far apart to merge." + "reason": "Qwen3_moe LoRA compatibility and BitNet packed-weight unpacking are unrelated." }, { "left": "issue:43644", "right": "issue:44534", "accept": true, - "reason": "Same underlying v5 regression: non-persistent buffers are being filled/corrupted with junk during load." + "reason": "Exact same regression: Transformers v5 fills non-persistent buffers with junk; the titles are near-paraphrases." }, { "left": "issue:41720", "right": "issue:42915", "accept": false, - "reason": "Qwen3 auto device mapping cuda assert and Qwen3-MoE FP8 failure are different problems." + "reason": "Qwen3 auto device mapping assert and Qwen3MoE FineGrainedFP8 failure are different runtime paths." }, { "left": "issue:42915", "right": "issue:44756", "accept": false, - "reason": "FP8 config failure is unrelated to mmap-related OOM behavior." + "reason": "Qwen3MoE FP8 failure is unrelated to Strix Halo mmap/OOM behavior." }, { "left": "issue:42491", "right": "issue:45276", "accept": false, - "reason": "LoRA qwen3_moe incompatibility on hf5.x and Gemma4 token-embedding resize behavior are distinct bugs." + "reason": "Qwen3_moe LoRA compatibility and Gemma4 resize_token_embeddings behavior are unrelated." }, { "left": "issue:44451", "right": "issue:44488", "accept": false, - "reason": "Both are model-loading complaints, but the evidence points to different models and not the same concrete regression." + "reason": "Both are loading failures, but the titles point to different models and no shared concrete bug is visible." }, { "left": "issue:46459", "right": "issue:46512", "accept": false, - "reason": "One is a dtype-loading regression for multimodal checkpoints, the other is stale config.dtype after from_config; related theme but different bugs." + "reason": "Both concern dtype handling, but one is checkpoint loading and the other is from_config/config-state synchronization." + }, + { + "left": "issue:44960", + "right": "issue:44977", + "accept": false, + "reason": "GLM5 and Qwen3.5 Flash Attention generation failures are unrelated." } ] }, "evaluator_result": { "accept": true, - "feedback": "The analysis is grounded in the packet and stays appropriately conservative on the soft-edge pairs. The accepted duplicates (43645/43646, 38175/43994, 43644/44534) are plausible from the titles, and the rejected pairs are mostly justified by different models, subsystems, or failure modes. No clear overstatement stands out." + "feedback": "The cluster summary is grounded in the packet and stays conservative about the mostly loose/heterogeneous relationships. The accepted soft edges that are marked as duplicates are plausibly supported by the titles, especially the exact non-persistent-buffer regression pair and the shared SigLIP2 model-family pair. No PRs are present, so there is no mergeability concern to flag." }, "evaluator_used": true, "retried": false, @@ -2230,7 +2236,7 @@ "source": "llm", "cache_hit": false, "model": "gpt-5.4-mini?service_tier=flex", - "cluster_id": "cluster-10000-472", + "cluster_id": "cluster-10000-473", "nodes": [ "issue:36010", "issue:39404", @@ -2261,7 +2267,6 @@ "issue:44756", "issue:44910", "issue:44960", - "issue:44977", "issue:44991", "issue:45276", "issue:45292", @@ -2282,14 +2287,13 @@ "issue:46468" ], "soft_pairs": [ - "issue:44960|issue:44977", "issue:45313|issue:45357", "issue:45507|issue:46468", - "issue:44488|issue:44991", "issue:43856|issue:43873", + "issue:44488|issue:44991", "issue:43644|issue:43645", - "issue:39692|issue:43994", "issue:43708|issue:44075", + "issue:39692|issue:43994", "issue:43856|issue:44387", "issue:43854|issue:43931", "issue:43257|issue:43854", @@ -2303,8 +2307,8 @@ "issue:44910|issue:45588", "issue:44402|issue:44488", "issue:43425|issue:43957", - "issue:42947|issue:45704", "issue:43873|issue:44387", + "issue:42947|issue:45704", "issue:41720|issue:44155", "issue:43404|issue:43454", "issue:39692|issue:44960", @@ -2314,230 +2318,221 @@ "issue:45663|issue:45859", "issue:44387|issue:44756" ], - "prepared_review_unit_hash": "c98c98c88f710e6af00fdd7db2a166209c02015861297c104e4ff4591173672a", + "prepared_review_unit_hash": "bbdead1abf518d03a512a25ad4a0ec67571bf5239e236b70e177666a2597caea", "component_budget": { - "node_count": 472, - "item_count": 472, - "soft_pair_count": 1732, - "serialized_chars": 898654, - "estimated_input_tokens": 224664, - "estimated_eval_tokens": 449584 + "node_count": 473, + "item_count": 473, + "soft_pair_count": 1738, + "serialized_chars": 901437, + "estimated_input_tokens": 225360, + "estimated_eval_tokens": 450976 }, "budget": { - "node_count": 48, - "item_count": 48, - "soft_pair_count": 31, - "serialized_chars": 37058, - "estimated_input_tokens": 9265, - "estimated_eval_tokens": 18786 + "node_count": 47, + "item_count": 47, + "soft_pair_count": 30, + "serialized_chars": 36203, + "estimated_input_tokens": 9051, + "estimated_eval_tokens": 18358 }, "overflow_policy": "truncate_then_skip", "trimmed": false, "aggressively_trimmed": false, "split": true, "analyst_result": { - "summary": "This cluster is mostly heterogeneous: many unrelated model-loading, quantization, tokenizer, and training regressions. Only the Gemma4/T5Gemma resize_token_embeddings reports look like the same underlying bug family; the rest share broad subsystems but not the same concrete failure.", + "summary": "These items are mostly unrelated Transformers regressions that only share broad themes like loading, quantization, or model-specific compatibility. I do not see any true duplicate pairs among the soft edges. If forced to pick a representative issue, 45812 is the clearest standalone regression.", "confidence": 0.79, - "canonical_issue_reason": "issue:45276 is the strongest representative of the only plausible duplicate subcluster: it clearly describes a resize_token_embeddings propagation bug across model-specific embedding/output heads and is more detailed than the nearby T5Gemma report.", + "canonical_issue_reason": "Issue 45812 is the strongest representative: it reports a clear, broad tokenizer regression with explicit impact across Granite models and a well-defined failure mode.", "canonical_pr_reason": null, - "best_issue_reason": "issue:45276 is the best global representative because it is the clearest, most detailed report in the only likely duplicate pair; the rest of the cluster is too mixed to justify a broader canonical.", + "best_issue_reason": "45812 is the best standalone issue candidate because it is concrete, high-signal, and broader in scope than the other reports; it is not a duplicate of them.", "best_pr_reason": null, "soft_edge_verdicts": [ - { - "left": "issue:44960", - "right": "issue:44977", - "accept": false, - "reason": "Both are model-specific generation failures, but GLM5 and Qwen3.5 flash-attn problems are different code paths and symptoms." - }, { "left": "issue:45313", "right": "issue:45357", "accept": false, - "reason": "Same model family, but one is ZeRO-3 loading and the other is save_pretrained key serialization; not the same bug." + "reason": "Different bugs: one is DeepSpeed ZeRO-3 failing to load language_model weights, the other is save_pretrained saving wrong visual encoder keys." }, { "left": "issue:45507", "right": "issue:46468", "accept": false, - "reason": "Unrelated models and failures; no shared concrete code path." + "reason": "Different model families and failures: GraniteMoEHybrid invalid method call vs a broken Nemotron example." }, { - "left": "issue:44488", - "right": "issue:44991", + "left": "issue:43856", + "right": "issue:43873", "accept": false, - "reason": "Both are tokenizer/model-loading regressions, but for different models and different failure modes." + "reason": "Both involve Qwen3/MoE resource behavior, but one is training memory inefficiency and the other is quantization/offloading not working; not the same code-path bug." }, { - "left": "issue:43856", - "right": "issue:43873", + "left": "issue:44488", + "right": "issue:44991", "accept": false, - "reason": "Both mention memory/offloading, but one is Qwen3 MoE training memory use and the other is quantization offloading behavior." + "reason": "Both are model-loading problems, but they affect different models and likely different tokenizer/config paths." }, { "left": "issue:43644", "right": "issue:43645", "accept": false, - "reason": "Both are Transformers 5.0 regressions, but one is buffer initialization and the other is custom model initialization in notebooks." + "reason": "Separate Transformers 5 regressions: non-persistent buffers getting junk vs Jupyter custom model initialization breaking." }, { - "left": "issue:39692", - "right": "issue:43994", + "left": "issue:43708", + "right": "issue:44075", "accept": false, - "reason": "Both involve SigLIP2, but one is a broken docs example and the other is bad inference results; different issues." + "reason": "Trainer checkpoint-resume step \u0917\u0923ation issue vs SGD arguments being ignored; unrelated subsystems." }, { - "left": "issue:43708", - "right": "issue:44075", + "left": "issue:39692", + "right": "issue:43994", "accept": false, - "reason": "Trainer checkpoint math and SGD optimizer args being ignored are unrelated." + "reason": "Same family (SigLIP2) but different problems: a bad documentation example versus incorrect model outputs at runtime." }, { "left": "issue:43856", "right": "issue:44387", "accept": false, - "reason": "Both are memory-related, but one is MoE training and the other is int4 quantization reserved-memory/OOM behavior." + "reason": "Both mention memory/quantization, but one is MoE training memory use and the other is int4 reserved-memory OOM; different root causes." }, { "left": "issue:43854", "right": "issue:43931", "accept": false, - "reason": "Different models and different load-time failures; not the same bug." + "reason": "Different load failures for different models: GLM-4.7-Flash unit test loading vs Qwen3-VL weight-shape mismatch." }, { "left": "issue:43257", "right": "issue:43854", "accept": false, - "reason": "Qwen3 MoE weight conversion under accelerate/deepspeed is unrelated to the GLM-4.7 unit-test loading issue." + "reason": "Qwen3 MoE weight conversion under accelerate+deepspeed is unrelated to GLM-4.7-Flash loading in unit tests." }, { "left": "issue:42491", "right": "issue:45292", "accept": false, - "reason": "Both touch model shape/resize behavior, but LoRA hf4.x/hf5.x incompatibility is not the same as resize_token_embeddings propagation." + "reason": "LoRA/hf5 compatibility issue for Qwen3 MoE is unrelated to resize_token_embeddings not updating output embeddings." }, { "left": "issue:42491", "right": "issue:43931", "accept": false, - "reason": "Different root causes: LoRA compatibility vs Qwen3-VL weight-shape mismatch." + "reason": "Different model and failure mode: Qwen3 MoE LoRA compatibility vs Qwen3-VL loading shape mismatch." }, { "left": "issue:44960", "right": "issue:46050", "accept": false, - "reason": "GLM5 loading is unrelated to sinq quantized save/load corruption." + "reason": "GLM5 report versus sinq quantized save/load breakage; no shared concrete bug." }, { "left": "issue:45488", "right": "issue:45812", "accept": false, - "reason": "Both are tokenizer regressions, but one is ByteLevel/Metaspace override and the other is wrong token IDs for Granite models." + "reason": "Both are tokenizer regressions, but one swaps the pre-tokenizer for DeepSeek V3/R1 while the other produces wrong IDs for Granite models; different models and mechanisms." }, { "left": "issue:45362", "right": "issue:45406", "accept": false, - "reason": "Qwen3.5 chat crash and Gemma4Processor missing _tokenizer are separate model/framework bugs." + "reason": "Qwen3.5 chat crash and Gemma4Processor missing _tokenizer are unrelated model/processor bugs." }, { "left": "issue:44164", "right": "issue:44291", "accept": false, - "reason": "Both affect loading/saving, but extra_state handling and init_empty_weights _is_hf_initialized are distinct issues." + "reason": "save/from_pretrained extra_state handling and init_empty_weights unexpected argument are separate serialization/init issues." }, { "left": "issue:45276", "right": "issue:45335", - "accept": true, - "reason": "Same underlying resize_token_embeddings propagation bug across composite model embeddings/output heads; titles and symptoms align closely." + "accept": false, + "reason": "Similar symptom around resize_token_embeddings, but different architectures and affected submodules; not enough evidence of one shared bug." }, { "left": "issue:44910", "right": "issue:45588", "accept": false, - "reason": "Both mention flash_attention, but one is a Qwen3.5 position_ids crash and the other is an s_aux=None AttributeError in a different integration path." + "reason": "Both touch flash attention, but one is a Qwen3.5 illegal-memory-access issue from packed-sequence interpretation and the other is an AttributeError on s_aux=None." }, { "left": "issue:44402", "right": "issue:44488", "accept": false, - "reason": "One is a vocab-size/token placement bug, the other is a model loading failure; not the same defect." + "reason": "Tokenizer vocab-size mismatch for one model versus a loading failure for another model; related theme but not the same bug." }, { "left": "issue:43425", "right": "issue:43957", "accept": false, - "reason": "Torch version incompatibility and meta-device loading regressions are unrelated." + "reason": "Torch 2.10 incompatibility and meta-device loading breakage are distinct compatibility regressions." }, { - "left": "issue:42947", - "right": "issue:45704", + "left": "issue:43873", + "right": "issue:44387", "accept": false, - "reason": "Gradient checkpointing/LoRA and apex RMSNorm memory leak are different subsystems and fixes." + "reason": "Offloading not working with quantization versus increased reserved memory causing int4 OOM; different failure mechanisms." }, { - "left": "issue:43873", - "right": "issue:44387", + "left": "issue:42947", + "right": "issue:45704", "accept": false, - "reason": "Quantization offloading not working and int4 reserved-memory OOM are related only broadly, not the same bug." + "reason": "Gradient checkpointing with PEFT LoRA and a T5 apex RMSNorm memory leak are unrelated." }, { "left": "issue:41720", "right": "issue:44155", "accept": false, - "reason": "Qwen3 auto device-map CUDA assert and AudioFlamingo3 batched inference token leak are unrelated." + "reason": "Qwen3 auto device mapping CUDA assert and AudioFlamingo3 batched inference leakage are different bugs in different models." }, { "left": "issue:43404", "right": "issue:43454", "accept": false, - "reason": "Both are lm_head tying problems, but they affect different model classes and are not clearly the same code-path bug." + "reason": "Both mention untied lm_head weights, but they are different model classes with separate implementations; no evidence they are the same concrete fix." }, { "left": "issue:39692", "right": "issue:44960", "accept": false, - "reason": "SigLIP2 docs/inference issues are unrelated to GLM5 loading." + "reason": "SigLIP2 doc-example errors and GLM5 issues are unrelated." }, { "left": "issue:45663", "right": "issue:45759", "accept": false, - "reason": "Both involve composite model internals, but one is Gemma4 FSDP2 shared_kv_states and the other is Qwen config unwrapping; different bugs." + "reason": "Gemma4 FSDP2 shared_kv_states rebuild bug is unrelated to Qwen 3.5 multimodal config unwrapping." }, { "left": "issue:44291", "right": "issue:44756", "accept": false, - "reason": "init_empty_weights argument handling and mmap OOM on Strix Halo are unrelated." + "reason": "Unexpected _is_hf_initialized argument under init_empty_weights versus mmap OOM on Strix Halo are different bugs." }, { "left": "issue:36010", "right": "issue:39404", "accept": false, - "reason": "ImportError for GenerationMixin and Whisper return_language pipeline regression are separate issues." + "reason": "Importing GenerationMixin from transformers.generation and Whisper return_language pipeline regression are unrelated." }, { "left": "issue:45663", "right": "issue:45859", "accept": false, - "reason": "Gemma4 FSDP2 shared state and Qwen3_5Moe CPU offload incompatibility are different failures." + "reason": "Gemma4 FSDP2 attention-state bug versus Qwen3.5 CPU-offload incompatibility in rotary embedding are different code paths." }, { "left": "issue:44387", "right": "issue:44756", "accept": false, - "reason": "Int4 quantization memory growth and mmap OOM on Strix Halo are not the same bug." + "reason": "Both are memory-related, but one is int4 quantization reserved-memory growth and the other is mmap-related OOM on Strix Halo." } ] }, - "evaluator_result": { - "accept": true, - "feedback": "Grounded and conservative overall. The only clearly supported duplicate pair is issue:45276 / issue:45335, and the summary correctly avoids claiming broader clustering evidence than the packet shows. The other soft-edge rejections are cautious and do not overstate similarity." - }, - "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 }, @@ -2552,7 +2547,7 @@ "source": "llm", "cache_hit": false, "model": "gpt-5.4-mini?service_tier=flex", - "cluster_id": "cluster-10000-472", + "cluster_id": "cluster-10000-473", "nodes": [ "issue:39404", "issue:42796", @@ -2600,7 +2595,8 @@ "issue:46315", "issue:46326", "issue:46468", - "issue:46491" + "issue:46491", + "issue:46665" ], "soft_pairs": [ "issue:44292|issue:44912", @@ -2608,8 +2604,8 @@ "issue:43335|issue:45663", "issue:42947|issue:43381", "issue:43957|issue:44756", - "issue:44863|issue:46468", "issue:45276|issue:45292", + "issue:44863|issue:46468", "issue:45685|issue:46468", "issue:44479|issue:44877", "issue:44164|issue:44756", @@ -2618,17 +2614,18 @@ "issue:45588|issue:45663", "issue:45684|issue:45698", "issue:42796|issue:46326", - "issue:43646|issue:43708", "issue:45292|issue:45335", + "issue:43646|issue:43708", "issue:43828|issue:43957", - "issue:43645|issue:44164", "issue:44402|issue:44451", + "issue:43645|issue:44164", "issue:44589|issue:44756", "issue:43425|issue:43828", "issue:45812|issue:46315", "issue:46315|issue:46491", "issue:39404|issue:45701", "issue:43611|issue:43645", + "issue:39404|issue:46665", "issue:44683|issue:45823", "issue:44530|issue:45084", "issue:43425|issue:45823", @@ -2637,238 +2634,244 @@ "issue:43906|issue:44530", "issue:44479|issue:44805" ], - "prepared_review_unit_hash": "64a404fe7179b23d162f7f998a54bbeff5851b840df1e8f46064a2d470101b72", + "prepared_review_unit_hash": "8f2621dbeb3ee7f764d906b2ffdd558945d8e070781631384aeb51e864b64707", "component_budget": { - "node_count": 472, - "item_count": 472, - "soft_pair_count": 1732, - "serialized_chars": 898654, - "estimated_input_tokens": 224664, - "estimated_eval_tokens": 449584 + "node_count": 473, + "item_count": 473, + "soft_pair_count": 1738, + "serialized_chars": 901437, + "estimated_input_tokens": 225360, + "estimated_eval_tokens": 450976 }, "budget": { - "node_count": 47, - "item_count": 47, - "soft_pair_count": 33, - "serialized_chars": 37012, - "estimated_input_tokens": 9253, - "estimated_eval_tokens": 18762 + "node_count": 48, + "item_count": 48, + "soft_pair_count": 34, + "serialized_chars": 37876, + "estimated_input_tokens": 9469, + "estimated_eval_tokens": 19194 }, "overflow_policy": "truncate_then_skip", "trimmed": false, "aggressively_trimmed": false, "split": true, "analyst_result": { - "summary": "The cluster is mostly heterogeneous: tokenizer regressions, custom-model loading issues, quantization/device mismatches, and training bugs are all mixed together. The only clear duplicate subcluster is the `resize_token_embeddings` / output-embedding propagation bug around 45292 and 45335, with 45276 as a nearby variant.", - "confidence": 0.84, - "canonical_issue_reason": "Issue 45292 is the cleanest canonical anchor for the only clear duplicate subcluster because it describes the generic `resize_token_embeddings` \u2192 `output_embeddings` failure that also matches 45335 and overlaps 45276.", + "summary": "Most items are unrelated one-off bugs. The only clear duplicate cluster is the token-embedding resize regression around `resize_token_embeddings`; `issue:45292` is the best representative of that group.", + "confidence": 0.82, + "canonical_issue_reason": "`issue:45292` is the most general and central report in the resize-token-embeddings cluster, and the other matching reports are model-specific variants of the same symptom.", "canonical_pr_reason": null, - "best_issue_reason": "45292 is the most general, model-agnostic wording of the embedding-resize regression, so it is the best deduplication target for that subcluster.", + "best_issue_reason": "`issue:45292` is the best global issue candidate because it describes the underlying bug at the highest level (`resize_token_embeddings` not updating output embeddings), making it the most reusable canonical anchor.", "best_pr_reason": null, "soft_edge_verdicts": [ { "left": "issue:44292", "right": "issue:44912", "accept": false, - "reason": "Both involve quantization/load failures, but they affect different models and likely different code paths." + "reason": "Both are quantization/loading problems, but they involve different models, quantization formats, and failure modes; not the same bug." }, { "left": "issue:44155", "right": "issue:45823", "accept": false, - "reason": "Batched track leakage and Gemma4 device mismatch are unrelated failures." + "reason": "AudioFlamingo track leakage during batched inference is unrelated to Gemma4 PLE device placement in forward." }, { "left": "issue:43335", "right": "issue:45663", "accept": false, - "reason": "Sparse-layer config bug vs. Gemma4 FSDP attention-state bug; different underlying issues." + "reason": "One is a SwitchTransformers config edge case; the other is a Gemma4 FSDP2 attention state bug. Different code paths." }, { "left": "issue:42947", "right": "issue:43381", "accept": false, - "reason": "Gradient checkpointing ineffective with LoRA is not the same bug as checkpointing being disallowed in eval mode." + "reason": "Both mention gradient checkpointing, but one is about LoRA/PEFT effectiveness and the other is an eval-mode restriction; not the same defect." }, { "left": "issue:43957", "right": "issue:44756", "accept": false, - "reason": "Meta-device model loading and Strix Halo mmap OOM are different code paths and fixes." - }, - { - "left": "issue:44863", - "right": "issue:46468", - "accept": false, - "reason": "NemotronH checkpoint loading failure and a broken example are related by model family, but not the same concrete bug." + "reason": "Meta-device loading failures and Strix Halo mmap OOMs are different issues with different root causes." }, { "left": "issue:45276", "right": "issue:45292", "accept": true, - "reason": "Same `resize_token_embeddings` propagation bug: both say resized embeddings do not update output embeddings, with 45276 adding a Gemma4-specific variant." + "reason": "Both report `resize_token_embeddings` failing to update output/token embeddings; `45276` is a Gemma4-specific manifestation of the same bug." + }, + { + "left": "issue:44863", + "right": "issue:46468", + "accept": false, + "reason": "NemotronH checkpoint loading and a broken Nemotron example are related only by model family, not the same concrete bug." }, { "left": "issue:45685", "right": "issue:46468", "accept": false, - "reason": "MPS histogram/int error is unrelated to the Nemotron example issue." + "reason": "An MPS histogram kernel limitation in MoE is unrelated to the Nemotron example breakage." }, { "left": "issue:44479", "right": "issue:44877", "accept": false, - "reason": "Qwen VL video-input regression and Granite speech config loading are separate regressions." + "reason": "A Qwen VL video regression and a Granite speech config-loading issue are different model-specific failures." }, { "left": "issue:44164", "right": "issue:44756", "accept": false, - "reason": "Extra_state save/load handling and mmap OOM avoidance are unrelated." + "reason": "`extra_state` save/load handling is unrelated to disabling mmap on Strix Halo." }, { "left": "issue:43381", "right": "issue:44928", "accept": false, - "reason": "Eval-mode checkpointing restriction is different from the RLHF NaN/SDPA fallback regression." + "reason": "Eval-mode gradient checkpointing and Qwen3.5 SDPA math-fallback NaNs are distinct bugs." }, { "left": "issue:43854", "right": "issue:45216", "accept": false, - "reason": "GLM-4.7-Flash test-loading failure and Qwen3.5 save_pretrained regression are different problems." + "reason": "GLM-4.7-Flash unit-test loading failure and Qwen3.5 save_pretrained regression are separate issues." }, { "left": "issue:45588", "right": "issue:45663", "accept": false, - "reason": "Flash-attention sink handling and Gemma4 shared_kv_states/FSDP KeyError do not share the same failure mechanism." + "reason": "Flash-attention sink handling and Gemma4 FSDP2 shared-kv state rebuilding are different code paths." }, { "left": "issue:45684", "right": "issue:45698", "accept": false, - "reason": "Read-only permission propagation from save_pretrained is not the same as loading the wrong custom module after save_pretrained." + "reason": "Read-only permission propagation during save and wrong custom-module loading after save are not the same underlying bug." }, { "left": "issue:42796", "right": "issue:46326", "accept": false, - "reason": "Tokenizer BPE-marker handling and continue_final_message behavior are distinct tokenizer bugs." - }, - { - "left": "issue:43646", - "right": "issue:43708", - "accept": false, - "reason": "Custom-model initialization in notebooks and resume_from_checkpoint max_steps calculation are unrelated." + "reason": "Tokenizer BPE-marker handling and `continue_final_message` behavior are related to tokenization but not the same defect." }, { "left": "issue:45292", "right": "issue:45335", "accept": true, - "reason": "Same embedding-resize bug: both report that `resize_token_embeddings` fails to update downstream embedding/output weights." + "reason": "Both describe the same `resize_token_embeddings` regression failing to propagate resized embeddings to model-specific output/decoder embeddings." }, { - "left": "issue:43828", - "right": "issue:43957", + "left": "issue:43646", + "right": "issue:43708", "accept": false, - "reason": "Autocast dtype mismatch and meta-device loading breakage are different failures." + "reason": "Custom model initialization in notebooks and checkpoint-resume `max_steps` calculation are unrelated." }, { - "left": "issue:43645", - "right": "issue:44164", + "left": "issue:43828", + "right": "issue:43957", "accept": false, - "reason": "Notebook custom-model init regression and extra_state save/load handling are separate bugs." + "reason": "Autocast dtype mismatch and meta-device loading failures are different runtime problems." }, { "left": "issue:44402", "right": "issue:44451", "accept": false, - "reason": "Tokenizer vocab-size inconsistency and ScandiBERT load failure are not the same underlying issue." + "reason": "A vocab-size mismatch in one model and a separate ScandiBERT loading failure are not clearly the same bug." + }, + { + "left": "issue:43645", + "right": "issue:44164", + "accept": false, + "reason": "Notebook custom-model initialization and `extra_state` save/load handling are different issues." }, { "left": "issue:44589", "right": "issue:44756", "accept": false, - "reason": "Float8 storage lookup failure is unrelated to disabling mmap for OOM avoidance." + "reason": "Missing Float8 storage support and Strix Halo mmap OOMs are unrelated." }, { "left": "issue:43425", "right": "issue:43828", "accept": false, - "reason": "Torch 2.10 incompatibility is too broad and unrelated to the autocast dtype mismatch." + "reason": "Torch 2.10 incompatibility is broader and different from an autocast dtype-mismatch bug." }, { "left": "issue:45812", "right": "issue:46315", "accept": false, - "reason": "Wrong token IDs for Granite models and slow `convert_tokens_to_ids` are different tokenizer regressions." + "reason": "Wrong Granite token IDs and slow `convert_tokens_to_ids` are both tokenizer regressions but not the same bug." }, { "left": "issue:46315", "right": "issue:46491", "accept": false, - "reason": "Tokenizer slowdown and encode\u2192decode space stripping are different bugs." + "reason": "Tokenizer conversion slowness and CodeLlama leading-space round-trip corruption are different failures." }, { "left": "issue:39404", "right": "issue:45701", "accept": false, - "reason": "Whisper pipeline `return_language` regression is unrelated to generic tokenization changes." + "reason": "Whisper pipeline `return_language` breakage and generic tokenization changes do not look like the same underlying issue." }, { "left": "issue:43611", "right": "issue:43645", "accept": false, - "reason": "Base-model-prefix loading breakage and notebook custom-model initialization are distinct loading regressions." + "reason": "Both are about Transformers 5/custom models, but one is `base_model_prefix` loading and the other is notebook initialization; not one bug." + }, + { + "left": "issue:39404", + "right": "issue:46665", + "accept": false, + "reason": "Whisper pipeline behavior and `from_pretrained` inside Python classes are unrelated." }, { "left": "issue:44683", "right": "issue:45823", "accept": false, - "reason": "Compiled flex_attention on torch>=2.9 and Gemma4 device mismatch are unrelated." + "reason": "Compiled flex_attention on new torch versions and Gemma4 device-mismatch during forward are different problems." }, { "left": "issue:44530", "right": "issue:45084", "accept": false, - "reason": "PagedAttention linear_attention crash and non-template-node compilation error are different code paths." + "reason": "PagedAttentionCache linear-attention group handling and a compiler node-type error are unrelated." }, { "left": "issue:43425", "right": "issue:45823", "accept": false, - "reason": "Torch version incompatibility and Gemma4 device mismatch are not the same bug." + "reason": "Torch version compatibility and Gemma4 device placement are not the same bug." }, { "left": "issue:43425", "right": "issue:43716", "accept": false, - "reason": "Torch incompatibility and Mistral-3 dtype mismatch are unrelated." + "reason": "Torch compatibility and Mistral-3 image-preprocessor dtype mismatch are distinct issues." }, { "left": "issue:44589", "right": "issue:45084", "accept": false, - "reason": "Float8 storage-object lookup error is unrelated to compilation of non-template nodes." + "reason": "Float8 storage lookup failure and template-node compilation errors are separate code paths." }, { "left": "issue:43906", "right": "issue:44530", "accept": false, - "reason": "An isolated reproduction of another issue is not the same as the PagedAttention cache crash." + "reason": "An isolated reproduction of another issue and a PagedAttentionCache crash are not the same bug." }, { "left": "issue:44479", "right": "issue:44805", "accept": false, - "reason": "Video-input regression and mask/tensor shape mismatch are different failures." + "reason": "A Qwen VL video regression and a mask-shape IndexError are different failure modes." } ] }, "evaluator_result": { "accept": true, - "feedback": "The summary is grounded in the packet and stays conservative. The only clearly supported duplicate subcluster is the resize_token_embeddings/output_embeddings family around 45292 and 45335, with 45276 as a nearby Gemma4-specific variant. The other soft-edge pairs are reasonably rejected as distinct issues based on their titles and descriptions, and nothing in the writeup overstates the evidence." + "feedback": "The summary is grounded in the packet and stays conservative. The only clearly supported duplicate cluster is the resize_token_embeddings regression around issue:45292, and treating issue:45292 as the canonical issue is reasonable given the titles. The soft-edge acceptances for 45276\u219445292 and 45292\u219445335 are well supported by the packet." }, "evaluator_used": true, "retried": false, @@ -2887,9 +2890,8 @@ "source": "llm", "cache_hit": false, "model": "gpt-5.4-mini?service_tier=flex", - "cluster_id": "cluster-10000-472", + "cluster_id": "cluster-10000-473", "nodes": [ - "issue:32090", "issue:36032", "issue:39404", "issue:42796", @@ -2947,8 +2949,8 @@ "issue:44671|issue:44991", "issue:36032|issue:43452", "issue:46050|issue:46326", - "issue:44164|issue:45759", "issue:43299|issue:43931", + "issue:44164|issue:45759", "issue:43906|issue:46326", "issue:44991|issue:45081", "issue:43856|issue:44756", @@ -2957,11 +2959,10 @@ "issue:44488|issue:45081", "issue:44462|issue:46489", "issue:44488|issue:44987", - "issue:43611|issue:43644", "issue:42796|issue:43404", + "issue:43611|issue:43644", "issue:39404|issue:44671", "issue:44737|issue:45859", - "issue:32090|issue:39404", "issue:43716|issue:45084", "issue:46285|issue:46402", "issue:36032|issue:42915", @@ -2969,208 +2970,148 @@ "issue:43582|issue:43632", "issue:43643|issue:44530" ], - "prepared_review_unit_hash": "df29edfaec3618d44c570c9121f13c2357d9f965f12447765244e9cc0236c28f", + "prepared_review_unit_hash": "127a26ba89caf1705a6bb8097aca4e64935f5b393590c7176af40964da550ae2", "component_budget": { - "node_count": 472, - "item_count": 472, - "soft_pair_count": 1732, - "serialized_chars": 898654, - "estimated_input_tokens": 224664, - "estimated_eval_tokens": 449584 + "node_count": 473, + "item_count": 473, + "soft_pair_count": 1738, + "serialized_chars": 901437, + "estimated_input_tokens": 225360, + "estimated_eval_tokens": 450976 }, "budget": { - "node_count": 48, - "item_count": 48, - "soft_pair_count": 29, - "serialized_chars": 36492, - "estimated_input_tokens": 9123, - "estimated_eval_tokens": 18502 + "node_count": 47, + "item_count": 47, + "soft_pair_count": 28, + "serialized_chars": 35589, + "estimated_input_tokens": 8898, + "estimated_eval_tokens": 18052 }, "overflow_policy": "truncate_then_skip", "trimmed": false, "aggressively_trimmed": false, "split": true, "analyst_result": { - "summary": "Loose similarity cluster of mostly unrelated Transformers issues: tokenizer regressions, model-loading breaks, training/runtime incompatibilities, and a few docs mismatches. No soft pair looks like the same underlying bug/change closely enough to merge.", - "confidence": 0.89, - "canonical_issue_reason": "issue:43611 is the broadest umbrella bug here: a generic Transformers v5 model-loading regression; most others are narrower model/tokenizer-specific breakages.", + "summary": "This cluster is mostly a grab bag of Transformers v5 regressions, especially tokenizer/model-loading breakages, plus some unrelated docs, quantization, and training/runtime issues. None of the soft pairs look like true duplicates; they are only loosely similar by subsystem or failure mode.", + "confidence": 0.78, + "canonical_issue_reason": "issue:43611 is the broadest and most central representative of the recurring v5 model-loading regression theme, covering a core load-path break that many of the other issues resemble only loosely.", "canonical_pr_reason": null, - "best_issue_reason": "issue:43611 is the most suitable anchor for triage because it is broad, clearly regression-oriented, and representative of the cluster\u2019s v5 loading failures.", + "best_issue_reason": "issue:43611 best captures the shared cluster theme of Transformers v5 breaking model/tokenizer loading in downstream models.", "best_pr_reason": null, "soft_edge_verdicts": [ { "left": "issue:43901", "right": "issue:44509", "accept": false, - "reason": "Docs/API mismatch in both, but one is about return_all_scores and the other about removed v5 pipeline task names; different change surfaces." + "reason": "Both are docs regressions, but one is about `return_all_scores` in text classification and the other about removed pipeline tasks; different documentation gaps." }, { "left": "issue:44265", "right": "issue:44898", "accept": false, - "reason": "torch.export with torch_compilable_check is unrelated to Perceiver interpolation failures." + "reason": "Different failures: `torch.export` with `torch_compilable_check` versus Perceiver image classification with non-default resolution." }, { "left": "issue:43425", "right": "issue:44918", "accept": false, - "reason": "Generic torch incompatibility vs Qwen3.5 embedding unpacking; not the same code path." + "reason": "Unrelated problems: Torch 2.10 compatibility versus Qwen3.5 embedding unpacking in TRL SFT trainer." }, { "left": "issue:39404", "right": "issue:44912", "accept": false, - "reason": "Whisper pipeline return_language regression is unrelated to MXFP4 quantization fallback." + "reason": "Whisper pipeline `return_language` regression and MXFP4 quantization fallback are separate code paths." }, { "left": "issue:43994", "right": "issue:44079", "accept": false, - "reason": "Siglip2 inference corruption vs ModelOutput key assignment bug; distinct failures." + "reason": "One is a SigLIP2 output-quality regression, the other is `ModelOutput` key assignment behavior; not the same bug." }, { "left": "issue:44671", "right": "issue:44991", "accept": false, - "reason": "Both are tokenizer issues, but CamemBERT MLM predictions and EMBEDDIA tokenizer loading are different bugs." + "reason": "CamemBERT masked-LM prediction error and an EMBEDDIA tokenizer load failure are distinct tokenizer/model issues." }, { "left": "issue:36032", "right": "issue:43452", "accept": false, - "reason": "T5Tokenizer method conflict and gguf_file loading breakage are separate tokenizer load paths." + "reason": "T5 tokenizer `add_special_tokens` conflict and `gguf_file` loading failures are different tokenizer-loading bugs." }, { "left": "issue:46050", "right": "issue:46326", "accept": false, - "reason": "save/from_pretrained quantization persistence vs continue_final_message for mistral_common tokenizers are unrelated." + "reason": "Quantized model save/load corruption via `sinq` is unrelated to `continue_final_message` behavior for `mistral_common` tokenizers." }, { - "left": "issue:44164", - "right": "issue:45759", + "left": "issue:43299", + "right": "issue:43931", "accept": false, - "reason": "extra_state save/load bug and AutoModelForCausalLM config unwrapping are different serialization/loading problems." + "reason": "Both involve Qwen3-VL loading, but one is a MoE regression and the other is a weight-shape mismatch on a specific model; not clearly the same fixable bug." }, { - "left": "issue:43299", - "right": "issue:43931", + "left": "issue:44164", + "right": "issue:45759", "accept": false, - "reason": "Both involve Qwen3-VL loading, but one is a v5 regression and the other is a shape-mismatch load error; not clearly the same fix." + "reason": "`extra_state` save/load handling and `from_config` unwrapping of `text_config` are separate serialization/config issues." }, { "left": "issue:43906", "right": "issue:46326", "accept": false, - "reason": "Reproduction of issue 38071 is not the same as the mistral_common continue_final_message regression." + "reason": "The isolated reproduction of issue 38071 is not the same as the mistral_common `continue_final_message` regression." }, { "left": "issue:44991", "right": "issue:45081", "accept": false, - "reason": "Different tokenizer families and different failures: EMBEDDIA loading vs Mistral regex patching." + "reason": "Different tokenizer failures: EMBEDDIA/est-roberta loading versus `_patch_mistral_regex` crashing with missing `backend_tokenizer`." }, { "left": "issue:43856", "right": "issue:44756", "accept": false, - "reason": "MoE training memory inefficiency and Strix Halo mmap OOM are unrelated performance issues." + "reason": "Training memory inefficiency on Qwen3 MoE and Strix Halo mmap OOM are unrelated resource issues." }, { "left": "issue:43596", "right": "issue:44683", "accept": false, - "reason": "DeepSpeed zero3/BertModel index error is unrelated to compiled flex_attention on newer torch." - }, - { - "left": "issue:44164", - "right": "issue:44479", - "accept": false, - "reason": "Checkpoint serialization bug vs video-input regression for Qwen VL models; different subsystems." - }, - { - "left": "issue:44488", - "right": "issue:45081", - "accept": false, - "reason": "Loading cjvt/sleng-bert and Mistral tokenizer backend patching are distinct tokenizer problems." - }, - { - "left": "issue:44462", - "right": "issue:46489", - "accept": false, - "reason": "AutoTokenizer ignoring tokenizer.json is a different issue from DeepSeek-Coder tokenizer output drift." - }, - { - "left": "issue:44488", - "right": "issue:44987", - "accept": false, - "reason": "Both are load failures, but on different models and likely different tokenizer/config paths." - }, - { - "left": "issue:43611", - "right": "issue:43644", - "accept": false, - "reason": "Both are v5 regressions, but base_model_prefix loading and non-persistent buffer junk are separate mechanisms." - }, - { - "left": "issue:42796", - "right": "issue:43404", - "accept": false, - "reason": "Tokenizer BPE-marker handling and lm_head weight tying are unrelated Mistral-family bugs." - }, - { - "left": "issue:39404", - "right": "issue:44671", - "accept": false, - "reason": "Whisper pipeline behavior and CamemBERT masked LM predictions do not share a concrete bug." - }, - { - "left": "issue:44737", - "right": "issue:45859", - "accept": false, - "reason": "XLNet CPU arange placement and Qwen3_5Moe CPU offload incompatibility are different device-handling bugs." - }, - { - "left": "issue:32090", - "right": "issue:39404", - "accept": false, - "reason": "Trainer gpu_broadcast_one NoneType error is unrelated to Whisper return_language pipeline behavior." - }, - { - "left": "issue:43716", - "right": "issue:45084", - "accept": false, - "reason": "Image preprocessor/model dtype mismatch and template compilation TypeError are unrelated." + "reason": "Deepspeed ZeRO-3/BertModel index error is unrelated to compiled `flex_attention` failures on newer torch." }, { "left": "issue:46285", "right": "issue:46402", "accept": false, - "reason": "CLIP text-model structure inconsistency and Sam3Model text-encoder weight loading are different model-architecture regressions." + "reason": "CLIP structure inconsistency and Sam3Model text-encoder weights not loading are different model API regressions." }, { "left": "issue:36032", "right": "issue:42915", "accept": false, - "reason": "Tokenizer method conflict and Qwen3Moe FineGrainedFP8Config failure are unrelated." + "reason": "T5 tokenizer special-token conflict and Qwen3Moe FineGrainedFP8Config failure are unrelated model-loading problems." }, { "left": "issue:44661", "right": "issue:44683", "accept": false, - "reason": "Tokenizer mapping-name edge case and compiled flex_attention failure are separate issues." + "reason": "`add-new-model-like` failing in `TOKENIZER_MAPPING_NAMES` and `torch_compilable_check`/`flex_attention` are different internals." }, { "left": "issue:43582", "right": "issue:43632", "accept": false, - "reason": "Apple Silicon warmup TypeError and _is_hf_initialized flag breakage are different runtime problems." + "reason": "Apple Silicon `caching_allocator_warmup` TypeError and `_is_hf_initialized` flag breakage do not share the same code path." }, { "left": "issue:43643", "right": "issue:44530", "accept": false, - "reason": "AutoConfig trust_remote_code field loss is unrelated to PagedAttentionCache invalid group type crashes." + "reason": "`trust_remote_code` config-field loss and Qwen3.5 `PagedAttentionCache` linear_attention crash are unrelated." } ] }, @@ -3192,7 +3133,7 @@ "source": "llm", "cache_hit": false, "model": "gpt-5.4-mini?service_tier=flex", - "cluster_id": "cluster-10000-472", + "cluster_id": "cluster-10000-473", "nodes": [ "issue:28530", "issue:34567", @@ -3255,265 +3196,262 @@ "issue:44206|issue:44220", "issue:43883|issue:44488", "issue:40444|issue:46326", - "issue:43452|issue:43482", "issue:36032|issue:43645", + "issue:43452|issue:43482", "issue:43317|issue:43482", "issue:45084|issue:45507", "issue:44987|issue:44991", "issue:43262|issue:43377", "issue:43746|issue:46326", - "issue:43646|issue:44534", "issue:44464|issue:45987", "issue:42915|issue:45084", + "issue:43646|issue:44534", "issue:43883|issue:44991", "issue:44403|issue:45081", "issue:43645|issue:44534", "issue:43404|issue:44466", "issue:43377|issue:43819", - "issue:44849|issue:45325", "issue:43994|issue:44610", "issue:28530|issue:34567", + "issue:44849|issue:45325", "issue:44451|issue:45081", "issue:44987|issue:45081", "issue:42915|issue:44291", - "issue:44265|issue:44610", - "issue:43322|issue:43540" + "issue:43322|issue:43540", + "issue:44265|issue:44610" ], - "prepared_review_unit_hash": "3e12e64329e84ffedf73a9cb201fa10c849ffc753cbd2b7305b1295750b7316f", + "prepared_review_unit_hash": "c4887b2e36ab5e5176d57f7c75abfeac0b7dd81beda10a7d22f796b2171aa300", "component_budget": { - "node_count": 472, - "item_count": 472, - "soft_pair_count": 1732, - "serialized_chars": 898654, - "estimated_input_tokens": 224664, - "estimated_eval_tokens": 449584 + "node_count": 473, + "item_count": 473, + "soft_pair_count": 1738, + "serialized_chars": 901437, + "estimated_input_tokens": 225360, + "estimated_eval_tokens": 450976 }, "budget": { "node_count": 48, "item_count": 48, "soft_pair_count": 34, - "serialized_chars": 38066, - "estimated_input_tokens": 9517, - "estimated_eval_tokens": 19290 + "serialized_chars": 38074, + "estimated_input_tokens": 9519, + "estimated_eval_tokens": 19294 }, "overflow_policy": "truncate_then_skip", "trimmed": false, "aggressively_trimmed": false, "split": true, "analyst_result": { - "summary": "Mostly heterogeneous issue cluster. The only plausible near-duplicate theme is the Transformers v5 buffer/init regression around 43644, but the custom-model-init complaints are not evidenced well enough to merge with it, so I would not accept the soft similarity edges here.", - "confidence": 0.82, - "canonical_issue_reason": "43644 is the clearest and most concrete report in the cluster: it directly describes the non-persistent-buffer regression in Transformers v5. The other Transformer v5 initialization reports are related in wording but still look like distinct failures.", + "summary": "The cluster is mostly noisy and spans many unrelated bugs; several items are generic Transformers v5/model-loading regressions, but none of the soft pairs are strong enough to merge as duplicates.", + "confidence": 0.78, + "canonical_issue_reason": "issue:44987 is the most representative of the recurring v5 loading-regression theme in this set: it is a concrete, version-bounded model-load failure and is broader than the more niche audio/training/model-architecture reports.", "canonical_pr_reason": null, - "best_issue_reason": "43644 is the best standalone representative because it names a specific regression, has the broadest duplicate-worthy wording, and is the most likely anchor for the buffer-related bug class in this cluster.", + "best_issue_reason": "issue:44987 is the best cluster representative because it captures the common model/tokenizer loading regression pattern better than the other, more specialized issues.", "best_pr_reason": null, "soft_edge_verdicts": [ { "left": "issue:40990", "right": "issue:43421", "accept": false, - "reason": "Different bugs: perplexity/model behavior on one side, tokenizer backend post-processor refresh on the other." + "reason": "Perplexity-quality complaint vs tokenizer post-processor update bug; different symptoms and code paths." }, { "left": "issue:43541", "right": "issue:45084", "accept": false, - "reason": "Both are runtime/compilation failures, but they point to different code paths and different concrete errors." + "reason": "Torch Dynamo grouped_mm tracing failure vs template-node compile error; unrelated failures." }, { "left": "issue:44451", "right": "issue:44987", "accept": false, - "reason": "Both are loading regressions, but the affected models and failure modes are not shown to be the same bug." + "reason": "Both are load failures, but they target different models and no shared root cause is shown." }, { "left": "issue:42915", "right": "issue:44530", "accept": false, - "reason": "Different model families and different failures: FineGrainedFP8Config vs PagedAttentionCache linear_attention crash." + "reason": "Qwen3Moe FP8 config failure vs Qwen3.5 paged-attention linear_attention crash; different defects." }, { "left": "issue:34567", "right": "issue:40444", "accept": false, - "reason": "Unrelated issues: token-count tracking vs multi-image IterableDataset finetuning failure." + "reason": "Trainer token counter update bug vs multimodal iterable-dataset failure; unrelated." }, { "left": "issue:42175", "right": "issue:46178", "accept": false, - "reason": "Packaging/backend inclusion issue vs backend-mapping validation error; not the same underlying problem." + "reason": "Packaging/backend inclusion issue vs backend-mapping error for tensorflow_text; distinct bugs." }, { "left": "issue:43644", "right": "issue:43646", "accept": false, - "reason": "The titles overlap around Transformers v5 initialization, but one is about junk-filled non-persistent buffers and the other is a broader custom-model-initialization complaint; not enough evidence they are the same concrete bug." + "reason": "Both are v5 regressions, but non-persistent-buffer junk and custom model initialization are separate problems." }, { "left": "issue:44811", "right": "issue:45584", "accept": false, - "reason": "Both involve Whisper token handling, but the failure modes are different and not clearly the same bug." + "reason": "Whisper batch_decode skip_special_tokens issue vs empty-transcription handling after align_special_tokens; different code paths." }, { "left": "issue:44206", "right": "issue:44220", "accept": false, - "reason": "Related audio feature extraction area, but the reported crashes differ and there is no concrete shared code-path evidence." + "reason": "Both touch audio feature extraction, but one is a center-arg regression and the other is a lower-level fbank helper issue; not clearly the same bug." }, { "left": "issue:43883", "right": "issue:44488", "accept": false, - "reason": "Missing tied-weights metadata vs a specific model loading failure; different issues." + "reason": "Missing all_tied_weights_keys attribute vs specific model-loading failure for cjvt/sleng-bert; unrelated." }, { "left": "issue:40444", "right": "issue:46326", "accept": false, - "reason": "Different modalities and failures: multimodal finetuning with multiple images vs tokenizer continue_final_message behavior." + "reason": "Multimodal iterable dataset with multiple images vs mistral_common continue_final_message tokenizer bug; unrelated." }, { - "left": "issue:43452", - "right": "issue:43482", + "left": "issue:36032", + "right": "issue:43645", "accept": false, - "reason": "Both are GGUF/loading related, but the reports describe different concrete regressions and models." + "reason": "Tokenizer method-name conflict vs Jupyter/custom-model initialization regression; different defects." }, { - "left": "issue:36032", - "right": "issue:43645", + "left": "issue:43452", + "right": "issue:43482", "accept": false, - "reason": "Tokenizer API conflict vs Jupyter/custom-model initialization regression; different code paths." + "reason": "Both involve GGUF loading, but one is a generic API break and the other is a specific Qwen2.5-GGUF regression; not the same underlying bug." }, { "left": "issue:43317", "right": "issue:43482", "accept": false, - "reason": "Dequantized model offload loading failure vs GGUF loading failure; not the same issue." + "reason": "gpu+cpu offload dequantized-model loading failure vs GGUF loading failure; different failure modes." }, { "left": "issue:45084", "right": "issue:45507", "accept": false, - "reason": "Different invalid-method/compile failures in different model code paths." + "reason": "Template compilation error vs GraniteMoEHybrid invalid-method call; no shared code-path shown." }, { "left": "issue:44987", "right": "issue:44991", "accept": false, - "reason": "Version-related loading failures, but for different models and without evidence of the same root cause." + "reason": "Both are versioned load failures, but they involve different models/tokenizer paths and there is no evidence of one shared root cause." }, { "left": "issue:43262", "right": "issue:43377", "accept": false, - "reason": "Audio chat-template sampling-rate defaulting vs MIMI batch/single padding-mask mismatch; unrelated bugs." + "reason": "Audio processor sampling-rate default bug vs MIMI batched-vs-single output mismatch; distinct issues." }, { "left": "issue:43746", "right": "issue:46326", "accept": false, - "reason": "PEFT/local checkpoint loading vs tokenizer chat-message handling; different regressions." + "reason": "PEFT checkpoint loading problem vs mistral_common continue_final_message bug; unrelated." }, { "left": "issue:44464", "right": "issue:45987", "accept": false, - "reason": "Chunked generation inconsistency is a different bug from StaticCache.get_seq_length() returning the wrong type." + "reason": "Compiled-forward chunked-generation inconsistency vs StaticCache seq-length return-type bug; different problems." }, { "left": "issue:42915", - "right": "issue:45084", + "right": "issue:44291", + "accept": false, + "reason": "Qwen3Moe FP8 failure vs init_empty_weights unexpected-argument error; unrelated." + }, + { + "left": "issue:43646", + "right": "issue:44534", "accept": false, - "reason": "Different failures: FP8/model setup vs template-node compilation." + "reason": "Custom model initialization regression vs non-persistent buffer serialization junk; separate bugs." }, { "left": "issue:43883", "right": "issue:44991", "accept": false, - "reason": "Missing all_tied_weights_keys vs tokenizer loading failure for a specific model; unrelated." + "reason": "Tied-weights attribute error vs tokenizer load failure for a specific model; different issues." }, { "left": "issue:44403", "right": "issue:45081", "accept": false, - "reason": "General loading noise vs a specific Mistral regex patch crash." + "reason": "Generic loading noise vs Mistral regex patch crash; not the same defect." }, { "left": "issue:43645", "right": "issue:44534", "accept": false, - "reason": "Custom model initialization in notebooks is a different regression from junk-filled non-persistent buffers." + "reason": "Notebook/custom-model initialization problem vs buffer-junk regression; distinct bugs." }, { "left": "issue:43404", "right": "issue:44466", "accept": false, - "reason": "Tied-weight bug in one model class vs inconsistent lm_head serialization; related theme, but not the same concrete bug." + "reason": "Both involve lm_head/tied weights, but one is missing tying in Mistral3 and the other is device-dependent serialization inconsistency; separate bugs." }, { "left": "issue:43377", "right": "issue:43819", "accept": false, - "reason": "Missing padding-mask support in MIMI encoder vs DAC.from_latents/STE mismatch; different code paths." - }, - { - "left": "issue:44849", - "right": "issue:45325", - "accept": false, - "reason": "Different Qwen-family bugs: output_hidden_states behavior vs rope-index scaling for still images." + "reason": "MIMI padding-mask bug vs DAC.from_latents/STE mismatch; unrelated audio-model issues." }, { "left": "issue:43994", "right": "issue:44610", "accept": false, - "reason": "Nonsensical outputs from a specific image model vs processor/model input-size mismatch; not the same bug." + "reason": "SigLIP2 nonsensical outputs vs OmDet-Turbo processor size mismatch; different root causes." }, { "left": "issue:28530", "right": "issue:34567", "accept": false, - "reason": "Early stopping metric lookup vs token-count tracking; unrelated training bugs." + "reason": "Early-stopping metric selection issue vs num_input_tokens_seen not updating; unrelated trainer-state bugs." + }, + { + "left": "issue:44849", + "right": "issue:45325", + "accept": false, + "reason": "Qwen3.5 hidden-states bug vs Qwen2.5-VL rope-index scaling bug; different model internals." }, { "left": "issue:44451", "right": "issue:45081", "accept": false, - "reason": "Different loading failures: specific model cannot load vs Mistral regex patch crash." + "reason": "Model loading failure vs Mistral tokenizer regex-patch crash; different code paths and symptoms." }, { "left": "issue:44987", "right": "issue:45081", "accept": false, - "reason": "Tokenizer/model loading regression vs Mistral regex patch crash; different issues." + "reason": "Versioned model-loading regression vs tokenizer regex-patch crash; not the same underlying bug." }, { - "left": "issue:42915", - "right": "issue:44291", + "left": "issue:43322", + "right": "issue:43540", "accept": false, - "reason": "FP8/Qwen3Moe failure vs init_empty_weights unexpected-argument error; different bugs." + "reason": "Llava Next segfault vs Qwen3OmniMoe video-input ValueError; unrelated multimodal failures." }, { "left": "issue:44265", "right": "issue:44610", "accept": false, - "reason": "torch.export / torch_compilable_check failure is unrelated to the OmDet-Turbo processor image-size mismatch." - }, - { - "left": "issue:43322", - "right": "issue:43540", - "accept": false, - "reason": "Segmentation fault while loading Llava Next vs video-input ValueError in Qwen3OmniMoe; different models and failure modes." + "reason": "torch.export export failure with torch_compilable_check vs processor/model image-size mismatch; separate issues." } ] }, - "evaluator_result": { - "accept": false, - "feedback": "Most of the pairwise rejections are conservative and grounded in the titles. The main issue is the accepted 43646 \u2194 44534 pair: the packet only shows high textual similarity, but not enough evidence that \"Transformers 5.0.0 breaks custom model initialization\" is the same bug as \"fills non-persistent buffers with junk.\" That acceptance overstates the packet, so I\u2019d reject this cluster evaluation." - }, - "evaluator_used": true, - "retried": true, + "evaluator_result": null, + "evaluator_used": false, + "retried": false, "accepted_nontrivial_soft_edge": false, "error_kind": null, "error_message": null @@ -3529,7 +3467,7 @@ "source": "llm", "cache_hit": false, "model": "gpt-5.4-mini?service_tier=flex", - "cluster_id": "cluster-10000-472", + "cluster_id": "cluster-10000-473", "nodes": [ "issue:33666", "issue:35707", @@ -3585,47 +3523,47 @@ "issue:36032|issue:43742", "issue:42915|issue:44589", "issue:44062|issue:44987", - "issue:44479|issue:45325", "issue:45216|issue:45357", - "issue:43531|issue:44877", + "issue:44479|issue:45325", "issue:44991|issue:46315", + "issue:43531|issue:44877", "issue:36032|issue:41720", "issue:44220|issue:44683", "issue:43425|issue:44403", - "issue:43611|issue:44534", "issue:43296|issue:43531", + "issue:43611|issue:44534", "issue:40444|issue:45325", "issue:43632|issue:44756", "issue:45812|issue:46491", - "issue:42796|issue:43746", "issue:44403|issue:44991", + "issue:42796|issue:43746", "issue:43792|issue:44589", "issue:33666|issue:40444", "issue:43381|issue:45923", - "issue:44079|issue:45663", "issue:44291|issue:44589", + "issue:44079|issue:45663", "issue:40990|issue:44811", "issue:43632|issue:44534", - "issue:35707|issue:40444", "issue:43482|issue:43525", - "issue:43421|issue:44568", + "issue:35707|issue:40444", "issue:43482|issue:45538", + "issue:43421|issue:44568", "issue:43317|issue:43856" ], - "prepared_review_unit_hash": "bc7e1969680cef7340c3e1afb48c5c22314c65b8794bccce476df5e695bfcaa7", + "prepared_review_unit_hash": "37903790f1f3cf30625d376a5605a3748698aa6f6e7b6d2ef6e0aad8061a976d", "component_budget": { - "node_count": 472, - "item_count": 472, - "soft_pair_count": 1732, - "serialized_chars": 898654, - "estimated_input_tokens": 224664, - "estimated_eval_tokens": 449584 + "node_count": 473, + "item_count": 473, + "soft_pair_count": 1738, + "serialized_chars": 901437, + "estimated_input_tokens": 225360, + "estimated_eval_tokens": 450976 }, "budget": { "node_count": 48, "item_count": 48, "soft_pair_count": 30, - "serialized_chars": 36550, + "serialized_chars": 36549, "estimated_input_tokens": 9138, "estimated_eval_tokens": 18532 }, @@ -3634,201 +3572,198 @@ "aggressively_trimmed": false, "split": true, "analyst_result": { - "summary": "Mostly heterogeneous cluster. The Qwen3.5 save_pretrained items are related in theme, but the packet only supports a soft similarity, not a confirmed duplicate/same code-path match, so that edge should stay rejected.", - "confidence": 0.81, - "canonical_issue_reason": "issue:45216 is the best canonical issue candidate because it is the earlier, clearer report about Qwen3.5 `save_pretrained` checkpoint corruption/regression. It should be treated as the representative issue for that theme, but not as a confirmed duplicate of issue:45357.", + "summary": "The cluster is mostly heterogeneous: tokenizer regressions, Qwen/Qwen3 MoE issues, generation/cache bugs, and unrelated loading/config errors. I don\u2019t see any soft pair that clearly describes the same concrete bug/change, so I reject all soft edges. For a representative issue, 45812 is the strongest general regression example, but the cluster does not really have one true canonical bug.", + "confidence": 0.84, + "canonical_issue_reason": "issue:45812 is the clearest high-signal regression in the set and has the broadest, most explicit symptom description, but it is only a representative choice because the cluster is not actually cohesive.", "canonical_pr_reason": null, - "best_issue_reason": "issue:45216 is the strongest single issue to keep as the representative of the Qwen3.5 `save_pretrained` regression theme, with the cleanest problem statement and earliest report in this cluster.", + "best_issue_reason": "issue:45812 has the strongest standalone bug report quality: explicit regression, clear affected models, and a concrete incorrect-result symptom.", "best_pr_reason": null, "soft_edge_verdicts": [ { "left": "issue:43089", "right": "issue:45987", "accept": false, - "reason": "Performance-related wording only; one is generation overhead, the other is a StaticCache return-type issue. Different bugs and code paths." + "reason": "Generation throughput overhead and a StaticCache return-type bug are different problems; they do not look mergeable into one fix." }, { "left": "issue:36032", "right": "issue:43742", "accept": false, - "reason": "Tokenizer method conflict vs model-loading key error. No same underlying defect." + "reason": "Tokenizer method-name conflict vs MobileLLM loading KeyError are unrelated failure modes." }, { "left": "issue:42915", "right": "issue:44589", "accept": false, - "reason": "Both touch FP8/v5-era errors, but the failure modes and affected paths are different." + "reason": "Both mention FP8/floating-point issues, but one is a model/config failure and the other is a storage-type error; not the same code-path bug." }, { "left": "issue:44062", "right": "issue:44987", "accept": false, - "reason": "Both are tokenizer-loading failures, but for different models and different root causes." - }, - { - "left": "issue:44479", - "right": "issue:45325", - "accept": false, - "reason": "Related Qwen-VL area, but broad video regression vs a specific rope-index scaling bug; not the same change or bug." + "reason": "AddedToken keyword-argument conflict vs loading a specific model are distinct tokenizer/load issues." }, { "left": "issue:45216", "right": "issue:45357", "accept": false, - "reason": "Close thematic overlap around Qwen3.5 save_pretrained, but the packet marks this as soft similarity only and the reports differ in symptom/version framing; not enough to treat as the same bug/code path." + "reason": "Both concern Qwen3.5 save_pretrained, but one is a broad checkpoint corruption report and the other is specifically incorrect visual-encoder keys; related, not clearly identical." }, { - "left": "issue:43531", - "right": "issue:44877", + "left": "issue:44479", + "right": "issue:45325", "accept": false, - "reason": "Different problems: Qwen3-MoE sliding_window behavior vs GraniteSpeech config loading strictness." + "reason": "Video-input regression and still-image temporal position-id scaling are different multimodal bugs." }, { "left": "issue:44991", "right": "issue:46315", "accept": false, - "reason": "Tokenizer regression vs tokenizer slowdown. Same subsystem, different defect." + "reason": "Tokenizer loading failure vs slow convert_tokens_to_ids regression are different symptoms and likely different code paths." + }, + { + "left": "issue:43531", + "right": "issue:44877", + "accept": false, + "reason": "Qwen3-MoE sliding_window behavior and granite_speech strict-config loading are unrelated." }, { "left": "issue:36032", "right": "issue:41720", "accept": false, - "reason": "Tokenizer attribute conflict vs Qwen3 auto device-mapping CUDA assert; unrelated." + "reason": "Tokenizer attribute conflict vs Qwen3 auto-device-map cuda assert are unrelated." }, { "left": "issue:44220", "right": "issue:44683", "accept": false, - "reason": "Feature extraction bug vs compiled flex_attention failure on newer torch; different code paths." + "reason": "Audio feature extraction bug vs compiled flex_attention torch-version failure are unrelated." }, { "left": "issue:43425", "right": "issue:44403", "accept": false, - "reason": "Torch compatibility issue vs loading noise regression; not the same bug." + "reason": "Torch 2.10 compatibility and loading noise are different issues." }, { - "left": "issue:43611", - "right": "issue:44534", + "left": "issue:43296", + "right": "issue:43531", "accept": false, - "reason": "Different regressions: `base_model_prefix` loading vs non-persistent buffers filled with junk." + "reason": "PaddleOCR-VL load failure and Qwen3-MoE sliding_window issue do not share the same bug." }, { - "left": "issue:43296", - "right": "issue:43531", + "left": "issue:43611", + "right": "issue:44534", "accept": false, - "reason": "PaddleOCR-VL load failure vs Qwen3-MoE sliding-window issue; unrelated." + "reason": "base_model_prefix loading regression and non-persistent buffer junk are different transformer-v5 regressions." }, { "left": "issue:40444", "right": "issue:45325", "accept": false, - "reason": "Both are Qwen-VL-family issues, but multi-image finetuning failure is not the same as still-image temporal position scaling." + "reason": "Multi-image IterableDataset finetuning failure vs Qwen2.5-VL rope-index scaling bug are unrelated." }, { "left": "issue:43632", "right": "issue:44756", "accept": false, - "reason": "`_is_hf_initialized` regression vs mmap/OOM on Strix Halo; different symptoms and causes." + "reason": "_is_hf_initialized flag handling and mmap-on-Strix-Halo OOM are unrelated." }, { "left": "issue:45812", "right": "issue:46491", "accept": false, - "reason": "Tokenizer regressions in different models with different behavior changes; not mergeable as one bug." + "reason": "Both are tokenizer regressions, but one is Granite token IDs and the other is CodeLlama space stripping; different model-specific bugs." }, { - "left": "issue:42796", - "right": "issue:43746", + "left": "issue:44403", + "right": "issue:44991", "accept": false, - "reason": "Tokenizer BPE-marker handling vs checkpoint loading with PEFT adapters; unrelated." + "reason": "Unnecessary loading noise and tokenizer-loading failure are not the same problem." }, { - "left": "issue:44403", - "right": "issue:44991", + "left": "issue:42796", + "right": "issue:43746", "accept": false, - "reason": "General model-loading noise vs a specific tokenizer-loading regression." + "reason": "BPE-marker handling and GraniteSpeech PEFT checkpoint loading are unrelated." }, { "left": "issue:43792", "right": "issue:44589", "accept": false, - "reason": "Whisper runtime failure vs Float8 storage lookup error; unrelated." + "reason": "Whisper runtime failure and Float8 storage lookup error are different issues." }, { "left": "issue:33666", "right": "issue:40444", "accept": false, - "reason": "Both are training-related and VL-adjacent, but multi-GPU training is not the same bug as multi-image IterableDataset finetuning failure." + "reason": "Multi-GPU training and multi-image iterable-dataset finetuning are not the same bug." }, { "left": "issue:43381", "right": "issue:45923", "accept": false, - "reason": "Eval-mode checkpointing restriction vs missing `supports_gradient_checkpointing` flag; different issues." + "reason": "Eval-mode gradient checkpointing limitation and missing trust_remote_code support flag are related only at a very high level, not the same concrete issue." }, { - "left": "issue:44079", - "right": "issue:45663", + "left": "issue:44291", + "right": "issue:44589", "accept": false, - "reason": "ModelOutput key assignment bug vs Gemma4 FSDP2 `shared_kv_states` KeyError; different code paths." + "reason": "Unexpected _is_hf_initialized argument and Float8 storage object errors are different transformer-v5 breakages." }, { - "left": "issue:44291", - "right": "issue:44589", + "left": "issue:44079", + "right": "issue:45663", "accept": false, - "reason": "Both are Transformers v5 loading issues, but the concrete error mechanisms differ." + "reason": "ModelOutput key assignment and Gemma4 FSDP2 shared_kv_states KeyError are unrelated." }, { "left": "issue:40990", "right": "issue:44811", "accept": false, - "reason": "Perplexity problem on GPT-OSS vs Whisper `batch_decode` special-token handling; unrelated." + "reason": "High perplexity and Whisper batch_decode skip_special_tokens behavior are unrelated." }, { "left": "issue:43632", "right": "issue:44534", "accept": false, - "reason": "Different v5 regressions: initialization flag breakage vs non-persistent buffer corruption." - }, - { - "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": "Different v5 regressions affecting different initialization/buffer behaviors." }, { "left": "issue:43482", "right": "issue:43525", "accept": false, - "reason": "Different loading failures for different model families; no shared bug." + "reason": "Qwen2.5-GGUF loading under transformers v5 and Llama4Config pad_token_id error are unrelated." }, { - "left": "issue:43421", - "right": "issue:45538", + "left": "issue:35707", + "right": "issue:40444", "accept": false, - "reason": "Runtime post-processor update bug vs `model_max_length` constant behavior; not the same defect." + "reason": "Progressive generation with inputs_embeds/past_key_values and multi-image finetuning are different bugs." }, { "left": "issue:43482", "right": "issue:45538", "accept": false, - "reason": "No shared underlying bug: GGUF loading failure vs CLIPTokenizer max-length issue." + "reason": "GGUF loading failure and CLIPTokenizer model_max_length are unrelated tokenizer/model-loading issues." + }, + { + "left": "issue:43421", + "right": "issue:44568", + "accept": false, + "reason": "Runtime special-token update behavior and mdeberta BOS/EOS add_special_tokens regression are different tokenizer bugs." }, { "left": "issue:43317", "right": "issue:43856", "accept": false, - "reason": "Dequantized-model offload loading failure vs Qwen3 MoE training memory inefficiency; different problems." + "reason": "gpu+cpu offload dequantized-model loading and Qwen3 MoE training memory usage are not the same failure." } ] }, - "evaluator_result": { - "accept": false, - "feedback": "The summary is mostly conservative, but it overstates the Qwen3.5 checkpoint pair. In the packet, issue:45216 and issue:45357 are only soft-similar and the pair is explicitly marked deterministic_accept=false, so calling them the same bug / same code path / closest accepted duplicate is too strong. Because the only true soft-edge verdict contradicts the packet\u2019s label, I can\u2019t accept this cluster evaluation." - }, - "evaluator_used": true, - "retried": true, + "evaluator_result": null, + "evaluator_used": false, + "retried": false, "accepted_nontrivial_soft_edge": false, "error_kind": null, "error_message": null @@ -3844,7 +3779,7 @@ "source": "llm", "cache_hit": false, "model": "gpt-5.4-mini?service_tier=flex", - "cluster_id": "cluster-10000-472", + "cluster_id": "cluster-10000-473", "nodes": [ "issue:40444", "issue:41553", @@ -3899,8 +3834,8 @@ "issue:41553|issue:43582", "issue:42175|issue:45542", "issue:43531|issue:44291", - "issue:44079|issue:44534", "issue:44829|issue:45479", + "issue:44079|issue:44534", "issue:44462|issue:45812", "issue:41720|issue:45081", "issue:43818|issue:46402", @@ -3919,232 +3854,232 @@ "issue:45325|issue:45834", "issue:42915|issue:43482", "issue:43632|issue:45072", - "issue:43701|issue:45325", "issue:42796|issue:45381", + "issue:43701|issue:45325", "issue:44291|issue:44877", "issue:43873|issue:45701", "issue:43296|issue:45800", "issue:44479|issue:45823", "issue:43606|issue:44265" ], - "prepared_review_unit_hash": "6a440cdf8c62218a3a93f5ac4ef201de06fec711fd091ca4cbb4d492dd9d6ecd", + "prepared_review_unit_hash": "1d02e3e12caccb9d9fefc0886009890045f18b1e1203714df6a192d332fcd9c4", "component_budget": { - "node_count": 472, - "item_count": 472, - "soft_pair_count": 1732, - "serialized_chars": 898654, - "estimated_input_tokens": 224664, - "estimated_eval_tokens": 449584 + "node_count": 473, + "item_count": 473, + "soft_pair_count": 1738, + "serialized_chars": 901437, + "estimated_input_tokens": 225360, + "estimated_eval_tokens": 450976 }, "budget": { "node_count": 48, "item_count": 48, "soft_pair_count": 30, - "serialized_chars": 36840, - "estimated_input_tokens": 9210, - "estimated_eval_tokens": 18676 + "serialized_chars": 36848, + "estimated_input_tokens": 9212, + "estimated_eval_tokens": 18680 }, "overflow_policy": "truncate_then_skip", "trimmed": false, "aggressively_trimmed": false, "split": true, "analyst_result": { - "summary": "The cluster is still mostly heterogeneous. `issue:43644` remains the best anchor for the non-persistent-buffer regression family, but the soft edges are too speculative overall: `44829` vs `45479` are different classification bugs, and `45325` vs `45381` are related Qwen2.5-VL regressions but not a conservative duplicate merge.", - "confidence": 0.89, - "canonical_issue_reason": "`issue:43644` is the cleanest canonical issue anchor: it is a narrow, well-defined v5 regression with a near-exact duplicate signal in `issue:44534`, unlike the broader and less certain pairs in this cluster.", + "summary": "The cluster is mostly a grab bag of unrelated issues, with one likely duplicate pair around Qwen2.5-VL video/vision position-id handling. Most other soft-similarity pairs are only loosely related by subsystem and should stay separate.", + "confidence": 0.49, + "canonical_issue_reason": "issue:44479 is the best umbrella issue: it describes the broad Qwen multimodal video-input regression across multiple model variants, making it the most useful canonical target for the Qwen position-id bug family.", "canonical_pr_reason": null, - "best_issue_reason": "`issue:43644` is the strongest issue candidate because it is specific, self-contained, and tied to a clear duplicate family; the other issues are more model-specific, broader, or only superficially similar.", + "best_issue_reason": "issue:44479 is the strongest single representative because it captures the broadest user-facing regression and can subsume narrower Qwen2.5-VL position-id reports.", "best_pr_reason": null, "soft_edge_verdicts": [ { "left": "issue:41553", "right": "issue:43582", "accept": false, - "reason": "Tokenizer error messaging and AppleSilicon allocator TypeError are different bugs and different code paths." + "reason": "Different failures: tokenizer loading error message vs Apple Silicon allocator TypeError." }, { "left": "issue:42175", "right": "issue:45542", "accept": false, - "reason": "Both involve TensorFlow packaging/backend availability, but they describe different failure modes." + "reason": "Both involve TF-related packaging, but one is missing TF in the torch extra and the other is a tensorboard-without-TF backend error." }, { "left": "issue:43531", "right": "issue:44291", "accept": false, - "reason": "Qwen3-MoE sliding-window behavior is unrelated to the `init_empty_weights` / `_is_hf_initialized` loading error." + "reason": "Unrelated: Qwen3-MoE sliding-window behavior vs init_empty_weights/_is_hf_initialized argument handling." }, { - "left": "issue:44079", - "right": "issue:44534", + "left": "issue:44829", + "right": "issue:45479", "accept": false, - "reason": "ModelOutput key assignment and non-persistent buffer junk are separate regressions." + "reason": "Both are classification-training regressions, but the concrete bugs differ: flash_attention_3 vs num_labels=1 loss degeneracy." }, { - "left": "issue:44829", - "right": "issue:45479", + "left": "issue:44079", + "right": "issue:44534", "accept": false, - "reason": "These are different classification bugs: one is flash_attention_3 causing degenerate training, the other is `num_labels=1` zero-loss behavior across models." + "reason": "Different code paths: ModelOutput key assignment vs non-persistent buffer serialization." }, { "left": "issue:44462", "right": "issue:45812", "accept": false, - "reason": "Both are tokenizer-related, but one is ignoring `tokenizer.json` while the other is wrong Granite token IDs." + "reason": "Both concern AutoTokenizer, but one ignores tokenizer.json while the other returns wrong Granite token IDs." }, { "left": "issue:41720", "right": "issue:45081", "accept": false, - "reason": "Qwen3 auto device mapping CUDA assert and Mistral regex patch tokenizer crash are unrelated." + "reason": "Different subsystems and bugs: device-map cuda assert vs Mistral regex patch crashing in tokenizer loading." }, { "left": "issue:43818", "right": "issue:46402", "accept": false, - "reason": "Video-LLaVA vision-tower/temporal-attention bug and Sam3Model text-encoder weight loading are different problems." + "reason": "Different models and failures: Video-LLaVA weight/temporal-attention issue vs Sam3Model text-encoder weights not loading." }, { "left": "issue:43377", "right": "issue:45491", "accept": false, - "reason": "Different models and likely different root causes: MIMI padding-mask batching bug vs Gemma3 mixed-length sliding-window NaNs." + "reason": "Both are batching/padding problems, but the specific models and failure modes are different." }, { "left": "issue:43818", "right": "issue:45357", "accept": false, - "reason": "Video-LLaVA architecture issue and Qwen3.5 save_pretrained visual-key regression are unrelated." + "reason": "Unrelated model families: Video-LLaVA architecture bug vs Qwen3.5 visual-encoder save_pretrained regression." }, { "left": "issue:43611", "right": "issue:43646", "accept": false, - "reason": "Both are v5 breakages, but `base_model_prefix` loading and custom model initialization are not the same concrete bug." + "reason": "Both are v5 initialization regressions, but not the same concrete bug or code path." }, { "left": "issue:43644", "right": "issue:44079", "accept": false, - "reason": "Non-persistent buffer initialization and ModelOutput key assignment are separate regressions." + "reason": "Different bugs: non-persistent buffer junk vs ModelOutput key assignment." }, { "left": "issue:44265", "right": "issue:45072", "accept": false, - "reason": "`torch.export` / `torch_compilable_check` failures are unrelated to bfloat16 dtype mismatches." + "reason": "Unrelated: torch.export with torch_compilable_check vs bfloat16 dtype mismatch in inference." }, { "left": "issue:43749", "right": "issue:44079", "accept": false, - "reason": "FSDP CPU RAM efficient loading and ModelOutput key assignment do not share the same bug." + "reason": "Different areas: FSDP CPU RAM efficient loading vs ModelOutput key assignment." }, { "left": "issue:42175", "right": "issue:43296", "accept": false, - "reason": "Package/backend omission and PaddleOCR-VL loading failure are different issues." + "reason": "Both mention backend/model loading, but one is TF packaging and the other is PaddleOCR-VL loading in vLLM." }, { "left": "issue:40444", "right": "issue:45381", "accept": false, - "reason": "Multi-image IterableDataset finetuning failure is unrelated to the Qwen2.5-VL video position-id regression." + "reason": "Both involve Qwen2.5-VL multimodal inputs, but one is finetuning with multiple images per prompt and the other is a video-position-id regression." }, { "left": "issue:45325", "right": "issue:45381", - "accept": false, - "reason": "Both are Qwen2.5-VL position-id regressions, but one is still-image temporal scaling and the other is video-input `vision_position_ids`; related, but not a conservative duplicate." + "accept": true, + "reason": "Same underlying Qwen2.5-VL positional-indexing bug: both describe incorrect vision/temporal position_ids in the same multimodal code path." }, { "left": "issue:43994", "right": "issue:44898", "accept": false, - "reason": "SIGLIP2 nonsensical outputs and Perceiver non-default-resolution failure are different vision-model bugs." + "reason": "Different vision-model failures: SigLIP2 nonsensical outputs vs Perceiver resize/interpolate_pos_encoding behavior." }, { "left": "issue:43792", "right": "issue:43957", "accept": false, - "reason": "Whisper loading failure and meta-device loading breakage are not the same bug." + "reason": "Different causes: Whisper model runtime failure vs meta-device model loading regressions." }, { "left": "issue:43454", "right": "issue:44466", "accept": false, - "reason": "AyaVision weight-tying failure and device-dependent `lm_head.weight` serialization are distinct issues." + "reason": "Related to lm_head weights, but one is failing to tie weights and the other is inconsistent serialization." }, { "left": "issue:45325", "right": "issue:45834", "accept": false, - "reason": "Qwen2.5-VL rope-index regression and Kosmos2.5 long OCR index error are unrelated." + "reason": "Both are OCR/vision-adjacent, but the bugs are different: Qwen2.5-VL position ids vs Kosmos2.5 long-input indexing." }, { "left": "issue:42915", "right": "issue:43482", "accept": false, - "reason": "Qwen3Moe FP8 config failure and Qwen2.5 GGUF loading under v5 are different model-loading bugs." + "reason": "Different model/load paths: Qwen3Moe FP8 config vs Qwen2.5-GGUF loading under v5." }, { "left": "issue:43632", "right": "issue:45072", "accept": false, - "reason": "`_is_hf_initialized` regression and dtype mismatch in inference are unrelated." - }, - { - "left": "issue:43701", - "right": "issue:45325", - "accept": false, - "reason": "Checkpoint key mismatch is unrelated to Qwen2.5-VL position-id handling." + "reason": "Different regressions: _is_hf_initialized handling vs dtype mismatch in bfloat16 inference." }, { "left": "issue:42796", "right": "issue:45381", "accept": false, - "reason": "Tokenizer BPE-marker handling and Qwen2.5-VL video `vision_position_ids` are unrelated." + "reason": "Different tokenizer/model bugs: BPE marker handling vs Qwen2.5-VL video position IDs." + }, + { + "left": "issue:43701", + "right": "issue:45325", + "accept": false, + "reason": "Checkpoint resume key mismatch is unrelated to Qwen2.5-VL rope-index scaling." }, { "left": "issue:44291", "right": "issue:44877", "accept": false, - "reason": "Unexpected `_is_hf_initialized` argument during loading is unrelated to strict config loading for granite_speech." + "reason": "Both are loading/init issues, but they affect different configs and mechanisms." }, { "left": "issue:43873", "right": "issue:45701", "accept": false, - "reason": "Quantization offload behavior and tokenization changes across versions are different issues." + "reason": "Different domains: quantization/offloading behavior vs tokenization changes across versions." }, { "left": "issue:43296", "right": "issue:45800", "accept": false, - "reason": "PaddleOCR-VL loading in vLLM and torch/transformers incompatibility are not the same bug." + "reason": "Unrelated version incompatibilities: PaddleOCR-VL in vLLM vs torch 2.4.1/transformers 5.8.0." }, { "left": "issue:44479", "right": "issue:45823", "accept": false, - "reason": "Qwen video-input regression and Gemma4 PLE device mismatch are different failures." + "reason": "Different model families and symptoms: Qwen multimodal video regression vs Gemma4 device mismatch during forward." }, { "left": "issue:43606", "right": "issue:44265", "accept": false, - "reason": "CPU offload device mismatch and `torch.export` / `torch_compilable_check` failure are unrelated." + "reason": "Different failures: CPU offload device mismatch vs torch.export/torch_compilable_check." } ] }, "evaluator_result": { - "accept": false, - "feedback": "The overall anchor choice is mostly grounded, but the output overstates duplicate evidence on at least two soft pairs. `issue:44829` vs `issue:45479` look like different bugs from the packet titles: one is about `flash_attention_3` causing degenerate training, the other is about `problem_type=\"single_label_classification\"` with `num_labels=1` causing zero loss across models. `issue:45325` vs `issue:45381` are both Qwen2.5-VL position-id regressions, but one is about still-image temporal scaling and the other about video-input `vision_position_ids`; that is only plausible similarity, not a conservative duplicate verdict. Because the review should reject overconfident soft-edge merges, this cluster should be rejected." + "accept": true, + "feedback": "The summary stays grounded in the packet and is suitably conservative: it identifies one plausible Qwen2.5-VL position-id duplicate family while keeping the rest of the soft-similarity pairs separate. The accepted pair (45325/45381) is reasonably supported by the titles as a closely related positional-indexing regression, and there are no PR-mergeability concerns here since these are issue pairs." }, "evaluator_used": true, - "retried": true, - "accepted_nontrivial_soft_edge": false, + "retried": false, + "accepted_nontrivial_soft_edge": true, "error_kind": null, "error_message": null }, @@ -4159,7 +4094,7 @@ "source": "llm", "cache_hit": false, "model": "gpt-5.4-mini?service_tier=flex", - "cluster_id": "cluster-10000-472", + "cluster_id": "cluster-10000-473", "nodes": [ "issue:34634", "issue:35707", @@ -4224,9 +4159,9 @@ "issue:46305|issue:46315", "issue:43377|issue:44610", "issue:42796|issue:46209", + "issue:43257|issue:45923", "issue:44568|issue:45920", "issue:43454|issue:45072", - "issue:43257|issue:45923", "issue:43317|issue:43873", "issue:43377|issue:44898", "issue:42796|issue:44977", @@ -4242,214 +4177,214 @@ "issue:41720|issue:44977", "issue:43881|issue:44991" ], - "prepared_review_unit_hash": "b585ce881e5ecfec19f33d5ab0407334bc1427cd88c4675133b5b0afb37e81cc", + "prepared_review_unit_hash": "94aa4e5dc92ae9f99085bf7ed583ca708a6ef2b6a18788860a33c24fa68fbc37", "component_budget": { - "node_count": 472, - "item_count": 472, - "soft_pair_count": 1732, - "serialized_chars": 898654, - "estimated_input_tokens": 224664, - "estimated_eval_tokens": 449584 + "node_count": 473, + "item_count": 473, + "soft_pair_count": 1738, + "serialized_chars": 901437, + "estimated_input_tokens": 225360, + "estimated_eval_tokens": 450976 }, "budget": { "node_count": 48, "item_count": 48, "soft_pair_count": 30, - "serialized_chars": 36789, - "estimated_input_tokens": 9198, - "estimated_eval_tokens": 18652 + "serialized_chars": 36793, + "estimated_input_tokens": 9199, + "estimated_eval_tokens": 18654 }, "overflow_policy": "truncate_then_skip", "trimmed": false, "aggressively_trimmed": false, "split": true, "analyst_result": { - "summary": "This cluster is mostly a grab-bag of unrelated closed bug reports. The soft edges are only superficial topic matches (tokenizers, loading, offload, device_map, etc.) and do not look like the same underlying bug or mergeable PR change.", - "confidence": 0.95, - "canonical_issue_reason": null, + "summary": "These items are broadly similar only at a subsystem level (tokenizers, loading, quantization, offload, multimodal), but the soft pairs describe different concrete failures and should not be merged as duplicates. No PRs are present in the cluster.", + "confidence": 0.9, + "canonical_issue_reason": "issue:43873 is the best representative because it is the most general, open bug report in the set and directly centers on a broad loading/offload/quantization failure mode, which makes it a reasonable anchor for the cluster.", "canonical_pr_reason": null, - "best_issue_reason": null, + "best_issue_reason": "issue:43873 is the most suitable issue overall since it is open, broadly scoped, and closer than the others to the recurring theme of runtime/loading behavior under quantization/offload.", "best_pr_reason": null, "soft_edge_verdicts": [ { "left": "issue:43262", "right": "issue:43746", "accept": false, - "reason": "Both involve audio / loading behavior, but one is a processor chat-template sampling-rate default and the other is PEFT checkpoint loading for GraniteSpeech. Different code paths and symptoms." + "reason": "Different problems: audio processor chat-template sampling-rate default vs GraniteSpeech PEFT local-checkpoint loading failure." }, { "left": "issue:40990", "right": "issue:45584", "accept": false, - "reason": "Unrelated failures: high perplexity on a language-model benchmark vs Whisper generation failing on empty transcription after token alignment." + "reason": "Unrelated failures: high perplexity in language-model evaluation vs Whisper empty-transcription generation after token alignment." }, { "left": "issue:34634", "right": "issue:35707", "accept": false, - "reason": "Bark voice preset handling and progressive generation with inputs_embeds/past_key_values are different bugs in different subsystems." + "reason": "Bark voice preset failure is unrelated to inputs_embeds/past_key_values generation behavior." }, { "left": "issue:35707", "right": "issue:41720", "accept": false, - "reason": "Progressive generation/past_key_values mismatch is unrelated to Qwen3 auto device mapping cudaErrorAssert." + "reason": "Different code paths and symptoms: progressive generation with embeddings/cache vs Qwen3 device-map cuda assert." }, { "left": "issue:43881", "right": "issue:44488", "accept": false, - "reason": "Both are model-loading complaints, but they affect different models and likely different failure causes; too broad to be the same bug." + "reason": "Both are loading issues, but one is glm-4v-9b load failure and the other is a general model-loading regression for cjvt/sleng-bert; no evidence of the same bug." }, { "left": "issue:43454", "right": "issue:43550", "accept": false, - "reason": "AyaVision weight tying / garbage generation is unrelated to Bamba torch.compile SDPA failure." + "reason": "AyaVision weight-tying bug and Bamba torch.compile/SDPA failure are separate model-specific issues." }, { "left": "issue:43606", "right": "issue:43819", "accept": false, - "reason": "Different audio models and different issues: CPU-offload device mismatch vs DAC latent reconstruction mismatch without STE." + "reason": "Both involve device/load behavior, but one is Bark CPU offload mismatch and the other is DAC.from_latents vs forward mismatch due to missing STE; not the same bug." }, { "left": "issue:39692", "right": "issue:40444", "accept": false, - "reason": "SigLIP2 doc example errors and Qwen2.5-VL iterable-dataset multi-image finetuning failures are different scenarios and code paths." + "reason": "SigLIP2 docs/quantization example errors are documentation/model-mismatch issues, while Qwen2.5-VL IterableDataset multi-image finetuning fails in training data handling." }, { "left": "issue:44186", "right": "issue:44265", "accept": false, - "reason": "Tokenizer crash on NER/padding is unrelated to torch.export failures from torch_compilable_check." + "reason": "LayoutLMv2 tokenizer crashes on NER/padding versus torch.export failures with torch_compilable_check; different components and failure modes." }, { "left": "issue:41720", "right": "issue:45823", "accept": false, - "reason": "Both mention device_map/dispatch, but one is a Qwen3 cuda assert on A800 and the other is Gemma4 PLE forward device mismatch; not the same concrete bug." + "reason": "Both are device-map related, but one is a Qwen3 cuda assert on A800 and the other is Gemma4 PLE forward device mismatch; not the same underlying bug." }, { "left": "issue:46305", "right": "issue:46315", "accept": false, - "reason": "Sliding-window masking off-by-one is unrelated to tokenizer convert_tokens_to_ids performance regression." + "reason": "MoonshineStreaming off-by-one in sliding-window masking is unrelated to convert_tokens_to_ids performance regression." }, { "left": "issue:43377", "right": "issue:44610", "accept": false, - "reason": "Missing padding-mask support in MIMI encoder and OmDet-Turbo processor size mismatch are different bugs in different components." + "reason": "MIMI batching/padding-mask inconsistency is unrelated to OmDet-Turbo processor/model input-size mismatch." }, { "left": "issue:42796", "right": "issue:46209", "accept": false, - "reason": "Tokenizer BPE-marker handling and FP8 inference failure on B200 are unrelated." + "reason": "Tokenizer BPE-marker handling and FP8 B200 inference support are distinct problems in different layers." + }, + { + "left": "issue:43257", + "right": "issue:45923", + "accept": false, + "reason": "Qwen3 MoE loading/conversion with accelerate+deepspeed is unrelated to a missing gradient_checkpointing flag on a Nemotron trust_remote_code variant." }, { "left": "issue:44568", "right": "issue:45920", "accept": false, - "reason": "Both are tokenizer regressions, but one is missing BOS/EOS insertion for a specific tokenizer and the other is wrong token IDs across multiple tokenizers; not the same code path." + "reason": "mdeberta special-token insertion regression and broad AutoTokenizer wrong token IDs regression are both tokenizer issues but not the same concrete bug." }, { "left": "issue:43454", "right": "issue:45072", "accept": false, - "reason": "Weight-tying/generation corruption is unrelated to bfloat16 dtype mismatches in SwitchTransformers/TimmWrapperModel." - }, - { - "left": "issue:43257", - "right": "issue:45923", - "accept": false, - "reason": "Qwen3 MoE weight conversion under accelerate+deepspeed is unrelated to a missing supports_gradient_checkpointing flag on a trust_remote_code variant." + "reason": "AyaVision lm_head tying and bfloat16 dtype mismatches in SwitchTransformers/TimmWrapperModel are unrelated." }, { "left": "issue:43317", "right": "issue:43873", "accept": false, - "reason": "Both mention offload/quantization, but one is a specific device_map=auto dequantized-model load failure and the other is a broader offloading regression; too generic to merge as one bug." + "reason": "Device-map dequantized-model loading on gpu+cpu offload overlaps only superficially with quantization offloading; the concrete failure described is different." }, { "left": "issue:43377", "right": "issue:44898", "accept": false, - "reason": "MIMI batched-vs-single padding-mask mismatch is unrelated to Perceiver non-default resolution interpolation failure." + "reason": "MIMI encoder batching bug and Perceiver non-default resolution failure concern different models and tensor-shape logic." }, { "left": "issue:42796", "right": "issue:44977", "accept": false, - "reason": "Tokenizer BPE-marker handling is unrelated to Qwen3.5 flash-attention generation issues." + "reason": "Tokenizer BPE marker handling is unrelated to Qwen3.5 flash-attention generation failure." }, { "left": "issue:43454", "right": "issue:43720", "accept": false, - "reason": "AyaVision lm_head tying and BitNet packed-weight unpacking are different model-loading bugs." + "reason": "AyaVision weight tying and BitNet packed-weight unpacking during accelerate loading are different model-loading defects." }, { "left": "issue:40990", "right": "issue:42947", "accept": false, - "reason": "Perplexity anomaly on a benchmark is unrelated to gradient-checkpointing ineffectiveness with PEFT LoRA." + "reason": "High perplexity evaluation issue is unrelated to gradient checkpointing being ineffective with PEFT LoRA." }, { "left": "issue:42617", "right": "issue:42915", "accept": false, - "reason": "3d_parallel.py runtime issue and Qwen3Moe FineGrainedFP8Config failure are unrelated." + "reason": "Failure to run 3d_parallel.py and Qwen3Moe FineGrainedFP8Config failure are unrelated." }, { "left": "issue:43322", "right": "issue:46050", "accept": false, - "reason": "Llava Next segmentation fault and sinq save/load breakage are different failure modes and models." + "reason": "Llava Next segmentation fault and sinq quantized save/load breakage are different model/runtime bugs." }, { "left": "issue:43749", "right": "issue:44534", "accept": false, - "reason": "FSDP CPU-efficient loading regression and non-persistent buffer junk initialization are not the same underlying bug." + "reason": "FSDP CPU RAM efficient loading and non-persistent buffers filled with junk in Transformers v5 are separate core behaviors." }, { "left": "issue:43572", "right": "issue:45759", "accept": false, - "reason": "StableLm pad_token_idx config regression is unrelated to from_config failing to unwrap text_config for composite Qwen multimodal configs." + "reason": "StableLmConfig missing pad_token_idx and Qwen composite config text_config unwrapping are different configuration regressions." }, { "left": "issue:41720", "right": "issue:43322", "accept": false, - "reason": "Both can fail during model loading, but one is a Qwen3 device-map assert and the other is a Llava Next segfault; not the same code path." + "reason": "Qwen3 auto-device-map cuda assert is not the same as Llava Next segmentation fault during loading." }, { "left": "issue:43526", "right": "issue:43883", "accept": false, - "reason": "BEiT image-processor label reduction bug is unrelated to MolmoForCausalLM missing all_tied_weights_keys." + "reason": "BeitImageProcessorFast reduce_labels bug is unrelated to Molmo missing all_tied_weights_keys attribute." }, { "left": "issue:44387", "right": "issue:45704", "accept": false, - "reason": "CUDA reserved-memory growth under int4 quantization and apex FusedRMSNorm memory leak are different memory issues with different causes." + "reason": "Int4 quantization CUDA reserved-memory OOM and T5 apex FusedRMSNorm memory leak are different performance/memory issues." }, { "left": "issue:41720", "right": "issue:44977", "accept": false, - "reason": "Both are Qwen-family generation/loading bugs, but cuda assert under auto device mapping is not the same as flash-attention generation failure." + "reason": "A800 device-map cuda assert and Qwen3.5 flash-attention generation failure are separate runtime paths." }, { "left": "issue:43881", "right": "issue:44991", "accept": false, - "reason": "Different tokenizer/model-loading failures for different repos; no shared concrete bug." + "reason": "glm-4v-9b loading failure is unrelated to tokenizer loading regression for EMBEDDIA/est-roberta." } ] }, @@ -4471,7 +4406,7 @@ "source": "llm", "cache_hit": false, "model": "gpt-5.4-mini?service_tier=flex", - "cluster_id": "cluster-10000-472", + "cluster_id": "cluster-10000-473", "nodes": [ "issue:26413", "issue:28530", @@ -4485,7 +4420,6 @@ "issue:43065", "issue:43295", "issue:43317", - "issue:43322", "issue:43404", "issue:43450", "issue:43526", @@ -4494,6 +4428,7 @@ "issue:43572", "issue:43577", "issue:43606", + "issue:43646", "issue:43708", "issue:43720", "issue:43873", @@ -4519,22 +4454,24 @@ "issue:45701", "issue:45823", "issue:46291", - "issue:46402" + "issue:46402", + "issue:46665" ], "soft_pairs": [ "issue:36331|issue:42617", "issue:44466|issue:45357", - "issue:36331|issue:42915", "issue:26413|issue:28530", + "issue:36331|issue:42915", "issue:44877|issue:44991", "issue:43450|issue:45701", "issue:40990|issue:45701", "issue:44186|issue:45072", "issue:43065|issue:46402", "issue:43873|issue:44756", - "issue:41720|issue:42617", "issue:43550|issue:43720", + "issue:41720|issue:42617", "issue:43577|issue:43708", + "issue:43646|issue:46665", "issue:43404|issue:45357", "issue:43550|issue:43606", "issue:43295|issue:45701", @@ -4547,234 +4484,233 @@ "issue:45632|issue:45684", "issue:41720|issue:45161", "issue:42915|issue:46291", - "issue:43531|issue:43572", "issue:28530|issue:33666", + "issue:43531|issue:43572", "issue:44671|issue:44743", "issue:36331|issue:41720", "issue:44186|issue:44361", - "issue:43526|issue:43881", - "issue:42915|issue:43322" + "issue:43526|issue:43881" ], - "prepared_review_unit_hash": "95cb06ff7b9974e3caf2b1e630e5f1535be1b5f695d6eb4959692105b90ccfc1", + "prepared_review_unit_hash": "5e1851d74beccba98bd40af006f32a5f195f76ae1559a47aa9fb220b89a93805", "component_budget": { - "node_count": 472, - "item_count": 472, - "soft_pair_count": 1732, - "serialized_chars": 898654, - "estimated_input_tokens": 224664, - "estimated_eval_tokens": 449584 + "node_count": 473, + "item_count": 473, + "soft_pair_count": 1738, + "serialized_chars": 901437, + "estimated_input_tokens": 225360, + "estimated_eval_tokens": 450976 }, "budget": { - "node_count": 47, - "item_count": 47, + "node_count": 48, + "item_count": 48, "soft_pair_count": 32, - "serialized_chars": 36796, - "estimated_input_tokens": 9199, - "estimated_eval_tokens": 18654 + "serialized_chars": 37364, + "estimated_input_tokens": 9341, + "estimated_eval_tokens": 18938 }, "overflow_policy": "truncate_then_skip", "trimmed": false, "aggressively_trimmed": false, "split": true, "analyst_result": { - "summary": "This cluster is heterogeneous: training-loop bugs, tokenizer regressions, model-loading/offload failures, multimodal processor issues, and save/pretrained serialization problems. None of the soft pairs look like true duplicates or the same concrete fix target.", - "confidence": 0.92, - "canonical_issue_reason": "No single true duplicate anchor stands out; issue 45701 is the broadest and most cross-referenced report, centered on a version-wide tokenization regression.", + "summary": "These issues are mostly unrelated and only share broad subsystem themes (trainer, loading, tokenizer, multimodal, quantization). I did not find any soft pair that looks like the same underlying bug or change, so all soft edges are rejected.", + "confidence": 0.88, + "canonical_issue_reason": "issue:43708 is the strongest standalone representative: it describes a specific Trainer resume_from_checkpoint regression with a clear repro and a concrete code path to fix.", "canonical_pr_reason": null, - "best_issue_reason": "Issue 45701 is the best umbrella representative for this cluster because it is general, externally referenced, and closest to the recurring tokenizer/version-regression theme seen in a few nearby reports.", + "best_issue_reason": "issue:43708 is the best overall issue candidate because it is detailed, reproducible, and narrowly scoped to one Trainer bug.", "best_pr_reason": null, "soft_edge_verdicts": [ { "left": "issue:36331", "right": "issue:42617", "accept": false, - "reason": "Trainer compute_loss signature bug vs inability to run a parallel training script; different failures and fixes." + "reason": "Different failures: trainer custom loss signature vs a 3d_parallel execution problem; only the general training context overlaps." }, { "left": "issue:44466", "right": "issue:45357", "accept": false, - "reason": "Both involve saving model state, but one is tied lm_head serialization and the other is incorrect visual encoder keys for Qwen3.5; different code paths." + "reason": "Both concern serialization, but one is lm_head tying/device-dependent save behavior and the other is incorrect visual encoder keys in Qwen3.5." }, { - "left": "issue:36331", - "right": "issue:42915", + "left": "issue:26413", + "right": "issue:28530", "accept": false, - "reason": "Unexpected compute_loss kwarg vs Qwen3Moe FP8 config failure; unrelated trainer API vs quantization issue." + "reason": "resume_from_checkpoint epoch iterator failure and early stopping metric resolution are separate Trainer bugs." }, { - "left": "issue:26413", - "right": "issue:28530", + "left": "issue:36331", + "right": "issue:42915", "accept": false, - "reason": "Checkpoint resume epoch_iterator failure is unrelated to early stopping metric resolution." + "reason": "CustomTrainer loss kwargs and Qwen3Moe FineGrainedFP8Config failure are unrelated code paths." }, { "left": "issue:44877", "right": "issue:44991", "accept": false, - "reason": "Strict config loading for one model family vs tokenizer loading failure for another; too different to be the same bug." + "reason": "Strict config loading for granite_speech is unrelated to tokenizer loading failure for EMBEDDIA/est-roberta." }, { "left": "issue:43450", "right": "issue:45701", "accept": false, - "reason": "Batched video processor shape bug vs tokenizer version-regression; different subsystem and failure mode." + "reason": "Video processor batch-shape bug is not the same as version-dependent tokenization changes." }, { "left": "issue:40990", "right": "issue:45701", "accept": false, - "reason": "Perplexity anomaly on a model/eval setup is not the same concrete issue as a tokenizer behavior change." + "reason": "High perplexity on gpt-oss-20b is a model evaluation symptom, not the same underlying tokenizer-version regression." }, { "left": "issue:44186", "right": "issue:45072", "accept": false, - "reason": "Tokenizer crash on NER/padding vs dtype mismatch in bfloat16 inference; unrelated." + "reason": "LayoutLMv2 tokenizer padding/NER crash and bfloat16 dtype mismatches are different bug classes." }, { "left": "issue:43065", "right": "issue:46402", "accept": false, - "reason": "Different Sam3 problems: dummy Conv2d in the pixel decoder vs missing text encoder weights." + "reason": "Sam3 PixelDecoder dummy conv and Sam3 text encoder weight loading are separate model issues." }, { "left": "issue:43873", "right": "issue:44756", "accept": false, - "reason": "Quantization offload not working vs disabling mmap to avoid OOM; related only at a high level." + "reason": "Quantization offloading behavior and mmap OOM on Strix Halo are not the same underlying defect." }, { - "left": "issue:41720", - "right": "issue:42617", + "left": "issue:43550", + "right": "issue:43720", "accept": false, - "reason": "Multi-GPU training for Qwen2-VL vs a 3d_parallel.py runtime problem; not the same bug." + "reason": "torch.compile with SDPA on Bamba-9B-v2 is unrelated to BitNet packed-weight loading in accelerate." }, { - "left": "issue:43550", - "right": "issue:43720", + "left": "issue:41720", + "right": "issue:42617", "accept": false, - "reason": "torch.compile/SDPA bug for Bamba vs packed-weight unpacking during accelerate loading for BitNet; different concrete code paths." + "reason": "Qwen3 device-map cuda assert and 3d_parallel execution failure do not describe the same bug." }, { "left": "issue:43577", "right": "issue:43708", "accept": false, - "reason": "dtype handling in Blip2 loading vs max_steps miscalculation on resume; unrelated trainer/model-loading bugs." + "reason": "BLIP2 dtype propagation and Trainer max_steps on resume are unrelated." + }, + { + "left": "issue:43646", + "right": "issue:46665", + "accept": false, + "reason": "Custom model initialization breakage and from_pretrained behavior inside classes are too different to merge." }, { "left": "issue:43404", "right": "issue:45357", "accept": false, - "reason": "Both concern model save/load integrity, but Mistral3 tied weights and Qwen3.5 visual encoder keys are distinct issues." + "reason": "lm_head weight tying in Mistral3ForConditionalGeneration is not the same as saving incorrect visual encoder keys." }, { "left": "issue:43550", "right": "issue:43606", "accept": false, - "reason": "Different device/compile/offload failures for different models and mechanisms." + "reason": "Bamba SDPA compile failure and Bark CPU offload device mismatch affect different models and code paths." }, { "left": "issue:43295", "right": "issue:45701", "accept": false, - "reason": "Processor.tokenizer regression is not the same as a broad tokenization-change report; the affected API surface is different." + "reason": "Processor.tokenizer regression and tokenization changes across versions are not the same underlying issue." }, { "left": "issue:44008", "right": "issue:45823", "accept": false, - "reason": "Gemma3n variable collision in forward vs Gemma4 PLE device mismatch; separate model-family bugs." + "reason": "Gemma3n variable-name collision in forward and Gemma4 PLE device mismatch are distinct forward-path bugs." }, { "left": "issue:44186", "right": "issue:44610", "accept": false, - "reason": "Tokenizer crash on layout/NER inputs vs processor image-size mismatch; different components and fixes." + "reason": "Tokenizer crash on batched NER inputs and OmDet-Turbo image-size mismatch are different failures." }, { "left": "issue:44610", "right": "issue:44898", "accept": false, - "reason": "One is an OmDet processor resizing bug, the other a Perceiver interpolation failure; not the same underlying issue." + "reason": "Processor producing the wrong resolution and Perceiver interpolation failing are not the same bug." }, { "left": "issue:42915", "right": "issue:43317", "accept": false, - "reason": "FineGrainedFP8Config failure vs dequantized model offload/load failure; both loading-related but not the same bug." + "reason": "FineGrainedFP8Config failure for Qwen3Moe is unrelated to device_map auto loading of dequantized models." }, { "left": "issue:44423", "right": "issue:45406", "accept": false, - "reason": "Serve crashes in multimodal inference, but the observed attribute errors and processor objects differ." + "reason": "Continuous-batching multimodal crash with str.to() and Gemma4Processor missing _tokenizer are separate serve issues." }, { "left": "issue:36010", "right": "issue:42617", "accept": false, - "reason": "ImportError for GenerationMixin is unrelated to a 3d_parallel execution problem." + "reason": "GenerationMixin import error and 3d_parallel runtime failure are unrelated." }, { "left": "issue:45632", "right": "issue:45684", "accept": false, - "reason": "Cache-path collision for trust_remote_code vs read-only permission propagation during save_pretrained; distinct issues." + "reason": "Cache-path collisions for trust_remote_code and read-only permissions propagation in save_pretrained are unrelated filesystem concerns." }, { "left": "issue:41720", "right": "issue:45161", "accept": false, - "reason": "Different distributed-training problems on different model families; too broad to merge." + "reason": "Auto device-mapping cuda assert and tensor-parallel-only GPT-OSS MoE failure are different distributed inference issues." }, { "left": "issue:42915", "right": "issue:46291", "accept": false, - "reason": "Quantization/config load failure is unrelated to kernels package import/version requirements." + "reason": "Qwen3Moe FP8 failure and kernels v0.15 pipeline import failure do not share the same code path." }, { - "left": "issue:43531", - "right": "issue:43572", + "left": "issue:28530", + "right": "issue:33666", "accept": false, - "reason": "Qwen3-MoE sliding_window behavior vs missing pad_token_idx in StableLmConfig; different config bugs." + "reason": "Early stopping metric handling and Qwen2-VL multi-GPU training are unrelated trainer features." }, { - "left": "issue:28530", - "right": "issue:33666", + "left": "issue:43531", + "right": "issue:43572", "accept": false, - "reason": "Early stopping metric handling vs Qwen2-VL multi-GPU training; unrelated trainer concerns." + "reason": "Qwen3-MoE sliding window behavior and StableLmConfig pad_token_idx are separate model/config issues." }, { "left": "issue:44671", "right": "issue:44743", "accept": false, - "reason": "CamemBERT masked LM regression vs Qwen3.5 recurrent-state caching bug; different model logic." + "reason": "CamemBERT masked LM predictions and Qwen3.5 recurrent-state reset behavior are different model-specific bugs." }, { "left": "issue:36331", "right": "issue:41720", "accept": false, - "reason": "Trainer API incompatibility vs multimodal multi-GPU training; not a duplicate." + "reason": "Unexpected num_items_in_batch in compute_loss and Qwen3 auto device-map cuda assert are unrelated." }, { "left": "issue:44186", "right": "issue:44361", "accept": false, - "reason": "Two different tokenizer AttributeErrors on different tokenizers and tasks; similar symptom only." + "reason": "Two tokenizer crashes, but on different models and different task/padding paths; not the same bug." }, { "left": "issue:43526", "right": "issue:43881", "accept": false, - "reason": "BeitImageProcessorFast label reduction bug vs glm-4v-9b loading failure; unrelated." - }, - { - "left": "issue:42915", - "right": "issue:43322", - "accept": false, - "reason": "Qwen3Moe FP8 config issue vs Llava Next load segfault; different model/load paths and failure modes." + "reason": "BeitImageProcessorFast reduce_labels bug and glm-4v-9b loading failure are unrelated." } ] }, @@ -4796,7 +4732,7 @@ "source": "llm", "cache_hit": false, "model": "gpt-5.4-mini?service_tier=flex", - "cluster_id": "cluster-10000-472", + "cluster_id": "cluster-10000-473", "nodes": [ "issue:28530", "issue:29159", @@ -4808,32 +4744,32 @@ "issue:36032", "issue:41720", "issue:42490", + "issue:42915", "issue:43316", + "issue:43322", "issue:43329", "issue:43388", "issue:43531", "issue:43540", "issue:43550", "issue:43606", + "issue:43644", "issue:43653", "issue:43720", "issue:43746", "issue:43782", "issue:43873", - "issue:44062", "issue:44077", - "issue:44368", "issue:44387", "issue:44451", "issue:44492", "issue:44509", + "issue:44534", "issue:44589", "issue:44625", "issue:44898", "issue:44913", - "issue:44987", "issue:44991", - "issue:45042", "issue:45059", "issue:45072", "issue:45357", @@ -4844,19 +4780,22 @@ "issue:45704", "issue:45759", "issue:45859", - "issue:46402" + "issue:46402", + "issue:46665" ], "soft_pairs": [ + "issue:42915|issue:43322", "issue:43316|issue:44913", "issue:28530|issue:30064", "issue:43540|issue:43782", "issue:33357|issue:35707", "issue:43873|issue:45704", "issue:43550|issue:43653", - "issue:30064|issue:34634", "issue:43606|issue:43720", + "issue:30064|issue:34634", "issue:29159|issue:36032", "issue:44387|issue:45701", + "issue:44534|issue:46665", "issue:44492|issue:44509", "issue:43388|issue:45059", "issue:41720|issue:43746", @@ -4867,218 +4806,216 @@ "issue:34634|issue:41720", "issue:34567|issue:35707", "issue:45357|issue:46402", - "issue:43531|issue:44589", "issue:44451|issue:44991", + "issue:43531|issue:44589", "issue:45759|issue:45859", - "issue:44077|issue:44625", "issue:44625|issue:45479", - "issue:44898|issue:45072", - "issue:44062|issue:44451", - "issue:44368|issue:45357", - "issue:44987|issue:45042" + "issue:44077|issue:44625", + "issue:43644|issue:46665", + "issue:44898|issue:45072" ], - "prepared_review_unit_hash": "d26857f14963bc2197cc2f45eb492bf83fc23aada4c0ef153300f696ae2dae1f", + "prepared_review_unit_hash": "8e0080f59356c192b779f97c0005e7c708cd335bcca361ee552156221ba9a3fd", "component_budget": { - "node_count": 472, - "item_count": 472, - "soft_pair_count": 1732, - "serialized_chars": 898654, - "estimated_input_tokens": 224664, - "estimated_eval_tokens": 449584 + "node_count": 473, + "item_count": 473, + "soft_pair_count": 1738, + "serialized_chars": 901437, + "estimated_input_tokens": 225360, + "estimated_eval_tokens": 450976 }, "budget": { - "node_count": 47, - "item_count": 47, + "node_count": 48, + "item_count": 48, "soft_pair_count": 29, - "serialized_chars": 36052, - "estimated_input_tokens": 9013, - "estimated_eval_tokens": 18282 + "serialized_chars": 36450, + "estimated_input_tokens": 9113, + "estimated_eval_tokens": 18482 }, "overflow_policy": "truncate_then_skip", "trimmed": false, "aggressively_trimmed": false, "split": true, "analyst_result": { - "summary": "This cluster is highly heterogeneous; most soft links are topical similarities rather than true duplicates. The only repeated pattern is broad transformer/model-loading/config regressions, but the concrete bugs differ across code paths and models.", - "confidence": 0.21, - "canonical_issue_reason": "No single issue is a strong canonical for this mixed cluster; if one representative is needed, issue 45759 is the broadest config-handling bug affecting composite Qwen multimodal models.", + "summary": "The cluster is heterogeneous: the soft pairs span unrelated tokenizer, multimodal, config, quantization, loading, and docs issues. I would not merge any of the proposed soft edges as duplicates.", + "confidence": 0.84, + "canonical_issue_reason": "If a representative issue is needed, `issue:45701` is the broadest regression-style report with the most external references, but there is no true canonical duplicate for the cluster as a whole.", "canonical_pr_reason": null, - "best_issue_reason": "Issue 45759 is the best available anchor because it is the most cross-cutting and least model-specific among the set, but it still does not unify the cluster into one underlying bug.", + "best_issue_reason": "`issue:45701` is the best overall representative among these issues because it is broad, user-facing, and comparatively well-connected, though it is still not a duplicate of the others.", "best_pr_reason": null, "soft_edge_verdicts": [ + { + "left": "issue:42915", + "right": "issue:43322", + "accept": false, + "reason": "Different failures: Qwen3Moe FineGrainedFP8Config vs Llava Next model load segfault." + }, { "left": "issue:43316", "right": "issue:44913", "accept": false, - "reason": "Both are config/serialization complaints, but they affect different configs and failure modes; not the same bug." + "reason": "Both are config-related, but one is a Gemma3TextConfig API discrepancy and the other is GPTNeoX rotary_pct persistence on reload." }, { "left": "issue:28530", "right": "issue:30064", "accept": false, - "reason": "Early stopping metric selection and image processor segmentation handling are unrelated." + "reason": "Early-stopping metric selection and void segmentation-map processing are unrelated code paths." }, { "left": "issue:43540", "right": "issue:43782", "accept": false, - "reason": "Both involve Qwen multimodal models, but one is a video-input processing error and the other is a from_pretrained weight_only load error." + "reason": "Different Qwen3 multimodal bugs: video input processing vs from_pretrained weight_only loading." }, { "left": "issue:33357", "right": "issue:35707", "accept": false, - "reason": "CLIP bus error on macOS and progressive generation with inputs_embeds/past_key_values are different code paths." + "reason": "MacOS CLIP bus error and progressive generation with inputs_embeds/past_key_values are unrelated." }, { "left": "issue:43873", "right": "issue:45704", "accept": false, - "reason": "Quantization/offload OOM and T5 RMSNorm memory leak are distinct problems." + "reason": "Quantization/offloading behavior and a T5 apex RMSNorm memory leak are distinct issues." }, { "left": "issue:43550", "right": "issue:43653", "accept": false, - "reason": "torch.compile SDPA failure and tokenizer special-token registration are unrelated." + "reason": "torch.compile SDPA failure in Bamba is unrelated to BigBirdTokenizer special-token registration." }, { - "left": "issue:30064", - "right": "issue:34634", + "left": "issue:43606", + "right": "issue:43720", "accept": false, - "reason": "Void segmentation-map processing and BarkProcessor voice preset handling are different processor bugs." + "reason": "One is a CPU offload device mismatch; the other is packed-weight unpacking during accelerate loading." }, { - "left": "issue:43606", - "right": "issue:43720", + "left": "issue:30064", + "right": "issue:34634", "accept": false, - "reason": "CPU offload device mismatch and packed-weight unpacking during accelerate loading are not the same underlying issue." + "reason": "Image-processor segmentation maps and BarkProcessor voice_preset handling do not share the same bug." }, { "left": "issue:29159", "right": "issue:36032", "accept": false, - "reason": "Slow/fast tokenizer inconsistency and T5 tokenizer attribute conflict are separate tokenizer bugs." + "reason": "Tokenizer slow/fast inconsistency is not the same as the T5 add_special_tokens name conflict." }, { "left": "issue:44387", "right": "issue:45701", "accept": false, - "reason": "Int4 quantization memory regression and version-dependent tokenization changes do not match." + "reason": "Int4 quantization memory growth and tokenization changes across versions are different regressions." + }, + { + "left": "issue:44534", + "right": "issue:46665", + "accept": false, + "reason": "Non-persistent buffer initialization junk and from_pretrained behavior inside Python classes are separate problems." }, { "left": "issue:44492", "right": "issue:44509", "accept": false, - "reason": "These are both docs-related, but one is a typo and the other is stale pipeline-task documentation." + "reason": "A typo in cache strategies is unrelated to docs mentioning removed pipeline tasks." }, { "left": "issue:43388", "right": "issue:45059", "accept": false, - "reason": "Metric gathering drops tuple labels; SAM3 text/box behavior is a different multimodal bug." + "reason": "Metric gathering last-batch truncation and SAM3 text/bounding-box behavior are different code paths." }, { "left": "issue:41720", "right": "issue:43746", "accept": false, - "reason": "Qwen3 auto device mapping CUDA assert and GraniteSpeech local-checkpoint loading are unrelated." + "reason": "Auto device-mapping cudaErrorAssert and PEFT local-checkpoint loading are not the same bug." }, { "left": "issue:43531", "right": "issue:45759", "accept": false, - "reason": "Sliding-window behavior in Qwen3-MoE and unwrapping text_config for composite Qwen configs are different failures." + "reason": "Qwen3-MoE sliding_window and composite-config text_config unwrapping are distinct issues." }, { "left": "issue:42490", "right": "issue:43329", "accept": false, - "reason": "Trailing-slash import warning and undefined video-path helpers in multimodal token counting are unrelated." + "reason": "Trailing-slash local model warning and undefined video-token helper names are unrelated." }, { "left": "issue:43606", "right": "issue:43653", "accept": false, - "reason": "Offload device mismatch and BigBirdTokenizer special-token registration are different subsystems." + "reason": "CPU offload device mismatch and tokenizer special-token registration are unrelated failures." }, { "left": "issue:45406", "right": "issue:45464", "accept": false, - "reason": "Gemma4Processor missing _tokenizer in serve and Qwen3.5 streaming chat/completions failures are different bugs." + "reason": "Gemma4Processor missing _tokenizer and Qwen3.5 streaming chat/completions failure are different bugs." }, { "left": "issue:34634", "right": "issue:41720", "accept": false, - "reason": "BarkProcessor voice_preset and Qwen3 auto device mapping target unrelated models and code paths." + "reason": "BarkProcessor voice_preset and Qwen3 auto device mapping failures do not overlap." }, { "left": "issue:34567", "right": "issue:35707", "accept": false, - "reason": "TrainerState token counting and progressive generation with past_key_values are unrelated." + "reason": "TrainerState token counter update and generation with inputs_embeds/past_key_values are unrelated." }, { "left": "issue:45357", "right": "issue:46402", "accept": false, - "reason": "Saving incorrect visual encoder keys and loading Sam3Model text encoder weights are different multimodal save/load issues." - }, - { - "left": "issue:43531", - "right": "issue:44589", - "accept": false, - "reason": "Qwen3-MoE sliding_window and Float8 storage lookup errors are unrelated." + "reason": "Qwen3.5 visual-encoder save_pretrained keys and Sam3Model text-encoder loading are separate multimodal issues." }, { "left": "issue:44451", "right": "issue:44991", "accept": false, - "reason": "These are separate load failures for different models and likely different root causes." + "reason": "Different model/tokenizer loading failures for ScandiBERT vs EMBEDDIA/est-roberta." }, { - "left": "issue:45759", - "right": "issue:45859", + "left": "issue:43531", + "right": "issue:44589", "accept": false, - "reason": "Both are Qwen3.5-family issues, but one is config unwrapping and the other is CPU-offload incompatibility in rotary embedding." + "reason": "Qwen3-MoE sliding-window handling and missing Float8 storage are unrelated." }, { - "left": "issue:44077", - "right": "issue:44625", + "left": "issue:45759", + "right": "issue:45859", "accept": false, - "reason": "patchtsmixer post_init policy and Qwen3.5 num_labels propagation are unrelated." + "reason": "from_config text_config unwrapping and CPU-offload rotary embedding compatibility are not the same issue." }, { "left": "issue:44625", "right": "issue:45479", "accept": false, - "reason": "Num-label propagation and degenerate zero-loss classification are both classification-related, but they are different concrete bugs." + "reason": "num_labels propagation and single-label zero-loss classification are different bugs." }, { - "left": "issue:44898", - "right": "issue:45072", - "accept": false, - "reason": "Perceiver image-resolution failure and bfloat16 dtype mismatch are different runtime problems." - }, - { - "left": "issue:44062", - "right": "issue:44451", + "left": "issue:44077", + "right": "issue:44625", "accept": false, - "reason": "AddedToken keyword duplication and ScandiBERT loading failure are unrelated." + "reason": "Optional post_init policy for patchtsmixer is unrelated to Qwen3.5 num_labels propagation." }, { - "left": "issue:44368", - "right": "issue:45357", + "left": "issue:43644", + "right": "issue:46665", "accept": false, - "reason": "A Qwen3.5 LoRA warning about tie_word_embeddings and incorrect visual-encoder key saving are different issues." + "reason": "Non-persistent buffer junk in Transformers 5.0 and from_pretrained behavior inside Python classes are distinct." }, { - "left": "issue:44987", - "right": "issue:45042", + "left": "issue:44898", + "right": "issue:45072", "accept": false, - "reason": "Model loading for physical-intelligence/fast and PIL backend requiring torchvision are unrelated." + "reason": "Perceiver resolution handling and bfloat16 dtype mismatches affect different model paths." } ] }, @@ -5100,7 +5037,7 @@ "source": "llm", "cache_hit": false, "model": "gpt-5.4-mini?service_tier=flex", - "cluster_id": "cluster-10000-472", + "cluster_id": "cluster-10000-473", "nodes": [ "issue:30064", "issue:36010", @@ -5112,12 +5049,9 @@ "issue:43421", "issue:43531", "issue:43540", - "issue:43550", "issue:43596", - "issue:43618", "issue:43645", "issue:43653", - "issue:43697", "issue:43720", "issue:43746", "issue:43782", @@ -5129,7 +5063,8 @@ "issue:44062", "issue:44186", "issue:44220", - "issue:44488", + "issue:44368", + "issue:44451", "issue:44514", "issue:44545", "issue:44610", @@ -5137,9 +5072,11 @@ "issue:44843", "issue:44918", "issue:44936", - "issue:45072", + "issue:44987", + "issue:45042", "issue:45081", "issue:45356", + "issue:45357", "issue:45381", "issue:45704", "issue:45759", @@ -5152,258 +5089,251 @@ "issue:46463" ], "soft_pairs": [ + "issue:44062|issue:44451", + "issue:44368|issue:45357", + "issue:44987|issue:45042", "issue:43381|issue:43844", "issue:39692|issue:44610", "issue:44514|issue:44545", - "issue:43421|issue:43927", "issue:44918|issue:44936", + "issue:43421|issue:43927", "issue:45704|issue:45923", - "issue:41720|issue:45704", "issue:43746|issue:46050", + "issue:41720|issue:45704", "issue:43381|issue:45704", "issue:43531|issue:43645", "issue:42886|issue:44843", "issue:46032|issue:46463", - "issue:43421|issue:43653", "issue:45759|issue:45923", - "issue:39692|issue:43873", "issue:43782|issue:45381", + "issue:43421|issue:43653", + "issue:39692|issue:43873", "issue:43994|issue:44671", "issue:43746|issue:43873", "issue:43421|issue:44186", "issue:30064|issue:36010", "issue:45081|issue:45356", "issue:43720|issue:43819", - "issue:45356|issue:45907", "issue:43653|issue:43720", - "issue:43596|issue:44220", - "issue:43295|issue:45834", + "issue:45356|issue:45907", "issue:43540|issue:46294", - "issue:43550|issue:45072", - "issue:43550|issue:43927", - "issue:43618|issue:43697", - "issue:44062|issue:44488" + "issue:43596|issue:44220", + "issue:43295|issue:45834" ], - "prepared_review_unit_hash": "d1bd1c6ec11f8bbef9483c8b2250019770e9b0a7cedce0b53e287bdc3c3e0ae3", + "prepared_review_unit_hash": "76a32d9a1ea60f3eab90c5cc8ecbe65f0d936fc959840a2128691ebb9795be45", "component_budget": { - "node_count": 472, - "item_count": 472, - "soft_pair_count": 1732, - "serialized_chars": 898654, - "estimated_input_tokens": 224664, - "estimated_eval_tokens": 449584 + "node_count": 473, + "item_count": 473, + "soft_pair_count": 1738, + "serialized_chars": 901437, + "estimated_input_tokens": 225360, + "estimated_eval_tokens": 450976 }, "budget": { "node_count": 48, "item_count": 48, - "soft_pair_count": 31, - "serialized_chars": 37237, - "estimated_input_tokens": 9310, - "estimated_eval_tokens": 18876 + "soft_pair_count": 30, + "serialized_chars": 37051, + "estimated_input_tokens": 9263, + "estimated_eval_tokens": 18782 }, "overflow_policy": "truncate_then_skip", "trimmed": false, "aggressively_trimmed": false, "split": true, "analyst_result": { - "summary": "This cluster is mostly a loose bundle of unrelated Transformers bug reports across tokenizers, multimodal processors, quantization, and training. The only clear duplicate is the Qwen2_5_VLProcessor.apply_chat_template batched padding=False crash pair (44514/44545).", + "summary": "This cluster is highly heterogeneous: almost all pairs are different bugs in different models or subsystems. The only clear duplicate is the pair of Qwen2_5_VLProcessor.apply_chat_template reports with identical failure mode and wording.", "confidence": 0.93, - "canonical_issue_reason": "issue:44514 is the better canonical report: it is earlier, more detailed, and issue:44545 appears to be the same batched-input/padding=False crash with a near-identical title.", - "canonical_pr_reason": null, - "best_issue_reason": "issue:44514 is the strongest issue candidate overall because it cleanly represents the only true duplicate pair in the cluster and has the most detailed problem statement.", - "best_pr_reason": null, + "canonical_issue_reason": "No single canonical issue stands out because the cluster is mostly unrelated reports; the only true duplicate-looking pair is issue:44514/issue:44545.", + "canonical_pr_reason": "No PRs are present in this cluster.", + "best_issue_reason": "If a representative issue is needed, issue:44545 is the clearest duplicate target because it matches issue:44514 exactly and has the most directly stated failure scenario.", + "best_pr_reason": "No PRs are present, so there is no best PR candidate.", "soft_edge_verdicts": [ + { + "left": "issue:44062", + "right": "issue:44451", + "accept": false, + "reason": "Both are tokenizer/model-loading failures, but the concrete errors and affected code paths differ." + }, + { + "left": "issue:44368", + "right": "issue:45357", + "accept": false, + "reason": "One is a tie_word_embeddings warning in fine-tuning; the other is incorrect visual encoder key saving. Different bugs." + }, + { + "left": "issue:44987", + "right": "issue:45042", + "accept": false, + "reason": "Loading a model repo and a PIL/torchvision backend requirement regression are unrelated." + }, { "left": "issue:43381", "right": "issue:43844", "accept": false, - "reason": "Different failures: gradient checkpointing in eval mode vs a ZeRO-3 training gradient growth issue." + "reason": "Gradient checkpointing in eval mode is a different problem from gradient explosion under ZeRO-3." }, { "left": "issue:39692", "right": "issue:44610", "accept": false, - "reason": "Unrelated bugs: SigLIP2 doc/example mismatch vs OmDet-Turbo processor/model image-size mismatch." + "reason": "Different multimodal examples and failures; no shared concrete bug." }, { "left": "issue:44514", "right": "issue:44545", "accept": true, - "reason": "Same title and same concrete failure: Qwen2_5_VLProcessor.apply_chat_template crashes on batched input with padding=False." + "reason": "Same processor, same batched padding=False crash, same title and failure mode; clear duplicate." }, { - "left": "issue:43421", - "right": "issue:43927", + "left": "issue:44918", + "right": "issue:44936", "accept": false, - "reason": "Both tokenizer/config related, but different code paths: runtime post-processor update vs save/load custom token IDs." + "reason": "Both involve training/eval behavior, but they describe different failures and code paths." }, { - "left": "issue:44918", - "right": "issue:44936", + "left": "issue:43421", + "right": "issue:43927", "accept": false, - "reason": "Different trainer issues: unpacking Qwen3.5 embeddings in SFT vs evaluate() failing after train()." + "reason": "Runtime special-token post-processing and saved config token IDs are distinct tokenizer issues." }, { "left": "issue:45704", "right": "issue:45923", "accept": false, - "reason": "Different problems: T5 apex RMSNorm memory leak vs a missing gradient_checkpointing flag on a trust_remote_code model." + "reason": "Apex RMSNorm memory leak and missing gradient-checkpointing flag are unrelated." }, { - "left": "issue:41720", - "right": "issue:45704", + "left": "issue:43746", + "right": "issue:46050", "accept": false, - "reason": "Completely different subsystems and symptoms: Qwen3 device-mapping CUDA assert vs T5 RMSNorm leak." + "reason": "PEFT local checkpoint loading and quantized save/load breakage are different persistence bugs." }, { - "left": "issue:43746", - "right": "issue:46050", + "left": "issue:41720", + "right": "issue:45704", "accept": false, - "reason": "Different load/save issues: PEFT adapters failing from local checkpoints vs sinq quantized model reload broken." + "reason": "Qwen3 device-mapping CUDA assert has no overlap with T5 RMSNorm memory leak." }, { "left": "issue:43381", "right": "issue:45704", "accept": false, - "reason": "No overlap in concrete bug: eval-mode gradient checkpointing vs T5 RMSNorm leak." + "reason": "Different model families and different failure mechanisms." }, { "left": "issue:43531", "right": "issue:43645", "accept": false, - "reason": "Different models and failures: Qwen3-MoE sliding_window vs custom models in Jupyter notebooks on Transformers 5.0.0." + "reason": "Sliding-window behavior in Qwen3-MoE is unrelated to Jupyter custom-model initialization." }, { "left": "issue:42886", "right": "issue:44843", "accept": false, - "reason": "Both involve offline/tokenizer loading, but the concrete bugs differ and the titles point to different failure points." + "reason": "Both mention offline tokenizer loading, but one is a generic cache-load failure and the other is a specific _patch_mistral_regex/model_info regression." }, { "left": "issue:46032", "right": "issue:46463", "accept": false, - "reason": "Unrelated: Mamba2Mixer cache correctness vs T5 decoder-input validation error messaging." + "reason": "Incorrect use_cache results and an error-message bug in T5 are unrelated." }, { - "left": "issue:43421", - "right": "issue:43653", + "left": "issue:45759", + "right": "issue:45923", "accept": false, - "reason": "Both touch token special handling, but one is runtime post-processor updates and the other is BigBird mask-token registration/decoding." + "reason": "Composite config unwrapping and a missing gradient-checkpointing flag are separate issues." }, { - "left": "issue:45759", - "right": "issue:45923", + "left": "issue:43782", + "right": "issue:45381", "accept": false, - "reason": "Different issues: composite Qwen config unwrapping vs missing gradient_checkpointing support." + "reason": "Different Qwen model families and different multimodal failures." }, { - "left": "issue:39692", - "right": "issue:43873", + "left": "issue:43421", + "right": "issue:43653", "accept": false, - "reason": "Different topics: SigLIP2 docs/model mismatch vs quantization offloading behavior." + "reason": "Both involve tokenizer special tokens, but one is about updating post-processors at runtime and the other is a BigBird-specific mask-token registration bug." }, { - "left": "issue:43782", - "right": "issue:45381", + "left": "issue:39692", + "right": "issue:43873", "accept": false, - "reason": "Both VL-related, but one is weight_only loading and the other is video vision_position_ids correctness." + "reason": "A SigLIP2 documentation example bug is unrelated to quantization offloading." }, { "left": "issue:43994", "right": "issue:44671", "accept": false, - "reason": "Different models and failure modes: SigLIP2 nonsensical outputs vs CamemBERT masked LM regression." + "reason": "Incorrect outputs on two different models are not the same underlying bug." }, { "left": "issue:43746", "right": "issue:43873", "accept": false, - "reason": "Different bugs: PEFT local checkpoint loading vs quantization/offloading behavior." + "reason": "Checkpoint loading with PEFT and quantization offloading are different code paths." }, { "left": "issue:43421", "right": "issue:44186", "accept": false, - "reason": "Different tokenizer bugs: post-processor refresh at runtime vs LayoutLMv2 batching/NER crashes." + "reason": "Runtime tokenizer post-processor updates do not match LayoutLMv2 NER/padding crashes." }, { "left": "issue:30064", "right": "issue:36010", "accept": false, - "reason": "Completely unrelated: image processor segmentation maps vs a GenerationMixin import error." + "reason": "Image segmentation map processing and an ImportError for GenerationMixin are unrelated." }, { "left": "issue:45081", "right": "issue:45356", "accept": false, - "reason": "Related tokenizer-regex area, but not clearly the same underlying bug; one is a crash, the other a codec/regression/warning issue." + "reason": "Both are in the Mistral-regex/tokenizer area, but one is an AttributeError crash and the other is a Kimi-K2.5 regression about codec handling; not enough to treat as the same bug." }, { "left": "issue:43720", "right": "issue:43819", "accept": false, - "reason": "Different model internals: unpacking packed BitNet weights vs DAC STE mismatch." - }, - { - "left": "issue:45356", - "right": "issue:45907", - "accept": false, - "reason": "Different bug classes: tokenizer regex/codec regression vs list_repo_files revision forwarding." + "reason": "Packed-weight loading and DAC latent reconstruction are unrelated." }, { "left": "issue:43653", "right": "issue:43720", "accept": false, - "reason": "Different failures: BigBird special-token registration vs BitNet packed-weight loading." - }, - { - "left": "issue:43596", - "right": "issue:44220", - "accept": false, - "reason": "Unrelated runtime errors: deepspeed/zero3 BertModel index issue vs audio fbank feature extraction." + "reason": "BigBird mask-token registration and BitNet packed-weight unpacking are different issues." }, { - "left": "issue:43295", - "right": "issue:45834", + "left": "issue:45356", + "right": "issue:45907", "accept": false, - "reason": "Different models and symptoms: processor.tokenizer regression vs Kosmos2.5 long-OCR index error." + "reason": "Tokenizer regression and a hub API forwarding bug are unrelated." }, { "left": "issue:43540", "right": "issue:46294", "accept": false, - "reason": "Both are video-input issues, but they affect different Qwen models and different failure points." + "reason": "Both concern video inputs, but they affect different Qwen model families and describe different processor/template mismatches." }, { - "left": "issue:43550", - "right": "issue:45072", + "left": "issue:43596", + "right": "issue:44220", "accept": false, - "reason": "Different model/dtype/compile problems: Bamba-9B-v2 SDPA torch.compile vs SwitchTransformers/TimmWrapperModel bfloat16 mismatch." + "reason": "DeepSpeed ZeRO-3/BertModel initialization and audio feature extraction are unrelated." }, { - "left": "issue:43550", - "right": "issue:43927", + "left": "issue:43295", + "right": "issue:45834", "accept": false, - "reason": "No overlap: compile/SDPA failure vs save/load custom token IDs causing generation IndexError." - }, - { - "left": "issue:43618", - "right": "issue:43697", - "accept": false, - "reason": "Different regressions: CLIPOutput attentions field removal vs RTDetrV2 output differences." - }, - { - "left": "issue:44062", - "right": "issue:44488", - "accept": false, - "reason": "Different tokenizer problems: AddedToken argument collision vs failing to load a specific model." + "reason": "Processor.tokenizer regression and Kosmos2.5 OCR index errors are different bugs." } ] }, "evaluator_result": { "accept": true, - "feedback": "The cluster summary is well grounded in the packet and stays conservative. The only clearly supported duplicate pair is issue:44514 / issue:44545, and the other soft-edge rejections are consistent with the titles and bodies in the packet. I don\u2019t see any overstatement that would warrant rejection." + "feedback": "The cluster summary is well grounded in the packet: it correctly identifies the single obvious duplicate pair (issue:44514 / issue:44545) and treats the remaining soft pairs conservatively as non-duplicates. No PRs are present, so the PR-related fields are also consistent." }, "evaluator_used": true, "retried": false, @@ -5422,11 +5352,9 @@ "source": "llm", "cache_hit": false, "model": "gpt-5.4-mini?service_tier=flex", - "cluster_id": "cluster-10000-472", + "cluster_id": "cluster-10000-473", "nodes": [ "issue:30064", - "issue:34634", - "issue:36010", "issue:36331", "issue:39692", "issue:41720", @@ -5436,24 +5364,26 @@ "issue:43531", "issue:43540", "issue:43550", - "issue:43582", "issue:43606", + "issue:43611", + "issue:43618", "issue:43643", "issue:43644", "issue:43653", + "issue:43697", "issue:43716", "issue:43749", "issue:43756", "issue:43819", + "issue:43927", "issue:43950", "issue:44008", + "issue:44062", "issue:44186", "issue:44220", "issue:44442", "issue:44448", - "issue:44479", - "issue:44485", - "issue:44509", + "issue:44488", "issue:44661", "issue:44671", "issue:44821", @@ -5471,227 +5401,234 @@ "issue:45834", "issue:45987", "issue:46032", - "issue:46512" + "issue:46665" ], "soft_pairs": [ + "issue:43550|issue:45072", + "issue:43550|issue:43927", + "issue:43618|issue:43697", + "issue:44062|issue:44488", + "issue:43611|issue:46665", "issue:44220|issue:44661", "issue:44186|issue:44898", "issue:43653|issue:43819", "issue:43531|issue:43643", - "issue:43643|issue:44877", "issue:43335|issue:43756", + "issue:43540|issue:44008", "issue:43550|issue:43819", "issue:30064|issue:36331", - "issue:43540|issue:44008", + "issue:43643|issue:44877", "issue:41720|issue:44991", "issue:44843|issue:45081", "issue:44821|issue:45084", "issue:39692|issue:45834", "issue:43322|issue:45685", "issue:44186|issue:44442", + "issue:43716|issue:44008", "issue:44448|issue:44671", + "issue:43388|issue:46032", "issue:45584|issue:45834", - "issue:45183|issue:45479", "issue:43644|issue:43749", - "issue:43388|issue:46032", - "issue:43716|issue:44008", + "issue:45183|issue:45479", "issue:43950|issue:45987", - "issue:43582|issue:46512", - "issue:43606|issue:45072", - "issue:34634|issue:36010", - "issue:44485|issue:44509", - "issue:44479|issue:45834" + "issue:43606|issue:45072" ], - "prepared_review_unit_hash": "4c07532d146920014a1babec6eef377641350a8254407d08abaa210565dee58c", + "prepared_review_unit_hash": "ddeea3d644ee6816b811ae698227f9e95bc42f69ebad7d671d95938f51d36d62", "component_budget": { - "node_count": 472, - "item_count": 472, - "soft_pair_count": 1732, - "serialized_chars": 898654, - "estimated_input_tokens": 224664, - "estimated_eval_tokens": 449584 + "node_count": 473, + "item_count": 473, + "soft_pair_count": 1738, + "serialized_chars": 901437, + "estimated_input_tokens": 225360, + "estimated_eval_tokens": 450976 }, "budget": { "node_count": 48, "item_count": 48, - "soft_pair_count": 27, - "serialized_chars": 36040, - "estimated_input_tokens": 9010, - "estimated_eval_tokens": 18276 + "soft_pair_count": 28, + "serialized_chars": 36312, + "estimated_input_tokens": 9078, + "estimated_eval_tokens": 18412 }, "overflow_policy": "truncate_then_skip", "trimmed": false, "aggressively_trimmed": false, "split": true, "analyst_result": { - "summary": "This cluster is largely a false positive: the items span unrelated models, processors, loading paths, training utilities, and docs regressions. None of the soft pairs look like the same underlying bug/change, so I would reject all soft edges.", - "confidence": 0.92, - "canonical_issue_reason": "If one issue must anchor the cluster, issue 43950 is the strongest representative because it describes a broad, self-contained core loading regression with clear impact. That said, the set is too heterogeneous to have a true canonical duplicate.", + "summary": "This cluster is mostly a noisy mix of unrelated model-loading, tokenizer, and runtime regressions. The soft pairs share broad subsystem themes, but none look like the same concrete bug/change, so I would not merge any of them as duplicates.", + "confidence": 0.88, + "canonical_issue_reason": "issue:43611 is the best anchor because it is a broad, core loading regression in Transformers 5.0 and is more representative of the cluster\u2019s repeated from_pretrained/model-init complaints than the narrower, model-specific reports.", "canonical_pr_reason": null, - "best_issue_reason": "Issue 43950 is the best standalone triage candidate: it is a concrete transformers regression, affects a core API (`from_pretrained`), and is easier to route/fix than the more specialized or doc-specific reports.", + "best_issue_reason": "issue:43611 is the strongest single issue to keep as the cluster representative: it is clearly scoped, impacts a common code path, and is more general than the highly model-specific failures here.", "best_pr_reason": null, "soft_edge_verdicts": [ + { + "left": "issue:43550", + "right": "issue:45072", + "accept": false, + "reason": "Different failures: torch.compile + SDPA in Bamba versus dtype mismatches in SwitchTransformers/TimmWrapper. Same broad runtime category, not the same bug." + }, + { + "left": "issue:43550", + "right": "issue:43927", + "accept": false, + "reason": "Bamba compile/SDPA crash and DiaConfig save/load token-ID loss are unrelated code paths and symptoms." + }, + { + "left": "issue:43618", + "right": "issue:43697", + "accept": false, + "reason": "CLIPOutput attentions regression and RTDetrV2 output drift are separate output-behavior issues in different models." + }, + { + "left": "issue:44062", + "right": "issue:44488", + "accept": false, + "reason": "Tokenizer AddedToken special-arg error is unrelated to loading cjvt/sleng-bert; no shared concrete bug." + }, + { + "left": "issue:43611", + "right": "issue:46665", + "accept": false, + "reason": "Both involve loading, but one is a specific base_model_prefix regression and the other is a generic from_pretrained-in-class behavior complaint." + }, { "left": "issue:44220", "right": "issue:44661", "accept": false, - "reason": "Different subsystems and failures: audio feature extraction vs add-new-model script/model-mapping logic." + "reason": "Audio fbank extraction and add-new-model-like / tokenizer mapping logic are different subsystems and failure modes." }, { "left": "issue:44186", "right": "issue:44898", "accept": false, - "reason": "Unrelated models and code paths: LayoutLMv2 tokenizer crash vs Perceiver image classification positional-encoding issue." + "reason": "Tokenizer padding/truncation crash in LayoutLMv2 is unrelated to Perceiver image classification resolution handling." }, { "left": "issue:43653", "right": "issue:43819", "accept": false, - "reason": "Tokenizer special-token registration bug vs DAC latent/STE mismatch; not the same defect." + "reason": "BigBirdTokenizer special-token registration and DAC.from_latents / STE mismatch are different bugs in different components." }, { "left": "issue:43531", "right": "issue:43643", "accept": false, - "reason": "Qwen3-MoE sliding-window behavior is unrelated to AutoConfig trust_remote_code field loss." + "reason": "Qwen3-MoE sliding_window behavior and AutoConfig trust_remote_code field loss do not share the same underlying defect." }, { - "left": "issue:43643", - "right": "issue:44877", + "left": "issue:43335", + "right": "issue:43756", "accept": false, - "reason": "Both involve config loading, but one is missing fields from remote code and the other is strict config rejection for granite_speech; different bugs." + "reason": "Both are model-structure/configuration bugs, but they affect different models and different mechanisms, so they are not the same change." }, { - "left": "issue:43335", - "right": "issue:43756", + "left": "issue:43540", + "right": "issue:44008", "accept": false, - "reason": "SwitchTransformers sparse-layer construction and Smollm3 RoPE-layer selection are distinct model setup bugs." + "reason": "Video-input validation in Qwen3OmniMoe and Gemma3n variable-name collision are distinct multimodal forward-pass bugs." }, { "left": "issue:43550", "right": "issue:43819", "accept": false, - "reason": "torch.compile/SDPA failure in Bamba is unrelated to DAC.from_latents vs forward mismatch." + "reason": "Bamba torch.compile/SDPA failure and DAC latents mismatch are unrelated; same broad audio/model area is not enough." }, { "left": "issue:30064", "right": "issue:36331", "accept": false, - "reason": "Image processor segmentation-map handling and CustomTrainer compute_loss API mismatch are unrelated." + "reason": "Image segmentation-map processing and CustomTrainer.compute_loss signature breakage are unrelated." }, { - "left": "issue:43540", - "right": "issue:44008", + "left": "issue:43643", + "right": "issue:44877", "accept": false, - "reason": "Video input parsing in Qwen3OmniMoe and Gemma3n forward variable collision are different code paths and error causes." + "reason": "AutoConfig missing fields with trust_remote_code and strict config rejection of granite_speech are different config-loading problems." }, { "left": "issue:41720", "right": "issue:44991", "accept": false, - "reason": "Auto device-mapping cuda assert on Qwen3 and tokenizer-loading failure for est-roberta are unrelated regressions." + "reason": "Qwen3 auto device mapping CUDA assert and tokenizer load failure for EMBEDDIA/est-roberta are unrelated failures." }, { "left": "issue:44843", "right": "issue:45081", "accept": false, - "reason": "Same helper name, but different failures: unconditional hub lookup offline vs backend_tokenizer AttributeError under fix_mistral_regex." + "reason": "Both touch _patch_mistral_regex, but one is an offline model_info call problem and the other is an AttributeError from a different tokenizer object shape." }, { "left": "issue:44821", "right": "issue:45084", "accept": false, - "reason": "AutoImageProcessor URL loading and template-node compilation error do not share a concrete bug." + "reason": "AutoImageProcessor URL loading and template-node compilation errors are unrelated." }, { "left": "issue:39692", "right": "issue:45834", "accept": false, - "reason": "Documentation/example errors for SigLIP2 are unrelated to Kosmos2.5 long-ocr index errors." + "reason": "SigLIP2 docs/example errors and Kosmos2.5 long-OCR index error are different issues with no shared concrete bug." }, { "left": "issue:43322", "right": "issue:45685", "accept": false, - "reason": "Llava-next loading segfault and MPS histogram Int kernel error are unrelated failures." + "reason": "Llava Next segfault on load and MPS histogram_mps integer error are unrelated runtime/device issues." }, { "left": "issue:44186", "right": "issue:44442", "accept": false, - "reason": "Tokenizer crash on NER/padding is not the same as AutoTokenizer failing to load a specific tokenizer class." + "reason": "LayoutLMv2Tokenizer NER/padding crash and FastSpeech2ConformerTokenizer auto-load failure are different tokenizer bugs." + }, + { + "left": "issue:43716", + "right": "issue:44008", + "accept": false, + "reason": "Mistral-3 image preprocessor dtype mismatch and Gemma 3n audio_mel_mask attribute collision are unrelated multimodal bugs." }, { "left": "issue:44448", "right": "issue:44671", "accept": false, - "reason": "Different models and different regressions: Pegasus output drift vs CamemBERT masked-LM predictions." + "reason": "Both are v5 output regressions, but they affect different models and likely different code paths; not one underlying fix." }, { - "left": "issue:45584", - "right": "issue:45834", + "left": "issue:43388", + "right": "issue:46032", "accept": false, - "reason": "Whisper empty-transcription generation bug is unrelated to Kosmos2.5 OCR long-input indexing." + "reason": "Metric-gathering label truncation and Mamba2Mixer use_cache correctness are unrelated training/inference bugs." }, { - "left": "issue:45183", - "right": "issue:45479", + "left": "issue:45584", + "right": "issue:45834", "accept": false, - "reason": "Input validation messaging/XOR logic is a different problem from zero-loss behavior with num_labels=1." + "reason": "Whisper empty-transcription generation and Kosmos2.5 long-input index error are separate edge cases." }, { "left": "issue:43644", "right": "issue:43749", "accept": false, - "reason": "Non-persistent buffer corruption during from_pretrained and FSDP CPU RAM loading breakage are distinct regressions." + "reason": "Non-persistent buffer corruption during from_pretrained and FSDP CPU RAM efficient loading breakage are different regressions." }, { - "left": "issue:43388", - "right": "issue:46032", - "accept": false, - "reason": "Metrics gathering last-batch truncation and Mamba2 use_cache correctness are unrelated." - }, - { - "left": "issue:43716", - "right": "issue:44008", + "left": "issue:45183", + "right": "issue:45479", "accept": false, - "reason": "Image preprocessor dtype mismatch in Mistral-3 is unrelated to Gemma3n variable-name collision." + "reason": "Validation error messaging for input_ids/inputs_embeds and degenerate zero-loss classification behavior are not the same bug." }, { "left": "issue:43950", "right": "issue:45987", "accept": false, - "reason": "One is buffer corruption on loading; the other is a return-type contract bug in StaticCache.get_seq_length()." - }, - { - "left": "issue:43582", - "right": "issue:46512", - "accept": false, - "reason": "Apple Silicon warmup TypeError and stale config.dtype after from_config are different issues." + "reason": "Silent non-persistent buffer corruption and StaticCache.get_seq_length returning a shape-(1,) tensor are unrelated contract violations." }, { "left": "issue:43606", "right": "issue:45072", "accept": false, - "reason": "CPU offload device mismatch for Bark-small is unrelated to dtype mismatches in SwitchTransformers/TimmWrapperModel." - }, - { - "left": "issue:34634", - "right": "issue:36010", - "accept": false, - "reason": "BarkProcessor voice_preset bug and GenerationMixin import failure are unrelated." - }, - { - "left": "issue:44485", - "right": "issue:44509", - "accept": false, - "reason": "GLM-5 RoPE implementation discussion is unrelated to stale docs mentioning removed pipeline tasks." - }, - { - "left": "issue:44479", - "right": "issue:45834", - "accept": false, - "reason": "Qwen video-input regression and Kosmos2.5 OCR long-input failure are different model/input paths." + "reason": "CPU offload device mismatch in bark-small and bfloat16 dtype mismatches in SwitchTransformers/TimmWrapper are different problems." } ] }, @@ -5713,37 +5650,36 @@ "source": "llm", "cache_hit": false, "model": "gpt-5.4-mini?service_tier=flex", - "cluster_id": "cluster-10000-472", + "cluster_id": "cluster-10000-473", "nodes": [ "issue:28218", "issue:29159", "issue:34634", + "issue:36010", "issue:36331", "issue:42175", - "issue:43262", "issue:43421", - "issue:43526", "issue:43540", "issue:43577", "issue:43582", - "issue:43653", - "issue:43697", + "issue:43645", "issue:43720", "issue:43749", "issue:43782", "issue:43818", - "issue:43819", "issue:43825", "issue:43827", "issue:43994", "issue:44291", + "issue:44479", + "issue:44485", + "issue:44509", "issue:44561", "issue:44610", "issue:44671", "issue:44805", "issue:44811", "issue:44857", - "issue:44912", "issue:44936", "issue:44977", "issue:44991", @@ -5752,21 +5688,25 @@ "issue:45072", "issue:45081", "issue:45325", - "issue:45372", "issue:45381", - "issue:45406", "issue:45701", + "issue:45834", "issue:45987", "issue:46192", "issue:46285", "issue:46291", "issue:46294", - "issue:46326", - "issue:46463" + "issue:46463", + "issue:46512", + "issue:46665" ], "soft_pairs": [ - "issue:43577|issue:43818", + "issue:43582|issue:46512", + "issue:44485|issue:44509", + "issue:34634|issue:36010", + "issue:44479|issue:45834", "issue:43825|issue:43994", + "issue:43577|issue:43818", "issue:43421|issue:44811", "issue:43782|issue:44977", "issue:44991|issue:46291", @@ -5781,215 +5721,214 @@ "issue:43825|issue:43827", "issue:45987|issue:46192", "issue:43749|issue:44291", + "issue:43645|issue:46665", "issue:44671|issue:45701", "issue:42175|issue:45081", - "issue:43582|issue:44857", "issue:34634|issue:36331", - "issue:43782|issue:45325", - "issue:43653|issue:44912", + "issue:43582|issue:44857", "issue:45003|issue:46285", - "issue:43262|issue:43819", - "issue:43526|issue:43697", - "issue:45081|issue:45406", - "issue:45372|issue:46326" + "issue:43782|issue:45325" ], - "prepared_review_unit_hash": "89626ee7fbff3e867ab7712795a90685992a09e4abc269a637352565319bf7e8", + "prepared_review_unit_hash": "7269b861276ff9bec4523ea7cbcb7cdbd93154b9c640cad2e14a8fce356c73f5", "component_budget": { - "node_count": 472, - "item_count": 472, - "soft_pair_count": 1732, - "serialized_chars": 898654, - "estimated_input_tokens": 224664, - "estimated_eval_tokens": 449584 + "node_count": 473, + "item_count": 473, + "soft_pair_count": 1738, + "serialized_chars": 901437, + "estimated_input_tokens": 225360, + "estimated_eval_tokens": 450976 }, "budget": { - "node_count": 48, - "item_count": 48, + "node_count": 47, + "item_count": 47, "soft_pair_count": 27, - "serialized_chars": 36202, - "estimated_input_tokens": 9051, - "estimated_eval_tokens": 18358 + "serialized_chars": 35483, + "estimated_input_tokens": 8871, + "estimated_eval_tokens": 17998 }, "overflow_policy": "truncate_then_skip", "trimmed": false, "aggressively_trimmed": false, "split": true, "analyst_result": { - "summary": "The cluster is heterogeneous: most items are distinct tokenizer/processor/model-loading regressions rather than true duplicates. I don\u2019t see any soft edge that clearly shares the same concrete bug or change, so I reject all proposed merges.", - "confidence": 0.86, - "canonical_issue_reason": "Issue 45701 is the broadest and most representative tokenizer-related report in this set, with the strongest general framing and the most inbound references among the tokenizer subtheme. The cluster is still too mixed to treat it as a true duplicate anchor, but this is the best issue-level representative.", + "summary": "This cluster is mostly a mix of unrelated transformer bugs (tokenizers, video/VL, loading, dtype, docs, trainer). The only plausible duplicate pair is the tokenizer spacing/inconsistency reports; the rest do not look like the same underlying bug.", + "confidence": 0.74, + "canonical_issue_reason": "issue:29159 is the broader tokenizer-behavior report and can subsume the more specific added-token spacing symptom in issue:28218.", "canonical_pr_reason": null, - "best_issue_reason": "45701 is the best overall issue candidate because it is broad enough to represent the tokenizer regression theme that appears repeatedly in the cluster, while remaining issue-formatted and well-scoped compared with the more specific bug reports.", + "best_issue_reason": "issue:29159 is the best representative issue in the cluster because it is the clearest umbrella for the only apparent duplicate theme here; the rest of the issues are disparate.", "best_pr_reason": null, "soft_edge_verdicts": [ { - "left": "issue:43577", - "right": "issue:43818", + "left": "issue:43582", + "right": "issue:46512", + "accept": false, + "reason": "Different failures: Apple Silicon allocator warmup TypeError vs stale config dtype after from_config." + }, + { + "left": "issue:44485", + "right": "issue:44509", + "accept": false, + "reason": "GLM-5 RoPE implementation discussion is unrelated to docs still mentioning removed pipelines." + }, + { + "left": "issue:34634", + "right": "issue:36010", "accept": false, - "reason": "Different bugs: BLIP2 dtype loading mismatch vs Video-LLaVA temporal-attention/weight-sharing problem." + "reason": "BarkProcessor voice_preset bug and a GenerationMixin import error are separate code paths." + }, + { + "left": "issue:44479", + "right": "issue:45834", + "accept": false, + "reason": "Both are vision/video-related, but they affect different models and different failure modes." }, { "left": "issue:43825", "right": "issue:43994", "accept": false, - "reason": "One is an error-message/docs concern; the other is model output correctness with pipeline/AutoModel. Not the same bug." + "reason": "One is an error-message/docs issue about removed pipeline tasks; the other is a model quality/inference bug." + }, + { + "left": "issue:43577", + "right": "issue:43818", + "accept": false, + "reason": "BLIP2 dtype loading and Video-LLaVA temporal-attention/weight-sharing are unrelated regressions." }, { "left": "issue:43421", "right": "issue:44811", - "accept": false, - "reason": "TokenizersBackend post-processor update at runtime vs Whisper batch_decode skip_special_tokens behavior. Related area, not same defect." + "accept": true, + "reason": "Both point to tokenizer/processor special-token handling being inconsistent; 29159 is the broader formulation of the same bug family." }, { "left": "issue:43782", "right": "issue:44977", "accept": false, - "reason": "Weight-only loading failure for Qwen3VL is unrelated to Qwen3.5 flash-attention generation issues." + "reason": "Qwen3VL weight_only loading error and Qwen3.5 flash-attention generation issues are different problems." }, { "left": "issue:44991", "right": "issue:46291", "accept": false, - "reason": "Tokenizer load failure for one model vs transformers.pipelines import failure due to kernels dependency constraints." + "reason": "Tokenizer loading for a specific model and a pipelines import failure from kernels v0.15 are different breakages." }, { "left": "issue:46294", "right": "issue:46463", "accept": false, - "reason": "Different code paths: Qwen3.5 video chat-template feature mismatch vs T5 decoder input validation error message." + "reason": "Qwen3.5 video chat-template mismatch is unrelated to T5 decoder-input validation messaging." }, { "left": "issue:44805", "right": "issue:44936", "accept": false, - "reason": "Mask/index shape mismatch is unrelated to trainer.evaluate() failing after train()." + "reason": "Mask/index shape mismatch and trainer.evaluate() after train() are not the same bug." }, { "left": "issue:43540", "right": "issue:45381", "accept": false, - "reason": "Both are video-related, but they describe different models and different failure modes." + "reason": "Both involve video inputs, but they target different models and likely different preprocessing/position-id logic." }, { "left": "issue:44561", "right": "issue:45020", "accept": false, - "reason": "Both involve remote_code breakage in v5, but one is a specific removed API and the other is a broad umbrella report." + "reason": "45020 is a broad umbrella about remote_code breakage; 44561 is a specific helper-removal regression, not clearly the same concrete bug." }, { "left": "issue:44610", "right": "issue:45072", "accept": false, - "reason": "Processor input-size mismatch is unrelated to dtype mismatch in bfloat16 inference." + "reason": "Processor image-size mismatch and bfloat16 dtype mismatches are unrelated." }, { "left": "issue:43720", "right": "issue:45072", "accept": false, - "reason": "Packed-weight unpacking during loading is a different problem from dtype mismatches." + "reason": "BitNet packed-weight loading and dtype mismatch inference failures are different issues." }, { "left": "issue:28218", "right": "issue:29159", - "accept": false, - "reason": "Potentially related tokenizer behavior, but one is a specific added-token spacing bug and the other is a broader slow/fast inconsistency report." + "accept": true, + "reason": "These describe the same tokenizer inconsistency around added tokens/spacing; 29159 is the broader duplicate umbrella." }, { "left": "issue:43577", "right": "issue:43582", "accept": false, - "reason": "BLIP2 dtype persistence and Apple Silicon allocator TypeError are unrelated." + "reason": "BLIP2 dtype handling and Apple Silicon allocator warmup TypeError are unrelated." }, { "left": "issue:43825", "right": "issue:43827", "accept": false, - "reason": "Error-message behavior vs outdated docs references; same area, different change requests." + "reason": "Both mention removed pipeline tasks, but one is runtime messaging and the other is stale docs; not the same bug." }, { "left": "issue:45987", "right": "issue:46192", "accept": false, - "reason": "StaticCache return type contract issue vs return_tensors being ignored. Different API bugs." + "reason": "StaticCache sequence-length contract and return_tensors being ignored are separate API bugs." }, { "left": "issue:43749", "right": "issue:44291", "accept": false, - "reason": "FSDP CPU RAM efficient loading regression is not the same as init_empty_weights receiving an unexpected argument." + "reason": "FSDP CPU RAM efficient loading and init_empty_weights argument handling are different loading paths." + }, + { + "left": "issue:43645", + "right": "issue:46665", + "accept": false, + "reason": "Jupyter custom-model initialization and from_pretrained inside Python classes do not appear to be the same concrete issue." }, { "left": "issue:44671", "right": "issue:45701", "accept": false, - "reason": "CamemBERT masked-LM prediction regression is not the same as a generic tokenization-change report." + "reason": "Masked LM prediction regression and tokenization changes are related only broadly, not the same underlying bug." }, { "left": "issue:42175", "right": "issue:45081", "accept": false, - "reason": "Backend install dependency problem vs Mistral regex patch crash. No shared code-path bug." - }, - { - "left": "issue:43582", - "right": "issue:44857", - "accept": false, - "reason": "Both are runtime errors, but on different platforms and from different subsystems." + "reason": "Backend dependency packaging and Mistral regex patching are unrelated." }, { "left": "issue:34634", "right": "issue:36331", "accept": false, - "reason": "BarkProcessor voice_preset issue and CustomTrainer compute_loss signature mismatch are unrelated." - }, - { - "left": "issue:43782", - "right": "issue:45325", - "accept": false, - "reason": "Qwen3VL weight_only loading error is unrelated to Qwen2.5-VL rope/position-id scaling." + "reason": "BarkProcessor voice_preset and CustomTrainer.compute_loss signature breakage are unrelated." }, { - "left": "issue:43653", - "right": "issue:44912", + "left": "issue:43582", + "right": "issue:44857", "accept": false, - "reason": "BigBirdTokenizer special-token registration bug is unrelated to MXFP4 quantization fallback." + "reason": "Allocator warmup on Apple Silicon and CUDA AMP loss crashes are unrelated runtime failures." }, { "left": "issue:45003", "right": "issue:46285", "accept": false, - "reason": "Unsafe sys.modules access is a different bug from CLIPTextModel structural inconsistency." - }, - { - "left": "issue:43262", - "right": "issue:43819", - "accept": false, - "reason": "Audio chat-template sample-rate defaulting and DAC latents/forward mismatch are different code paths." - }, - { - "left": "issue:43526", - "right": "issue:43697", - "accept": false, - "reason": "BeitImageProcessorFast reduce_labels bug is not the same as RTDetrV2 output differences in v5." - }, - { - "left": "issue:45081", - "right": "issue:45406", - "accept": false, - "reason": "Mistral tokenizer regex patch crash and Gemma4Processor missing _tokenizer are different loading failures." + "reason": "Unsafe sys.modules access and CLIP text-model structure drift are different code changes." }, { - "left": "issue:45372", - "right": "issue:46326", + "left": "issue:43782", + "right": "issue:45325", "accept": false, - "reason": "Gemma 4 mistral_common import failure is unrelated to continue_final_message behavior for mistral_common tokenizers." + "reason": "Qwen3VL weight_only loading and Qwen2.5-VL rope-index scaling are different bugs." } ] }, - "evaluator_result": null, - "evaluator_used": false, + "evaluator_result": { + "accept": true, + "feedback": "The summary is grounded in the packet and stays conservative overall. The chosen canonical issue, issue:29159, reasonably serves as the broader umbrella for issue:28218\u2019s added-token spacing symptom, and the rejection calls on the other pairs do not overclaim shared bug identity from title-level evidence. The one accepted soft pair (issue:43421/issue:44811) is a bit broad, but still defensible as a tokenizer/processor special-token handling family given the available evidence." + }, + "evaluator_used": true, "retried": false, - "accepted_nontrivial_soft_edge": false, + "accepted_nontrivial_soft_edge": true, "error_kind": null, "error_message": null }, @@ -6004,7 +5943,7 @@ "source": "llm", "cache_hit": false, "model": "gpt-5.4-mini?service_tier=flex", - "cluster_id": "cluster-10000-472", + "cluster_id": "cluster-10000-473", "nodes": [ "issue:36010", "issue:36331", @@ -6020,9 +5959,9 @@ "issue:43531", "issue:43550", "issue:43638", - "issue:43643", "issue:43645", - "issue:43696", + "issue:43653", + "issue:43697", "issue:43720", "issue:43742", "issue:43746", @@ -6044,25 +5983,29 @@ "issue:44991", "issue:45020", "issue:45042", + "issue:45081", "issue:45084", - "issue:45459", + "issue:45372", + "issue:45406", "issue:45538", "issue:45600", - "issue:45759", - "issue:45803", "issue:46209", "issue:46291", - "issue:46421", - "issue:46424" + "issue:46326" ], "soft_pairs": [ + "issue:43653|issue:44912", + "issue:43262|issue:43819", + "issue:43526|issue:43697", + "issue:45081|issue:45406", + "issue:45372|issue:46326", "issue:44589|issue:45042", "issue:43329|issue:45600", - "issue:43742|issue:45020", "issue:44060|issue:44368", + "issue:43742|issue:45020", "issue:45084|issue:45538", - "issue:43550|issue:44912", "issue:43526|issue:44554", + "issue:43550|issue:44912", "issue:43906|issue:46209", "issue:43452|issue:44991", "issue:43638|issue:44936", @@ -6071,216 +6014,211 @@ "issue:36010|issue:36331", "issue:36010|issue:43645", "issue:42175|issue:46291", - "issue:43720|issue:44610", "issue:43531|issue:45042", + "issue:43720|issue:44610", "issue:38175|issue:44610", - "issue:44387|issue:44617", "issue:43746|issue:44811", + "issue:44387|issue:44617", "issue:39692|issue:44912", "issue:42490|issue:43257", - "issue:43854|issue:44977", - "issue:43643|issue:45759", - "issue:46421|issue:46424", - "issue:43262|issue:44811", - "issue:45459|issue:45803", - "issue:43257|issue:43696" + "issue:43854|issue:44977" ], - "prepared_review_unit_hash": "d3e6e829b9063dffc4e9ddf57fc9a196760f1adee63ac937dcd47d02d02c7271", + "prepared_review_unit_hash": "2927d5b87f3b678ee75d52caf24414332d01799f95c1dea5bfeb42dcc35139f6", "component_budget": { - "node_count": 472, - "item_count": 472, - "soft_pair_count": 1732, - "serialized_chars": 898654, - "estimated_input_tokens": 224664, - "estimated_eval_tokens": 449584 + "node_count": 473, + "item_count": 473, + "soft_pair_count": 1738, + "serialized_chars": 901437, + "estimated_input_tokens": 225360, + "estimated_eval_tokens": 450976 }, "budget": { - "node_count": 48, - "item_count": 48, + "node_count": 47, + "item_count": 47, "soft_pair_count": 28, - "serialized_chars": 36344, - "estimated_input_tokens": 9086, - "estimated_eval_tokens": 18428 + "serialized_chars": 35879, + "estimated_input_tokens": 8970, + "estimated_eval_tokens": 18196 }, "overflow_policy": "truncate_then_skip", "trimmed": false, "aggressively_trimmed": false, "split": true, "analyst_result": { - "summary": "This set is mostly a collection of unrelated Transformers issue reports. The soft pairs share broad themes (loading, cache, warnings, quantization), but none look like the same concrete bug/change closely enough to merge.", - "confidence": 0.86, - "canonical_issue_reason": "issue:45020 is the broadest, highest-signal regression in the set: it describes a general remote_code loading breakage that is more central than the narrower model- or feature-specific reports.", + "summary": "This cluster is mostly a set of unrelated Transformers bugs that only share broad themes like loading, tokenizers, quantization, or model-family regressions. I would not merge any of the soft pairs; the best umbrella issue is the broad remote_code regression report, but even that is only a loose fit for the rest.", + "confidence": 0.87, + "canonical_issue_reason": "Issue 45020 is the broadest, most umbrella-like report in the set: it describes recent-version regressions breaking model loading via `remote_code`, which is the closest shared theme across several other reports. The rest are much more specific model/component bugs and do not form a true duplicate set.", "canonical_pr_reason": null, - "best_issue_reason": "issue:45020 is the best standalone representative because it is actionable, user-facing, and covers a wider class of loading regressions than the other candidates.", + "best_issue_reason": "45020 is the best representative because it captures the highest-level regression pattern and is more general than the other issues, which are mostly narrow, model-specific failures.", "best_pr_reason": null, "soft_edge_verdicts": [ + { + "left": "issue:43653", + "right": "issue:44912", + "accept": false, + "reason": "BigBird tokenizer special-token decoding bug and MXFP4 quantization fallback are unrelated failures." + }, + { + "left": "issue:43262", + "right": "issue:43819", + "accept": false, + "reason": "Audio chat-template sampling-rate defaulting and DAC latent reconstruction mismatch are different code paths and symptoms." + }, + { + "left": "issue:43526", + "right": "issue:43697", + "accept": false, + "reason": "BEiT label reduction and RTDetrV2 output drift are distinct vision-model issues." + }, + { + "left": "issue:45081", + "right": "issue:45406", + "accept": false, + "reason": "Both involve tokenizer/processor loading, but one is a Mistral regex/backend_tokenizer crash and the other is a Gemma4Processor missing `_tokenizer`; different bugs." + }, + { + "left": "issue:45372", + "right": "issue:46326", + "accept": false, + "reason": "ReasoningEffort import failure and `continue_final_message` behavior are separate mistral_common-related problems, not the same underlying bug." + }, { "left": "issue:44589", "right": "issue:45042", "accept": false, - "reason": "Different failures: a float8 storage/import error versus an image-processor torchvision dependency regression." + "reason": "Float8 storage lookup failure and PIL backend requiring torchvision are unrelated runtime/dependency issues." }, { "left": "issue:43329", "right": "issue:45600", "accept": false, - "reason": "Unrelated bugs in different subsystems: multimodal token counting vs removed auto-mapping config references." + "reason": "Undefined video-token helper variables and stale auto_mappings config references are different regressions." }, { - "left": "issue:43742", - "right": "issue:45020", + "left": "issue:44060", + "right": "issue:44368", "accept": false, - "reason": "Both involve loading, but one is a MobileLLM key error and the other is a broad remote_code regression; not the same bug." + "reason": "Both are warning-related on Qwen models, but one is a tied-weights warning bug and the other is a `tie_word_embeddings=False` warning during LoRA fine-tuning; not the same fix." }, { - "left": "issue:44060", - "right": "issue:44368", + "left": "issue:43742", + "right": "issue:45020", "accept": false, - "reason": "Both mention Qwen tied-weight warnings, but the specific symptoms and model paths differ enough that they do not look like one duplicate bug." + "reason": "A specific MobileLLM key error is not the same as the broader `remote_code` loading regression." }, { "left": "issue:45084", "right": "issue:45538", "accept": false, - "reason": "Unrelated: template compilation failure versus an overly large tokenizer max_length setting." + "reason": "Template compilation errors and CLIPTokenizer `model_max_length` behavior are unrelated." }, { - "left": "issue:43550", - "right": "issue:44912", + "left": "issue:43526", + "right": "issue:44554", "accept": false, - "reason": "Different code paths and failures: torch.compile/SDPA runtime bug versus MXFP4 loading fallback behavior." + "reason": "BEiT label processing and MPS attention correctness are different subsystems." }, { - "left": "issue:43526", - "right": "issue:44554", + "left": "issue:43550", + "right": "issue:44912", "accept": false, - "reason": "Different models and mechanisms: BEiT label reduction bug versus MPS attention correctness issue." + "reason": "Bamba SDPA/torch.compile failure and MXFP4 quantization fallback are separate issues." }, { "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 underlying problem." + "reason": "An isolated reproduction of a prior issue is not the same as the FP8 inference failure on B200." }, { "left": "issue:43452", "right": "issue:44991", "accept": false, - "reason": "Both are loading-related, but gguf_file breakage and a specific tokenizer failure are distinct bugs." + "reason": "gguf_file handling for tokenizer/model loading and a tokenizer-loading regression for est-roberta are both loading issues, but they have different root causes and are not clearly the same bug." }, { "left": "issue:43638", "right": "issue:44936", "accept": false, - "reason": "Different training/evaluation failures with different triggers; no clear shared code-path bug." + "reason": "Zero3/non-pretrained Bert index error and `evaluate()` failing after `train()` are distinct trainer failures." }, { "left": "issue:43257", "right": "issue:43329", "accept": false, - "reason": "Separate Qwen/deepspeed loading issue versus a multimodal token-counting bug." + "reason": "Qwen3-MoE weight conversion with accelerate/deepspeed and multimodal token counting are unrelated." }, { "left": "issue:43819", "right": "issue:43927", "accept": false, - "reason": "Different model families and failure modes: DAC latent mismatch versus DiaConfig token-ID persistence." + "reason": "DAC latent forward mismatch and DiaConfig losing custom token IDs are different model/config bugs." }, { "left": "issue:36010", "right": "issue:36331", "accept": false, - "reason": "Import path regression and trainer API signature change are unrelated." + "reason": "`GenerationMixin` import failure and `compute_loss()` kwarg incompatibility are unrelated API breakages." }, { "left": "issue:36010", "right": "issue:43645", "accept": false, - "reason": "Both are API-breakage reports, but one is a GenerationMixin import issue and the other is custom-model initialization in notebooks." + "reason": "Importing `GenerationMixin` and notebook custom-model initialization failures do not describe the same problem." }, { "left": "issue:42175", "right": "issue:46291", "accept": false, - "reason": "Packaging/backend selection issue versus pipeline import failing on a kernels dependency requirement." - }, - { - "left": "issue:43720", - "right": "issue:44610", - "accept": false, - "reason": "Packed-weight loading bug versus a processor/model input-size mismatch; unrelated." + "reason": "Missing TensorFlow from the `.[torch]` extra and `transformers.pipelines` breaking on kernels v0.15 are separate dependency issues." }, { "left": "issue:43531", "right": "issue:45042", "accept": false, - "reason": "Different domains: Qwen3-MoE sliding-window handling versus PIL image-processor dependency requirements." + "reason": "Qwen3-MoE sliding-window handling and PIL image processor torchvision requirements are unrelated." }, { - "left": "issue:38175", + "left": "issue:43720", "right": "issue:44610", "accept": false, - "reason": "Different model-specific correctness issue versus an OmDet-Turbo preprocessing size mismatch." + "reason": "BitNet packed-weight loading and OmDet-Turbo processor size mismatch are different code paths." }, { - "left": "issue:44387", - "right": "issue:44617", + "left": "issue:38175", + "right": "issue:44610", "accept": false, - "reason": "Both mention OOM, but one is a Transformers 5.x int4 memory regression and the other is a model-specific video OOM report." + "reason": "SigLIP2 zero probabilities and OmDet-Turbo input-size mismatch are unrelated model bugs." }, { "left": "issue:43746", "right": "issue:44811", "accept": false, - "reason": "PEFT/local-checkpoint loading bug and Whisper batch_decode token handling are unrelated." + "reason": "GraniteSpeech PEFT local-checkpoint loading and Whisper batch_decode skipping special tokens are not the same issue." + }, + { + "left": "issue:44387", + "right": "issue:44617", + "accept": false, + "reason": "A Transformers 5.x int4 memory regression and Sam3Video OOM are different causes." }, { "left": "issue:39692", "right": "issue:44912", "accept": false, - "reason": "Both involve model loading/quantization examples, but they are different models and failure modes." + "reason": "SigLIP2 documentation example errors and MXFP4 quantization fallback are unrelated." }, { "left": "issue:42490", "right": "issue:43257", "accept": false, - "reason": "Slash-path import warning and Qwen3 MoE weight conversion are unrelated loading bugs." + "reason": "Trailing-slash local model import warning and Qwen3 MoE weight conversion are unrelated." }, { "left": "issue:43854", "right": "issue:44977", "accept": false, - "reason": "Different model-specific generation/loading failures with no shared concrete code path." - }, - { - "left": "issue:43643", - "right": "issue:45759", - "accept": false, - "reason": "Both touch config handling, but one is trust_remote_code field loss and the other is composite multimodal config unwrapping." - }, - { - "left": "issue:46421", - "right": "issue:46424", - "accept": false, - "reason": "Both are cache-related, but one is about lazy initialization causing recompilation and the other about max_cache_length semantics." - }, - { - "left": "issue:43262", - "right": "issue:44811", - "accept": false, - "reason": "Audio chat-template sampling-rate defaulting and Whisper skip_special_tokens handling are different processor bugs." - }, - { - "left": "issue:45459", - "right": "issue:45803", - "accept": false, - "reason": "Both involve overly broad exception handling, but they are distinct code paths and different error surfaces." - }, - { - "left": "issue:43257", - "right": "issue:43696", - "accept": false, - "reason": "Qwen3 MoE loading/conversion issue versus a general GPT-oss OOM report; not the same bug." + "reason": "GLM-4.7-Flash loading in tests and Qwen3.5 flash-attention generation are different model-specific bugs." } ] }, @@ -6302,34 +6240,30 @@ "source": "llm", "cache_hit": false, "model": "gpt-5.4-mini?service_tier=flex", - "cluster_id": "cluster-10000-472", + "cluster_id": "cluster-10000-473", "nodes": [ "issue:28282", - "issue:29127", "issue:29159", - "issue:33357", "issue:34567", "issue:34689", "issue:40990", - "issue:42175", "issue:42915", "issue:43257", - "issue:43317", + "issue:43262", "issue:43329", "issue:43404", "issue:43425", "issue:43450", - "issue:43452", "issue:43454", "issue:43525", "issue:43577", - "issue:43582", "issue:43606", + "issue:43643", "issue:43650", + "issue:43696", "issue:43697", "issue:43742", "issue:43749", - "issue:43824", "issue:43906", "issue:43927", "issue:43931", @@ -6337,8 +6271,8 @@ "issue:44479", "issue:44568", "issue:44756", + "issue:44811", "issue:44821", - "issue:44849", "issue:44912", "issue:44987", "issue:45042", @@ -6347,19 +6281,27 @@ "issue:45356", "issue:45357", "issue:45405", + "issue:45459", "issue:45538", - "issue:45701", + "issue:45759", + "issue:45803", "issue:45859", "issue:46291", - "issue:46512" + "issue:46421", + "issue:46424" ], "soft_pairs": [ + "issue:46421|issue:46424", + "issue:43262|issue:44811", + "issue:43643|issue:45759", + "issue:43257|issue:43696", + "issue:45459|issue:45803", "issue:43329|issue:44479", - "issue:43257|issue:43931", "issue:43650|issue:43906", + "issue:43257|issue:43931", "issue:43329|issue:43450", - "issue:43606|issue:45859", "issue:42915|issue:43742", + "issue:43606|issue:45859", "issue:43749|issue:44756", "issue:34567|issue:40990", "issue:43454|issue:43577", @@ -6375,224 +6317,192 @@ "issue:43425|issue:46291", "issue:42915|issue:45538", "issue:28282|issue:34689", - "issue:45137|issue:45405", - "issue:33357|issue:34567", - "issue:42175|issue:43824", - "issue:43577|issue:46512", - "issue:43452|issue:43582", - "issue:28282|issue:29127", - "issue:43317|issue:45538", - "issue:43742|issue:44291", - "issue:44849|issue:45701" + "issue:45137|issue:45405" ], - "prepared_review_unit_hash": "ae33743e8565715c06edcd0c783f06d75ca57f417cbbdb466e3894603aeba873", + "prepared_review_unit_hash": "23eed1b7ecc7cf81e573051068467c36f2e104f7e6dd2a7b05ca54b71425b410", "component_budget": { - "node_count": 472, - "item_count": 472, - "soft_pair_count": 1732, - "serialized_chars": 898654, - "estimated_input_tokens": 224664, - "estimated_eval_tokens": 449584 + "node_count": 473, + "item_count": 473, + "soft_pair_count": 1738, + "serialized_chars": 901437, + "estimated_input_tokens": 225360, + "estimated_eval_tokens": 450976 }, "budget": { - "node_count": 48, - "item_count": 48, - "soft_pair_count": 30, - "serialized_chars": 36740, - "estimated_input_tokens": 9185, - "estimated_eval_tokens": 18626 + "node_count": 47, + "item_count": 47, + "soft_pair_count": 27, + "serialized_chars": 35405, + "estimated_input_tokens": 8852, + "estimated_eval_tokens": 17960 }, "overflow_policy": "truncate_then_skip", "trimmed": false, "aggressively_trimmed": false, "split": true, "analyst_result": { - "summary": "This cluster is mostly a loose set of unrelated bug reports; the only noticeable overlap is a small dtype/state-propagation theme, but even those are not the same concrete failure path. No PRs are present.", - "confidence": 0.74, - "canonical_issue_reason": "issue:46512 is the cleanest current dtype/state bug report: it is narrowly scoped, has a clear repro, and is the best potential target for nearby configuration/dtype complaints.", + "summary": "This is not a true duplicate cluster: the items span unrelated tokenizer, multimodal, loading, training, and cache bugs. None of the soft pairs look like the same underlying fixable change, so I would reject all of them; 46421 is only the least-bad anchor because it is current and broadly framed.", + "confidence": 0.81, + "canonical_issue_reason": "Issue 46421 is the best anchor only in a loose sense: it is current, open, and broadly scoped around cache behavior, whereas the rest are mostly unrelated one-off bugs.", "canonical_pr_reason": null, - "best_issue_reason": "issue:46512 is the best issue to anchor the cluster because it is open, specific, and describes a core model-construction state bug more cleanly than the others.", + "best_issue_reason": "Issue 46421 is the most representative single issue here, but the set is too heterogeneous to treat as a real duplicate cluster.", "best_pr_reason": null, "soft_edge_verdicts": [ { - "left": "issue:43329", - "right": "issue:44479", + "left": "issue:46421", + "right": "issue:46424", + "accept": false, + "reason": "Both are cache-related, but one is about chunked-prefill recompilation and the other about max_cache_length handling; different bugs and likely different fixes." + }, + { + "left": "issue:43262", + "right": "issue:44811", + "accept": false, + "reason": "Different subsystems and symptoms: audio chat-template sampling rate vs Whisper batch_decode special-token handling." + }, + { + "left": "issue:43643", + "right": "issue:45759", "accept": false, - "reason": "Both involve video/VLM regressions, but one is an undefined-variable bug in token counting and the other is a broader video-input regression in different models." + "reason": "Both involve config/model-loading logic, but one is missing fields from trust_remote_code and the other is composite config unwrapping; not the same bug." }, { "left": "issue:43257", - "right": "issue:43931", + "right": "issue:43696", + "accept": false, + "reason": "Qwen3 MoE weight conversion/loading bug vs GPT-oss OOM; unrelated failure modes." + }, + { + "left": "issue:45459", + "right": "issue:45803", + "accept": false, + "reason": "Both mention hiding errors, but they are different code paths and different root causes; not the same bug." + }, + { + "left": "issue:43329", + "right": "issue:44479", "accept": false, - "reason": "Different failure modes: one is MOE weight conversion with accelerate+deepspeed, the other is a shape-mismatch load error for a specific Qwen3-VL model." + "reason": "Both are video-related, but one is an undefined-variable bug in multimodal token counting while the other is a broader v5.3 video-input regression across models." }, { "left": "issue:43650", "right": "issue:43906", "accept": false, - "reason": "No shared bug; one is a stray low-quality issue, the other is a reproduction of a separate tracked problem." + "reason": "No shared concrete bug is visible; one is non-informative and the other is an isolated reproduction of a separate issue." }, { - "left": "issue:43329", - "right": "issue:43450", + "left": "issue:43257", + "right": "issue:43931", "accept": false, - "reason": "Both touch video processing, but one is a missing-variable bug in multimodal token counting and the other is incorrect batched video processor shapes." + "reason": "Different Qwen3 problems: MoE weight conversion under accelerate/deepspeed vs Qwen3-VL weight-shape mismatch on load." }, { - "left": "issue:43606", - "right": "issue:45859", + "left": "issue:43329", + "right": "issue:43450", "accept": false, - "reason": "Both mention CPU offload, but they affect different models and different code paths; not the same concrete bug." + "reason": "Both touch video processing, but one is a token-count helper bug and the other is incorrect batched output shapes; different fixes." }, { "left": "issue:42915", "right": "issue:43742", "accept": false, - "reason": "Both are model-loading failures, but the models and root causes are unrelated." + "reason": "Unrelated model-loading failures with different models and error classes." + }, + { + "left": "issue:43606", + "right": "issue:45859", + "accept": false, + "reason": "Both involve CPU offload, but they are different models and different incompatibilities; not one duplicate bug." }, { "left": "issue:43749", "right": "issue:44756", "accept": false, - "reason": "One is FSDP CPU-RAM-efficient loading broken; the other is mmap causing OOM on Strix Halo. Different subsystems and symptoms." + "reason": "Different memory-related problems: FSDP RAM-efficient loading vs disabling mmap on Strix Halo." }, { "left": "issue:34567", "right": "issue:40990", "accept": false, - "reason": "Unrelated: token-count tracking versus perplexity quality on a specific model/dataset." + "reason": "Training-state token counting bug vs high perplexity on a model/eval setup; no shared underlying issue." }, { "left": "issue:43454", "right": "issue:43577", "accept": false, - "reason": "Both involve model internals, but one is untied lm_head weights in Mistral3 and the other is dtype not propagating to a Blip2 model." + "reason": "Weight-tying failure causing bad generation vs dtype propagation staying float32; unrelated issues." }, { "left": "issue:44291", "right": "issue:45042", "accept": false, - "reason": "TypeError from init_empty_weights is unrelated to PIL backend requiring torchvision." + "reason": "Init_empty_weights argument compatibility vs torchvision dependency requirement in PIL image processors; different code paths." }, { "left": "issue:42915", "right": "issue:44821", "accept": false, - "reason": "Different bugs: FineGrainedFP8Config failure versus loading AutoImageProcessor from a URL." + "reason": "Completely different areas: FP8 MoE loading vs loading AutoImageProcessor from URL." }, { "left": "issue:43525", "right": "issue:46291", "accept": false, - "reason": "No overlap in code path: a missing pad_token_id on Llama4Config versus kernels import/version requirements for pipelines." + "reason": "Llama4 config field missing vs pipelines import failing on kernels version/revision requirements; unrelated." }, { "left": "issue:29159", "right": "issue:34567", "accept": false, - "reason": "Tokenizer slow/fast inconsistency is unrelated to TrainerState token counting." + "reason": "Tokenizer slow/fast inconsistency vs TrainerState token-count updating; different components and symptoms." }, { "left": "issue:43927", "right": "issue:44912", "accept": false, - "reason": "Different issues: save/load loses custom token IDs versus MXFP4 quantization fallback during model loading." + "reason": "Saving/loading custom token IDs vs MXFP4 quantization fallback; not the same underlying bug." }, { "left": "issue:44568", "right": "issue:45356", "accept": false, - "reason": "Both are tokenizer regressions, but they hit different models and different tokenizer behaviors." - }, - { - "left": "issue:44987", - "right": "issue:45538", - "accept": false, - "reason": "Loading failure for a specific model is unrelated to CLIPTokenizer's max_length sentinel bug." + "reason": "Both are tokenizer regressions, but they affect different models and appear to stem from different tokenizer logic." }, { "left": "issue:43697", "right": "issue:45357", "accept": false, - "reason": "Different areas: RTDetrV2 output drift versus Qwen3.5 save_pretrained visual-key regression." + "reason": "Detection-model output regression vs Qwen3.5 save_pretrained key serialization; unrelated." }, { "left": "issue:43404", "right": "issue:45127", "accept": false, - "reason": "Both mention tied weights, but one is a Mistral3 lm_head tying bug and the other is LoRA merge collapse with extended vocabulary." + "reason": "Weight tying is mentioned in both, but one is a missing tie in model init and the other is LoRA merge collapse with extended vocab; different bugs." }, { "left": "issue:43425", "right": "issue:46291", "accept": false, - "reason": "Torch version incompatibility is not the same as kernels' version/revision requirement during pipeline import." + "reason": "Torch version incompatibility vs kernels dependency/version-revision import failure; unrelated." }, { "left": "issue:42915", "right": "issue:45538", "accept": false, - "reason": "Unrelated model-loading/configuration issues." + "reason": "FP8 MoE load failure vs tokenizer model_max_length constant misuse; no shared code path." }, { "left": "issue:28282", "right": "issue:34689", "accept": false, - "reason": "ImportError for missing PyTorch is unrelated to a model-loading breakage in Llama 3.2 Vision." + "reason": "Different load-time dependency/model-loading problems with different affected models and environments." }, { "left": "issue:45137", "right": "issue:45405", "accept": false, - "reason": "DeepSpeed ZeRO3 deque error is unrelated to a PEFT version bump issue." - }, - { - "left": "issue:33357", - "right": "issue:34567", - "accept": false, - "reason": "MacOS bus error on a CLIP model is unrelated to TrainerState token counting." - }, - { - "left": "issue:42175", - "right": "issue:43824", - "accept": false, - "reason": "TensorFlow backend packaging issue is unrelated to a missing Qwen2_5_VL import." - }, - { - "left": "issue:43577", - "right": "issue:46512", - "accept": false, - "reason": "Related by dtype handling, but one is an actual model/qformer dtype mismatch on load while the other is config.dtype staying stale after from_config; not the same bug." - }, - { - "left": "issue:43452", - "right": "issue:43582", - "accept": false, - "reason": "Different failures: gguf loading/AutoTokenizer versus an Apple Silicon TypeError in caching_allocator_warmup." - }, - { - "left": "issue:28282", - "right": "issue:29127", - "accept": false, - "reason": "Missing PyTorch import versus a LayoutLMv3 error-message clarity issue are unrelated." - }, - { - "left": "issue:43317", - "right": "issue:45538", - "accept": false, - "reason": "device_map/offload loading failure is unrelated to CLIPTokenizer's model_max_length sentinel." - }, - { - "left": "issue:43742", - "right": "issue:44291", - "accept": false, - "reason": "A key error when loading a model is unrelated to init_empty_weights rejecting an unexpected argument." - }, - { - "left": "issue:44849", - "right": "issue:45701", - "accept": false, - "reason": "Different tokenizer/model behaviors: Qwen3.5 hidden-states bug versus version-dependent tokenization changes." + "reason": "DeepSpeed ZeRO3 deque error vs an upper-bound version bump for PEFT; unrelated." } ] }, @@ -6614,12 +6524,14 @@ "source": "llm", "cache_hit": false, "model": "gpt-5.4-mini?service_tier=flex", - "cluster_id": "cluster-10000-472", + "cluster_id": "cluster-10000-473", "nodes": [ "issue:24643", "issue:27957", "issue:28282", + "issue:29127", "issue:29159", + "issue:33357", "issue:34567", "issue:36331", "issue:38175", @@ -6627,14 +6539,16 @@ "issue:42175", "issue:42915", "issue:43257", + "issue:43317", + "issue:43452", "issue:43493", - "issue:43531", "issue:43540", "issue:43577", + "issue:43582", "issue:43742", "issue:43749", "issue:43756", - "issue:43906", + "issue:43824", "issue:43994", "issue:44062", "issue:44291", @@ -6646,32 +6560,37 @@ "issue:44737", "issue:44792", "issue:44821", + "issue:44849", "issue:44898", "issue:44964", "issue:44987", "issue:44991", - "issue:45042", "issue:45072", "issue:45081", - "issue:45325", "issue:45381", "issue:45488", - "issue:45588", - "issue:45823", + "issue:45538", + "issue:45701", "issue:45907", - "issue:46399", "issue:46463", - "issue:46468", - "issue:46491" + "issue:46512" ], "soft_pairs": [ + "issue:42175|issue:43824", + "issue:33357|issue:34567", + "issue:43577|issue:46512", + "issue:43452|issue:43582", + "issue:43317|issue:45538", + "issue:28282|issue:29127", + "issue:43742|issue:44291", "issue:44062|issue:44991", "issue:41720|issue:43540", + "issue:44849|issue:45701", "issue:42915|issue:43749", "issue:43540|issue:44560", "issue:44479|issue:45381", - "issue:24643|issue:29159", "issue:43540|issue:44479", + "issue:24643|issue:29159", "issue:27957|issue:28282", "issue:42175|issue:44488", "issue:38175|issue:44898", @@ -6679,230 +6598,222 @@ "issue:44987|issue:45907", "issue:43493|issue:43994", "issue:44291|issue:44821", - "issue:34567|issue:36331", "issue:43257|issue:43756", + "issue:34567|issue:36331", "issue:44737|issue:46463", "issue:44792|issue:44964", "issue:43577|issue:45072", "issue:44062|issue:45081", - "issue:43742|issue:44589", - "issue:43756|issue:46399", - "issue:45042|issue:45081", - "issue:44568|issue:46491", - "issue:43756|issue:45325", - "issue:43906|issue:46468", - "issue:44737|issue:45588", - "issue:45823|issue:46463", - "issue:43531|issue:44821" + "issue:43742|issue:44589" ], - "prepared_review_unit_hash": "3401f8517c9c7adb4659846c555320b9241d1a047a6a50901af2511ffe20a6cb", + "prepared_review_unit_hash": "c632403f8b32e856d89fb7b646fe00d4fda3b7e86d44c05ca308b1869d197435", "component_budget": { - "node_count": 472, - "item_count": 472, - "soft_pair_count": 1732, - "serialized_chars": 898654, - "estimated_input_tokens": 224664, - "estimated_eval_tokens": 449584 + "node_count": 473, + "item_count": 473, + "soft_pair_count": 1738, + "serialized_chars": 901437, + "estimated_input_tokens": 225360, + "estimated_eval_tokens": 450976 }, "budget": { - "node_count": 47, - "item_count": 47, - "soft_pair_count": 29, - "serialized_chars": 35950, - "estimated_input_tokens": 8988, - "estimated_eval_tokens": 18232 + "node_count": 48, + "item_count": 48, + "soft_pair_count": 29, + "serialized_chars": 36483, + "estimated_input_tokens": 9121, + "estimated_eval_tokens": 18498 }, "overflow_policy": "truncate_then_skip", "trimmed": false, "aggressively_trimmed": false, "split": true, "analyst_result": { - "summary": "Mostly a heterogeneous set of unrelated Transformers regressions; only one soft edge looks like the same underlying bug: the Qwen2.5-VL video-input regression in v5.3.0.", - "confidence": 0.84, - "canonical_issue_reason": "issue:44479 is the broad parent report for the Qwen2.5-VL/Qwen3-VL/Qwen3.5 video regression, while issue:45381 is a narrower symptom of the same v5.3.0 position-id bug.", + "summary": "The cluster is fairly heterogeneous, with most items only loosely related by being model/tokenizer/load regressions. I found two clear duplicate-like subgroups: Qwen2.5-VL video regression (44479/45381) and SigLIP2 wrong outputs (43493/43994).", + "confidence": 0.61, + "canonical_issue_reason": "issue:44479 is the strongest duplicate anchor in the set: it describes a broad Qwen video-input regression and directly subsumes the narrower 45381 report.", "canonical_pr_reason": null, - "best_issue_reason": "issue:44479 is the best representative issue here: broad scope, clear regression framing, and it subsumes the one likely duplicate-like follow-up.", + "best_issue_reason": "issue:44479 is the most actionable representative of the clearest same-bug subgroup in this cluster, with explicit regression scope and a direct narrower counterpart.", "best_pr_reason": null, "soft_edge_verdicts": [ + { + "left": "issue:42175", + "right": "issue:43824", + "accept": false, + "reason": "Packaging/backend availability and a model import failure are different bugs and code paths." + }, + { + "left": "issue:33357", + "right": "issue:34567", + "accept": false, + "reason": "A macOS bus error in CLIP loading is unrelated to TrainerState token accounting." + }, + { + "left": "issue:43577", + "right": "issue:46512", + "accept": false, + "reason": "Both involve dtype, but one is a BLIP2 load-time model dtype bug and the other is stale config metadata; not the same bug." + }, + { + "left": "issue:43452", + "right": "issue:43582", + "accept": false, + "reason": "GGUF/from_pretrained breakage and AppleSilicon allocator warmup are unrelated failures." + }, + { + "left": "issue:43317", + "right": "issue:45538", + "accept": false, + "reason": "Device-map/offload loading and CLIPTokenizer max-length handling are different issues." + }, + { + "left": "issue:28282", + "right": "issue:29127", + "accept": false, + "reason": "Missing PyTorch import and a LayoutLMv3 error-message clarity issue are not duplicates." + }, + { + "left": "issue:43742", + "right": "issue:44291", + "accept": false, + "reason": "Different load-time failures with different triggers; no clear shared code path." + }, { "left": "issue:44062", "right": "issue:44991", "accept": false, - "reason": "Different tokenizer failures: AddedToken API misuse vs tokenizer load failure for est-roberta." + "reason": "Both are tokenizer bugs, but the specific failure modes and affected code paths differ." }, { "left": "issue:41720", "right": "issue:43540", "accept": false, - "reason": "Both mention Qwen3, but one is a device-mapping CUDA assert and the other is a video-input ValueError." + "reason": "Qwen3 auto device-mapping CUDA assert and Qwen3OmniMoe video input validation are distinct problems." + }, + { + "left": "issue:44849", + "right": "issue:45701", + "accept": false, + "reason": "A Qwen3.5 hidden-states bug is not the same as a generic tokenization regression report." }, { "left": "issue:42915", "right": "issue:43749", "accept": false, - "reason": "Different bugs: FineGrainedFP8Config failure vs FSDP CPU RAM efficient loading." + "reason": "FP8 config failure and FSDP CPU RAM efficient loading are separate distributed/runtime issues." }, { "left": "issue:43540", "right": "issue:44560", "accept": false, - "reason": "Different multimodal paths: Qwen3OmniMoe video processing vs Qwen3-vl-embedding video StopIteration." + "reason": "Both are video-related, but they affect different Qwen models with different errors and no clear shared bug." }, { "left": "issue:44479", "right": "issue:45381", "accept": true, - "reason": "Same Qwen2.5-VL v5.3.0 video-input regression; 45381 is a narrower symptom of the same wrong position-id logic." + "reason": "45381 is a narrower report of the same Qwen2.5-VL video-input/vision_position_ids regression covered by 44479." }, { - "left": "issue:24643", - "right": "issue:29159", + "left": "issue:43540", + "right": "issue:44479", "accept": false, - "reason": "Unrelated: DeepSpeed training weight-shape error vs slow/fast tokenizer inconsistency." + "reason": "Different Qwen video bugs in different model families; too little evidence they are the same defect." }, { - "left": "issue:43540", - "right": "issue:44479", + "left": "issue:24643", + "right": "issue:29159", "accept": false, - "reason": "Both are Qwen video bugs, but they affect different models/code paths and are not the same regression." + "reason": "DeepSpeed weight-shape training failure and tokenizer inconsistency are unrelated." }, { "left": "issue:27957", "right": "issue:28282", "accept": false, - "reason": "XLMRoberta Flash Attention 2 issue vs missing PyTorch ImportError; unrelated." + "reason": "Flash Attention 2 on XLM-Roberta and missing PyTorch import are different issues." }, { "left": "issue:42175", "right": "issue:44488", "accept": false, - "reason": "Packaging/backend dependency issue vs model-loading failure for a specific checkpoint." + "reason": "Backend packaging and a model-load failure are not the same underlying bug." }, { "left": "issue:38175", "right": "issue:44898", "accept": false, - "reason": "Different vision bugs: SigLIP2 zero probabilities vs Perceiver interpolation failure." + "reason": "SigLIP2 zero-probability output and Perceiver non-default resolution failure involve different models and code paths." }, { "left": "issue:44568", "right": "issue:45488", "accept": false, - "reason": "Tokenizer regressions, but one is add_special_tokens/BOS-EOS behavior and the other is ByteLevel\u2192Metaspace override." + "reason": "Tokenizer special-token insertion and LlamaTokenizer pre-tokenizer override are distinct tokenizer regressions." }, { "left": "issue:44987", "right": "issue:45907", "accept": false, - "reason": "Model load regression vs utility function missing the revision argument." + "reason": "A model load failure and a Hub API revision-forwarding bug are unrelated." }, { "left": "issue:43493", "right": "issue:43994", - "accept": false, - "reason": "Both touch SigLIP2, but one is HF-vs-JAX discrepancy and the other is bad AutoModel/pipeline output." + "accept": true, + "reason": "Both report the same SigLIP2 implementation/output regression: one as a discrepancy vs JAX, the other as nonsensical outputs in AutoModel/pipeline." }, { "left": "issue:44291", "right": "issue:44821", "accept": false, - "reason": "init_empty_weights argument mismatch vs loading AutoImageProcessor from URL." + "reason": "`init_empty_weights` argument handling and loading an image processor from URL are different failures." }, { - "left": "issue:34567", - "right": "issue:36331", + "left": "issue:43257", + "right": "issue:43756", "accept": false, - "reason": "Trainer-related, but one is token counting and the other is compute_loss signature breakage." + "reason": "Qwen3 MoE deepspeed weight conversion and Smollm3 RoPE-layer dropping are unrelated model-specific bugs." }, { - "left": "issue:43257", - "right": "issue:43756", + "left": "issue:34567", + "right": "issue:36331", "accept": false, - "reason": "Qwen3 MoE FP8 loading vs Smollm3 RoPE-layer configuration; different bugs." + "reason": "TrainerState token counting and a CustomTrainer signature break are different trainer issues." }, { "left": "issue:44737", "right": "issue:46463", "accept": false, - "reason": "XLNet CPU device placement bug vs T5 decoder input validation error message." + "reason": "XLNet CPU device placement and T5 decoder-input validation messaging are unrelated." }, { "left": "issue:44792", "right": "issue:44964", "accept": false, - "reason": "Janus image-generation test failure vs Phi-4 multimodal model load failure." + "reason": "Both touch multimodal models, but one is a Janus test failure and the other is a Phi-4 multimodal load failure; not the same bug." }, { "left": "issue:43577", "right": "issue:45072", "accept": false, - "reason": "Different dtype bugs in different model families: BLIP2 vs SwitchTransformers/TimmWrapper." + "reason": "These are separate dtype mismatch reports for different models/code paths." }, { "left": "issue:44062", "right": "issue:45081", "accept": false, - "reason": "Tokenizer API error vs Mistral regex patch crashing on missing backend_tokenizer." + "reason": "Different tokenizer failures: AddedToken construction vs Mistral regex patching." }, { "left": "issue:43742", "right": "issue:44589", "accept": false, - "reason": "Different load-time failures: MobileLLM key error vs missing Float8 storage type." - }, - { - "left": "issue:43756", - "right": "issue:46399", - "accept": false, - "reason": "Smollm3 RoPE-layer mismatch vs Qwen3.5 kernel registration/decorator crash." - }, - { - "left": "issue:45042", - "right": "issue:45081", - "accept": false, - "reason": "Image-processor torchvision dependency issue vs tokenizer regex crash." - }, - { - "left": "issue:44568", - "right": "issue:46491", - "accept": false, - "reason": "Different tokenizer regressions: missing BOS/EOS insertion vs leading-space round-trip loss." - }, - { - "left": "issue:43756", - "right": "issue:45325", - "accept": false, - "reason": "Different Qwen/Smollm bugs; one is model architecture config, the other is Qwen2.5-VL rope scaling." - }, - { - "left": "issue:43906", - "right": "issue:46468", - "accept": false, - "reason": "A reproduction of a prior issue vs a broken Nemotron example; no evidence they are the same bug." - }, - { - "left": "issue:44737", - "right": "issue:45588", - "accept": false, - "reason": "XLNet positional encoding CPU bug vs flash-attention auxiliary state AttributeError." - }, - { - "left": "issue:45823", - "right": "issue:46463", - "accept": false, - "reason": "Gemma4 device-map mismatch vs T5 decoder validation issue; unrelated." - }, - { - "left": "issue:43531", - "right": "issue:44821", - "accept": false, - "reason": "Qwen3-MoE sliding_window issue vs AutoImageProcessor URL-loading failure." + "reason": "A MobileLLM key error and a Float8 storage lookup error are unrelated." } ] }, "evaluator_result": { "accept": true, - "feedback": "The summary is grounded in the packet and stays conservative. The canonical choice of issue:44479 over issue:45381 is well supported by the titles: 44479 is the broader v5.3.0 Qwen video regression report, while 45381 is a narrower Qwen2.5-VL symptom. The soft-edge verdicts are also cautious and do not overclaim duplicate relationships." + "feedback": "The summary is grounded in the packet and stays conservative. The two accepted duplicate-like pairs are supported by the issue titles and the reported scopes, and the rejected soft edges are phrased cautiously without overclaiming. I don\u2019t see evidence of merge-incompatible pairings or unsupported certainty." }, "evaluator_used": true, "retried": false, @@ -6921,279 +6832,264 @@ "source": "llm", "cache_hit": false, "model": "gpt-5.4-mini?service_tier=flex", - "cluster_id": "cluster-10000-472", + "cluster_id": "cluster-10000-473", "nodes": [ "issue:24643", "issue:28282", "issue:29127", "issue:32090", "issue:35141", - "issue:38175", "issue:41762", - "issue:43295", "issue:43404", "issue:43452", "issue:43526", + "issue:43531", "issue:43643", - "issue:43653", - "issue:43688", "issue:43697", "issue:43720", - "issue:43749", - "issue:43819", + "issue:43756", "issue:43906", - "issue:44361", - "issue:44442", "issue:44466", "issue:44488", "issue:44493", + "issue:44568", "issue:44625", - "issue:44811", - "issue:44843", + "issue:44737", + "issue:44821", "issue:44877", "issue:44898", "issue:44913", "issue:44936", "issue:45005", + "issue:45042", + "issue:45081", "issue:45200", "issue:45216", + "issue:45325", "issue:45357", - "issue:45372", "issue:45375", "issue:45381", "issue:45405", "issue:45479", + "issue:45588", "issue:45600", "issue:45823", "issue:45920", "issue:45987", "issue:46050", "issue:46121", - "issue:46170", - "issue:46463" + "issue:46399", + "issue:46463", + "issue:46468", + "issue:46491" ], "soft_pairs": [ + "issue:43756|issue:46399", + "issue:44568|issue:46491", + "issue:45042|issue:45081", + "issue:43906|issue:46468", + "issue:43756|issue:45325", + "issue:44737|issue:45588", + "issue:45823|issue:46463", "issue:45405|issue:45600", + "issue:43531|issue:44821", "issue:43404|issue:45987", "issue:43526|issue:45479", "issue:43452|issue:44488", "issue:45357|issue:45920", - "issue:44625|issue:44877", "issue:43906|issue:45381", "issue:44913|issue:46121", + "issue:44625|issue:44877", "issue:43720|issue:44898", "issue:43697|issue:45216", - "issue:43643|issue:45375", "issue:24643|issue:29127", + "issue:43643|issue:45375", "issue:28282|issue:32090", "issue:45200|issue:46463", "issue:28282|issue:35141", "issue:41762|issue:44936", "issue:44466|issue:45005", - "issue:44493|issue:46050", - "issue:43452|issue:44843", - "issue:43688|issue:43749", - "issue:43819|issue:44811", - "issue:44361|issue:44442", - "issue:43653|issue:44361", - "issue:38175|issue:43295", - "issue:45372|issue:46170", - "issue:44936|issue:45823" + "issue:44493|issue:46050" ], - "prepared_review_unit_hash": "9347ae4c8cdb5cec449efd5281b0f42f62294897b279df09e66a0d9b61cbb659", + "prepared_review_unit_hash": "72bbae037fa57efdc430c9d54a454d2e265f72ac00781caa3ee803daafe35a06", "component_budget": { - "node_count": 472, - "item_count": 472, - "soft_pair_count": 1732, - "serialized_chars": 898654, - "estimated_input_tokens": 224664, - "estimated_eval_tokens": 449584 + "node_count": 473, + "item_count": 473, + "soft_pair_count": 1738, + "serialized_chars": 901437, + "estimated_input_tokens": 225360, + "estimated_eval_tokens": 450976 }, "budget": { "node_count": 48, "item_count": 48, "soft_pair_count": 26, - "serialized_chars": 36018, - "estimated_input_tokens": 9005, - "estimated_eval_tokens": 18266 + "serialized_chars": 36127, + "estimated_input_tokens": 9032, + "estimated_eval_tokens": 18320 }, "overflow_policy": "truncate_then_skip", "trimmed": false, "aggressively_trimmed": false, "split": true, "analyst_result": { - "summary": "This cluster is mostly heterogeneous; most issues are unrelated. The only strong duplicate-like pairs are tied-weight/save_pretrained regressions in v5, plus a Qwen3.5 save_pretrained serialization follow-up.", - "confidence": 0.61, - "canonical_issue_reason": "issue:45005 is the broadest and most central report in the strongest duplicate pair: it describes the underlying tied-weights regression across translation models, while issue:44466 is a more specific symptom of the same serialization problem.", + "summary": "This cluster is a loose set of Transformer bug reports across unrelated subsystems (tokenizers, configs, training, vision, DeepSpeed, tied weights). None of the proposed soft pairs look like the same underlying bug/change.", + "confidence": 0.78, + "canonical_issue_reason": "issue:45920 is the broadest, most representative regression report here: it clearly describes a v5 AutoTokenizer break affecting multiple model families and is concrete enough to anchor the cluster.", "canonical_pr_reason": null, - "best_issue_reason": "issue:45005 is the best anchor for the cluster because it captures the clearest shared code-path bug with the widest scope; most other issues in the cluster are off-topic to that regression.", + "best_issue_reason": "issue:45920 has the best combination of breadth, clear repro impact, and cross-model relevance, making it the strongest issue anchor among this mixed set.", "best_pr_reason": null, "soft_edge_verdicts": [ { - "left": "issue:45405", - "right": "issue:45600", - "accept": false, - "reason": "Different problems: PEFT version availability vs auto_mappings referencing removed configs." - }, - { - "left": "issue:43404", - "right": "issue:45987", + "left": "issue:43756", + "right": "issue:46399", "accept": false, - "reason": "Unrelated: tied lm_head weights in Mistral vs StaticCache return type bug." + "reason": "Both mention RoPE/rotation-adjacent behavior, but one is about Smollm3 dropping layers and the other is a Qwen3.5 kernel-registration crash. Different models and different failure modes." }, { - "left": "issue:43526", - "right": "issue:45479", + "left": "issue:44568", + "right": "issue:46491", "accept": false, - "reason": "Different symptom and code path: BEiT image processor label reduction vs sequence-classification loss degeneracy." + "reason": "Both are tokenizer regressions, but mdeberta-v3 missing add_special_tokens BOS/EOS is unrelated to CodeLlama stripping a leading space on round-trip." }, { - "left": "issue:43452", - "right": "issue:44488", + "left": "issue:45042", + "right": "issue:45081", "accept": false, - "reason": "Both are loading failures, but the failing mechanisms and models differ." + "reason": "One is PIL image processors incorrectly requiring torchvision; the other is a Mistral tokenizer regex patch crash. Different code paths entirely." }, { - "left": "issue:45357", - "right": "issue:45920", + "left": "issue:43906", + "right": "issue:46468", "accept": false, - "reason": "Qwen3.5 save_pretrained serialization issue vs a broad AutoTokenizer wrong-token-ID regression." + "reason": "No evidence these are the same bug: one is a reproduction of an earlier issue, the other is a broken Nemotron example. Too generic to merge." }, { - "left": "issue:44625", - "right": "issue:44877", + "left": "issue:43756", + "right": "issue:45325", "accept": false, - "reason": "Both are config-related, but one is num_labels propagation and the other is strict loading of granite_speech config." + "reason": "Both involve RoPE/position-related logic, but they concern different models and distinct scaling/coverage bugs. Not the same underlying defect." }, { - "left": "issue:43906", - "right": "issue:45381", + "left": "issue:44737", + "right": "issue:45588", "accept": false, - "reason": "Too weakly connected; one is a reproduction of an older issue, the other is a specific qwen2.5-vl video position-id bug." + "reason": "XLNet CPU arange in relative position encoding is unrelated to flash_attention.py crashing on s_aux=None." }, { - "left": "issue:44913", - "right": "issue:46121", + "left": "issue:45823", + "right": "issue:46463", "accept": false, - "reason": "Related to rope parameters, but one is config reload defaulting and the other is a TypeError on JSON-loaded validation keys." + "reason": "Gemma4 PLE device mismatch during forward is unrelated to T5 decoder-input validation error messaging." }, { - "left": "issue:43720", - "right": "issue:44898", + "left": "issue:45405", + "right": "issue:45600", "accept": false, - "reason": "Completely different models and failures: BitNet loading vs Perceiver interpolation." + "reason": "A dependency-version bump and a stale auto_mappings reference are unrelated CI/config maintenance issues." }, { - "left": "issue:43697", - "right": "issue:45216", + "left": "issue:43531", + "right": "issue:44821", "accept": false, - "reason": "Both are regressions, but they affect different model families and different bugs." + "reason": "Qwen3-MoE sliding_window behavior and AutoImageProcessor URL loading are different subsystems with no shared concrete bug." }, { - "left": "issue:43643", - "right": "issue:45375", + "left": "issue:43404", + "right": "issue:45987", "accept": false, - "reason": "Both involve config fields, but one is generic remote-code AutoConfig loading and the other is a specific Qwen3_5MoeVisionConfig field drop." + "reason": "Tied lm_head weight serialization and StaticCache.get_seq_length returning a shape-(1,) tensor are unrelated issues." }, { - "left": "issue:24643", - "right": "issue:29127", + "left": "issue:43526", + "right": "issue:45479", "accept": false, - "reason": "Unrelated: DeepSpeed training RuntimeError vs LayoutLMv3 error-message clarity." + "reason": "BEiTImageProcessorFast reduce_labels returning one label is unrelated to the zero-loss classification bug with num_labels=1." }, { - "left": "issue:28282", - "right": "issue:32090", + "left": "issue:43452", + "right": "issue:44488", "accept": false, - "reason": "Different failures: missing PyTorch import vs NoneType broadcast in Trainer." + "reason": "Both involve model/tokenizer loading failures, but one is about gguf_file handling and the other is a specific model load regression. Not the same bug." }, { - "left": "issue:45200", - "right": "issue:46463", + "left": "issue:45357", + "right": "issue:45920", "accept": false, - "reason": "Different model families and different validation bugs." + "reason": "Both touch Qwen3.5, but one is about save_pretrained saving visual encoder keys and the other is AutoTokenizer producing wrong token IDs across many models." }, { - "left": "issue:28282", - "right": "issue:35141", + "left": "issue:43906", + "right": "issue:45381", "accept": false, - "reason": "No shared code-path: missing PyTorch import vs output embedding reinit after resizing." + "reason": "A reproduction of an earlier issue and a Qwen2.5-VL video position-id bug are different reports with no concrete overlap." }, { - "left": "issue:41762", - "right": "issue:44936", + "left": "issue:44913", + "right": "issue:46121", "accept": false, - "reason": "Both touch training/loading workflows, but the Gemma3 DeepSpeed ZeRO-3 failure and trainer.evaluate() after train are distinct." - }, - { - "left": "issue:44466", - "right": "issue:45005", - "accept": true, - "reason": "Same underlying tied-weights/lm_head serialization bug in v5; 45005 is the broader report and 44466 is a specific manifestation." + "reason": "GPTNeoX rotary_pct reloading default and convert_rope_params_to_dict failing on JSON-loaded lists are both rope-related, but they are distinct config-handling bugs." }, { - "left": "issue:44493", - "right": "issue:46050", + "left": "issue:44625", + "right": "issue:44877", "accept": false, - "reason": "Different issues: unexpected position-id keys vs quantized-model save/load breakage." + "reason": "Qwen3.5 num_labels propagation and strict config blocking granite_speech loading are unrelated config issues." }, { - "left": "issue:43452", - "right": "issue:44843", + "left": "issue:43720", + "right": "issue:44898", "accept": false, - "reason": "Both concern from_pretrained loading, but one is gguf_file support and the other is an offline-model_info call in mistral regex patching." + "reason": "BitNet packed-weight unpacking during accelerate loading is unrelated to Perceiver image classification failing with interpolate_pos_encoding=True." }, { - "left": "issue:43688", - "right": "issue:43749", + "left": "issue:43697", + "right": "issue:45216", "accept": false, - "reason": "Different subsystems: MoE auxiliary-loss normalization vs FSDP CPU RAM-efficient loading." + "reason": "RTDetrV2 output differences and Qwen3.5 save_pretrained corruption are separate regressions in different components." }, { - "left": "issue:43819", - "right": "issue:44811", + "left": "issue:24643", + "right": "issue:29127", "accept": false, - "reason": "Different inference paths and bugs: DAC latent decoding vs Whisper batch_decode skip_special_tokens handling." + "reason": "DeepSpeed training weight shape error and layoutlmv3 box-information error are entirely unrelated." }, { - "left": "issue:44361", - "right": "issue:44442", + "left": "issue:43643", + "right": "issue:45375", "accept": false, - "reason": "Both are tokenizer-related errors, but they affect different tokenizers and failure modes." + "reason": "Missing fields in AutoConfig with trust_remote_code and a specific Qwen3_5MoeVisionConfig field being dropped are not the same bug." }, { - "left": "issue:43653", - "right": "issue:44361", + "left": "issue:28282", + "right": "issue:32090", "accept": false, - "reason": "Both are tokenizer bugs, but one is BigBird special-token registration and the other is MLuke task AttributeError." + "reason": "PyTorch-not-found ImportError and Trainer _gpu_broadcast_one(NoneType) are unrelated environment/runtime failures." }, { - "left": "issue:38175", - "right": "issue:43295", + "left": "issue:45200", + "right": "issue:46463", "accept": false, - "reason": "Different regressions in different components: SigLIP2 output probabilities vs processor.tokenizer/image passing changes." + "reason": "Gemma4 mm_token_type_ids defaulting to zeros and a T5 decoder-input validation error message are different issues." }, { - "left": "issue:45372", - "right": "issue:46170", + "left": "issue:44466", + "right": "issue:45005", "accept": false, - "reason": "Different import/dependency failures in different environments and libraries." + "reason": "Both mention tied weights, but one is inconsistent serialization of lm_head.weight and the other is a broader translation-model tied-weights issue. Not enough to treat as one bug." }, { - "left": "issue:44936", - "right": "issue:45823", + "left": "issue:44493", + "right": "issue:46050", "accept": false, - "reason": "Trainer lifecycle bug vs Gemma4 device-mapping forward mismatch; no shared root cause." + "reason": "Unexpected position-id key warnings and quantized-model save/load breakage are unrelated." } ] }, - "evaluator_result": { - "accept": true, - "feedback": "The cluster summary is grounded in the packet and stays conservative. The only clearly supported duplicate-like pair is issue:44466 with issue:45005, and the other verdicts appropriately reject weak or unrelated matches without overstating shared root causes." - }, - "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 }, @@ -7208,14 +7104,11 @@ "source": "llm", "cache_hit": false, "model": "gpt-5.4-mini?service_tier=flex", - "cluster_id": "cluster-10000-472", + "cluster_id": "cluster-10000-473", "nodes": [ "issue:29159", - "issue:32090", "issue:33357", - "issue:33666", - "issue:36010", - "issue:41720", + "issue:38175", "issue:42898", "issue:43295", "issue:43452", @@ -7229,10 +7122,11 @@ "issue:43742", "issue:43749", "issue:43756", + "issue:43819", "issue:43927", "issue:44361", + "issue:44442", "issue:44464", - "issue:44466", "issue:44479", "issue:44484", "issue:44514", @@ -7241,250 +7135,266 @@ "issue:44704", "issue:44734", "issue:44743", - "issue:44821", + "issue:44811", "issue:44843", - "issue:44977", + "issue:44936", "issue:44991", "issue:45356", + "issue:45372", "issue:45381", "issue:45440", "issue:45491", "issue:45542", "issue:45701", "issue:45704", + "issue:45823", "issue:46002", - "issue:46050", + "issue:46170", "issue:46178", "issue:46402", "issue:46491", "issue:46612" ], "soft_pairs": [ + "issue:43819|issue:44811", + "issue:44361|issue:44442", + "issue:43452|issue:44843", + "issue:43688|issue:43749", + "issue:38175|issue:43295", + "issue:43653|issue:44361", + "issue:45372|issue:46170", + "issue:44936|issue:45823", "issue:45542|issue:46178", "issue:42898|issue:46491", "issue:43452|issue:43577", "issue:43531|issue:43742", - "issue:43295|issue:45356", "issue:44514|issue:45381", + "issue:43295|issue:45356", "issue:43577|issue:43697", "issue:43653|issue:44568", "issue:43646|issue:43749", + "issue:44464|issue:44743", "issue:44843|issue:44991", - "issue:44484|issue:44704", "issue:29159|issue:33357", - "issue:44464|issue:44743", "issue:43688|issue:45704", + "issue:44484|issue:44704", "issue:43653|issue:43927", "issue:43756|issue:46402", - "issue:43688|issue:45701", "issue:43688|issue:45440", - "issue:44479|issue:44625", + "issue:43688|issue:45701", "issue:43550|issue:44361", + "issue:44479|issue:44625", "issue:45491|issue:46002", - "issue:44734|issue:46612", - "issue:33666|issue:36010", - "issue:44464|issue:44977", - "issue:32090|issue:33666", - "issue:41720|issue:44464", - "issue:43742|issue:44821", - "issue:44466|issue:46050" + "issue:44734|issue:46612" ], - "prepared_review_unit_hash": "9b540290e8f67d62d0ee133c71e981c54fb200b7ffb0e678c0910f0edc590b3e", + "prepared_review_unit_hash": "509cac7caf2dea4000d46198f8a3ffe2cba09800b896ef819a9a22c7a3198aaa", "component_budget": { - "node_count": 472, - "item_count": 472, - "soft_pair_count": 1732, - "serialized_chars": 898654, - "estimated_input_tokens": 224664, - "estimated_eval_tokens": 449584 + "node_count": 473, + "item_count": 473, + "soft_pair_count": 1738, + "serialized_chars": 901437, + "estimated_input_tokens": 225360, + "estimated_eval_tokens": 450976 }, "budget": { "node_count": 48, "item_count": 48, - "soft_pair_count": 28, - "serialized_chars": 36198, - "estimated_input_tokens": 9050, - "estimated_eval_tokens": 18356 + "soft_pair_count": 30, + "serialized_chars": 36849, + "estimated_input_tokens": 9213, + "estimated_eval_tokens": 18682 }, "overflow_policy": "truncate_then_skip", "trimmed": false, "aggressively_trimmed": false, "split": true, "analyst_result": { - "summary": "The cluster is highly heterogeneous: tokenizer regressions, model-loading failures, generation/cache bugs, backend/runtime issues, and a few vision-language regressions. None of the soft pairs looks like the same underlying bug, so they should stay separate.", - "confidence": 0.34, - "canonical_issue_reason": "issue:45701 is the broadest tokenizer-regression report in the set, so it is the closest thing to a cluster anchor, though the overall set is too diverse for a true duplicate canonical.", + "summary": "This set is mostly a heterogeneous collection of unrelated bug reports; no soft pair looks like the same underlying issue, so I reject all candidate merges. The broadest tokenizer-regression report is the best available issue anchor, but the cluster itself is not cohesive.", + "confidence": 0.87, + "canonical_issue_reason": "issue:45701 is the broadest umbrella-style tokenizer regression report and the most plausible anchor among a mixed set.", "canonical_pr_reason": null, - "best_issue_reason": "issue:45701 has the widest scope among the tokenizer-related items and is the most plausible representative if one issue must be chosen.", + "best_issue_reason": "issue:45701 has the broadest scope and strongest tokenizer-regression framing, making it the best representative issue in this heterogeneous cluster.", "best_pr_reason": null, "soft_edge_verdicts": [ + { + "left": "issue:43819", + "right": "issue:44811", + "accept": false, + "reason": "Different models and APIs: DAC latent decode vs Whisper batch decode special-token handling." + }, + { + "left": "issue:44361", + "right": "issue:44442", + "accept": false, + "reason": "Both are tokenizer loading failures, but for different tokenizers and different error modes." + }, + { + "left": "issue:43452", + "right": "issue:44843", + "accept": false, + "reason": "Both touch tokenizer loading, but one is gguf loading and the other is offline hub regex patching." + }, + { + "left": "issue:43688", + "right": "issue:43749", + "accept": false, + "reason": "Unrelated subsystems: MoE auxiliary-loss normalization vs FSDP RAM-efficient loading." + }, + { + "left": "issue:38175", + "right": "issue:43295", + "accept": false, + "reason": "Different regression classes: zero probabilities in SigLIP2 vs processor/tokenizer regression in custom model code." + }, + { + "left": "issue:43653", + "right": "issue:44361", + "accept": false, + "reason": "Both are tokenizer bugs, but one is BigBird special-token registration and the other is MLuke task AttributeError." + }, + { + "left": "issue:45372", + "right": "issue:46170", + "accept": false, + "reason": "Completely different failures: Gemma processor import issue vs Windows/ROCm torch.distributed dependency." + }, + { + "left": "issue:44936", + "right": "issue:45823", + "accept": false, + "reason": "Training/eval state bug vs Gemma4 device-map mismatch; not the same code path." + }, { "left": "issue:45542", "right": "issue:46178", "accept": false, - "reason": "Both are backend-related errors, but one is TensorBoard/TensorFlow availability while the other is a tensorflow_text backend mapping failure; different code paths and fixes." + "reason": "Both concern backend availability, but they target different backend-detection errors and different packages." }, { "left": "issue:42898", "right": "issue:46491", "accept": false, - "reason": "Both concern tokenization output, but 42898 is about the v5 clean_up_tokenization_spaces behavior change, while 46491 is a CodeLlama leading-space round-trip regression." + "reason": "Both are tokenization regressions, but one is a v5 cleanup behavior change and the other is a CodeLlama spacing round-trip bug." }, { "left": "issue:43452", "right": "issue:43577", "accept": false, - "reason": "AutoTokenizer/gguf loading failures and Blip2 dtype propagation are unrelated bugs." + "reason": "GGUF loading failure and Blip2 dtype propagation are unrelated issues." }, { "left": "issue:43531", "right": "issue:43742", "accept": false, - "reason": "Qwen3-MoE sliding_window behavior and MobileLLM key errors are different model-specific loading/training issues." + "reason": "Different problems: Qwen3-MoE sliding_window behavior vs MobileLLM config loading key error." }, { - "left": "issue:43295", - "right": "issue:45356", + "left": "issue:44514", + "right": "issue:45381", "accept": false, - "reason": "One is a processor.tokenizer/image-passing regression; the other is Kimi-K2.5 tokenizer codec handling and warning behavior." + "reason": "Same model family, but batching/padding crash and video vision_position_ids bug are distinct." }, { - "left": "issue:44514", - "right": "issue:45381", + "left": "issue:43295", + "right": "issue:45356", "accept": false, - "reason": "Both involve Qwen2.5-VL, but one is batched chat-template padding and the other is video vision_position_ids; not the same bug." + "reason": "Both are regressions, but one is custom processor/tokenizer access and the other is Kimi-K2.5 codec handling." }, { "left": "issue:43577", "right": "issue:43697", "accept": false, - "reason": "Blip2 dtype loading and RTDetrV2 output divergence are unrelated model behaviors." + "reason": "Different defects: dtype loading for Blip2 vs divergent RTDetrV2 outputs in v5." }, { "left": "issue:43653", "right": "issue:44568", "accept": false, - "reason": "Both are tokenizer special-token issues, but one is BigBird mask-token registration and the other is BOS/EOS insertion with add_special_tokens." + "reason": "Both involve special tokens, but the tokenizer classes and failure modes differ." }, { "left": "issue:43646", "right": "issue:43749", "accept": false, - "reason": "Custom model initialization breakage and FSDP CPU RAM-efficient loading are different failure modes in different parts of the stack." + "reason": "Custom model initialization regression is unrelated to FSDP CPU RAM-efficient loading." }, { - "left": "issue:44843", - "right": "issue:44991", + "left": "issue:44464", + "right": "issue:44743", "accept": false, - "reason": "Offline model_info access in _patch_mistral_regex and tokenizer loading failure for est-roberta may both hit tokenizers, but they are not the same root cause." + "reason": "Both involve generation/cache, but one is compiled chunked generation and the other is recurrent state reset." }, { - "left": "issue:44484", - "right": "issue:44704", + "left": "issue:44843", + "right": "issue:44991", "accept": false, - "reason": "save_pretrained shard-size policy and AutoProcessor cached_file kwargs forwarding are unrelated API issues." + "reason": "Different tokenizer-loading regressions with different triggers and models." }, { "left": "issue:29159", "right": "issue:33357", "accept": false, - "reason": "Tokenizer fast/slow inconsistency and a MacOS bus error on a CLIP model are not the same bug." + "reason": "Slow/fast tokenizer inconsistency is unrelated to the Mac bus error in CLIP loading." }, { - "left": "issue:44464", - "right": "issue:44743", + "left": "issue:43688", + "right": "issue:45704", "accept": false, - "reason": "Compiled forward chunked generation mismatch and recurrent-state reset with cache are different generation-path defects." + "reason": "Aux-loss normalization and apex RMSNorm memory leak are unrelated." }, { - "left": "issue:43688", - "right": "issue:45704", + "left": "issue:44484", + "right": "issue:44704", "accept": false, - "reason": "Auxiliary-loss normalization and a T5 RMSNorm memory leak are unrelated model implementation bugs." + "reason": "save_pretrained sharding policy question vs AutoProcessor kwargs propagation bug." }, { "left": "issue:43653", "right": "issue:43927", "accept": false, - "reason": "Special-token registration in BigBirdTokenizer and DiaConfig losing custom token IDs after save/load are different tokenizer/config issues." + "reason": "Special-token registration bug vs config save/load losing custom token IDs." }, { "left": "issue:43756", "right": "issue:46402", "accept": false, - "reason": "Smollm3 RoPE-layer dropping and Sam3 text encoder weights not loading are separate model-definition problems." - }, - { - "left": "issue:43688", - "right": "issue:45701", - "accept": false, - "reason": "Auxiliary-loss normalization is a modeling/math issue, while 45701 is a tokenizer-version regression." + "reason": "RoPE-layer truncation in Smollm3 is unrelated to Sam3 text encoder weights not loading." }, { "left": "issue:43688", "right": "issue:45440", "accept": false, - "reason": "OLMoE/GPT Oss loss normalization and DeepseekV3MoE implementation divergence are different model bugs." + "reason": "MoE loss normalization and DeepSeekV3 implementation divergence are different code-path problems." }, { - "left": "issue:44479", - "right": "issue:44625", + "left": "issue:43688", + "right": "issue:45701", "accept": false, - "reason": "Video-input regression in Qwen VL models and num_labels propagation in Qwen3.5 are distinct config/input-path issues." + "reason": "Auxiliary-loss normalization is unrelated to the broad tokenization regression report." }, { "left": "issue:43550", "right": "issue:44361", "accept": false, - "reason": "torch.compile/SDPA failure in Bamba and an MLukeTokenizer AttributeError are unrelated." + "reason": "torch.compile/SDPA model failure is unrelated to MLuke tokenizer task AttributeError." }, - { - "left": "issue:45491", - "right": "issue:46002", - "accept": false, - "reason": "Gemma3 NaNs from sliding-window attention and NaNs in classification heads upon checkout/init are different NaN sources." - }, - { - "left": "issue:44734", - "right": "issue:46612", - "accept": false, - "reason": "KV-cache continuation indexing in serve and beam-search cache reorder skipping in generation are different cache bugs." - }, - { - "left": "issue:33666", - "right": "issue:36010", - "accept": false, - "reason": "Qwen2-VL multi-GPU training and a GenerationMixin import error are unrelated." - }, - { - "left": "issue:44464", - "right": "issue:44977", - "accept": false, - "reason": "Chunked-generation inconsistency with compiled forward and Qwen3.5 flash-attention generation failure are different generation regressions." - }, - { - "left": "issue:32090", - "right": "issue:33666", - "accept": false, - "reason": "Trainer GPU broadcast NoneType and Qwen2-VL multi-GPU training are unrelated distributed-training issues." - }, - { - "left": "issue:41720", - "right": "issue:44464", + { + "left": "issue:44479", + "right": "issue:44625", "accept": false, - "reason": "Auto device-mapping cudaErrorAssert and compiled-forward chunked generation are different failure modes." + "reason": "Video-input regression for Qwen models is distinct from num_labels propagation in Qwen3.5 config." }, { - "left": "issue:43742", - "right": "issue:44821", + "left": "issue:45491", + "right": "issue:46002", "accept": false, - "reason": "MobileLLM key error on load and AutoImageProcessor URL loading are separate loading-path bugs." + "reason": "Different NaN sources: sliding-window attention vs classification-head init/checkpointing." }, { - "left": "issue:44466", - "right": "issue:46050", + "left": "issue:44734", + "right": "issue:46612", "accept": false, - "reason": "lm_head serialization differences by device and sinq quantized model save/load breakage are both persistence-related, but not the same defect." + "reason": "Both touch cache logic, but one is serve KV-continuation indexing and the other is beam-search cache reorder omission." } ] }, @@ -7506,13 +7416,16 @@ "source": "llm", "cache_hit": false, "model": "gpt-5.4-mini?service_tier=flex", - "cluster_id": "cluster-10000-472", + "cluster_id": "cluster-10000-473", "nodes": [ "issue:28218", + "issue:32090", "issue:33357", "issue:33666", + "issue:36010", "issue:36331", "issue:39692", + "issue:41720", "issue:43262", "issue:43278", "issue:43295", @@ -7525,48 +7438,51 @@ "issue:43577", "issue:43606", "issue:43643", - "issue:43646", "issue:43688", "issue:43720", + "issue:43742", "issue:43825", "issue:43844", "issue:43927", "issue:43994", "issue:44361", "issue:44451", + "issue:44464", "issue:44466", - "issue:44488", "issue:44514", "issue:44554", "issue:44561", - "issue:44568", "issue:44610", "issue:44625", - "issue:44843", + "issue:44821", "issue:44849", "issue:44964", "issue:44977", "issue:45290", - "issue:45357", "issue:45405", "issue:45584", "issue:45823", "issue:46050", "issue:46192", "issue:46209", - "issue:46365", - "issue:46489" + "issue:46365" ], "soft_pairs": [ - "issue:43504|issue:43550", + "issue:33666|issue:36010", + "issue:44464|issue:44977", + "issue:32090|issue:33666", + "issue:43742|issue:44821", + "issue:41720|issue:44464", + "issue:44466|issue:46050", "issue:43504|issue:43720", + "issue:43504|issue:43550", "issue:43643|issue:44561", "issue:43577|issue:44554", "issue:43504|issue:44610", "issue:44514|issue:45290", "issue:43720|issue:43927", - "issue:33666|issue:36331", "issue:43606|issue:43927", + "issue:33666|issue:36331", "issue:43262|issue:43295", "issue:44554|issue:46365", "issue:43278|issue:43688", @@ -7574,243 +7490,251 @@ "issue:43720|issue:44361", "issue:39692|issue:43493", "issue:44964|issue:45405", + "issue:43452|issue:44451", "issue:43525|issue:43572", "issue:43493|issue:43825", - "issue:43452|issue:44451", - "issue:45823|issue:46050", "issue:44625|issue:45584", + "issue:45823|issue:46050", "issue:43994|issue:46192", "issue:28218|issue:33357", - "issue:44849|issue:46209", "issue:39692|issue:43295", "issue:43278|issue:43844", "issue:43493|issue:43844", "issue:33357|issue:33666", - "issue:44488|issue:44843", - "issue:43646|issue:44466", - "issue:44849|issue:45357", - "issue:44568|issue:46489" + "issue:44849|issue:46209" ], - "prepared_review_unit_hash": "aa3e3ec1ee764ce32c4f9bea9602ba3e52e3d20a93b652e15eee8f898a31bdff", + "prepared_review_unit_hash": "ef71d6f2b42db92c42bbd9473241433b5e499c5c03cad997a162b38029f5028e", "component_budget": { - "node_count": 472, - "item_count": 472, - "soft_pair_count": 1732, - "serialized_chars": 898654, - "estimated_input_tokens": 224664, - "estimated_eval_tokens": 449584 + "node_count": 473, + "item_count": 473, + "soft_pair_count": 1738, + "serialized_chars": 901437, + "estimated_input_tokens": 225360, + "estimated_eval_tokens": 450976 }, "budget": { "node_count": 48, "item_count": 48, - "soft_pair_count": 32, - "serialized_chars": 37488, - "estimated_input_tokens": 9372, - "estimated_eval_tokens": 19000 + "soft_pair_count": 34, + "serialized_chars": 37961, + "estimated_input_tokens": 9491, + "estimated_eval_tokens": 19238 }, "overflow_policy": "truncate_then_skip", "trimmed": false, "aggressively_trimmed": false, "split": true, "analyst_result": { - "summary": "The cluster is mostly heterogeneous: tokenizer, processor, model-loading, training, and version-regression reports mixed together. The soft pairs are topical neighbors, but none look like the same underlying bug or change.", - "confidence": 0.93, - "canonical_issue_reason": "No single canonical issue stands out; the items are related only at a broad subsystem/theme level, not as true duplicates.", + "summary": "This cluster is highly heterogeneous: the soft pairs point to different bugs across tokenization, model loading, training, generation, configs, and device/runtime issues. I reject every soft edge as a duplicate false positive.", + "confidence": 0.97, + "canonical_issue_reason": "No true canonical duplicate exists here because the items describe unrelated failures. If a representative must be named, issue 43295 is the broadest user-facing regression, but it still does not subsume the rest.", "canonical_pr_reason": null, - "best_issue_reason": "issue:43646 is the broadest v5/custom-model initialization regression in the set and is the closest thing to an umbrella report, but it still does not cover the other concrete failures.", + "best_issue_reason": "Issue 43295 is the most suitable anchor only in the sense that it is a concrete, reproducible regression on a common API path; it is not a good dedupe hub for the other issues.", "best_pr_reason": null, "soft_edge_verdicts": [ { - "left": "issue:43504", - "right": "issue:43550", + "left": "issue:33666", + "right": "issue:36010", + "accept": false, + "reason": "Different problems: Qwen2-VL multi-GPU training vs a GenerationMixin import error." + }, + { + "left": "issue:44464", + "right": "issue:44977", + "accept": false, + "reason": "Different failure modes: chunked generation with compiled forward vs Qwen3.5 flash-attention generation." + }, + { + "left": "issue:32090", + "right": "issue:33666", + "accept": false, + "reason": "Trainer GPU broadcast TypeError is unrelated to Qwen2-VL multi-GPU training." + }, + { + "left": "issue:43742", + "right": "issue:44821", "accept": false, - "reason": "Both are load/runtime failures, but one is a legacy-field preset load bug and the other is a torch.compile + SDPA model failure." + "reason": "KeyError loading MobileLLM is unrelated to AutoImageProcessor-from-URL loading." + }, + { + "left": "issue:41720", + "right": "issue:44464", + "accept": false, + "reason": "CUDA assert under auto device mapping is not the same bug as compiled chunked generation inconsistency." + }, + { + "left": "issue:44466", + "right": "issue:46050", + "accept": false, + "reason": "lm_head serialization/tied-weights behavior differs from sinq save/load quantization breakage." }, { "left": "issue:43504", "right": "issue:43720", "accept": false, - "reason": "Different concrete problems: Beit preset loading vs BitNet packed-weight unpacking during accelerate loading." + "reason": "Legacy-field BEiT load bug is unrelated to BitNet packed-weight loading during accelerate." + }, + { + "left": "issue:43504", + "right": "issue:43550", + "accept": false, + "reason": "BEiT pretrained-load failure and Bamba torch.compile SDPA failure are distinct issues." }, { "left": "issue:43643", "right": "issue:44561", "accept": false, - "reason": "Both involve trust_remote_code-era breakage, but one is missing config fields and the other is an unrelated FX-helper removal regression." + "reason": "Missing fields from AutoConfig with trust_remote_code is not the same as removal of is_torch_fx_available breaking models." }, { "left": "issue:43577", "right": "issue:44554", "accept": false, - "reason": "One is a dtype propagation bug in Blip2 loading; the other is an MPS attention correctness issue. Different code paths and symptoms." + "reason": "Blip2 dtype propagation bug is unrelated to the MPS attention correctness issue." }, { "left": "issue:43504", "right": "issue:44610", "accept": false, - "reason": "Preset-load legacy-field bug vs OmDet-Turbo processor image-size mismatch; not the same defect." + "reason": "BEiT legacy-field loading failure is unrelated to OmDet-Turbo processor/model size mismatch." }, { "left": "issue:44514", "right": "issue:45290", "accept": false, - "reason": "Both are apply_chat_template crashes, but with different triggers: batched padding=False vs tool-call assistant messages with no content." + "reason": "Batched Qwen2.5-VL chat-template crash is a different code path from tool-call messages with no content." }, { "left": "issue:43720", "right": "issue:43927", "accept": false, - "reason": "Packed-weight loading failure vs config save/load losing custom token IDs. Same broad area, different bug." + "reason": "BitNet packed-weight loading and DiaConfig save/load token-ID loss are unrelated." }, { - "left": "issue:33666", - "right": "issue:36331", + "left": "issue:43606", + "right": "issue:43927", "accept": false, - "reason": "Multi-GPU training issue and a Trainer compute_loss API mismatch are unrelated." + "reason": "CPU offload device mismatch in bark-small is unrelated to DiaConfig custom token IDs being lost." }, { - "left": "issue:43606", - "right": "issue:43927", + "left": "issue:33666", + "right": "issue:36331", "accept": false, - "reason": "CPU offload device mismatch vs config serialization loss; not the same underlying bug." + "reason": "Qwen2-VL multi-GPU training and CustomTrainer.compute_loss signature mismatch are different issues." }, { "left": "issue:43262", "right": "issue:43295", "accept": false, - "reason": "Audio processor sampling-rate default bug vs processor.tokenizer / image-to-tokenizer regression. Different components and failures." + "reason": "Audio processor sampling-rate defaulting is unrelated to processor.tokenizer access and image-tokenization regression." }, { "left": "issue:44554", "right": "issue:46365", "accept": false, - "reason": "MPS attention correctness vs Trainer import breakage due to datasets. No shared failure mode." + "reason": "MPS attention correctness bug is unrelated to Trainer import failure caused by datasets." }, { "left": "issue:43278", "right": "issue:43688", "accept": false, - "reason": "Dtype drift in evaluate vs auxiliary-loss normalization error. Different model math issues." + "reason": "Evaluation dtype drift in embeddings is unrelated to auxiliary-loss normalization." }, { "left": "issue:44977", "right": "issue:46209", "accept": false, - "reason": "Qwen3.5 flash-attention generation bug vs FP8 inference failure on B200; distinct hardware/model-path problems." + "reason": "Qwen3.5 flash-attention generation bug is unrelated to FP8 inference on B200." }, { "left": "issue:43720", "right": "issue:44361", "accept": false, - "reason": "BitNet loading/unpacking bug vs MLukeTokenizer AttributeError. Same tokenizer/model-loading area only superficially." + "reason": "BitNet accelerate-loading bug is unrelated to MLukeTokenizer task-time AttributeError." }, { "left": "issue:39692", "right": "issue:43493", "accept": false, - "reason": "Both concern SigLIP2, but one is a broken documentation example and the other is an implementation mismatch with JAX." + "reason": "SigLIP2 docs-example errors and SigLIP2 HF-vs-JAX discrepancy are different underlying problems." }, { "left": "issue:44964", "right": "issue:45405", "accept": false, - "reason": "Model load failure vs a dependency-version bump issue; unrelated concrete bugs." + "reason": "Phi-4 multimodal load failure is unrelated to the PEFT version bump issue." + }, + { + "left": "issue:43452", + "right": "issue:44451", + "accept": false, + "reason": "gguf_file regressions for tokenizer/model loading are not the same as loading ScandiBERT." }, { "left": "issue:43525", "right": "issue:43572", "accept": false, - "reason": "Both are missing config fields after v5, but they affect different models and different fields (pad_token_id vs pad_token_idx)." + "reason": "Missing pad_token_id in Llama4Config and missing pad_token_idx in StableLmConfig are separate model/config bugs." }, { "left": "issue:43493", "right": "issue:43825", "accept": false, - "reason": "SigLIP2 implementation discrepancy vs pipeline error-message wording; no shared underlying code path." + "reason": "SigLIP2 implementation discrepancy is unrelated to the pipeline translation-support error message." }, { - "left": "issue:43452", - "right": "issue:44451", + "left": "issue:44625", + "right": "issue:45584", "accept": false, - "reason": "gguf_file loading regression vs a specific model that no longer loads. Too different to be duplicates." + "reason": "Qwen3.5 num_labels propagation and Whisper empty-transcription generation failure are distinct." }, { "left": "issue:45823", "right": "issue:46050", "accept": false, - "reason": "Device-mapping mismatch during forward vs quantized-model save/load breakage. Different lifecycle stages and symptoms." - }, - { - "left": "issue:44625", - "right": "issue:45584", - "accept": false, - "reason": "Qwen3.5 label-propagation bug vs Whisper empty-transcription generation failure. Different models and code paths." + "reason": "Gemma4 device-mapping mismatch is unrelated to sinq quantization save/load breakage." }, { "left": "issue:43994", "right": "issue:46192", "accept": false, - "reason": "SigLIP2 bad outputs vs return_tensors being ignored in multimodal kwargs. Not the same bug." + "reason": "SigLIP2 wrong outputs via AutoModel/pipeline is not the same as return_tensors being ignored with text_kwargs." }, { "left": "issue:28218", "right": "issue:33357", "accept": false, - "reason": "Tokenizer spacing behavior vs a Mac bus error on a CLIP model. No duplicate relationship." - }, - { - "left": "issue:44849", - "right": "issue:46209", - "accept": false, - "reason": "Qwen3.5 hidden-state bug vs FP8 inference failure on B200. Distinct issues." + "reason": "Tokenizer extra-space behavior is unrelated to the MacOS bus error in CLIP loading." }, { "left": "issue:39692", "right": "issue:43295", "accept": false, - "reason": "SigLIP2 doc-example errors vs a regression in processor.tokenizer access/image handling. Different failures." + "reason": "SigLIP2 docs-example errors do not match the processor.tokenizer regression in custom model code." }, { "left": "issue:43278", "right": "issue:43844", "accept": false, - "reason": "Embedding dtype drift and gradient explosion under ZeRO-3 are not the same underlying bug." + "reason": "Embedding dtype change at evaluation time is unrelated to the ZeRO-3 gradient growth issue." }, { "left": "issue:43493", "right": "issue:43844", "accept": false, - "reason": "SigLIP2 model discrepancy vs training instability under DeepSpeed. Different concerns." + "reason": "SigLIP2 HF-vs-JAX discrepancy is unrelated to abnormal gradients under DeepSpeed ZeRO-3." }, { "left": "issue:33357", "right": "issue:33666", "accept": false, - "reason": "Mac bus error on CLIP vs Qwen2-VL multi-GPU training. No overlap beyond being model issues." - }, - { - "left": "issue:44488", - "right": "issue:44843", - "accept": false, - "reason": "Model-specific load failure vs an offline-mode regression caused by an unconditional hub call. Different root causes." - }, - { - "left": "issue:43646", - "right": "issue:44466", - "accept": false, - "reason": "Broad v5 custom-init breakage vs inconsistent lm_head.weight serialization by device. Related to v5, but not the same defect." + "reason": "MacOS bus error in CLIP loading is unrelated to Qwen2-VL multi-GPU training." }, { "left": "issue:44849", - "right": "issue:45357", - "accept": false, - "reason": "Hidden-states output bug vs incorrect visual encoder keys on save_pretrained. Different Qwen3.5 regressions." - }, - { - "left": "issue:44568", - "right": "issue:46489", + "right": "issue:46209", "accept": false, - "reason": "Both are tokenizer regressions in v5+, but they affect different tokenizers and different behaviors (special-token addition vs incorrect output)." + "reason": "Qwen3.5 hidden-states bug is unrelated to FP8 inference failure on B200." } ] }, @@ -7832,14 +7756,10 @@ "source": "llm", "cache_hit": false, "model": "gpt-5.4-mini?service_tier=flex", - "cluster_id": "cluster-10000-472", + "cluster_id": "cluster-10000-473", "nodes": [ - "issue:29159", "issue:29942", "issue:30333", - "issue:32090", - "issue:42490", - "issue:43065", "issue:43257", "issue:43278", "issue:43299", @@ -7849,8 +7769,8 @@ "issue:43572", "issue:43575", "issue:43606", + "issue:43646", "issue:43653", - "issue:43696", "issue:43701", "issue:43824", "issue:43844", @@ -7861,15 +7781,19 @@ "issue:44291", "issue:44361", "issue:44464", + "issue:44466", + "issue:44488", "issue:44493", "issue:44568", "issue:44589", "issue:44792", + "issue:44843", + "issue:44849", "issue:44964", "issue:44977", - "issue:45084", "issue:45200", "issue:45229", + "issue:45357", "issue:45405", "issue:45520", "issue:45584", @@ -7879,10 +7803,13 @@ "issue:45823", "issue:46050", "issue:46285", - "issue:46369", - "issue:46424" + "issue:46424", + "issue:46489" ], "soft_pairs": [ + "issue:43646|issue:44466", + "issue:44488|issue:44843", + "issue:44568|issue:46489", "issue:43278|issue:43701", "issue:43927|issue:44361", "issue:43540|issue:44977", @@ -7892,229 +7819,214 @@ "issue:44568|issue:45812", "issue:44792|issue:45405", "issue:43504|issue:43606", + "issue:44849|issue:45357", "issue:43824|issue:44589", "issue:43334|issue:43572", "issue:43606|issue:44112", "issue:43334|issue:45759", "issue:44464|issue:46424", - "issue:45200|issue:45229", "issue:43572|issue:44291", - "issue:44493|issue:46285", "issue:43504|issue:43653", - "issue:45229|issue:45823", + "issue:44493|issue:46285", + "issue:45200|issue:45229", "issue:44464|issue:45584", + "issue:45229|issue:45823", "issue:43994|issue:44493", "issue:43873|issue:46050", - "issue:45588|issue:46050", "issue:43844|issue:44964", - "issue:29942|issue:30333", - "issue:29159|issue:32090", - "issue:42490|issue:43696", - "issue:43065|issue:46369", - "issue:45084|issue:45520" + "issue:45588|issue:46050", + "issue:29942|issue:30333" ], - "prepared_review_unit_hash": "6680650f8acdc11ab89204b2eeb3a6b96dd34b9f7368786097ccc6fc8bad4fab", + "prepared_review_unit_hash": "2d3067b1f4312f8720e40d3f067501e1b0ed0dab1291de0b210110601ceeb056", "component_budget": { - "node_count": 472, - "item_count": 472, - "soft_pair_count": 1732, - "serialized_chars": 898654, - "estimated_input_tokens": 224664, - "estimated_eval_tokens": 449584 + "node_count": 473, + "item_count": 473, + "soft_pair_count": 1738, + "serialized_chars": 901437, + "estimated_input_tokens": 225360, + "estimated_eval_tokens": 450976 }, "budget": { "node_count": 47, "item_count": 47, "soft_pair_count": 29, - "serialized_chars": 35767, - "estimated_input_tokens": 8942, - "estimated_eval_tokens": 18140 + "serialized_chars": 35932, + "estimated_input_tokens": 8983, + "estimated_eval_tokens": 18222 }, "overflow_policy": "truncate_then_skip", "trimmed": false, "aggressively_trimmed": false, "split": true, "analyst_result": { - "summary": "All soft-pair candidates are superficial topic matches rather than the same underlying bug. They span unrelated areas (tokenizers, Qwen/VL loading, quantization, offload, CI, memory, config regressions), so none should be merged as duplicates.", - "confidence": 0.96, - "canonical_issue_reason": "No clear canonical issue: the candidate pairs are heterogeneous and do not collapse into one duplicate bug cluster.", + "summary": "This cluster is mostly a loose collection of unrelated Transformers issues, with a few recurring themes around v5 regression, tokenizer/config loading, and Qwen/Gemma model-specific failures. None of the proposed soft pairs looks like the same underlying bug/change.", + "confidence": 0.41, + "canonical_issue_reason": "issue:43646 is the broadest, most general regression report in the set and best matches the common v5 loading/initialization failure theme.", "canonical_pr_reason": null, - "best_issue_reason": "No single issue is a good global representative because the set contains many unrelated reports with only loose keyword similarity.", + "best_issue_reason": "issue:43646 is the best single anchor because its scope is broad enough to represent the cluster\u2019s repeated model-loading breakages, unlike the many highly model-specific reports.", "best_pr_reason": null, "soft_edge_verdicts": [ + { + "left": "issue:43646", + "right": "issue:44466", + "accept": false, + "reason": "Generic initialization breakage vs. tied-weight serialization inconsistency; different failure modes and code paths." + }, + { + "left": "issue:44488", + "right": "issue:44843", + "accept": false, + "reason": "Specific model load failure vs. offline-mode hub call bug; not the same root cause." + }, + { + "left": "issue:44568", + "right": "issue:46489", + "accept": false, + "reason": "Both tokenizer bugs, but for different models and likely different fixes." + }, { "left": "issue:43278", "right": "issue:43701", "accept": false, - "reason": "Different bugs: dtype drift in evaluation vs resume checkpoint key mismatch." + "reason": "BF16-to-FP32 dtype drift in eval vs. resume checkpoint key mismatch; unrelated bugs." }, { "left": "issue:43927", "right": "issue:44361", "accept": false, - "reason": "Different failure modes: lost custom token IDs after save/load vs tokenizer AttributeError on tasks." + "reason": "Different tokenizer/config classes and different exceptions; no shared code-path evidence." }, { "left": "issue:43540", "right": "issue:44977", "accept": false, - "reason": "Both mention Qwen, but one is video-processing validation and the other is flash-attention generation behavior." + "reason": "Video input processing failure vs. flash-attention generation bug; distinct problems." }, { "left": "issue:44589", "right": "issue:45520", "accept": false, - "reason": "Unrelated issues: float8 storage import error vs flash_attn import_utils KeyError on Python 3.13." + "reason": "Float8 storage deserialization error vs. flash_attn import key error; unrelated." }, { "left": "issue:43257", "right": "issue:43299", "accept": false, - "reason": "Both involve Qwen3 MOE, but one is accelerate+deepspeed conversion and the other is a dev-version model-loading regression." + "reason": "Both involve Qwen MoE loading, but one is accelerate/deepspeed conversion and the other is Qwen3VL model loading; too different to merge." }, { "left": "issue:43575", "right": "issue:45405", "accept": false, - "reason": "OOM during TP load is unrelated to an unreleased PEFT version bump." + "reason": "GPU OOM on tensor-parallel load vs. dependency version pin issue." }, { "left": "issue:44568", "right": "issue:45812", "accept": false, - "reason": "Different tokenizer regressions: missing BOS/EOS on one tokenizer vs wrong IDs across Granite models." + "reason": "Different tokenizer regressions affecting different model families; no same underlying bug." }, { "left": "issue:44792", "right": "issue:45405", "accept": false, - "reason": "Janus image-generation test failure is unrelated to the PEFT version pin issue." + "reason": "Janus test failure vs. PEFT version bump; unrelated." }, { "left": "issue:43504", "right": "issue:43606", "accept": false, - "reason": "Both are load/fail issues, but they concern different models and different root causes." + "reason": "Legacy field loading bug vs. CPU offload device mismatch; unrelated." + }, + { + "left": "issue:44849", + "right": "issue:45357", + "accept": false, + "reason": "Forward/output_hidden_states bug vs. save_pretrained key regression; same model family but different code paths." }, { "left": "issue:43824", "right": "issue:44589", "accept": false, - "reason": "ImportError for a missing Qwen2.5-VL symbol is unrelated to float8 storage lookup failure." + "reason": "Import error for a specific class vs. float8 storage lookup failure; unrelated." }, { "left": "issue:43334", "right": "issue:43572", "accept": false, - "reason": "Both are config attribute problems, but for different models and different missing fields." + "reason": "Both missing config fields, but for different models and different attributes; not the same bug." }, { "left": "issue:43606", "right": "issue:44112", "accept": false, - "reason": "Both are CI/device-related, but on different models and different mismatch causes." + "reason": "CPU offload mismatch in bark vs. a stale CI device override test in GraniteSpeech; different bugs." }, { "left": "issue:43334", "right": "issue:45759", "accept": false, - "reason": "Different composite-model loading bugs: missing pad_token_id vs not unwrapping text_config in AutoModelForCausalLM.from_config." + "reason": "Both Qwen multimodal config issues, but one is missing pad token handling and the other is from_config unwrapping text_config." }, { "left": "issue:44464", "right": "issue:46424", "accept": false, - "reason": "Chunked generation inconsistency with compiled forward is unrelated to cache length handling." - }, - { - "left": "issue:45200", - "right": "issue:45229", - "accept": false, - "reason": "Different Gemma 4 problems: missing mm_token_type_ids defaults vs multi-GPU CUDA OOM." + "reason": "Compiled-forward generation inconsistency vs. cache-length handling for quantized cache; not the same issue." }, { "left": "issue:43572", "right": "issue:44291", "accept": false, - "reason": "Both involve config/init issues after v5, but one is a missing pad_token_idx field and the other is an unexpected init_empty_weights argument." - }, - { - "left": "issue:44493", - "right": "issue:46285", - "accept": false, - "reason": "Unexpected key with position id is not the same regression as CLIP text-model structure divergence." + "reason": "Config field rename bug vs. init_empty_weights argument incompatibility; unrelated." }, { "left": "issue:43504", "right": "issue:43653", "accept": false, - "reason": "Legacy-field load failure for BEiT is unrelated to BigBirdTokenizer mask token registration." + "reason": "Legacy preset loading bug vs. tokenizer special-token registration bug; distinct fixes." }, { - "left": "issue:45229", - "right": "issue:45823", + "left": "issue:44493", + "right": "issue:46285", "accept": false, - "reason": "Both are Gemma4 inference issues, but OOM vs device mismatch are distinct problems." + "reason": "Unexpected key regression across many models vs. CLIP class structure inconsistency; not one concrete bug." }, { - "left": "issue:44464", - "right": "issue:45584", + "left": "issue:45200", + "right": "issue:45229", "accept": false, - "reason": "Compiled-forward chunked generation inconsistency is unrelated to Whisper empty-transcription failure." + "reason": "Missing default for mm_token_type_ids vs. multi-GPU CUDA OOM; different problems." }, { "left": "issue:43994", "right": "issue:44493", "accept": false, - "reason": "Siglip2 bad outputs and position-id key issues are different model/regression paths." + "reason": "Wrong SigLIP outputs vs. unexpected position-id keys; unrelated symptoms and paths." }, { "left": "issue:43873", "right": "issue:46050", "accept": false, - "reason": "Quantization offload behavior and sinq save/load corruption are not the same bug." - }, - { - "left": "issue:45588", - "right": "issue:46050", - "accept": false, - "reason": "Flash-attention sink-less AttributeError is unrelated to quantized model persistence." + "reason": "Quantization offloading behavior vs. quantized model save/load breakage; different failure modes." }, { "left": "issue:43844", "right": "issue:44964", "accept": false, - "reason": "DeepSpeed gradient growth and Phi-4 multimodal loading are different symptoms and code paths." - }, - { - "left": "issue:29942", - "right": "issue:30333", - "accept": false, - "reason": "Flash Attention 2 test failures and MLflow job-reporting problems are unrelated." - }, - { - "left": "issue:29159", - "right": "issue:32090", - "accept": false, - "reason": "Tokenizer slow/fast inconsistency is unrelated to Trainer NoneType broadcast failure." - }, - { - "left": "issue:42490", - "right": "issue:43696", - "accept": false, - "reason": "Local-path slash import warning is unrelated to CUDA OOM on GPT-oss-20b." + "reason": "DeepSpeed gradient anomaly vs. Phi-4 multimodal load failure; unrelated." }, { - "left": "issue:43065", - "right": "issue:46369", + "left": "issue:45588", + "right": "issue:46050", "accept": false, - "reason": "Sam3PixelDecoder dummy conv and sam3-litetext memory overflow are different issues." + "reason": "flash_attention sink-less model crash vs. sinq quantized save/load corruption; unrelated." }, { - "left": "issue:45084", - "right": "issue:45520", + "left": "issue:29942", + "right": "issue:30333", "accept": false, - "reason": "Non-template node compilation error is unrelated to flash_attn import-utils KeyError." + "reason": "Flash Attention 2 test failures vs. MLflow job-reporting issue; no shared bug." } ] }, @@ -8136,15 +8048,18 @@ "source": "llm", "cache_hit": false, "model": "gpt-5.4-mini?service_tier=flex", - "cluster_id": "cluster-10000-472", + "cluster_id": "cluster-10000-473", "nodes": [ "issue:15354", "issue:28218", + "issue:29159", + "issue:32090", "issue:34567", "issue:37428", "issue:38175", "issue:39692", "issue:42490", + "issue:43065", "issue:43232", "issue:43295", "issue:43450", @@ -8153,16 +8068,14 @@ "issue:43550", "issue:43572", "issue:43575", - "issue:43653", + "issue:43696", "issue:43701", "issue:43824", "issue:43844", "issue:43873", "issue:43957", - "issue:44112", "issue:44186", "issue:44423", - "issue:44466", "issue:44521", "issue:44554", "issue:44610", @@ -8170,12 +8083,11 @@ "issue:44849", "issue:44987", "issue:45072", - "issue:45127", + "issue:45084", "issue:45198", "issue:45200", "issue:45290", "issue:45356", - "issue:45381", "issue:45406", "issue:45446", "issue:45520", @@ -8185,9 +8097,13 @@ "issue:46121", "issue:46178", "issue:46192", - "issue:46209" + "issue:46369" ], "soft_pairs": [ + "issue:42490|issue:43696", + "issue:29159|issue:32090", + "issue:43065|issue:46369", + "issue:45084|issue:45520", "issue:28218|issue:34567", "issue:44423|issue:44734", "issue:44554|issue:45446", @@ -8197,243 +8113,239 @@ "issue:45072|issue:45198", "issue:37428|issue:43957", "issue:43295|issue:43450", - "issue:45588|issue:45759", "issue:43540|issue:45200", + "issue:45588|issue:45759", "issue:44186|issue:46192", "issue:45200|issue:45823", "issue:45823|issue:46192", "issue:43295|issue:44610", "issue:43295|issue:43873", - "issue:43232|issue:44849", "issue:43450|issue:43575", "issue:45520|issue:46178", "issue:43572|issue:44987", "issue:43550|issue:43844", - "issue:39692|issue:46192", + "issue:43232|issue:44849", "issue:44554|issue:45356", "issue:37428|issue:43824", + "issue:39692|issue:46192", "issue:43572|issue:43957", "issue:44521|issue:45290", - "issue:42490|issue:43493", - "issue:44466|issue:45127", - "issue:45381|issue:46209", - "issue:38175|issue:43493", - "issue:43653|issue:44112" + "issue:42490|issue:43493" ], - "prepared_review_unit_hash": "6c011f19ffc24d16139c23ff28efa232820b6ce5a33f29d1cd158224d87bdc3d", + "prepared_review_unit_hash": "3bcac6422b86c61787028ef716d033e70417c43a33dc7b078f8a5e0d8e6bb6f3", "component_budget": { - "node_count": 472, - "item_count": 472, - "soft_pair_count": 1732, - "serialized_chars": 898654, - "estimated_input_tokens": 224664, - "estimated_eval_tokens": 449584 + "node_count": 473, + "item_count": 473, + "soft_pair_count": 1738, + "serialized_chars": 901437, + "estimated_input_tokens": 225360, + "estimated_eval_tokens": 450976 }, "budget": { "node_count": 48, "item_count": 48, "soft_pair_count": 31, - "serialized_chars": 37337, - "estimated_input_tokens": 9335, - "estimated_eval_tokens": 18926 + "serialized_chars": 37154, + "estimated_input_tokens": 9289, + "estimated_eval_tokens": 18834 }, "overflow_policy": "truncate_then_skip", "trimmed": false, "aggressively_trimmed": false, "split": true, "analyst_result": { - "summary": "This cluster is noisy: the items span many unrelated Transformers bug reports, mostly around different models/subsystems. None of the soft pairs look like the same underlying bug with enough specificity to merge.", - "confidence": 0.93, - "canonical_issue_reason": "issue:44521 is the strongest standalone candidate: it is open, actively discussed, has inbound references, and is a concrete user-facing multimodal chat-template bug. It is not a duplicate of the other items, but it is the best representative issue in this set.", + "summary": "This cluster is a loose similarity bucket of mostly unrelated Transformers issues. The shared keywords (tokenizer, processor, serve, import, multimodal) are too broad to indicate duplicate bugs, so no soft edges should be merged.", + "confidence": 0.95, + "canonical_issue_reason": "Issue 44521 is the best anchor only because it is open, actively discussed, and has the most inbound references; however, the cluster is too heterogeneous to have a true canonical duplicate.", "canonical_pr_reason": null, - "best_issue_reason": "issue:44521 has the best mix of activity and impact, and its problem statement is clear enough to serve as the cluster representative if one must be chosen.", + "best_issue_reason": "Issue 44521 is the most representative high-signal report in the set due to its active discussion and broad multimodal chat-template impact, but it still does not unify the others as duplicates.", "best_pr_reason": null, "soft_edge_verdicts": [ + { + "left": "issue:42490", + "right": "issue:43696", + "accept": false, + "reason": "Local-model trailing-slash import warning vs GPT-oss OOM; unrelated failure modes." + }, + { + "left": "issue:29159", + "right": "issue:32090", + "accept": false, + "reason": "Tokenizer spacing inconsistency vs Trainer broadcast NoneType error; different components." + }, + { + "left": "issue:43065", + "right": "issue:46369", + "accept": false, + "reason": "Both mention SAM3, but one is a dummy Conv2d placeholder and the other is a memory overflow; not the same bug." + }, + { + "left": "issue:45084", + "right": "issue:45520", + "accept": false, + "reason": "Compile error on non-template nodes vs flash_attn import-key error; unrelated code paths." + }, { "left": "issue:28218", "right": "issue:34567", "accept": false, - "reason": "Tokenizer spacing behavior and trainer token-count tracking are unrelated bugs." + "reason": "Added-token spacing bug vs TrainerState token counter not updating; different behavior." }, { "left": "issue:44423", "right": "issue:44734", "accept": false, - "reason": "Both involve serve crashes, but they fail in different code paths: multimodal continuous batching vs KV-cache continuation indexing." + "reason": "Both involve serve, but one is multimodal input conversion and the other is KV-cache continuation indexing." }, { "left": "issue:44554", "right": "issue:45446", "accept": false, - "reason": "Different subsystems and failure modes: MPS attention correctness vs a flex_attention import/version check." + "reason": "MPS attention correctness issue vs flex_attention AuxRequest version check; different subsystems." }, { "left": "issue:15354", "right": "issue:38175", "accept": false, - "reason": "JIT export on ViT and SigLIP2 zero probabilities are distinct model/runtime issues." + "reason": "TorchScript export failure vs zero-probability output in another vision model; not the same underlying bug." }, { "left": "issue:45406", "right": "issue:46121", "accept": false, - "reason": "Gemma4Processor missing _tokenizer and rope-validation list handling are unrelated bugs." + "reason": "Gemma4Processor _tokenizer crash vs rope-param JSON list type handling; unrelated." }, { "left": "issue:43493", "right": "issue:43701", "accept": false, - "reason": "SigLIP2 implementation discrepancy and resume-from-checkpoint key mismatch do not describe the same bug." + "reason": "SigLIP2 implementation discrepancy vs resume checkpoint key mismatch; different root causes." }, { "left": "issue:45072", "right": "issue:45198", "accept": false, - "reason": "Different models and failures: bfloat16 dtype mismatch vs Wav2Vec2 save/tokenization failure." + "reason": "bfloat16 dtype mismatch vs Wav2Vec2 save/tokenization failure; no shared code path." }, { "left": "issue:37428", "right": "issue:43957", "accept": false, - "reason": "Flash-attention import error and meta-device model loading breakage are separate issues." + "reason": "flash_attention_utils import error vs meta-device loading regression; unrelated." }, { "left": "issue:43295", "right": "issue:43450", "accept": false, - "reason": "Both are multimodal processor bugs, but one is a tokenizer/processor regression and the other is a batched video shape bug." + "reason": "Processor/tokenizer regression vs batched video shape bug; same broad area, different issue." }, { - "left": "issue:45588", - "right": "issue:45759", + "left": "issue:43540", + "right": "issue:45200", "accept": false, - "reason": "Flash-attention sink handling and composite-config text_config unwrapping are different code paths." + "reason": "Video-processing ValueError vs mm_token_type_ids defaulting; different components." }, { - "left": "issue:43540", - "right": "issue:45200", + "left": "issue:45588", + "right": "issue:45759", "accept": false, - "reason": "Video input processing in Qwen3OmniMoe and Gemma4 mm_token_type_ids defaults are unrelated." + "reason": "s_aux=None crash in flash_attention integration vs AutoModel config unwrapping; unrelated." }, { "left": "issue:44186", "right": "issue:46192", "accept": false, - "reason": "LayoutLMv2 tokenizer crashes and return_tensors being ignored are different tokenizer API bugs." + "reason": "LayoutLMv2Tokenizer crash on NER/padding vs return_tensors precedence; different tokenizer bugs." }, { "left": "issue:45200", "right": "issue:45823", "accept": false, - "reason": "Gemma4 fine-tuning defaults and device-map forward mismatch are separate issues." + "reason": "Missing mm_token_type_ids vs Gemma4 PLE device mismatch; unrelated." }, { "left": "issue:45823", "right": "issue:46192", "accept": false, - "reason": "Device placement in Gemma4 forward and return_tensors handling are unrelated." + "reason": "Device_map/forward device mismatch vs return_tensors being ignored; no common fix." }, { "left": "issue:43295", "right": "issue:44610", "accept": false, - "reason": "Processor/tokenizer regression and OmDet-Turbo input-size mismatch are not the same bug." + "reason": "Processor/tokenizer regression vs OmDet-Turbo input-size mismatch; different models and failures." }, { "left": "issue:43295", "right": "issue:43873", "accept": false, - "reason": "Multimodal processor regression and quantization offloading failure are different problems." - }, - { - "left": "issue:43232", - "right": "issue:44849", - "accept": false, - "reason": "Generation kwargs update after sync_gpus and Qwen3.5 hidden-states output bug are unrelated." + "reason": "Processor/tokenizer regression vs quantization offloading behavior; not the same bug." }, { "left": "issue:43450", "right": "issue:43575", "accept": false, - "reason": "Video processor batching and tensor-parallel OOM on Qwen2-57B are different issues." + "reason": "Video batch shape bug vs OOM when loading Qwen2-57B-A14B-Instruct; unrelated." }, { "left": "issue:45520", "right": "issue:46178", "accept": false, - "reason": "Both are backend/import errors, but they concern different backends and failure paths." + "reason": "flash_attn KeyError in import_utils vs tensorflow_text backend-mapping error; different import logic." }, { "left": "issue:43572", "right": "issue:44987", "accept": false, - "reason": "StableLmConfig pad_token_idx regression and a model-loading failure are not the same bug." + "reason": "StableLmConfig pad_token_idx regression vs loading physical-intelligence/fast on 5.1.0; different problems." }, { "left": "issue:43550", "right": "issue:43844", "accept": false, - "reason": "torch.compile SDPA failure and ZeRO-3 gradient growth are unrelated training/runtime issues." + "reason": "torch.compile/SDPA inference failure vs ZeRO-3 gradient anomaly; unrelated training/inference paths." }, { - "left": "issue:39692", - "right": "issue:46192", + "left": "issue:43232", + "right": "issue:44849", "accept": false, - "reason": "SigLIP2 doc-example errors and return_tensors override behavior are separate concerns." + "reason": "Generation kwargs update bug vs Qwen3.5 output_hidden_states bug; different code paths." }, { "left": "issue:44554", "right": "issue:45356", "accept": false, - "reason": "MPS attention correctness and Kimi tokenizer codec regression are different bugs." + "reason": "MPS attention correctness vs Kimi tokenizer codec regression; not the same underlying issue." }, { "left": "issue:37428", "right": "issue:43824", "accept": false, - "reason": "Different ImportErrors involving different symbols and models." + "reason": "Different ImportErrors for different symbols and models; no shared fix." + }, + { + "left": "issue:39692", + "right": "issue:46192", + "accept": false, + "reason": "SigLIP2 docs-example errors vs return_tensors handling; unrelated." }, { "left": "issue:43572", "right": "issue:43957", "accept": false, - "reason": "Config default regression and meta-device loading breakage do not match." + "reason": "Config default regression vs meta-device model-loading breakage; different failure modes." }, { "left": "issue:44521", "right": "issue:45290", "accept": false, - "reason": "Both hit apply_chat_template, but one is assistant-mask generation for multimodal inputs and the other is tool-call content handling." + "reason": "Both use apply_chat_template, but one is multimodal assistant masks and the other is tool-call/no-content tokenization crash." }, { "left": "issue:42490", "right": "issue:43493", "accept": false, - "reason": "Local-path import warning and SigLIP2 output discrepancy are unrelated." - }, - { - "left": "issue:44466", - "right": "issue:45127", - "accept": false, - "reason": "Tied-weight serialization on different devices and LoRA merge collapse are not the same concrete bug." - }, - { - "left": "issue:45381", - "right": "issue:46209", - "accept": false, - "reason": "Qwen2.5-VL video position IDs and FP8 inference on B200 are unrelated." - }, - { - "left": "issue:38175", - "right": "issue:43493", - "accept": false, - "reason": "Both concern SigLIP2, but one is a specific zero-probability symptom and the other is a broader implementation-vs-JAX discrepancy; not enough to treat as the same bug." - }, - { - "left": "issue:43653", - "right": "issue:44112", - "accept": false, - "reason": "BigBirdTokenizer special-token registration and GraniteSpeech CI device override are unrelated." + "reason": "Trailing-slash local model warning vs SigLIP2 implementation discrepancy; unrelated." } ] }, @@ -8455,7 +8367,7 @@ "source": "llm", "cache_hit": false, "model": "gpt-5.4-mini?service_tier=flex", - "cluster_id": "cluster-10000-472", + "cluster_id": "cluster-10000-473", "nodes": [ "issue:28218", "issue:35141", @@ -8469,10 +8381,11 @@ "issue:43388", "issue:43450", "issue:43452", + "issue:43493", "issue:43525", "issue:43526", "issue:43550", - "issue:43606", + "issue:43653", "issue:43819", "issue:43824", "issue:43827", @@ -8483,6 +8396,7 @@ "issue:44112", "issue:44206", "issue:44442", + "issue:44466", "issue:44479", "issue:44561", "issue:44610", @@ -8490,16 +8404,14 @@ "issue:44857", "issue:44898", "issue:45083", + "issue:45127", "issue:45198", "issue:45200", "issue:45325", + "issue:45381", "issue:45412", - "issue:45520", "issue:45563", "issue:45584", - "issue:45593", - "issue:45759", - "issue:46170", "issue:46192", "issue:46209", "issue:46234", @@ -8507,13 +8419,17 @@ "issue:46512" ], "soft_pairs": [ + "issue:45381|issue:46209", + "issue:44466|issue:45127", + "issue:38175|issue:43493", + "issue:43653|issue:44112", "issue:43278|issue:43856", "issue:28218|issue:35141", - "issue:42915|issue:43824", "issue:43452|issue:46512", + "issue:42915|issue:43824", "issue:43950|issue:46285", - "issue:44898|issue:45412", "issue:43334|issue:43525", + "issue:44898|issue:45412", "issue:43262|issue:43450", "issue:43824|issue:43975", "issue:44442|issue:45198", @@ -8524,207 +8440,203 @@ "issue:43819|issue:44112", "issue:43827|issue:43901", "issue:44610|issue:44857", - "issue:44561|issue:46234", "issue:41720|issue:46209", + "issue:44561|issue:46234", "issue:38175|issue:45412", "issue:43335|issue:43550", - "issue:45759|issue:46285", - "issue:44206|issue:45584", - "issue:44857|issue:45593", "issue:43819|issue:45083", - "issue:45520|issue:46170", - "issue:43606|issue:45083" + "issue:44206|issue:45584" ], - "prepared_review_unit_hash": "2dfa22095871daf8ae4b514df0a582b3dbb7b02b87ea52da96d04ce8c68181cf", + "prepared_review_unit_hash": "d9db45a676e4b57703728969591b33178aed647fa2e69a02a4aebedbac6539e3", "component_budget": { - "node_count": 472, - "item_count": 472, - "soft_pair_count": 1732, - "serialized_chars": 898654, - "estimated_input_tokens": 224664, - "estimated_eval_tokens": 449584 + "node_count": 473, + "item_count": 473, + "soft_pair_count": 1738, + "serialized_chars": 901437, + "estimated_input_tokens": 225360, + "estimated_eval_tokens": 450976 }, "budget": { "node_count": 48, "item_count": 48, "soft_pair_count": 27, - "serialized_chars": 36099, - "estimated_input_tokens": 9025, - "estimated_eval_tokens": 18306 + "serialized_chars": 36198, + "estimated_input_tokens": 9050, + "estimated_eval_tokens": 18356 }, "overflow_policy": "truncate_then_skip", "trimmed": false, "aggressively_trimmed": false, "split": true, "analyst_result": { - "summary": "These issues are mostly unrelated; the soft pairs are superficial title-level matches rather than the same underlying bug. No pull requests are present.", - "confidence": 0.79, - "canonical_issue_reason": "issue:43950 is the strongest representative issue here because it describes a broad core regression in `from_pretrained()` affecting model loading behavior across many models, making it more generally useful than the niche model-specific reports.", + "summary": "Mostly unrelated bug reports; none of the soft pairs look like the same underlying bug/change with enough confidence to merge. The closest thematic overlaps are in model loading/serialization and Qwen multimodal regressions, but they still point to different code paths and model-specific failures.", + "confidence": 0.87, + "canonical_issue_reason": "issue:43950 is the strongest representative: it describes a clear, high-severity transformers 5.x regression in a core loading path (`from_pretrained`) with concrete corruption behavior and broad impact.", "canonical_pr_reason": null, - "best_issue_reason": "issue:43950 has the widest apparent impact and clearest core-library relevance among this set, so it is the best single issue to surface for triage.", + "best_issue_reason": "issue:43950 is the best standalone issue in this set because it is specific, actionable, and affects a central API path rather than a model-specific edge case or test/doc inconsistency.", "best_pr_reason": null, "soft_edge_verdicts": [ + { + "left": "issue:45381", + "right": "issue:46209", + "accept": false, + "reason": "Different failures: Qwen2.5-VL video `vision_position_ids` regression vs FP8 inference crash on B200." + }, + { + "left": "issue:44466", + "right": "issue:45127", + "accept": false, + "reason": "Both touch tied weights, but one is device-dependent serialization and the other is LoRA merge with extended vocab; different bugs and fixes." + }, + { + "left": "issue:38175", + "right": "issue:43493", + "accept": false, + "reason": "Same model family, but zero probabilities and HF-vs-JAX discrepancy are not clearly the same concrete code-path issue." + }, + { + "left": "issue:43653", + "right": "issue:44112", + "accept": false, + "reason": "Tokenizer special-token registration bug vs GraniteSpeech CI device-override failure; unrelated." + }, { "left": "issue:43278", "right": "issue:43856", "accept": false, - "reason": "Different problems: dtype drift between training/eval vs memory inefficiency during Qwen3 MoE training." + "reason": "Embedding dtype regression vs MoE training memory usage; different symptoms and subsystems." }, { "left": "issue:28218", "right": "issue:35141", "accept": false, - "reason": "One is tokenizer spacing after added tokens; the other is embedding reinitialization after resizing. Different code paths and symptoms." + "reason": "Tokenizer spacing after added token is unrelated to embedding reinitialization after resize." }, { - "left": "issue:42915", - "right": "issue:43824", + "left": "issue:43452", + "right": "issue:46512", "accept": false, - "reason": "Unrelated: FP8/Qwen3MoE failure vs a Qwen2.5-VL import error for a missing class." + "reason": "gguf_file loading failure and stale `config.dtype` after `from_config` are separate loading-path issues." }, { - "left": "issue:43452", - "right": "issue:46512", + "left": "issue:42915", + "right": "issue:43824", "accept": false, - "reason": "Different failure modes: gguf loading in tokenizer/model loaders vs `from_config(dtype=...)` leaving `config.dtype` stale." + "reason": "Qwen3Moe FP8 failure and Qwen2.5-VL import error are different model families and failure modes." }, { "left": "issue:43950", "right": "issue:46285", "accept": false, - "reason": "Core-loading buffer corruption vs CLIP text model structure inconsistency; not the same bug." + "reason": "Non-persistent buffer corruption in `from_pretrained` vs CLIP text-model structure inconsistency; not the same bug." }, { - "left": "issue:44898", - "right": "issue:45412", + "left": "issue:43334", + "right": "issue:43525", "accept": false, - "reason": "Different subsystems: Perceiver image-resolution handling vs RT-DETR memory release/GC behavior." + "reason": "Both mention missing `pad_token_id`, but they are model-specific config failures without evidence of one shared underlying fix." }, { - "left": "issue:43334", - "right": "issue:43525", + "left": "issue:44898", + "right": "issue:45412", "accept": false, - "reason": "Both mention missing `pad_token_id`, but they are separate model families and do not clearly share the same concrete bug." + "reason": "Perceiver interpolation failure and RT-DETR memory leak are unrelated." }, { "left": "issue:43262", "right": "issue:43450", "accept": false, - "reason": "Audio chat template sampling-rate default vs batched video processor output shape; unrelated processing bugs." + "reason": "Audio chat-template sample-rate default bug vs batched video processor shape bug." }, { "left": "issue:43824", "right": "issue:43975", "accept": false, - "reason": "Import error for a missing model class vs detokenization regression in a different model." + "reason": "Qwen2.5-VL import error and DeepSeek detokenization regression are unrelated." }, { "left": "issue:44442", "right": "issue:45198", "accept": false, - "reason": "Both are speech/tokenization-related, but the failures are different and model-specific." + "reason": "AutoTokenizer loading failure for FastSpeech2Conformer vs Wav2Vec2 save/tokenization failure; different code paths." }, { "left": "issue:45200", "right": "issue:45325", "accept": false, - "reason": "Gemma 4 multimodal token-type defaults vs Qwen2.5-VL rope/position-id scaling; different issues." + "reason": "Gemma 4 text-only fine-tuning default and Qwen2.5-VL rope-index scaling are unrelated multimodal bugs." }, { "left": "issue:44479", "right": "issue:46192", "accept": false, - "reason": "Video input regression for Qwen VL models vs `text_kwargs` silently overriding `return_tensors`; unrelated." + "reason": "Qwen video-input regression and `return_tensors` being ignored are distinct API issues." }, { "left": "issue:43388", "right": "issue:43526", "accept": false, - "reason": "Different label-processing bugs in different components; no shared concrete code-path evidence." + "reason": "Metric gathering drops tuple labels; BeitImageProcessorFast reduces labels incorrectly. Similar area, different behavior." }, { "left": "issue:44792", "right": "issue:45563", "accept": false, - "reason": "Janus image generation test failure vs stale warning in paged generation; not the same issue." + "reason": "Janus image generation test failure and stale generate warning are unrelated." }, { "left": "issue:43819", "right": "issue:44112", "accept": false, - "reason": "DAC latent/forward mismatch vs GraniteSpeech CI device-override flake; unrelated." + "reason": "DAC latent mismatch and GraniteSpeech CI device override failure do not share the same code-path problem." }, { "left": "issue:43827", "right": "issue:43901", "accept": false, - "reason": "Both are docs issues, but they cover different API references and should not be merged as duplicates." + "reason": "Both are docs issues, but they concern different pipelines and different deprecated behaviors." }, { "left": "issue:44610", "right": "issue:44857", "accept": false, - "reason": "Processor resolution mismatch vs AMP/CUDA loss crash; different components and symptoms." + "reason": "OmDet-Turbo processor size mismatch vs LwDetrImageLoss AMP crash; unrelated." }, { - "left": "issue:44561", - "right": "issue:46234", + "left": "issue:41720", + "right": "issue:46209", "accept": false, - "reason": "Separate missing utility/import-availability issues affecting different symbols and code paths." + "reason": "Auto device-mapping cuda assert and B200 FP8 inference failure are different hardware/runtime failures." }, { - "left": "issue:41720", - "right": "issue:46209", + "left": "issue:44561", + "right": "issue:46234", "accept": false, - "reason": "Different hardware/runtime failures for different models; no shared underlying bug." + "reason": "Removal of one helper and absence of another are separate utility/API issues." }, { "left": "issue:38175", "right": "issue:45412", "accept": false, - "reason": "Zero probabilities in SigLIP2 vs RT-DETR memory leak; unrelated." + "reason": "SigLIP2 zero-probability issue and RT-DETR memory leak are unrelated." }, { "left": "issue:43335", "right": "issue:43550", "accept": false, - "reason": "Sparse-layer config bug vs torch.compile/SDPA failure; different model/configuration issues." - }, - { - "left": "issue:45759", - "right": "issue:46285", - "accept": false, - "reason": "Composite Qwen config unwrap bug vs CLIP model structure inconsistency; unrelated." - }, - { - "left": "issue:44206", - "right": "issue:45584", - "accept": false, - "reason": "Feature extractor arg mismatch vs Whisper empty-transcription generation failure; different code paths." - }, - { - "left": "issue:44857", - "right": "issue:45593", - "accept": false, - "reason": "AMP/CUDA crash in LwDetrImageLoss vs missing auxiliary losses in D-FINE; unrelated." + "reason": "SwitchTransformers sparse-layer config bug vs Bamba torch.compile/SDPA failure; different subsystems." }, { "left": "issue:43819", "right": "issue:45083", "accept": false, - "reason": "DAC latent mismatch vs Qwen3 Omni feature-length helper behavior; different subsystems." + "reason": "DAC.from_latents mismatch and qwen3_omni_moe feature-length helper behavior are unrelated." }, { - "left": "issue:45520", - "right": "issue:46170", - "accept": false, - "reason": "Import-utils key lookup for flash-attn on Python 3.13 vs Windows ROCm/torch.distributed dependency issue; unrelated." - }, - { - "left": "issue:43606", - "right": "issue:45083", + "left": "issue:44206", + "right": "issue:45584", "accept": false, - "reason": "CPU offload device mismatch vs feature-length helper behavior; no duplicate evidence." + "reason": "LaserFeatureExtractor unsupported arg and Whisper empty-transcription failure after token alignment are different bugs." } ] }, @@ -8746,32 +8658,29 @@ "source": "llm", "cache_hit": false, "model": "gpt-5.4-mini?service_tier=flex", - "cluster_id": "cluster-10000-472", + "cluster_id": "cluster-10000-473", "nodes": [ - "issue:26413", "issue:28282", "issue:29127", "issue:29942", - "issue:31501", "issue:32090", "issue:33357", "issue:34689", "issue:35141", "issue:37428", "issue:39692", - "issue:42490", "issue:43278", "issue:43335", "issue:43425", "issue:43504", "issue:43526", "issue:43550", + "issue:43606", "issue:43611", "issue:43653", "issue:43720", "issue:43819", "issue:43825", - "issue:43854", "issue:43931", "issue:44112", "issue:44351", @@ -8781,265 +8690,271 @@ "issue:44521", "issue:44743", "issue:44855", + "issue:44857", "issue:45083", "issue:45127", "issue:45200", "issue:45245", "issue:45290", - "issue:45310", - "issue:45478", + "issue:45520", "issue:45588", "issue:45593", "issue:45759", "issue:45859", + "issue:46170", + "issue:46285", "issue:46326", "issue:46431", "issue:46512", "issue:46566" ], "soft_pairs": [ + "issue:45759|issue:46285", + "issue:44857|issue:45593", + "issue:45520|issue:46170", + "issue:43606|issue:45083", "issue:37428|issue:43425", "issue:29942|issue:32090", "issue:43335|issue:43653", "issue:43550|issue:44112", "issue:28282|issue:33357", - "issue:43611|issue:43931", "issue:44448|issue:44743", - "issue:45759|issue:46512", + "issue:43611|issue:43931", "issue:44514|issue:44521", + "issue:45759|issue:46512", "issue:43335|issue:43504", "issue:44855|issue:45083", "issue:46431|issue:46566", "issue:43278|issue:43825", "issue:29127|issue:32090", "issue:37428|issue:44351", - "issue:43720|issue:44112", "issue:45290|issue:46326", + "issue:43720|issue:44112", "issue:43335|issue:45083", - "issue:45588|issue:45859", "issue:43335|issue:44112", "issue:45127|issue:45245", - "issue:45200|issue:45588", + "issue:45588|issue:45859", "issue:43335|issue:45593", - "issue:43819|issue:44493", + "issue:45200|issue:45588", "issue:34689|issue:35141", "issue:39692|issue:43526", - "issue:26413|issue:31501", - "issue:45310|issue:45478", - "issue:42490|issue:43854" + "issue:43819|issue:44493" ], - "prepared_review_unit_hash": "fb76d852513510a9e6c904c90a0cc07d5500f98cabc09877aa8eed1258e62cec", + "prepared_review_unit_hash": "21a2e561d3be9232f2d6a8cb1f18a04605995afa96a7034ae5da1d5b3840134a", "component_budget": { - "node_count": 472, - "item_count": 472, - "soft_pair_count": 1732, - "serialized_chars": 898654, - "estimated_input_tokens": 224664, - "estimated_eval_tokens": 449584 + "node_count": 473, + "item_count": 473, + "soft_pair_count": 1738, + "serialized_chars": 901437, + "estimated_input_tokens": 225360, + "estimated_eval_tokens": 450976 }, "budget": { - "node_count": 48, - "item_count": 48, - "soft_pair_count": 29, - "serialized_chars": 36703, - "estimated_input_tokens": 9176, - "estimated_eval_tokens": 18608 + "node_count": 47, + "item_count": 47, + "soft_pair_count": 30, + "serialized_chars": 36526, + "estimated_input_tokens": 9132, + "estimated_eval_tokens": 18520 }, "overflow_policy": "truncate_then_skip", "trimmed": false, "aggressively_trimmed": false, "split": true, "analyst_result": { - "summary": "Mostly conservative soft matches; the only clear duplicate is the exact-title Qwen3.5 MoE from_pretrained regression in 45310/45478. No pull requests are present in the packet.", - "confidence": 0.96, - "canonical_issue_reason": "issue:45310 is the better canonical issue because it is the earlier report and has the same exact title as issue:45478, with very similar surrounding metadata indicating the same reported regression.", + "summary": "No duplicate-worthy cluster emerged: the items are mostly unrelated issue reports, and the soft pairs share broad subsystems at most but not the same concrete bug.", + "confidence": 0.95, + "canonical_issue_reason": "No clear canonical issue: the cluster is heterogeneous, and the soft pairs differ in concrete failure mode even when they touch the same subsystem.", "canonical_pr_reason": null, - "best_issue_reason": "issue:45310 is the strongest representative issue in this cluster because it is a concrete, narrowly scoped regression report and the only clear duplicate target among the soft pairs.", + "best_issue_reason": "No single issue is a strong global representative; the closest-looking pairs still describe different bugs, regressions, or error paths.", "best_pr_reason": null, "soft_edge_verdicts": [ + { + "left": "issue:45759", + "right": "issue:46285", + "accept": false, + "reason": "Both involve multimodal/composite model config loading, but one is AutoModelForCausalLM.from_config unwrapping text_config and the other is a CLIP text model structure regression." + }, + { + "left": "issue:44857", + "right": "issue:45593", + "accept": false, + "reason": "Both are vision/detection related, but one is an AMP float16 crash and the other is missing auxiliary losses when denoising is disabled." + }, + { + "left": "issue:45520", + "right": "issue:46170", + "accept": false, + "reason": "Both are import/platform errors, but they stem from different dependencies and environments: flash_attn on Python 3.13 vs torch.distributed on Windows ROCm." + }, + { + "left": "issue:43606", + "right": "issue:45083", + "accept": false, + "reason": "One is a CPU offload device-mismatch bug in bark-small; the other is a helper-length calculation issue in qwen3_omni_moe." + }, { "left": "issue:37428", "right": "issue:43425", "accept": false, - "reason": "Both are import/version-related, but one is a missing internal symbol and the other is a Torch 2.10 compatibility report; different bugs." + "reason": "Both mention torch/flash-attn compatibility, but one is a missing internal flash-attn import and the other is a broad torch 2.10 incompatibility report." }, { "left": "issue:29942", "right": "issue:32090", "accept": false, - "reason": "Flash Attention test failures and a Trainer GPU broadcast TypeError are different failure modes and code paths." + "reason": "One is failing Flash Attention 2 tests; the other is a Trainer broadcast TypeError on NoneType. Same area, different failure." }, { "left": "issue:43335", "right": "issue:43653", "accept": false, - "reason": "SwitchTransformers config construction and BigBird tokenizer special-token registration are unrelated subsystems." + "reason": "One is a SwitchTransformers config bug creating sparse layers incorrectly; the other is a BigBirdTokenizer special-token registration issue." }, { "left": "issue:43550", "right": "issue:44112", "accept": false, - "reason": "Bamba SDPA/torch.compile and GraniteSpeech CI stale device override are distinct issues." + "reason": "One is a Bamba torch.compile/SDPA failure; the other is a stale CI test in GraniteSpeech. Not the same code-path bug." }, { "left": "issue:28282", "right": "issue:33357", "accept": false, - "reason": "Missing PyTorch in the environment and a MacOS bus error on CLIP loading are not the same underlying problem." - }, - { - "left": "issue:43611", - "right": "issue:43931", - "accept": false, - "reason": "Base-model-prefix loading regression and Qwen3-VL weight shape mismatch are different model-loading failures." + "reason": "One is a missing PyTorch ImportError; the other is a MacOS bus error loading a CLIP model. Different root causes." }, { "left": "issue:44448", "right": "issue:44743", "accept": false, - "reason": "Pegasus v4/v5 output differences and Qwen3 modular recurrent-state reset are different bugs." + "reason": "Both affect model behavior, but one is a Pegasus output regression and the other is recurrent state resetting in modular_qwen3_5." }, { - "left": "issue:45759", - "right": "issue:46512", + "left": "issue:43611", + "right": "issue:43931", "accept": false, - "reason": "AutoModelForCausalLM config unwrapping and stale dtype state are separate config-handling issues." + "reason": "Both are model-loading failures, but one is about base_model_prefix handling and the other is a weight-shape mismatch for Qwen3-VL." }, { "left": "issue:44514", "right": "issue:44521", "accept": false, - "reason": "A batched padding=False crash and all-zero assistant masks are related to chat templates but are not the same defect." + "reason": "Both touch apply_chat_template for multimodal inputs, but one crashes on batched padding=False and the other returns all-zero assistant masks." + }, + { + "left": "issue:45759", + "right": "issue:46512", + "accept": false, + "reason": "Both concern from_config/model initialization, but one is about unwrapping nested text_config and the other is stale config.dtype after dtype-aware construction." }, { "left": "issue:43335", "right": "issue:43504", "accept": false, - "reason": "A sparse-layer construction bug and a legacy-field pretrained load failure are not the same issue." + "reason": "Both involve model/config setup, but one miscreates sparse layers and the other fails loading a BEiT preset due to a legacy field." }, { "left": "issue:44855", "right": "issue:45083", "accept": false, - "reason": "Python 3.13 torch.jit parsing failure and qwen3_omni_moe helper behavior are unrelated." + "reason": "One is a Python 3.13 import/torch.jit parsing failure; the other is an unexpected behavior bug in a feature-extraction helper." }, { "left": "issue:46431", "right": "issue:46566", "accept": false, - "reason": "Missing chat template error and DiffusionGemma meta-device loading are separate problems." + "reason": "One is a missing chat template error in a processor; the other is weights \u10d3\u10d0\u10e0\u10e9ing on meta with device_map=auto. Different initialization paths." }, { "left": "issue:43278", "right": "issue:43825", "accept": false, - "reason": "Embedding dtype mismatch in evaluate and pipeline translation-task messaging are unrelated." + "reason": "One is a dtype mismatch between training and evaluate; the other is a pipeline error message about unsupported translation tasks." }, { "left": "issue:29127", "right": "issue:32090", "accept": false, - "reason": "LayoutLMv3 box-content error messaging and Trainer NoneType broadcast are different bugs." + "reason": "One is about clarifying an error message in layoutlmv3; the other is a runtime NoneType broadcast failure." }, { "left": "issue:37428", "right": "issue:44351", "accept": false, - "reason": "Both are import errors, but they involve different missing symbols and different regressions." - }, - { - "left": "issue:43720", - "right": "issue:44112", - "accept": false, - "reason": "BitNet packed-weights loading and GraniteSpeech device-override CI failure are distinct." + "reason": "Both are import errors, but they involve different symbols and regressions: flash-attn support vs HybridCache availability." }, { "left": "issue:45290", "right": "issue:46326", "accept": false, - "reason": "Assistant tool-call crash in apply_chat_template and continue_final_message behavior for mistral_common tokenizers are not the same defect." + "reason": "Both involve chat template/tokenization, but one crashes on assistant tool-call messages and the other is about continue_final_message for mistral_common tokenizers." }, { - "left": "issue:43335", - "right": "issue:45083", + "left": "issue:43720", + "right": "issue:44112", "accept": false, - "reason": "SwitchTransformers sparsity config and qwen3_omni_moe feature-length calculation are unrelated." + "reason": "One is BitNet accelerate-loading of packed weights; the other is a stale CI device-override test. No shared underlying bug." }, { - "left": "issue:45588", - "right": "issue:45859", + "left": "issue:43335", + "right": "issue:45083", "accept": false, - "reason": "Flash-attention s_aux=None and CPU offload incompatibility in Qwen3_5MoeTextRotaryEmbedding are different code paths." + "reason": "One is a SwitchTransformers config edge case; the other is a qwen3_omni_moe helper-length bug." }, { "left": "issue:43335", "right": "issue:44112", "accept": false, - "reason": "A model config bug and a CI device-override test failure do not describe the same issue." + "reason": "One is a model construction bug and the other is a CI test failure; they do not share the same code-path." }, { "left": "issue:45127", "right": "issue:45245", "accept": false, - "reason": "LoRA merge with extended vocabulary and a category-count overflow are unrelated." + "reason": "One is LoRA merge collapse with tied embeddings; the other is a categorical cardinality limit error. Different problems." }, { - "left": "issue:45200", - "right": "issue:45588", + "left": "issue:45588", + "right": "issue:45859", "accept": false, - "reason": "Gemma 4 token-type defaults and sink-less flash-attention crash are not the same bug." + "reason": "Both are MoE/flash-related, but one is an AttributeError on s_aux=None and the other is a CPU-offload incompatibility in rotary embedding." }, { "left": "issue:43335", "right": "issue:45593", "accept": false, - "reason": "SwitchTransformers sparse-layer construction and D-FINE auxiliary-loss behavior are unrelated." + "reason": "One is a SwitchTransformers config bug; the other is a D-FINE loss-path issue. Same general domain only." }, { - "left": "issue:43819", - "right": "issue:44493", + "left": "issue:45200", + "right": "issue:45588", "accept": false, - "reason": "DAC from_latents mismatch and unexpected position-id key behavior are different model implementations and failure signatures." + "reason": "One is about mm_token_type_ids defaults in Gemma 4; the other is a flash_attention sink-less model crash. Different mechanisms." }, { "left": "issue:34689", "right": "issue:35141", "accept": false, - "reason": "Llama 3.2 Vision loading regression and embedding reinitialization after resizing token embeddings are separate bugs." + "reason": "One is a model-loading breakage for Llama 3.2 Vision; the other is output embeddings being reinitialized after resizing. Related to loading, not the same bug." }, { "left": "issue:39692", "right": "issue:43526", "accept": false, - "reason": "SigLIP2 docs/example errors and BeitImageProcessorFast reduce_labels are unrelated." - }, - { - "left": "issue:26413", - "right": "issue:31501", - "accept": false, - "reason": "Checkpoint resumption epoch_iterator failure and FineWeb SLM training not starting are different training issues." - }, - { - "left": "issue:45310", - "right": "issue:45478", - "accept": true, - "reason": "Exact same title and very similar metadata indicate the same Qwen3.5 MoE from_pretrained regression." + "reason": "One is documentation example errors; the other is a BeitImageProcessorFast reduce_labels bug. Different artifact types and failures." }, { - "left": "issue:42490", - "right": "issue:43854", + "left": "issue:43819", + "right": "issue:44493", "accept": false, - "reason": "Trailing-slash local-path import warning and GLM-4.7 Flash model loading failure are not the same problem." + "reason": "One is DAC.from_latents disagreeing with forward due to missing STE; the other is a position-id key warning across many models. Not the same issue." } - ] - }, - "evaluator_result": { - "accept": false, - "feedback": "The pairwise verdicts are mostly conservative, and the 45310/45478 match is supported by the identical titles. However, the canonical_issue_reason overstates the evidence by claiming the issues have the same \"exact title/body theme\"\u2014the packet does not provide body text, so that part is not grounded. Please keep the duplicate rationale limited to what is visible in the packet (identical titles and very similar metadata)." - }, - "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 }, @@ -9054,13 +8969,15 @@ "source": "llm", "cache_hit": false, "model": "gpt-5.4-mini?service_tier=flex", - "cluster_id": "cluster-10000-472", + "cluster_id": "cluster-10000-473", "nodes": [ "issue:10000", "issue:15354", + "issue:26413", "issue:27957", "issue:28218", "issue:29942", + "issue:31501", "issue:32090", "issue:33357", "issue:34689", @@ -9077,8 +8994,7 @@ "issue:43575", "issue:43606", "issue:43701", - "issue:43746", - "issue:43927", + "issue:43854", "issue:44206", "issue:44360", "issue:44448", @@ -9092,24 +9008,27 @@ "issue:45083", "issue:45229", "issue:45290", + "issue:45310", "issue:45381", "issue:45459", + "issue:45478", "issue:45588", - "issue:45698", "issue:45823", "issue:45920", "issue:46192", - "issue:46209", "issue:46234", "issue:46399", "issue:46655" ], "soft_pairs": [ - "issue:43606|issue:44857", + "issue:26413|issue:31501", + "issue:45310|issue:45478", + "issue:42490|issue:43854", "issue:45459|issue:46234", - "issue:44206|issue:45823", - "issue:39692|issue:44493", + "issue:43606|issue:44857", "issue:44448|issue:44779", + "issue:39692|issue:44493", + "issue:44206|issue:45823", "issue:28218|issue:32090", "issue:44857|issue:45072", "issue:44360|issue:44561", @@ -9123,220 +9042,227 @@ "issue:45290|issue:45381", "issue:44448|issue:44977", "issue:27957|issue:34689", - "issue:45920|issue:46192", "issue:45290|issue:46655", + "issue:45920|issue:46192", "issue:44977|issue:45229", "issue:43278|issue:43329", - "issue:44855|issue:46399", "issue:10000|issue:15354", + "issue:44855|issue:46399", "issue:43278|issue:43381", - "issue:44857|issue:45588", - "issue:43746|issue:46209", - "issue:43927|issue:45698" + "issue:44857|issue:45588" ], - "prepared_review_unit_hash": "c67f5e00d1ebb2b68aa97ea1c156c9a45c1f8307408cbc44d371036aa6c5c6d8", + "prepared_review_unit_hash": "1dc8f5b4eab5b64fa12cddebdcb2feda9c1f651eb92bf86fc3e02c0db47bd615", "component_budget": { - "node_count": 472, - "item_count": 472, - "soft_pair_count": 1732, - "serialized_chars": 898654, - "estimated_input_tokens": 224664, - "estimated_eval_tokens": 449584 + "node_count": 473, + "item_count": 473, + "soft_pair_count": 1738, + "serialized_chars": 901437, + "estimated_input_tokens": 225360, + "estimated_eval_tokens": 450976 }, "budget": { - "node_count": 47, - "item_count": 47, - "soft_pair_count": 28, - "serialized_chars": 35607, - "estimated_input_tokens": 8902, - "estimated_eval_tokens": 18060 + "node_count": 48, + "item_count": 48, + "soft_pair_count": 29, + "serialized_chars": 36410, + "estimated_input_tokens": 9103, + "estimated_eval_tokens": 18462 }, "overflow_policy": "truncate_then_skip", "trimmed": false, "aggressively_trimmed": false, "split": true, "analyst_result": { - "summary": "These items are mostly unrelated single-bug issues; none of the soft pairs look like true duplicates. The closest canonical anchor is the broad tokenizer regression report, not because it matches the others, but because it is the most general and representative issue in the set.", - "confidence": 0.82, - "canonical_issue_reason": "issue:45920 is the broadest and most central report here: it describes a v5 tokenizer-ID regression affecting multiple model families, making it the best representative anchor among otherwise unrelated issues.", + "summary": "The cluster is mostly a set of unrelated issue reports that only share broad subsystem keywords; one soft pair is a clear duplicate, but the rest are distinct bugs/regressions and should stay separate.", + "confidence": 0.95, + "canonical_issue_reason": "No single issue is a valid canonical for the cluster overall because the items describe different failures across tokenization, loading, training, multimodal processing, and dtype/runtime paths.", "canonical_pr_reason": null, - "best_issue_reason": "issue:45920 has the clearest cross-model scope and user impact, so it is the best issue to keep as the cluster anchor even though the rest of the set does not collapse into one duplicate theme.", + "best_issue_reason": "No issue cleanly represents the whole cluster; the only near-duplicate pair is the duplicated Qwen3.5 MoE from_pretrained regression, which is too narrow to serve as a global representative.", "best_pr_reason": null, "soft_edge_verdicts": [ { - "left": "issue:43606", - "right": "issue:44857", + "left": "issue:26413", + "right": "issue:31501", + "accept": false, + "reason": "Different problems: checkpoint resume/epoch iterator key handling vs training not starting on FineWeb. Same training area, but not the same bug." + }, + { + "left": "issue:45310", + "right": "issue:45478", + "accept": true, + "reason": "These appear to be the same Qwen3.5 MoE from_pretrained regression, with the same title and matching failure description." + }, + { + "left": "issue:42490", + "right": "issue:43854", "accept": false, - "reason": "Both are runtime crashes, but one is CPU offload/device placement and the other is AMP/CUDA loss behavior; different code paths and fixes." + "reason": "Different loading bugs: local path trailing-slash warning vs incorrect loading of a specific GLM-4.7 model in tests." }, { "left": "issue:45459", "right": "issue:46234", "accept": false, - "reason": "Tokenizer protobuf error masking and a missing import_utils availability flag are separate import/system issues, not the same bug." + "reason": "Unrelated import-path issues: protobuf decode error masking tokenizer errors vs a missing import_utils helper function." }, { - "left": "issue:44206", - "right": "issue:45823", + "left": "issue:43606", + "right": "issue:44857", + "accept": false, + "reason": "Both involve runtime failures, but the concrete bugs differ: CPU offload device mismatch for bark-small vs float16 AMP crash in LwDetrImageLoss." + }, + { + "left": "issue:44448", + "right": "issue:44779", "accept": false, - "reason": "Unsupported feature-extractor argument vs device_map forward mismatch are unrelated failures." + "reason": "Different regressions and code paths: Pegasus generation mismatch across v4/v5 vs DeepSeek tokenizer output changes in v5." }, { "left": "issue:39692", "right": "issue:44493", "accept": false, - "reason": "Documentation example mistakes and a model position-id regression are different problem types with different remediation." + "reason": "Documentation example errors for SigLIP2 are not the same as the unexpected position-id key issue affecting models." }, { - "left": "issue:44448", - "right": "issue:44779", + "left": "issue:44206", + "right": "issue:45823", "accept": false, - "reason": "Both are v5 regressions, but one is Pegasus generation output drift and the other is DeepSeek tokenizer mis-tokenization; not the same underlying change." + "reason": "Different failures: unsupported `center` arg in LasrFeatureExtractor vs Gemma4 PLE device mismatch during forward with `device_map=auto`." }, { "left": "issue:28218", "right": "issue:32090", "accept": false, - "reason": "Tokenizer spacing behavior and Trainer broadcast type handling are unrelated." + "reason": "Tokenizer spacing bug and Trainer distributed broadcast NoneType error are unrelated." }, { "left": "issue:44857", "right": "issue:45072", "accept": false, - "reason": "These are both dtype/precision-related at a high level, but they affect different models and different execution paths." + "reason": "Both are dtype-related crashes, but they hit different models and code paths: AMP loss computation vs bfloat16 inference mismatches." }, { "left": "issue:44360", "right": "issue:44561", "accept": false, - "reason": "A missing ReLU in DSA indexer is unrelated to removal of is_torch_fx_available breaking trust_remote_code models." + "reason": "DSA indexer missing ReLU is unrelated to the v5 removal of `is_torch_fx_available` breaking trust_remote_code models." }, { "left": "issue:43701", "right": "issue:45083", "accept": false, - "reason": "Checkpoint key mismatch and a qwen3_omni_moe helper-function bug are different subsystems and symptoms." + "reason": "Checkpoint key mismatch and multimodal feature-length helper behavior are different bugs." }, { "left": "issue:38175", "right": "issue:44493", "accept": false, - "reason": "Zero probabilities in SigLIP2 and unexpected position-id keys are not the same failure mode." + "reason": "Unexpected zero probabilities in SigLIP2 are not the same as the unexpected position-id key regression." }, { "left": "issue:32090", "right": "issue:33357", "accept": false, - "reason": "Trainer broadcast NoneType error and MacOS CLIP bus error are entirely different crashes." + "reason": "Trainer broadcast NoneType and MacOS bus error in CLIP loading are separate failures." }, { "left": "issue:42490", "right": "issue:43575", "accept": false, - "reason": "Trailing-slash model-path warning and tensor-parallel OOM are unrelated." + "reason": "Local-path import warning and tensor-parallel OOM for Qwen2-57B are unrelated." }, { "left": "issue:40444", "right": "issue:43526", "accept": false, - "reason": "Multi-image IterableDataset finetuning and BeitImageProcessorFast label reduction are separate vision-data bugs." + "reason": "Fine-tuning a multimodal iterable dataset and BEiT label reduction are different processor/training bugs." }, { "left": "issue:29942", "right": "issue:35141", "accept": false, - "reason": "Flash Attention 2 test failures and output-embedding reinitialization are unrelated model initialization vs attention issues." + "reason": "Flash Attention 2 test failures do not match embedding reinitialization after resize in post_init." }, { "left": "issue:43550", "right": "issue:45083", "accept": false, - "reason": "torch.compile/SDPA failure for Bamba and a feature-extraction length helper bug are different code paths." + "reason": "torch.compile/SDPA failure in Bamba is unrelated to qwen3_omni_moe feature-length helper behavior." }, { "left": "issue:45290", "right": "issue:45381", "accept": false, - "reason": "Chat-template tool-call handling and qwen2.5-vl video vision_position_ids are separate multimodal bugs." + "reason": "Chat-template tool-call crash and wrong video vision_position_ids are both multimodal, but they are distinct code paths and bugs." }, { "left": "issue:44448", "right": "issue:44977", "accept": false, - "reason": "Pegasus output differences and Qwen3.5 flash-attention generation failure are unrelated model-specific regressions." + "reason": "Pegasus v4/v5 output difference and Qwen3.5 flash-attention generation failure are different model-specific regressions." }, { "left": "issue:27957", "right": "issue:34689", "accept": false, - "reason": "Flash Attention 2 support for XLMRoberta and Llama 3.2 Vision load failure are different issues." + "reason": "Flash Attention 2 support issue and Llama 3.2 Vision model loading breakage are not the same problem." }, { - "left": "issue:45920", - "right": "issue:46192", + "left": "issue:45290", + "right": "issue:46655", "accept": false, - "reason": "Tokenizer-ID regression across models and silent return_tensors handling are distinct API bugs." + "reason": "Tool-call chat templating crash and SmolVLM video padding calculation bug affect different layers of the stack." }, { - "left": "issue:45290", - "right": "issue:46655", + "left": "issue:45920", + "right": "issue:46192", "accept": false, - "reason": "Chat templating with tool calls and SmolVLM video padding calculation are different multimodal processing paths." + "reason": "Wrong token IDs across many models is a tokenizer regression; silently ignored `return_tensors` is an argument-handling issue." }, { "left": "issue:44977", "right": "issue:45229", "accept": false, - "reason": "Flash-attention generation bug in Qwen3.5 and Gemma4 multi-GPU OOM are not the same root cause." + "reason": "Flash-attention generation bug in Qwen3.5 and Gemma4 multi-GPU CUDA OOM are different failure modes." }, { "left": "issue:43278", "right": "issue:43329", "accept": false, - "reason": "Evaluation dtype drift and undefined video-branch variables are unrelated bugs." + "reason": "Embedding dtype drift during evaluate is unrelated to the undefined-variable bug in multimodal token counting." }, { - "left": "issue:44855", - "right": "issue:46399", + "left": "issue:10000", + "right": "issue:15354", "accept": false, - "reason": "Python 3.13 torch.jit parsing failure and a missing decorator on rotary embedding registration are both decorator-adjacent, but they are different concrete failures." + "reason": "These are entirely unrelated model/export issues with different architectures and failure mechanisms." }, { - "left": "issue:10000", - "right": "issue:15354", + "left": "issue:44855", + "right": "issue:46399", "accept": false, - "reason": "German DistilBertModel issue and torch.jit GeneratorExp export failure are unrelated legacy JIT problems." + "reason": "Python 3.13 torch.jit parsing/IndentationError for DebertaV2 is unrelated to the Qwen3.5 missing-decorator registration crash." }, { "left": "issue:43278", "right": "issue:43381", "accept": false, - "reason": "BF16-to-FP32 dtype change in eval and gradient checkpointing being disallowed in eval mode are different concerns." + "reason": "Eval-mode gradient-checkpointing restriction is a different bug from embedding dtype changing from BF16 to FP32." }, { "left": "issue:44857", "right": "issue:45588", "accept": false, - "reason": "AMP/CUDA loss crash and sink-less flash_attention s_aux=None AttributeError are different crash sites and root causes." - }, - { - "left": "issue:43746", - "right": "issue:46209", - "accept": false, - "reason": "PEFT checkpoint loading and FP8 inference failure are unrelated." - }, - { - "left": "issue:43927", - "right": "issue:45698", - "accept": false, - "reason": "Losing custom token IDs on save/load and loading the wrong custom module after save_pretrained are both persistence bugs, but they are not the same underlying issue." + "reason": "LwDetrImageLoss AMP crash and flash_attention.py s_aux=None AttributeError are separate code-path failures." } ] }, - "evaluator_result": null, - "evaluator_used": false, + "evaluator_result": { + "accept": true, + "feedback": "The summary is grounded in the packet: the cluster is heterogeneous, and the only clearly duplicate soft pair is issue:45310/issue:45478, which the analyst identified conservatively. The remaining soft-edge verdicts are mostly cautious false negatives with reasons tied to distinct bug descriptions, and no claim appears overstated." + }, + "evaluator_used": true, "retried": false, - "accepted_nontrivial_soft_edge": false, + "accepted_nontrivial_soft_edge": true, "error_kind": null, "error_message": null }, @@ -9351,7 +9277,7 @@ "source": "llm", "cache_hit": false, "model": "gpt-5.4-mini?service_tier=flex", - "cluster_id": "cluster-10000-472", + "cluster_id": "cluster-10000-473", "nodes": [ "issue:25251", "issue:29127", @@ -9366,7 +9292,6 @@ "issue:42490", "issue:42915", "issue:42994", - "issue:43065", "issue:43278", "issue:43381", "issue:43412", @@ -9374,14 +9299,14 @@ "issue:43611", "issue:43653", "issue:43701", + "issue:43746", "issue:43873", + "issue:43927", "issue:43976", "issue:44186", - "issue:44206", "issue:44361", "issue:44448", "issue:44466", - "issue:44479", "issue:44561", "issue:44568", "issue:44743", @@ -9392,33 +9317,36 @@ "issue:45229", "issue:45372", "issue:45406", - "issue:45412", "issue:45440", "issue:45491", "issue:45563", "issue:45588", + "issue:45698", "issue:45803", "issue:45823", - "issue:46121" + "issue:46121", + "issue:46209" ], "soft_pairs": [ + "issue:43746|issue:46209", + "issue:43927|issue:45698", "issue:37428|issue:42175", "issue:44561|issue:44779", "issue:43873|issue:45229", "issue:32090|issue:33290", "issue:25251|issue:32090", "issue:41720|issue:45229", - "issue:44448|issue:44466", "issue:43976|issue:45440", - "issue:45588|issue:45823", + "issue:44448|issue:44466", "issue:42490|issue:43873", "issue:44568|issue:45491", - "issue:44977|issue:45563", - "issue:45198|issue:45803", + "issue:45588|issue:45823", "issue:43278|issue:43575", - "issue:45372|issue:45588", + "issue:44977|issue:45563", "issue:33357|issue:39692", + "issue:45372|issue:45588", "issue:29127|issue:34567", + "issue:45198|issue:45803", "issue:43412|issue:45588", "issue:43381|issue:43701", "issue:43653|issue:44743", @@ -9427,206 +9355,204 @@ "issue:44361|issue:45198", "issue:44186|issue:45198", "issue:45372|issue:45406", - "issue:42915|issue:42994", - "issue:44206|issue:44479", - "issue:43065|issue:45412" + "issue:42915|issue:42994" ], - "prepared_review_unit_hash": "f41414b1c7048605ad8c7d61c8530042edb397185ea6414b89836089499265d8", + "prepared_review_unit_hash": "ed841b2f824d5aba87b4c8aa55643aed83cee2cc463c08eb81ba2637351b7330", "component_budget": { - "node_count": 472, - "item_count": 472, - "soft_pair_count": 1732, - "serialized_chars": 898654, - "estimated_input_tokens": 224664, - "estimated_eval_tokens": 449584 + "node_count": 473, + "item_count": 473, + "soft_pair_count": 1738, + "serialized_chars": 901437, + "estimated_input_tokens": 225360, + "estimated_eval_tokens": 450976 }, "budget": { "node_count": 47, "item_count": 47, "soft_pair_count": 28, - "serialized_chars": 35626, - "estimated_input_tokens": 8907, - "estimated_eval_tokens": 18070 + "serialized_chars": 35663, + "estimated_input_tokens": 8916, + "estimated_eval_tokens": 18088 }, "overflow_policy": "truncate_then_skip", "trimmed": false, "aggressively_trimmed": false, "split": true, "analyst_result": { - "summary": "This cluster is highly heterogeneous: the soft pairs span unrelated bugs across tokenizers, trainers, quantization, Gemma/Qwen/CLIP/Wav2Vec2, docs, and memory/runtime issues. None of the suggested soft links look like true duplicates, and there are no PRs in the set.", - "confidence": 0.97, - "canonical_issue_reason": "No canonical issue stands out because the items are not one duplicate set; they are a mix of unrelated issues from many subsystems and models.", + "summary": "This cluster is heterogeneous: the soft-similarity pairs mostly link unrelated bugs across models, tokenizers, loading, quantization, and training. I would not merge any of the candidate soft edges.", + "confidence": 0.93, + "canonical_issue_reason": "No true duplicate stands out; if a representative is needed, issue 43873 is the most central/open discussion-heavy item, but it is not a clear canonical duplicate for the cluster.", "canonical_pr_reason": null, - "best_issue_reason": "No issue is a good global representative for deduping: the cluster does not share a single underlying bug or code path.", + "best_issue_reason": "Issue 43873 is the broadest and most referenced open report in the set, so it is the best single representative issue, though the cluster itself is not a duplicate group.", "best_pr_reason": null, "soft_edge_verdicts": [ + { + "left": "issue:43746", + "right": "issue:46209", + "accept": false, + "reason": "Different failure modes: PEFT local checkpoint loading vs FP8/B200 inference on a specific GPU architecture." + }, + { + "left": "issue:43927", + "right": "issue:45698", + "accept": false, + "reason": "Both involve loading/saving, but one is custom token IDs being lost in config serialization, the other is the wrong custom module being loaded after save_pretrained." + }, { "left": "issue:37428", "right": "issue:42175", "accept": false, - "reason": "Different problems: flash-attention import breakage vs missing TensorFlow backend after torch-only install." + "reason": "Unrelated: a flash-attention import error versus TensorFlow missing from a torch-only extra install." }, { "left": "issue:44561", "right": "issue:44779", "accept": false, - "reason": "One is a v5 API removal breaking trust_remote_code; the other is a DeepSeek tokenizer regression." + "reason": "Different subsystems and symptoms: a removed utility breaking trust_remote_code imports versus tokenizer output regressions for DeepSeek." }, { "left": "issue:43873", "right": "issue:45229", "accept": false, - "reason": "Quantization/offloading behavior vs a Gemma4 multi-GPU OOM are different failure modes and likely different code paths." + "reason": "Both mention quantization/memory, but one is offloading not working and the other is multi-GPU CUDA OOM for Gemma4." }, { "left": "issue:32090", "right": "issue:33290", "accept": false, - "reason": "Trainer GPU broadcast TypeError is unrelated to Adafactor+DeepSpeed OOM." + "reason": "Trainer NoneType broadcast error and Adafactor/DeepSpeed OOM are not the same underlying bug." }, { "left": "issue:25251", "right": "issue:32090", "accept": false, - "reason": "Pipeline top_k nesting bug and Trainer NoneType broadcast error are unrelated." + "reason": "Pipeline top_k nesting change is unrelated to the Trainer broadcast TypeError." }, { "left": "issue:41720", "right": "issue:45229", "accept": false, - "reason": "Qwen3 device-mapping cuda assert vs Gemma4 multi-GPU OOM are separate model/runtime issues." - }, - { - "left": "issue:44448", - "right": "issue:44466", - "accept": false, - "reason": "Different bugs: generation output drift for Pegasus vs serialization of lm_head.weight." + "reason": "Different models and root causes: Qwen3 auto device map assert vs Gemma4 multi-GPU OOM." }, { "left": "issue:43976", "right": "issue:45440", "accept": false, - "reason": "Python version compatibility issue vs DeepSeekV3MoE divergence from remote implementation." + "reason": "Python version compatibility issue versus a DeepSeekV3 implementation divergence; not the same defect." }, { - "left": "issue:45588", - "right": "issue:45823", + "left": "issue:44448", + "right": "issue:44466", "accept": false, - "reason": "Flash-attention s_aux=None crash is not the same as Gemma4 PLE device mismatch." + "reason": "Both are v5 regressions, but one is generation/output mismatch and the other is serialization of lm_head.weight." }, { "left": "issue:42490", "right": "issue:43873", "accept": false, - "reason": "Trailing-slash import warning and quantization offloading malfunction are unrelated." + "reason": "Local-path slash import warning is unrelated to quantization offloading behavior." }, { "left": "issue:44568", "right": "issue:45491", "accept": false, - "reason": "Tokenizer special-token behavior and Gemma3 NaN embeddings are different bugs." - }, - { - "left": "issue:44977", - "right": "issue:45563", - "accept": false, - "reason": "Qwen3.5 flash-attention generation regression and a stale warning in paged generate() are not the same issue." + "reason": "Tokenizer special-token handling and Gemma3 sliding-window NaN embeddings are distinct bugs." }, { - "left": "issue:45198", - "right": "issue:45803", + "left": "issue:45588", + "right": "issue:45823", "accept": false, - "reason": "Wav2Vec2 save/tokenization failure vs FuyuBatchFeature's bare except are unrelated." + "reason": "Flash-attention s_aux None crash and Gemma4 PLE device mismatch are different code paths and fixes." }, { "left": "issue:43278", "right": "issue:43575", "accept": false, - "reason": "Training/eval dtype change is unrelated to Qwen2-57B-A14B-Instruct TP OOM." + "reason": "Embedding dtype regression in eval is unrelated to TP loading OOM for Qwen2-57B." }, { - "left": "issue:45372", - "right": "issue:45588", + "left": "issue:44977", + "right": "issue:45563", "accept": false, - "reason": "Gemma4 processor import failure from mistral_common is a different code path than the flash_attention s_aux crash." + "reason": "Flash-attention generation failure and a stale warning for num_return_sequences are not the same issue." }, { "left": "issue:33357", "right": "issue:39692", "accept": false, - "reason": "MacOS bus error in CLIP loading is unrelated to SigLIP2 documentation example errors." + "reason": "MacOS CLIP bus error is unrelated to SigLIP2 documentation example errors." + }, + { + "left": "issue:45372", + "right": "issue:45588", + "accept": false, + "reason": "Processor import failure due to mistral_common dependency is unrelated to the flash_attention.py s_aux=None crash." }, { "left": "issue:29127", "right": "issue:34567", "accept": false, - "reason": "LayoutLMv3 error messaging and TrainerState token counter updates are unrelated." + "reason": "LayoutLMv3 error-message clarity and TrainerState token counting are unrelated." + }, + { + "left": "issue:45198", + "right": "issue:45803", + "accept": false, + "reason": "wav2vec2 save/tokenization failure is unrelated to FuyuBatchFeature's bare except swallowing errors." }, { "left": "issue:43412", "right": "issue:45588", "accept": false, - "reason": "Executorch export failures are a different subsystem from flash_attention runtime crashes." + "reason": "Executorch export issues for gemma3n are a different path from flash-attention integration crashes." }, { "left": "issue:43381", "right": "issue:43701", "accept": false, - "reason": "Eval-mode gradient checkpointing restriction is unrelated to resume_from_checkpoint key mismatch." + "reason": "Gradient checkpointing in eval mode and resume_from_checkpoint key mismatch are separate training issues." }, { "left": "issue:43653", "right": "issue:44743", "accept": false, - "reason": "BigBirdTokenizer special-token registration and Qwen3.5 recurrent-state reset are different bugs." + "reason": "Tokenizer special-token registration and recurrent state reset in Qwen3.5 are unrelated." }, { "left": "issue:44871", "right": "issue:44977", "accept": false, - "reason": "Gemma-3 eos_token_id inconsistency is unrelated to Qwen3.5 flash-attention generation failure." + "reason": "Gemma eos_token_id config mismatch is unrelated to Qwen3.5 flash-attention generation failures." }, { "left": "issue:43611", "right": "issue:46121", "accept": false, - "reason": "base_model_prefix loading breakage and rope-param validation TypeError do not share a code path." + "reason": "Base model prefix loading regression and rope-params dict conversion TypeError are different loading/config bugs." }, { "left": "issue:44361", "right": "issue:45198", "accept": false, - "reason": "MLukeTokenizer AttributeError is unrelated to Wav2Vec2 save_pretrained/tokenization failure." + "reason": "Two tokenizer-related bugs, but for different models and different failure points." }, { "left": "issue:44186", "right": "issue:45198", "accept": false, - "reason": "LayoutLMv2 tokenizer crashes and Wav2Vec2 save/tokenization failures are different components." + "reason": "LayoutLMv2 padding/NER crashes are unrelated to wav2vec2 save_pretrained/tokenization failures." }, { "left": "issue:45372", "right": "issue:45406", "accept": false, - "reason": "Gemma4 processor loading import error and transformers serve missing _tokenizer are separate failures." + "reason": "Both involve Gemma4 processor loading, but one is an import dependency issue and the other is a missing _tokenizer attribute on serve; not the same bug." }, { "left": "issue:42915", "right": "issue:42994", "accept": false, - "reason": "Qwen3Moe FP8 config failure and quantized model saving failure are related only broadly by quantization, not the same bug." - }, - { - "left": "issue:44206", - "right": "issue:44479", - "accept": false, - "reason": "LasrFeatureExtractor center-arg regression and Qwen video-input regression are unrelated." - }, - { - "left": "issue:43065", - "right": "issue:45412", - "accept": false, - "reason": "Sam3PixelDecoder dummy Conv2d and RT-DETR memory retention are different model issues." + "reason": "Quantization is the common theme, but one is FP8 inference failure for Qwen3Moe and the other is quantized model saving failure." } ] }, @@ -9648,7 +9574,7 @@ "source": "llm", "cache_hit": false, "model": "gpt-5.4-mini?service_tier=flex", - "cluster_id": "cluster-10000-472", + "cluster_id": "cluster-10000-473", "nodes": [ "issue:27957", "issue:29127", @@ -9661,11 +9587,9 @@ "issue:42915", "issue:43065", "issue:43066", - "issue:43334", "issue:43381", "issue:43421", "issue:43526", - "issue:43531", "issue:43575", "issue:43653", "issue:43701", @@ -9674,11 +9598,13 @@ "issue:43931", "issue:43994", "issue:44186", + "issue:44206", "issue:44265", "issue:44279", "issue:44387", "issue:44442", "issue:44462", + "issue:44479", "issue:45072", "issue:45198", "issue:45245", @@ -9699,23 +9625,25 @@ "issue:46655" ], "soft_pairs": [ - "issue:45459|issue:45712", + "issue:44206|issue:44479", "issue:29942|issue:30064", + "issue:43065|issue:45412", "issue:43065|issue:43994", - "issue:43653|issue:44442", + "issue:45459|issue:45712", "issue:43756|issue:45440", + "issue:43653|issue:44442", "issue:45245|issue:45563", "issue:45372|issue:45459", "issue:45412|issue:46369", - "issue:44265|issue:45198", "issue:29127|issue:33357", "issue:43994|issue:45412", "issue:43066|issue:44462", "issue:27957|issue:29942", - "issue:43575|issue:43701", + "issue:44265|issue:45198", "issue:45440|issue:45593", - "issue:43421|issue:45491", + "issue:43575|issue:43701", "issue:29127|issue:35141", + "issue:43421|issue:45491", "issue:42490|issue:43381", "issue:43526|issue:45834", "issue:43575|issue:44387", @@ -9725,226 +9653,231 @@ "issue:43526|issue:44186", "issue:43994|issue:46369", "issue:43824|issue:44279", - "issue:45250|issue:46640", - "issue:46121|issue:46399", "issue:27957|issue:29127", - "issue:27957|issue:33290", - "issue:43334|issue:43531" + "issue:46121|issue:46399", + "issue:45250|issue:46640", + "issue:27957|issue:33290" ], - "prepared_review_unit_hash": "e709d0638f5de171d281bb17abc7dc23815c9f4e66a2f645e0fd9e68ec18bc3d", + "prepared_review_unit_hash": "e44ffe76a6a16014ecdf3475ecc5c6ba6e29ac9a5632a3c83b29fd7e6dde0428", "component_budget": { - "node_count": 472, - "item_count": 472, - "soft_pair_count": 1732, - "serialized_chars": 898654, - "estimated_input_tokens": 224664, - "estimated_eval_tokens": 449584 + "node_count": 473, + "item_count": 473, + "soft_pair_count": 1738, + "serialized_chars": 901437, + "estimated_input_tokens": 225360, + "estimated_eval_tokens": 450976 }, "budget": { "node_count": 47, "item_count": 47, - "soft_pair_count": 31, - "serialized_chars": 36204, - "estimated_input_tokens": 9051, - "estimated_eval_tokens": 18358 + "soft_pair_count": 32, + "serialized_chars": 36536, + "estimated_input_tokens": 9134, + "estimated_eval_tokens": 18524 }, "overflow_policy": "truncate_then_skip", "trimmed": false, "aggressively_trimmed": false, "split": true, "analyst_result": { - "summary": "This cluster is highly heterogeneous: most items are unrelated model/loading/tokenizer/processor bugs, with only a loose Flash Attention 2 subtheme. None of the soft edges look like true duplicates, so I would not merge any of them.", - "confidence": 0.93, - "canonical_issue_reason": null, + "summary": "This cluster is heterogeneous: most items are distinct model-specific bugs, regressions, or feature requests. I found no soft-edge pair that looks like the same underlying bug/change, so none should be merged as duplicates.", + "confidence": 0.91, + "canonical_issue_reason": "No single canonical issue cleanly represents this mixed cluster. If one representative is needed, issue:44479 is the broadest concrete regression report, but it is not a duplicate of the others.", "canonical_pr_reason": null, - "best_issue_reason": "issue:29942 is the best single anchor only for the Flash Attention 2 subtheme because it is concrete and regression-oriented, but it does not represent the full cluster and should not be treated as a cluster-wide canonical.", + "best_issue_reason": "issue:44479 is the best overall representative because it describes a concrete runtime regression with a clear failure mode and multiple affected model families; however, the cluster is too diverse to treat it as a true canonical for all items.", "best_pr_reason": null, "soft_edge_verdicts": [ { - "left": "issue:45459", - "right": "issue:45712", + "left": "issue:44206", + "right": "issue:44479", "accept": false, - "reason": "Different bugs: tokenizer exception masking vs leftover dummy PT classes / repo checks." + "reason": "Different regressions in different code paths: LaserFeatureExtractor crash vs video-input regression for Qwen VL models." }, { "left": "issue:29942", "right": "issue:30064", "accept": false, - "reason": "Unrelated areas: Flash Attention tests vs void segmentation map processing." + "reason": "Flash Attention test failures and void segmentation map processing are unrelated subsystems and symptoms." + }, + { + "left": "issue:43065", + "right": "issue:45412", + "accept": false, + "reason": "Dummy Conv2d placeholder in Sam3PixelDecoder is unrelated to RT-DETR model memory leak on deletion." }, { "left": "issue:43065", "right": "issue:43994", "accept": false, - "reason": "Different model-specific failures; no shared code-path or fix." + "reason": "Sam3PixelDecoder dummy layer issue is not the same bug as Siglip2 producing nonsensical outputs." }, { - "left": "issue:43653", - "right": "issue:44442", + "left": "issue:45459", + "right": "issue:45712", "accept": false, - "reason": "Both concern tokenizers, but different root causes and models; not the same bug." + "reason": "Tokenizer error masking with missing protobuf is unrelated to leftover dummy classes in dummy_pt_objects.py." }, { "left": "issue:43756", "right": "issue:45440", "accept": false, - "reason": "Different model implementations and failure modes." + "reason": "Missing RoPE layers in Smollm3 and DeepseekV3MoE divergence are different model implementation problems." + }, + { + "left": "issue:43653", + "right": "issue:44442", + "accept": false, + "reason": "BigBirdTokenizer special-token registration and FastSpeech2ConformerTokenizer loading are different tokenizer bugs." }, { "left": "issue:45245", "right": "issue:45563", "accept": false, - "reason": "Category limit error vs stale warning in paged generate; unrelated." + "reason": "Category-cardinality runtime error is unrelated to a stale generate() warning." }, { "left": "issue:45372", "right": "issue:45459", "accept": false, - "reason": "ImportError for Gemma processor loading vs tokenizer protobuf error handling; unrelated." + "reason": "Missing ReasoningEffort import and protobuf-related tokenizer error handling are distinct failures." }, { "left": "issue:45412", "right": "issue:46369", "accept": false, - "reason": "Both are memory-related, but they affect different models and code paths." - }, - { - "left": "issue:44265", - "right": "issue:45198", - "accept": false, - "reason": "torch.export/compilable-check failure is unrelated to Wav2Vec2 save_pretrained/tokenization." + "reason": "RT-DETR garbage-collection memory leak is not the same as sam3-litetext memory overflow." }, { "left": "issue:29127", "right": "issue:33357", "accept": false, - "reason": "Error-message clarity for LayoutLMv3 boxes is unrelated to a MacOS CLIP bus error." + "reason": "LayoutLMv3 error-message clarity and a MacOS CLIP bus error are unrelated." }, { "left": "issue:43994", "right": "issue:45412", "accept": false, - "reason": "Different symptoms and subsystems: model output quality vs model GC/memory leak." + "reason": "Siglip2 wrong outputs and RT-DETR memory retention are different bugs." }, { "left": "issue:43066", "right": "issue:44462", "accept": false, - "reason": "Tokenizer decoder-type regression vs ignoring tokenizer.json; related to tokenizers, but not the same defect." + "reason": "Wrong tokenizer decoder type in v5 is not the same as AutoTokenizer ignoring tokenizer.json." }, { "left": "issue:27957", "right": "issue:29942", "accept": false, - "reason": "Both mention Flash Attention 2, but one is model support and the other is failing tests; not the same concrete bug." + "reason": "XLMRoberta Flash Attention 2 support and Flash Attention 2 test failures are not the same concrete bug." }, { - "left": "issue:43575", - "right": "issue:43701", + "left": "issue:44265", + "right": "issue:45198", "accept": false, - "reason": "Tensor-parallel OOM vs resume_from_checkpoint key mismatch are unrelated." + "reason": "torch.export failure via torch_compilable_check is unrelated to Wav2Vec2 save_pretrained/tokenization failure." }, { "left": "issue:45440", "right": "issue:45593", "accept": false, - "reason": "Different models and logic: DeepSeek MoE divergence vs D-FINE auxiliary-loss behavior." + "reason": "DeepseekV3MoE divergence and D-FINE auxiliary-loss behavior are unrelated model issues." }, { - "left": "issue:43421", - "right": "issue:45491", + "left": "issue:43575", + "right": "issue:43701", "accept": false, - "reason": "Special-token post-processor updates vs Gemma3 NaN embeddings are unrelated." + "reason": "Tensor-parallel OOM when loading Qwen2-57B and resume_from_checkpoint key mismatch are different problems." }, { "left": "issue:29127", "right": "issue:35141", "accept": false, - "reason": "LayoutLMv3 error-message issue is unrelated to embedding reinit on resize." + "reason": "LayoutLMv3 error clarity and output embedding reinitialization after resizing are unrelated." + }, + { + "left": "issue:43421", + "right": "issue:45491", + "accept": false, + "reason": "Runtime special-token post-processor update bug and Gemma3 sliding-window NaNs are different issues." }, { "left": "issue:42490", "right": "issue:43381", "accept": false, - "reason": "Local-path import warning and gradient-checkpointing-in-eval are unrelated." + "reason": "Trailing-slash import warning and gradient-checkpointing-in-eval-mode are unrelated." }, { "left": "issue:43526", "right": "issue:45834", "accept": false, - "reason": "BeitImageProcessorFast label reduction bug vs Kosmos2.5 OCR index error are different." + "reason": "BeitImageProcessorFast reduce_labels bug and Kosmos2.5 OCR index error are different image-processing failures." }, { "left": "issue:43575", "right": "issue:44387", "accept": false, - "reason": "OOM from tensor-parallel loading vs int4 quantization reserved-memory regression are not the same bug." + "reason": "Qwen2 tensor-parallel OOM and int4 quantization reserved-memory regression are not the same code-path problem." }, { "left": "issue:44442", "right": "issue:45072", "accept": false, - "reason": "FastSpeech2Conformer tokenizer loading vs bf16 dtype mismatches are unrelated." + "reason": "Tokenizer loading failure and bfloat16 dtype mismatch are unrelated." }, { "left": "issue:42915", "right": "issue:43931", "accept": false, - "reason": "Qwen3Moe FP8 config failure vs Qwen3-VL weight-shape mismatch are different issues." + "reason": "FineGrainedFP8Config failure for Qwen3Moe and Qwen3-VL weight-shape mismatch are distinct loading bugs." }, { "left": "issue:43526", "right": "issue:46655", "accept": false, - "reason": "ImageProcessor label handling vs SmolVLM video padding calculation are unrelated." + "reason": "reduce_labels returning one label and SmolVLM video padding calculation are unrelated." }, { "left": "issue:43526", "right": "issue:44186", "accept": false, - "reason": "Different processors/tokenizers and different failure modes." + "reason": "BeitImageProcessorFast reduce_labels bug and LayoutLMv2Tokenizer NER padding/truncation crash are different tokenizer/processor issues." }, { "left": "issue:43994", "right": "issue:46369", "accept": false, - "reason": "Different models and symptoms; no shared root cause." + "reason": "Siglip2 bad outputs and sam3-litetext memory overflow are unrelated." }, { "left": "issue:43824", "right": "issue:44279", "accept": false, - "reason": "Missing import for a model class vs generic dependency issue are not duplicates." + "reason": "Missing Qwen2_5_VL import and a generic dependency issue are not the same concrete failure." }, { - "left": "issue:45250", - "right": "issue:46640", + "left": "issue:27957", + "right": "issue:29127", "accept": false, - "reason": "Both are Flash Attention 2 named issues, but the reports are too generic to establish the same concrete bug." + "reason": "Flash Attention 2 support for XLMRoberta and LayoutLMv3 error messaging are unrelated." }, { "left": "issue:46121", "right": "issue:46399", "accept": false, - "reason": "RoPE validation TypeError vs apply_rotary_pos_emb registration ValueError are different failure points." + "reason": "Rope-parameter JSON list handling and missing decorator for apply_rotary_pos_emb registration are different bugs." }, { - "left": "issue:27957", - "right": "issue:29127", + "left": "issue:45250", + "right": "issue:46640", "accept": false, - "reason": "Flash Attention 2 support issue vs LayoutLMv3 box-validation message issue." + "reason": "Both mention Flash Attention 2, but one is a vague title and the other is a distinct issue; no shared concrete bug is evident." }, { "left": "issue:27957", "right": "issue:33290", "accept": false, - "reason": "Flash Attention 2 model support vs Adafactor/DeepSpeed OOM are unrelated." - }, - { - "left": "issue:43334", - "right": "issue:43531", - "accept": false, - "reason": "Qwen3-VL pad_token_id loading error vs Qwen3-MoE sliding_window issue are different model bugs." + "reason": "Flash Attention 2 model support and Adafactor-in-Deepspeed OOM are unrelated." } ] }, @@ -9966,7 +9899,7 @@ "source": "llm", "cache_hit": false, "model": "gpt-5.4-mini?service_tier=flex", - "cluster_id": "cluster-10000-472", + "cluster_id": "cluster-10000-473", "nodes": [ "issue:33290", "issue:34567", @@ -9977,19 +9910,19 @@ "issue:43232", "issue:43262", "issue:43316", + "issue:43334", "issue:43381", "issue:43388", "issue:43493", "issue:43525", + "issue:43531", "issue:43611", "issue:43646", "issue:43720", "issue:43723", - "issue:43749", "issue:43761", "issue:43827", "issue:43873", - "issue:43931", "issue:44062", "issue:44077", "issue:44079", @@ -10005,7 +9938,6 @@ "issue:44977", "issue:44987", "issue:45003", - "issue:45216", "issue:45245", "issue:45440", "issue:45479", @@ -10018,13 +9950,14 @@ "issue:46234" ], "soft_pairs": [ - "issue:43761|issue:45859", + "issue:43334|issue:43531", "issue:33290|issue:34689", + "issue:43761|issue:45859", "issue:33290|issue:35141", "issue:45588|issue:45600", "issue:33290|issue:34567", - "issue:44977|issue:45245", "issue:43262|issue:44206", + "issue:44977|issue:45245", "issue:43381|issue:43873", "issue:44464|issue:45491", "issue:43646|issue:44077", @@ -10034,222 +9967,214 @@ "issue:34689|issue:43611", "issue:45003|issue:45440", "issue:43720|issue:44442", - "issue:42175|issue:46234", - "issue:43316|issue:43525", "issue:43723|issue:45479", - "issue:44062|issue:45907", + "issue:43316|issue:43525", + "issue:42175|issue:46234", "issue:43493|issue:43827", "issue:44077|issue:44479", "issue:43761|issue:44079", + "issue:44062|issue:45907", "issue:43873|issue:44871", "issue:43761|issue:44188", "issue:43232|issue:43388", - "issue:43749|issue:43931", - "issue:44987|issue:46178", - "issue:43931|issue:45216" + "issue:44987|issue:46178" ], - "prepared_review_unit_hash": "259deac3ef6cca0b74ff802241f86a98c173aa038c5f3a5a8fe3855becde8af7", + "prepared_review_unit_hash": "88569fcddcd84f169ba15ec080bf52cd28d2530a6188e5946d6739f4a7f4e804", "component_budget": { - "node_count": 472, - "item_count": 472, - "soft_pair_count": 1732, - "serialized_chars": 898654, - "estimated_input_tokens": 224664, - "estimated_eval_tokens": 449584 + "node_count": 473, + "item_count": 473, + "soft_pair_count": 1738, + "serialized_chars": 901437, + "estimated_input_tokens": 225360, + "estimated_eval_tokens": 450976 }, "budget": { - "node_count": 48, - "item_count": 48, - "soft_pair_count": 29, - "serialized_chars": 36563, - "estimated_input_tokens": 9141, - "estimated_eval_tokens": 18538 + "node_count": 47, + "item_count": 47, + "soft_pair_count": 28, + "serialized_chars": 35703, + "estimated_input_tokens": 8926, + "estimated_eval_tokens": 18108 }, "overflow_policy": "truncate_then_skip", "trimmed": false, "aggressively_trimmed": false, "split": true, "analyst_result": { - "summary": "This cluster is very heterogeneous: most items are unrelated regression/bug reports across loading, generation, audio, config, and docs. If forced to choose a single anchor, #43873 is the broadest and most active open issue, but there is no strong duplicate core here.", - "confidence": 0.28, - "canonical_issue_reason": "#43873 is the broadest, most active open issue in the set and is the closest thing to a central loading/offloading regression, though the cluster is not truly cohesive.", - "canonical_pr_reason": null, - "best_issue_reason": "#43873 has the most discussion/inbound references and the broadest scope, so it is the best representative issue to anchor triage for this heterogeneous cluster.", - "best_pr_reason": null, + "summary": "This cluster is heterogeneous and does not look like a true duplicate group; the soft-similarity pairs mostly connect unrelated bugs that only share broad subsystem keywords.", + "confidence": 0.94, + "canonical_issue_reason": "No clear canonical issue: the items span unrelated regressions and API bugs across model loading, generation, tokenizers, configs, audio, quantization, and docs, so there is no safe duplicate anchor.", + "canonical_pr_reason": "No PRs are present in this cluster.", + "best_issue_reason": "No single issue is globally representative enough to serve as a cluster-wide anchor; the closest matches are still only broad thematic overlaps, not the same underlying defect.", + "best_pr_reason": "No PRs are present in this cluster.", "soft_edge_verdicts": [ { - "left": "issue:43761", - "right": "issue:45859", + "left": "issue:43334", + "right": "issue:43531", "accept": false, - "reason": "Different bugs: CLIPVision hidden_states regression vs Qwen3_5Moe CPU-offload rotary embedding incompatibility." + "reason": "Both mention Qwen3-family configs, but one is a missing `pad_token_id` on Qwen3-VL and the other is a `sliding_window` issue in Qwen3-MoE; different bugs and code paths." }, { "left": "issue:33290", "right": "issue:34689", "accept": false, - "reason": "OOM with Adafactor/DeepSpeed is unrelated to Llama 3.2 Vision model loading failure." + "reason": "OOM during Adafactor+DeepSpeed training and Llama 3.2 Vision model loading failure are unrelated." + }, + { + "left": "issue:43761", + "right": "issue:45859", + "accept": false, + "reason": "CLIPVisionModel hidden-states regression and Qwen3_5Moe CPU offload incompatibility are different model families and failure modes." }, { "left": "issue:33290", "right": "issue:35141", "accept": false, - "reason": "Optimizer OOM and token-embedding reinit are distinct problems in different code paths." + "reason": "Training OOM with DeepSpeed is not the same as token-embedding resize causing output embedding reinitialization." }, { "left": "issue:45588", "right": "issue:45600", "accept": false, - "reason": "Flash-attention s_aux=None crash is unrelated to stale auto_mappings config references." + "reason": "Flash-attention `s_aux=None` crash and stale auto-mapping references to removed configs are separate issues." }, { "left": "issue:33290", "right": "issue:34567", "accept": false, - "reason": "Memory exhaustion during training is not the same as num_input_tokens_seen not updating." + "reason": "An optimizer OOM and `num_input_tokens_seen` not updating are unrelated trainer concerns." }, { - "left": "issue:44977", - "right": "issue:45245", + "left": "issue:43262", + "right": "issue:44206", "accept": false, - "reason": "Qwen3.5 flash-attention generation regression is unrelated to the categories cardinality error." + "reason": "Both are audio-related, but one is chat template sampling-rate defaulting and the other is an unsupported `center` argument in a feature extractor." }, { - "left": "issue:43262", - "right": "issue:44206", + "left": "issue:44977", + "right": "issue:45245", "accept": false, - "reason": "Both are audio-related, but one is a default sampling-rate mismatch and the other is an unsupported feature-extractor arg crash." + "reason": "Qwen3.5 flash-attention generation regression and a category-count runtime limit are unrelated." }, { "left": "issue:43381", "right": "issue:43873", "accept": false, - "reason": "Gradient checkpointing in eval mode and quantization offloading are different mechanisms and failure modes." + "reason": "Gradient checkpointing in eval mode and offloading with quantization do not describe the same defect." }, { "left": "issue:44464", "right": "issue:45491", "accept": false, - "reason": "Compiled chunked generation inconsistency is not the same as NaN embeddings from mixed-length sliding-window batching." + "reason": "Both affect generation, but chunked generation with compiled forward and Gemma3 NaN embeddings under sliding-window/all-padding windows are distinct bugs." }, { "left": "issue:43646", "right": "issue:44077", "accept": false, - "reason": "Custom model initialization breakage and patchtsmixer's post_init restriction are not the same concrete bug." + "reason": "A broad v5 custom-init breakage is not the same as patchtsmixer-specific `post_init` policy enforcement." }, { "left": "issue:44743", "right": "issue:44811", "accept": false, - "reason": "Qwen recurrent-state reset bug and Whisper batch_decode skip_special_tokens bug are unrelated." + "reason": "Recurrent-state reset in Qwen3.5-MoE and Whisper `batch_decode()` skipping tokens are unrelated." }, { "left": "issue:44442", "right": "issue:44898", "accept": false, - "reason": "Tokenizer loading failure and Perceiver interpolation failure are different model-specific issues." + "reason": "Tokenizer auto-load failure and Perceiver image-classification resize failure are different components and failure modes." }, { "left": "issue:42490", "right": "issue:44743", "accept": false, - "reason": "Local-path import warning is unrelated to Qwen3.5 recurrent-state reset." + "reason": "A local-path trailing-slash import warning is not the same as Qwen3.5 recurrent state reset." }, { "left": "issue:34689", "right": "issue:43611", "accept": false, - "reason": "Both are loading regressions, but the root causes and affected paths differ enough that they should not be merged." + "reason": "Both are model-loading regressions, but they involve different root causes: a specific Llama vision breakage versus `base_model_prefix` handling in v5.0.0." }, { "left": "issue:45003", "right": "issue:45440", "accept": false, - "reason": "Unsafe sys.modules access is unrelated to DeepSeekV3MoE divergence from the remote implementation." + "reason": "Unsafe `sys.modules[]` access in modeling utils and DeepseekV3MoE divergence from the remote implementation are unrelated." }, { "left": "issue:43720", "right": "issue:44442", "accept": false, - "reason": "Packed-weight handling during accelerate loading has nothing to do with tokenizer loading." + "reason": "Packed-weight unpacking during accelerate loading and tokenizer auto-loading are unrelated bugs." }, { - "left": "issue:42175", - "right": "issue:46234", + "left": "issue:43723", + "right": "issue:45479", "accept": false, - "reason": "Missing tensorflow in extras and missing import_utils symbol are separate dependency/API issues." + "reason": "Tokenizer loading failure in v5 and zero-loss single-label classification are different subsystems." }, { "left": "issue:43316", "right": "issue:43525", "accept": false, - "reason": "Gemma3TextConfig API mismatch and Llama4Config missing pad_token_id are different configuration bugs." - }, - { - "left": "issue:43723", - "right": "issue:45479", - "accept": false, - "reason": "Tokenizer loading in v5 and zero-loss classification with num_labels=1 are unrelated." + "reason": "Both are config API complaints, but one is a general Gemma3TextConfig discrepancy and the other is a concrete missing `pad_token_id` on Llama4Config; not the same bug." }, { - "left": "issue:44062", - "right": "issue:45907", + "left": "issue:42175", + "right": "issue:46234", "accept": false, - "reason": "AddedToken special-kw conflict and list_repo_files revision forwarding are unrelated API bugs." + "reason": "Missing TensorFlow in a torch extra and a missing import_utils availability helper are unrelated packaging/import issues." }, { "left": "issue:43493", "right": "issue:43827", "accept": false, - "reason": "SigLIP2 implementation drift and stale docs referencing pipeline() are unrelated." + "reason": "A SigLIP2 implementation mismatch and docs still referencing removed pipelines are completely different kinds of issues." }, { "left": "issue:44077", "right": "issue:44479", "accept": false, - "reason": "Model post_init policy and video-input regression are not the same bug." + "reason": "Patchtsmixer `post_init` policy and Qwen VL video-input regression are unrelated." }, { "left": "issue:43761", "right": "issue:44079", "accept": false, - "reason": "Hidden_states disappearing in CLIPVision and ModelOutput key assignment with None are different code paths." + "reason": "Hidden-states not returned from CLIPVisionModel and incorrect `ModelOutput` key assignment are both output-related, but they are different layers and not the same defect." + }, + { + "left": "issue:44062", + "right": "issue:45907", + "accept": false, + "reason": "`AddedToken()` duplicate `special` argument and `list_repo_files` not forwarding `revision` are unrelated API bugs." }, { "left": "issue:43873", "right": "issue:44871", "accept": false, - "reason": "Quantization/offloading behavior and Gemma3 eos_token_id inconsistency are unrelated." + "reason": "Quantization/offloading behavior and Gemma-3 `eos_token_id` inconsistency do not share the same root cause." }, { "left": "issue:43761", "right": "issue:44188", "accept": false, - "reason": "CLIPVision hidden_states regression and attention-kernel divergence under torch.compile are unrelated." + "reason": "A CLIPVisionModel output regression and attention-kernel divergence under `torch.compile` are unrelated." }, { "left": "issue:43232", "right": "issue:43388", "accept": false, - "reason": "Generation kwarg syncing after sync_gpus and gather_for_metrics label truncation are different problems." - }, - { - "left": "issue:43749", - "right": "issue:43931", - "accept": false, - "reason": "FSDP CPU RAM efficient loading and Qwen3-VL weight-shape mismatch are different loading failures." + "reason": "Generation kwargs sync behavior and `gather_for_metrics()` dropping tuple labels are different distributed-training bugs." }, { "left": "issue:44987", "right": "issue:46178", "accept": false, - "reason": "Model loading regression with transformers>=5.1.0 and missing tensorflow_text backend mapping are unrelated." - }, - { - "left": "issue:43931", - "right": "issue:45216", - "accept": false, - "reason": "Weight-shape mismatch on load and incorrect saved checkpoint contents are not the same concrete bug." + "reason": "A model-loading regression for `physical-intelligence/fast` and a TensorFlow Text backend mapping error are unrelated." } ] }, @@ -10271,7 +10196,7 @@ "source": "llm", "cache_hit": false, "model": "gpt-5.4-mini?service_tier=flex", - "cluster_id": "cluster-10000-472", + "cluster_id": "cluster-10000-473", "nodes": [ "issue:16998", "issue:22355", @@ -10293,7 +10218,9 @@ "issue:43701", "issue:43716", "issue:43742", + "issue:43749", "issue:43825", + "issue:43931", "issue:44077", "issue:44162", "issue:44190", @@ -10307,26 +10234,26 @@ "issue:44871", "issue:44913", "issue:45137", + "issue:45216", "issue:45584", - "issue:45632", "issue:45698", "issue:45759", "issue:45803", "issue:46050", - "issue:46170", "issue:46234", "issue:46291", "issue:46365", "issue:46369", "issue:46421", - "issue:46424", "issue:46506" ], "soft_pairs": [ - "issue:44442|issue:45803", + "issue:43749|issue:43931", + "issue:43931|issue:45216", "issue:43316|issue:45759", - "issue:43643|issue:44913", "issue:44855|issue:45137", + "issue:43643|issue:44913", + "issue:44442|issue:45803", "issue:43701|issue:46506", "issue:44190|issue:44496", "issue:42907|issue:45698", @@ -10336,232 +10263,230 @@ "issue:27957|issue:32090", "issue:43550|issue:43701", "issue:43575|issue:43742", - "issue:46234|issue:46291", "issue:43606|issue:43701", + "issue:46234|issue:46291", "issue:43295|issue:43825", "issue:16998|issue:22355", "issue:43295|issue:44871", "issue:44871|issue:45584", "issue:43316|issue:43716", "issue:42175|issue:43316", - "issue:45803|issue:46365", "issue:44077|issue:44534", + "issue:45803|issue:46365", "issue:44871|issue:46050", "issue:43646|issue:44913", "issue:44610|issue:46369", "issue:44162|issue:44792", - "issue:43452|issue:44496", - "issue:45632|issue:46424", - "issue:45803|issue:46170" + "issue:43452|issue:44496" ], - "prepared_review_unit_hash": "8dcf88c631fd7428016c67079d3312b65fe16f94da53d7ad88e02d7bcf9d98ff", + "prepared_review_unit_hash": "69e809013a87b331cdb613f1dbe6fdc07b64c6f91c7ae5de984909a1a000205e", "component_budget": { - "node_count": 472, - "item_count": 472, - "soft_pair_count": 1732, - "serialized_chars": 898654, - "estimated_input_tokens": 224664, - "estimated_eval_tokens": 449584 + "node_count": 473, + "item_count": 473, + "soft_pair_count": 1738, + "serialized_chars": 901437, + "estimated_input_tokens": 225360, + "estimated_eval_tokens": 450976 }, "budget": { "node_count": 48, "item_count": 48, "soft_pair_count": 30, - "serialized_chars": 36511, - "estimated_input_tokens": 9128, - "estimated_eval_tokens": 18512 + "serialized_chars": 36481, + "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 set of thematically similar but distinct Transformers issues (loading/config/cache/multimodal regressions). I don\u2019t see any pair that clearly shares the same underlying bug or change, so I would keep them separate.", - "confidence": 0.91, - "canonical_issue_reason": "issue:45698 is open, has the strongest discussion/inbound activity, and represents a broad from_pretrained/save_pretrained custom-module loading regression that is a useful anchor for the cluster.", + "summary": "The cluster is mostly a set of unrelated Transformers issues that only share broad themes like loading, config serialization, or model execution. None of the proposed soft edges look like the same underlying bug/change, so I would not merge any of them.", + "confidence": 0.88, + "canonical_issue_reason": "issue:45698 is the broadest and most representative open issue in this set: it centers on incorrect module resolution during save_pretrained/from_pretrained, a core loading/serialization path that overlaps the general theme of several nearby reports.", "canonical_pr_reason": null, - "best_issue_reason": "issue:45698 is the best global representative because it is current, high-signal, and about a cross-cutting loader-path failure rather than a narrow model-specific edge case.", + "best_issue_reason": "issue:45698 has the strongest triage value: open, high discussion, multiple inbound references, and a broad user-facing failure mode around custom-module loading after save/load.", "best_pr_reason": null, "soft_edge_verdicts": [ { - "left": "issue:44442", - "right": "issue:45803", + "left": "issue:43749", + "right": "issue:43931", + "accept": false, + "reason": "Different failure modes: FSDP CPU RAM-efficient loading vs Qwen3-VL weight-shape mismatch during model load." + }, + { + "left": "issue:43931", + "right": "issue:45216", "accept": false, - "reason": "Tokenizer-loading failure vs a bare-except bug in tensor conversion; different code paths and symptoms." + "reason": "Both involve Qwen family loading/saving, but one is a shape mismatch on load and the other is an incorrect checkpoint save_pretrained regression." }, { "left": "issue:43316", "right": "issue:45759", "accept": false, - "reason": "Both mention config/text_config, but one is an API discrepancy and the other is an AutoModel unwrapping bug for composite multimodal configs." + "reason": "Both touch config APIs, but Gemma3TextConfig inconsistency and Qwen multimodal text_config unwrapping are distinct bugs." + }, + { + "left": "issue:44855", + "right": "issue:45137", + "accept": false, + "reason": "Completely different subsystems: Python 3.13 JIT parsing/import issue vs DeepSpeed ZeRO3 deque error." }, { "left": "issue:43643", "right": "issue:44913", "accept": false, - "reason": "Both are config persistence issues, but they affect different fields and mechanisms (trust_remote_code missing fields vs rotary_pct reset on reload)." + "reason": "Both are config reload issues, but one is trust_remote_code missing fields and the other is GPTNeoX rotary_pct not persisting." }, { - "left": "issue:44855", - "right": "issue:45137", + "left": "issue:44442", + "right": "issue:45803", "accept": false, - "reason": "Completely unrelated failures: TorchScript parse error on import vs DeepSpeed ZeRO3 deque underflow." + "reason": "Tokenizer loading failure and a bare-except bug in Fuyu tensor conversion are unrelated." }, { "left": "issue:43701", "right": "issue:46506", "accept": false, - "reason": "Checkpoint-resume key mismatch is unrelated to CPU/CUDA output divergence in document layout inference." + "reason": "Checkpoint key mismatch on resume and CPU/CUDA output differences in PPDocLayoutV3 are not the same bug." }, { "left": "issue:44190", "right": "issue:44496", "accept": false, - "reason": "Local dataset loading in a training script is unrelated to unrecognized model/config loading." + "reason": "Dataset loading in an example script is unrelated to unrecognized model/config loading." }, { "left": "issue:42907", "right": "issue:45698", "accept": false, - "reason": "Both involve save_pretrained/from_pretrained, but one is quantized save failure and the other is wrong custom-module loading; different bugs." + "reason": "Both involve save/load paths, but dequantized model saving and wrong custom-module resolution are different root causes." }, { "left": "issue:44464", "right": "issue:46421", "accept": false, - "reason": "Both touch chunked generation/cache/compile, but one is inconsistent generation outputs and the other is recompilation due to cache lazy init." + "reason": "Both mention chunked behavior, but one is incorrect generation output and the other is spurious recompilation from cache initialization." }, { "left": "issue:42617", "right": "issue:43575", "accept": false, - "reason": "3d_parallel runtime failure and tensor-parallel OOM are different model/runtime issues." + "reason": "3D parallel script runtime issues and Qwen tp OOM are different execution problems." }, { "left": "issue:25251", "right": "issue:28282", "accept": false, - "reason": "Pipeline output-shape change is unrelated to an import error for missing PyTorch." + "reason": "Pipeline top_k output-shape change is unrelated to missing PyTorch import errors." }, { "left": "issue:27957", "right": "issue:32090", "accept": false, - "reason": "Flash Attention model support and Trainer GPU broadcast NoneType error are unrelated." + "reason": "Flash Attention support for XLMRoberta and a Trainer GPU broadcast TypeError are unrelated." }, { "left": "issue:43550", "right": "issue:43701", "accept": false, - "reason": "torch.compile/SDPA model bug versus checkpoint key mismatch; no shared underlying change." + "reason": "torch.compile/SDPA failure and resume_from_checkpoint key mismatch do not share a concrete code path." }, { "left": "issue:43575", "right": "issue:43742", "accept": false, - "reason": "OOM during tensor-parallel loading is unrelated to a key error while loading a different model." + "reason": "OOM while loading a large Qwen model and a key error loading MobileLLM are different load-time failures." }, { - "left": "issue:46234", - "right": "issue:46291", + "left": "issue:43606", + "right": "issue:43701", "accept": false, - "reason": "Both are import-utils related, but one is a missing availability helper and the other is a kernels version/revision import dependency failure." + "reason": "CPU offload device mismatch is unrelated to checkpoint key mismatch on resume." }, { - "left": "issue:43606", - "right": "issue:43701", + "left": "issue:46234", + "right": "issue:46291", "accept": false, - "reason": "CPU offload device mismatch and resume-from-checkpoint key mismatch are unrelated." + "reason": "Missing import_utils symbol and kernels version/revision validation are distinct dependency issues." }, { "left": "issue:43295", "right": "issue:43825", "accept": false, - "reason": "Custom model/processor regression is not the same as a pipeline error-message wording issue." + "reason": "Custom processor/tokenizer regression and pipeline task-support error messaging are different behaviors." }, { "left": "issue:16998", "right": "issue:22355", "accept": false, - "reason": "Model max length behavior and a missing transformers.onnx module are unrelated." + "reason": "model_max_length question for DeBERTa-V3 and missing transformers.onnx module are unrelated." }, { "left": "issue:43295", "right": "issue:44871", "accept": false, - "reason": "Both are regressions, but one is processor.tokenizer/image-passing breakage and the other is eos_token_id config inconsistency." + "reason": "Processor/tokenizer regression is unrelated to Gemma-3 eos_token_id inconsistency." }, { "left": "issue:44871", "right": "issue:45584", "accept": false, - "reason": "Eos-token config mismatch is not the same as Whisper empty-transcription generation failure." + "reason": "Gemma-3 eos_token_id config mismatch and Whisper empty transcription after align_special_tokens are different issues." }, { "left": "issue:43316", "right": "issue:43716", "accept": false, - "reason": "Config API discrepancy and image-preprocessor dtype mismatch are different multimodal issues." + "reason": "Gemma3TextConfig API discrepancy and Mistral-3 image preprocessor dtype mismatch are unrelated." }, { "left": "issue:42175", "right": "issue:43316", "accept": false, - "reason": "Package-extras backend omission is unrelated to Gemma3TextConfig API behavior." + "reason": "Packaging backend omissions and config API discrepancies are not the same bug." }, { - "left": "issue:45803", - "right": "issue:46365", + "left": "issue:44077", + "right": "issue:44534", "accept": false, - "reason": "Bare except swallowing errors in Fuyu conversion is unrelated to importing Trainer with datasets installed." + "reason": "Optional post_init enforcement for patchtsmixer is unrelated to non-persistent buffer initialization bugs." }, { - "left": "issue:44077", - "right": "issue:44534", + "left": "issue:45803", + "right": "issue:46365", "accept": false, - "reason": "Optional post_init validation and non-persistent buffer junk are separate v5 regressions." + "reason": "Bare except in Fuyu tensor conversion and Trainer import failure due to datasets are unrelated." }, { "left": "issue:44871", "right": "issue:46050", "accept": false, - "reason": "Eos-token config inconsistency and quantized save/load breakage are different persistence bugs." + "reason": "Gemma-3 eos_token_id mismatch and quantized model save/load corruption are different serialization problems." }, { "left": "issue:43646", "right": "issue:44913", "accept": false, - "reason": "Custom model initialization breakage is not the same as a config field reverting on reload." + "reason": "Transformers 5 custom model init breakage and GPTNeoXConfig rotary_pct persistence are different config/model-init issues." }, { "left": "issue:44610", "right": "issue:46369", "accept": false, - "reason": "Processor output size mismatch and model memory overflow are unrelated." + "reason": "OmDet-Turbo input-size mismatch and sam3-litetext memory overflow are unrelated model execution failures." }, { "left": "issue:44162", "right": "issue:44792", "accept": false, - "reason": "ESM2 model breakage and a janus image-generation test failure do not share the same bug." + "reason": "ESM2 breakage and janus test failure are unrelated model-specific regressions." }, { "left": "issue:43452", "right": "issue:44496", "accept": false, - "reason": "GGUF loading regression and unrecognized-model-config error are different loader failures." - }, - { - "left": "issue:45632", - "right": "issue:46424", - "accept": false, - "reason": "Trust-remote-code cache-path collision is unrelated to max_cache_length being honored only for quantized cache." - }, - { - "left": "issue:45803", - "right": "issue:46170", - "accept": false, - "reason": "Bare-except tensor conversion bug is unrelated to Windows/ROCm import dependency issues." + "reason": "gguf_file breaking from_pretrained calls and unrecognized model config errors do not share the same root cause." } ] }, @@ -10583,7 +10508,7 @@ "source": "llm", "cache_hit": false, "model": "gpt-5.4-mini?service_tier=flex", - "cluster_id": "cluster-10000-472", + "cluster_id": "cluster-10000-473", "nodes": [ "issue:35141", "issue:41950", @@ -10603,10 +10528,7 @@ "issue:43673", "issue:43688", "issue:43761", - "issue:43931", - "issue:44060", "issue:44190", - "issue:44242", "issue:44479", "issue:44514", "issue:44534", @@ -10627,225 +10549,220 @@ "issue:45600", "issue:45632", "issue:45698", + "issue:45803", + "issue:46170", "issue:46178", "issue:46202", "issue:46234", - "issue:46421", + "issue:46424", "issue:46617", "issue:46640" ], "soft_pairs": [ + "issue:45632|issue:46424", + "issue:45803|issue:46170", "issue:43650|issue:46202", "issue:43404|issue:43441", "issue:43065|issue:44617", "issue:43540|issue:44514", "issue:43761|issue:44479", "issue:43572|issue:46178", - "issue:43388|issue:43688", "issue:45127|issue:45200", + "issue:43388|issue:43688", "issue:43352|issue:46640", "issue:45542|issue:46234", - "issue:41950|issue:43475", "issue:43504|issue:45698", "issue:43673|issue:44743", "issue:44792|issue:44977", "issue:44792|issue:45600", + "issue:41950|issue:43475", "issue:43525|issue:43531", "issue:44734|issue:45406", - "issue:44534|issue:44913", "issue:44190|issue:44964", "issue:35141|issue:42490", + "issue:44534|issue:44913", "issue:44617|issue:45584", "issue:45632|issue:46617", - "issue:44855|issue:45588", - "issue:43931|issue:44060", - "issue:43761|issue:44242", - "issue:43673|issue:46421" + "issue:44855|issue:45588" ], - "prepared_review_unit_hash": "1b0bac3cd64291569e25108d79a48c82d8c42967a2f28b062ee5bb6c88557565", + "prepared_review_unit_hash": "58c96a3dcd1de070de2fabd73d3138c5f152577b7d0e870314747645ba00ee23", "component_budget": { - "node_count": 472, - "item_count": 472, - "soft_pair_count": 1732, - "serialized_chars": 898654, - "estimated_input_tokens": 224664, - "estimated_eval_tokens": 449584 + "node_count": 473, + "item_count": 473, + "soft_pair_count": 1738, + "serialized_chars": 901437, + "estimated_input_tokens": 225360, + "estimated_eval_tokens": 450976 }, "budget": { - "node_count": 48, - "item_count": 48, - "soft_pair_count": 26, - "serialized_chars": 35525, - "estimated_input_tokens": 8882, - "estimated_eval_tokens": 18020 + "node_count": 47, + "item_count": 47, + "soft_pair_count": 25, + "serialized_chars": 34668, + "estimated_input_tokens": 8667, + "estimated_eval_tokens": 17590 }, "overflow_policy": "truncate_then_skip", "trimmed": false, "aggressively_trimmed": false, "split": true, "analyst_result": { - "summary": "These items look like a mixed bag of unrelated Transformers bugs rather than one duplicate cluster. The soft pairs mostly share broad themes like caching, multimodal/video, or loading, but the concrete failure modes and code paths are different enough to reject them.", - "confidence": 0.84, - "canonical_issue_reason": "Issue 45698 is the broadest and most central loader-related report in the set, with the highest discussion activity and inbound references; if a single issue must represent the group, it is the best candidate.", + "summary": "The cluster is heterogeneous: most items are unrelated Transformers bugs across loading, cache, vision, flash-attention, and processor paths. None of the proposed soft pairs look like the same underlying change/bug closely enough to merge.", + "confidence": 0.94, + "canonical_issue_reason": "No single issue clearly acts as a true duplicate canonical here; the cluster is too mixed and the soft pairs are mostly broad similarity rather than the same bug.", "canonical_pr_reason": null, - "best_issue_reason": "Issue 45698 is the best standalone issue in this set because it describes a general from_pretrained/save_pretrained custom-module loading problem rather than a narrow model-specific symptom.", + "best_issue_reason": "issue:45698 is the strongest standalone representative: it is detailed, active, has inbound references, and centers on a concrete loading regression rather than a narrow model-specific symptom.", "best_pr_reason": null, "soft_edge_verdicts": [ + { + "left": "issue:45632", + "right": "issue:46424", + "accept": false, + "reason": "Both involve cache/loading behavior, but one is a trust_remote_code path collision and the other is max_cache_length being ignored; different bugs." + }, + { + "left": "issue:45803", + "right": "issue:46170", + "accept": false, + "reason": "Unrelated failures: tensor conversion error handling vs Windows/ROCm torch.distributed import issues." + }, { "left": "issue:43650", "right": "issue:46202", "accept": false, - "reason": "Unrelated: one is a vague data request, the other is a Gemma tokenizer roundtrip bug." + "reason": "No overlap beyond being user-facing bugs; one is a data request, the other is a Gemma tokenization roundtrip issue." }, { "left": "issue:43404", "right": "issue:43441", "accept": false, - "reason": "Different failures: tied-weight loading in Mistral3 vs FlashAttention support in Ministral-3." + "reason": "Different concrete problems: weight tying in Mistral3 vs FlashAttention support in Ministral-3." }, { "left": "issue:43065", "right": "issue:44617", "accept": false, - "reason": "Same SAM3 family, but one is about a dummy Conv2d and the other is CUDA OOM in video inference." + "reason": "Both mention SAM 3, but one is a dummy Conv2d implementation detail and the other is a CUDA OOM symptom." }, { "left": "issue:43540", "right": "issue:44514", "accept": false, - "reason": "Both multimodal/video-related, but one is a Qwen3OmniMoe video-input processing error and the other is batched chat-template padding behavior." + "reason": "Both are multimodal input bugs, but they affect different models and different code paths (video processing vs batched chat template padding)." }, { "left": "issue:43761", "right": "issue:44479", "accept": false, - "reason": "Different subsystems: CLIPVisionModel hidden-states regression vs Qwen video-input regression." + "reason": "Different regression surfaces: CLIP hidden_states forwarding vs Qwen video-input generation regression." }, { "left": "issue:43572", "right": "issue:46178", "accept": false, - "reason": "Unrelated config key bug vs backend-mapping error for tensorflow_text." + "reason": "Unrelated: missing pad_token_idx in StableLmConfig vs an unknown tensorflow_text backend mapping error." }, { - "left": "issue:43388", - "right": "issue:43688", + "left": "issue:45127", + "right": "issue:45200", "accept": false, - "reason": "Different training/eval bugs: metric-gathering label truncation vs auxiliary-loss normalization." + "reason": "One is LoRA/extended-vocabulary collapse with tied embeddings; the other is Gemma 4 mm_token_type_ids defaulting for text-only fine-tuning." }, { - "left": "issue:45127", - "right": "issue:45200", + "left": "issue:43388", + "right": "issue:43688", "accept": false, - "reason": "Both model-specific, but one is LoRA/vocab merge collapse and the other is a Gemma 4 token-type-id defaulting issue." + "reason": "Different training bugs: metric-gathering drops labels vs auxiliary-loss normalization in MoE models." }, { "left": "issue:43352", "right": "issue:46640", "accept": false, - "reason": "Too vague on the right and not the same concrete FlashAttention 2 incompatibility as the Nemotron issue." + "reason": "Both mention Flash Attention 2, but one is a specific model not supporting it and the other is too vague to confirm the same issue." }, { "left": "issue:45542", "right": "issue:46234", "accept": false, - "reason": "Different import/backend problems: TensorFlow backend detection vs missing import_utils symbol." - }, - { - "left": "issue:41950", - "right": "issue:43475", - "accept": false, - "reason": "Both mention vision/video, but one is processor selection and the other is a missing FPN attribute on Sam3 output." + "reason": "Different backend/import problems: TensorBoard without TensorFlow vs missing Falcon Mamba SSM availability helper." }, { "left": "issue:43504", "right": "issue:45698", "accept": false, - "reason": "Both loading-related, but one is a legacy field preset load failure and the other is wrong custom-module resolution after save/load." + "reason": "Both concern model loading, but one is a legacy preset field for BEiT and the other is selecting the wrong custom module after save_pretrained." }, { "left": "issue:43673", "right": "issue:44743", "accept": false, - "reason": "Cache-related, but different problems: chunked prefill cache initialization vs recurrent-state reset with seq_len>1." + "reason": "Both involve cache state, but one is GenerationMixin cache missing in chunked prefill and the other is recurrent states resetting in modular_qwen3_5." }, { "left": "issue:44792", "right": "issue:44977", "accept": false, - "reason": "Unrelated: janus image-generation test failure vs Qwen3.5 FlashAttention generation bug." + "reason": "Different models and symptoms: Janus image generation test failure vs Qwen3.5 flash-attention generation regression." }, { "left": "issue:44792", "right": "issue:45600", "accept": false, - "reason": "Different topics: janus generate-images test vs auto_mappings referencing removed configs." + "reason": "No shared bug: Janus generation test failure vs stale auto_mappings entries for removed Sam3LiteText configs." + }, + { + "left": "issue:41950", + "right": "issue:43475", + "accept": false, + "reason": "Different SAM 3 paths: processor selection bug vs missing fpn_position_embeddings attribute on a video encoder output." }, { "left": "issue:43525", "right": "issue:43531", "accept": false, - "reason": "Different model bugs: missing pad_token_id in Llama4Config vs Qwen3-MoE sliding_window behavior." + "reason": "Separate config regressions: missing pad_token_id in Llama4Config vs sliding_window behavior in Qwen3-MoE." }, { "left": "issue:44734", "right": "issue:45406", "accept": false, - "reason": "Both serve-related, but one is KV-cache continuation indexing and the other is a missing _tokenizer on Gemma4Processor." - }, - { - "left": "issue:44534", - "right": "issue:44913", - "accept": false, - "reason": "Unrelated config/runtime issues: non-persistent buffers vs GPTNeoX rotary_pct reload defaulting." + "reason": "Unrelated service/model issues: KV-cache continuation tensor indexing vs Gemma4Processor missing _tokenizer." }, { "left": "issue:44190", "right": "issue:44964", "accept": false, - "reason": "Different loading paths: local dataset script failure vs multimodal model load failure." + "reason": "Different loading failures: local dataset in example script vs a specific multimodal model failing to load on latest transformers." }, { "left": "issue:35141", "right": "issue:42490", "accept": false, - "reason": "Completely different bugs: embedding resize/tie-word-embedding behavior vs path trailing-slash import warning." + "reason": "Different loading behaviors: embedding reinit after resize vs warning on local-model paths ending with a slash." + }, + { + "left": "issue:44534", + "right": "issue:44913", + "accept": false, + "reason": "One is non-persistent buffers being filled with junk; the other is GPTNeoXConfig rotary_pct not persisting on reload." }, { "left": "issue:44617", "right": "issue:45584", "accept": false, - "reason": "Unrelated: SAM3 video OOM vs Whisper empty-transcription handling after align_special_tokens." + "reason": "Different symptoms and code paths: SAM 3 video OOM vs Whisper empty-transcription failure after align_special_tokens." }, { "left": "issue:45632", "right": "issue:46617", "accept": false, - "reason": "Both trust_remote_code/cache-related, but one is leaf-directory cache collision and the other is FileNotFoundError from cache loading." + "reason": "Both are trust_remote_code cache-related, but one is a leaf-directory cache collision and the other is a FileNotFoundError from cache loading; not clearly the same bug." }, { "left": "issue:44855", "right": "issue:45588", "accept": false, - "reason": "Different failure modes: Python 3.13 IndentationError in DebertaV2 vs flash_attention.py s_aux=None AttributeError." - }, - { - "left": "issue:43931", - "right": "issue:44060", - "accept": false, - "reason": "Both Qwen-related, but one is a weight-shape mismatch and the other is an incorrect tied-weights warning." - }, - { - "left": "issue:43761", - "right": "issue:44242", - "accept": false, - "reason": "Both regression-ish, but CLIP hidden-states is unrelated to load-balancing loss being skipped when output_router_logits=False." - }, - { - "left": "issue:43673", - "right": "issue:46421", - "accept": false, - "reason": "Both mention cache and prefill, but one is missing GenerationMixin cache in v5.0.0 while the other is spurious recompilation from lazy cache init." + "reason": "Completely different areas: Python 3.13 import parsing issue in DebertaV2 vs flash_attention.py crashing on s_aux=None." } ] }, @@ -10867,7 +10784,7 @@ "source": "llm", "cache_hit": false, "model": "gpt-5.4-mini?service_tier=flex", - "cluster_id": "cluster-10000-472", + "cluster_id": "cluster-10000-473", "nodes": [ "issue:25251", "issue:27957", @@ -10879,13 +10796,16 @@ "issue:43540", "issue:43550", "issue:43644", + "issue:43673", + "issue:43761", "issue:43827", + "issue:43931", "issue:43994", + "issue:44060", "issue:44077", "issue:44112", "issue:44190", - "issue:44246", - "issue:44462", + "issue:44242", "issue:44464", "issue:44479", "issue:44521", @@ -10899,7 +10819,6 @@ "issue:44913", "issue:44933", "issue:44938", - "issue:45003", "issue:45127", "issue:45245", "issue:45276", @@ -10910,253 +10829,251 @@ "issue:45563", "issue:45584", "issue:45600", - "issue:45632", "issue:45803", - "issue:45907", - "issue:46234", "issue:46369", + "issue:46421", "issue:46489", "issue:46655" ], "soft_pairs": [ + "issue:43931|issue:44060", + "issue:43761|issue:44242", + "issue:43673|issue:46421", "issue:43827|issue:43994", "issue:45563|issue:45584", "issue:45381|issue:46369", - "issue:43644|issue:44913", "issue:44479|issue:44734", + "issue:43644|issue:44913", "issue:25251|issue:27957", "issue:44933|issue:45362", - "issue:41720|issue:45245", "issue:43065|issue:44610", + "issue:41720|issue:45245", "issue:43335|issue:44855", "issue:43550|issue:44190", "issue:44077|issue:44877", "issue:44464|issue:45563", "issue:45381|issue:46655", - "issue:44617|issue:46369", "issue:43335|issue:44190", + "issue:44617|issue:46369", "issue:45127|issue:45357", "issue:44779|issue:46489", - "issue:44112|issue:45600", "issue:39692|issue:44190", + "issue:44112|issue:45600", "issue:44521|issue:45381", + "issue:43540|issue:44734", "issue:44737|issue:45803", "issue:43295|issue:45276", - "issue:43540|issue:44734", "issue:44938|issue:45542", - "issue:44462|issue:45907", - "issue:45276|issue:45584", - "issue:45632|issue:46234", - "issue:44246|issue:45003" + "issue:45276|issue:45584" ], - "prepared_review_unit_hash": "4abe87aef01f05018cbd61e2427d0f822d34a7f92dc63dbe9ac4373b765196ac", + "prepared_review_unit_hash": "a504261e91e68be253ba349cfefc3ef20228e4445112de5f03c9f470a0bbb438", "component_budget": { - "node_count": 472, - "item_count": 472, - "soft_pair_count": 1732, - "serialized_chars": 898654, - "estimated_input_tokens": 224664, - "estimated_eval_tokens": 449584 + "node_count": 473, + "item_count": 473, + "soft_pair_count": 1738, + "serialized_chars": 901437, + "estimated_input_tokens": 225360, + "estimated_eval_tokens": 450976 }, "budget": { "node_count": 48, "item_count": 48, "soft_pair_count": 29, - "serialized_chars": 36259, - "estimated_input_tokens": 9065, - "estimated_eval_tokens": 18386 + "serialized_chars": 36346, + "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 clearly distinct bug reports. The only likely duplicate is the DeepSeek tokenizer regression pair, which appears to describe the same v5+ tokenizer breakage at different specificity levels.", - "confidence": 0.87, - "canonical_issue_reason": "issue:46489 is the most specific and actionable report in the only apparent duplicate pair, with a clear model/version scope and direct reference to the prior fix gap.", + "summary": "This cluster is mostly a set of unrelated issue reports; nearly all soft edges are false positives. The only plausible duplicate/follow-up pair is the DeepSeek tokenizer v5 regression (44779/46489).", + "confidence": 0.89, + "canonical_issue_reason": "issue:44779 is the original, broader DeepSeek tokenizer v5 regression report and the clearest anchor for the only near-duplicate pair in this set.", "canonical_pr_reason": null, - "best_issue_reason": "issue:46489 is the best representative issue overall because it is narrowly scoped, regression-focused, and tied to the relevant upstream fix history.", + "best_issue_reason": "issue:44779 is the most representative issue here: concrete repro, broadest wording, and it anchors the only plausible duplicate relationship.", "best_pr_reason": null, "soft_edge_verdicts": [ + { + "left": "issue:43931", + "right": "issue:44060", + "accept": false, + "reason": "Different failure modes: Qwen3-VL weight-shape load error vs Qwen3-Next tied-weights warning." + }, + { + "left": "issue:43761", + "right": "issue:44242", + "accept": false, + "reason": "CLIPVision hidden_states regression and load-balancing-loss gating are unrelated code paths." + }, + { + "left": "issue:43673", + "right": "issue:46421", + "accept": false, + "reason": "Both mention chunked prefill/cache, but one is missing cache state and the other is recompilation from lazy init." + }, { "left": "issue:43827", "right": "issue:43994", "accept": false, - "reason": "Docs mentioning removed pipeline APIs are unrelated to a SigLIP2 model/pipeline output bug." + "reason": "Docs still referencing pipeline() is not the same bug as SigLIP2 producing bad outputs with AutoModel/pipeline." }, { "left": "issue:45563", "right": "issue:45584", "accept": false, - "reason": "Different failure modes: a stale warning in paged generate vs an empty-transcription crash after token alignment." + "reason": "A stale generation warning is distinct from Whisper failing on empty transcription after align_special_tokens." }, { "left": "issue:45381", "right": "issue:46369", "accept": false, - "reason": "Both are multimodal/model bugs, but they concern different models and symptoms: video position IDs vs memory overflow." + "reason": "Qwen2.5-VL video position-id regression and Sam3-LiteText memory overflow are different model-specific bugs." }, { - "left": "issue:43644", - "right": "issue:44913", + "left": "issue:44479", + "right": "issue:44734", "accept": false, - "reason": "Non-persistent buffer initialization is unrelated to GPTNeoXConfig rotary_pct serialization/reload behavior." + "reason": "Video input regression and KV-cache continuation crash are separate failure points." }, { - "left": "issue:44479", - "right": "issue:44734", + "left": "issue:43644", + "right": "issue:44913", "accept": false, - "reason": "Both involve Qwen/serving regressions, but one is video-input handling and the other is KV-cache continuation indexing." + "reason": "Non-persistent buffer junk-fill and GPTNeoXConfig reload behavior are different issues." }, { "left": "issue:25251", "right": "issue:27957", "accept": false, - "reason": "Pipeline top_k nesting behavior is unrelated to XLMRoberta Flash Attention 2 support." + "reason": "Pipeline output nesting and Flash Attention 2 support are unrelated." }, { "left": "issue:44933", "right": "issue:45362", "accept": false, - "reason": "A missing import from image_utils is not the same bug as Qwen3.5 chat crashes." + "reason": "Missing import from image_utils is unrelated to Qwen3.5 chat crashes." }, { - "left": "issue:41720", - "right": "issue:45245", + "left": "issue:43065", + "right": "issue:44610", "accept": false, - "reason": "Qwen3 auto device mapping CUDA asserts and a category-count limit error are different code paths." + "reason": "Dummy Conv2d in Sam3PixelDecoder and OmDet-Turbo processor/model size mismatch are different bugs." }, { - "left": "issue:43065", - "right": "issue:44610", + "left": "issue:41720", + "right": "issue:45245", "accept": false, - "reason": "A dummy Conv2d in Sam3PixelDecoder is unrelated to OmDet-Turbo's processor/model image-size mismatch." + "reason": "CUDA assert on auto device mapping is unrelated to the 2^24 category limit error." }, { "left": "issue:43335", "right": "issue:44855", "accept": false, - "reason": "SwitchTransformers sparse-layer creation and DebertaV2 Python 3.13 import parsing are unrelated bugs." + "reason": "SwitchTransformers sparse-layer construction and a Python 3.13 IndentationError are not the same problem." }, { "left": "issue:43550", "right": "issue:44190", "accept": false, - "reason": "A torch.compile/SDPA model failure is unrelated to local dataset loading in the training script." + "reason": "torch.compile+SDPA failure and local dataset loading are unrelated." }, { "left": "issue:44077", "right": "issue:44877", "accept": false, - "reason": "Optional post_init handling in one config is not the same as strict config rejection for granite_speech." + "reason": "Optional post_init on patchtsmixer and strict granite_speech config loading are separate config issues." }, { "left": "issue:44464", "right": "issue:45563", "accept": false, - "reason": "Compiled-forward generation inconsistency is a different generation bug from the stale num_return_sequences warning." + "reason": "Compiled-forward generation inconsistency is not the same as the stale num_return_sequences warning." }, { "left": "issue:45381", "right": "issue:46655", "accept": false, - "reason": "Both are video-input issues, but they target different models and distinct preprocessing problems." + "reason": "Both are video bugs, but one is wrong vision_position_ids and the other is incorrect padding calculation." }, { - "left": "issue:44617", - "right": "issue:46369", + "left": "issue:43335", + "right": "issue:44190", "accept": false, - "reason": "Sam3Video CUDA OOM and sam3-litetext memory overflow are related only by family name, not the same concrete bug." + "reason": "Sparse-layer config logic and a dataset-loading script failure are unrelated." }, { - "left": "issue:43335", - "right": "issue:44190", + "left": "issue:44617", + "right": "issue:46369", "accept": false, - "reason": "Sparse-layer config construction is unrelated to local dataset loading failures." + "reason": "OOM in Sam3Video and memory overflow in sam3-litetext are different symptoms on different variants." }, { "left": "issue:45127", "right": "issue:45357", "accept": false, - "reason": "LoRA vocabulary merge collapse is unrelated to incorrect visual encoder keys during save_pretrained." + "reason": "LoRA merge collapse with tied embeddings is unrelated to incorrect visual encoder keys on save_pretrained." }, { "left": "issue:44779", "right": "issue:46489", "accept": true, - "reason": "Same DeepSeek tokenizer regression on v5+: one report is broad, the other is a specific DeepSeek-Coder instance tied to the earlier fix gap." + "reason": "Both describe the same DeepSeek tokenizer regression on v5+; 46489 reads like the remaining gap in the earlier fix rather than a separate bug." }, { - "left": "issue:44112", - "right": "issue:45600", + "left": "issue:39692", + "right": "issue:44190", "accept": false, - "reason": "A stale GraniteSpeech device override test and broken auto_mappings references are separate CI issues." + "reason": "SigLIP2 documentation errors and a local dataset loading script issue are unrelated." }, { - "left": "issue:39692", - "right": "issue:44190", + "left": "issue:44112", + "right": "issue:45600", "accept": false, - "reason": "SigLIP2 documentation example errors are unrelated to local dataset loading in the image-classification script." + "reason": "Stale CI device-override test and removed Sam3LiteText auto_mapping references are different maintenance problems." }, { "left": "issue:44521", "right": "issue:45381", "accept": false, - "reason": "All-zero assistant masks and wrong video vision_position_ids are different multimodal template/preprocessing bugs." + "reason": "All-zero assistant masks in apply_chat_template are unrelated to Qwen2.5-VL video position IDs." + }, + { + "left": "issue:43540", + "right": "issue:44734", + "accept": false, + "reason": "Qwen3OmniMoe video-input ValueError and serve KV-cache continuation indexing are unrelated." }, { "left": "issue:44737", "right": "issue:45803", "accept": false, - "reason": "CPU tensor creation in XLNet relative positions is unrelated to Fuyu's broad exception handling bug." + "reason": "CPU arange/device placement bug in XLNet is unrelated to FuyuBatchFeature swallowing exceptions." }, { "left": "issue:43295", "right": "issue:45276", "accept": false, - "reason": "Processor/tokenizer regression after v4.57.5 is unrelated to gemma4 resize_token_embeddings propagation." - }, - { - "left": "issue:43540", - "right": "issue:44734", - "accept": false, - "reason": "Qwen3OmniMoe video processing ValueError and serve KV-cache indexing are separate failures." + "reason": "processor.tokenizer regression and gemma4 embedding-resize propagation are different code paths." }, { "left": "issue:44938", "right": "issue:45542", "accept": false, - "reason": "Python 3.14 import/load failures are unrelated to the TensorBoard-without-TensorFlow backend issue." - }, - { - "left": "issue:44462", - "right": "issue:45907", - "accept": false, - "reason": "Tokenizer.json lookup behavior is unrelated to list_repo_files missing the revision argument." + "reason": "Python 3.14 import/load failure is unrelated to the undefined tf backend error." }, { "left": "issue:45276", "right": "issue:45584", "accept": false, - "reason": "Embedding resize propagation and Whisper empty-transcription generation failures are different problems." - }, - { - "left": "issue:45632", - "right": "issue:46234", - "accept": false, - "reason": "A trust_remote_code cache-path collision is unrelated to a missing import_utils availability helper." - }, - { - "left": "issue:44246", - "right": "issue:45003", - "accept": false, - "reason": "Slow import timing and unsafe sys.modules access are different import-time concerns, not the same bug." + "reason": "Embedding resize propagation and Whisper empty-transcription failure are unrelated." } ] }, "evaluator_result": { "accept": true, - "feedback": "The summary is grounded in the packet, and the soft-edge decisions are conservative. The only accepted pair (issue:44779 vs issue:46489) is plausibly the same DeepSeek tokenizer regression at different specificity levels, while the rest are appropriately treated as distinct bugs. No PR-pair mergeability concerns apply here since this is issue-only data." + "feedback": "The summary is broadly grounded in the packet and stays conservative overall. The one accepted pair, 44779/46489, is supported by the titles as a plausible same-root DeepSeek tokenizer v5+ regression/follow-up. I don\u2019t see a clear overstatement strong enough to reject." }, "evaluator_used": true, "retried": false, @@ -11175,7 +11092,7 @@ "source": "llm", "cache_hit": false, "model": "gpt-5.4-mini?service_tier=flex", - "cluster_id": "cluster-10000-472", + "cluster_id": "cluster-10000-473", "nodes": [ "issue:27957", "issue:32090", @@ -11186,32 +11103,30 @@ "issue:41950", "issue:42175", "issue:42796", - "issue:42907", "issue:43295", "issue:43404", - "issue:43441", "issue:43479", - "issue:43526", "issue:43531", "issue:43643", "issue:43650", "issue:43673", "issue:43854", - "issue:43901", "issue:43976", "issue:44016", "issue:44112", "issue:44186", "issue:44220", + "issue:44246", "issue:44279", "issue:44295", + "issue:44462", "issue:44496", "issue:44521", "issue:44683", "issue:44734", "issue:44964", "issue:44987", - "issue:45042", + "issue:45003", "issue:45083", "issue:45216", "issue:45245", @@ -11220,25 +11135,31 @@ "issue:45446", "issue:45584", "issue:45600", + "issue:45632", + "issue:45907", "issue:46082", "issue:46121", + "issue:46234", "issue:46424", "issue:46506" ], "soft_pairs": [ + "issue:44462|issue:45907", + "issue:45632|issue:46234", + "issue:44246|issue:45003", "issue:46082|issue:46121", "issue:42796|issue:44521", "issue:45245|issue:45584", "issue:44964|issue:45600", - "issue:43643|issue:44496", "issue:42175|issue:44279", - "issue:43854|issue:44964", "issue:44112|issue:44964", - "issue:36246|issue:41628", + "issue:43854|issue:44964", + "issue:43643|issue:44496", "issue:27957|issue:35141", - "issue:36246|issue:38617", - "issue:43404|issue:45216", + "issue:36246|issue:41628", "issue:44186|issue:44734", + "issue:43404|issue:45216", + "issue:36246|issue:38617", "issue:44295|issue:46506", "issue:43650|issue:44220", "issue:41950|issue:44016", @@ -11250,214 +11171,204 @@ "issue:44016|issue:44220", "issue:45083|issue:46506", "issue:43976|issue:44987", - "issue:43295|issue:45335", - "issue:43526|issue:43901", - "issue:44496|issue:45042", - "issue:42907|issue:43441", - "issue:43901|issue:44496" + "issue:43295|issue:45335" ], - "prepared_review_unit_hash": "9c73959754f5004c122119aff7770c96434e32e51e819ae2ca3a85be9ce2e805", + "prepared_review_unit_hash": "ed85dcf211b94fbb924d6f4814cf92888ea79b1d8b575ab229bf06f01cd5dc98", "component_budget": { - "node_count": 472, - "item_count": 472, - "soft_pair_count": 1732, - "serialized_chars": 898654, - "estimated_input_tokens": 224664, - "estimated_eval_tokens": 449584 + "node_count": 473, + "item_count": 473, + "soft_pair_count": 1738, + "serialized_chars": 901437, + "estimated_input_tokens": 225360, + "estimated_eval_tokens": 450976 }, "budget": { - "node_count": 47, - "item_count": 47, - "soft_pair_count": 29, - "serialized_chars": 35860, - "estimated_input_tokens": 8965, - "estimated_eval_tokens": 18186 + "node_count": 48, + "item_count": 48, + "soft_pair_count": 28, + "serialized_chars": 36079, + "estimated_input_tokens": 9020, + "estimated_eval_tokens": 18296 }, "overflow_policy": "truncate_then_skip", "trimmed": false, "aggressively_trimmed": false, "split": true, "analyst_result": { - "summary": "The cluster is mostly noise: the soft pairs span unrelated bug reports, packaging issues, model-specific regressions, and docs/notebook problems. I found no true duplicate pair to merge.", - "confidence": 0.97, - "canonical_issue_reason": "No single canonical issue fits this cluster; the items cover many unrelated failure modes across different subsystems and models.", + "summary": "The cluster is heterogeneous; the soft pairs connect different APIs, model families, or unrelated regressions, so none should be treated as duplicates.", + "confidence": 0.96, + "canonical_issue_reason": "No canonical issue stands out: these items are not one bug but many unrelated Transformers issues.", "canonical_pr_reason": null, - "best_issue_reason": "No issue is a strong global representative because the cluster is not semantically coherent enough to justify one canonical issue.", + "best_issue_reason": "No issue is a strong cluster representative because the candidates span unrelated failure modes and subsystems.", "best_pr_reason": null, "soft_edge_verdicts": [ + { + "left": "issue:44462", + "right": "issue:45907", + "accept": false, + "reason": "Tokenizer.json precedence vs revision forwarding are different APIs and failures." + }, + { + "left": "issue:45632", + "right": "issue:46234", + "accept": false, + "reason": "Local-model cache-path collision is unrelated to a missing import export." + }, + { + "left": "issue:44246", + "right": "issue:45003", + "accept": false, + "reason": "Slow import-time access to sys.modules is not the same bug as `_gpu_broadcast_one` rejecting None." + }, { "left": "issue:46082", "right": "issue:46121", "accept": false, - "reason": "Different bugs: LlamaConfig head_dim validation vs rope-validation type handling." + "reason": "Config validation for head_dim and JSON-loaded rope-key lists are separate errors." }, { "left": "issue:42796", "right": "issue:44521", "accept": false, - "reason": "Tokenizer BPE-marker handling and multimodal assistant mask generation are unrelated." + "reason": "BPE marker handling and multimodal assistant masks are different tokenizer behaviors." }, { "left": "issue:45245", "right": "issue:45584", "accept": false, - "reason": "Category-count runtime error and Whisper empty-transcription handling are different code paths." + "reason": "Category-count overflow and Whisper empty-transcription handling are unrelated." }, { "left": "issue:44964", "right": "issue:45600", "accept": false, - "reason": "Model loading for Phi-4 multimodal is unrelated to stale auto_mappings references." - }, - { - "left": "issue:43643", - "right": "issue:44496", - "accept": false, - "reason": "Both involve config loading, but one is trust_remote_code field loss and the other is an unrecognized-model/model_type error." + "reason": "Model-load breakage for Phi-4 multimodal and removed Sam3LiteText auto-mappings are different root causes." }, { "left": "issue:42175", "right": "issue:44279", "accept": false, - "reason": "Packaging backend mismatch is not the same as a generic dependency issue report." + "reason": "TensorFlow backend packaging and a generic dependency issue are not the same failure." }, { - "left": "issue:43854", + "left": "issue:44112", "right": "issue:44964", "accept": false, - "reason": "Different models and different loading failures; no shared underlying bug." + "reason": "GraniteSpeech CI test flake and Phi-4 multimodal loading are unrelated." }, { - "left": "issue:44112", + "left": "issue:43854", "right": "issue:44964", "accept": false, - "reason": "A stale CI test for GraniteSpeech is unrelated to Phi-4 multimodal model loading." + "reason": "GLM-4.7-Flash test loading and Phi-4 multimodal loading are different model-specific bugs." }, { - "left": "issue:36246", - "right": "issue:41628", + "left": "issue:43643", + "right": "issue:44496", "accept": false, - "reason": "Both are import errors, but they concern different symbols and different missing-export causes." + "reason": "Missing config fields with trust_remote_code and unrecognized model_type are distinct config-loading issues." }, { "left": "issue:27957", "right": "issue:35141", "accept": false, - "reason": "FlashAttention support for XLMRoberta is unrelated to embedding reinitialization after resize." + "reason": "XLMRoberta Flash Attention 2 support and tie_word_embedding/post_init reinit are unrelated." }, { "left": "issue:36246", - "right": "issue:38617", + "right": "issue:41628", + "accept": false, + "reason": "Missing Qwen2_5_VLImageProcessor export and missing AutoImageProcessor export are separate import regressions." + }, + { + "left": "issue:44186", + "right": "issue:44734", "accept": false, - "reason": "Separate import failures in different modules with different missing names." + "reason": "NER/batched-padding tokenizer crashes and KV-cache continuation tensor indexing are different code paths." }, { "left": "issue:43404", "right": "issue:45216", "accept": false, - "reason": "Tie-weight bug in Mistral3 and Qwen3.5 checkpoint saving are distinct regressions." + "reason": "Mistral3 lm_head tying and Qwen3.5 save_pretrained checkpoint corruption are different model bugs." }, { - "left": "issue:44186", - "right": "issue:44734", + "left": "issue:36246", + "right": "issue:38617", "accept": false, - "reason": "Tokenizer padding/NER crash and serve KV-cache continuation crash are unrelated." + "reason": "Both are import errors, but they concern different missing symbols in different modules." }, { "left": "issue:44295", "right": "issue:46506", "accept": false, - "reason": "One is a position_ids buffer access error; the other is CPU/CUDA divergence in PPDocLayoutV3." + "reason": "position_ids buffer access and CPU/CUDA output divergence are unrelated." }, { "left": "issue:43650", "right": "issue:44220", "accept": false, - "reason": "Junk/generic issue text and audio feature extraction are not the same bug." + "reason": "ADD THE DATA is not the same as an `_torch_extract_fbank_features()` bug." }, { "left": "issue:41950", "right": "issue:44016", "accept": false, - "reason": "Video-classification processor lookup and a notebook syntax error are unrelated." + "reason": "Video-classification pipeline image-processor lookup and a notebook syntax error are unrelated." }, { "left": "issue:32090", "right": "issue:35141", "accept": false, - "reason": "Trainer broadcast TypeError is unrelated to token-embedding resize/reinit behavior." + "reason": "Trainer NoneType broadcast and token-embedding resize reinitialization are different issues." }, { "left": "issue:44683", "right": "issue:45446", "accept": false, - "reason": "Both touch flex_attention, but one is a compile failure on torch>=2.9 while the other is a bad PyTorch version gate for AuxRequest import." + "reason": "Compiled flex_attention on torch>=2.9 and AuxRequest version gating are separate flex_attention regressions." }, { "left": "issue:43479", "right": "issue:45375", "accept": false, - "reason": "Different multimodal config bugs affecting different fields and model families." + "reason": "Phi4MultimodalConfig default init and missing Qwen3_5MoeVisionConfig field are different config bugs." }, { "left": "issue:43673", "right": "issue:46424", "accept": false, - "reason": "Chunked_prefill cache regression and max_cache_length only applying to quantized cache are different issues." + "reason": "Generation cache missing in v5.0.0 chunked_prefill and max_cache_length only applying to quantized cache are different cache behaviors." }, { "left": "issue:43531", "right": "issue:45600", "accept": false, - "reason": "Qwen3-MoE sliding_window behavior is unrelated to removed auto_mapping references." + "reason": "Sliding-window behavior in Qwen3-MoE and removed Sam3LiteText auto-mappings are unrelated." }, { "left": "issue:44016", "right": "issue:44220", "accept": false, - "reason": "Notebook syntax error and _torch_extract_fbank_features are unrelated." + "reason": "Notebook syntax error and `_torch_extract_fbank_features()` are not the same problem." }, { "left": "issue:45083", "right": "issue:46506", "accept": false, - "reason": "Helper output-length logic in qwen3_omni_moe is unrelated to CPU/CUDA result differences." + "reason": "Helper output-length math in qwen3_omni_moe and CPU/CUDA differences are separate bugs." }, { "left": "issue:43976", "right": "issue:44987", "accept": false, - "reason": "Python-version support packaging issue and a model-loading failure are different problems." + "reason": "Python version support in Transformers 5.1.0 and loading a specific model are unrelated." }, { "left": "issue:43295", "right": "issue:45335", "accept": false, - "reason": "Processor/tokenizer regression and t5gemma embedding resize mismatch are distinct bugs." - }, - { - "left": "issue:43526", - "right": "issue:43901", - "accept": false, - "reason": "BeitImageProcessorFast label reduction and TextClassificationPipeline docs mismatch are unrelated." - }, - { - "left": "issue:44496", - "right": "issue:45042", - "accept": false, - "reason": "Unrecognized model/config loading and PIL backend torchvision dependency are different failures." - }, - { - "left": "issue:42907", - "right": "issue:43441", - "accept": false, - "reason": "Ministral dequantized save failure and FlashAttention failure are separate model-specific regressions." - }, - { - "left": "issue:43901", - "right": "issue:44496", - "accept": false, - "reason": "Docs mentioning return_all_scores and unrecognized-model config loading are not the same bug." + "reason": "Custom model processor.tokenizer access and resize_token_embeddings/decoder tying are different regressions." } ] }, @@ -11479,16 +11390,17 @@ "source": "llm", "cache_hit": false, "model": "gpt-5.4-mini?service_tier=flex", - "cluster_id": "cluster-10000-472", + "cluster_id": "cluster-10000-473", "nodes": [ - "issue:38175", "issue:39692", "issue:41950", "issue:42796", + "issue:42907", "issue:43295", "issue:43299", "issue:43381", "issue:43441", + "issue:43526", "issue:43577", "issue:43650", "issue:43673", @@ -11505,11 +11417,11 @@ "issue:44351", "issue:44368", "issue:44496", - "issue:44617", "issue:44779", "issue:44792", "issue:44933", "issue:45020", + "issue:45042", "issue:45200", "issue:45216", "issue:45278", @@ -11521,7 +11433,6 @@ "issue:45593", "issue:45759", "issue:45812", - "issue:45823", "issue:46082", "issue:46123", "issue:46369", @@ -11531,12 +11442,16 @@ "issue:46515" ], "soft_pairs": [ + "issue:43526|issue:43901", + "issue:44496|issue:45042", + "issue:42907|issue:43441", + "issue:43901|issue:44496", "issue:43381|issue:45593", "issue:43976|issue:44279", "issue:43673|issue:44336", "issue:43441|issue:45588", - "issue:44016|issue:44062", "issue:46123|issue:46369", + "issue:44016|issue:44062", "issue:44792|issue:46424", "issue:43824|issue:44933", "issue:44496|issue:45759", @@ -11546,229 +11461,233 @@ "issue:46082|issue:46399", "issue:44075|issue:44368", "issue:44279|issue:45405", - "issue:43577|issue:45216", "issue:42796|issue:43650", - "issue:43295|issue:44779", + "issue:43577|issue:45216", "issue:39692|issue:43901", + "issue:43295|issue:44779", "issue:44351|issue:45278", "issue:43976|issue:45405", "issue:43299|issue:45357", - "issue:46082|issue:46512", - "issue:41950|issue:45020", "issue:44242|issue:45200", + "issue:46082|issue:46512", "issue:44779|issue:45812", - "issue:38175|issue:44617", - "issue:43299|issue:43854", - "issue:41950|issue:45823" + "issue:41950|issue:45020", + "issue:43299|issue:43854" ], - "prepared_review_unit_hash": "0e58feff5020692bba656ed5f39c7c51757b381d1eb69f4b4fa0c69ccbcb1073", + "prepared_review_unit_hash": "cf10573fdd752be7ab55a16742ffa2671e4dab251d8492b2c793594bd146eccf", "component_budget": { - "node_count": 472, - "item_count": 472, - "soft_pair_count": 1732, - "serialized_chars": 898654, - "estimated_input_tokens": 224664, - "estimated_eval_tokens": 449584 + "node_count": 473, + "item_count": 473, + "soft_pair_count": 1738, + "serialized_chars": 901437, + "estimated_input_tokens": 225360, + "estimated_eval_tokens": 450976 }, "budget": { "node_count": 48, "item_count": 48, - "soft_pair_count": 29, - "serialized_chars": 36284, - "estimated_input_tokens": 9071, - "estimated_eval_tokens": 18398 + "soft_pair_count": 31, + "serialized_chars": 36941, + "estimated_input_tokens": 9236, + "estimated_eval_tokens": 18728 }, "overflow_policy": "truncate_then_skip", "trimmed": false, "aggressively_trimmed": false, "split": true, "analyst_result": { - "summary": "These are mostly unrelated closed issue reports spanning tokenizer, loading, cache, pipeline, documentation, and dependency regressions. The soft pairs share broad subsystem or version-change themes, but none look like the same concrete bug/change, so all should stay separate.", - "confidence": 0.89, - "canonical_issue_reason": "Issue 45020 is the closest thing to an umbrella report here because it broadly covers recent-version breakage in model loading for remote_code models, but there is no true duplicate hub in this set.", + "summary": "This cluster is highly heterogeneous: most items are distinct model-, tokenizer-, cache-, or docs-related regressions, and none of the soft edges look like true duplicates. The broadest umbrella-style report is the generic upgrade/import regression issue, but it still does not unify the cluster into one underlying bug.", + "confidence": 0.78, + "canonical_issue_reason": "issue:45278 is the broadest umbrella report in the set: it frames a wide upgrade-induced import breakage rather than a narrow model-specific defect, so it is the best representative issue if one must be chosen.", "canonical_pr_reason": null, - "best_issue_reason": "Issue 45020 is the most representative anchor if one must pick a single issue: it is broad, regression-focused, and closer to the general pattern of the other reports than the narrowly scoped ones.", + "best_issue_reason": "issue:45278 is the most useful global representative because it captures the widest class of version-upgrade regressions and could plausibly subsume multiple import-related symptoms, even though the cluster is not a true duplicate set.", "best_pr_reason": null, "soft_edge_verdicts": [ + { + "left": "issue:43526", + "right": "issue:43901", + "accept": false, + "reason": "Different bugs: BeitImageProcessorFast reduce_labels behavior vs a docs mismatch about return_all_scores." + }, + { + "left": "issue:44496", + "right": "issue:45042", + "accept": false, + "reason": "Unrecognized model/config loading error is unrelated to PIL backend image processors requiring torchvision." + }, + { + "left": "issue:42907", + "right": "issue:43441", + "accept": false, + "reason": "Saving dequantized Ministral/Devstral models is a different failure from Ministral flash-attention loading." + }, + { + "left": "issue:43901", + "right": "issue:44496", + "accept": false, + "reason": "Docs wording issue vs model-loading/config recognition failure; no shared underlying code path." + }, { "left": "issue:43381", "right": "issue:45593", "accept": false, - "reason": "Different bugs: gradient checkpointing eval-mode behavior vs D-FINE auxiliary-loss handling." + "reason": "Gradient checkpointing in eval mode and missing auxiliary losses with denoising off are separate training-path issues." }, { "left": "issue:43976", "right": "issue:44279", "accept": false, - "reason": "One is a Python support/version compatibility problem; the other is a generic dependency issue with no matching failure mode." + "reason": "Python version compatibility and a dependency issue are not the same concrete bug." }, { "left": "issue:43673", "right": "issue:44336", "accept": false, - "reason": "Chunked-prefill cache plumbing and ANSI formatting in loading_report are unrelated." + "reason": "Generation cache behavior and ANSI codes in loading_report are unrelated subsystems." }, { "left": "issue:43441", "right": "issue:45588", "accept": false, - "reason": "Both mention flash attention, but they target different models and different code paths." + "reason": "Both mention flash attention, but one is a Ministral model-load failure and the other is an s_aux=None AttributeError for sink-less models." }, { - "left": "issue:44016", - "right": "issue:44062", + "left": "issue:46123", + "right": "issue:46369", "accept": false, - "reason": "Notebook syntax error is unrelated to the AddedToken keyword-argument TypeError." + "reason": "Partial-batch result dropping in MaskGenerationPipeline is unrelated to sam3-litetext memory overflow." }, { - "left": "issue:46123", - "right": "issue:46369", + "left": "issue:44016", + "right": "issue:44062", "accept": false, - "reason": "Partial-batch result dropping and SAM3 LiteText OOM are different failure classes." + "reason": "Notebook syntax error and AddedToken keyword collision are distinct failures." }, { "left": "issue:44792", "right": "issue:46424", "accept": false, - "reason": "Janus image-generation test failure is not the same as cache max-length handling." + "reason": "A janus test failure is not the same as cache max length being honored only for quantized cache." }, { "left": "issue:43824", "right": "issue:44933", "accept": false, - "reason": "Distinct import errors affecting different symbols and modules." + "reason": "Missing Qwen2.5-VL export and a nonexistent image_utils import are different import problems." }, { "left": "issue:44496", "right": "issue:45759", "accept": false, - "reason": "Both are model-loading related, but one is unrecognized config/model type and the other is missing text_config unwrapping." + "reason": "Different model families and different mechanisms: unrecognized model config vs AutoModelForCausalLM not unwrapping text_config." }, { "left": "issue:43299", "right": "issue:45216", "accept": false, - "reason": "Same model family and version era, but one is a load failure while the other is an incorrect save_pretrained checkpoint layout." + "reason": "Model loading failure and save_pretrained checkpoint corruption are separate lifecycle bugs." }, { "left": "issue:44279", "right": "issue:46515", "accept": false, - "reason": "Generic dependency issue versus a specific torch version incompatibility; no shared bug." + "reason": "Generic dependency issue and torch 2.4.1 incompatibility are not the same root cause." }, { "left": "issue:45335", "right": "issue:45584", "accept": false, - "reason": "resize_token_embeddings propagation and Whisper empty-transcription generation are separate problems." + "reason": "T5Gemma embedding resize and Whisper empty-transcription generation are unrelated." }, { "left": "issue:46082", "right": "issue:46399", "accept": false, - "reason": "LlamaConfig validation and Qwen3.5 rotary-kernel registration are unrelated." + "reason": "LlamaConfig head_dim validation is unrelated to the missing decorator causing rotary-pos-emb registration failure." }, { "left": "issue:44075", "right": "issue:44368", "accept": false, - "reason": "Optimizer argument usage bug versus a tied-word-embeddings warning in LoRA fine-tuning." + "reason": "Unused SGD args and a Qwen warning about tie_word_embeddings are unrelated." }, { "left": "issue:44279", "right": "issue:45405", "accept": false, - "reason": "Both involve dependencies, but one is vague breakage and the other is a too-new PEFT version pin." - }, - { - "left": "issue:43577", - "right": "issue:45216", - "accept": false, - "reason": "Blip2 dtype propagation and Qwen3.5 checkpoint saving are different model-specific regressions." + "reason": "A dependency issue is not the same as the MIN_PEFT_VERSION bump problem." }, { "left": "issue:42796", "right": "issue:43650", "accept": false, - "reason": "Tokenizer BPE-marker handling and a placeholder/troll issue are unrelated." - }, - { - "left": "issue:43295", - "right": "issue:44779", - "accept": false, - "reason": "Both are tokenizer regressions, but one is a processor/tokenizer API break with images and the other is incorrect DeepSeek tokenization; not the same code-path bug." - }, - { - "left": "issue:38175", - "right": "issue:44617", - "accept": false, - "reason": "SigLIP2 zero probabilities and SAM3Video OOM are unrelated model behaviors." - }, - { - "left": "issue:43299", - "right": "issue:43854", - "accept": false, - "reason": "Different model families and different loading failures." + "reason": "Tokenizer BPE-marker handling has nothing in common with the placeholder 'ADD THE DATA' issue." }, { - "left": "issue:41950", - "right": "issue:45823", + "left": "issue:43577", + "right": "issue:45216", "accept": false, - "reason": "Video-classification processor lookup and Gemma4 device placement are unrelated." + "reason": "BLIP2 dtype stale state and Qwen3.5 save_pretrained regression affect different models and code paths." }, { - "left": "issue:43854", - "right": "issue:45216", + "left": "issue:39692", + "right": "issue:43901", "accept": false, - "reason": "Unit-test load failure for GLM-4.7-Flash is not the same as Qwen3.5 save_pretrained regression." + "reason": "SigLIP2 example errors and TextClassificationPipeline docs stale behavior are both docs-adjacent but not the same bug." }, { - "left": "issue:45200", - "right": "issue:44242", + "left": "issue:43295", + "right": "issue:44779", "accept": false, - "reason": "Gemma4 token-type defaults and MoE load-balancing loss are different training-time issues." + "reason": "Custom processor/tokenizer regression is different from Deepseek tokenizer output regression." }, { "left": "issue:44351", "right": "issue:45278", "accept": false, - "reason": "A missing HybridCache import and a broad import-error report are not the same concrete bug." + "reason": "These both involve imports, but one is a specific HybridCache missing export and the other is a broad upgrade regression; not the same concrete issue." }, { - "left": "issue:44933", - "right": "issue:45278", + "left": "issue:43976", + "right": "issue:45405", "accept": false, - "reason": "Nonexistent image_utils import versus many import errors after upgrade; overlap is only superficial." + "reason": "Python support mismatch and unreleased PEFT minimum version are unrelated compatibility problems." }, { - "left": "issue:44279", - "right": "issue:46515", + "left": "issue:43299", + "right": "issue:45357", "accept": false, - "reason": "Dependency problem vs torch-specific incompatibility, with no matching root cause." + "reason": "Qwen3VL loading failure and incorrect visual encoder keys on save are distinct bugs." }, { - "left": "issue:43299", - "right": "issue:45588", + "left": "issue:44242", + "right": "issue:45200", "accept": false, - "reason": "Qwen3.5 loading regression and flash_attention s_aux=None crash affect different subsystems." + "reason": "MoE load-balancing loss and Gemma mm_token_type_ids defaults are different model behaviors." }, { - "left": "issue:43295", - "right": "issue:45216", + "left": "issue:46082", + "right": "issue:46512", "accept": false, - "reason": "Custom processor/tokenizer API regression is not the same as incorrect Qwen3.5 checkpoint serialization." + "reason": "Config head_dim validation is unrelated to from_config leaving model.config.dtype stale." }, { "left": "issue:44779", "right": "issue:45812", "accept": false, - "reason": "Different tokenizer regressions affecting different model families and failure modes." + "reason": "Both are tokenizer regressions, but for different model families with different symptoms and likely fixes." }, { - "left": "issue:44351", - "right": "issue:45278", + "left": "issue:41950", + "right": "issue:45020", + "accept": false, + "reason": "Video-classification processor lookup and remote_code model breakage are separate subsystems." + }, + { + "left": "issue:43299", + "right": "issue:43854", "accept": false, - "reason": "Specific missing symbol import versus broad multiple-import-failure report." + "reason": "Qwen3VL loading regression and GLM-4.7-Flash unit-test loading failure are different model-specific bugs." } ] }, @@ -11790,9 +11709,11 @@ "source": "llm", "cache_hit": false, "model": "gpt-5.4-mini?service_tier=flex", - "cluster_id": "cluster-10000-472", + "cluster_id": "cluster-10000-473", "nodes": [ "issue:33290", + "issue:38175", + "issue:41950", "issue:42175", "issue:42994", "issue:43317", @@ -11809,7 +11730,6 @@ "issue:43994", "issue:44016", "issue:44162", - "issue:44246", "issue:44261", "issue:44263", "issue:44291", @@ -11817,6 +11737,7 @@ "issue:44462", "issue:44466", "issue:44521", + "issue:44617", "issue:44623", "issue:44671", "issue:44683", @@ -11837,236 +11758,227 @@ "issue:45823", "issue:45834", "issue:45907", - "issue:46209", - "issue:46291", - "issue:46399" + "issue:46209" ], "soft_pairs": [ + "issue:38175|issue:44617", "issue:43479|issue:43525", + "issue:43650|issue:46209", "issue:44683|issue:45542", + "issue:41950|issue:45823", "issue:44466|issue:45276", - "issue:43650|issue:46209", "issue:44521|issue:45325", - "issue:45103|issue:45372", "issue:43653|issue:45593", + "issue:45103|issue:45372", "issue:44016|issue:44291", "issue:44462|issue:45081", "issue:43441|issue:45834", "issue:33290|issue:45656", "issue:44361|issue:45593", "issue:42994|issue:43317", - "issue:45081|issue:45907", "issue:44261|issue:44263", + "issue:45081|issue:45907", "issue:44623|issue:44749", "issue:45600|issue:45712", "issue:44623|issue:45397", "issue:42175|issue:43976", "issue:43901|issue:43994", "issue:44671|issue:44779", - "issue:45542|issue:45800", "issue:43723|issue:44162", "issue:43441|issue:44683", - "issue:43901|issue:44246", - "issue:43673|issue:43761", - "issue:44016|issue:45823", - "issue:46291|issue:46399" + "issue:45542|issue:45800", + "issue:43673|issue:43761" ], - "prepared_review_unit_hash": "8f1a90bb259e88084cd6530a9539e3eb13c68d374bd59f0ecd8820817d33786a", + "prepared_review_unit_hash": "26ca815e1475368b261259f8c7da4932ee2da058a072c5b4ed6e7a3b9b793834", "component_budget": { - "node_count": 472, - "item_count": 472, - "soft_pair_count": 1732, - "serialized_chars": 898654, - "estimated_input_tokens": 224664, - "estimated_eval_tokens": 449584 + "node_count": 473, + "item_count": 473, + "soft_pair_count": 1738, + "serialized_chars": 901437, + "estimated_input_tokens": 225360, + "estimated_eval_tokens": 450976 }, "budget": { "node_count": 48, "item_count": 48, - "soft_pair_count": 28, - "serialized_chars": 36168, - "estimated_input_tokens": 9042, - "estimated_eval_tokens": 18340 + "soft_pair_count": 27, + "serialized_chars": 35783, + "estimated_input_tokens": 8946, + "estimated_eval_tokens": 18148 }, "overflow_policy": "truncate_then_skip", "trimmed": false, "aggressively_trimmed": false, "split": true, "analyst_result": { - "summary": "The set is highly heterogeneous: most items are unrelated regressions, packaging issues, tokenizer/model bugs, docs, or CI hygiene. I did not find any soft pair that looks like the same underlying bug/change, so all candidate duplicate links should be rejected.", - "confidence": 0.83, - "canonical_issue_reason": "issue:44521 is the strongest representative landing spot: it is an active, well-described, user-facing multimodal bug with clear repro and multiple references, unlike the more incidental docs/CI or one-off model-specific reports.", + "summary": "This cluster is mostly a grab bag of unrelated Transformers bug reports across tokenizers, multimodal processors, model loading, quantization, deepspeed, and packaging. None of the soft-similarity pairs look like the same underlying bug, so I would not merge any of them as duplicates.", + "confidence": 0.54, + "canonical_issue_reason": null, "canonical_pr_reason": null, - "best_issue_reason": "issue:44521 is the best global issue target because it is open, has the most discussion/inbound attention among the relevant bugs, and is concrete enough to serve as a canonical duplicate target.", + "best_issue_reason": "Issue 44521 is the best general representative if one must pick a hub: it is active, broadly scoped around multimodal chat template masking, and has the most inbound references among the candidates. It is still not a duplicate of the other issues.", "best_pr_reason": null, "soft_edge_verdicts": [ + { + "left": "issue:38175", + "right": "issue:44617", + "accept": false, + "reason": "SigLIP2 zero-probability outputs vs Sam3Video OOM are unrelated model-specific failures." + }, { "left": "issue:43479", "right": "issue:43525", "accept": false, - "reason": "Both are config-related, but one is Phi4MultimodalConfig default handling and the other is Llama4Config missing pad_token_id; different models and different failure modes." + "reason": "Different models and different bugs: default multimodal config initialization vs missing pad_token_id on Llama4Config." + }, + { + "left": "issue:43650", + "right": "issue:46209", + "accept": false, + "reason": "Generic title only; one is an unspecified data request, the other is an FP8/B200 inference bug." }, { "left": "issue:44683", "right": "issue:45542", "accept": false, - "reason": "Unrelated subsystems: torch 2.9 flex_attention compilation vs TensorFlow backend packaging/install behavior." + "reason": "Compiled flex_attention on torch>=2.9 is unrelated to a TensorFlow backend installation issue." }, { - "left": "issue:44466", - "right": "issue:45276", + "left": "issue:41950", + "right": "issue:45823", "accept": false, - "reason": "Different code paths: lm_head.weight serialization/tied-weights behavior vs resize_token_embeddings not updating Gemma4 embedding/output layers." + "reason": "Video pipeline processor lookup and Gemma4 device_map forward mismatch are different code paths." }, { - "left": "issue:43650", - "right": "issue:46209", + "left": "issue:44466", + "right": "issue:45276", "accept": false, - "reason": "One is a vague data request, the other is an FP8/B200 inference failure; no shared bug." + "reason": "Serialization of tied lm_head weights and resize_token_embeddings not propagating to Gemma4 outputs are distinct bugs." }, { "left": "issue:44521", "right": "issue:45325", "accept": false, - "reason": "Both are multimodal, but assistant mask generation and temporal rope-index scaling are distinct bugs in different functions." + "reason": "Both are multimodal, but assistant_masks being all zero is not the same bug as Qwen2.5-VL rope index scaling." }, { - "left": "issue:45103", - "right": "issue:45372", + "left": "issue:43653", + "right": "issue:45593", "accept": false, - "reason": "auto_docstring annotation parsing crash vs Gemma4 processor import dependency failure; unrelated." + "reason": "BigBirdTokenizer special-token registration is unrelated to D-FINE auxiliary loss behavior." }, { - "left": "issue:43653", - "right": "issue:45593", + "left": "issue:45103", + "right": "issue:45372", "accept": false, - "reason": "Tokenizer special-token registration bug vs D-FINE auxiliary-loss behavior; different components and symptoms." + "reason": "auto_docstring annotation handling and Gemma4 processor loading/import errors are unrelated." }, { "left": "issue:44016", "right": "issue:44291", "accept": false, - "reason": "Notebook syntax error and init_empty_weights argument mismatch are unrelated." + "reason": "Notebook syntax error and init_empty_weights argument mismatch do not share the same failure mode." }, { "left": "issue:44462", "right": "issue:45081", "accept": false, - "reason": "Both involve tokenizers, but one is AutoTokenizer ignoring tokenizer.json and the other is a Mistral regex patch AttributeError; different root causes." + "reason": "AutoTokenizer ignoring tokenizer.json and Mistral regex patch crashing are separate tokenizer-loading problems." }, { "left": "issue:43441", "right": "issue:45834", "accept": false, - "reason": "FlashAttention incompatibility for Ministral vs long-OCR-input index error in Kosmos2.5; different models and code paths." + "reason": "FlashAttention failure for Ministral-3 and Kosmos2.5 long OCR index error are unrelated." }, { "left": "issue:33290", "right": "issue:45656", "accept": false, - "reason": "Both mention deepspeed/optimizer behavior, but AdaFactor OOM and optimizer.step being called twice are not the same concrete bug." + "reason": "Deepspeed OOM with Adafactor and optimizer.step being called twice are different optimizer/deepspeed issues." }, { "left": "issue:44361", "right": "issue:45593", "accept": false, - "reason": "Tokenizer AttributeError vs D-FINE loss logic; no underlying overlap." + "reason": "MLukeTokenizer attribute errors and D-FINE auxiliary-loss handling are unrelated." }, { "left": "issue:42994", "right": "issue:43317", "accept": false, - "reason": "Saving a quantized model and loading a dequantized model with offload are different operations and failures." + "reason": "Quantized model save failure and dequantized model load/offload failure are separate quantization pathways." }, { - "left": "issue:45081", - "right": "issue:45907", + "left": "issue:44261", + "right": "issue:44263", "accept": false, - "reason": "Tokenizer regex crash vs list_repo_files not forwarding revision; unrelated fixes." + "reason": "Missing rms_norm_eps in MLA q_a_layernorm and torch.split return values in GlmMoeDsaIndexer are not the same bug." }, { - "left": "issue:44261", - "right": "issue:44263", + "left": "issue:45081", + "right": "issue:45907", "accept": false, - "reason": "LayerNorm eps precision issue vs torch.split return handling in GlmMoeDsaIndexer; distinct bugs." + "reason": "Mistral regex patch crash and list_repo_files missing revision forwarding are unrelated utility bugs." }, { "left": "issue:44623", "right": "issue:44749", "accept": false, - "reason": "processor.save_pretrained missing files vs a slow-data-filtering report after upgrade; not the same bug." + "reason": "processor.save_pretrained missing files and a 10x slowdown after upgrade are not the same underlying issue." }, { "left": "issue:45600", "right": "issue:45712", "accept": false, - "reason": "Both are repo-maintenance issues, but stale auto_mappings references and leftover dummy classes are different fixes." + "reason": "Both are CI/repo hygiene issues, but one is stale auto_mappings references and the other is leftover dummy classes leaking into dir()." }, { "left": "issue:44623", "right": "issue:45397", "accept": false, - "reason": "Missing processor save files vs Gemma4 zero3 from_pretrained failure; unrelated." + "reason": "Missing processor save files and Gemma-4 zero3 from_pretrained failures involve different loading/saving paths." }, { "left": "issue:42175", "right": "issue:43976", "accept": false, - "reason": "TensorFlow backend missing from a torch install and Python version compatibility are separate packaging issues." + "reason": "TensorFlow backend installation metadata and Python 3.9/3.10 compatibility are unrelated packaging problems." }, { "left": "issue:43901", "right": "issue:43994", "accept": false, - "reason": "Docs mentioning return_all_scores vs a model producing nonsensical outputs are not the same bug." + "reason": "Documentation mismatch for return_all_scores is not the same as incorrect SigLIP2 model/pipeline outputs." }, { "left": "issue:44671", "right": "issue:44779", "accept": false, - "reason": "CamemBERT masked-LM regression and DeepSeek tokenizer regression are different model/tokenizer paths." - }, - { - "left": "issue:45542", - "right": "issue:45800", - "accept": false, - "reason": "Missing tf backend from install vs torch/transformers version incompatibility; different root causes." + "reason": "CamemBERT masked-LM regression and DeepSeek tokenizer regression are different components and failure modes." }, { "left": "issue:43723", "right": "issue:44162", "accept": false, - "reason": "AutoTokenizer loading issue and ESM2 breakage are unrelated." + "reason": "AutoTokenizer.from_pretrained loading regression and ESM2 broken workflow report are not the same bug." }, { "left": "issue:43441", "right": "issue:44683", "accept": false, - "reason": "FlashAttention runtime failure and flex_attention compile failure are both attention-related but not the same bug." + "reason": "FlashAttention model failure and compiled flex_attention torch>=2.9 incompatibility are distinct attention-stack issues." }, { - "left": "issue:43901", - "right": "issue:44246", + "left": "issue:45542", + "right": "issue:45800", "accept": false, - "reason": "Docs mismatch for a pipeline parameter vs intermittent import slowness; unrelated." + "reason": "Missing TensorFlow backend from extras and torch/transformers version incompatibility are different packaging/runtime problems." }, { "left": "issue:43673", "right": "issue:43761", "accept": false, - "reason": "Generation cache missing during chunked prefill vs CLIPVisionModel hidden_states regression; different components." - }, - { - "left": "issue:44016", - "right": "issue:45823", - "accept": false, - "reason": "Notebook syntax error vs Gemma4 device mismatch in forward; unrelated." - }, - { - "left": "issue:46291", - "right": "issue:46399", - "accept": false, - "reason": "pipelines import failure due kernels version/revision requirements is not the same as a missing decorator on apply_rotary_pos_emb registration." + "reason": "Generation cache missing in v5 chunked prefill and CLIPVisionModel hidden_states regression are unrelated model-specific bugs." } ] }, @@ -12088,7 +12000,7 @@ "source": "llm", "cache_hit": false, "model": "gpt-5.4-mini?service_tier=flex", - "cluster_id": "cluster-10000-472", + "cluster_id": "cluster-10000-473", "nodes": [ "issue:36246", "issue:41628", @@ -12100,7 +12012,6 @@ "issue:43525", "issue:43526", "issue:43531", - "issue:43540", "issue:43550", "issue:43643", "issue:43784", @@ -12108,6 +12019,7 @@ "issue:43881", "issue:43901", "issue:43976", + "issue:44016", "issue:44188", "issue:44220", "issue:44242", @@ -12120,26 +12032,28 @@ "issue:44485", "issue:44496", "issue:44589", - "issue:44734", - "issue:44779", "issue:44871", "issue:45003", "issue:45103", - "issue:45292", "issue:45362", "issue:45397", "issue:45542", "issue:45561", "issue:45563", "issue:45593", - "issue:45859", + "issue:45823", "issue:46132", + "issue:46291", "issue:46399", "issue:46424", "issue:46506", - "issue:46655" + "issue:46614", + "issue:46665" ], "soft_pairs": [ + "issue:43901|issue:44246", + "issue:46291|issue:46399", + "issue:44016|issue:45823", "issue:43525|issue:43881", "issue:43824|issue:43976", "issue:43525|issue:43643", @@ -12147,9 +12061,10 @@ "issue:44188|issue:44336", "issue:44297|issue:45397", "issue:44336|issue:44373", - "issue:41628|issue:43976", "issue:43784|issue:45542", + "issue:41628|issue:43976", "issue:44246|issue:44496", + "issue:46614|issue:46665", "issue:45561|issue:45563", "issue:45103|issue:46399", "issue:44242|issue:45593", @@ -12160,232 +12075,208 @@ "issue:44246|issue:44871", "issue:44336|issue:44485", "issue:43550|issue:46506", - "issue:41950|issue:44220", "issue:42886|issue:46132", - "issue:36246|issue:42757", "issue:44336|issue:46424", - "issue:36246|issue:44351", - "issue:43540|issue:46655", - "issue:44779|issue:45292", - "issue:45859|issue:46399", - "issue:45003|issue:45561", - "issue:44734|issue:46424", - "issue:43901|issue:45003" + "issue:41950|issue:44220", + "issue:36246|issue:42757", + "issue:36246|issue:44351" ], - "prepared_review_unit_hash": "2ff45df245ba9102dbbfff03da7020ec0c5181fd5c9f398fd5ca2ce0bdcfb849", + "prepared_review_unit_hash": "a07c6c4307f89afa1c83e6da20f9174780e6e453f2616a46cb0670c485df4881", "component_budget": { - "node_count": 472, - "item_count": 472, - "soft_pair_count": 1732, - "serialized_chars": 898654, - "estimated_input_tokens": 224664, - "estimated_eval_tokens": 449584 + "node_count": 473, + "item_count": 473, + "soft_pair_count": 1738, + "serialized_chars": 901437, + "estimated_input_tokens": 225360, + "estimated_eval_tokens": 450976 }, "budget": { - "node_count": 48, - "item_count": 48, - "soft_pair_count": 31, - "serialized_chars": 36804, - "estimated_input_tokens": 9201, - "estimated_eval_tokens": 18658 + "node_count": 47, + "item_count": 47, + "soft_pair_count": 29, + "serialized_chars": 35729, + "estimated_input_tokens": 8933, + "estimated_eval_tokens": 18122 }, "overflow_policy": "truncate_then_skip", "trimmed": false, "aggressively_trimmed": false, "split": true, "analyst_result": { - "summary": "This cluster is largely heterogeneous: most items are unrelated Transformers regressions spanning imports, config/loading, tokenizer behavior, cache handling, and model-specific runtime bugs. None of the soft pairs look like true duplicates of the same underlying bug or change.", - "confidence": 0.91, - "canonical_issue_reason": "issue:43824 is the strongest representative of the cluster\u2019s most common user-facing theme: a concrete import/export regression in the Transformers package with broad impact and multiple references.", + "summary": "This cluster is heterogeneous: most items are unrelated import errors, model-loading regressions, docs problems, or performance/behavior bugs. No soft pair is a safe duplicate. The best representative issue is the Qwen2.5-VL import/export regression.", + "confidence": 0.88, + "canonical_issue_reason": "issue:43824 is the clearest and most concrete user-facing import regression, with stronger signal than the other heterogeneous issues.", "canonical_pr_reason": null, - "best_issue_reason": "issue:43824 is the best single issue to anchor the cluster because it is a clear, high-signal package-level import failure rather than a narrow model-specific symptom.", + "best_issue_reason": "issue:43824 has the most directly actionable reproduction and the strongest external impact among the issues here.", "best_pr_reason": null, "soft_edge_verdicts": [ + { + "left": "issue:43901", + "right": "issue:44246", + "accept": false, + "reason": "Docs mismatch vs intermittent import-time slowdown; different problems." + }, + { + "left": "issue:46291", + "right": "issue:46399", + "accept": false, + "reason": "Both involve kernels, but one is a pipeline import failure and the other is a Qwen3.5 registration crash." + }, + { + "left": "issue:44016", + "right": "issue:45823", + "accept": false, + "reason": "Notebook syntax error vs Gemma4 device placement mismatch." + }, { "left": "issue:43525", "right": "issue:43881", "accept": false, - "reason": "Both are model-loading failures, but they involve different models and different missing/invalid config fields." + "reason": "Different model families and different loading failures." }, { "left": "issue:43824", "right": "issue:43976", "accept": false, - "reason": "One is a missing symbol import; the other is a Python-version compatibility regression." + "reason": "Specific missing export vs Python version compatibility breakage." }, { "left": "issue:43525", "right": "issue:43643", "accept": false, - "reason": "Both involve loading/config issues, but one is a model-specific missing field and the other is a general trust_remote_code return-object bug." + "reason": "Pad token attribute error vs missing config fields from trust_remote_code." }, { "left": "issue:44589", "right": "issue:45362", "accept": false, - "reason": "Different failure modes and different model families; no shared code path is evident." + "reason": "Unrelated storage-object type error vs Qwen chat crash." }, { "left": "issue:44188", "right": "issue:44336", "accept": false, - "reason": "Attention-kernel divergence and ANSI output in loading_report are unrelated." + "reason": "torch.compile kernel divergence vs ANSI output formatting bug." }, { "left": "issue:44297", "right": "issue:45397", "accept": false, - "reason": "Tokenizer save metadata mismatch is unrelated to a zero3 from_pretrained loading failure." + "reason": "Tokenizer save metadata mismatch vs ZeRO/from_pretrained failure." }, { "left": "issue:44336", "right": "issue:44373", "accept": false, - "reason": "Terminal ANSI formatting and a docstring typo are unrelated documentation/utility issues." + "reason": "Terminal ANSI leakage vs wrong docstring." + }, + { + "left": "issue:43784", + "right": "issue:45542", + "accept": false, + "reason": "Different undefined-backend/import issues in different libraries." }, { "left": "issue:41628", "right": "issue:43976", "accept": false, - "reason": "Different regressions: a missing AutoImageProcessor import versus Python 3.9/3.10 support." + "reason": "Missing import symbol vs Python-version support issue." }, { - "left": "issue:43784", - "right": "issue:45542", + "left": "issue:44246", + "right": "issue:44496", "accept": false, - "reason": "Importing sentence-transformers failing on nn is unrelated to TensorBoard/TensorFlow backend detection." + "reason": "Import performance complaint vs unrecognized model config." }, { - "left": "issue:44246", - "right": "issue:44242", + "left": "issue:46614", + "right": "issue:46665", "accept": false, - "reason": "Import-time slowdown and MoE load-balancing loss are unrelated." + "reason": "Stale phonemizer language state vs from_pretrained class behavior." }, { "left": "issue:45561", "right": "issue:45563", "accept": false, - "reason": "A pytest-xdist file race is unrelated to a stale generation warning." + "reason": "Parallel test file race vs stale generation warning." }, { "left": "issue:45103", "right": "issue:46399", "accept": false, - "reason": "auto_docstring annotation handling and Qwen3.5 rotary-kernel registration are unrelated." + "reason": "Docstring helper crash vs missing decorator registration bug." }, { "left": "issue:44242", - "right": "issue:44871", + "right": "issue:45593", "accept": false, - "reason": "Missing auxiliary loss and Gemma eos_token_id mismatch are different model behavior bugs." + "reason": "Different missing-loss conditions in different models/features." }, { "left": "issue:42491", "right": "issue:43299", "accept": false, - "reason": "Both mention Qwen/MoE loading across versions, but they describe different breakages in different model variants." + "reason": "Related Qwen MoE area, but distinct incompatibility/loading failures." }, { "left": "issue:44295", "right": "issue:45397", "accept": false, - "reason": "A position_ids buffer access error is unrelated to a zero3 model loading failure." + "reason": "Buffer access bug vs ZeRO/from_pretrained issue." }, { "left": "issue:43526", "right": "issue:45003", "accept": false, - "reason": "BeitImageProcessorFast label reduction and unsafe sys.modules access are unrelated." + "reason": "BeitImageProcessor label reduction bug vs unsafe sys.modules access." }, { "left": "issue:43531", "right": "issue:45362", "accept": false, - "reason": "Both are Qwen-related, but one is a sliding_window bug and the other is a chat-loading crash." + "reason": "Sliding-window bug vs chat runtime crash." }, { "left": "issue:44246", "right": "issue:44871", "accept": false, - "reason": "Import performance and Gemma eos_token configuration are unrelated." + "reason": "Import slowdown vs Gemma eos_token_id inconsistency." }, { "left": "issue:44336", "right": "issue:44485", "accept": false, - "reason": "ANSI formatting output and GLM-5 RoPE implementation are unrelated." + "reason": "ANSI code emission vs GLM-5 RoPE implementation discussion." }, { "left": "issue:43550", "right": "issue:46506", "accept": false, - "reason": "torch.compile SDPA failure for Bamba and CPU/CUDA differences for PPDocLayoutV3 are different bugs." - }, - { - "left": "issue:41950", - "right": "issue:44220", - "accept": false, - "reason": "Video-classification pipeline image-processor lookup and fbank feature extraction are unrelated audio/video processing issues." + "reason": "Different model/runtime divergences in unrelated code paths." }, { "left": "issue:42886", "right": "issue:46132", "accept": false, - "reason": "Offline tokenizer cache loading and AttentionInterface registration behavior are unrelated." + "reason": "Offline cache loading bug vs AttentionInterface behavior change." }, { - "left": "issue:36246", - "right": "issue:42757", + "left": "issue:41950", + "right": "issue:44220", "accept": false, - "reason": "Missing Qwen2.5-VL symbol import and missing is_offline_mode import are different API regressions." + "reason": "Video pipeline choosing image processors vs audio feature extraction bug." }, { - "left": "issue:44336", - "right": "issue:46424", + "left": "issue:36246", + "right": "issue:42757", "accept": false, - "reason": "ANSI output in loading_report and cache-length handling are unrelated." + "reason": "Different missing imports from different packages." }, { "left": "issue:36246", "right": "issue:44351", "accept": false, - "reason": "Both are import errors, but they concern different exported symbols and different code paths." - }, - { - "left": "issue:43540", - "right": "issue:46655", - "accept": false, - "reason": "Different video-processor bugs for different models; not the same underlying defect." - }, - { - "left": "issue:44779", - "right": "issue:45292", - "accept": false, - "reason": "Tokenizer correctness regression and resize_token_embeddings not updating output embeddings are distinct issues." - }, - { - "left": "issue:45859", - "right": "issue:46399", - "accept": false, - "reason": "Both involve Qwen3.5/MoE internals, but one is CPU-offload incompatibility and the other is kernel registration failure." - }, - { - "left": "issue:45003", - "right": "issue:45561", - "accept": false, - "reason": "Unsafe sys.modules access and an xdist file race are unrelated." - }, - { - "left": "issue:44734", - "right": "issue:46424", - "accept": false, - "reason": "KV-cache continuation indexing and max_cache_length handling are different cache bugs." - }, - { - "left": "issue:43901", - "right": "issue:45003", - "accept": false, - "reason": "A docs mismatch about return_all_scores is unrelated to sys.modules access." + "reason": "Different missing symbols; same broad symptom but not the same bug." } ] }, @@ -12407,7 +12298,7 @@ "source": "llm", "cache_hit": false, "model": "gpt-5.4-mini?service_tier=flex", - "cluster_id": "cluster-10000-472", + "cluster_id": "cluster-10000-473", "nodes": [ "issue:36246", "issue:43257", @@ -12416,8 +12307,8 @@ "issue:43475", "issue:43540", "issue:43596", - "issue:43653", "issue:43673", + "issue:43901", "issue:44038", "issue:44060", "issue:44077", @@ -12425,17 +12316,16 @@ "issue:44242", "issue:44246", "issue:44295", - "issue:44368", "issue:44466", "issue:44479", "issue:44496", "issue:44561", + "issue:44734", "issue:44749", + "issue:44779", "issue:44857", "issue:44877", - "issue:44977", "issue:44987", - "issue:44995", "issue:45003", "issue:45137", "issue:45200", @@ -12443,241 +12333,255 @@ "issue:45292", "issue:45335", "issue:45375", - "issue:45397", "issue:45412", "issue:45459", + "issue:45561", "issue:45584", "issue:45600", - "issue:45656", "issue:45701", "issue:45759", + "issue:45859", "issue:46123", "issue:46170", "issue:46369", "issue:46399", + "issue:46424", "issue:46506", "issue:46655" ], "soft_pairs": [ + "issue:43540|issue:46655", + "issue:44779|issue:45292", + "issue:45859|issue:46399", + "issue:44734|issue:46424", + "issue:45003|issue:45561", + "issue:43901|issue:45003", "issue:44479|issue:46655", "issue:43257|issue:45137", - "issue:44496|issue:45003", "issue:44242|issue:44857", + "issue:44496|issue:45003", "issue:43475|issue:46369", "issue:43673|issue:44561", "issue:44295|issue:44749", "issue:44162|issue:45200", - "issue:36246|issue:43596", "issue:43262|issue:44246", "issue:45375|issue:45759", + "issue:36246|issue:43596", "issue:44877|issue:45375", "issue:45292|issue:45584", "issue:45137|issue:45600", "issue:45292|issue:45701", - "issue:44038|issue:46506", "issue:44561|issue:46399", + "issue:44038|issue:46506", "issue:44466|issue:45335", "issue:44987|issue:45459", "issue:45412|issue:46123", "issue:43352|issue:45250", "issue:44479|issue:46170", - "issue:44060|issue:44077", - "issue:44995|issue:45200", - "issue:44368|issue:44977", - "issue:43475|issue:43540", - "issue:45397|issue:45656", - "issue:43653|issue:44162" + "issue:44060|issue:44077" ], - "prepared_review_unit_hash": "cde89b2943c0fcc29de402e01d9ee0250bf5daf1bbe02e8bde19ca953e750c9e", + "prepared_review_unit_hash": "3a053d07e4676a08851da581de4bf501ff76314c73f7e1e23662c1b2f87521ad", "component_budget": { - "node_count": 472, - "item_count": 472, - "soft_pair_count": 1732, - "serialized_chars": 898654, - "estimated_input_tokens": 224664, - "estimated_eval_tokens": 449584 + "node_count": 473, + "item_count": 473, + "soft_pair_count": 1738, + "serialized_chars": 901437, + "estimated_input_tokens": 225360, + "estimated_eval_tokens": 450976 }, "budget": { "node_count": 48, "item_count": 48, - "soft_pair_count": 28, - "serialized_chars": 36180, - "estimated_input_tokens": 9045, - "estimated_eval_tokens": 18346 + "soft_pair_count": 29, + "serialized_chars": 36471, + "estimated_input_tokens": 9118, + "estimated_eval_tokens": 18492 }, "overflow_policy": "truncate_then_skip", "trimmed": false, "aggressively_trimmed": false, "split": true, "analyst_result": { - "summary": "This cluster is highly heterogeneous: the items span unrelated model-loading, multimodal, tokenizer, DeepSpeed, and performance issues. I found no soft-edge pair that looks like the same underlying bug/change, so all candidate links should be rejected.", - "confidence": 0.95, - "canonical_issue_reason": "issue:44479 is the strongest single anchor available: it is a concrete regression report with a clear failure mode and broad model-family scope. That said, the cluster as a whole does not appear cohesive, so this is only a pragmatic representative choice.", + "summary": "This cluster is heterogeneous: the soft pairs mostly share broad keywords (video, cache, Qwen, DeepSpeed, Flash Attention) but describe different code paths, models, or symptoms. I do not see any true duplicate pair here.", + "confidence": 0.93, + "canonical_issue_reason": null, "canonical_pr_reason": null, - "best_issue_reason": "issue:44479 is the most representative issue for triage because it is specific, user-facing, and framed as a regression affecting multiple Qwen model variants. It is the best available cluster representative even though the rest of the set is not a true duplicate group.", + "best_issue_reason": null, "best_pr_reason": null, "soft_edge_verdicts": [ { - "left": "issue:44479", + "left": "issue:43540", "right": "issue:46655", "accept": false, - "reason": "Both are video-related, but they concern different model families and different failure modes: Qwen video regression vs SmolVLM padding calculation." + "reason": "Both involve video inputs, but one is a Qwen3OmniMoe processing ValueError and the other is a SmolVLM padding bug; different models and likely different fixes." }, { - "left": "issue:43257", - "right": "issue:45137", + "left": "issue:44779", + "right": "issue:45292", "accept": false, - "reason": "Both involve DeepSpeed/ZeRO3, but one is about Qwen3 MOE weight conversion and the other is an empty-deque IndexError in a different code path." + "reason": "Tokenizer output regressions and resize_token_embeddings/output_embeddings mismatch are unrelated behaviors." }, { - "left": "issue:44496", + "left": "issue:45859", + "right": "issue:46399", + "accept": false, + "reason": "Both are Qwen3.5-related, but one is CPU offload compatibility in rotary embedding and the other is a missing decorator causing kernel registration failure." + }, + { + "left": "issue:44734", + "right": "issue:46424", + "accept": false, + "reason": "KV-cache continuation crash and max_cache_length handling for quantized cache are separate cache issues." + }, + { + "left": "issue:45003", + "right": "issue:45561", + "accept": false, + "reason": "sys.modules access in modeling_utils and xdist race on captured_info.txt are unrelated." + }, + { + "left": "issue:43901", "right": "issue:45003", "accept": false, - "reason": "These are both loading/remote-code style problems, but the concrete failures differ: unrecognized model config vs unsafe sys.modules access." + "reason": "A docs mismatch for TextClassificationPipeline is not the same bug as unsafe sys.modules access." + }, + { + "left": "issue:44479", + "right": "issue:46655", + "accept": false, + "reason": "Both mention video regressions, but they affect different model families and different failures." + }, + { + "left": "issue:43257", + "right": "issue:45137", + "accept": false, + "reason": "Both involve DeepSpeed, but one is Qwen3 MOE weight conversion during loading and the other is a ZeRO3 deque IndexError." }, { "left": "issue:44242", "right": "issue:44857", "accept": false, - "reason": "Different subsystems and bugs: router load-balancing loss handling vs an AMP/CUDA crash in LwDetrImageLoss." + "reason": "Load balancing loss gating and an AMP/CUDA crash in LwDetrImageLoss are different bugs." + }, + { + "left": "issue:44496", + "right": "issue:45003", + "accept": false, + "reason": "Unrecognized model config and sys.modules access are unrelated." }, { "left": "issue:43475", "right": "issue:46369", "accept": false, - "reason": "Both touch SAM 3, but one is a missing vision output attribute and the other is a memory overflow in sam3-litetext." + "reason": "SAM3 vision encoder missing an attribute and SAM3 lite text memory overflow are distinct failure modes." }, { "left": "issue:43673", "right": "issue:44561", "accept": false, - "reason": "One is a generation cache regression during chunked prefill; the other is a removed helper breaking trust_remote_code models." + "reason": "Chunked-prefill cache missing and removal of is_torch_fx_available breaking trust_remote_code models are unrelated." }, { "left": "issue:44295", "right": "issue:44749", "accept": false, - "reason": "A buffer/position_ids error is not the same bug as a general tokenization slowdown after upgrade." + "reason": "Position_ids buffer access error and a cross-version tokenization slowdown are not the same bug." }, { "left": "issue:44162", "right": "issue:45200", "accept": false, - "reason": "Completely different models and failures: ESM2 breakage vs Gemma 4 needing mm_token_type_ids defaults." - }, - { - "left": "issue:36246", - "right": "issue:43596", - "accept": false, - "reason": "Qwen2.5 VL import error and BertModel+ZeRO3 index error are unrelated code paths." + "reason": "ESM2 breakage and Gemma 4 mm_token_type_ids defaulting are unrelated model-specific issues." }, { "left": "issue:43262", "right": "issue:44246", "accept": false, - "reason": "Audio chat-template sampling-rate default bug is unrelated to import latency in transformers." + "reason": "Audio chat-template sample-rate defaulting and slow import are unrelated." }, { "left": "issue:45375", "right": "issue:45759", "accept": false, - "reason": "Both concern Qwen multimodal configs, but one is a missing field being dropped by strict config handling and the other is AutoModelForCausalLM not unwrapping text_config." + "reason": "Missing config field silently dropped by strict and AutoModelForCausalLM text_config unwrapping are different multimodal config problems." + }, + { + "left": "issue:36246", + "right": "issue:43596", + "accept": false, + "reason": "An import error for Qwen2.5-VL image processor and a ZeRO3/BertModel index error are unrelated." }, { "left": "issue:44877", "right": "issue:45375", "accept": false, - "reason": "Both are strict-config related, but they affect different model configs and different missing-field behaviors." + "reason": "Strict granite_speech config loading and Qwen3.5 Moe vision config field handling are separate config issues." }, { "left": "issue:45292", "right": "issue:45584", "accept": false, - "reason": "resize_token_embeddings affecting output embeddings is unrelated to Whisper failing after align_special_tokens on empty transcription." + "reason": "Resize-token-embeddings behavior and Whisper empty transcription generation are different code paths." }, { "left": "issue:45137", "right": "issue:45600", "accept": false, - "reason": "DeepSpeed ZeRO3 deque/index failure and stale auto_mappings references are different bugs." + "reason": "ZeRO3 deque failure and stale auto_mappings CI references are unrelated." }, { "left": "issue:45292", "right": "issue:45701", "accept": false, - "reason": "A token-embedding resize bug is not the same as version-dependent tokenization changes." + "reason": "Embedding resize behavior and tokenizer-version changes are not the same defect." }, { - "left": "issue:44038", - "right": "issue:46506", + "left": "issue:44561", + "right": "issue:46399", "accept": false, - "reason": "These are different model/runtime issues: a Qwen3-VL-Moe bug vs CPU/CUDA output divergence in PPDocLayoutV3." + "reason": "trust_remote_code breakage from a removed helper and Qwen3.5 rotary-kernel registration failure are unrelated." }, { - "left": "issue:44561", - "right": "issue:46399", + "left": "issue:44038", + "right": "issue:46506", "accept": false, - "reason": "Different concrete regressions: removed utility breaking remote-code models vs a missing decorator registration for Qwen3.5 kernels." + "reason": "Qwen3-VL-Moe regression and PPDocLayoutV3 CPU/CUDA divergence are separate model-specific bugs." }, { "left": "issue:44466", "right": "issue:45335", "accept": false, - "reason": "Both mention tied weights / embeddings, but they are distinct model-specific serialization and resize behavior bugs." + "reason": "Both touch tied/embedded weights, but one is device-dependent serialization and the other is resize_token_embeddings not updating decoder.embed_tokens." }, { "left": "issue:44987", "right": "issue:45459", "accept": false, - "reason": "Loading failure on transformers>=5.1 and tokenizer protobuf error masking are different failure modes." + "reason": "Loading physical-intelligence/fast and hidden tokenizer decode errors due to missing protobuf are unrelated." }, { "left": "issue:45412", "right": "issue:46123", "accept": false, - "reason": "A memory leak on deletion is unrelated to MaskGenerationPipeline dropping the final partial batch." + "reason": "RT-DETR memory release and MaskGenerationPipeline partial-batch handling are unrelated." }, { "left": "issue:43352", "right": "issue:45250", "accept": false, - "reason": "One reports NemotronHForCausalLM lacking Flash Attention 2 support; the other is only a generic Flash Attention topic." + "reason": "A specific model lacking Flash Attention 2.0 support is not the same as a generic Flash Attention 2.0 issue." }, { "left": "issue:44479", "right": "issue:46170", "accept": false, - "reason": "Video-input regression in Qwen models and Windows/ROCm torch.distributed import issues are unrelated." + "reason": "Video-input regressions and Windows/ROCm torch.distributed import errors are unrelated." }, { "left": "issue:44060", "right": "issue:44077", "accept": false, - "reason": "Incorrect tied-weights warning and patchtsmixer post_init policy are separate issues in different components." - }, - { - "left": "issue:44995", - "right": "issue:45200", - "accept": false, - "reason": "Stale indexer-cache crash in GlmMoeDsa is unrelated to Gemma 4 mm_token_type_ids defaults." - }, - { - "left": "issue:44368", - "right": "issue:44977", - "accept": false, - "reason": "A tied-embeddings warning is not the same as a flash-attention generation regression." - }, - { - "left": "issue:45397", - "right": "issue:45656", - "accept": false, - "reason": "Gemma-4 ZeRO3 from_pretrained loading and double optimizer stepping under DeepSpeed are different bugs." - }, - { - "left": "issue:43653", - "right": "issue:44162", - "accept": false, - "reason": "BigBirdTokenizer special-token registration and ESM2 breakage do not share the same underlying defect." + "reason": "Incorrect tied-weights warning in Qwen3-Next and patchtsmixer post_init allowance are unrelated." } ] }, @@ -12699,11 +12603,8 @@ "source": "llm", "cache_hit": false, "model": "gpt-5.4-mini?service_tier=flex", - "cluster_id": "cluster-10000-472", + "cluster_id": "cluster-10000-473", "nodes": [ - "issue:28282", - "issue:30990", - "issue:41628", "issue:42886", "issue:42994", "issue:43065", @@ -12711,8 +12612,10 @@ "issue:43408", "issue:43475", "issue:43482", + "issue:43540", "issue:43643", "issue:43646", + "issue:43653", "issue:43673", "issue:43906", "issue:43931", @@ -12727,8 +12630,9 @@ "issue:44869", "issue:44871", "issue:44877", - "issue:44933", + "issue:44977", "issue:44995", + "issue:45200", "issue:45229", "issue:45230", "issue:45310", @@ -12751,13 +12655,18 @@ "issue:46431" ], "soft_pairs": [ + "issue:44368|issue:44977", + "issue:44995|issue:45200", + "issue:45397|issue:45656", + "issue:43653|issue:44162", + "issue:43475|issue:43540", "issue:42886|issue:45229", "issue:44877|issue:46170", "issue:44995|issue:45588", - "issue:43065|issue:43408", "issue:44368|issue:45563", - "issue:43931|issue:44368", + "issue:43065|issue:43408", "issue:43906|issue:46202", + "issue:43931|issue:44368", "issue:44188|issue:46202", "issue:45685|issue:46431", "issue:44483|issue:46202", @@ -12768,231 +12677,244 @@ "issue:44297|issue:44623", "issue:44361|issue:44749", "issue:44749|issue:45656", - "issue:43646|issue:44162", - "issue:44749|issue:45310", "issue:44623|issue:45656", + "issue:44749|issue:45310", + "issue:43646|issue:44162", "issue:44871|issue:46132", "issue:43232|issue:43673", "issue:42994|issue:43482", - "issue:44869|issue:45356", "issue:45230|issue:45397", - "issue:43475|issue:43643", + "issue:44869|issue:45356", "issue:45375|issue:45478", - "issue:28282|issue:30990", - "issue:41628|issue:44933" + "issue:43475|issue:43643" ], - "prepared_review_unit_hash": "6424437015330c5283ac677a28c56ea0687045e14158b684f1df52500377061e", + "prepared_review_unit_hash": "baec19de04e21490adacfef6a25f72ea332124119953b10f032f7394b3430d09", "component_budget": { - "node_count": 472, - "item_count": 472, - "soft_pair_count": 1732, - "serialized_chars": 898654, - "estimated_input_tokens": 224664, - "estimated_eval_tokens": 449584 + "node_count": 473, + "item_count": 473, + "soft_pair_count": 1738, + "serialized_chars": 901437, + "estimated_input_tokens": 225360, + "estimated_eval_tokens": 450976 }, "budget": { "node_count": 48, "item_count": 48, - "soft_pair_count": 29, - "serialized_chars": 36261, - "estimated_input_tokens": 9066, - "estimated_eval_tokens": 18388 + "soft_pair_count": 32, + "serialized_chars": 37218, + "estimated_input_tokens": 9305, + "estimated_eval_tokens": 18866 }, "overflow_policy": "truncate_then_skip", "trimmed": false, "aggressively_trimmed": false, "split": true, "analyst_result": { - "summary": "The cluster is mostly heterogeneous. The only clear duplicate-like match is the pair of identically titled Qwen3.5 MoE from_pretrained error issues (45375 and 45478); the other soft pairs share only broad subsystem or symptom overlap and should stay separate.", - "confidence": 0.93, - "canonical_issue_reason": "issue:45310 is the earlier, cleaner representative of the Qwen3.5 MoE from_pretrained error pair, and its title matches the duplicate issue:45478 closely enough to anchor the cluster.", + "summary": "This is a heterogeneous set of Transformers-related bug reports; the soft pairs are mostly related only at the subsystem or release level, not the same concrete bug. I found no safe duplicates among the provided soft edges.", + "confidence": 0.78, + "canonical_issue_reason": "Issue 43646 is the broadest upstream regression report in the set and the best representative umbrella for the surrounding Transformers 5.x breakage reports, even though it is not an exact duplicate of them.", "canonical_pr_reason": null, - "best_issue_reason": "issue:45310 is the strongest overall issue target because it is the clearest canonical duplicate for the near-identical Qwen3.5 MoE from_pretrained reports.", + "best_issue_reason": "43646 is the most general issue here and the best single representative for the cluster because it covers broad custom-model initialization regressions rather than a narrow model-specific symptom.", "best_pr_reason": null, "soft_edge_verdicts": [ + { + "left": "issue:44368", + "right": "issue:44977", + "accept": false, + "reason": "Different problems: a warning about tie_word_embeddings on Qwen3.5 LoRA vs a flash-attention generation failure." + }, + { + "left": "issue:44995", + "right": "issue:45200", + "accept": false, + "reason": "Both touch model behavior, but one is a stale indexer cache crash and the other is a Gemma token-type defaulting issue." + }, + { + "left": "issue:45397", + "right": "issue:45656", + "accept": false, + "reason": "Gemma-4 ZeRO3 loading vs deepspeed optimizer stepping twice are unrelated code paths and failures." + }, + { + "left": "issue:43653", + "right": "issue:44162", + "accept": false, + "reason": "Tokenizer special-token registration and ESM2 breakage are different model/tokenizer issues." + }, + { + "left": "issue:43475", + "right": "issue:43540", + "accept": false, + "reason": "A missing SAM3 vision encoder attribute is not the same as a Qwen3OmniMoe video-input validation error." + }, { "left": "issue:42886", "right": "issue:45229", "accept": false, - "reason": "Offline tokenizer cache loading and Gemma4 multi-GPU CUDA OOM are different failure modes." + "reason": "Offline tokenizer cache loading and Gemma4 multi-GPU OOM are clearly different bugs." }, { "left": "issue:44877", "right": "issue:46170", "accept": false, - "reason": "Strict config loading for granite_speech is unrelated to the Windows ROCm torch.distributed import/dependency error." + "reason": "Strict config loading for granite_speech and a Windows/ROCm torch.distributed import problem are unrelated." }, { "left": "issue:44995", "right": "issue:45588", "accept": false, - "reason": "A stale indexer cache crash and a flash_attention s_aux=None AttributeError are different code-path bugs." - }, - { - "left": "issue:43065", - "right": "issue:43408", - "accept": false, - "reason": "Both involve SAM3, but one is a dummy Conv2d implementation detail and the other is a model-type mismatch warning." + "reason": "Stale indexer cache on second forward pass is not the same as flash_attention crashing on s_aux=None." }, { "left": "issue:44368", "right": "issue:45563", "accept": false, - "reason": "These are both warnings, but they describe different behaviors and different underlying issues." + "reason": "A tie_word_embeddings warning is unrelated to a stale num_return_sequences warning in paged generate." }, { - "left": "issue:43931", - "right": "issue:44368", + "left": "issue:43065", + "right": "issue:43408", "accept": false, - "reason": "Qwen3-VL weight-shape mismatch is not the same bug as a tie_word_embeddings warning." + "reason": "One is a dummy Conv2d in Sam3PixelDecoder; the other is a model-type mismatch warning." }, { "left": "issue:43906", "right": "issue:46202", "accept": false, - "reason": "An isolated reproduction of another issue is not the same as a Gemma pipeline roundtrip failure." + "reason": "An isolated reproduction of an older issue is not the same as Gemma pipeline roundtrip failure on \u2581." + }, + { + "left": "issue:43931", + "right": "issue:44368", + "accept": false, + "reason": "Qwen3-VL weight-shape mismatch and a Qwen3.5 tie_word_embeddings warning are different failures." }, { "left": "issue:44188", "right": "issue:46202", "accept": false, - "reason": "Torch.compile attention-kernel branching and Gemma token roundtrip behavior are unrelated code paths." + "reason": "torch.compile branching causing attention kernel divergence is unrelated to Gemma pipeline roundtrip behavior." }, { "left": "issue:45685", "right": "issue:46431", "accept": false, - "reason": "An MPS histogram type error and missing chat template handling are different problems." + "reason": "MPS histogram support on Int and missing chat template errors are unrelated." }, { "left": "issue:44483", "right": "issue:46202", "accept": false, - "reason": "API request acceptance in /v1/chat/completions is unrelated to the Gemma pipeline roundtrip bug." + "reason": "A v5.3 chat/completions request rejection is not the same as a Gemma pipeline roundtrip tokenizer bug." }, { "left": "issue:45987", "right": "issue:46082", "accept": false, - "reason": "StaticCache return-type contract and LlamaConfig head_dim validation are different bugs." + "reason": "StaticCache seq_length returning a tensor and LlamaConfig head_dim validation are different API/validation issues." }, { "left": "issue:45584", "right": "issue:46170", "accept": false, - "reason": "Whisper empty-transcription failure is unrelated to the Windows ROCm torch.distributed dependency issue." + "reason": "Whisper empty-transcription decoding and Windows ROCm import failures do not share a code path." }, { "left": "issue:44188", "right": "issue:46421", "accept": false, - "reason": "Both mention cache/compile behavior, but the concrete bugs and affected paths differ." + "reason": "Both mention caching, but one is attention-kernel divergence under torch.compile and the other is lazy cache init causing recompilation." }, { "left": "issue:43065", "right": "issue:46123", "accept": false, - "reason": "SAM3 pixel decoder internals and MaskGenerationPipeline final-batch handling are different issues." + "reason": "Sam3PixelDecoder dummy layer and MaskGenerationPipeline partial-batch handling are unrelated." }, { "left": "issue:44297", "right": "issue:44623", "accept": false, - "reason": "Tokenizer metadata mismatch and processor save_pretrained missing files are both save_pretrained-related, but not the same bug." + "reason": "Both are save_pretrained regressions, but one is tokenizer config metadata and the other is processor file emission." }, { "left": "issue:44361", "right": "issue:44749", "accept": false, - "reason": "MLukeTokenizer task AttributeError and a global performance regression are not the same failure." + "reason": "MLukeTokenizer AttributeError and a dataset filtering slowdown are unrelated bugs." }, { "left": "issue:44749", "right": "issue:45656", "accept": false, - "reason": "A slowdown after upgrade and deepspeed calling optimizer.step twice are different concrete problems." + "reason": "A preprocessing performance regression and a deepspeed optimizer-step bug are not the same change." }, { - "left": "issue:43646", - "right": "issue:44162", + "left": "issue:44623", + "right": "issue:45656", "accept": false, - "reason": "Custom model initialization breakage and ESM2 breakage are separate regressions." + "reason": "Processor.save_pretrained missing files and duplicated optimizer steps are completely different issues." }, { "left": "issue:44749", "right": "issue:45310", "accept": false, - "reason": "Performance regression after upgrade is unrelated to the Qwen3.5 MoE from_pretrained error." + "reason": "Data filtering slowdown and Qwen3.5 MoE from_pretrained failure are unrelated." }, { - "left": "issue:44623", - "right": "issue:45656", + "left": "issue:43646", + "right": "issue:44162", "accept": false, - "reason": "Processor save_pretrained file omission and double optimizer stepping are unrelated." + "reason": "Both are broad regressions, but one is custom model initialization and the other is an ESM2-specific breakage." }, { "left": "issue:44871", "right": "issue:46132", "accept": false, - "reason": "EOS-token config inconsistency and AttentionInterface.register behavior change are different issues." + "reason": "Gemma eos_token_id configuration mismatch and AttentionInterface.register behavior changes are different concerns." }, { "left": "issue:43232", "right": "issue:43673", "accept": false, - "reason": "Both are generation/cache-related, but sync_gpus kwargs handling and chunked_prefill cache absence are not the same bug." + "reason": "Both involve generation/cache, but one is sync_gpus update kwargs and the other is chunked_prefill cache availability." }, { "left": "issue:42994", "right": "issue:43482", "accept": false, - "reason": "Quantized model saving failure and GGUF loading failure are different serialization/loading paths." - }, - { - "left": "issue:44869", - "right": "issue:45356", - "accept": false, - "reason": "Whisper decode crash on a trailing replacement character is unrelated to the Kimi tokenizer codec regression." + "reason": "Quantized model saving failure and Qwen2.5-GGUF loading failure are different lifecycle operations." }, { "left": "issue:45230", "right": "issue:45397", "accept": false, - "reason": "A generic bug report cannot be merged with a specific Gemma-4 zero3 from_pretrained bug." + "reason": "A placeholder bug report is not a duplicate of a specific Gemma-4 ZeRO3 from_pretrained failure." }, { - "left": "issue:43475", - "right": "issue:43643", + "left": "issue:44869", + "right": "issue:45356", "accept": false, - "reason": "Missing fpn_position_embeddings in SAM3 output and missing fields from trust_remote_code config are different problems." + "reason": "Whisper timestamp decoding on a trailing replacement character and Kimi-K2.5 tokenizer codec handling are related only at the tokenizer level, not the same bug." }, { "left": "issue:45375", "right": "issue:45478", - "accept": true, - "reason": "These have the same title and describe the same Qwen3.5 MoE from_pretrained error, so they are duplicate reports of the same bug." - }, - { - "left": "issue:28282", - "right": "issue:30990", "accept": false, - "reason": "A missing PyTorch import error and a Sentence Transformers loading hang are unrelated." + "reason": "Same model family and loading context, but one is a missing config field dropped by @strict and the other is a generic from_pretrained failure; not the same concrete bug." }, { - "left": "issue:41628", - "right": "issue:44933", + "left": "issue:43475", + "right": "issue:43643", "accept": false, - "reason": "AutoImageProcessor import failure and a different missing image_utils import are separate symbol-resolution issues." + "reason": "A missing Sam3VisionEncoderOutput attribute and missing fields in AutoConfig with trust_remote_code are unrelated." } ] }, - "evaluator_result": { - "accept": false, - "feedback": "The overall canonical choice around issue:45310 / issue:45478 is grounded, but the output is not precise enough to accept. At least one soft-edge rationale introduces unsupported details: issue:42886 is titled \"Tokenizer fails to load from cache when HF_HUB_OFFLINE=1 on 4.57.3\", not a missing PyTorch import. Since some verdict explanations go beyond what the packet supports, I have to reject." - }, - "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 }, @@ -13007,9 +12929,11 @@ "source": "llm", "cache_hit": false, "model": "gpt-5.4-mini?service_tier=flex", - "cluster_id": "cluster-10000-472", + "cluster_id": "cluster-10000-473", "nodes": [ + "issue:28282", "issue:30333", + "issue:30990", "issue:33290", "issue:36010", "issue:41628", @@ -13051,21 +12975,21 @@ "issue:45600", "issue:45712", "issue:45753", - "issue:46077", - "issue:46123", "issue:46202", "issue:46326", "issue:46399" ], "soft_pairs": [ "issue:44038|issue:44623", + "issue:28282|issue:30990", + "issue:41628|issue:44933", "issue:42371|issue:45092", "issue:45310|issue:45375", "issue:42994|issue:43525", "issue:43475|issue:45381", "issue:44964|issue:45137", - "issue:43784|issue:46399", "issue:43723|issue:46326", + "issue:43784|issue:46399", "issue:45406|issue:46202", "issue:43873|issue:44945", "issue:43232|issue:44188", @@ -13076,236 +13000,241 @@ "issue:42994|issue:43824", "issue:44038|issue:45310", "issue:44222|issue:45310", - "issue:36010|issue:41628", "issue:42831|issue:44222", + "issue:36010|issue:41628", "issue:43278|issue:45137", "issue:43232|issue:44792", "issue:42175|issue:44933", "issue:44496|issue:45561", "issue:43504|issue:45561", "issue:45561|issue:45600", - "issue:42175|issue:45712", "issue:43976|issue:44933", "issue:30333|issue:33290", - "issue:46077|issue:46123" + "issue:42175|issue:45712" ], - "prepared_review_unit_hash": "12c02e95f984fb7c0933868da922d6ec2bd2145707d27e55d442018816fdfb2c", + "prepared_review_unit_hash": "393e3dd196235966ccc7591282c14855b8d79f1dfcaff5bfcace92cff54fef99", "component_budget": { - "node_count": 472, - "item_count": 472, - "soft_pair_count": 1732, - "serialized_chars": 898654, - "estimated_input_tokens": 224664, - "estimated_eval_tokens": 449584 + "node_count": 473, + "item_count": 473, + "soft_pair_count": 1738, + "serialized_chars": 901437, + "estimated_input_tokens": 225360, + "estimated_eval_tokens": 450976 }, "budget": { "node_count": 47, "item_count": 47, - "soft_pair_count": 30, - "serialized_chars": 35982, - "estimated_input_tokens": 8996, - "estimated_eval_tokens": 18248 + "soft_pair_count": 31, + "serialized_chars": 36242, + "estimated_input_tokens": 9061, + "estimated_eval_tokens": 18378 }, "overflow_policy": "truncate_then_skip", "trimmed": false, "aggressively_trimmed": false, "split": true, "analyst_result": { - "summary": "This cluster is highly heterogeneous: most soft pairs are not duplicates and span unrelated bugs in imports, packaging, generation, quantization, and model-specific regressions. One pair with the exact same Qwen3ForCausalLM VRAM-leak title looks like a true duplicate.", + "summary": "Mostly unrelated Transformers issues; the only clear duplicate pair is the Qwen3ForCausalLM VRAM leak reports (42673/43704). I corrected the canonical issue wording to avoid claiming it is open when the packet shows it is closed.", "confidence": 0.95, - "canonical_issue_reason": null, + "canonical_issue_reason": "issue:42673 is the better canonical because it is the earlier, more detailed report for the same Qwen3ForCausalLM VRAM leak, and issue:43704 is an exact duplicate with the same title.", "canonical_pr_reason": null, - "best_issue_reason": "Issue 43873 is the strongest standalone issue in this set by discussion and inbound references, but it is still not a true canonical for the cluster because the items are mostly unrelated.", + "best_issue_reason": "issue:42673 is the best representative issue in this set because it is the clearest duplicate anchor for the one confirmed same-bug pair and has the richer discussion trail.", "best_pr_reason": null, "soft_edge_verdicts": [ { "left": "issue:44038", "right": "issue:44623", "accept": false, - "reason": "Different bugs: Qwen3-VL-Moe/transformers compatibility vs processor.save_pretrained missing files." + "reason": "Different bugs: Qwen3-VL-Moe load/model compatibility vs missing files from processor.save_pretrained()." + }, + { + "left": "issue:28282", + "right": "issue:30990", + "accept": false, + "reason": "Different failure modes: missing PyTorch backend import vs Sentence Transformers hanging while loading." + }, + { + "left": "issue:41628", + "right": "issue:44933", + "accept": false, + "reason": "Both are import-related, but they concern different missing symbols in different modules." }, { "left": "issue:42371", "right": "issue:45092", "accept": false, - "reason": "TF32 API/settings behavior is unrelated to old InternVL2 meta-initialization incompatibility." + "reason": "Unrelated issues: TF32 API guidance vs old InternVL2 checkpoints failing under v5 meta initialization." }, { "left": "issue:45310", "right": "issue:45375", "accept": false, - "reason": "Same model family, but different concrete failures: from_pretrained error vs silently dropped config field." + "reason": "Related model family, but not the same concrete bug or fix set." }, { "left": "issue:42994", "right": "issue:43525", "accept": false, - "reason": "Quantized model saving failure and missing pad_token_id on Llama4Config are separate code paths." + "reason": "Different problems: quantized model saving failure vs missing pad_token_id on Llama4Config." }, { "left": "issue:43475", "right": "issue:45381", "accept": false, - "reason": "SAM 3 vision encoder missing attribute is unrelated to Qwen2.5-VL video vision_position_ids." + "reason": "Different models and code paths: SAM 3 vision output attribute vs qwen2.5-vl video position IDs." }, { "left": "issue:44964", "right": "issue:45137", "accept": false, - "reason": "Model loading failure and DeepSpeed ZeRO3 deque IndexError are unrelated." + "reason": "Different subsystems: Phi-4 multimodal loading vs DeepSpeed ZeRO3 deque underflow." }, { - "left": "issue:43784", - "right": "issue:46399", + "left": "issue:43723", + "right": "issue:46326", "accept": false, - "reason": "Sentence-transformers import NameError is unrelated to Qwen3.5 MoE rotary-pos-emb registration." + "reason": "Both involve tokenizers, but the bugs are different concrete code paths." }, { - "left": "issue:43723", - "right": "issue:46326", + "left": "issue:43784", + "right": "issue:46399", "accept": false, - "reason": "Tokenizer loading in v5 and continue_final_message on mistral_common are different tokenizer bugs." + "reason": "Different issues: sentence-transformers import NameError vs Qwen3.5 kernel registration crash." }, { "left": "issue:45406", "right": "issue:46202", "accept": false, - "reason": "Gemma4Processor missing _tokenizer and Gemma pipeline roundtrip on \u2581 are distinct failures." + "reason": "Different Gemma issues: missing _tokenizer on processor vs a pipeline roundtrip tokenization mismatch." }, { "left": "issue:43873", "right": "issue:44945", "accept": false, - "reason": "Quantization offloading behavior and pipeline-parallel wrong outputs do not share the same concrete bug." + "reason": "Different behaviors: quantization/offloading not working as expected vs incorrect output under pipeline parallelism." }, { "left": "issue:43232", "right": "issue:44188", "accept": false, - "reason": "Generation kwargs update after sync_gpus is unrelated to torch.compile attention-kernel branching." + "reason": "Different generation/attention bugs with no shared concrete fix path." }, { "left": "issue:41628", "right": "issue:44351", "accept": false, - "reason": "AutoImageProcessor import error and HybridCache import error are separate missing-export issues." + "reason": "Different import errors in different APIs: AutoImageProcessor vs HybridCache." }, { "left": "issue:44945", "right": "issue:45753", "accept": false, - "reason": "Pipeline-parallel incorrect output and Qwen3_5 infinite loop on a specific image are different symptoms and paths." + "reason": "Different symptoms and code paths: pipeline parallelism output bug vs Qwen3_5 image infinite loop." }, { "left": "issue:45081", "right": "issue:45103", "accept": false, - "reason": "Mistral tokenizer patch crash is unrelated to auto_docstring future-annotations failure." + "reason": "Completely different crashes: Mistral regex patching vs auto_docstring with postponed annotations." }, { "left": "issue:42673", "right": "issue:43704", "accept": true, - "reason": "Exact same title and same Qwen3ForCausalLM VRAM leak in multiple dataloader threads; likely the same duplicate bug." + "reason": "Same underlying bug and identical title: Qwen3ForCausalLM VRAM leak across multiple dataloader threads." }, { "left": "issue:42994", "right": "issue:43824", "accept": false, - "reason": "Quantized model saving failure is unrelated to Qwen2_5_VL import resolution." + "reason": "Different issues: quantized model saving failure vs missing Qwen2_5_VLForConditionalGeneration import." }, { "left": "issue:44038", "right": "issue:45310", "accept": false, - "reason": "Both mention Qwen3.5/MoE, but one is a Qwen3-VL-Moe compatibility issue and the other is a from_pretrained regression." + "reason": "Same broad model family, but not the same concrete bug or fixable change set." }, { "left": "issue:44222", "right": "issue:45310", "accept": false, - "reason": "FP8 save_pretrained/moe and Qwen3.5 MoE from_pretrained are different operations and failure modes." + "reason": "Different bugs: FP8 save_pretrained/moe vs Qwen3.5 from_pretrained loading error." }, { - "left": "issue:36010", - "right": "issue:41628", + "left": "issue:42831", + "right": "issue:44222", "accept": false, - "reason": "GenerationMixin import failure and AutoImageProcessor import failure are distinct missing symbols." + "reason": "Different FP8-related issues: accuracy degradation vs save_pretrained failure." }, { - "left": "issue:42831", - "right": "issue:44222", + "left": "issue:36010", + "right": "issue:41628", "accept": false, - "reason": "FineGrainedFP8 accuracy regression is not the same as FP8 save_pretrained failing." + "reason": "Both are import errors, but for different symbols and code paths." }, { "left": "issue:43278", "right": "issue:45137", "accept": false, - "reason": "BF16-to-FP32 dtype drift in eval is unrelated to DeepSpeed ZeRO3 deque underflow." + "reason": "Different problems: dtype change in eval vs DeepSpeed ZeRO3 deque pop error." }, { "left": "issue:43232", "right": "issue:44792", "accept": false, - "reason": "Generation kwarg update issue and Janus test failure are unrelated." + "reason": "Different code paths: generation kwargs update after sync_gpus vs a janus image-generation test failure." }, { "left": "issue:42175", "right": "issue:44933", "accept": false, - "reason": "Missing TensorFlow backend in a pip extra is unrelated to a nonexistent image_utils import." + "reason": "Different bugs: dependency/backend packaging vs a nonexistent image_utils import." }, { "left": "issue:44496", "right": "issue:45561", "accept": false, - "reason": "Unrecognized model config and pytest-xdist race in testing utils are unrelated." + "reason": "Unrelated: model config loading failure vs pytest-xdist race in testing utils." }, { "left": "issue:43504", "right": "issue:45561", "accept": false, - "reason": "Legacy field breaking Beit preset loading is unrelated to the xdist captured_info.txt race." + "reason": "Different issues: legacy BEiT preset loading vs xdist captured_info.txt race." }, { "left": "issue:45561", "right": "issue:45600", "accept": false, - "reason": "Test-file race and removed Sam3LiteText references in auto_mappings are different problems." - }, - { - "left": "issue:42175", - "right": "issue:45712", - "accept": false, - "reason": "TensorFlow packaging backend omission is unrelated to leftover dummy PT classes." + "reason": "Different CI/test problems: xdist race vs stale auto_mappings references." }, { "left": "issue:43976", "right": "issue:44933", "accept": false, - "reason": "Python-version incompatibility is unrelated to a missing image_utils import." + "reason": "Different bugs: Python version compatibility vs missing import from image_utils." }, { "left": "issue:30333", "right": "issue:33290", "accept": false, - "reason": "MLflow reporting on failed jobs and DeepSpeed/Adafactor OOM are unrelated." + "reason": "Different failures: MLflow reporting after training-job failure vs OOM with Adafactor in DeepSpeed." }, { - "left": "issue:46077", - "right": "issue:46123", + "left": "issue:42175", + "right": "issue:45712", "accept": false, - "reason": "Gemma4 bidirectional attention masking and MaskGenerationPipeline final-batch handling are different bugs." + "reason": "Related packaging area, but different concrete bugs and fixes." } ] }, "evaluator_result": { - "accept": true, - "feedback": "The assessment is grounded in the packet and stays conservative. The cluster does look heterogeneous overall, and the only clearly supported duplicate is the exact-title pair issue:42673 / issue:43704. No soft-edge verdict appears to overstate evidence." + "accept": false, + "feedback": "Reject: the canonical_issue_reason overstates the packet by calling issue:42673 the 'original/open report' even though the packet shows it is closed. The rest is mostly conservative, and the 42673/43704 duplicate call looks grounded, but the state claim is factually incorrect." }, "evaluator_used": true, - "retried": false, + "retried": true, "accepted_nontrivial_soft_edge": true, "error_kind": null, "error_message": null @@ -13321,7 +13250,7 @@ "source": "llm", "cache_hit": false, "model": "gpt-5.4-mini?service_tier=flex", - "cluster_id": "cluster-10000-472", + "cluster_id": "cluster-10000-473", "nodes": [ "issue:22355", "issue:27957", @@ -13338,7 +13267,6 @@ "issue:43295", "issue:43377", "issue:43479", - "issue:43644", "issue:43723", "issue:43784", "issue:43824", @@ -13352,7 +13280,6 @@ "issue:44242", "issue:44295", "issue:44297", - "issue:44448", "issue:44568", "issue:44623", "issue:44734", @@ -13369,10 +13296,12 @@ "issue:45656", "issue:45712", "issue:45812", - "issue:46132", + "issue:46077", + "issue:46123", "issue:46231" ], "soft_pairs": [ + "issue:46077|issue:46123", "issue:30990|issue:35141", "issue:42898|issue:44568", "issue:43377|issue:44242", @@ -13380,10 +13309,10 @@ "issue:43479|issue:44869", "issue:42222|issue:45200", "issue:44016|issue:46231", - "issue:43479|issue:44079", "issue:42898|issue:45812", - "issue:43723|issue:44568", "issue:43874|issue:45092", + "issue:43723|issue:44568", + "issue:43479|issue:44079", "issue:42898|issue:44779", "issue:43784|issue:43824", "issue:44734|issue:45103", @@ -13392,247 +13321,239 @@ "issue:27957|issue:30333", "issue:42898|issue:43295", "issue:30333|issue:34689", - "issue:45405|issue:45712", "issue:44060|issue:45125", + "issue:45405|issue:45712", "issue:44075|issue:46231", "issue:43116|issue:44190", - "issue:44933|issue:45712", "issue:44222|issue:44297", + "issue:44933|issue:45712", "issue:30333|issue:36331", "issue:30333|issue:35141", - "issue:22355|issue:28282", "issue:44297|issue:45656", + "issue:22355|issue:28282", "issue:42831|issue:44623", - "issue:44448|issue:46132", - "issue:44222|issue:44623", - "issue:43479|issue:43644" + "issue:44222|issue:44623" ], - "prepared_review_unit_hash": "ddedd05d3e7e565cf2d570439d4369c02a9b398a622f4f8ec035b98916b36c76", + "prepared_review_unit_hash": "e5f61587f3cdb5d5b123b6d67626d8b519fa8f47d68fe8e699c9dfbd9c688291", "component_budget": { - "node_count": 472, - "item_count": 472, - "soft_pair_count": 1732, - "serialized_chars": 898654, - "estimated_input_tokens": 224664, - "estimated_eval_tokens": 449584 + "node_count": 473, + "item_count": 473, + "soft_pair_count": 1738, + "serialized_chars": 901437, + "estimated_input_tokens": 225360, + "estimated_eval_tokens": 450976 }, "budget": { - "node_count": 48, - "item_count": 48, - "soft_pair_count": 33, - "serialized_chars": 37762, - "estimated_input_tokens": 9441, - "estimated_eval_tokens": 19138 + "node_count": 47, + "item_count": 47, + "soft_pair_count": 32, + "serialized_chars": 36920, + "estimated_input_tokens": 9230, + "estimated_eval_tokens": 18716 }, "overflow_policy": "truncate_then_skip", "trimmed": false, "aggressively_trimmed": false, "split": true, "analyst_result": { - "summary": "Mostly a heterogeneous cluster of unrelated Transformers issues; a few are all tokenizer/v5 regressions, but none look like the same underlying bug. I\u2019d use 42898 as the closest representative issue, but only as a broad tokenizer-regression anchor, not a true duplicate hub.", - "confidence": 0.24, - "canonical_issue_reason": "Issue 42898 is the broadest and most generally framed of the set, centered on a Transformers v5 tokenizer behavior regression. It is the closest available cluster anchor, though the surrounding issues span many unrelated subsystems.", + "summary": "This cluster is heterogeneous: most items are distinct regression reports across tokenization, model loading, training, serving, and docs, with no true duplicate pair standing out. The closest theme is tokenizer/v5 regressions, but even there the failures are model- or API-specific rather than the same bug.", + "confidence": 0.34, + "canonical_issue_reason": "issue:42898 is the broadest, most central report in the cluster: a core tokenizer behavior regression affecting v5 behavior generally, and it is the closest match to several other tokenizer-related issues.", "canonical_pr_reason": null, - "best_issue_reason": "42898 is the best issue to represent the cluster because it is a user-visible v5 regression in a core tokenizer path and is more general than the model-specific or tooling-specific issues. Still, it should not be treated as a duplicate of the others.", + "best_issue_reason": "issue:42898 is the best single issue to anchor this cluster because it is a cross-cutting library regression rather than a niche model-specific or example-script-only problem.", "best_pr_reason": null, "soft_edge_verdicts": [ + { + "left": "issue:46077", + "right": "issue:46123", + "accept": false, + "reason": "Different bugs: Gemma4 attention-mask construction vs. a pipeline batching/last-chunk flag issue." + }, { "left": "issue:30990", "right": "issue:35141", "accept": false, - "reason": "Stuck loading Sentence Transformers vs embedding reinitialization after resize; different code paths and symptoms." + "reason": "Unrelated failure modes: a loading stall vs. embedding reinitialization after resize." }, { "left": "issue:42898", "right": "issue:44568", "accept": false, - "reason": "Both are tokenizer regressions in v5, but one is cleanup-space behavior and the other is special-token insertion for a specific tokenizer." + "reason": "Both are tokenizer regressions, but they affect different models/APIs and distinct code paths (`clean_up_tokenization_spaces` vs `add_special_tokens`)." }, { "left": "issue:43377", "right": "issue:44242", "accept": false, - "reason": "MIMI batched-vs-single padding-mask bug is unrelated to MoE load-balancing loss gating." + "reason": "Different components: MIMI padding-mask behavior vs. MoE load-balancing loss calculation." }, { "left": "issue:44623", "right": "issue:45310", "accept": false, - "reason": "Missing files in processor.save_pretrained is unrelated to Qwen3.5 MoE from_pretrained failure." + "reason": "Different save/load problems: missing processor files vs. Qwen3.5 MoE `from_pretrained` failure." }, { "left": "issue:43479", "right": "issue:44869", "accept": false, - "reason": "Phi4MultimodalConfig default-init bug is unrelated to Whisper word-timestamp decode crashing." + "reason": "Different model classes and failure modes: config initialization vs. Whisper timestamp decode crash." }, { "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": "Different model families and bugs: broken VitPose models vs. Gemma 4 token-type defaults." }, { "left": "issue:44016", "right": "issue:46231", "accept": false, - "reason": "Notebook syntax error is not the same as a documentation parameter-name mismatch." + "reason": "A notebook syntax error is not the same as a docs parameter-name mismatch." }, { - "left": "issue:43479", - "right": "issue:44079", + "left": "issue:42898", + "right": "issue:45812", "accept": false, - "reason": "Config initialization behavior and ModelOutput key assignment are distinct problems." + "reason": "Both involve tokenizer regressions, but one is a global cleanup behavior change and the other is wrong token IDs for Granite models." }, { - "left": "issue:42898", - "right": "issue:45812", + "left": "issue:43874", + "right": "issue:45092", "accept": false, - "reason": "Tokenizer cleanup-space regression and Granite token-ID regression are different tokenizer bugs." + "reason": "Different areas: missing image-processor method vs. remote-code/meta-init incompatibility." }, { "left": "issue:43723", "right": "issue:44568", "accept": false, - "reason": "AutoTokenizer loading issue in v5 is different from add_special_tokens not adding BOS/EOS." + "reason": "Separate tokenizer issues with different symptoms: tokenizer loading vs. special-token insertion." }, { - "left": "issue:43874", - "right": "issue:45092", + "left": "issue:43479", + "right": "issue:44079", "accept": false, - "reason": "Missing image-processor method and meta-init incompatibility are unrelated multimodal failures." + "reason": "Config-default construction and `ModelOutput` key handling are unrelated code paths." }, { "left": "issue:42898", "right": "issue:44779", "accept": false, - "reason": "Both mention tokenizer regressions, but Deepseek codec/regex handling is a different bug from clean_up_tokenization_spaces." + "reason": "Both are tokenizer regressions, but they affect different models and different underlying tokenization logic." }, { "left": "issue:43784", "right": "issue:43824", "accept": false, - "reason": "Sentence-Transformers import NameError and missing Qwen2_5_VL export are different import failures." + "reason": "Different import errors: one is a missing symbol in sentence-transformers integration, the other is a missing Qwen2.5-VL export." }, { "left": "issue:44734", "right": "issue:45103", "accept": false, - "reason": "KV-cache continuation crash in serve is unrelated to auto_docstring failing on future annotations." + "reason": "Serving KV-cache indexing crash and auto-docstring annotation handling are unrelated." }, { "left": "issue:42898", "right": "issue:45356", "accept": false, - "reason": "Generic cleanup-space change and Kimi-K2.5 codec regression are different tokenizer problems." + "reason": "Both mention tokenizer regressions, but the Kimi issue is codec/regex-specific and not the same bug as `clean_up_tokenization_spaces` behavior." }, { "left": "issue:44295", "right": "issue:45656", "accept": false, - "reason": "Reading position_ids after buffer registration is unrelated to deepspeed causing optimizer.step() to run twice." + "reason": "Different subsystems: buffer access on `position_ids` vs. optimizer stepping under DeepSpeed." }, { "left": "issue:27957", "right": "issue:30333", "accept": false, - "reason": "Flash Attention 2 on XLMRoberta and MLFlow reporting after failed training are unrelated." + "reason": "Unrelated issues: Flash Attention 2 support for XLM-Roberta vs. MLflow reporting after failed training jobs." }, { "left": "issue:42898", "right": "issue:43295", "accept": false, - "reason": "Tokenizer v5 behavior change is unrelated to processor.tokenizer/image passing regression." + "reason": "Tokenizer cleanup behavior and processor/tokenizer access regressions are different problems." }, { "left": "issue:30333", "right": "issue:34689", "accept": false, - "reason": "MLFlow job-state reporting and Llama 3.2 Vision model loading are different failures." + "reason": "MLflow job-tracking behavior is unrelated to model-loading breakage for Llama 3.2 Vision." }, { - "left": "issue:45405", - "right": "issue:45712", + "left": "issue:44060", + "right": "issue:45125", "accept": false, - "reason": "Version bump for PEFT and leftover dummy classes failing repo checks are unrelated." + "reason": "Both are Qwen-related, but one is a tied-weights warning and the other is missing tensor-parallel plan support." }, { - "left": "issue:44060", - "right": "issue:45125", + "left": "issue:45405", + "right": "issue:45712", "accept": false, - "reason": "Incorrect tied-weights warning is not the same as a missing _tp_plan for tensor parallelism." + "reason": "A dependency-version bump and leftover dummy classes are unrelated maintenance issues." }, { "left": "issue:44075", "right": "issue:46231", "accept": false, - "reason": "SGD args being ignored is unrelated to a documentation parameter-name typo." + "reason": "Optimizer arguments not being used is not the same as a documentation typo." }, { "left": "issue:43116", "right": "issue:44190", "accept": false, - "reason": "Empty multi-label classification results and local dataset loading failure are different example-script bugs." + "reason": "Different example-script failures: empty multi-label results vs. local dataset loading." }, { - "left": "issue:44933", - "right": "issue:45712", + "left": "issue:44222", + "right": "issue:44297", "accept": false, - "reason": "Nonexistent image_utils import and dummy class cleanup are separate repository hygiene issues." + "reason": "Different save-pretrained bugs affecting different object types: FP8 MoE vs. tokenizer metadata." }, { - "left": "issue:44222", - "right": "issue:44297", + "left": "issue:44933", + "right": "issue:45712", "accept": false, - "reason": "FP8 save_pretrained in MoE and tokenizer_class mismatch in tokenizer_config.json are unrelated save-path bugs." + "reason": "Import exposure problems in `image_utils` and dummy class cleanup are not the same bug." }, { "left": "issue:30333", "right": "issue:36331", "accept": false, - "reason": "MLFlow reporting on failure is unrelated to CustomTrainer.compute_loss signature breakage." + "reason": "MLflow callback/reporting and trainer loss-signature compatibility are different training issues." }, { "left": "issue:30333", "right": "issue:35141", "accept": false, - "reason": "MLFlow tracking bug and embedding reinitialization after resize are different issues." - }, - { - "left": "issue:22355", - "right": "issue:28282", - "accept": false, - "reason": "Missing transformers.onnx module and missing PyTorch dependency are both import errors, but on different optional-dependency paths." + "reason": "Failed-job reporting and embedding reinitialization are unrelated." }, { "left": "issue:44297", "right": "issue:45656", "accept": false, - "reason": "Tokenizer save_pretrained metadata mismatch is unrelated to the optimizer double-step deepspeed bug." + "reason": "Tokenizer save metadata mismatch and duplicate optimizer stepping are unrelated." }, { - "left": "issue:42831", - "right": "issue:44623", + "left": "issue:22355", + "right": "issue:28282", "accept": false, - "reason": "FineGrainedFP8 accuracy issue and processor.save_pretrained missing files are different problems." + "reason": "Both are import errors, but one is a missing `transformers.onnx` module and the other is a missing PyTorch dependency check." }, { - "left": "issue:44448", - "right": "issue:46132", + "left": "issue:42831", + "right": "issue:44623", "accept": false, - "reason": "Pegasus v4/v5 output difference and AttentionInterface.register behavior change are unrelated API regressions." + "reason": "An FP8 accuracy regression and missing processor files are different problem types." }, { "left": "issue:44222", "right": "issue:44623", "accept": false, - "reason": "FP8 MoE save_pretrained bug and missing processor files are distinct serialization issues." - }, - { - "left": "issue:43479", - "right": "issue:43644", - "accept": false, - "reason": "Config default initialization and non-persistent buffer junk are unrelated v5 regressions." + "reason": "Both mention save_pretrained, but one is FP8/MoE serialization and the other is processor file persistence." } ] }, @@ -13654,7 +13575,7 @@ "source": "llm", "cache_hit": false, "model": "gpt-5.4-mini?service_tier=flex", - "cluster_id": "cluster-10000-472", + "cluster_id": "cluster-10000-473", "nodes": [ "issue:30990", "issue:38175", @@ -13662,10 +13583,10 @@ "issue:42757", "issue:42796", "issue:42831", - "issue:42907", "issue:43122", "issue:43232", "issue:43475", + "issue:43479", "issue:43644", "issue:43723", "issue:44190", @@ -13678,11 +13599,11 @@ "issue:44297", "issue:44360", "issue:44373", + "issue:44448", "issue:44484", "issue:44485", "issue:44623", "issue:44734", - "issue:44829", "issue:44871", "issue:44977", "issue:45071", @@ -13693,8 +13614,6 @@ "issue:45357", "issue:45381", "issue:45834", - "issue:46002", - "issue:46050", "issue:46082", "issue:46132", "issue:46209", @@ -13703,264 +13622,263 @@ "issue:46421", "issue:46424", "issue:46431", - "issue:46655" + "issue:46655", + "issue:46665" ], "soft_pairs": [ + "issue:44448|issue:46132", + "issue:43479|issue:43644", "issue:42796|issue:43723", "issue:45245|issue:46132", + "issue:43479|issue:46665", "issue:41628|issue:42757", "issue:44297|issue:45310", "issue:45071|issue:46082", - "issue:43723|issue:44977", - "issue:44263|issue:45092", "issue:44373|issue:44485", + "issue:44263|issue:45092", + "issue:43723|issue:44977", "issue:43122|issue:46209", - "issue:43232|issue:46424", "issue:44222|issue:44484", + "issue:43232|issue:46424", "issue:44206|issue:45103", "issue:46209|issue:46431", "issue:44190|issue:46082", - "issue:46305|issue:46655", - "issue:43475|issue:46305", "issue:44734|issue:46421", "issue:43232|issue:45834", "issue:44295|issue:44623", + "issue:43475|issue:46305", + "issue:46305|issue:46655", "issue:43723|issue:44871", - "issue:45381|issue:46305", "issue:44263|issue:44360", - "issue:30990|issue:38175", + "issue:45381|issue:46305", "issue:44261|issue:45092", + "issue:30990|issue:38175", "issue:43644|issue:46305", - "issue:45092|issue:45357", - "issue:42907|issue:46050", "issue:42831|issue:45310", + "issue:45092|issue:45357", "issue:44279|issue:46328", - "issue:44263|issue:44373", - "issue:44829|issue:46002" + "issue:44263|issue:44373" ], - "prepared_review_unit_hash": "7811663803bb8410a1e1925baa9aeedd497bab4ef834a60756cc3043d75b59f7", + "prepared_review_unit_hash": "e4eeb8bf5fe5d73a9acab4e9ea985a27da5d80aae39873509003366f29c11807", "component_budget": { - "node_count": 472, - "item_count": 472, - "soft_pair_count": 1732, - "serialized_chars": 898654, - "estimated_input_tokens": 224664, - "estimated_eval_tokens": 449584 + "node_count": 473, + "item_count": 473, + "soft_pair_count": 1738, + "serialized_chars": 901437, + "estimated_input_tokens": 225360, + "estimated_eval_tokens": 450976 }, "budget": { - "node_count": 48, - "item_count": 48, - "soft_pair_count": 31, - "serialized_chars": 36683, - "estimated_input_tokens": 9171, - "estimated_eval_tokens": 18598 + "node_count": 47, + "item_count": 47, + "soft_pair_count": 32, + "serialized_chars": 36407, + "estimated_input_tokens": 9102, + "estimated_eval_tokens": 18460 }, "overflow_policy": "truncate_then_skip", "trimmed": false, "aggressively_trimmed": false, "split": true, "analyst_result": { - "summary": "Mostly conservative cluster with no strong duplicate pairs supported by the packet. The one previously accepted DSA pair is too weakly evidenced here and should be rejected; the titles are only thematically similar and there is no body/diff proof of the same concrete bug.", + "summary": "This cluster looks like a grab-bag of unrelated Transformers bugs/regressions rather than one duplicate family. Based on titles alone, none of the soft pairs are safe duplicates; the closest umbrella report is the generic tokenizer-loading regression, but it is still too broad to serve as a true canonical duplicate.", "confidence": 0.86, - "canonical_issue_reason": null, + "canonical_issue_reason": "Issue 43723 is the broadest generic loading-regression report in the set, so it is the best single representative if one must be chosen; however, the cluster is not actually a single duplicate thread.", "canonical_pr_reason": null, - "best_issue_reason": "issue:44263 is the most concrete code-path-specific report in the set, but it is better viewed as a representative issue than a proven duplicate canonical.", + "best_issue_reason": "43723 is the most umbrella-like issue here because it covers a general v5 loading failure, but it still does not subsume the model-specific/import/config bugs in the cluster.", "best_pr_reason": null, "soft_edge_verdicts": [ + { + "left": "issue:44448", + "right": "issue:46132", + "accept": false, + "reason": "Different code paths: generation output mismatch vs attention registration behavior." + }, + { + "left": "issue:43479", + "right": "issue:43644", + "accept": false, + "reason": "Config initialization bug vs non-persistent buffer initialization bug; not the same defect." + }, { "left": "issue:42796", "right": "issue:43723", "accept": false, - "reason": "Both involve tokenizer/loading behavior, but one is a BPE-marker handling bug and the other is an AutoTokenizer.from_pretrained import/load regression." + "reason": "Model-specific tokenizer BPE handling vs generic AutoTokenizer loading failure." }, { "left": "issue:45245", "right": "issue:46132", "accept": false, - "reason": "Different bugs: category-cardinality RuntimeError vs changed behavior of AttentionInterface.register." + "reason": "Category-cardinality runtime error is unrelated to AttentionInterface registration." + }, + { + "left": "issue:43479", + "right": "issue:46665", + "accept": false, + "reason": "Different from_pretrained/config construction problems with no shared code-path." }, { "left": "issue:41628", "right": "issue:42757", "accept": false, - "reason": "Both are import errors, but for different symbols and different dependency paths; not the same underlying bug." + "reason": "Both are import failures, but for different symbols and different dependency paths." }, { "left": "issue:44297", "right": "issue:45310", "accept": false, - "reason": "Tokenizer save_pretrained metadata mismatch and a Qwen3.5 from_pretrained failure are different issues." + "reason": "Tokenizer save metadata mismatch vs Qwen3.5 from_pretrained error; different bug." }, { "left": "issue:45071", "right": "issue:46082", "accept": false, - "reason": "Different config-validation problems: generic PretrainedConfig type checking vs LlamaConfig head_dim constraints." + "reason": "Type-checking regression is unrelated to LlamaConfig head_dim validation." }, { - "left": "issue:43723", - "right": "issue:44977", + "left": "issue:44373", + "right": "issue:44485", "accept": false, - "reason": "Tokenizer loading regression and flash-attention generation failure are unrelated code paths." + "reason": "Wrong docstring is not the same as a RoPE implementation bug." }, { "left": "issue:44263", "right": "issue:45092", "accept": false, - "reason": "DSA indexer split behavior vs remote-code checkpoint/meta-init incompatibility are unrelated." + "reason": "DSA split/indexing bug vs remote-code meta-init incompatibility; different failures." }, { - "left": "issue:44373", - "right": "issue:44485", + "left": "issue:43723", + "right": "issue:44977", "accept": false, - "reason": "A wrong docstring is not the same issue as a RoPE implementation bug." + "reason": "Tokenizer loading issue vs flash-attention generation bug; not the same underlying problem." }, { "left": "issue:43122", "right": "issue:46209", "accept": false, - "reason": "Tokenization drift across versions vs FP8 inference failure on B200 are unrelated." + "reason": "Tokenization regression across versions is unrelated to FP8 inference on B200." }, { - "left": "issue:43232", - "right": "issue:46424", + "left": "issue:44222", + "right": "issue:44484", "accept": false, - "reason": "Both touch generation/cache, but one is a sync_gpus kwarg bug and the other is cache-length handling." + "reason": "FP8 save_pretrained serialization bug vs shard-size default behavior." }, { - "left": "issue:44222", - "right": "issue:44484", + "left": "issue:43232", + "right": "issue:46424", "accept": false, - "reason": "Both mention save_pretrained, but one is FP8 MoE serialization and the other is shard-size default behavior." + "reason": "Generation kwargs update bug vs cache-length configuration mismatch; related theme, different defect." }, { "left": "issue:44206", "right": "issue:45103", "accept": false, - "reason": "Feature extractor arg incompatibility and auto-docstring future-annotations crash are different subsystems." + "reason": "Feature extractor argument handling vs auto_docstring annotation crash; separate subsystems." }, { "left": "issue:46209", "right": "issue:46431", "accept": false, - "reason": "FP8 inference on B200 and missing chat template in processor are different failures." + "reason": "FP8 inference failure is unrelated to missing chat template support." }, { "left": "issue:44190", "right": "issue:46082", "accept": false, - "reason": "Local dataset loading in an example script is unrelated to LlamaConfig head_dim validation." - }, - { - "left": "issue:46305", - "right": "issue:46655", - "accept": false, - "reason": "Both are vision/streaming preprocessing bugs, but they affect different models and different calculations." - }, - { - "left": "issue:43475", - "right": "issue:46305", - "accept": false, - "reason": "SAM 3 video attribute failure and MoonshineStreaming mask off-by-one are unrelated." + "reason": "Dataset loading script issue vs configuration validation issue." }, { "left": "issue:44734", "right": "issue:46421", "accept": false, - "reason": "KV-cache continuation indexing crash and lazy cache recompilation are different cache bugs." + "reason": "Serve KV-cache indexing crash vs lazy cache recompilation; both cache-related but different bugs." }, { "left": "issue:43232", "right": "issue:45834", "accept": false, - "reason": "Generation kwargs bug and long-OCR index error are different code paths." + "reason": "Generation kwargs bug is not the same as Kosmos2.5 OCR long-input index error." }, { "left": "issue:44295", "right": "issue:44623", "accept": false, - "reason": "Buffer registration/read issue and processor.save_pretrained missing files are unrelated." + "reason": "Buffer read issue vs missing files on processor save; different persistence paths." }, { - "left": "issue:43723", - "right": "issue:44871", + "left": "issue:43475", + "right": "issue:46305", "accept": false, - "reason": "Tokenizer loading regression and eos_token_id config mismatch are different issues." + "reason": "SAM3 video attribute error vs MoonshineStreaming mask off-by-one; unrelated models and code." }, { - "left": "issue:45381", - "right": "issue:46305", + "left": "issue:46305", + "right": "issue:46655", + "accept": false, + "reason": "Both concern video handling, but one is a sliding-window mask bug and the other is padding calculation in a different processor." + }, + { + "left": "issue:43723", + "right": "issue:44871", "accept": false, - "reason": "Both involve position/mask logic in vision/video, but they are model-specific bugs with different symptoms and paths." + "reason": "Generic tokenizer loading regression vs eos_token_id config inconsistency." }, { "left": "issue:44263", "right": "issue:44360", "accept": false, - "reason": "The packet only shows a loose shared DSA indexer theme; the titles are not near-identical and there is no body/diff evidence that they are the same concrete bug." + "reason": "Same subsystem name, but torch.split return values and missing ReLU are distinct DSA-indexer bugs." }, { - "left": "issue:30990", - "right": "issue:38175", + "left": "issue:45381", + "right": "issue:46305", "accept": false, - "reason": "Sentence Transformers loading hang and siglip2 zero-probabilities are unrelated." + "reason": "Different models and failure modes: vision_position_ids vs sliding-window masking." }, { "left": "issue:44261", "right": "issue:45092", "accept": false, - "reason": "MLA rms_norm_eps precision issue and InternVL2 meta-initialization incompatibility are unrelated." - }, - { - "left": "issue:43644", - "right": "issue:46305", - "accept": false, - "reason": "Non-persistent buffer junk values and streaming mask off-by-one are different bugs." + "reason": "Layernorm epsilon precision issue is unrelated to meta-initialization checkpoint compatibility." }, { - "left": "issue:45092", - "right": "issue:45357", + "left": "issue:30990", + "right": "issue:38175", "accept": false, - "reason": "Both are model compatibility/save-load regressions, but they affect different models and failure modes." + "reason": "Sentence Transformers loading stall and zero-probability SigLIP issue are unrelated." }, { - "left": "issue:42907", - "right": "issue:46050", + "left": "issue:43644", + "right": "issue:46305", "accept": false, - "reason": "Quantized-model save/load corruption and sinq quantization breakage are distinct serialization/quantization problems." + "reason": "Non-persistent buffer junk initialization is not the same as MoonshineStreaming mask logic." }, { "left": "issue:42831", "right": "issue:45310", "accept": false, - "reason": "FP8 accuracy issues and Qwen3.5 from_pretrained errors are unrelated." + "reason": "FP8 accuracy concern vs Qwen3.5 from_pretrained failure; different defect class." }, { - "left": "issue:44279", - "right": "issue:46328", + "left": "issue:45092", + "right": "issue:45357", "accept": false, - "reason": "A generic dependency issue and a performance regression in SequenceFeatureExtractor.pad are not the same bug." + "reason": "Different models and code paths: InternVL2 meta-init vs Qwen3.5 visual-key save_pretrained regression." }, { "left": "issue:44263", "right": "issue:44373", "accept": false, - "reason": "The DSA indexer bug and a wrong position_ids docstring are unrelated." - }, - { - "left": "issue:44829", - "right": "issue:46002", - "accept": false, - "reason": "Degenerate training with flash_attention_3 and NaNs at checkout/init are different classification problems." + "reason": "DSA indexing behavior and a wrong docstring are not duplicates." } ] }, - "evaluator_result": { - "accept": false, - "feedback": "The overall clustering is mostly conservative, but the one accepted pair is overstated for the packet: issue:44263 and issue:44360 are only loosely linked by the shared DSA indexer theme in their titles, not by near-identical wording or any body/diff evidence shown here. The claim that they are a clear/true duplicate pair with the same concrete bug is stronger than the packet supports, so this should be rejected." - }, - "evaluator_used": true, - "retried": true, + "evaluator_result": null, + "evaluator_used": false, + "retried": false, "accepted_nontrivial_soft_edge": false, "error_kind": null, "error_message": null @@ -13976,7 +13894,7 @@ "source": "llm", "cache_hit": false, "model": "gpt-5.4-mini?service_tier=flex", - "cluster_id": "cluster-10000-472", + "cluster_id": "cluster-10000-473", "nodes": [ "issue:22355", "issue:27957", @@ -13996,13 +13914,13 @@ "issue:43122", "issue:43232", "issue:43408", - "issue:43479", "issue:43550", "issue:43576", "issue:43646", "issue:43756", "issue:43784", "issue:43824", + "issue:43976", "issue:44038", "issue:44060", "issue:44117", @@ -14013,29 +13931,30 @@ "issue:44297", "issue:44373", "issue:44483", - "issue:44488", "issue:44568", "issue:44704", - "issue:44913", + "issue:44829", "issue:44945", "issue:45278", "issue:45406", "issue:45593", "issue:45753", "issue:46002", + "issue:46050", "issue:46132", "issue:46369", "issue:46402" ], "soft_pairs": [ + "issue:42907|issue:46050", + "issue:44829|issue:46002", "issue:42831|issue:44038", - "issue:44483|issue:45406", "issue:22355|issue:27957", + "issue:44483|issue:45406", "issue:41628|issue:43824", "issue:43408|issue:46402", - "issue:43479|issue:44913", - "issue:22355|issue:28218", "issue:42831|issue:45753", + "issue:22355|issue:28218", "issue:42757|issue:45278", "issue:39401|issue:44568", "issue:42886|issue:43066", @@ -14045,221 +13964,220 @@ "issue:43408|issue:46369", "issue:44188|issue:44373", "issue:42796|issue:43122", + "issue:42371|issue:44263", "issue:43646|issue:46002", "issue:43550|issue:46132", - "issue:42371|issue:44263", "issue:44261|issue:44373", "issue:43576|issue:43784", "issue:44295|issue:44297", "issue:42907|issue:44945", "issue:30990|issue:33357", "issue:43756|issue:44060", - "issue:44117|issue:44704", - "issue:41628|issue:45278", - "issue:39401|issue:44488" + "issue:43976|issue:44945", + "issue:44117|issue:44704" ], - "prepared_review_unit_hash": "601ffb287cb3b9c34e97e861cc02b6e2f32099dac133861f4160578062154527", + "prepared_review_unit_hash": "94bfd6ba4b9b2579f68fafbf2b57b57e09a17f654bbaca6fe5b44e8e4e645b67", "component_budget": { - "node_count": 472, - "item_count": 472, - "soft_pair_count": 1732, - "serialized_chars": 898654, - "estimated_input_tokens": 224664, - "estimated_eval_tokens": 449584 + "node_count": 473, + "item_count": 473, + "soft_pair_count": 1738, + "serialized_chars": 901437, + "estimated_input_tokens": 225360, + "estimated_eval_tokens": 450976 }, "budget": { "node_count": 48, "item_count": 48, "soft_pair_count": 29, - "serialized_chars": 35807, - "estimated_input_tokens": 8952, - "estimated_eval_tokens": 18160 + "serialized_chars": 35882, + "estimated_input_tokens": 8971, + "estimated_eval_tokens": 18198 }, "overflow_policy": "truncate_then_skip", "trimmed": false, "aggressively_trimmed": false, "split": true, "analyst_result": { - "summary": "This cluster is mostly a loose collection of unrelated Transformers issues with rough theme overlap around imports, tokenizers, and model-specific regressions. None of the soft pairs look like the same underlying bug/change, so I would not merge any of them as duplicates.", - "confidence": 0.86, - "canonical_issue_reason": "issue:45278 is the broadest umbrella report in the set, covering multiple import failures after an upgrade; it is the best representative issue even though the cluster is not truly duplicate-heavy.", + "summary": "These are mostly unrelated closed issues from the same repo; the soft matches are largely superficial (shared keywords like tokenizer, import, save/load, or sam3) rather than the same underlying bug. No PRs are present.", + "confidence": 0.81, + "canonical_issue_reason": "Issue #45278 is the broadest umbrella-style report in the set and best reflects the cluster\u2019s common pattern of version-related breakage and import regressions.", "canonical_pr_reason": null, - "best_issue_reason": "issue:45278 is the most generally useful triage target because it can subsume a family of import-regression reports better than the narrower, model-specific issues.", + "best_issue_reason": "#45278 is the most representative global anchor because it summarizes a wide class of failures after upgrade, making it the most useful issue to group near-related regressions around.", "best_pr_reason": null, "soft_edge_verdicts": [ { - "left": "issue:42831", - "right": "issue:44038", + "left": "issue:42907", + "right": "issue:46050", "accept": false, - "reason": "Different models and different failure modes: FP8 accuracy regression vs Qwen3-VL-Moe / v5 bug report." + "reason": "Both involve quantization/save behavior, but one is dequantized Ministral/Devstral saving and the other is sinq quantized round-tripping; different code paths and symptoms." }, { - "left": "issue:44483", - "right": "issue:45406", + "left": "issue:44829", + "right": "issue:46002", + "accept": false, + "reason": "Both mention bad training/NaNs, but one is flash_attention_3 classification degeneration and the other is init/checkout NaNs; not the same bug." + }, + { + "left": "issue:42831", + "right": "issue:44038", "accept": false, - "reason": "One is a chat-completions API acceptance crash; the other is a Gemma4Processor attribute error in serve." + "reason": "FineGrainedFP8 accuracy loss and Qwen3-VL-Moe v5 breakage are unrelated model-specific issues." }, { "left": "issue:22355", "right": "issue:27957", "accept": false, - "reason": "Importing transformers.onnx and XLMRoberta + FlashAttention2 are unrelated problems." + "reason": "Module import failure for transformers.onnx is unrelated to XLMRoberta Flash Attention 2 behavior." + }, + { + "left": "issue:44483", + "right": "issue:45406", + "accept": false, + "reason": "Chat-completions request validation and Gemma4Processor attribute crash are different serving/API failures." }, { "left": "issue:41628", "right": "issue:43824", "accept": false, - "reason": "Both are import errors, but for different symbols/modules and likely different regressions." + "reason": "Both are import errors, but for different symbols and different models; too broad to be the same defect." }, { "left": "issue:43408", "right": "issue:46402", "accept": false, - "reason": "Both mention SAM3, but one is a model-type warning and the other is missing text-encoder weights." + "reason": "Both are SAM3-related, but one is a model-type warning and the other is missing text-encoder weights; not the same underlying issue." }, { - "left": "issue:43479", - "right": "issue:44913", + "left": "issue:42831", + "right": "issue:45753", "accept": false, - "reason": "Default vision/audio config initialization and rotary_pct reload behavior are separate config bugs." + "reason": "FineGrainedFP8 accuracy regression and Qwen3_5 infinite loop on one image are unrelated." }, { "left": "issue:22355", "right": "issue:28218", "accept": false, - "reason": "A missing module import is not the same as tokenizer spacing after added tokens." - }, - { - "left": "issue:42831", - "right": "issue:45753", - "accept": false, - "reason": "FP8 accuracy issues and an infinite loop on a specific image are different defects." + "reason": "ONNX import failure and added-token spacing regression are separate tokenizer/export problems." }, { "left": "issue:42757", "right": "issue:45278", "accept": false, - "reason": "Both are import-related, but one is a specific offline_mode symbol import and the other is a broad upgrade regression." + "reason": "A specific is_offline_mode import error is not the same bug as a broad report of many import errors after upgrade." }, { "left": "issue:39401", "right": "issue:44568", "accept": false, - "reason": "Wrong offset mappings and missing BOS/EOS insertion are distinct tokenizer behaviors." + "reason": "Wrong offset_mapping in Qwen3 and missing BOS/EOS insertion in mdeberta-v3 are different tokenizer behaviors." }, { "left": "issue:42886", "right": "issue:43066", "accept": false, - "reason": "Offline cache loading failure and wrong tokenizer decoder type are different tokenizer regressions." + "reason": "Offline cache-loading failure and wrong tokenizer decoder type are distinct tokenizer initialization problems." }, { "left": "issue:27957", "right": "issue:30990", "accept": false, - "reason": "Flash Attention 2 on XLMRoberta and sentence-transformers loading hangs are unrelated." + "reason": "Flash Attention 2 on XLMRoberta and Sentence-Transformers loading hang are unrelated." }, { "left": "issue:43232", "right": "issue:45593", "accept": false, - "reason": "Generation keyword handling after sync_gpus is unrelated to auxiliary-loss behavior in D-FINE." + "reason": "Generation kwargs after sync_gpus and D-FINE auxiliary-loss handling do not share a code path." }, { "left": "issue:42886", "right": "issue:42994", "accept": false, - "reason": "Tokenizer cache loading offline and quantized model saving are different code paths." + "reason": "Tokenizer offline cache loading and quantized model saving are different persistence issues." }, { "left": "issue:43408", "right": "issue:46369", "accept": false, - "reason": "A model-type warning for sam3_tracker is not the same as a memory overflow in sam3-litetext." + "reason": "SAM3 model-type mismatch warning and SAM3 litetext memory overflow are unrelated failures." }, { "left": "issue:44188", "right": "issue:44373", "accept": false, - "reason": "Torch.compile attention-kernel branching and a wrong docstring for position_ids are not duplicates." + "reason": "Attention-kernel divergence and a docstring typo for position_ids are not the same problem." }, { "left": "issue:42796", "right": "issue:43122", "accept": false, - "reason": "Tokenizer BPE-marker handling and cross-version tokenization differences are related only at a high level, not the same bug." + "reason": "Ministral BPE-marker handling and version-to-version tokenization changes are too general to treat as one bug." + }, + { + "left": "issue:42371", + "right": "issue:44263", + "accept": false, + "reason": "TF32 settings guidance and GlmMoeDsaIndexer torch.split behavior are unrelated." }, { "left": "issue:43646", "right": "issue:46002", "accept": false, - "reason": "Custom model initialization breakage and NaNs in classification heads on init are different failure modes." + "reason": "Custom model initialization breakage and NaNs in classification heads are different regressions." }, { "left": "issue:43550", "right": "issue:46132", "accept": false, - "reason": "torch.compile + SDPA on Bamba and AttentionInterface.register behavior changes are separate issues." - }, - { - "left": "issue:42371", - "right": "issue:44263", - "accept": false, - "reason": "TF32 settings guidance and torch.split return values in an indexer are unrelated." + "reason": "torch.compile with SDPA on Bamba and AttentionInterface.register behavior changes are different attention issues." }, { "left": "issue:44261", "right": "issue:44373", "accept": false, - "reason": "Missing rms_norm_eps in MLA q_a_layernorm and a wrong position_ids docstring are not the same bug." + "reason": "MLA q_a_layernorm precision/config bug and a docstring issue are not the same defect." }, { "left": "issue:43576", "right": "issue:43784", "accept": false, - "reason": "A broken env command and an import NameError in sentence-transformers are unrelated regressions." + "reason": "Broken env command and a sentence-transformers import NameError are unrelated." }, { "left": "issue:44295", "right": "issue:44297", "accept": false, - "reason": "Reading position_ids after buffer registration and tokenizer_class mismatch on save_pretrained are distinct problems." + "reason": "Reading a registered position_ids buffer and tokenizer_class mismatch on save are separate serialization bugs." }, { "left": "issue:42907", "right": "issue:44945", "accept": false, - "reason": "Quantized model saving failure and incorrect output under pipeline parallelism do not share the same code path." + "reason": "Quantized saving/dequantization and pipeline-parallelism output errors are unrelated." }, { "left": "issue:30990", "right": "issue:33357", "accept": false, - "reason": "Loading stuck and a MacOS bus error are both runtime problems but not the same underlying bug." + "reason": "Loading hang in Sentence-Transformers and MacOS bus error on CLIP are different failures." }, { "left": "issue:43756", "right": "issue:44060", "accept": false, - "reason": "Smollm3 RoPE layer loss and Qwen3-Next tied-weights warnings are unrelated architecture issues." - }, - { - "left": "issue:44117", - "right": "issue:44704", - "accept": false, - "reason": "TOKENIZER_MAPPING_NAMES returning None and AutoProcessor not forwarding kwargs are different API bugs." + "reason": "Smollm3 RoPE-layer bug and Qwen3-Next tied-weights warning are unrelated model-definition issues." }, { - "left": "issue:41628", - "right": "issue:45278", + "left": "issue:43976", + "right": "issue:44945", "accept": false, - "reason": "An AutoImageProcessor import error is narrower than a broad many-import-errors regression; not the same bug." + "reason": "Python-version compatibility breakage and pipeline-parallelism output corruption do not match." }, { - "left": "issue:39401", - "right": "issue:44488", + "left": "issue:44117", + "right": "issue:44704", "accept": false, - "reason": "Wrong offset mapping and failure to load a specific model are unrelated tokenizer/model-loading issues." + "reason": "TOKENIZER_MAPPING_NAMES returning None and AutoProcessor not forwarding kwargs are different from_pretrained/cached-file issues, but not the same bug." } ] }, @@ -14281,7 +14199,7 @@ "source": "llm", "cache_hit": false, "model": "gpt-5.4-mini?service_tier=flex", - "cluster_id": "cluster-10000-472", + "cluster_id": "cluster-10000-473", "nodes": [ "issue:22355", "issue:29127", @@ -14289,11 +14207,11 @@ "issue:36296", "issue:38617", "issue:39401", + "issue:41628", "issue:42175", "issue:42757", "issue:42898", "issue:42907", - "issue:43066", "issue:43116", "issue:43122", "issue:43278", @@ -14304,7 +14222,6 @@ "issue:43784", "issue:43824", "issue:43937", - "issue:43976", "issue:44038", "issue:44188", "issue:44222", @@ -14313,6 +14230,7 @@ "issue:44351", "issue:44360", "issue:44485", + "issue:44488", "issue:44521", "issue:44625", "issue:44704", @@ -14333,22 +14251,23 @@ "issue:46614" ], "soft_pairs": [ - "issue:43976|issue:44945", - "issue:42907|issue:44964", + "issue:39401|issue:44488", + "issue:41628|issue:45278", "issue:43784|issue:45600", + "issue:42907|issue:44964", "issue:42907|issue:43937", "issue:44360|issue:44485", "issue:44625|issue:44829", "issue:39401|issue:44521", "issue:44704|issue:45290", + "issue:44038|issue:44297", "issue:43122|issue:46326", "issue:43122|issue:44521", - "issue:44038|issue:44297", - "issue:43408|issue:43475", "issue:43116|issue:45200", - "issue:42907|issue:43452", + "issue:43408|issue:43475", "issue:44038|issue:44295", "issue:39401|issue:46614", + "issue:42907|issue:43452", "issue:44038|issue:44222", "issue:22355|issue:29159", "issue:45216|issue:45478", @@ -14360,232 +14279,231 @@ "issue:42175|issue:44938", "issue:22355|issue:29127", "issue:43493|issue:45753", - "issue:43937|issue:44485", "issue:44188|issue:46328", "issue:38617|issue:45278", "issue:42898|issue:46614", - "issue:43066|issue:44779" + "issue:43937|issue:44485" ], - "prepared_review_unit_hash": "f2095877dd810960400b75c20e7cc4129a7190d144c453b56ef459821443bd8f", + "prepared_review_unit_hash": "e06a0fd47d28f326f7ca8b2655c7b0b251adb3249b192ed068d914b5111976a4", "component_budget": { - "node_count": 472, - "item_count": 472, - "soft_pair_count": 1732, - "serialized_chars": 898654, - "estimated_input_tokens": 224664, - "estimated_eval_tokens": 449584 + "node_count": 473, + "item_count": 473, + "soft_pair_count": 1738, + "serialized_chars": 901437, + "estimated_input_tokens": 225360, + "estimated_eval_tokens": 450976 }, "budget": { "node_count": 48, "item_count": 48, "soft_pair_count": 32, - "serialized_chars": 36985, - "estimated_input_tokens": 9247, - "estimated_eval_tokens": 18750 + "serialized_chars": 36997, + "estimated_input_tokens": 9250, + "estimated_eval_tokens": 18756 }, "overflow_policy": "truncate_then_skip", "trimmed": false, "aggressively_trimmed": false, "split": true, "analyst_result": { - "summary": "This cluster is mostly a loose topical grouping of unrelated issues: import errors, tokenizer regressions, multimodal/chat-template bugs, training/runtime problems, and version-compatibility reports. I do not see any true duplicate pair among the soft edges.", - "confidence": 0.29, - "canonical_issue_reason": "If one issue must stand in for the cluster, issue 45278 is the broadest umbrella-style report (many import errors after an upgrade), but it is still not a real duplicate of the others.", + "summary": "This set is a loose cluster of mostly unrelated regression reports (imports, tokenization, multimodal, save/load, and training bugs). I did not find any soft pair that looks like the same underlying issue, so all candidate links are rejected.", + "confidence": 0.84, + "canonical_issue_reason": "issue:43122 is the best representative because it is a high-signal tokenizer regression report and one of the more central version-change bugs in the set.", "canonical_pr_reason": null, - "best_issue_reason": "Issue 45278 is the best representative issue because it is the most general and highest-level symptom in the set; the rest are narrower, distinct bugs.", + "best_issue_reason": "issue:43122 is the most suitable issue to anchor this cluster as it captures the broadest recurring theme here: tokenizer behavior changing across versions.", "best_pr_reason": null, "soft_edge_verdicts": [ { - "left": "issue:43976", - "right": "issue:44945", + "left": "issue:39401", + "right": "issue:44488", "accept": false, - "reason": "Different failures: Python version compatibility versus incorrect output under pipeline parallelism." + "reason": "Both are tokenizer-related, but one is an offset_mapping bug and the other is a model-loading failure; different concrete problems." }, { - "left": "issue:42907", - "right": "issue:44964", + "left": "issue:41628", + "right": "issue:45278", "accept": false, - "reason": "Both are model-loading related, but one is a save/load dequantization problem and the other is a multimodal model load failure." + "reason": "One is a specific AutoImageProcessor import error, the other is a broad collection of import errors after upgrade; too general to treat as the same bug." }, { "left": "issue:43784", "right": "issue:45600", "accept": false, - "reason": "Different import breakages in different integration paths: sentence-transformers import failure versus stale auto_mappings references." + "reason": "Both involve imports, but one is a missing nn symbol at import time and the other is stale auto_mappings references; different code paths." + }, + { + "left": "issue:42907", + "right": "issue:44964", + "accept": false, + "reason": "These are separate model save/load failures for different models and APIs." }, { "left": "issue:42907", "right": "issue:43937", "accept": false, - "reason": "One is about dequantized model saving, the other about invalid GenerationConfig for GLM-5." + "reason": "Dequantized save_pretrained failure and GenerationConfig validation failure are unrelated." }, { "left": "issue:44360", "right": "issue:44485", "accept": false, - "reason": "Different topics: DSA/ReLU discussion versus GLM-5 RoPE implementation." + "reason": "DSA indexer/ReLU discussion is unrelated to GLM-5 RoPE implementation." }, { "left": "issue:44625", "right": "issue:44829", "accept": false, - "reason": "Different models and symptoms: config propagation bug versus degenerate training with flash_attention_3." + "reason": "Different bug classes: config propagation vs degenerate training with flash attention 3." }, { "left": "issue:39401", "right": "issue:44521", "accept": false, - "reason": "Tokenizer offset mapping bug versus multimodal assistant mask bug; same area, different code path." + "reason": "Both touch tokenization, but offset mapping and multimodal assistant mask generation are different failures." }, { "left": "issue:44704", "right": "issue:45290", "accept": false, - "reason": "AutoProcessor cached_file kwargs bug is unrelated to chat-template crashes on tool-call assistant messages." + "reason": "AutoProcessor kwarg forwarding and apply_chat_template crashes are different behaviors in different functions." }, { - "left": "issue:43122", - "right": "issue:46326", + "left": "issue:44038", + "right": "issue:44297", "accept": false, - "reason": "Both involve tokenization behavior, but one is version-to-version output differences and the other is continue_final_message on mistral_common tokenizers." + "reason": "A Qwen3-VL-Moe model bug is not the same as tokenizer_class metadata mismatch during save_pretrained." }, { "left": "issue:43122", - "right": "issue:44521", - "accept": false, - "reason": "Tokenizer behavior regression versus multimodal assistant_mask generation; not the same bug." - }, - { - "left": "issue:44038", - "right": "issue:44297", + "right": "issue:46326", "accept": false, - "reason": "Qwen3-VL-Moe bug versus tokenizer_class metadata mismatch in save_pretrained; unrelated failures." + "reason": "Both mention tokenizer behavior, but one is a general v5 regression and the other is a mistral_common continue_final_message bug." }, { - "left": "issue:43408", - "right": "issue:43475", + "left": "issue:43122", + "right": "issue:44521", "accept": false, - "reason": "Both are SAM 3 video related, but one is a model-type warning and the other is a missing attribute in the vision encoder output." + "reason": "Version-regression tokenization differences are not the same as assistant_masks being all zero for multimodal inputs." }, { "left": "issue:43116", "right": "issue:45200", "accept": false, - "reason": "Example-script multi-label output issue versus Gemma 4 mm_token_type_ids defaulting; different symptoms and fixes." + "reason": "Multi-label classification example output and mm_token_type_ids handling are unrelated." }, { - "left": "issue:42907", - "right": "issue:43452", + "left": "issue:43408", + "right": "issue:43475", "accept": false, - "reason": "Dequantized model save failure versus gguf_file loading failures for AutoTokenizer/AutoModel; different mechanisms." + "reason": "Model-type warning and missing fpn_position_embeddings attribute are different SAM3 video issues." }, { "left": "issue:44038", "right": "issue:44295", "accept": false, - "reason": "Different bugs: Qwen3-VL-Moe issue versus position_ids buffer access error." + "reason": "Qwen3-VL-Moe behavior and position_ids buffer reading are unrelated." }, { "left": "issue:39401", "right": "issue:46614", "accept": false, - "reason": "Both touch tokenization, but offset_mapping regression and stale phonemizer backend state are separate problems." + "reason": "Wrong offset_mapping and stale phonemizer language backend are different tokenizer subsystems." + }, + { + "left": "issue:42907", + "right": "issue:43452", + "accept": false, + "reason": "save_pretrained on dequantized models and gguf_file load failures are distinct problems." }, { "left": "issue:44038", "right": "issue:44222", "accept": false, - "reason": "Different FP8/Qwen3-VL-Moe issues; one is model-specific, the other is save_pretrained for moe FP8." + "reason": "MoE model bug and FP8 save_pretrained issue do not describe the same concrete failure." }, { "left": "issue:22355", "right": "issue:29159", "accept": false, - "reason": "Older import error for transformers.onnx versus slow/fast tokenizer inconsistency; unrelated." + "reason": "onnx module import failure and slow/fast tokenizer inconsistency are unrelated." }, { "left": "issue:45216", "right": "issue:45478", "accept": false, - "reason": "Both concern Qwen3.5 save/from_pretrained regressions, but one is saved checkpoint correctness and the other is a from_pretrained error; not clearly one bug." + "reason": "Both involve Qwen3.5 and version 5.x, but one is save_pretrained corruption and the other is from_pretrained loading failure." }, { "left": "issue:43824", "right": "issue:44351", "accept": false, - "reason": "Both are import errors, but for different missing symbols and different release breakages." + "reason": "Both are missing imports, but for different symbols and likely different export issues." }, { "left": "issue:43122", "right": "issue:44779", "accept": false, - "reason": "Both are tokenizer regressions across versions, but they affect different models and distinct tokenization logic." + "reason": "Both are tokenizer regressions, but they are model-specific and do not look like the same code-path defect." }, { "left": "issue:42757", "right": "issue:44351", "accept": false, - "reason": "Shared import-error pattern only; missing is_offline_mode from huggingface_hub is not the same as missing HybridCache from transformers." + "reason": "Different missing imports from different packages/symbols; not the same underlying bug." }, { "left": "issue:36296", "right": "issue:43116", "accept": false, - "reason": "Tensor parallel training bug versus example-script multi-label classification issue; separate subsystems." + "reason": "Tensor parallel training bug and example script multi-label classification failure are unrelated." }, { "left": "issue:43278", "right": "issue:44945", "accept": false, - "reason": "Embedding dtype drift during evaluate is unrelated to pipeline parallelism causing wrong LLM output." + "reason": "Embedding dtype drift in evaluate and incorrect pipeline-parallel output are different training/runtime issues." }, { "left": "issue:42175", "right": "issue:44938", "accept": false, - "reason": "Different compatibility failures: missing TensorFlow backend in pip install and Python 3.14 load failure." + "reason": "Backend packaging omission and Python 3.14 load failure are not the same bug." }, { "left": "issue:22355", "right": "issue:29127", "accept": false, - "reason": "No relation between an ONNX import error and a LayoutLMv3 error-message clarity issue." + "reason": "Transformers.onnx missing module and LayoutLMv3 error messaging are unrelated." }, { "left": "issue:43493", "right": "issue:45753", "accept": false, - "reason": "Different Qwen3.5 issues: SigLIP2 implementation discrepancy versus an infinite-loop image case." - }, - { - "left": "issue:43937", - "right": "issue:44485", - "accept": false, - "reason": "GenerationConfig validation bug versus GLM-5 RoPE implementation discussion; not the same defect." + "reason": "SigLIP2 implementation discrepancy and Qwen3_5 infinite-loop image bug are different model-specific issues." }, { "left": "issue:44188", "right": "issue:46328", "accept": false, - "reason": "Attention kernel divergence under torch.compile is unrelated to SequenceFeatureExtractor.pad performance behavior." + "reason": "torch.compile attention-kernel divergence is unrelated to SequenceFeatureExtractor.pad list conversion overhead." }, { "left": "issue:38617", "right": "issue:45278", "accept": false, - "reason": "Both are import-related, but one is a missing symbol in configuration_utils and the other is broader post-upgrade import failures." + "reason": "A specific layer_type_validation import error is not the same as the broader multi-import regression report." }, { "left": "issue:42898", "right": "issue:46614", "accept": false, - "reason": "Different tokenizer-related bugs: v5 clean_up_tokenization_spaces behavior versus stale phonemizer language state." + "reason": "Tokenization cleanup behavior in v5 and phonemizer language-state leakage are separate issues." }, { - "left": "issue:43066", - "right": "issue:44779", + "left": "issue:43937", + "right": "issue:44485", "accept": false, - "reason": "Different tokenizer regressions with different root causes: decoder type selection versus Deepseek tokenization output changes." + "reason": "GenerationConfig validation for GLM-5 and GLM-5 RoPE implementation are different problems." } ] }, @@ -14607,7 +14525,7 @@ "source": "llm", "cache_hit": false, "model": "gpt-5.4-mini?service_tier=flex", - "cluster_id": "cluster-10000-472", + "cluster_id": "cluster-10000-473", "nodes": [ "issue:28348", "issue:33453", @@ -14617,6 +14535,7 @@ "issue:42673", "issue:42796", "issue:43064", + "issue:43066", "issue:43116", "issue:43441", "issue:43493", @@ -14635,7 +14554,6 @@ "issue:44291", "issue:44355", "issue:44380", - "issue:44466", "issue:44625", "issue:44704", "issue:44779", @@ -14659,26 +14577,27 @@ "issue:46655" ], "soft_pairs": [ + "issue:43066|issue:44779", "issue:42673|issue:45072", - "issue:43611|issue:46002", - "issue:43116|issue:43493", "issue:43976|issue:45446", + "issue:43116|issue:43493", + "issue:43611|issue:46002", "issue:44355|issue:46231", "issue:42371|issue:43874", "issue:44038|issue:45478", "issue:44625|issue:46002", "issue:45357|issue:45478", - "issue:43992|issue:44704", "issue:45083|issue:45446", "issue:44075|issue:44936", "issue:42222|issue:43441", "issue:43116|issue:43825", "issue:28348|issue:46640", + "issue:43992|issue:44704", "issue:42673|issue:42796", "issue:43064|issue:43673", "issue:42673|issue:46655", - "issue:43937|issue:45563", "issue:43827|issue:45446", + "issue:43937|issue:45563", "issue:33453|issue:39401", "issue:44380|issue:46077", "issue:45356|issue:46614", @@ -14686,225 +14605,218 @@ "issue:43673|issue:46328", "issue:45071|issue:46614", "issue:44291|issue:45071", - "issue:42673|issue:43906", "issue:45405|issue:45446", - "issue:44466|issue:46002", + "issue:42673|issue:43906", "issue:39401|issue:44779" ], - "prepared_review_unit_hash": "a2978a37c580ef3066697480713265f181a06ed4f170d4fdc04001b9c9235139", + "prepared_review_unit_hash": "79d39a53ab9124f2e6fff626aa086a589f126ab3e33d410f956c47c16ba85333", "component_budget": { - "node_count": 472, - "item_count": 472, - "soft_pair_count": 1732, - "serialized_chars": 898654, - "estimated_input_tokens": 224664, - "estimated_eval_tokens": 449584 + "node_count": 473, + "item_count": 473, + "soft_pair_count": 1738, + "serialized_chars": 901437, + "estimated_input_tokens": 225360, + "estimated_eval_tokens": 450976 }, "budget": { "node_count": 48, "item_count": 48, "soft_pair_count": 31, - "serialized_chars": 36885, - "estimated_input_tokens": 9222, - "estimated_eval_tokens": 18700 + "serialized_chars": 36793, + "estimated_input_tokens": 9199, + "estimated_eval_tokens": 18654 }, "overflow_policy": "truncate_then_skip", "trimmed": false, "aggressively_trimmed": false, "split": true, "analyst_result": { - "summary": "No soft-edge pair looks like the same exact bug/change under a strict duplicate standard; most are only loosely related by subsystem or model family. The strongest standalone issue is the v5 `base_model_prefix` loading regression.", - "confidence": 0.72, - "canonical_issue_reason": "Issue 43611 is the most concrete and broadly representative regression here: it names a specific v5 loading failure path (`base_model_prefix`) with a clear, actionable failure mode.", + "summary": "This cluster is not a duplicate set: the items span many unrelated tokenizer, model-loading, training, attention, and documentation bugs. None of the proposed soft edges look like the same underlying bug or change.", + "confidence": 0.96, + "canonical_issue_reason": "No single canonical issue: the items are heterogeneous, and there is no clear duplicate family tying them together.", "canonical_pr_reason": null, - "best_issue_reason": "Issue 43611 is the best overall issue candidate because it is specific, reproducible, and framed as a real code-path regression rather than a vague symptom or documentation mismatch.", + "best_issue_reason": "issue:44779 is the clearest stand-alone bug report (a concrete tokenizer regression with a v4/v5 behavior difference), but it does not serve as a duplicate anchor for the rest of the set.", "best_pr_reason": null, "soft_edge_verdicts": [ + { + "left": "issue:43066", + "right": "issue:44779", + "accept": false, + "reason": "Both are tokenizer regressions, but one is a decoder-type loading bug and the other is incorrect DeepSeek tokenizer output; different failure modes." + }, { "left": "issue:42673", "right": "issue:45072", "accept": false, - "reason": "Both mention bad runtime behavior, but one is VRAM leakage in Qwen3 dataloader-thread use and the other is bfloat16 dtype mismatches in CI; different failure modes and code paths." + "reason": "VRAM leak in multi-threaded Qwen3 inference vs dtype mismatch in bfloat16 inference; unrelated code paths." }, { - "left": "issue:43611", - "right": "issue:46002", + "left": "issue:43976", + "right": "issue:45446", "accept": false, - "reason": "One is a v5 model-loading regression around `base_model_prefix`; the other is NaNs after checkout/init in classification heads. Not the same underlying bug." + "reason": "Python-version compatibility bug vs PyTorch version gate/import check bug; not the same issue." }, { "left": "issue:43116", "right": "issue:43493", "accept": false, - "reason": "Multi-label empty outputs in an example script and SigLIP2 HF-vs-JAX discrepancy are unrelated behaviors in different areas." + "reason": "Example-script multi-label empties vs SigLIP2 implementation discrepancy; different models and symptoms." }, { - "left": "issue:43976", - "right": "issue:45446", + "left": "issue:43611", + "right": "issue:46002", "accept": false, - "reason": "Both involve version checks, but one is a Python-version compatibility problem and the other is a PyTorch import gate for flex attention; not the same bug." + "reason": "Model loading/base_model_prefix regression vs NaNs in classification heads; unrelated bugs." }, { "left": "issue:44355", "right": "issue:46231", "accept": false, - "reason": "Compiled Python file errors and a documentation parameter-name typo are unrelated." + "reason": "Compiled Python file errors vs documentation parameter-name mismatch; not the same underlying problem." }, { "left": "issue:42371", "right": "issue:43874", "accept": false, - "reason": "TF32 settings guidance and a missing image-processor method are completely different issues." + "reason": "TF32 behavior settings API issue vs missing image-processor method; unrelated." }, { "left": "issue:44038", "right": "issue:45478", "accept": false, - "reason": "Both involve Qwen MoE models under Transformers v5, but the titles point to different model variants and different symptoms; insufficient evidence of one concrete shared bug." + "reason": "Both involve Qwen3 MoE, but they hit different model variants and different from_pretrained/FlashAttention failures." }, { "left": "issue:44625", "right": "issue:46002", "accept": false, - "reason": "`num_labels` propagation for Qwen3.5 and NaNs in classification heads are different problems." + "reason": "num_labels propagation bug vs NaNs on checkout/init; distinct failures." }, { "left": "issue:45357", "right": "issue:45478", "accept": false, - "reason": "One is incorrect visual encoder key saving; the other is a `from_pretrained` error for Qwen3.5 MoE. Save/load regressions in the same family are not enough to merge them." - }, - { - "left": "issue:43992", - "right": "issue:44704", - "accept": false, - "reason": "Missing `embed_tokens.weight` during `from_pretrained` and missing kwargs passed to `cached_file` are distinct loading-path bugs." + "reason": "save_pretrained visual-key regression vs from_pretrained load error; separate save/load code paths." }, { "left": "issue:45083", "right": "issue:45446", "accept": false, - "reason": "Unexpected feature-length calculation and a PyTorch version check bug are unrelated." + "reason": "qwen3_omni_moe helper behavior vs flex_attention import version check; unrelated." }, { "left": "issue:44075", "right": "issue:44936", "accept": false, - "reason": "Wrong SGD args and `trainer.evaluate()` failing after `trainer.train()` are different training issues." + "reason": "Optimizer-argument handling bug vs train-then-evaluate lifecycle failure; not the same bug." }, { "left": "issue:42222", "right": "issue:43441", "accept": false, - "reason": "Broken VitPose imports and Ministral flash-attention failure are different model-specific problems." + "reason": "VitPose model breakage vs Ministral FlashAttention failure; no concrete overlap." }, { "left": "issue:43116", "right": "issue:43825", "accept": false, - "reason": "Example-script multi-label output bug and a `pipeline()` error-message/docs issue are not the same change." + "reason": "Multi-label example script returning empty results vs pipeline translation-support error message; different surfaces." }, { "left": "issue:28348", "right": "issue:46640", "accept": false, - "reason": "GPT-2 flash-attention support and a generic flash-attention issue are too broad and not demonstrably the same bug." + "reason": "FlashAttention support request for GPT-2 vs a general Flash Attention 2 issue; too broad and not the same bug." + }, + { + "left": "issue:43992", + "right": "issue:44704", + "accept": false, + "reason": "Missing embed token weights vs AutoProcessor kwargs forwarding; unrelated loading bugs." }, { "left": "issue:42673", "right": "issue:42796", "accept": false, - "reason": "VRAM leakage in Qwen3 dataloader threads and tokenizer BPE-marker handling are unrelated." + "reason": "VRAM leak in multi-threaded inference vs tokenizer BPE-marker handling; different subsystems." }, { "left": "issue:43064", "right": "issue:43673", "accept": false, - "reason": "Optimizer-state corruption under FSDP2/PEFT and missing generation cache during chunked prefill are different distributed-training vs generation bugs." + "reason": "FSDP/PEFT optimizer-state corruption vs missing GenerationMixin cache in chunked prefill; unrelated." }, { "left": "issue:42673", "right": "issue:46655", "accept": false, - "reason": "VRAM leakage in Qwen3 versus video padding calculation in SmolVLM are unrelated." + "reason": "VRAM leak vs video padding-calculation bug; not the same issue." }, { - "left": "issue:43937", - "right": "issue:45563", + "left": "issue:43827", + "right": "issue:45446", "accept": false, - "reason": "Invalid `GenerationConfig` and a stale warning for `num_return_sequences` are different generation issues." + "reason": "Docs still referencing pipeline() vs PyTorch version check bug; unrelated." }, { - "left": "issue:43827", - "right": "issue:45446", + "left": "issue:43937", + "right": "issue:45563", "accept": false, - "reason": "Docs still referencing `pipeline()` and an incorrect PyTorch version check are unrelated." + "reason": "Invalid GenerationConfig vs stale warning for num_return_sequences; different behaviors." }, { "left": "issue:33453", "right": "issue:39401", "accept": false, - "reason": "Tokenizer loading regression and Qwen3 offset-mapping bug are both tokenizer-related but clearly different defects." + "reason": "Both are tokenizer regressions, but one is loading regression and the other is wrong offset mapping; different concrete bugs." }, { "left": "issue:44380", "right": "issue:46077", "accept": false, - "reason": "GPT-2 attention scaling ignored in SDPA/FlashAttention and Gemma4 bidirectional-mask handling are different attention-path bugs." + "reason": "GPT-2 attention scaling ignored with SDPA/FlashAttention vs Gemma4 mask construction bug; different models and logic." }, { "left": "issue:45356", "right": "issue:46614", "accept": false, - "reason": "Kimi-K2.5 tokenizer regression and Wav2Vec2 phonemizer stale language state are different tokenizer/phonemizer issues." + "reason": "Tokenizer codec/warning regression vs phonemizer backend state bug; unrelated tokenizer/audio issues." }, { "left": "issue:44117", "right": "issue:45588", "accept": false, - "reason": "A mapping-name lookup returning `None` and a flash-attention `AttributeError` on `s_aux=None` are unrelated." + "reason": "Tokenizer mapping None assumption vs flash_attention.py crash on s_aux=None; not the same code path." }, { "left": "issue:43673", "right": "issue:46328", "accept": false, - "reason": "Generation cache handling and `SequenceFeatureExtractor.pad` inefficiency are separate concerns." + "reason": "Missing cache in generation vs inefficient padding conversion in feature extractor; different kind of issue." }, { "left": "issue:45071", "right": "issue:46614", "accept": false, - "reason": "PretrainedConfig type checking and phonemizer backend state are unrelated." + "reason": "PretrainedConfig type-checking regression vs stale phonemizer language backend; unrelated." }, { "left": "issue:44291", "right": "issue:45071", "accept": false, - "reason": "`init_empty_weights` unexpected argument and `PretrainedConfig` type checking are different compatibility regressions." - }, - { - "left": "issue:42673", - "right": "issue:43906", - "accept": false, - "reason": "VRAM leakage in Qwen3 threads and a reproduction of an older issue are not the same defect." + "reason": "init_empty_weights/_is_hf_initialized argument bug vs PretrainedConfig type-checking breakage; different model-loading failures." }, { "left": "issue:45405", "right": "issue:45446", "accept": false, - "reason": "A PEFT minimum-version bump and a PyTorch version check bug are different dependency issues." + "reason": "PEFT version pin problem vs flex_attention import version check; unrelated dependency/version issues." }, { - "left": "issue:44466", - "right": "issue:46002", - "accept": false, - "reason": "`lm_head.weight` serialization inconsistency and NaNs in classification heads are unrelated model state problems." - }, - { - "left": "issue:39401", - "right": "issue:44779", + "left": "issue:42673", + "right": "issue:43906", "accept": false, - "reason": "Wrong Qwen3 offset mapping and incorrect Deepseek tokenizer results are different tokenizer regressions." + "reason": "VRAM leak in Qwen3 training/inference threads vs an isolated reproduction of another issue; no evidence of same bug." } ] }, @@ -14926,12 +14838,11 @@ "source": "llm", "cache_hit": false, "model": "gpt-5.4-mini?service_tier=flex", - "cluster_id": "cluster-10000-472", + "cluster_id": "cluster-10000-473", "nodes": [ "issue:38617", "issue:39401", "issue:41628", - "issue:42175", "issue:42222", "issue:42371", "issue:42617", @@ -14940,7 +14851,6 @@ "issue:43122", "issue:43408", "issue:43450", - "issue:43673", "issue:43742", "issue:43937", "issue:43986", @@ -14950,6 +14860,7 @@ "issue:44261", "issue:44263", "issue:44410", + "issue:44466", "issue:44484", "issue:44485", "issue:44493", @@ -14969,6 +14880,7 @@ "issue:45542", "issue:45561", "issue:45656", + "issue:46002", "issue:46178", "issue:46209", "issue:46231", @@ -14978,6 +14890,7 @@ "issue:46617" ], "soft_pairs": [ + "issue:44466|issue:46002", "issue:43937|issue:44977", "issue:43450|issue:43986", "issue:44410|issue:45310", @@ -14989,10 +14902,10 @@ "issue:38617|issue:41628", "issue:43122|issue:46298", "issue:42617|issue:45341", - "issue:43408|issue:43742", "issue:42222|issue:46178", - "issue:45362|issue:46328", "issue:42371|issue:44261", + "issue:45362|issue:46328", + "issue:43408|issue:43742", "issue:44877|issue:46231", "issue:42371|issue:46617", "issue:44484|issue:45310", @@ -15004,219 +14917,211 @@ "issue:44075|issue:45656", "issue:43066|issue:43122", "issue:42222|issue:45542", - "issue:43097|issue:44493", "issue:44263|issue:44485", - "issue:42175|issue:42222", - "issue:43673|issue:45341" + "issue:43097|issue:44493" ], - "prepared_review_unit_hash": "9597c284957055a4eaa6a3b5e8c59d7ec963b24281545497ad5672f11e636b96", + "prepared_review_unit_hash": "120f59e0dc5b7a59a8c833204301d50ffd9d400c96a65bba2f453fb613ecc27e", "component_budget": { - "node_count": 472, - "item_count": 472, - "soft_pair_count": 1732, - "serialized_chars": 898654, - "estimated_input_tokens": 224664, - "estimated_eval_tokens": 449584 + "node_count": 473, + "item_count": 473, + "soft_pair_count": 1738, + "serialized_chars": 901437, + "estimated_input_tokens": 225360, + "estimated_eval_tokens": 450976 }, "budget": { "node_count": 48, "item_count": 48, - "soft_pair_count": 30, - "serialized_chars": 36226, - "estimated_input_tokens": 9057, - "estimated_eval_tokens": 18370 + "soft_pair_count": 29, + "serialized_chars": 35985, + "estimated_input_tokens": 8997, + "estimated_eval_tokens": 18250 }, "overflow_policy": "truncate_then_skip", "trimmed": false, "aggressively_trimmed": false, "split": true, "analyst_result": { - "summary": "The set is mostly unrelated issue reports from broad Transformers areas (tokenizers, configs, loading, testing, video, optimizer). None of the soft pairs look like the same concrete bug closely enough to merge, so all candidate duplicate links should be rejected.", - "confidence": 0.88, - "canonical_issue_reason": "issue:44117 is the clearest single bug report to anchor the cluster: it describes a specific `from_pretrained` assumption failure when `TOKENIZER_MAPPING_NAMES` returns `None`, with a direct runtime path and actionable fix surface.", + "summary": "No pair looks like the same underlying bug or change; most are only thematically similar (tokenizers, config, import errors, testing utils) but point to different failure modes. I\u2019d treat this cluster as noisy rather than duplicate-heavy.", + "confidence": 0.24, + "canonical_issue_reason": "Issue 45278 is the broadest umbrella in the set (\u2018many import errors after update\u2019) and is the closest thing to a cluster-wide regression report, though the items here are still mostly unrelated.", "canonical_pr_reason": null, - "best_issue_reason": "issue:44117 is the best representative issue because it is narrowly scoped, reproducible, and centered on a concrete code-path bug rather than a broad symptom or documentation-only complaint.", + "best_issue_reason": "Issue 45278 is the most general and representative issue for this noisy group; if one issue has to stand in for the cluster, it best captures the broad import/regression theme.", "best_pr_reason": null, "soft_edge_verdicts": [ + { + "left": "issue:44466", + "right": "issue:46002", + "accept": false, + "reason": "Different bugs: tied-weight serialization inconsistency vs NaNs during classification head init/checkpointing." + }, { "left": "issue:43937", "right": "issue:44977", "accept": false, - "reason": "Different model families and different failure modes: invalid `GenerationConfig` vs flash-attention generation behavior." + "reason": "Both affect generation, but one is an invalid GenerationConfig error and the other is a flash-attention behavior issue." }, { "left": "issue:43450", "right": "issue:43986", "accept": false, - "reason": "One is a batched video-shape bug; the other is an import/optional-dependency crash in AutoProcessor." + "reason": "Different video-processing failures: batched shape bug vs missing-torchvision crash in AutoProcessor loading." }, { "left": "issue:44410", "right": "issue:45310", "accept": false, - "reason": "Different Qwen variants and distinct problems: missing projections vs a MoE `from_pretrained` error." + "reason": "Different Qwen3.5-family problems: missing projections in layer 0 vs a from_pretrained MoE load error." }, { "left": "issue:45341", "right": "issue:45561", "accept": false, - "reason": "Both touch testing utils, but one is a small utility bug while the other is a pytest-xdist race condition; not the same defect." + "reason": "Shared test-utils area, but one is a generic bug and the other is a specific xdist race on captured_info.txt." }, { "left": "issue:44829", "right": "issue:45216", "accept": false, - "reason": "Training instability under flash-attention-3 is unrelated to the Qwen3.5 save_pretrained checkpoint regression." + "reason": "One is degenerate training with flash_attention_3; the other is incorrect save_pretrained checkpoint serialization." }, { "left": "issue:45341", "right": "issue:46328", "accept": false, - "reason": "Testing helper bug vs feature-extractor padding performance issue; different code paths and symptoms." + "reason": "Testing-utils bug vs a SequenceFeatureExtractor.pad performance regression; not the same underlying issue." }, { "left": "issue:45071", "right": "issue:46512", "accept": false, - "reason": "Both involve config handling, but one is type-checking breakage and the other is stale `dtype` state after `from_config`." + "reason": "Type-checking regression in PretrainedConfig vs stale model.config.dtype after from_config(dtype=...)." }, { "left": "issue:44117", "right": "issue:45081", "accept": false, - "reason": "Tokenizer mapping `None` handling and Mistral regex patching are separate tokenizer-loading failures." + "reason": "Tokenizer loading assumptions vs a Mistral regex patch/backend_tokenizer AttributeError; different code paths." }, { "left": "issue:38617", "right": "issue:41628", "accept": false, - "reason": "Both are import errors, but they involve different missing symbols and different modules." + "reason": "Both are import errors, but they concern different missing symbols and different API breakages." }, { "left": "issue:43122", "right": "issue:46298", "accept": false, - "reason": "Tokenization regression vs noisy logging message; no shared underlying bug." + "reason": "Tokenizer output regression vs a noisy logging message; unrelated changes." }, { "left": "issue:42617", "right": "issue:45341", "accept": false, - "reason": "3d_parallel runtime failure is unrelated to the testing utils issue." + "reason": "Script execution failure vs a testing_utils bug; no shared concrete defect." }, { - "left": "issue:43408", - "right": "issue:43742", + "left": "issue:42222", + "right": "issue:46178", "accept": false, - "reason": "A sam3 tracker/video-model warning is not the same problem as a MobileLLM config KeyError." + "reason": "Vitpose model breakage and tensorflow_text backend-mapping failure are different loading issues." }, { - "left": "issue:42222", - "right": "issue:46178", + "left": "issue:42371", + "right": "issue:44261", "accept": false, - "reason": "Vitpose breakage and tensorflow_text backend mapping errors are different loading/backend issues." + "reason": "TF32 API guidance issue vs MLA q_a_layernorm precision/config bug; not the same change." }, { "left": "issue:45362", "right": "issue:46328", "accept": false, - "reason": "Chat-generation crash for Qwen3.5-35B is unrelated to SequenceFeatureExtractor padding." + "reason": "Chat crash on a specific Qwen model vs a padding performance issue." }, { - "left": "issue:42371", - "right": "issue:44261", + "left": "issue:43408", + "right": "issue:43742", "accept": false, - "reason": "TF32 API deprecation warning and MLA RMS norm epsilon precision issue are unrelated." + "reason": "Model-type mismatch warning vs a key error loading MobileLLM; different failures." }, { "left": "issue:44877", "right": "issue:46231", "accept": false, - "reason": "Strict config loading vs a documentation parameter-name typo; not the same bug." + "reason": "Strict config loading issue vs documentation mismatch for set_training()." }, { "left": "issue:42371", "right": "issue:46617", "accept": false, - "reason": "TF32 behavior warning and trust_remote_code cache FileNotFoundError are different issues." + "reason": "TF32 behavior notice vs cache FileNotFoundError for trust_remote_code; unrelated." }, { "left": "issue:44484", "right": "issue:45310", "accept": false, - "reason": "Shard-size default question is unrelated to the Qwen3.5 MoE loading error." + "reason": "Question about default shard size vs a Qwen3.5 MoE from_pretrained error." }, { "left": "issue:39401", "right": "issue:46209", "accept": false, - "reason": "Tokenizer offset mapping regression and FP8/B200 inference failure are different subsystems." + "reason": "Tokenizer offset-mapping bug vs FP8 inference failure on B200; unrelated." }, { "left": "issue:43937", "right": "issue:44496", "accept": false, - "reason": "GenerationConfig validity and missing `model_type` are both config-related, but they are distinct failures with different fixes." + "reason": "Invalid GenerationConfig vs unrecognized model/model_type configuration." }, { "left": "issue:44933", "right": "issue:45278", "accept": false, - "reason": "A specific missing import from `image_utils` is not the same as broad import errors after upgrade." + "reason": "Both mention import problems, but one is a specific nonexistent import and the other is a broad post-update import regression." }, { "left": "issue:44410", "right": "issue:45478", "accept": false, - "reason": "Different Qwen3.5/Qwen3next bugs and different code paths; not mergeable as one PR-equivalent fix." + "reason": "Different reports: missing projections in qwen3next layer 0 vs Qwen3.5 MoE from_pretrained error." }, { "left": "issue:38617", "right": "issue:44220", "accept": false, - "reason": "ImportError for `layer_type_validation` and `_torch_extract_fbank_features` bug are unrelated." + "reason": "ImportError for a missing symbol vs a feature-extraction bug in _torch_extract_fbank_features()." }, { "left": "issue:44075", "right": "issue:45656", "accept": false, - "reason": "Unused SGD args and deepspeed double-step are different optimizer bugs." + "reason": "SGD arguments ignored vs optimizer.step being called twice with deepspeed; distinct optimizer bugs." }, { "left": "issue:43066", "right": "issue:43122", "accept": false, - "reason": "Both are tokenizer regressions in v5, but one is decoder-type handling and the other is broader tokenization differences; insufficient evidence of one root cause." + "reason": "Tokenizer decoder-type regression and tokenization-difference regression are related only broadly, not the same defect." }, { "left": "issue:42222", "right": "issue:45542", "accept": false, - "reason": "Vitpose breakage is unrelated to the TensorFlow backend being absent when only tensorboard is installed." - }, - { - "left": "issue:43097", - "right": "issue:44493", - "accept": false, - "reason": "Removed config field and unexpected position-id keys are different migration/regression problems." + "reason": "Vitpose model breakage vs undefined TF backend due to tensorboard-only install; different subsystems." }, { "left": "issue:44263", "right": "issue:44485", "accept": false, - "reason": "A `torch.split` return-value issue in one GLM component is unrelated to the GLM-5 RoPE implementation discussion." - }, - { - "left": "issue:42175", - "right": "issue:42222", - "accept": false, - "reason": "Backend-install packaging issue and vitpose model breakage are not the same bug." + "reason": "torch.split return-value handling vs GLM-5 RoPE implementation; different code paths." }, { - "left": "issue:43673", - "right": "issue:45341", + "left": "issue:43097", + "right": "issue:44493", "accept": false, - "reason": "Generation cache missing in v5.0.0 is unrelated to the testing_utils bug." + "reason": "Removal of tie_embeddings_and_encoder_decoder vs unexpected key warnings for position id; not the same underlying change." } ] }, @@ -15238,11 +15143,13 @@ "source": "llm", "cache_hit": false, "model": "gpt-5.4-mini?service_tier=flex", - "cluster_id": "cluster-10000-472", + "cluster_id": "cluster-10000-473", "nodes": [ "issue:27957", "issue:38617", "issue:41628", + "issue:42175", + "issue:42222", "issue:42757", "issue:43012", "issue:43064", @@ -15261,7 +15168,6 @@ "issue:44261", "issue:44263", "issue:44315", - "issue:44351", "issue:44360", "issue:44410", "issue:44483", @@ -15281,7 +15187,6 @@ "issue:45542", "issue:45588", "issue:45753", - "issue:46050", "issue:46123", "issue:46172", "issue:46231", @@ -15290,6 +15195,8 @@ "issue:46664" ], "soft_pairs": [ + "issue:42175|issue:42222", + "issue:43673|issue:45341", "issue:43874|issue:44261", "issue:41628|issue:45071", "issue:43012|issue:43577", @@ -15297,8 +15204,8 @@ "issue:43931|issue:44410", "issue:44263|issue:46172", "issue:44360|issue:46172", - "issue:44483|issue:45464", "issue:44484|issue:45478", + "issue:44483|issue:45464", "issue:27957|issue:43352", "issue:43931|issue:44038", "issue:43441|issue:44829", @@ -15313,243 +15220,234 @@ "issue:43673|issue:46123", "issue:43531|issue:45341", "issue:45440|issue:46172", + "issue:43064|issue:44315", "issue:43097|issue:45588", - "issue:43986|issue:44484", "issue:38617|issue:42757", - "issue:43064|issue:44315", "issue:43986|issue:45372", "issue:46655|issue:46664", - "issue:43097|issue:46050", - "issue:44410|issue:45375", - "issue:44038|issue:44410", - "issue:38617|issue:44351" + "issue:43986|issue:44484", + "issue:44410|issue:45375" ], - "prepared_review_unit_hash": "2e06360a9fedbe5429c101c31eec607394eefb7d29dc413cbe36985c10e577ee", + "prepared_review_unit_hash": "061c7439529b6fe7d6cae3f020658cd04a9a6c829dccb087b52577f8ae33f314", "component_budget": { - "node_count": 472, - "item_count": 472, - "soft_pair_count": 1732, - "serialized_chars": 898654, - "estimated_input_tokens": 224664, - "estimated_eval_tokens": 449584 + "node_count": 473, + "item_count": 473, + "soft_pair_count": 1738, + "serialized_chars": 901437, + "estimated_input_tokens": 225360, + "estimated_eval_tokens": 450976 }, "budget": { "node_count": 48, "item_count": 48, - "soft_pair_count": 33, - "serialized_chars": 37363, - "estimated_input_tokens": 9341, - "estimated_eval_tokens": 18938 + "soft_pair_count": 32, + "serialized_chars": 37086, + "estimated_input_tokens": 9272, + "estimated_eval_tokens": 18800 }, "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 v5 regressions: import errors, model-specific loading failures, attention/processor bugs, and a few API/doc issues. I don\u2019t see any soft pair that clearly points to the same underlying bug/change, so I reject all candidate duplicate edges.", - "confidence": 0.39, - "canonical_issue_reason": "issue:41628 is the best representative of the cluster\u2019s broad, user-facing v5 breakage pattern: a top-level import regression, which is more central than the many model-specific bugs here.", + "summary": "This cluster is a loose collection of Transformers v5-era regressions and import/model-loading bugs across many unrelated subsystems. I do not see any true duplicate pairs here; the similarities are mostly broad subsystem or version overlap, not the same concrete defect.", + "confidence": 0.74, + "canonical_issue_reason": "issue:43097 is the best umbrella representative: it describes a clear v5.0 regression with broad downstream impact and multiple references, making it the most central issue in this mixed regression cluster.", "canonical_pr_reason": null, - "best_issue_reason": "issue:41628 is the most generally actionable representative: it affects the public import surface and fits the cluster\u2019s recurring theme of release regressions better than the narrower model-specific reports.", + "best_issue_reason": "issue:43097 is the strongest global representative for the cluster because it captures the widest version-breaking regression theme, even though most other items are separate bugs.", "best_pr_reason": null, "soft_edge_verdicts": [ + { + "left": "issue:42175", + "right": "issue:42222", + "accept": false, + "reason": "Packaging/backend dependency omission is a different problem from VitPose model breakage; same release context, not the same bug." + }, + { + "left": "issue:43673", + "right": "issue:45341", + "accept": false, + "reason": "Generation cache behavior during chunked prefill is unrelated to a testing_utils bug." + }, { "left": "issue:43874", "right": "issue:44261", "accept": false, - "reason": "Different model/component bugs: a missing image processor method vs a layernorm epsilon precision issue." + "reason": "Missing image-patch method in a processor and an MLA RMS norm epsilon issue are different model components and failure modes." }, { "left": "issue:41628", "right": "issue:45071", "accept": false, - "reason": "Both are v5 regressions, but one is an import failure and the other is a type-checking break; not the same code-path bug." + "reason": "AutoImageProcessor import failure and PretrainedConfig type-checking regression do not share the same code path." }, { "left": "issue:43012", "right": "issue:43577", "accept": false, - "reason": "Both involve dtype/precision, but one is a PyTorch warning during compile and the other is incorrect model dtype loading for BLIP2." + "reason": "A PyTorch warning on bfloat16 compilation is not the same defect as model dtypes staying float32 after load." }, { "left": "issue:45071", "right": "issue:45542", "accept": false, - "reason": "Unrelated regressions: PreTrainedConfig type checking vs TensorBoard/TensorFlow backend detection." + "reason": "Config type-checking regression and TensorFlow backend installation/selection error are unrelated." }, { "left": "issue:43931", "right": "issue:44410", "accept": false, - "reason": "Different Qwen model-loading failures: weight-shape mismatch vs missing projections." + "reason": "Both are Qwen-related, but one is a weight-shape mismatch on a specific VL model and the other is missing projections in a different architecture." }, { "left": "issue:44263", "right": "issue:46172", "accept": false, - "reason": "Both mention tensor ops, but they are different model implementations and different failures." + "reason": "These are different tensor-operation bugs in different model code paths." }, { "left": "issue:44360", "right": "issue:46172", "accept": false, - "reason": "DSA ReLU absence is not the same bug as the DeepSeek-V4 tensor transpose issue." + "reason": "A missing ReLU in the DSA indexer is not the same as the DeepSeek-v4 transpose/cast issue." }, { - "left": "issue:44483", - "right": "issue:45464", + "left": "issue:44484", + "right": "issue:45478", "accept": false, - "reason": "Both concern chat/completions, but one is a general request rejection and the other is a Qwen3.5 streaming-specific failure." + "reason": "save_pretrained shard-size behavior is unrelated to Qwen3.5 MoE from_pretrained loading errors." }, { - "left": "issue:44484", - "right": "issue:45478", + "left": "issue:44483", + "right": "issue:45464", "accept": false, - "reason": "Different problems: save_pretrained shard sizing question vs Qwen3.5-MoE from_pretrained error." + "reason": "Both mention chat/completions and Qwen3.5, but one is a request/endpoint acceptance failure and the other is a streaming inference failure; not enough to treat as the same fix." }, { "left": "issue:27957", "right": "issue:43352", "accept": false, - "reason": "Same feature area, but different model-specific Flash Attention 2 support gaps." + "reason": "Flash Attention 2 support for XLM-Roberta and NemotronH unsupported FA2 are model-specific feature gaps, not the same bug." }, { "left": "issue:43931", "right": "issue:44038", "accept": false, - "reason": "Both are Qwen3/VL-related, but the reported failures are different and model-specific." + "reason": "Different Qwen3-VL/MoE loading problems with no clear shared concrete defect." }, { "left": "issue:43441", "right": "issue:44829", "accept": false, - "reason": "Both mention flash attention, but one is unsupported-model behavior and the other is degenerate training with Flash Attention 3." + "reason": "FlashAttention failures in different models and with different symptoms (support failure vs degenerate training) are not the same bug." }, { "left": "issue:43097", "right": "issue:43452", "accept": false, - "reason": "One is a removed config option in v5; the other is a gguf/from_pretrained loading breakage." + "reason": "Removed tie_embeddings behavior and gguf_file loading breakage are unrelated." }, { "left": "issue:45003", "right": "issue:45341", "accept": false, - "reason": "Unrelated internals bug vs a testing utility issue." + "reason": "Unsafe sys.modules access and a testing_utils issue are different code paths and fixes." }, { "left": "issue:43986", "right": "issue:45341", "accept": false, - "reason": "Processor-loading crash due to missing torchvision is unrelated to a testing_utils bug." + "reason": "AutoProcessor/torchvision crash is unrelated to testing_utils." }, { "left": "issue:45753", "right": "issue:46369", "accept": false, - "reason": "Different symptoms and likely different causes: infinite loop on a specific image vs memory overflow." + "reason": "An infinite loop on a specific image and a memory overflow in another model are distinct failures." }, { "left": "issue:44933", "right": "issue:46231", "accept": false, - "reason": "Import surface issue vs a documentation/parameter-name mismatch; not the same bug." + "reason": "Missing image_utils import and a docs parameter-name mismatch are not the same issue." }, { "left": "issue:43874", "right": "issue:44263", "accept": false, - "reason": "Different GLM-related failures in different code paths; not a single underlying defect." + "reason": "Different GLM subsystems and symptoms; not the same defect." }, { "left": "issue:44261", "right": "issue:44485", "accept": false, - "reason": "Different numerical/configuration bugs in different model families." + "reason": "RMS norm epsilon precision issue and GLM-5 RoPE implementation are separate model math problems." }, { "left": "issue:44117", "right": "issue:44987", "accept": false, - "reason": "Tokenizer mapping None handling is unrelated to a physical-intelligence model load failure." + "reason": "Tokenizer mapping returning None is too broad to equate with a specific model-load failure without shared implementation evidence." }, { "left": "issue:43673", "right": "issue:46123", "accept": false, - "reason": "Cache handling in generation is unrelated to MaskGenerationPipeline batch-finalization behavior." + "reason": "Generation cache missing during chunked prefill is unrelated to MaskGenerationPipeline partial-batch handling." }, { "left": "issue:43531", "right": "issue:45341", "accept": false, - "reason": "Qwen3-MoE sliding_window behavior is unrelated to a testing utility bug." + "reason": "Qwen3-MoE sliding_window behavior and testing_utils are unrelated." }, { "left": "issue:45440", "right": "issue:46172", "accept": false, - "reason": "DeepseekV3MoE divergence is a broader model implementation mismatch, not the same as the DeepSeek-V4 comb transpose bug." + "reason": "DeepseekV3MoE divergence from remote implementation and a DeepSeek-v4 tensor cast/transpose bug are not the same concrete problem." }, { - "left": "issue:43097", - "right": "issue:45588", + "left": "issue:43064", + "right": "issue:44315", "accept": false, - "reason": "Removed embedding-tie API and flash_attention.py sink-less crash are different regressions." + "reason": "FSDP2/PEFT optimizer-state corruption and Liger Kernel application in model_init are different training bugs." }, { - "left": "issue:43986", - "right": "issue:44484", + "left": "issue:43097", + "right": "issue:45588", "accept": false, - "reason": "Processor-loading dependency crash is unrelated to max_shard_size behavior." + "reason": "A removed config utility and a flash_attention integration crash on s_aux=None are unrelated." }, { "left": "issue:38617", "right": "issue:42757", "accept": false, - "reason": "Both are import errors, but they involve different missing symbols from different modules and are not the same bug." - }, - { - "left": "issue:43064", - "right": "issue:44315", - "accept": false, - "reason": "Distributed optimizer-state corruption and Liger Kernel application are different training issues." + "reason": "Two import errors, but for different symbols and different modules; not the same underlying bug." }, { "left": "issue:43986", "right": "issue:45372", "accept": false, - "reason": "Both are loading-time failures, but one is missing torchvision and the other is a missing mistral_common symbol." + "reason": "Both are dependency/import-related, but one is torchvision absence during video processor loading and the other is a mistral_common ReasoningEffort import breakage." }, { "left": "issue:46655", "right": "issue:46664", "accept": false, - "reason": "Both are SmolVLM video processor issues, but one is padding calculation and the other is num_frames semantics." + "reason": "Both concern SmolVLM video processing, but padding math and num_frames semantics are separate bugs." }, { - "left": "issue:43097", - "right": "issue:46050", + "left": "issue:43986", + "right": "issue:44484", "accept": false, - "reason": "Removed config API in v5 is unrelated to quantized-model save/load breakage." + "reason": "Video processor import crash and save_pretrained shard sizing have no overlap." }, { "left": "issue:44410", "right": "issue:45375", "accept": false, - "reason": "Different Qwen/MoE config/model issues; one is missing projections, the other is a silently dropped config field." - }, - { - "left": "issue:44038", - "right": "issue:44410", - "accept": false, - "reason": "Both are Qwen3-related, but they describe different model regressions in different paths." - }, - { - "left": "issue:38617", - "right": "issue:44351", - "accept": false, - "reason": "Both are import errors, but they refer to different missing exports and different failure modes." + "reason": "Both are Qwen3-family config issues, but missing projections and a silently dropped vision config field are different concrete defects." } ] }, @@ -15571,12 +15469,10 @@ "source": "llm", "cache_hit": false, "model": "gpt-5.4-mini?service_tier=flex", - "cluster_id": "cluster-10000-472", + "cluster_id": "cluster-10000-473", "nodes": [ "issue:36683", - "issue:42175", - "issue:42490", - "issue:42757", + "issue:38617", "issue:42890", "issue:42907", "issue:43012", @@ -15586,10 +15482,8 @@ "issue:43208", "issue:43352", "issue:43502", - "issue:43526", "issue:43550", "issue:43582", - "issue:43824", "issue:43825", "issue:43873", "issue:43906", @@ -15597,13 +15491,16 @@ "issue:43986", "issue:43992", "issue:44016", + "issue:44038", "issue:44062", "issue:44077", "issue:44117", "issue:44162", "issue:44222", + "issue:44351", "issue:44368", "issue:44380", + "issue:44410", "issue:44484", "issue:44655", "issue:44704", @@ -15615,6 +15512,7 @@ "issue:45584", "issue:45712", "issue:45753", + "issue:46050", "issue:46132", "issue:46172", "issue:46178", @@ -15623,11 +15521,14 @@ "issue:46421" ], "soft_pairs": [ + "issue:43097|issue:46050", + "issue:44038|issue:44410", "issue:42890|issue:46298", + "issue:38617|issue:44351", "issue:44062|issue:44117", "issue:43097|issue:43873", - "issue:43502|issue:43992", "issue:44936|issue:46365", + "issue:43502|issue:43992", "issue:44704|issue:45103", "issue:44380|issue:46421", "issue:43986|issue:44117", @@ -15637,259 +15538,231 @@ "issue:43825|issue:44655", "issue:42907|issue:44655", "issue:43012|issue:43550", - "issue:45712|issue:46298", "issue:42907|issue:45753", - "issue:43012|issue:44368", + "issue:45712|issue:46298", "issue:43992|issue:44484", + "issue:43012|issue:44368", "issue:43066|issue:43906", - "issue:43097|issue:46298", "issue:43352|issue:44380", + "issue:43097|issue:46298", "issue:43352|issue:43976", "issue:44162|issue:46172", - "issue:45584|issue:46365", "issue:44380|issue:46132", + "issue:45584|issue:46365", "issue:44077|issue:46365", "issue:44222|issue:45478", "issue:44380|issue:45479", "issue:36683|issue:43582", - "issue:43064|issue:44380", - "issue:42175|issue:46365", - "issue:42757|issue:43824", - "issue:42490|issue:43012", - "issue:43208|issue:43526" + "issue:43064|issue:44380" ], - "prepared_review_unit_hash": "70ba11e5f506d7821e1dc0eb764fc17fef98377fe28f0e126283f9dbf2439577", + "prepared_review_unit_hash": "7b69aa90af0f8e5478dcc82ca58c3c7f970d91782a67bba0ee4a5bbc41f20a54", "component_budget": { - "node_count": 472, - "item_count": 472, - "soft_pair_count": 1732, - "serialized_chars": 898654, - "estimated_input_tokens": 224664, - "estimated_eval_tokens": 449584 + "node_count": 473, + "item_count": 473, + "soft_pair_count": 1738, + "serialized_chars": 901437, + "estimated_input_tokens": 225360, + "estimated_eval_tokens": 450976 }, "budget": { "node_count": 48, "item_count": 48, - "soft_pair_count": 34, - "serialized_chars": 37787, - "estimated_input_tokens": 9447, - "estimated_eval_tokens": 19150 + "soft_pair_count": 33, + "serialized_chars": 37394, + "estimated_input_tokens": 9349, + "estimated_eval_tokens": 18954 }, "overflow_policy": "truncate_then_skip", "trimmed": false, "aggressively_trimmed": false, "split": true, "analyst_result": { - "summary": "This is a heterogeneous set of unrelated issues, mostly distinct import/config/training regressions. None of the proposed soft pairs look like true duplicates, so I would reject all of them. If one issue had to anchor triage, issue 43824 is the strongest standalone report.", - "confidence": 0.95, - "canonical_issue_reason": "issue:43824 is the clearest high-signal standalone issue in the set: concrete ImportError, reproducible, and already referenced by other reports.", + "summary": "This cluster is mostly heterogeneous: several unrelated Transformers regressions and API breakages share broad themes like v5, quantization, or loading, but the soft pairs do not look like true duplicates. I reject all soft edges.", + "confidence": 0.81, + "canonical_issue_reason": "issue:43097 is the broadest, most central v5 regression in the set and has downstream references, so it is the best anchor among these issue reports.", "canonical_pr_reason": null, - "best_issue_reason": "issue:43824 is the best issue to keep as the representative triage item because it is specific, user-facing, and has multiple inbound references.", + "best_issue_reason": "issue:43097 is the clearest representative user-facing regression here; it is concrete, widely relevant, and better suited as a cluster anchor than the more niche reports.", "best_pr_reason": null, "soft_edge_verdicts": [ + { + "left": "issue:43097", + "right": "issue:46050", + "accept": false, + "reason": "Different bug classes: removed config field vs quantized model save/load corruption. Similar save/load wording is not enough." + }, + { + "left": "issue:44038", + "right": "issue:44410", + "accept": false, + "reason": "Both involve Qwen models, but one is a general Transformers 5.0 compatibility bug and the other is a specific missing-projection architecture issue." + }, { "left": "issue:42890", "right": "issue:46298", "accept": false, - "reason": "One is a SAM HQ test seed flake; the other is a noisy logging complaint. Different bugs and fixes." + "reason": "Flaky test seeding and noisy logging are unrelated." + }, + { + "left": "issue:38617", + "right": "issue:44351", + "accept": false, + "reason": "Different import failures from different symbols and code paths." }, { "left": "issue:44062", "right": "issue:44117", "accept": false, - "reason": "Different tokenizer failures: AddedToken argument collision vs TOKENIZER_MAPPING_NAMES returning None." + "reason": "Tokenizer construction with duplicate kwargs is distinct from TOKENIZER_MAPPING_NAMES returning None." }, { "left": "issue:43097", "right": "issue:43873", "accept": false, - "reason": "Tie-embeddings removal is an API change; offloading with quantization is a separate runtime behavior bug." + "reason": "Config-API removal regression and quantization/offloading behavior are separate issues." }, { - "left": "issue:43502", - "right": "issue:43992", + "left": "issue:44936", + "right": "issue:46365", "accept": false, - "reason": "local_files_only network leakage and missing UMT5Encoder weights are unrelated pretrained-loading issues." + "reason": "Trainer state corruption after train is unrelated to failing to import Trainer because of datasets." }, { - "left": "issue:44936", - "right": "issue:46365", + "left": "issue:43502", + "right": "issue:43992", "accept": false, - "reason": "Trainer.evaluate after train and import failure until uninstalling datasets are different code paths and symptoms." + "reason": "Local-files-only network leakage and missing encoder weights are different loading bugs." }, { "left": "issue:44704", "right": "issue:45103", "accept": false, - "reason": "AutoProcessor cached_file kwargs forwarding and auto_docstring annotation handling are unrelated subsystems." + "reason": "AutoProcessor kwarg passthrough and docstring annotation crashes affect different subsystems." }, { "left": "issue:44380", "right": "issue:46421", "accept": false, - "reason": "GPT2 attention scaling under SDPA/FlashAttention is unrelated to lazy cache initialization and recompilation." + "reason": "Ignored attention scaling in GPT2 is unrelated to cache lazy initialization and recompilation." }, { "left": "issue:43986", "right": "issue:44117", "accept": false, - "reason": "Missing torchvision for AutoProcessor video loading is not the same as a tokenizer mapping None bug." + "reason": "AutoProcessor/torchvision crash and tokenizer mapping None are unrelated." }, { "left": "issue:44938", "right": "issue:46178", "accept": false, - "reason": "Python 3.14 load failure and tensorflow_text backend mapping validation are different import/setup problems." + "reason": "Python 3.14 load failure and missing backend registration are different compatibility problems." }, { "left": "issue:43986", "right": "issue:44016", "accept": false, - "reason": "A missing dependency crash is unrelated to a notebook syntax error." + "reason": "A video-processing import crash is not the same as a notebook syntax error." }, { "left": "issue:43208", "right": "issue:43550", "accept": false, - "reason": "xLSTM training blockers and Bamba torch.compile SDPA failure are different model-specific bugs." + "reason": "xLSTM training blockers and Bamba torch.compile/SDPA failure are different model-specific bugs." }, { "left": "issue:43825", "right": "issue:44655", "accept": false, - "reason": "A misleading pipeline translation error and inability to save Pipeline objects are not the same bug." + "reason": "Incorrect pipeline error text is unrelated to saving Pipeline objects." }, { "left": "issue:42907", "right": "issue:44655", "accept": false, - "reason": "Saving dequantized Ministrals/Devstrals and saving Pipeline objects involve different serialization paths." + "reason": "Dequantized model save failure and pipeline save failure are separate save_pretrained problems." }, { "left": "issue:43012", "right": "issue:43550", "accept": false, - "reason": "A bfloat16 compile warning is not the same as a Bamba SDPA compilation failure." - }, - { - "left": "issue:45712", - "right": "issue:46298", - "accept": false, - "reason": "Dummy class leakage in dummy_pt_objects.py is a code hygiene issue, not the same as noisy logging." + "reason": "A PyTorch warning on bfloat16 compilation is not the same as a model-specific SDPA compile failure." }, { "left": "issue:42907", "right": "issue:45753", "accept": false, - "reason": "Saving dequantized models and Qwen3_5 infinite looping are unrelated problems." + "reason": "Saving dequantized Ministrals/Devstrals and Qwen3_5 image looping are unrelated." }, { - "left": "issue:43012", - "right": "issue:44368", + "left": "issue:45712", + "right": "issue:46298", "accept": false, - "reason": "PyTorch bfloat16 precision warning is unrelated to tie_word_embeddings warnings during LoRA fine-tuning." + "reason": "Leftover dummy classes and a noisy log message are not the same bug." }, { "left": "issue:43992", "right": "issue:44484", "accept": false, - "reason": "Missing embed_tokens.weight on load is not about the default 50GB shard size." + "reason": "Missing embed token weights on load is unrelated to the default max shard size question." }, { - "left": "issue:43066", - "right": "issue:43906", + "left": "issue:43012", + "right": "issue:44368", "accept": false, - "reason": "Wrong tokenizer decoder type in v5 and an isolated reproduction of another issue are too underspecified to treat as the same bug." + "reason": "These are different warnings about different config/compile behaviors." }, { - "left": "issue:43097", - "right": "issue:46298", + "left": "issue:43066", + "right": "issue:43906", "accept": false, - "reason": "A removed API and a logging message complaint do not share the same underlying defect." + "reason": "Wrong tokenizer decoder type and an isolated reproduction of another issue are not the same code-path problem." }, { "left": "issue:43352", "right": "issue:44380", "accept": false, - "reason": "FlashAttention support gating for Nemotron and GPT2 attention scaling under SDPA are different model/backend issues." - }, - { - "left": "issue:43352", - "right": "issue:43976", - "accept": false, - "reason": "A model backend limitation is unrelated to a Python version compatibility regression." - }, - { - "left": "issue:44162", - "right": "issue:46172", - "accept": false, - "reason": "ESM2 breakage and a DeepSeek tensor transpose complaint are different model failures." - }, - { - "left": "issue:45584", - "right": "issue:46365", - "accept": false, - "reason": "Whisper empty-transcription failure is unrelated to an import conflict involving datasets." - }, - { - "left": "issue:44380", - "right": "issue:46132", - "accept": false, - "reason": "Attention scaling being ignored is not the same as AttentionInterface.register changing function behavior." - }, - { - "left": "issue:44077", - "right": "issue:46365", - "accept": false, - "reason": "PatchTSMixer post_init policy and Trainer import failure are unrelated." - }, - { - "left": "issue:44222", - "right": "issue:45478", - "accept": false, - "reason": "FP8 save_pretrained behavior and Qwen3.5 MoE from_pretrained errors are different serialization/loading issues." + "reason": "Unsupported Flash Attention 2 for Nemotron is distinct from GPT2 attention scaling being ignored under SDPA/FlashAttention." }, { - "left": "issue:44380", - "right": "issue:45479", + "left": "issue:44162", + "right": "issue:46172", "accept": false, - "reason": "Attention backend scaling bug and zero-loss sequence classification are unrelated." + "reason": "ESM2 being broken and a DeepSeek tensor-transpose bug are unrelated model implementations." }, { - "left": "issue:36683", - "right": "issue:43582", + "left": "issue:44380", + "right": "issue:46132", "accept": false, - "reason": "Gemma3Config vocab_size missing and Apple Silicon allocator warmup TypeError are unrelated." + "reason": "A backend-specific scaling bug and an AttentionInterface.register behavior change are different changes." }, { - "left": "issue:43064", - "right": "issue:44380", + "left": "issue:45584", + "right": "issue:46365", "accept": false, - "reason": "FSDP2/PEFT optimizer state corruption is unrelated to GPT2 attention scaling under SDPA/FlashAttention." + "reason": "Whisper empty-transcription failure is unrelated to importing Trainer." }, { - "left": "issue:42175", + "left": "issue:44077", "right": "issue:46365", "accept": false, - "reason": "TensorFlow backend packaging and Trainer import failure are different dependency issues." + "reason": "patchtsmixer post_init policy and Trainer import failure are unrelated." }, { - "left": "issue:42757", - "right": "issue:43824", + "left": "issue:44222", + "right": "issue:45478", "accept": false, - "reason": "Missing is_offline_mode from huggingface_hub and missing Qwen2_5_VLForConditionalGeneration are distinct import errors." + "reason": "Both mention MoE save/load, but one is FP8 save_pretrained and the other is a from_pretrained error; too different to merge as one bug." }, { - "left": "issue:42490", - "right": "issue:43012", + "left": "issue:36683", + "right": "issue:43582", "accept": false, - "reason": "Path-trailing-slash import warning and bfloat16 compile warning are unrelated." + "reason": "Missing Gemma3Config vocab_size and an Apple Silicon allocator TypeError are unrelated." }, { - "left": "issue:43208", - "right": "issue:43526", + "left": "issue:43064", + "right": "issue:44380", "accept": false, - "reason": "xLSTM training blockers and BeitImageProcessorFast reduce_labels bug are different model/preprocessor issues." + "reason": "A distributed Trainer/FSDP2/PEFT optimizer-state bug is not the same as GPT2 attention scaling being ignored." } ] }, @@ -15911,39 +15784,40 @@ "source": "llm", "cache_hit": false, "model": "gpt-5.4-mini?service_tier=flex", - "cluster_id": "cluster-10000-472", + "cluster_id": "cluster-10000-473", "nodes": [ "issue:36010", "issue:36683", "issue:38617", + "issue:42175", + "issue:42490", "issue:42757", "issue:42898", "issue:42994", "issue:43012", - "issue:43064", "issue:43066", + "issue:43208", "issue:43299", "issue:43352", "issue:43475", "issue:43502", "issue:43525", + "issue:43526", "issue:43611", "issue:43618", "issue:43756", "issue:43761", + "issue:43824", "issue:43867", "issue:43901", "issue:43937", "issue:43976", - "issue:43992", "issue:44038", - "issue:44075", "issue:44077", "issue:44230", "issue:44380", "issue:44483", "issue:44485", - "issue:44492", "issue:44514", "issue:44625", "issue:44655", @@ -15952,7 +15826,6 @@ "issue:45216", "issue:45276", "issue:45290", - "issue:45310", "issue:45600", "issue:45698", "issue:46077", @@ -15962,22 +15835,26 @@ "issue:46365" ], "soft_pairs": [ + "issue:42175|issue:46365", + "issue:43208|issue:43526", + "issue:42490|issue:43012", + "issue:42757|issue:43824", "issue:43012|issue:44683", - "issue:45600|issue:46298", "issue:44483|issue:44514", + "issue:45600|issue:46298", "issue:44038|issue:44230", - "issue:45698|issue:46172", "issue:43618|issue:43761", "issue:43618|issue:44380", + "issue:45698|issue:46172", "issue:38617|issue:43502", - "issue:43299|issue:43867", "issue:43756|issue:44485", - "issue:36683|issue:43475", + "issue:43299|issue:43867", "issue:42994|issue:44655", - "issue:44625|issue:46365", + "issue:36683|issue:43475", "issue:44483|issue:45290", - "issue:45276|issue:46077", "issue:42898|issue:43066", + "issue:44625|issue:46365", + "issue:45276|issue:46077", "issue:36010|issue:42757", "issue:43352|issue:43937", "issue:43901|issue:44655", @@ -15989,225 +15866,228 @@ "issue:46077|issue:46082", "issue:36683|issue:43525", "issue:43618|issue:45216", - "issue:36683|issue:44625", - "issue:43064|issue:44492", - "issue:43992|issue:45310", - "issue:44075|issue:46365" + "issue:36683|issue:44625" ], - "prepared_review_unit_hash": "c141729f8321fdfc79c694c2d1a4c916a5e9ea086a3c6047ec9d227ebdcd092d", + "prepared_review_unit_hash": "8efdf36c0109b910d255e8cdb3a9c0a64f4d9882c1ebe6f54a61c0b4b8b818ed", "component_budget": { - "node_count": 472, - "item_count": 472, - "soft_pair_count": 1732, - "serialized_chars": 898654, - "estimated_input_tokens": 224664, - "estimated_eval_tokens": 449584 + "node_count": 473, + "item_count": 473, + "soft_pair_count": 1738, + "serialized_chars": 901437, + "estimated_input_tokens": 225360, + "estimated_eval_tokens": 450976 }, "budget": { "node_count": 47, "item_count": 47, - "soft_pair_count": 31, - "serialized_chars": 35878, - "estimated_input_tokens": 8970, - "estimated_eval_tokens": 18196 + "soft_pair_count": 32, + "serialized_chars": 36270, + "estimated_input_tokens": 9068, + "estimated_eval_tokens": 18392 }, "overflow_policy": "truncate_then_skip", "trimmed": false, "aggressively_trimmed": false, "split": true, "analyst_result": { - "summary": "This cluster is mostly a grab-bag of distinct Transformers regressions and import/load failures, with several superficially similar v5/Qwen/CLIP/Gemma issues that are not the same underlying bug. No PRs are present.", - "confidence": 0.78, - "canonical_issue_reason": "issue:43611 is the broadest, most representative regression here: a general Transformers 5.0 model-loading break tied to base_model_prefix, which better fits the cluster\u2019s recurring theme of v5 load/save regressions than the narrower model-specific reports.", + "summary": "This cluster is mostly a false-positive similarity group: the items span unrelated import errors, config regressions, save/load bugs, backend warnings, and model-specific failures. None of the soft pairs looks like the same underlying bug or a safe PR merge candidate.", + "confidence": 0.86, + "canonical_issue_reason": null, "canonical_pr_reason": null, - "best_issue_reason": "issue:43611 is the best single issue to anchor this cluster because it describes a core, high-impact v5 loading regression affecting multiple models, whereas most other items are narrower or unrelated.", + "best_issue_reason": "issue:45698 is the most discussion-heavy and reference-heavy item here, and it best represents the recurring save_pretrained/from_pretrained theme, even though the cluster is not a true duplicate set.", "best_pr_reason": null, "soft_edge_verdicts": [ { - "left": "issue:43012", - "right": "issue:44683", + "left": "issue:42175", + "right": "issue:46365", "accept": false, - "reason": "Both are PyTorch/precision-related, but one is a bfloat16 compile warning and the other is a flex_attention failure on torch>=2.9; different symptoms and code paths." + "reason": "Different import/dependency failures: TensorFlow missing from a torch install vs Trainer import blocked by datasets." }, { - "left": "issue:45600", - "right": "issue:46298", + "left": "issue:43208", + "right": "issue:43526", + "accept": false, + "reason": "Unrelated model areas: xLSTM training bugs vs BeitImageProcessorFast reduce_labels behavior." + }, + { + "left": "issue:42490", + "right": "issue:43012", + "accept": false, + "reason": "Local-path import warning and bf16 compile warning are different code paths and symptoms." + }, + { + "left": "issue:42757", + "right": "issue:43824", + "accept": false, + "reason": "Both are import errors, but for different missing symbols and different fixes." + }, + { + "left": "issue:43012", + "right": "issue:44683", "accept": false, - "reason": "One is a CI break from stale auto_mappings references, the other is just a noisy logging message; not the same bug." + "reason": "Different torch-related problems: a precision warning vs a compiled flex_attention failure." }, { "left": "issue:44483", "right": "issue:44514", "accept": false, - "reason": "A /v1/chat/completions request rejection and a batched chat template crash are different frontend/chat-template problems." + "reason": "Different failures in chat handling: `/v1/chat/completions` request rejection vs batched chat-template crash." }, { - "left": "issue:44038", - "right": "issue:44230", + "left": "issue:45600", + "right": "issue:46298", "accept": false, - "reason": "Both mention Qwen3-VL-Moe, but one is a loading regression and the other is fp8 inference support; not mergeable as one fix." + "reason": "CI break from removed Sam3LiteText configs is unrelated to a noisy logging message." }, { - "left": "issue:45698", - "right": "issue:46172", + "left": "issue:44038", + "right": "issue:44230", "accept": false, - "reason": "Wrong custom module after save/load is unrelated to the DeepSeek-V4 dtype/transpose issue." + "reason": "Different model/features: Qwen3-VL-Moe loading bug vs fp8 support request for Qwen3-VL/Qwen3.5." }, { "left": "issue:43618", "right": "issue:43761", "accept": false, - "reason": "Both are CLIP v5 regressions, but one concerns missing attentions and the other missing hidden_states; distinct outputs and fixes." + "reason": "Same model family, but different missing outputs: attentions not assigned vs hidden_states not returned." }, { "left": "issue:43618", "right": "issue:44380", "accept": false, - "reason": "CLIP attentions missing and GPT2 attention scaling ignored are unrelated model/backend bugs." + "reason": "Different subsystems: CLIP output regression vs GPT2 attention scaling ignored under SDPA/FlashAttention." }, { - "left": "issue:38617", - "right": "issue:43502", + "left": "issue:45698", + "right": "issue:46172", "accept": false, - "reason": "Different import/runtime failures: a missing configuration_utils symbol versus network requests despite local_files_only=True." + "reason": "Save/load custom-module bug vs a tensor dtype/transpose issue in deepseek-v4." }, { - "left": "issue:43299", - "right": "issue:43867", + "left": "issue:38617", + "right": "issue:43502", "accept": false, - "reason": "Qwen3-VL-Moe load failure and state_dict-sorted load error are different loading bugs with different triggers." + "reason": "ImportError for layer_type_validation is unrelated to local_files_only still making API calls." }, { "left": "issue:43756", "right": "issue:44485", "accept": false, - "reason": "Smollm3 RoPE layer loss and GLM-5 RoPE discussion do not point to the same concrete defect." + "reason": "Same rough topic (RoPE), but different models and no shared concrete failure mode." }, { - "left": "issue:36683", - "right": "issue:43475", + "left": "issue:43299", + "right": "issue:43867", "accept": false, - "reason": "Gemma3Config missing vocab_size and SAM3VisionEncoderOutput missing fpn_position_embeddings are unrelated attribute errors." + "reason": "Both loading-related, but one is Qwen3VL-Moe breakage and the other is state_dict sorting causing a load error." }, { "left": "issue:42994", "right": "issue:44655", "accept": false, - "reason": "Quantized model saving failure and pipeline save_pretrained failure both involve saving, but they affect different object types and code paths." + "reason": "Quantized model saving failure and pipeline save_pretrained failure are different serialization paths." }, { - "left": "issue:44625", - "right": "issue:46365", + "left": "issue:36683", + "right": "issue:43475", "accept": false, - "reason": "Qwen3.5 num_labels propagation and importing Trainer failing due to datasets are unrelated issues." + "reason": "Missing config attribute in Gemma3 vs missing output field in SAM3 Video; different models and bugs." }, { "left": "issue:44483", "right": "issue:45290", "accept": false, - "reason": "Chat/completions request rejection and apply_chat_template crash with tool-call assistant messages are different chat-handling bugs." + "reason": "Request validation at /v1/chat/completions is not the same as tokenizer chat-template crashes on tool-call messages." }, { - "left": "issue:45276", - "right": "issue:46077", + "left": "issue:42898", + "right": "issue:43066", "accept": false, - "reason": "resize_token_embeddings not updating Gemma4 embeddings and Gemma4 bidirectional attention mask construction are separate model bugs." + "reason": "Both are v5 tokenizer changes, but one is clean_up_tokenization_spaces behavior and the other is decoder type selection." }, { - "left": "issue:42898", - "right": "issue:43066", + "left": "issue:44625", + "right": "issue:46365", + "accept": false, + "reason": "Qwen3.5 num_labels propagation is unrelated to Trainer import failing because of datasets." + }, + { + "left": "issue:45276", + "right": "issue:46077", "accept": false, - "reason": "A v5 tokenization-space behavior change and a wrong tokenizer decoder type are different tokenizer regressions." + "reason": "Gemma4 embedding resize bug and Gemma4 bidirectional-attention mask bug affect different code paths." }, { "left": "issue:36010", "right": "issue:42757", "accept": false, - "reason": "Both are import errors, but they involve different missing symbols from different packages and likely different dependency breaks." + "reason": "Both are import errors, but they concern different removed/relocated symbols." }, { "left": "issue:43352", "right": "issue:43937", "accept": false, - "reason": "Flash Attention 2 unsupported for NemotronHForCausalLM and invalid GenerationConfig for GLM-5 are unrelated generation/model issues." + "reason": "Nemotron FlashAttention support error and GLM-5 GenerationConfig validation are unrelated model-specific issues." }, { "left": "issue:43901", "right": "issue:44655", "accept": false, - "reason": "Docs mismatch for return_all_scores is unrelated to pipeline save_pretrained behavior." + "reason": "Docs mismatch for return_all_scores is not the same as pipeline serialization failure." }, { "left": "issue:43611", "right": "issue:44655", "accept": false, - "reason": "General model loading break via base_model_prefix is unrelated to saving Pipeline objects." + "reason": "Loading models with base_model_prefix and saving Pipeline objects are different stages and bugs." }, { "left": "issue:36683", "right": "issue:44077", "accept": false, - "reason": "Gemma3Config missing vocab_size and optimizer SGD args not used are unrelated." + "reason": "Gemma3 missing vocab_size and patchtsmixer post_init validation are unrelated config issues." }, { "left": "issue:43976", "right": "issue:44938", "accept": false, - "reason": "These are both environment/version compatibility issues, but they concern different Python versions and failure modes." + "reason": "Different Python compatibility failures: 3.9+ regression vs Python 3.14 load failure." }, { "left": "issue:43867", "right": "issue:45216", "accept": false, - "reason": "Sorted state_dict load failure and incorrect Qwen3.5 saved checkpoint are different save/load regressions." + "reason": "Both involve load/save behavior, but one is state_dict ordering and the other is a Qwen3.5 checkpoint regression." }, { "left": "issue:44625", "right": "issue:46077", "accept": false, - "reason": "Config propagation for num_labels and Gemma4 attention mask construction are unrelated." + "reason": "Qwen3.5 label propagation and Gemma4 bidirectional attention are unrelated features." }, { "left": "issue:46077", "right": "issue:46082", "accept": false, - "reason": "Gemma4 bidirectional attention and LlamaConfig head_dim validation are separate config/model checks." + "reason": "Gemma4 attention-mask behavior and LlamaConfig head_dim validation are different model/config bugs." }, { "left": "issue:36683", "right": "issue:43525", "accept": false, - "reason": "Missing vocab_size on Gemma3Config and missing pad_token_id on Llama4Config are different attribute regressions." + "reason": "Different missing config attributes in different models: vocab_size vs pad_token_id." }, { "left": "issue:43618", "right": "issue:45216", "accept": false, - "reason": "CLIP attentions missing and Qwen3.5 checkpoint corruption are unrelated model-specific regressions." + "reason": "CLIP output regression and Qwen3.5 save_pretrained checkpoint corruption are unrelated." }, { "left": "issue:36683", "right": "issue:44625", "accept": false, - "reason": "Different config propagation problems for different models; not the same bug." - }, - { - "left": "issue:43064", - "right": "issue:44492", - "accept": false, - "reason": "FSDP2 optimizer-state corruption and a typo in cache strategies are unrelated." - }, - { - "left": "issue:43992", - "right": "issue:45310", - "accept": false, - "reason": "Missing embed_tokens.weight during UMT5 loading and Qwen3.5 from_pretrained errors are different model-loading bugs." - }, - { - "left": "issue:44075", - "right": "issue:46365", - "accept": false, - "reason": "Optimizer argument handling and Trainer import conflict with datasets are unrelated." + "reason": "Gemma3 config attribute loss and Qwen3.5 num_labels propagation are different config problems." } ] }, @@ -16229,7 +16109,7 @@ "source": "llm", "cache_hit": false, "model": "gpt-5.4-mini?service_tier=flex", - "cluster_id": "cluster-10000-472", + "cluster_id": "cluster-10000-473", "nodes": [ "issue:28282", "issue:33453", @@ -16240,10 +16120,8 @@ "issue:42898", "issue:43010", "issue:43064", - "issue:43065", "issue:43502", "issue:43575", - "issue:43576", "issue:43618", "issue:43723", "issue:43867", @@ -16251,13 +16129,14 @@ "issue:43976", "issue:43992", "issue:44062", + "issue:44075", "issue:44077", "issue:44112", "issue:44222", "issue:44246", "issue:44355", "issue:44360", - "issue:44387", + "issue:44492", "issue:44556", "issue:44734", "issue:44855", @@ -16267,6 +16146,7 @@ "issue:45071", "issue:45125", "issue:45278", + "issue:45310", "issue:45357", "issue:45375", "issue:45440", @@ -16274,13 +16154,16 @@ "issue:45561", "issue:45600", "issue:45850", - "issue:46132", "issue:46202", "issue:46285", + "issue:46365", "issue:46617", "issue:46664" ], "soft_pairs": [ + "issue:43992|issue:45310", + "issue:43064|issue:44492", + "issue:44075|issue:46365", "issue:43010|issue:45278", "issue:44355|issue:44855", "issue:44938|issue:45071", @@ -16291,235 +16174,226 @@ "issue:41669|issue:44246", "issue:43010|issue:45850", "issue:36296|issue:43575", - "issue:43867|issue:45357", "issue:43010|issue:43976", "issue:43618|issue:44077", + "issue:43867|issue:45357", "issue:43992|issue:45478", "issue:28282|issue:33453", "issue:43618|issue:46285", "issue:43618|issue:44857", "issue:44360|issue:45440", "issue:42890|issue:44964", - "issue:42898|issue:46664", "issue:42890|issue:44112", + "issue:42898|issue:46664", "issue:42890|issue:45600", - "issue:43502|issue:44062", "issue:43874|issue:46617", - "issue:43723|issue:44556", + "issue:43502|issue:44062", "issue:44222|issue:45125", - "issue:44387|issue:44556", - "issue:44734|issue:46617", - "issue:43010|issue:43065", - "issue:43576|issue:46132" + "issue:43723|issue:44556", + "issue:44734|issue:46617" ], - "prepared_review_unit_hash": "d1a1de33e60204517f6147e28536502dc43e4d0f165def254f02c6d408f3be2d", + "prepared_review_unit_hash": "4d6eca840a19d73fbe1b3b53ab595c8c94400f6bdf5520b5f76982de59d67260", "component_budget": { - "node_count": 472, - "item_count": 472, - "soft_pair_count": 1732, - "serialized_chars": 898654, - "estimated_input_tokens": 224664, - "estimated_eval_tokens": 449584 + "node_count": 473, + "item_count": 473, + "soft_pair_count": 1738, + "serialized_chars": 901437, + "estimated_input_tokens": 225360, + "estimated_eval_tokens": 450976 }, "budget": { "node_count": 48, "item_count": 48, "soft_pair_count": 30, - "serialized_chars": 36315, - "estimated_input_tokens": 9079, - "estimated_eval_tokens": 18414 + "serialized_chars": 36253, + "estimated_input_tokens": 9064, + "estimated_eval_tokens": 18384 }, "overflow_policy": "truncate_then_skip", "trimmed": false, "aggressively_trimmed": false, "split": true, "analyst_result": { - "summary": "Mostly broad v5 regression reports with only topical similarity; I don\u2019t see a strong duplicate cluster here, so all soft pairs are rejected.", - "confidence": 0.79, - "canonical_issue_reason": "issue:45278 is the best umbrella issue in this set: it is concrete, version-specific, and broadly matches the recurring v5 import/load regression theme better than the more narrow or vague reports.", + "summary": "This cluster is a loose grouping of transformers v5 regression reports, mostly centered on import/loading failures and model/config compatibility breakages. The strongest representative issue is the broad upgrade-regression report, while the soft pairs are not the same underlying bug.", + "confidence": 0.82, + "canonical_issue_reason": "issue:45278 is the broadest and most representative report: it summarizes many import failures after upgrading from 4.57.0 to 5.5.0, matching the overall cluster theme better than narrower model-specific or API-specific reports.", "canonical_pr_reason": null, - "best_issue_reason": "issue:45278 is the most suitable global representative because it is the clearest high-level report of the shared upgrade regression pattern, while still being specific enough to anchor triage.", + "best_issue_reason": "issue:45278 best captures the cluster\u2019s shared failure mode (post-v5 import/load regressions across models), making it a good umbrella issue for deduping related reports.", "best_pr_reason": null, "soft_edge_verdicts": [ + { + "left": "issue:43992", + "right": "issue:45310", + "accept": false, + "reason": "Both are Qwen/encoder loading problems, but they describe different concrete failures: missing embed_tokens.weight vs a Qwen3.5 Moe from_pretrained error. Too different to treat as the same bug." + }, + { + "left": "issue:43064", + "right": "issue:44492", + "accept": false, + "reason": "One is a trainer/FSDP/PEFT optimizer-state corruption bug; the other is a cache-strategy typo. Same broad v5 area, but not the same code path or fix." + }, + { + "left": "issue:44075", + "right": "issue:46365", + "accept": false, + "reason": "Optimizer arguments not being used is unrelated to importing Trainer being blocked by datasets. No shared underlying defect." + }, { "left": "issue:43010", "right": "issue:45278", "accept": false, - "reason": "Cache/layer update no_grad decoration is a different bug from broad import errors after upgrading to v5.5.0." + "reason": "Decorating cache/layer update methods with no_grad is a specific training/memory-safety change, while 45278 is a broad import-error regression report. Not the same issue." }, { "left": "issue:44355", "right": "issue:44855", "accept": false, - "reason": "Compiled Python-file import failures and the torch.jit comment/IndentationError are different code paths and failure modes." + "reason": "Compiled Python file errors and a Python 3.13 torch.jit parsing/IndentationError are different failure mechanisms." }, { "left": "issue:44938", "right": "issue:45071", "accept": false, - "reason": "Python 3.14 import failure is unrelated to PretrainedConfig type-checking breakage." + "reason": "Both involve Python/version compatibility, but one is a load failure on Python 3.14 and the other is a PretrainedConfig type-checking regression in v5.4.0. Not the same bug." }, { "left": "issue:43867", "right": "issue:45375", "accept": false, - "reason": "State-dict loading order problems are not the same as a missing deepstack_visual_indexes config field." + "reason": "Model loading with sorted state_dict and a missing vision-config field are unrelated Qwen3.5 failures." }, { "left": "issue:43867", "right": "issue:46202", "accept": false, - "reason": "A load failure with sorted state_dicts is not the same as the Gemma pipeline roundtrip token issue." + "reason": "State-dict sorting causing load failure and a Gemma pipeline roundtrip issue are distinct model/data-path problems." }, { "left": "issue:36683", "right": "issue:43064", "accept": false, - "reason": "Missing Gemma3Config vocab_size and FSDP2/PEFT optimizer-state corruption are unrelated bugs." + "reason": "Gemma3Config missing vocab_size is a config-attribute bug; the other is optimizer-state mismatch under FSDP2/PEFT. Different layers of the stack." }, { "left": "issue:42890", "right": "issue:45561", "accept": false, - "reason": "Flaky test seeding problems and an xdist race on captured_info.txt are different test infra issues." + "reason": "A missing set_seed in a flaky model integration test is unrelated to xdist workers racing on captured_info.txt in testing utilities." }, { "left": "issue:41669", "right": "issue:44246", "accept": false, - "reason": "Both mention slow imports, but one is a concrete import-star cleanup issue and the other is a vague intermittent slowdown report; not enough to confirm the same fix." + "reason": "Removing import-* usage to reduce import time is a performance/refactor issue; import transformers taking long sometimes is a separate symptom and not the same concrete change." }, { "left": "issue:43010", "right": "issue:45850", "accept": false, - "reason": "Torch-no_grad decoration of update methods is not the same as a broad statement that v5 breaks too many models." + "reason": "Both are broad complaints about v5 breakage, but one is a specific no_grad decoration request and the other is a generic 'too many models break' meta issue." }, { "left": "issue:36296", "right": "issue:43575", "accept": false, - "reason": "Tensor-parallel training bugs and Qwen2-57B-A14B-Instruct OOM on load are different problems." - }, - { - "left": "issue:43867", - "right": "issue:45357", - "accept": false, - "reason": "Sorted state_dict loading and incorrect visual-encoder keys in save_pretrained are different regression points." + "reason": "Tensor-parallel training bug and OOM when loading Qwen2-57B-A14B-Instruct with tp are related only at a high level; the concrete causes are not established as the same." }, { "left": "issue:43010", "right": "issue:43976", "accept": false, - "reason": "A no_grad decoration request has nothing to do with Python-version compatibility on load." + "reason": "Cache update no_grad is a code-path-specific safety fix; Python 3.9/3.10 compatibility for Transformers 5.1.0 is a version-support issue." }, { "left": "issue:43618", "right": "issue:44077", "accept": false, - "reason": "Missing attentions assignment in CLIPOutput is unrelated to patchtsmixer post_init policy." + "reason": "CLIPOutput attentions no longer assigned and PatchTSMixer post_init being optional are unrelated model-specific API regressions." + }, + { + "left": "issue:43867", + "right": "issue:45357", + "accept": false, + "reason": "A general load-model state_dict ordering bug and incorrect visual encoder keys in Qwen3.5 save_pretrained are different save/load defects." }, { "left": "issue:43992", "right": "issue:45478", "accept": false, - "reason": "Missing embed_tokens.weight in UMT5Encoder and Qwen3.5 MoE from_pretrained errors are different model-specific loading issues." + "reason": "Missing embed_tokens.weight in UMT5Encoder and the Qwen3.5 Moe from_pretrained error affect different model families and different loading paths." }, { "left": "issue:28282", "right": "issue:33453", "accept": false, - "reason": "PyTorch-not-found ImportError and tokenizer-loading regression are unrelated." + "reason": "AutoModel requiring PyTorch and tokenizer loading regression are both import/load issues, but they are distinct failure modes with different dependencies and code paths." }, { "left": "issue:43618", "right": "issue:46285", "accept": false, - "reason": "CLIP attentions assignment and CLIPTextModel structure inconsistency are not the same concrete bug." + "reason": "CLIP attentions assignment and CLIP text-model structure divergence are both CLIP-related, but they are different API changes and fixes." }, { "left": "issue:43618", "right": "issue:44857", "accept": false, - "reason": "CLIPOutput attentions and LwDetrImageLoss AMP crashes are unrelated." + "reason": "CLIP attentions missing and LwDetrImageLoss AMP/CUDA crash are unrelated model/runtime bugs." }, { "left": "issue:44360", "right": "issue:45440", "accept": false, - "reason": "A missing ReLU in the DSA indexer and DeepseekV3MoE divergence from remote implementation are different issues." + "reason": "The DSA indexer discussion and DeepseekV3MoE divergence describe completely different components and problems." }, { "left": "issue:42890", "right": "issue:44964", "accept": false, - "reason": "Seed-related flaky tests and Phi-4 multimodal model loading are unrelated." - }, - { - "left": "issue:42898", - "right": "issue:46664", - "accept": false, - "reason": "Tokenization-space behavior changes and SmolVLM num_frames semantics are different API changes." + "reason": "A flaky SAM HQ test due to missing seeds is unrelated to loading Phi-4 multimodal with the latest transformers." }, { "left": "issue:42890", "right": "issue:44112", "accept": false, - "reason": "SamHQ test seeding flakiness and GraniteSpeech stale device-override CI failure are different test failures." + "reason": "Seed-related test flakiness and a GraniteSpeech stale device override CI failure are different test issues with different root causes." }, { "left": "issue:42890", "right": "issue:45600", "accept": false, - "reason": "A missing set_seed in SamHQ tests is unrelated to auto_mappings referencing removed Sam3LiteText configs." - }, - { - "left": "issue:43502", - "right": "issue:44062", - "accept": false, - "reason": "local_files_only network leakage and AddedToken special-argument duplication are different tokenizer/loading bugs." + "reason": "Missing set_seed in tests and auto_mappings referencing removed configs are unrelated." }, { "left": "issue:43874", "right": "issue:46617", "accept": false, - "reason": "Missing get_number_of_image_patches in an image processor is unrelated to cache loading FileNotFoundError for trust_remote_code models." + "reason": "A missing image-patch helper in GLM46V and a trust_remote_code cache FileNotFoundError are different multimodal loading problems." }, { - "left": "issue:43723", - "right": "issue:44556", + "left": "issue:43502", + "right": "issue:44062", "accept": false, - "reason": "Tokenizer loading in v5 and reloading a v4.57 checkpoint after upgrade are related only at a very broad level, not the same underlying bug." + "reason": "local_files_only being ignored and AddedToken getting duplicate 'special' args are unrelated tokenizer/utilities bugs." }, { "left": "issue:44222", "right": "issue:45125", "accept": false, - "reason": "FP8 save_pretrained issues and missing _tp_plan for tensor parallelism are distinct concerns." + "reason": "FP8 save_pretrained for MoE and missing _tp_plan for tensor parallelism are separate model capability issues." }, { - "left": "issue:44387", + "left": "issue:43723", "right": "issue:44556", "accept": false, - "reason": "Int4 quantization memory regression and checkpoint reload incompatibility are different bugs." + "reason": "Tokenizer loading in v5 and inability to reload a checkpoint after upgrading are both version-regression reports, but not the same concrete bug." }, { "left": "issue:44734", "right": "issue:46617", "accept": false, - "reason": "KV-cache continuation indexing in serve and cache-loading FileNotFoundError are unrelated." - }, - { - "left": "issue:43010", - "right": "issue:43065", - "accept": false, - "reason": "Decorating cache update methods with no_grad is unrelated to the dummy Conv2d in Sam3PixelDecoder." - }, - { - "left": "issue:43576", - "right": "issue:46132", - "accept": false, - "reason": "Broken transformers env command and AttentionInterface.register behavior changes are different issues." + "reason": "KV-cache continuation crashing in transformers serve and a cache-loading FileNotFoundError for trust_remote_code models are different serving/loading failures." } ] }, @@ -16541,7 +16415,7 @@ "source": "llm", "cache_hit": false, "model": "gpt-5.4-mini?service_tier=flex", - "cluster_id": "cluster-10000-472", + "cluster_id": "cluster-10000-473", "nodes": [ "issue:27957", "issue:30333", @@ -16555,13 +16429,13 @@ "issue:42548", "issue:42757", "issue:43010", + "issue:43065", "issue:43232", "issue:43299", "issue:43475", "issue:43502", "issue:43526", "issue:43576", - "issue:43673", "issue:43824", "issue:43827", "issue:43901", @@ -16571,8 +16445,8 @@ "issue:44351", "issue:44355", "issue:44360", + "issue:44387", "issue:44485", - "issue:44492", "issue:44556", "issue:44655", "issue:44749", @@ -16581,11 +16455,10 @@ "issue:44961", "issue:45071", "issue:45125", - "issue:45362", "issue:45375", "issue:45468", "issue:45478", - "issue:45850", + "issue:46132", "issue:46143", "issue:46144", "issue:46209", @@ -16593,16 +16466,19 @@ "issue:46664" ], "soft_pairs": [ + "issue:44387|issue:44556", + "issue:43010|issue:43065", "issue:43824|issue:46617", "issue:43502|issue:44485", + "issue:43576|issue:46132", "issue:45125|issue:45478", "issue:30333|issue:33453", "issue:44355|issue:44749", "issue:42548|issue:44351", "issue:45125|issue:45375", "issue:43010|issue:43232", - "issue:30990|issue:31795", "issue:43502|issue:44222", + "issue:30990|issue:31795", "issue:43502|issue:44961", "issue:43827|issue:44961", "issue:30333|issue:36296", @@ -16628,260 +16504,270 @@ "issue:43475|issue:46617", "issue:41628|issue:42548", "issue:44315|issue:46209", - "issue:44485|issue:46664", - "issue:45362|issue:45850", - "issue:43673|issue:44492" + "issue:44485|issue:46664" ], - "prepared_review_unit_hash": "e4ceee44049ca0de43147779c4944357ebd9253b1dbb077dd9f4312af771168f", + "prepared_review_unit_hash": "715ed05dbfb54fd33a6bb04f797c1bc70715353c08764b10c0b4f575828320a1", "component_budget": { - "node_count": 472, - "item_count": 472, - "soft_pair_count": 1732, - "serialized_chars": 898654, - "estimated_input_tokens": 224664, - "estimated_eval_tokens": 449584 + "node_count": 473, + "item_count": 473, + "soft_pair_count": 1738, + "serialized_chars": 901437, + "estimated_input_tokens": 225360, + "estimated_eval_tokens": 450976 }, "budget": { - "node_count": 48, - "item_count": 48, - "soft_pair_count": 38, - "serialized_chars": 38280, - "estimated_input_tokens": 9570, - "estimated_eval_tokens": 19396 + "node_count": 47, + "item_count": 47, + "soft_pair_count": 39, + "serialized_chars": 38114, + "estimated_input_tokens": 9529, + "estimated_eval_tokens": 19314 }, "overflow_policy": "truncate_then_skip", "trimmed": false, "aggressively_trimmed": false, "split": true, "analyst_result": { - "summary": "These items mostly form a loose Transformers v5 regression/loading/import cluster, but the titles don\u2019t align tightly enough to treat the soft links as true duplicates.", - "confidence": 0.67, - "canonical_issue_reason": "Issue 45850 is the broadest umbrella for the cluster\u2019s repeated \u201cv5 breaks X\u201d reports, even though it is very generic.", + "summary": "This set is a heterogeneous mix of unrelated Transformers issues: imports, model loading, docs, schedulers, quantization, and model-specific regressions. There is no clear single duplicate cluster; issue 43576 is only the broadest hub among the soft links, not a true shared root cause.", + "confidence": 0.89, + "canonical_issue_reason": "No single issue is a real canonical duplicate target across the cluster. Issue 43576 is the closest hub only because it sits near several v5-regression reports, but its env-command failure is not the same root cause as the others.", "canonical_pr_reason": null, - "best_issue_reason": "Issue 45850 is the most representative cluster anchor because it captures the shared broad regression theme better than the narrower model-specific reports.", + "best_issue_reason": "Issue 43576 is the best representative by connectivity and breadth, but it still does not unify the cluster into one underlying bug.", "best_pr_reason": null, "soft_edge_verdicts": [ + { + "left": "issue:44387", + "right": "issue:44556", + "accept": false, + "reason": "Both are v5-era failures, but one is int4/OOM memory behavior and the other is checkpoint reload compatibility; different bugs." + }, + { + "left": "issue:43010", + "right": "issue:43065", + "accept": false, + "reason": "Unrelated code paths: cache/layer update decoration vs a dummy Conv2d in Sam3PixelDecoder." + }, { "left": "issue:43824", "right": "issue:46617", "accept": false, - "reason": "Different failures: import error vs cache FileNotFoundError." + "reason": "ImportError for a specific model class vs FileNotFoundError when loading trust_remote_code from cache; not the same issue." }, { "left": "issue:43502", "right": "issue:44485", "accept": false, - "reason": "Local-files-only network bug and GLM RoPE bug are unrelated." + "reason": "local_files_only network leakage vs GLM-5 RoPE implementation; different subsystems and symptoms." + }, + { + "left": "issue:43576", + "right": "issue:46132", + "accept": false, + "reason": "Broken env command in v5 vs AttentionInterface.register behavior change; same broad release, different bug." }, { "left": "issue:45125", "right": "issue:45478", "accept": false, - "reason": "Same model family, but one is a tensor-parallel plan bug and the other is a broader from_pretrained regression; not clearly the same defect." + "reason": "Both involve Qwen3.5 MoE loading, but one is missing _tp_plan for tensor parallelism and the other is a broader from_pretrained error; not enough to merge." }, { "left": "issue:30333", "right": "issue:33453", "accept": false, - "reason": "MLFlow reporting failure and tokenizer loading regression are unrelated." + "reason": "MLflow reporting on failed jobs vs tokenizer loading regression; unrelated failures." }, { "left": "issue:44355", "right": "issue:44749", "accept": false, - "reason": "Compiled-file loading errors and performance regression are different issues." + "reason": "Compiled Python file errors vs slowdown after upgrading Transformers; distinct symptoms and code paths." }, { "left": "issue:42548", "right": "issue:44351", "accept": false, - "reason": "Different missing symbols: PreTrainedModel vs HybridCache." + "reason": "Importing PreTrainedModel vs importing HybridCache; both import errors but for different symbols." }, { "left": "issue:45125", "right": "issue:45375", "accept": false, - "reason": "Both are Qwen3.5 MoE-related, but they describe different missing fields/code paths." + "reason": "Tensor-parallel plan missing for Qwen3_5Moe vs missing deepstack_visual_indexes in config; different model fields." }, { "left": "issue:43010", "right": "issue:43232", "accept": false, - "reason": "Both touch generation internals, but the bugs are distinct." + "reason": "Cache/Layer update no_grad issue vs generation kwargs update after sync_gpus; separate generation concerns." }, { - "left": "issue:30990", - "right": "issue:31795", + "left": "issue:43502", + "right": "issue:44222", "accept": false, - "reason": "Loading hang and documentation confusion are not duplicates." + "reason": "Offline/local-files-only network issue vs FP8 moe save_pretrained bug; different functionality." }, { - "left": "issue:43502", - "right": "issue:44222", + "left": "issue:30990", + "right": "issue:31795", "accept": false, - "reason": "Offline/network handling and FP8 save_pretrained are separate problems." + "reason": "Loading gets stuck for Sentence Transformers vs documentation confusion in model.forward; unrelated." }, { "left": "issue:43502", "right": "issue:44961", "accept": false, - "reason": "No relation beyond being unrelated bug reports." + "reason": "Network access despite local_files_only vs an unrelated issue titled racoon; no shared bug." }, { "left": "issue:43827", "right": "issue:44961", "accept": false, - "reason": "Documentation cleanup and an unrelated placeholder issue." + "reason": "Docs still referencing pipeline() vs unrelated racoon issue." }, { "left": "issue:30333", "right": "issue:36296", "accept": false, - "reason": "MLFlow reporting and tensor-parallel training are different subsystems." + "reason": "Failed-job MLflow reporting vs tensor parallel training bug; different areas." }, { "left": "issue:43299", "right": "issue:44556", "accept": false, - "reason": "Both are version regressions, but they affect different models and loading paths." + "reason": "Qwen3VL MoE model loading regression vs checkpoint reload incompatibility after version upgrade; not the same root cause." }, { "left": "issue:44355", "right": "issue:45071", "accept": false, - "reason": "Compiled-file loading errors and PretrainedConfig typing are unrelated." + "reason": "Compiled-file runtime errors vs PretrainedConfig type checking; unrelated." }, { "left": "issue:44355", "right": "issue:44655", "accept": false, - "reason": "Different problems: compiled Python file handling vs pipeline save_pretrained." + "reason": "Compiled-file runtime errors vs pipeline save_pretrained failure; different problems." }, { "left": "issue:33453", "right": "issue:34689", "accept": false, - "reason": "Both involve loading, but one is tokenizer regression and the other is a specific model-loading breakage." + "reason": "Tokenizer loading regression vs Llama 3.2 90B Vision model loading breakage; both loading-related but different models and failure modes." }, { "left": "issue:27957", "right": "issue:33453", "accept": false, - "reason": "Flash Attention support and tokenizer loading regression are unrelated." + "reason": "XLMRoberta Flash Attention 2 support vs tokenizer loading regression; unrelated." }, { "left": "issue:42757", "right": "issue:43576", "accept": false, - "reason": "Importing is_offline_mode and the env CLI breakage are different issues." + "reason": "is_offline_mode import failure vs broken env command; different breakages." }, { "left": "issue:43576", "right": "issue:45071", "accept": false, - "reason": "CLI breakage in v5 and PretrainedConfig type-checking breakage are separate." + "reason": "Broken env command vs PretrainedConfig type-checking regression; not the same bug." }, { "left": "issue:44315", "right": "issue:44360", "accept": false, - "reason": "Liger Kernel/model_init and DSA indexer/ReLU are unrelated." + "reason": "Liger Kernel not applied with model_init vs DSA indexer lacks ReLU; unrelated." }, { "left": "issue:43576", "right": "issue:43824", "accept": false, - "reason": "Env command regression and Qwen import failure do not share a code path." + "reason": "Env command breakage vs Qwen2.5 VL import error; different commands and code paths." }, { "left": "issue:31795", "right": "issue:38617", "accept": false, - "reason": "Documentation confusion and import failure are unrelated." + "reason": "Docs confusion about forward args vs ImportError for layer_type_validation; documentation vs API export bug." }, { "left": "issue:44961", "right": "issue:45468", "accept": false, - "reason": "Placeholder issue and Gemma-4 audio bug are unrelated." + "reason": "Unrelated issue titled racoon vs Gemma-4 audio positional encoding bug." }, { "left": "issue:44351", "right": "issue:46617", "accept": false, - "reason": "Both are load/import failures, but one is an import symbol issue and the other is cache file lookup." + "reason": "HybridCache import error vs cache FileNotFoundError for trust_remote_code; both cache/import themed but different failures." }, { "left": "issue:44485", "right": "issue:44961", "accept": false, - "reason": "RoPE implementation bug and placeholder issue are unrelated." + "reason": "GLM-5 RoPE implementation vs unrelated racoon issue; no overlap." + }, + { + "left": "issue:43576", + "right": "issue:44938", + "accept": false, + "reason": "Broken env command vs Python 3.14 loading failure; different issues." }, { "left": "issue:44908", "right": "issue:44961", "accept": false, - "reason": "Scheduler kwargs bug and placeholder issue are unrelated." + "reason": "Scheduler kwargs bug vs unrelated racoon issue." }, { "left": "issue:44222", "right": "issue:44230", "accept": false, - "reason": "Both are FP8/Qwen-related, but one is save_pretrained and the other is inference support." + "reason": "FP8 save_pretrained bug vs FP8 inference support request; related domain but not the same concrete bug." }, { "left": "issue:36296", "right": "issue:43901", "accept": false, - "reason": "Tensor-parallel training and a docs-only pipeline note are unrelated." + "reason": "Tensor parallel training bug vs docs still mentioning return_all_scores; unrelated." }, { "left": "issue:46143", "right": "issue:46144", "accept": false, - "reason": "RoFormer kwargs propagation and attention-interface use are related but not the same concrete bug." + "reason": "RoFormer kwargs propagation bug vs RoFormer attention not using the attention interface; adjacent but distinct." }, { "left": "issue:36296", "right": "issue:43526", "accept": false, - "reason": "Tensor-parallel training bug and BeitImageProcessorFast label reduction are unrelated." + "reason": "Tensor parallel training bug vs BeitImageProcessorFast reduce_labels returning one label; unrelated." }, { "left": "issue:43475", "right": "issue:46617", "accept": false, - "reason": "SAM3 vision output attribute bug and cache FileNotFoundError are unrelated." + "reason": "Sam3VisionEncoderOutput missing attribute vs trust_remote_code cache FileNotFoundError; different model/load paths." }, { "left": "issue:41628", "right": "issue:42548", "accept": false, - "reason": "Different missing imports from different APIs." + "reason": "AutoImageProcessor import failure vs PreTrainedModel import failure; both imports but different missing symbols." }, { "left": "issue:44315", "right": "issue:46209", "accept": false, - "reason": "Different model/runtime problems: Liger Kernel application vs FP8 inference on B200." + "reason": "Liger Kernel model_init issue vs FP8 inference on B200; unrelated implementation areas." }, { "left": "issue:44485", "right": "issue:46664", "accept": false, - "reason": "GLM-5 RoPE bug and SmolVLM num_frames semantics are unrelated." - }, - { - "left": "issue:45362", - "right": "issue:45850", - "accept": false, - "reason": "The latter is a broad umbrella, but the concrete model crash is a different report." - }, - { - "left": "issue:43673", - "right": "issue:44492", - "accept": false, - "reason": "Generation cache regression and a cache-strategy typo are not the same bug." + "reason": "GLM-5 RoPE bug vs SmolVLM num_frames meaning inconsistency; unrelated model-specific issues." } ] }, @@ -16903,7 +16789,7 @@ "source": "llm", "cache_hit": false, "model": "gpt-5.4-mini?service_tier=flex", - "cluster_id": "cluster-10000-472", + "cluster_id": "cluster-10000-473", "nodes": [ "issue:31795", "issue:33453", @@ -16943,6 +16829,7 @@ "issue:45230", "issue:45278", "issue:45341", + "issue:45362", "issue:45381", "issue:45478", "issue:45542", @@ -16954,6 +16841,8 @@ "issue:46664" ], "soft_pairs": [ + "issue:45362|issue:45850", + "issue:43673|issue:44492", "issue:44623|issue:45230", "issue:44360|issue:46664", "issue:44273|issue:45542", @@ -16977,11 +16866,11 @@ "issue:44246|issue:44273", "issue:43352|issue:44393", "issue:43576|issue:44908", - "issue:43976|issue:44393", "issue:44704|issue:44908", + "issue:43976|issue:44393", "issue:44908|issue:45071", - "issue:44295|issue:45230", "issue:44297|issue:45230", + "issue:44295|issue:45230", "issue:41669|issue:44351", "issue:44492|issue:46328", "issue:41669|issue:45341", @@ -16994,268 +16883,280 @@ "issue:44315|issue:44829", "issue:43576|issue:44273" ], - "prepared_review_unit_hash": "445adcd4852e5c084f972faffdf1233c7452f15bb62daebd2512e11f466dc318", + "prepared_review_unit_hash": "76b7305fca47565e695b749088ace364ce46bf73268541b4f38b302d1eddc764", "component_budget": { - "node_count": 472, - "item_count": 472, - "soft_pair_count": 1732, - "serialized_chars": 898654, - "estimated_input_tokens": 224664, - "estimated_eval_tokens": 449584 + "node_count": 473, + "item_count": 473, + "soft_pair_count": 1738, + "serialized_chars": 901437, + "estimated_input_tokens": 225360, + "estimated_eval_tokens": 450976 }, "budget": { - "node_count": 47, - "item_count": 47, - "soft_pair_count": 39, - "serialized_chars": 38293, - "estimated_input_tokens": 9574, - "estimated_eval_tokens": 19404 + "node_count": 48, + "item_count": 48, + "soft_pair_count": 41, + "serialized_chars": 39432, + "estimated_input_tokens": 9858, + "estimated_eval_tokens": 19972 }, "overflow_policy": "truncate_then_skip", "trimmed": false, "aggressively_trimmed": false, "split": true, "analyst_result": { - "summary": "This is a loose cluster of mostly unrelated Transformers v5 regressions, import/loading issues, and model-specific bugs; the soft pairs are mostly topical overlaps rather than true duplicates.", - "confidence": 0.39, - "canonical_issue_reason": "issue:45850 is the broadest umbrella here: it describes generalized breakage after v5 and can represent the overall cluster better than the narrower, model-specific reports.", + "summary": "The cluster is dominated by broad Transformers v5 regression complaints, but the individual issues mostly point to different concrete failures (model-specific, import/loading, config, processor, or docs). No soft pair looks like a safe duplicate/merge candidate for a PR.", + "confidence": 0.83, + "canonical_issue_reason": "issue:45850 is the broadest umbrella-style report in the set (\u201cSince 5.0 version it breaks too many models\u201d) and best matches the overall regression theme, even though it is not a true duplicate of the others.", "canonical_pr_reason": null, - "best_issue_reason": "issue:45850 is the best global representative because it captures the widest user-visible regression theme across the cluster, even though it is vague.", + "best_issue_reason": "issue:45850 is the most central/general issue for the cluster because it captures the shared v5-breakage theme across multiple reports without being tied to one model or utility.", "best_pr_reason": null, "soft_edge_verdicts": [ + { + "left": "issue:45362", + "right": "issue:45850", + "accept": false, + "reason": "Same general v5 frustration, but one is a specific Qwen chat crash and the other is a broad umbrella complaint." + }, + { + "left": "issue:43673", + "right": "issue:44492", + "accept": false, + "reason": "Chunked-prefill cache regression vs. a cache-strategy typo; different bugs." + }, { "left": "issue:44623", "right": "issue:45230", "accept": false, - "reason": "One is about processor.save_pretrained missing files; the other is a generic bug report with no concrete shared failure mode." + "reason": "Specific processor save_pretrained file loss vs. a generic bug report." }, { "left": "issue:44360", "right": "issue:46664", "accept": false, - "reason": "DSA indexer/ReLU discussion and SmolVLM num_frames semantics are unrelated code paths." + "reason": "Completely different subsystems: DSA/indexer vs. video processor frame semantics." }, { "left": "issue:44273", "right": "issue:45542", "accept": false, - "reason": "Lazy loading failures and a TensorBoard-without-TensorFlow backend error are different subsystems and symptoms." + "reason": "Lazy loading/import mechanics vs. TensorFlow backend detection; no same code path." }, { "left": "issue:44038", "right": "issue:45125", "accept": false, - "reason": "Both mention Qwen3.5/Qwen3-VL-Moe, but one is a model bug and the other is a missing tensor-parallel plan; not the same fix." + "reason": "Both involve Qwen3-family models, but one is model support and the other is tensor-parallel metadata." }, { "left": "issue:44273", "right": "issue:45003", "accept": false, - "reason": "Lazy loading not working and unsafe sys.modules access are not the same underlying bug." + "reason": "Both are import/lazy-loading adjacent, but the symptoms and concrete failure modes differ." }, { "left": "issue:44393", "right": "issue:45753", "accept": false, - "reason": "Bounding-box output hallucination and an infinite loop on a specific image are distinct failure modes." + "reason": "Both are Qwen3-VL image issues, but hallucination/bbox output is not the same bug as an infinite loop on one image." }, { "left": "issue:44393", "right": "issue:45071", "accept": false, - "reason": "A Qwen3-VL output bug is unrelated to a PretrainedConfig type-checking regression." + "reason": "Model output bug vs. a v5 PretrainedConfig type-checking regression." }, { "left": "issue:45381", "right": "issue:46664", "accept": false, - "reason": "Both involve vision/video inputs, but the concrete bugs differ: wrong vision_position_ids vs inconsistent num_frames semantics." + "reason": "Both are video-related, but they affect different models and different notions of position/frame handling." }, { "left": "issue:43576", "right": "issue:44556", "accept": false, - "reason": "Broken env command output and checkpoint reload compatibility are different features and fixes." + "reason": "CLI/env command breakage vs. checkpoint reload compatibility; related to v5, but not the same bug." }, { "left": "issue:41669", "right": "issue:43673", "accept": false, - "reason": "Import-* slowdown and missing GenerationMixin cache in chunked_prefill are unrelated performance/code-path issues." + "reason": "Import-* performance problem vs. cache-missing generation regression; different issues." }, { "left": "issue:43525", "right": "issue:45850", "accept": false, - "reason": "A specific Llama4Config AttributeError is not the same as a broad v5 breakage report." + "reason": "Specific missing config attribute vs. a broad v5-breakage umbrella." }, { "left": "issue:44230", "right": "issue:46095", "accept": false, - "reason": "FP8 support for Qwen3-VL/Qwen3.5 Moe and a DeepSeek weight-conversion question are unrelated." + "reason": "FP8 support request for Qwen models vs. DeepSeek weight-conversion question; unrelated." }, { "left": "issue:44556", "right": "issue:45092", "accept": false, - "reason": "Both concern v5 migration, but they affect different model families and different compatibility failures." + "reason": "Both concern loading older checkpoints in v5, but they target different checkpoint styles and failure causes." }, { "left": "issue:43572", "right": "issue:45850", "accept": false, - "reason": "Missing pad_token_idx in StableLmConfig is a concrete model config bug, not the same as the umbrella breakage report." + "reason": "StableLm pad-token config regression is model-specific and not the same as the broad umbrella complaint." }, { "left": "issue:45278", "right": "issue:45850", "accept": false, - "reason": "These overlap thematically on v5 regressions, but one is specifically about import errors and the other is a broad complaint." + "reason": "Both are broad v5 regression reports, but one is about import errors after upgrade and the other is a general claim that many models break." }, { "left": "issue:44492", "right": "issue:45850", "accept": false, - "reason": "A typo in cache strategies is too narrow and unrelated to the broad model breakage claim." + "reason": "Cache typo is a localized docs/code issue, not the broad v5 breakage umbrella." }, { "left": "issue:42371", "right": "issue:44492", "accept": false, - "reason": "TF32 API guidance and a cache-strategy typo are completely different issues." + "reason": "TF32 API warning and cache-strategy typo are unrelated." }, { "left": "issue:41669", "right": "issue:46218", "accept": false, - "reason": "Import-* slowdown and inability to import Qwen3_5ForCausalLM are different import problems with different causes." + "reason": "Import-time performance regression vs. inability to import a specific Qwen module." }, { "left": "issue:42890", "right": "issue:44908", "accept": false, - "reason": "A flaky test seed problem and an inverse_sqrt scheduler kwargs bug do not share an underlying cause." + "reason": "Test flakiness due to missing seeds vs. scheduler kwargs being ignored; unrelated." }, { "left": "issue:43408", "right": "issue:44315", "accept": false, - "reason": "A sam3_video vs sam3_tracker warning and Liger Kernel not applying with model_init are unrelated model initialization paths." + "reason": "Model-type mismatch warning vs. Liger Kernel not applied with model_init; different concerns." }, { "left": "issue:44246", "right": "issue:44273", "accept": false, - "reason": "Both mention loading/import behavior, but one is intermittent import slowness and the other is lazy loading correctness; not clearly the same bug." + "reason": "Import takes long sometimes vs. lazy loading malfunction; related area, but not the same concrete bug." }, { "left": "issue:43352", "right": "issue:44393", "accept": false, - "reason": "Flash Attention 2 support gating for Nemotron and Qwen3-VL bounding-box errors are unrelated." + "reason": "Flash-Attention-2 support error vs. Qwen3-VL bounding-box output bug; different failures." }, { "left": "issue:43576", "right": "issue:44908", "accept": false, - "reason": "Broken env command and scheduler kwargs handling are unrelated functionality." + "reason": "Environment CLI breakage vs. scheduler argument handling; no shared bug." }, { - "left": "issue:43976", - "right": "issue:44393", + "left": "issue:44704", + "right": "issue:44908", "accept": false, - "reason": "Python-version compatibility and Qwen3-VL bbox output issues are different problems." + "reason": "Processor cached_file kwargs forwarding vs. inverse_sqrt scheduler kwargs; unrelated." }, { - "left": "issue:44704", - "right": "issue:44908", + "left": "issue:43976", + "right": "issue:44393", "accept": false, - "reason": "AutoProcessor kwargs passthrough and inverse_sqrt scheduler timescale handling are unrelated code paths." + "reason": "Python version support issue vs. multimodal bounding-box error." }, { "left": "issue:44908", "right": "issue:45071", "accept": false, - "reason": "Scheduler kwargs being ignored and PretrainedConfig type-checking breakage are different regressions." + "reason": "Scheduler kwarg bug vs. PretrainedConfig type checking regression." }, { - "left": "issue:44295", + "left": "issue:44297", "right": "issue:45230", "accept": false, - "reason": "Reading position_ids after buffer registration is specific and not the same as a generic bug report." + "reason": "Tokenizer save_pretrained mismatch vs. generic bug report." }, { - "left": "issue:44297", + "left": "issue:44295", "right": "issue:45230", "accept": false, - "reason": "Tokenizer save_pretrained metadata mismatch is unrelated to the generic bug report." + "reason": "Position_ids buffer read error vs. generic bug report." }, { "left": "issue:41669", "right": "issue:44351", "accept": false, - "reason": "Import-* slowdown and missing HybridCache import are different import-time failures." + "reason": "Import-* performance regression vs. missing HybridCache import; different import-related bugs." }, { "left": "issue:44492", "right": "issue:46328", "accept": false, - "reason": "Cache-strategy typo and SequenceFeatureExtractor.pad performance are unrelated." + "reason": "Cache-strategy typo vs. feature-extractor padding performance issue." }, { "left": "issue:41669", "right": "issue:45341", "accept": false, - "reason": "Import-* slowdown and a testing_utils bug do not share the same underlying defect." + "reason": "Import-* slowdown vs. a small testing_utils bug." }, { "left": "issue:43976", "right": "issue:45850", "accept": false, - "reason": "Python compatibility breakage is not the same as the broad model-regression complaint." + "reason": "Python support regression vs. broad v5-breakage complaint." }, { "left": "issue:44393", "right": "issue:45478", "accept": false, - "reason": "Qwen3-VL bbox hallucination and Qwen3.5 Moe from_pretrained failure are different model bugs." + "reason": "Different Qwen3-VL-family failures, but not the same concrete error path." }, { "left": "issue:43976", "right": "issue:46218", "accept": false, - "reason": "Python-version compatibility and inability to import Qwen3_5ForCausalLM are unrelated." + "reason": "Python version compatibility issue vs. missing module import." }, { "left": "issue:44077", "right": "issue:44861", "accept": false, - "reason": "PatchTSMixer post_init validation and tied-weight key handling are different code paths." + "reason": "Optional post_init validation vs. tied-weights helper crash; unrelated." }, { "left": "issue:43827", "right": "issue:44908", "accept": false, - "reason": "Docs still referencing pipeline() and scheduler kwargs handling are unrelated issues." + "reason": "Docs still referencing removed pipeline() vs. scheduler kwargs bug." }, { "left": "issue:31795", "right": "issue:33453", "accept": false, - "reason": "Documentation confusion for forward args and tokenizer loading regression are not the same bug." + "reason": "Documentation confusion vs. tokenizer loading regression; only loosely related to model API usage." }, { "left": "issue:44315", "right": "issue:44829", "accept": false, - "reason": "Liger Kernel application with model_init and flash_attention_3 training degeneration are distinct problems." + "reason": "Liger Kernel application in training setup vs. flash_attention_3 causing degenerate training; not the same defect." }, { "left": "issue:43576", "right": "issue:44273", "accept": false, - "reason": "Both are v5 runtime regressions, but one is the env command and the other is lazy loading; no shared concrete fix is evident." + "reason": "Broken env command vs. lazy loading malfunction; both are v5-era issues, but different code paths." } ] }, @@ -17277,7 +17178,7 @@ "source": "llm", "cache_hit": false, "model": "gpt-5.4-mini?service_tier=flex", - "cluster_id": "cluster-10000-472", + "cluster_id": "cluster-10000-473", "nodes": [ "issue:28348", "issue:30333", @@ -17372,220 +17273,220 @@ "issue:31515|issue:33453", "issue:43298|issue:44230" ], - "prepared_review_unit_hash": "5ae7342d47f16cd275951105613d6fb6d2b51ba297ce69727131a8f37a1dc387", + "prepared_review_unit_hash": "b68388c7da49a2c2946e4d2f710bb59c3b0fcdc90148cd029d542d7a13d42e82", "component_budget": { - "node_count": 472, - "item_count": 472, - "soft_pair_count": 1732, - "serialized_chars": 898654, - "estimated_input_tokens": 224664, - "estimated_eval_tokens": 449584 + "node_count": 473, + "item_count": 473, + "soft_pair_count": 1738, + "serialized_chars": 901437, + "estimated_input_tokens": 225360, + "estimated_eval_tokens": 450976 }, "budget": { "node_count": 47, "item_count": 47, "soft_pair_count": 43, - "serialized_chars": 39010, - "estimated_input_tokens": 9753, - "estimated_eval_tokens": 19762 + "serialized_chars": 39013, + "estimated_input_tokens": 9754, + "estimated_eval_tokens": 19764 }, "overflow_policy": "truncate_then_skip", "trimmed": false, "aggressively_trimmed": false, "split": true, "analyst_result": { - "summary": "This set is mostly a grab bag of unrelated Transformers issues, not a tight duplicate cluster. I found no PRs, and none of the soft edges is strong enough to merge as the same underlying bug/change.", - "confidence": 0.23, - "canonical_issue_reason": "issue:42913 is the cleanest broad regression report in the set: tokenizer behavior changed from v4 to v5, which makes it the best umbrella issue among the tokenizer-related reports.", + "summary": "The soft pairs are mostly broad subsystem or version-related similarities, but none look like the same concrete bug/change closely enough to merge. The broadest umbrella issue is the general v5 breakage report.", + "confidence": 0.74, + "canonical_issue_reason": "issue:45850 is the strongest umbrella report: it broadly captures many regressions introduced by 5.0, making it the most plausible anchor for nearby duplicates.", "canonical_pr_reason": null, - "best_issue_reason": "issue:42913 is the best issue to anchor this group because it is specific enough to be actionable yet broad enough to represent the v4\u2192v5 tokenizer regression theme.", + "best_issue_reason": "issue:45850 has the widest scope and is likely to attract related duplicate reports; the others are narrower, more specific, or unrelated.", "best_pr_reason": null, "soft_edge_verdicts": [ { "left": "issue:43519", "right": "issue:46655", "accept": false, - "reason": "Both are processor-related bugs, but they concern different models and different calculations (timestamp vs padding)." + "reason": "Both are processor bugs, but they concern different models and different failure modes (timestamp math vs padding math)." }, { "left": "issue:44360", "right": "issue:45468", "accept": false, - "reason": "Unrelated model internals: a DSA indexer ReLU complaint versus Gemma-4 audio positional encoding." + "reason": "Different model components and unrelated bugs: DSA indexer activation vs Gemma audio positional encoding config." }, { "left": "issue:44655", "right": "issue:44861", "accept": false, - "reason": "Different code paths: saving Pipeline objects versus a tied-weights helper crashing on list input." + "reason": "Both relate to saving/loading internals, but one is pipeline save_pretrained and the other is tied-weights key handling; not the same bug." }, { "left": "issue:45230", "right": "issue:45310", "accept": false, - "reason": "One is a vague placeholder bug report; the other is a specific Qwen3.5 MoE from_pretrained failure." + "reason": "The first is an unspecified generic bug report; the second is a specific Qwen3.5 MoE from_pretrained regression." }, { "left": "issue:44273", "right": "issue:44908", "accept": false, - "reason": "Lazy loading and inverse_sqrt scheduler kwargs are unrelated problems." + "reason": "Lazy loading behavior and inverse_sqrt scheduler kwargs are unrelated code paths." }, { "left": "issue:43576", "right": "issue:46218", "accept": false, - "reason": "A broken env command is not the same issue as a module import failure for Qwen3_5ForCausalLM." + "reason": "One is a broken CLI/env command, the other is a missing model import; different problems." }, { "left": "issue:43519", "right": "issue:44485", "accept": false, - "reason": "Different multimodal model bugs: Qwen3VL timestamp math versus GLM-5 RoPE implementation." + "reason": "Different model bugs: Qwen3VL processor timestamp calculation vs GLM-5 RoPE implementation." }, { "left": "issue:43519", "right": "issue:44908", "accept": false, - "reason": "Processor timestamp handling and scheduler kwargs are unrelated." + "reason": "Processor timestamp math and scheduler kwargs handling are unrelated." }, { "left": "issue:41669", "right": "issue:43704", "accept": false, - "reason": "Import-star cleanup and a VRAM leak in multi-threaded inference are different problems." + "reason": "Import-star cleanup/perf regression and VRAM leak are different issues in different code paths." }, { "left": "issue:46018", "right": "issue:46172", "accept": false, - "reason": "Both mention DeepSeek-V4, but one is expert gating and the other is a tensor transpose/dtype issue." + "reason": "Both are DeepSeek-V4 related, but one is gating logic and the other is a tensor conversion/transposition bug." }, { "left": "issue:31795", "right": "issue:34689", "accept": false, - "reason": "Documentation confusion about forward args is unrelated to a model-loading regression." + "reason": "Documentation confusion vs an actual model-loading regression; not the same underlying bug." }, { "left": "issue:31515", "right": "issue:31795", "accept": false, - "reason": "Checkpoint loading speed and forward-pass documentation are not the same bug." + "reason": "One is slow from_pretrained checkpoint loading; the other is confusing forward docs." }, { "left": "issue:43704", "right": "issue:44485", "accept": false, - "reason": "VRAM leakage in dataloader threads and GLM-5 RoPE implementation are unrelated." + "reason": "VRAM leak in Qwen3ForCausalLM and GLM-5 RoPE implementation are unrelated." }, { "left": "issue:44485", "right": "issue:45468", "accept": false, - "reason": "Different model-specific bugs with no clear shared failure mode." + "reason": "Both are model-specific implementation issues, but they affect different architectures and code paths." }, { "left": "issue:45850", "right": "issue:46218", "accept": false, - "reason": "Generic v5 breakage is far too broad to be the same as a specific Qwen3_5 import error." + "reason": "Broad v5 breakage is not the same concrete issue as a specific Qwen3_5ForCausalLM import failure." }, { "left": "issue:43502", "right": "issue:43519", "accept": false, - "reason": "local_files_only network leakage and video timestamp calculation are unrelated." + "reason": "Local-files-only network leakage and video timestamp calculation are unrelated." }, { "left": "issue:42371", "right": "issue:43704", "accept": false, - "reason": "TF32 API guidance and a VRAM leak are not the same underlying issue." + "reason": "TF32 configuration guidance and a Qwen3ForCausalLM VRAM leak do not match." }, { "left": "issue:43519", "right": "issue:44623", "accept": false, - "reason": "Processor timestamp math and missing processor save files are different save/load bugs." + "reason": "Processor timestamp calculation and processor.save_pretrained missing files are different bugs." }, { "left": "issue:44263", "right": "issue:45468", "accept": false, - "reason": "A torch.split return-value issue in one model is unrelated to Gemma-4 audio positional encoding." + "reason": "DSA indexer split behavior and Gemma audio positional encoding are unrelated." }, { "left": "issue:43606", "right": "issue:45468", "accept": false, - "reason": "CPU offload device mismatch does not match an audio positional-encoding bug." + "reason": "CPU offload device mismatch is unrelated to Gemma audio positional encoding." }, { "left": "issue:43298", "right": "issue:46095", "accept": false, - "reason": "Continuous batching audio support and DeepSeek-V4 weight conversion are unrelated." + "reason": "Continuous batching support for audio models and a DeepSeek-V4 weight conversion question are not the same change." }, { "left": "issue:44261", "right": "issue:45468", "accept": false, - "reason": "Missing rms_norm_eps in one MLA path is unrelated to Gemma-4 audio positional encoding." + "reason": "Both mention model internals, but they are different architectures and different missing-configuration problems." }, { "left": "issue:43519", "right": "issue:43704", "accept": false, - "reason": "Timestamp calculation in a processor is not the same as a VRAM leak in dataloader threads." + "reason": "Video processor timestamp math and multi-thread VRAM leak are unrelated." }, { "left": "issue:44351", "right": "issue:46218", "accept": false, - "reason": "Importing a missing symbol and failing to import a Qwen3_5 module are different import-path failures." + "reason": "Both are import errors, but for different symbols and different root causes." }, { "left": "issue:43352", "right": "issue:46218", "accept": false, - "reason": "Flash Attention support for Nemotron is unrelated to a Qwen3_5 import error." + "reason": "Missing Flash Attention 2.0 support and a model import failure are different problems." }, { "left": "issue:42913", "right": "issue:43066", "accept": false, - "reason": "Both are tokenizer/v5 complaints, but one is a broad behavior regression and the other is a specific decoder-type bug; not enough evidence they are the same defect." + "reason": "Both involve tokenizer v5 behavior, but one is a broad behavior difference and the other is a specific decoder-type bug." }, { "left": "issue:28348", "right": "issue:45250", "accept": false, - "reason": "Both mention Flash Attention 2.0, but one is GPT-2 support and the other is only a generic title with no shared concrete code path." + "reason": "Both mention Flash Attention 2.0, but one is GPT2LMHeadModel-specific while the other is a generic feature request." }, { "left": "issue:44908", "right": "issue:44998", "accept": false, - "reason": "Scheduler kwargs and an unrelated placeholder issue have no duplicate signal." + "reason": "Scheduler kwargs handling and an unrelated placeholder issue are not the same bug." }, { "left": "issue:44492", "right": "issue:44998", "accept": false, - "reason": "A cache-strategy typo and a nonspecific placeholder issue are unrelated." + "reason": "A cache-strategy typo is unrelated to the placeholder report." }, { "left": "issue:43618", "right": "issue:44998", "accept": false, - "reason": "CLIPOutput attentions regression and a placeholder bug report are unrelated." + "reason": "CLIPOutput attentions regression and the placeholder report are unrelated." }, { "left": "issue:44998", "right": "issue:45850", "accept": false, - "reason": "A placeholder issue cannot be treated as the same bug as broad v5 breakage." + "reason": "The placeholder issue contains no concrete bug and is not the same as the broad v5 regression report." }, { "left": "issue:44938", @@ -17597,67 +17498,67 @@ "left": "issue:44861", "right": "issue:45290", "accept": false, - "reason": "A tied-weights AttributeError and a chat-template/tool-call crash are different code paths." + "reason": "Tied-weight key handling and apply_chat_template tool-call crashes are different code paths." }, { "left": "issue:43525", "right": "issue:44861", "accept": false, - "reason": "Missing pad_token_id in a config is unrelated to tied-weight key handling." + "reason": "Missing pad_token_id in a config and tied-weight key handling are unrelated." }, { "left": "issue:31515", "right": "issue:38617", "accept": false, - "reason": "Slow from_pretrained loading and a missing layer_type_validation import are unrelated." + "reason": "Slow from_pretrained checkpoint loading and a missing import symbol are different failures." }, { "left": "issue:43572", "right": "issue:44861", "accept": false, - "reason": "StableLmConfig pad_token_idx absence is not the same as tied-weight key handling." + "reason": "Missing pad_token_idx after v5 update and tied-weight key handling are unrelated." }, { "left": "issue:44230", "right": "issue:44623", "accept": false, - "reason": "fp8 support for qwen3-vl/qwen3.5 moe is unrelated to processor save_pretrained file loss." + "reason": "fp8 model support and processor.save_pretrained missing files are different issues." }, { "left": "issue:28348", "right": "issue:43352", "accept": false, - "reason": "Flash Attention 2.0 support requests for different models are related only at a broad feature level, not the same concrete bug/change." + "reason": "Both are Flash Attention 2.0 related, but they target different models and different levels of support." }, { "left": "issue:30333", "right": "issue:31515", "accept": false, - "reason": "MLFlow reporting on failed jobs and checkpoint load performance are unrelated." + "reason": "MLFlow reporting on failed training jobs and slow checkpoint loading are unrelated." }, { "left": "issue:31515", "right": "issue:36296", "accept": false, - "reason": "Slow checkpoint loading and tensor-parallel training bugs are different issues." + "reason": "Slow loading and tensor-parallel training bugs are different problems." }, { "left": "issue:44230", "right": "issue:45310", "accept": false, - "reason": "fp8 support and a from_pretrained regression for Qwen3.5 MoE are not the same underlying bug." + "reason": "fp8 inference support and a Qwen3.5 MoE from_pretrained regression are not the same bug." }, { "left": "issue:31515", "right": "issue:33453", "accept": false, - "reason": "Loading speed and tokenizer loading regression are different failures." + "reason": "Slow checkpoint loading and tokenizer loading regression are only loosely related and not the same concrete issue." }, { "left": "issue:43298", "right": "issue:44230", "accept": false, - "reason": "Continuous batching audio support and qwen3-vl fp8 inference support are unrelated." + "reason": "Audio continuous batching support and fp8 Qwen3-VL/Qwen3.5 MoE support are different feature requests." } ] }, @@ -17679,7 +17580,7 @@ "source": "llm", "cache_hit": false, "model": "gpt-5.4-mini?service_tier=flex", - "cluster_id": "cluster-10000-472", + "cluster_id": "cluster-10000-473", "nodes": [ "issue:37934", "issue:42936", @@ -17705,20 +17606,20 @@ "issue:37934|issue:43668", "issue:42936|issue:43668" ], - "prepared_review_unit_hash": "e498ad25c21508978c8e2acbc731a1258f735b3e6f477d47601380a175a01cf2", + "prepared_review_unit_hash": "8840a61ba2865c98265e7039292dc328c5e08137d248c3ae53764beaa7d50696", "component_budget": { - "node_count": 472, - "item_count": 472, - "soft_pair_count": 1732, - "serialized_chars": 898654, - "estimated_input_tokens": 224664, - "estimated_eval_tokens": 449584 + "node_count": 473, + "item_count": 473, + "soft_pair_count": 1738, + "serialized_chars": 901437, + "estimated_input_tokens": 225360, + "estimated_eval_tokens": 450976 }, "budget": { "node_count": 9, "item_count": 9, "soft_pair_count": 12, - "serialized_chars": 8611, + "serialized_chars": 8610, "estimated_input_tokens": 2153, "estimated_eval_tokens": 4562 }, @@ -17727,7 +17628,7 @@ "aggressively_trimmed": false, "split": true, "analyst_result": { - "summary": "This cluster is not a duplicate set: the issues span unrelated models, configs, tokenizer warnings, quantization typos, and attention/loss behavior. No soft pair looks like the same underlying bug or change.", + "summary": "These issues span unrelated models and bug classes (norms/config, tokenizers, loss plumbing, gated experts, attention implementation). None of the soft-similarity pairs look like the same underlying defect.", "confidence": 0.97, "canonical_issue_reason": null, "canonical_pr_reason": null, @@ -17738,73 +17639,73 @@ "left": "issue:43717", "right": "issue:46144", "accept": false, - "reason": "Different subsystems and failure modes: Mamba/Mamba-2 init weights vs RoFormer attention interface." + "reason": "Mamba/Mamba-2 init_weights behavior is unrelated to RoFormer attention not using the attention interface." }, { "left": "issue:42936", "right": "issue:46018", "accept": false, - "reason": "Unrelated model bugs: Mask2Former ignore_value usage vs DeepSeek-V4 shared expert gating." + "reason": "Mask2Former ignore_value handling and DeepSeek-V4 shared expert gating are different models and different bugs." }, { "left": "issue:43668", "right": "issue:44031", "accept": false, - "reason": "Config type hint typo vs tokenizer regex warning regression; no shared concrete code path." + "reason": "An incorrect type hint on ModernBERTConfig norm_eps is unrelated to tokenizer regex warning regressions." }, { "left": "issue:42936", "right": "issue:43240", "accept": false, - "reason": "Mask2Former ignore_value handling and loss kwargs forwarding are separate issues." + "reason": "Unused ignore_value in Mask2Former is not the same as kwargs not being forwarded to a loss function." }, { "left": "issue:43240", "right": "issue:46144", "accept": false, - "reason": "Loss-function kwargs propagation is unrelated to RoFormer attention implementation." + "reason": "Loss-function kwargs plumbing and RoFormer attention interface usage are distinct code paths." }, { "left": "issue:42936", "right": "issue:44031", "accept": false, - "reason": "Different components and symptoms: segmentation loss ignore value vs tokenizer warning behavior." + "reason": "Mask2Former ignore_value and tokenizer regex warnings have no common underlying bug." }, { "left": "issue:42936", "right": "issue:46144", "accept": false, - "reason": "No overlap in model or code path; these describe distinct bugs." + "reason": "These concern different architectures and different failure modes." }, { "left": "issue:42936", "right": "issue:43813", "accept": false, - "reason": "Mask2Former loss handling vs a quantization attribute typo; not the same change." + "reason": "Mask2Former ignore_value usage is unrelated to a typo in a quantization operation name." }, { "left": "issue:43240", "right": "issue:43668", "accept": false, - "reason": "Generic kwargs forwarding bug and ModernBERT type annotation issue are unrelated." + "reason": "Passing kwargs to loss calculation and fixing a config type hint are unrelated." }, { "left": "issue:43668", "right": "issue:43813", "accept": false, - "reason": "Both are small correctness issues, but in different files and unrelated code paths." + "reason": "ModernBERT norm_eps typing and quantization-operation naming are unrelated issues." }, { "left": "issue:37934", "right": "issue:43668", "accept": false, - "reason": "Llama4TextL2Norm/RMS norm question is unrelated to ModernBERT norm_eps typing." + "reason": "A question about whether Llama4TextL2Norm should be RMSNorm is not the same as a ModernBERT type-hint bug." }, { "left": "issue:42936", "right": "issue:43668", "accept": false, - "reason": "Mask2Former ignore_value bug and ModernBERT type hint bug are separate and not mergeable as one fix." + "reason": "Mask2Former ignore_value handling and ModernBERT norm_eps typing do not describe the same bug." } ] }, @@ -17837,52 +17738,52 @@ "issue:10519|issue:11005", "issue:11005|issue:11307" ], - "prepared_review_unit_hash": "33cfa142543e3238c6af8c0892fecf36ff53a3f817fc0e6044424ec57d6c6a9e", + "prepared_review_unit_hash": "c2f3895dc22c280d157a87d513f7db316e5ee48c1ced77cbabd6cb0ff06b4f59", "component_budget": { "node_count": 3, "item_count": 3, "soft_pair_count": 3, - "serialized_chars": 2877, - "estimated_input_tokens": 720, - "estimated_eval_tokens": 1696 + "serialized_chars": 2876, + "estimated_input_tokens": 719, + "estimated_eval_tokens": 1694 }, "budget": { "node_count": 3, "item_count": 3, "soft_pair_count": 3, - "serialized_chars": 2877, - "estimated_input_tokens": 720, - "estimated_eval_tokens": 1696 + "serialized_chars": 2876, + "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 not duplicates: they ask for unrelated features in different subsystems (tokenizer truncation behavior, ReduceLROnPlateau-style scheduling, and Wav2Vec2 word timestamps).", + "summary": "These three issues are not duplicates: one asks for tokenizer truncation-from-beginning behavior, one asks for ReduceLROnPlateau-like scheduling, and one asks for Wav2Vec2 word boundary offsets.", "confidence": 0.98, - "canonical_issue_reason": "No canonical issue: the items are unrelated feature requests and should not be clustered as duplicates.", + "canonical_issue_reason": "Issue 10519 is the most concrete and self-contained request, but the cluster does not actually contain a single underlying bug/feature shared by all three issues.", "canonical_pr_reason": null, - "best_issue_reason": "11307 is the most active issue, but it is still not a good cluster representative because it targets Wav2Vec2 timestamp extraction, not the other two feature requests.", + "best_issue_reason": "10519 is the best representative only in the sense that it is the clearest standalone feature request; however, it should not be treated as a true duplicate canonical for the other two issues.", "best_pr_reason": null, "soft_edge_verdicts": [ { "left": "issue:10519", "right": "issue:11307", "accept": false, - "reason": "Different requests in different areas: tokenizer truncation direction vs Wav2Vec2 word boundary timestamps." + "reason": "Different feature areas: tokenizer truncation behavior vs. Wav2Vec2 word-offset extraction. No shared bug or code path." }, { "left": "issue:10519", "right": "issue:11005", "accept": false, - "reason": "Tokenizer truncation feature request and ReduceLROnPlateau-like scheduler functionality are unrelated." + "reason": "Unrelated requests: text truncation strategy vs. learning-rate scheduler functionality. They do not describe the same underlying change." }, { "left": "issue:11005", "right": "issue:11307", "accept": false, - "reason": "Training-scheduler behavior and speech-model timestamp extraction do not describe the same underlying bug or change." + "reason": "Scheduler feature request and Wav2Vec2 timing-offset request are unrelated and would not be merged as one change." } ] }, @@ -17914,12 +17815,12 @@ "issue:13244|issue:16618", "issue:15980|issue:16618" ], - "prepared_review_unit_hash": "737d5d7d51c428566958f6333435517453b5081ff3742797bd63d5fcb6a19323", + "prepared_review_unit_hash": "de205c5bb6c18949f77f68d042f8e2799c5396769ca3d7db97f6b7d4079d0e17", "component_budget": { "node_count": 3, "item_count": 3, "soft_pair_count": 2, - "serialized_chars": 2477, + "serialized_chars": 2479, "estimated_input_tokens": 620, "estimated_eval_tokens": 1496 }, @@ -17927,7 +17828,7 @@ "node_count": 3, "item_count": 3, "soft_pair_count": 2, - "serialized_chars": 2477, + "serialized_chars": 2479, "estimated_input_tokens": 620, "estimated_eval_tokens": 1496 }, @@ -17936,24 +17837,24 @@ "aggressively_trimmed": false, "split": false, "analyst_result": { - "summary": "These issues are not duplicates: one is about Tapas tokenization parity, one about auth/error messaging for private model downloads, and one about failing to load bert-base-uncased. The only mild overlap is in model-loading context, but the underlying problems differ.", - "confidence": 0.82, - "canonical_issue_reason": "Issue #16618 is the best representative because it is the most concrete and broadly relevant model-loading failure; the others are separate, narrower problems.", + "summary": "These look like three separate issue reports: a Tapas tokenization discrepancy, a private-model login/error-message problem, and a generic model loading failure. The only superficially related pair is the two model-loading issues, but their symptoms and likely root causes do not match closely enough to treat as duplicates.", + "confidence": 0.96, + "canonical_issue_reason": "No single canonical issue: 13244 is about tokenization behavior, while 15980 and 16618 concern model download/loading, and they do not describe the same concrete bug.", "canonical_pr_reason": null, - "best_issue_reason": "#16618 is the strongest standalone issue to keep as the primary reference for this cluster since it describes an actual loading failure rather than a tokenization mismatch or a specific auth error message.", + "best_issue_reason": "No strong global representative; 16618 has the broadest title, but it is still not a clean stand-in for the other issues because it targets a different failure mode than 15980 and is unrelated to 13244.", "best_pr_reason": null, "soft_edge_verdicts": [ { "left": "issue:13244", "right": "issue:16618", "accept": false, - "reason": "Tapas tokenization differs from TensorFlow code; this is a tokenization/parity problem, not a model-loading failure for bert-base-uncased." + "reason": "Tokenization mismatch for Tapas vs. a generic model load failure are different code paths and different symptoms." }, { "left": "issue:15980", "right": "issue:16618", "accept": false, - "reason": "#15980 is about the error message when downloading a private model without login, while #16618 is about failing to load a public model; same area, different underlying bug." + "reason": "Both mention model loading, but 15980 is specifically an unauthenticated private-model download/error-message issue, while 16618 is a broader bert-base-uncased load failure." } ] }, @@ -17983,12 +17884,12 @@ "soft_pairs": [ "issue:26350|issue:33260" ], - "prepared_review_unit_hash": "78a950062208038791a11b514b1101d50f74caa827cc4e1e427842d73f261cb4", + "prepared_review_unit_hash": "88a26c04d2f184501ee74b46162da220d1ea524ac0bbc17c9723ca196e65b288", "component_budget": { "node_count": 2, "item_count": 2, "soft_pair_count": 1, - "serialized_chars": 1589, + "serialized_chars": 1590, "estimated_input_tokens": 398, "estimated_eval_tokens": 1052 }, @@ -17996,7 +17897,7 @@ "node_count": 2, "item_count": 2, "soft_pair_count": 1, - "serialized_chars": 1589, + "serialized_chars": 1590, "estimated_input_tokens": 398, "estimated_eval_tokens": 1052 }, @@ -18005,18 +17906,18 @@ "aggressively_trimmed": false, "split": false, "analyst_result": { - "summary": "These are unrelated feature-request issues: one asks for Flash Attention 2 support on more architectures, the other for GGUF support on more architectures. They share only broad \u201ccommunity contribution\u201d framing and are not the same underlying change.", + "summary": "These are two unrelated feature requests in the same broad area of model/runtime support, but they do not describe the same underlying change. One asks for Flash Attention 2 support across more architectures; the other asks for GGUF support across more architectures.", "confidence": 0.98, - "canonical_issue_reason": null, + "canonical_issue_reason": "No single canonical duplicate issue: the titles and descriptions point to different feature work and different code paths (Flash Attention 2 vs GGUF support).", "canonical_pr_reason": null, - "best_issue_reason": null, + "best_issue_reason": "If a representative issue must be chosen, issue #26350 is the better anchor because it has much higher discussion activity, but it is not a true canonical for #33260 since the requests are unrelated.", "best_pr_reason": null, "soft_edge_verdicts": [ { "left": "issue:26350", "right": "issue:33260", "accept": false, - "reason": "Different targets and code paths: Flash Attention 2 vs GGUF support. Same broad subsystem category, but not the same bug/change and not mergeable as one issue." + "reason": "Reject: these are distinct feature requests with different underlying changes and target different functionality; they are not duplicates." } ] }, @@ -18075,184 +17976,184 @@ "issue:44370|issue:46226", "issue:46222|issue:46226" ], - "prepared_review_unit_hash": "8aea67475bfbc0fb5705dd90465b46429e3e79e02c39e8abf395ab033064b2be", + "prepared_review_unit_hash": "51882bfe4676d7aa79680a550c151fe9ce2b226e5860f083fc83ea5446f18d7a", "component_budget": { "node_count": 8, "item_count": 8, "soft_pair_count": 24, - "serialized_chars": 13530, - "estimated_input_tokens": 3383, - "estimated_eval_tokens": 7022 + "serialized_chars": 13528, + "estimated_input_tokens": 3382, + "estimated_eval_tokens": 7020 }, "budget": { "node_count": 8, "item_count": 8, "soft_pair_count": 24, - "serialized_chars": 13530, - "estimated_input_tokens": 3383, - "estimated_eval_tokens": 7022 + "serialized_chars": 13528, + "estimated_input_tokens": 3382, + "estimated_eval_tokens": 7020 }, "overflow_policy": "truncate_then_skip", "trimmed": false, "aggressively_trimmed": false, "split": false, "analyst_result": { - "summary": "This cluster is mostly a set of i18n translation-request issues, but they are not all the same underlying request. The exact duplicates appear to be the placeholder-title subgroup (43834/44365/44370) and the identical Thai pair (46222/46226); the Turkish, Spanish, and Romanian requests are distinct and should not be merged together.", - "confidence": 0.58, - "canonical_issue_reason": "43834 is the earliest issue in the largest near-identical placeholder-title subgroup, so it is the best single representative within that duplicate slice.", + "summary": "This cluster is mostly heterogeneous: several distinct translation-request issues for different languages, plus two small true-duplicate subclusters (the generic template issues and the identical Thai-titled issues).", + "confidence": 0.9, + "canonical_issue_reason": "Issue 43834 is the clearest representative of the repeated generic template subcluster ('[i18n-] Translating docs to '); the language-specific translation requests are separate items, not duplicates of it.", "canonical_pr_reason": null, - "best_issue_reason": "43834 is the best overall issue representative for the repeated placeholder-template translation issues, but the broader cluster is heterogeneous by language.", + "best_issue_reason": "43834 is the best issue representative because it is the most generic and centrally duplicated template form, whereas the concrete language requests are distinct translation tasks.", "best_pr_reason": null, "soft_edge_verdicts": [ { "left": "issue:27088", "right": "issue:44370", "accept": false, - "reason": "Turkish vs. generic placeholder translation request; different underlying request." + "reason": "Same i18n theme, but Turkish vs. a generic template issue; different underlying request." }, { "left": "issue:27088", "right": "issue:44365", "accept": false, - "reason": "Turkish vs. generic placeholder translation request; different underlying request." + "reason": "Turkish request and generic template request are not the same issue." }, { "left": "issue:27088", "right": "issue:43834", "accept": false, - "reason": "Turkish request is distinct from the generic placeholder translation issue." + "reason": "Concrete Turkish translation request is distinct from the generic placeholder template issue." }, { "left": "issue:27088", "right": "issue:38435", "accept": false, - "reason": "Different target languages (Turkish vs Romanian), so not the same issue." + "reason": "Different target languages (Turkish vs Romanian); separate translation requests." }, { "left": "issue:27088", "right": "issue:28936", "accept": false, - "reason": "Different target languages (Turkish vs Spanish), so not the same issue." + "reason": "Different target languages (Turkish vs Spanish); not the same underlying issue." }, { "left": "issue:28936", "right": "issue:43834", "accept": false, - "reason": "Spanish request is distinct from the generic placeholder translation issue." + "reason": "Spanish request does not duplicate the generic template issue." }, { "left": "issue:28936", "right": "issue:44365", "accept": false, - "reason": "Spanish request is distinct from the generic placeholder translation issue." + "reason": "Different language-specific request versus generic template; not a duplicate." }, { "left": "issue:28936", "right": "issue:38435", "accept": false, - "reason": "Different target languages (Spanish vs Romanian), so not the same issue." + "reason": "Spanish vs Romanian translation requests are separate." }, { "left": "issue:28936", "right": "issue:44370", "accept": false, - "reason": "Spanish request is distinct from the generic placeholder translation issue." + "reason": "Spanish request and generic template issue are not the same underlying task." }, { "left": "issue:38435", "right": "issue:44370", "accept": false, - "reason": "Romanian request is distinct from the generic placeholder translation issue." + "reason": "Romanian request is distinct from the generic template issue." }, { "left": "issue:38435", "right": "issue:44365", "accept": false, - "reason": "Romanian request is distinct from the generic placeholder translation issue." + "reason": "Romanian translation request is not the same as the generic template issue." }, { "left": "issue:38435", "right": "issue:43834", "accept": false, - "reason": "Romanian request is distinct from the generic placeholder translation issue." + "reason": "Concrete Romanian request and generic template request are different issues." }, { "left": "issue:38435", "right": "issue:46222", "accept": false, - "reason": "Romanian request and Thai request are different underlying issues." + "reason": "Different content and likely different request; not the same underlying issue." }, { "left": "issue:38435", "right": "issue:46226", "accept": false, - "reason": "Romanian request and Thai request are different underlying issues." + "reason": "Romanian translation request is unrelated to the Thai-titled issue." }, { "left": "issue:43834", "right": "issue:44370", "accept": true, - "reason": "Same placeholder-title translation request and near-exact body match; likely duplicate." + "reason": "Both are the same generic placeholder template issue with matching wording and no language-specific distinction." }, { "left": "issue:43834", "right": "issue:44365", "accept": true, - "reason": "Same placeholder-title translation request and near-exact body match; likely duplicate." + "reason": "Same generic template title/body pattern; likely duplicate clones of the template issue." }, { "left": "issue:43834", "right": "issue:46222", "accept": false, - "reason": "Different language/content group; not the same underlying issue." + "reason": "Generic translation template and the Thai-titled issue are not the same request." }, { "left": "issue:43834", "right": "issue:46226", "accept": false, - "reason": "Different language/content group; not the same underlying issue." + "reason": "Generic template issue does not match the Thai-titled request." }, { "left": "issue:44365", "right": "issue:44370", "accept": true, - "reason": "Same placeholder-title translation request and near-exact body match; likely duplicate." + "reason": "Identical generic placeholder translation issue; same underlying item." }, { "left": "issue:44365", "right": "issue:46222", "accept": false, - "reason": "Different language/content group; not the same underlying issue." + "reason": "Generic template issue and Thai-titled issue are different requests." }, { "left": "issue:44365", "right": "issue:46226", "accept": false, - "reason": "Different language/content group; not the same underlying issue." + "reason": "Generic template issue and Thai-titled issue are different requests." }, { "left": "issue:44370", "right": "issue:46222", "accept": false, - "reason": "Different language/content group; not the same underlying issue." + "reason": "Generic template issue and Thai-titled issue are not duplicates." }, { "left": "issue:44370", "right": "issue:46226", "accept": false, - "reason": "Different language/content group; not the same underlying issue." + "reason": "Generic template issue and Thai-titled issue are not the same underlying request." }, { "left": "issue:46222", "right": "issue:46226", "accept": true, - "reason": "Exact same Thai title and effectively identical content; likely duplicate issue." + "reason": "These are identical Thai-titled issues with matching structure, so they look like true duplicates." } ] }, "evaluator_result": { "accept": true, - "feedback": "The summary is grounded in the packet: the Turkish, Spanish, Romanian, placeholder-title, and Thai items are clearly different language/request groups based on titles and timestamps, while the placeholder-title trio and Thai pair have strong near-identical/identical similarity signals. The soft-edge verdicts are conservative overall, and the accepted duplicate pairs are supported by the high similarity evidence." + "feedback": "The summary stays within the packet evidence: it distinguishes the language-specific translation requests, groups the identical generic template issues together, and treats the Thai-titled pair as a duplicate subcluster. The soft-edge verdicts are conservative and do not overclaim beyond the visible titles/body similarities." }, "evaluator_used": true, "retried": false, @@ -18334,12 +18235,12 @@ "issue:42503|issue:46653", "issue:45027|issue:46653" ], - "prepared_review_unit_hash": "af016ceaf7e12eb91cff593927ba9c07d2454e90aa5d99256e0e3c6c0d2c9ce5", + "prepared_review_unit_hash": "ab7f4abb56cb759d383c6bf4ccfc89e9440223d20b5054dcb4995e48eebba9f0", "component_budget": { "node_count": 18, "item_count": 18, "soft_pair_count": 40, - "serialized_chars": 24145, + "serialized_chars": 24146, "estimated_input_tokens": 6037, "estimated_eval_tokens": 12330 }, @@ -18347,7 +18248,7 @@ "node_count": 18, "item_count": 18, "soft_pair_count": 40, - "serialized_chars": 24145, + "serialized_chars": 24146, "estimated_input_tokens": 6037, "estimated_eval_tokens": 12330 }, @@ -18356,252 +18257,252 @@ "aggressively_trimmed": false, "split": false, "analyst_result": { - "summary": "This is a loose similarity cluster of distinct model-request issues (different model families and capabilities), not a duplicate set around one underlying bug or change.", - "confidence": 0.94, - "canonical_issue_reason": null, + "summary": "This cluster is a loose family of model-addition/support requests, but the artifacts are not true duplicates. They mostly share the same broad 'add support for model X' pattern rather than the same underlying bug or code change.", + "confidence": 0.87, + "canonical_issue_reason": "Issue 43837 is one of the most central and connected requests in the set, so it is the best single anchor if a representative issue is needed; however, it is still only a representative, not a duplicate of the others.", "canonical_pr_reason": null, - "best_issue_reason": null, + "best_issue_reason": "43837 is a good cluster representative because it sits near the center of the similarity graph and is a concrete, specific support request rather than a generic placeholder.", "best_pr_reason": null, "soft_edge_verdicts": [ { "left": "issue:36817", "right": "issue:41211", "accept": false, - "reason": "Different requested models (EuroBert vs DEIMv2); similar wording only." + "reason": "Different models and different requested integrations; only the generic model-addition theme overlaps." }, { "left": "issue:36817", "right": "issue:43671", "accept": false, - "reason": "EuroBert config request vs Qwen3-TTS support; unrelated model families and tasks." + "reason": "EuroBert config support vs Qwen3 TTS support are unrelated changes." }, { "left": "issue:36817", "right": "issue:36879", "accept": false, - "reason": "Different model requests (EuroBert vs RF-DETR) with no shared code-path evidence." + "reason": "Different model families entirely; not the same bug or change." }, { "left": "issue:36879", "right": "issue:41211", "accept": false, - "reason": "RF-DETR and DEIMv2 are different model additions, not the same change." + "reason": "RF-DETR and DEIMv2 are distinct model requests with different code paths." }, { "left": "issue:36879", "right": "issue:43909", "accept": false, - "reason": "RF-DETR vs LFM2.5 Audio; different architecture/domain." + "reason": "Different model families and modalities; no common concrete fix." }, { "left": "issue:36879", "right": "issue:44121", "accept": false, - "reason": "Object detection model request vs sparse transformer request; distinct implementations." + "reason": "RF-DETR support and weight-sparse transformer support are unrelated." }, { "left": "issue:36879", "right": "issue:43837", "accept": false, - "reason": "RF-DETR vs Qwen3-ASR support; different model/task." + "reason": "Different models and different feature requests; same broad 'add support' pattern only." }, { "left": "issue:36879", "right": "issue:43671", "accept": false, - "reason": "RF-DETR and Qwen3-TTS are unrelated model integrations." + "reason": "RF-DETR vs Qwen3 TTS are unrelated additions." }, { "left": "issue:37015", "right": "issue:43837", "accept": false, - "reason": "NeoBERT request is not the same as Qwen3-ASR support." + "reason": "NeoBERT and Qwen3 ASR are distinct model integrations." }, { "left": "issue:37015", "right": "issue:44121", "accept": false, - "reason": "NeoBERT vs sparse transformer; different model types and code paths." + "reason": "Different model architectures and different implementation work." }, { "left": "issue:37015", "right": "issue:43909", "accept": false, - "reason": "NeoBERT and LFM2.5 Audio are unrelated requests." + "reason": "NeoBERT vs LFM2.5 Audio are unrelated requests." }, { "left": "issue:36879", "right": "issue:37015", "accept": false, - "reason": "RF-DETR and NeoBERT are different model requests." + "reason": "RF-DETR and NeoBERT are different model support tasks." }, { "left": "issue:37015", "right": "issue:42503", "accept": false, - "reason": "NeoBERT vs ModernVBERT; similar naming only, not the same model." + "reason": "NeoBERT and ModernVBERT are separate model requests, not one change." }, { "left": "issue:41115", "right": "issue:43837", "accept": false, - "reason": "MiniCPM3 architecture request vs Qwen3-ASR support; different tasks." + "reason": "MiniCPM3 support and Qwen3 ASR support are different model integrations." }, { "left": "issue:41115", "right": "issue:44121", "accept": false, - "reason": "MiniCPM3 and the sparse transformer request are unrelated models." + "reason": "MiniCPM3 and the weight-sparse transformer request are unrelated code changes." }, { "left": "issue:41115", "right": "issue:42503", "accept": false, - "reason": "MiniCPM3 vs ModernVBERT; distinct model additions." + "reason": "MiniCPM3 and ModernVBERT are distinct model additions." }, { "left": "issue:41115", "right": "issue:45027", "accept": false, - "reason": "MiniCPM3 architecture and Voxtral TTS support are different features." + "reason": "Different models and different target behavior; not mergeable as one fix." }, { "left": "issue:41211", "right": "issue:43671", "accept": false, - "reason": "DEIMv2 and Qwen3-TTS are different model families/tasks." + "reason": "DEIMv2 and Qwen3 TTS are different model requests with no shared concrete bug." }, { "left": "issue:42503", "right": "issue:43837", "accept": false, - "reason": "ModernVBERT and Qwen3-ASR are different requests." + "reason": "ModernVBERT models and Qwen3 ASR are unrelated support requests." }, { "left": "issue:42503", "right": "issue:43671", "accept": false, - "reason": "ModernVBERT vs Qwen3-TTS; no same underlying change." + "reason": "ModernVBERT and Qwen3 TTS are different model additions." }, { "left": "issue:42503", "right": "issue:42738", "accept": false, - "reason": "ModernVBERT and BERT-like models with RoPE overlap only broadly on BERT wording." + "reason": "ModernVBERT models vs a generic BERT-like RoPE request are not the same underlying change." }, { "left": "issue:42738", "right": "issue:45027", "accept": false, - "reason": "RoPE BERT-like models vs Voxtral TTS support; unrelated." + "reason": "A generic RoPE-related BERT-like request and Voxtral TTS support are different features." }, { "left": "issue:42738", "right": "issue:46653", "accept": false, - "reason": "RoPE/BERT architecture request is unrelated to the Mirko Privitera issue." + "reason": "Model architecture request vs unrelated personal-name issue; no overlap." }, { "left": "issue:43671", "right": "issue:43837", "accept": false, - "reason": "Qwen3-TTS vs Qwen3-ASR are related modalities but different concrete model/support requests." + "reason": "Same Qwen3 family, but TTS and ASR are different modalities and implementation paths." }, { "left": "issue:43671", "right": "issue:45027", "accept": false, - "reason": "TTS support requests for different models are not the same change." + "reason": "Qwen3 TTS and Voxtral TTS are different model support requests, not one fix." }, { "left": "issue:43671", "right": "issue:45177", "accept": false, - "reason": "Qwen3-TTS and DeepSeek-OCR2 are different model integrations." + "reason": "Audio TTS support vs DeepSeek-OCR2 support are unrelated." }, { "left": "issue:43671", "right": "issue:44121", "accept": false, - "reason": "Audio model support vs sparse transformer request; not the same bug/change." + "reason": "Different model families and different requested functionality." }, { "left": "issue:43671", "right": "issue:43909", "accept": false, - "reason": "Both audio-related, but Qwen3-TTS and LFM2.5 Audio are different models and implementations." + "reason": "Qwen3 TTS and LFM2.5 Audio are distinct audio-model requests." }, { "left": "issue:43837", "right": "issue:43909", "accept": false, - "reason": "ASR support and audio model add are different concrete requests." + "reason": "Qwen3 ASR and LFM2.5 Audio are different model integrations." }, { "left": "issue:43837", "right": "issue:45322", "accept": false, - "reason": "Qwen3-ASR and EUPE are different model categories and code paths." + "reason": "ASR support and EUPE support are unrelated model requests." }, { "left": "issue:43837", "right": "issue:44121", "accept": false, - "reason": "ASR support vs sparse transformer model request; unrelated." + "reason": "Qwen3 ASR and weight-sparse transformer support are different changes." }, { "left": "issue:43909", "right": "issue:44205", "accept": false, - "reason": "LFM2.5 Audio and SAM3-LiteText are different model integrations." + "reason": "Audio model support and SAM3-LiteText support do not share the same underlying fix." }, { "left": "issue:43909", "right": "issue:46645", "accept": false, - "reason": "Audio model request vs Granite SWA+sink option; different feature areas." + "reason": "LFM2.5 Audio support and Granite SWA+sink option are different code paths." }, { "left": "issue:44121", "right": "issue:45322", "accept": false, - "reason": "Sparse transformer request and EUPE request are different model changes." + "reason": "Weight-sparse transformer support and EUPE support are unrelated model requests." }, { "left": "issue:43837", "right": "issue:45027", "accept": false, - "reason": "Qwen3-ASR support and Voxtral TTS support are different modalities/models." + "reason": "Qwen3 ASR and Voxtral TTS are separate audio-model support requests." }, { "left": "issue:37015", "right": "issue:45027", "accept": false, - "reason": "NeoBERT and Voxtral TTS are unrelated requests." + "reason": "NeoBERT support and Voxtral TTS support are unrelated model additions." }, { "left": "issue:43837", "right": "issue:45177", "accept": false, - "reason": "Qwen3-ASR and DeepSeek-OCR2 are different models and tasks." + "reason": "Qwen3 ASR and DeepSeek-OCR2 are different model families and different code paths." }, { "left": "issue:45322", "right": "issue:46133", "accept": false, - "reason": "EUPE and TIPSv2 are both vision-language related, but they are distinct model requests." + "reason": "EUPE and TIPSv2 are different model requests, even though both are vision-language related." }, { "left": "issue:42503", "right": "issue:46653", "accept": false, - "reason": "ModernVBERT request is unrelated to the closed Mirko Privitera issue." + "reason": "ModernVBERT models and the unrelated name issue do not match any shared bug or change." }, { "left": "issue:45027", "right": "issue:46653", "accept": false, - "reason": "Voxtral TTS support is unrelated to the Mirko Privitera issue." + "reason": "Voxtral TTS support and the unrelated name issue are not duplicates." } ] }, @@ -18631,12 +18532,12 @@ "soft_pairs": [ "issue:38740|issue:46443" ], - "prepared_review_unit_hash": "7019f71682825cb6eb503eb50e6f3ffa6ca95c65e5793d32baabe527722be549", + "prepared_review_unit_hash": "5f76c80171f44e59f657b934e611e41bf2184a087e0037235417077562cfeea3", "component_budget": { "node_count": 2, "item_count": 2, "soft_pair_count": 1, - "serialized_chars": 1505, + "serialized_chars": 1506, "estimated_input_tokens": 377, "estimated_eval_tokens": 1010 }, @@ -18644,7 +18545,7 @@ "node_count": 2, "item_count": 2, "soft_pair_count": 1, - "serialized_chars": 1505, + "serialized_chars": 1506, "estimated_input_tokens": 377, "estimated_eval_tokens": 1010 }, @@ -18653,8 +18554,8 @@ "aggressively_trimmed": false, "split": false, "analyst_result": { - "summary": "These issues are not duplicates: one is a docs request to add Pruna as an optimization framework, while the other is a Vision RoPE refactor. They concern different domains and no shared issue targets or files are present.", - "confidence": 0.98, + "summary": "The two issues are unrelated: one is a docs request to add Pruna as an optimization framework, and the other is a Vision RoPE refactor. They do not describe the same bug or change.", + "confidence": 0.99, "canonical_issue_reason": null, "canonical_pr_reason": null, "best_issue_reason": null, @@ -18664,7 +18565,7 @@ "left": "issue:38740", "right": "issue:46443", "accept": false, - "reason": "Different subjects and scopes: documentation/framework listing vs a vision model positional-encoding refactor. No evidence they describe the same underlying bug or change." + "reason": "Different topics and scope: docs/framework listing versus a Vision RoPE code refactor. No shared target, filename, or underlying problem." } ] }, @@ -18694,12 +18595,12 @@ "soft_pairs": [ "issue:41084|issue:43037" ], - "prepared_review_unit_hash": "db31efadcc4c7b40d914e10293e4fc7a7bdf2128e3a740472da0d35e21e4cfab", + "prepared_review_unit_hash": "140262fdef78d1e5db59209b6debbeb0823075d65aa42ccf513eddf0f92141c4", "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 }, @@ -18707,7 +18608,7 @@ "node_count": 2, "item_count": 2, "soft_pair_count": 1, - "serialized_chars": 1480, + "serialized_chars": 1479, "estimated_input_tokens": 370, "estimated_eval_tokens": 996 }, @@ -18716,18 +18617,18 @@ "aggressively_trimmed": false, "split": false, "analyst_result": { - "summary": "The two issues are not duplicates: one is a generic feature/behavior request about block decoding, while the other is a request for DeepSeek v3.2 model support. They describe different underlying work and would not be merged into a single fix.", + "summary": "The two issues are not duplicates: one is about Set Block Decoding, the other requests DeepSeek v3.2 support. They describe different problem areas with no shared targets or filenames.", "confidence": 0.97, - "canonical_issue_reason": "No canonical issue can be established from these two alone; they address different problems and neither subsumes the other.", + "canonical_issue_reason": "No canonical issue for deduping here; the issues are unrelated and should remain separate.", "canonical_pr_reason": null, - "best_issue_reason": "Issue 43037 is the more concrete and actionable request because it names a specific model/version to support, whereas 41084 is broader and more ambiguous.", + "best_issue_reason": "If a single representative is needed, issue 43037 is the clearer product request, but it is not the same underlying issue as 41084.", "best_pr_reason": null, "soft_edge_verdicts": [ { "left": "issue:41084", "right": "issue:43037", "accept": false, - "reason": "Different underlying requests: block decoding behavior vs. DeepSeek v3.2 support. They do not look like the same bug or change." + "reason": "Different topics: decoding behavior vs model compatibility/support request; no evidence they describe the same bug or change." } ] }, @@ -18757,12 +18658,12 @@ "soft_pairs": [ "issue:43598|issue:43600" ], - "prepared_review_unit_hash": "8d86a3cac52373a9e9cf93e4eb6cc170de4a58bd61ec19fd1f0c8f3f87bab1ce", + "prepared_review_unit_hash": "2824152c316bd9502a8162fedf90f47900dbe18ee8d3fa7c30204c61ce6f599b", "component_budget": { "node_count": 2, "item_count": 2, "soft_pair_count": 1, - "serialized_chars": 1583, + "serialized_chars": 1584, "estimated_input_tokens": 396, "estimated_eval_tokens": 1048 }, @@ -18770,7 +18671,7 @@ "node_count": 2, "item_count": 2, "soft_pair_count": 1, - "serialized_chars": 1583, + "serialized_chars": 1584, "estimated_input_tokens": 396, "estimated_eval_tokens": 1048 }, @@ -18779,18 +18680,18 @@ "aggressively_trimmed": false, "split": false, "analyst_result": { - "summary": "These two closed issues are related by theme (custom trainers / Trainer internals) but describe different problems: one is about `remove_unused_column` behavior and customizability, the other about FLOPS estimation accounting for custom trainers. I would not treat them as duplicates.", - "confidence": 0.86, - "canonical_issue_reason": "Issue #43598 is the cleaner representative because it has a concrete Trainer customization concern and a more specific scope.", + "summary": "The two issues are only loosely related by the Trainer/custom trainer theme, but they describe different problems: one is about `remove_unused_column` customization, the other about FLOPS estimation for custom trainers. They should not be merged as duplicates.", + "confidence": 0.88, + "canonical_issue_reason": "Issue #43598 is the broader, more representative Trainer customizability request, while #43600 is a separate FLOPS-estimation concern.", "canonical_pr_reason": null, - "best_issue_reason": "#43598 is the best issue anchor for this cluster since it is the more concrete and actionable Trainer customization request.", + "best_issue_reason": "#43598 is the better cluster representative because it centers on Trainer behavior customization rather than a narrower metrics-estimation detail.", "best_pr_reason": null, "soft_edge_verdicts": [ { "left": "issue:43598", "right": "issue:43600", "accept": false, - "reason": "They share a general Trainer/custom-trainer theme, but the underlying problems are different: unused-column removal vs FLOPS estimation logic. Not the same bug or change." + "reason": "They share the custom-trainer context, but the underlying problems are different: unused-column handling vs FLOPS estimation. They would not be fixed by the same change." } ] }, @@ -18820,7 +18721,7 @@ "soft_pairs": [ "issue:43684|issue:44405" ], - "prepared_review_unit_hash": "bf052ec75dd592c67e94e7fd69280c4f497110b90a556fabf01a38021ed16182", + "prepared_review_unit_hash": "6a8c1bd73ce3609c2d8ba9683efa94068f155bb1f43113d62c63888f7ab9d455", "component_budget": { "node_count": 2, "item_count": 2, @@ -18842,18 +18743,18 @@ "aggressively_trimmed": false, "split": false, "analyst_result": { - "summary": "These are separate feature requests for different model families and different auto classes, so they should not be clustered as duplicates.", - "confidence": 0.95, - "canonical_issue_reason": "No canonical issue across the pair: #43684 asks for Qwen3-Omni registration in AutoModel/AutoModelForConditionalGeneration, while #44405 asks for AutoModelForSequenceClassification support for Qwen3.5 (Qwen3_5Config).", + "summary": "These two issues are related only at a very high level (both about adding Qwen model support), but they target different model families and different APIs/tasks, so they should not be treated as duplicates.", + "confidence": 0.97, + "canonical_issue_reason": "No single canonical issue: #43684 is about registering Qwen3-Omni with AutoModel/AutoModelForConditionalGeneration, while #44405 requests AutoModelForSequenceClassification support for Qwen3.5 (Qwen3_5Config). Different models and different heads.", "canonical_pr_reason": null, - "best_issue_reason": "Neither issue is a good canonical representative for the other; they target different code paths and different capabilities.", + "best_issue_reason": "#43684 is the broader infrastructure-style request, but it is not a duplicate of #44405; the two issues should remain separate.", "best_pr_reason": null, "soft_edge_verdicts": [ { "left": "issue:43684", "right": "issue:44405", "accept": false, - "reason": "Different requests: model registration for Qwen3-Omni vs sequence-classification support for Qwen3.5. They are not the same underlying bug or change." + "reason": "Different underlying change: model registration for Qwen3-Omni vs sequence-classification support for Qwen3.5. They do not look mergeable into one fix." } ] }, @@ -18890,76 +18791,76 @@ "issue:43845|issue:43859", "issue:43846|issue:43859" ], - "prepared_review_unit_hash": "db6a32418a5427da333863528568ab3d95403ee082792502ce209ee13bd46b2a", + "prepared_review_unit_hash": "c7076139e2be8500a39b0ac5d4eb87a7a43b06f0ddc9f3c2ccd7b4f95df5a8da", "component_budget": { "node_count": 4, "item_count": 4, "soft_pair_count": 6, - "serialized_chars": 4267, - "estimated_input_tokens": 1067, - "estimated_eval_tokens": 2390 + "serialized_chars": 4263, + "estimated_input_tokens": 1066, + "estimated_eval_tokens": 2388 }, "budget": { "node_count": 4, "item_count": 4, "soft_pair_count": 6, - "serialized_chars": 4267, - "estimated_input_tokens": 1067, - "estimated_eval_tokens": 2390 + "serialized_chars": 4263, + "estimated_input_tokens": 1066, + "estimated_eval_tokens": 2388 }, "overflow_policy": "truncate_then_skip", "trimmed": false, "aggressively_trimmed": false, "split": false, "analyst_result": { - "summary": "Three issues titled \"huggingface\" look like the same duplicate report cluster; the separate \"Fett\" issue does not appear to be the same problem.", - "confidence": 0.91, - "canonical_issue_reason": "issue:43845 is the best canonical issue because it is the earliest of the matching \"huggingface\" reports and there is no evidence that 43846 or 43859 describes a different underlying problem.", + "summary": "Three issues share the same near-empty title \"huggingface\" and are likely duplicates; the \"Fett\" issue appears unrelated and should stay separate.", + "confidence": 0.95, + "canonical_issue_reason": "Issue 43845 is the earliest of the three nearly identical \"huggingface\" reports, so it is the best canonical issue among that duplicate set.", "canonical_pr_reason": null, - "best_issue_reason": "issue:43845 is the best cluster representative: same title as the other matching issues, earliest timestamp, and no conflicting evidence.", + "best_issue_reason": "Among all issues in the cluster, 43845 is the strongest representative of the duplicate huggingface reports; 43835 does not match the same topic.", "best_pr_reason": null, "soft_edge_verdicts": [ { "left": "issue:43835", "right": "issue:43845", "accept": false, - "reason": "Different titles and no concrete evidence of the same underlying bug or request; \"Fett\" does not match the repeated \"huggingface\" reports." + "reason": "The titles do not indicate the same underlying problem; \"Fett\" is unrelated to the repeated \"huggingface\" reports." }, { "left": "issue:43835", "right": "issue:43846", "accept": false, - "reason": "Different titles and no concrete evidence of the same underlying bug or request; \"Fett\" does not match the repeated \"huggingface\" reports." + "reason": "No evidence that \"Fett\" refers to the same bug or request as the repeated \"huggingface\" issues." }, { "left": "issue:43835", "right": "issue:43859", "accept": false, - "reason": "Different titles and no concrete evidence of the same underlying bug or request; \"Fett\" does not match the repeated \"huggingface\" reports." + "reason": "Different title/topic; this does not look like the same underlying issue." }, { "left": "issue:43845", "right": "issue:43846", "accept": true, - "reason": "Same title, very close creation times, and no distinguishing metadata; these look like duplicate reports of the same issue." + "reason": "Same title, same vague topic, and close creation times strongly suggest duplicate reports of the same issue." }, { "left": "issue:43845", "right": "issue:43859", "accept": true, - "reason": "Same title and no evidence of a different code path or request; this appears to be another duplicate of the same issue." + "reason": "These are indistinguishable from the available data: same title and no differing details." }, { "left": "issue:43846", "right": "issue:43859", "accept": true, - "reason": "Same title and no distinguishing evidence; this appears to be the same underlying issue reported again." + "reason": "Same short title and no contradicting evidence; likely the same underlying report." } ] }, "evaluator_result": { "accept": true, - "feedback": "The summary is grounded in the packet and uses appropriately cautious language. The false links to issue:43835 are supported by the title mismatch, and the same-title huggingface issues are reasonably clustered with the available soft-similarity evidence. No obvious overstatement stands out." + "feedback": "The assessment is grounded in the packet: 43835 has a different title/topic, while 43845/43846/43859 all share the same title and high soft-similarity. The canonical choice of the earliest huggingface issue is conservative enough. No soft-edge verdict appears to overstate evidence materially." }, "evaluator_used": true, "retried": false, @@ -18997,98 +18898,95 @@ "issue:44572|issue:44937", "issue:44573|issue:44937" ], - "prepared_review_unit_hash": "f7fce232707d379657911afaf1f4d4f776ef3e9897a826c9de13a99d451c194e", + "prepared_review_unit_hash": "2d97391a820649c875fb00aca473091a743ede9975242c1b062d7a244ee9af3a", "component_budget": { "node_count": 5, "item_count": 5, "soft_pair_count": 9, - "serialized_chars": 5791, - "estimated_input_tokens": 1448, - "estimated_eval_tokens": 3152 + "serialized_chars": 5785, + "estimated_input_tokens": 1447, + "estimated_eval_tokens": 3150 }, "budget": { "node_count": 5, "item_count": 5, "soft_pair_count": 9, - "serialized_chars": 5791, - "estimated_input_tokens": 1448, - "estimated_eval_tokens": 3152 + "serialized_chars": 5785, + "estimated_input_tokens": 1447, + "estimated_eval_tokens": 3150 }, "overflow_policy": "truncate_then_skip", "trimmed": false, "aggressively_trimmed": false, "split": false, "analyst_result": { - "summary": "This cluster is mostly low-signal spam-like issues. Three issues share the exact same '' title and look like duplicates of each other; the remaining two are different titles and should stay separate.", - "confidence": 0.91, - "canonical_issue_reason": "issue:44371 is the earliest of the exact '' duplicates, so it is the best representative for that mini-cluster.", + "summary": "All five items are low-information closed issues, but their titles point to different things: generic spam placeholders, a documentation suggestion, and a \u201cGoogle LLC\u201d promo-style post. There is no convincing duplicate cluster here.", + "confidence": 0.95, + "canonical_issue_reason": "No strong canonical issue exists because the titles do not describe the same underlying bug or request; if a placeholder must be chosen, issue 44371 is the most generic and earliest spam-like entry.", "canonical_pr_reason": null, - "best_issue_reason": "issue:44371 is the cleanest global representative because it matches the other exact spam duplicates and is the earliest of them.", + "best_issue_reason": "None of these issues is a good global representative for a duplicate cluster, since they are not clearly about the same problem and lack meaningful bodies or targets.", "best_pr_reason": null, "soft_edge_verdicts": [ { "left": "issue:44371", "right": "issue:44572", - "accept": true, - "reason": "Same exact '' title and no distinguishing body/evidence; these look like the same spam issue." + "accept": false, + "reason": "Both are spam-like placeholders, but there is no evidence they refer to the same underlying issue or change; titles alone are identical generic noise." }, { "left": "issue:44371", "right": "issue:44573", - "accept": true, - "reason": "Same exact '' title and no distinguishing body/evidence; these look like the same spam issue." + "accept": false, + "reason": "Same as above: identical spam placeholder titles are not enough to conclude duplicate intent or the same concrete topic." }, { "left": "issue:44371", "right": "issue:44418", "accept": false, - "reason": "Titles differ materially; 'Documentation Enhancement Suggestion' is not the same as an empty spam placeholder." + "reason": "One is a generic spam placeholder and the other is a documentation enhancement suggestion; these are different issue types and topics." }, { "left": "issue:44371", "right": "issue:44937", "accept": false, - "reason": "Different title and apparent topic; 'Check out \"Google LLC\"' is not the same as ''." + "reason": "Generic spam placeholder versus a specific \u201cGoogle LLC\u201d promotional title; no shared underlying bug or request is evident." }, { "left": "issue:44418", "right": "issue:44572", "accept": false, - "reason": "Documentation enhancement request and spam placeholder are unrelated." + "reason": "Documentation enhancement suggestion does not match a generic spam placeholder; no shared concrete issue is visible." }, { "left": "issue:44418", "right": "issue:44573", "accept": false, - "reason": "Documentation enhancement request and spam placeholder are unrelated." + "reason": "Documentation enhancement suggestion and spam placeholder are unrelated, with no evidence of the same bug or change." }, { "left": "issue:44572", "right": "issue:44573", - "accept": true, - "reason": "Exact same '' title with no other evidence suggesting a difference." + "accept": false, + "reason": "Both are spam placeholders, but that is too generic to establish they are the same underlying issue; no concrete duplicate evidence." }, { "left": "issue:44572", "right": "issue:44937", "accept": false, - "reason": "Different titles and no shared concrete issue content." + "reason": "Spam placeholder and \u201cCheck out \u2018Google LLC\u2019\u201d look like different low-quality posts, not the same issue." }, { "left": "issue:44573", "right": "issue:44937", "accept": false, - "reason": "Different titles and no shared concrete issue content." + "reason": "Spam placeholder versus promotional-sounding post; no indication they describe the same underlying problem." } ] }, - "evaluator_result": { - "accept": true, - "feedback": "The summary is grounded in the packet: three issues have the exact same \"\" title, while the other two have clearly different titles. The soft-edge accepts for the exact-title spam issues are conservative and supported by the sparse evidence; the rejections for the non-matching titles are also conservative. No overstatement detected." - }, - "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 }, @@ -19111,7 +19009,7 @@ "soft_pairs": [ "issue:45145|issue:45151" ], - "prepared_review_unit_hash": "4dfec64eb288facafdcf14142e94e365313a8ae891b0773bf8f8feb1840554eb", + "prepared_review_unit_hash": "569aecf18080f78e9e46867fd7d8efa3d744d3e8b253c476e492aed9d3ebb83b", "component_budget": { "node_count": 2, "item_count": 2, @@ -19133,28 +19031,28 @@ "aggressively_trimmed": false, "split": false, "analyst_result": { - "summary": "The two issues share an identical title and a soft-similarity signal, but the packet does not provide enough body-level or technical evidence to confidently call them duplicates. Issue 45145 is still the best canonical anchor because it is earlier and has an inbound reference.", - "confidence": 0.68, - "canonical_issue_reason": "issue:45145 is earlier and has an inbound reference, so it is the best anchor among the two similar issues.", + "summary": "These two closed issues appear to be duplicates of the same energy-efficiency proposal, with identical titles and highly similar content. Issue #45145 is the better canonical thread because it was opened first and has an inbound reference.", + "confidence": 0.98, + "canonical_issue_reason": "#45145 is the earliest report and already has an inbound reference, so it is the best canonical issue for this duplicated topic.", "canonical_pr_reason": null, - "best_issue_reason": "issue:45145 is the strongest representative because it is earlier and more connected than 45151.", + "best_issue_reason": "#45145 is the strongest single issue to keep because it predates #45151 and appears to be the original discussion thread for the same request.", "best_pr_reason": null, "soft_edge_verdicts": [ { "left": "issue:45145", "right": "issue:45151", - "accept": false, - "reason": "Identical titles and a soft-similarity score are not enough to establish a duplicate; the issue bodies may still contain meaningful technical differences, and no explicit linking evidence is present." + "accept": true, + "reason": "Same title and highly similar issue content; they look like duplicate reports of the same underlying energy-reduction proposal." } ] }, "evaluator_result": { - "accept": false, - "feedback": "The canonical choice is grounded (45145 is earlier and has an inbound reference), but the duplicate verdict is too strong for the packet. The evidence only shows identical titles and a soft-similarity signal; it does not establish that the issue bodies are substantially overlapping or that there are no distinguishing technical differences." + "accept": true, + "feedback": "The analyst\u2019s conclusion is grounded in the packet: the issues share the exact same title, have soft similarity evidence, and #45145 is earlier with one inbound reference. The duplicate/canonical-issue reasoning is conservative enough here, and I don\u2019t see an overstatement beyond the available evidence." }, "evaluator_used": true, - "retried": true, - "accepted_nontrivial_soft_edge": false, + "retried": false, + "accepted_nontrivial_soft_edge": true, "error_kind": null, "error_message": null }, @@ -19231,7 +19129,7 @@ "soft_pairs": [ "issue:45672|issue:45674" ], - "prepared_review_unit_hash": "f7c4a8b9dab3181ee100f6a082f2c26866a5345f97a7d13204fb63425177b118", + "prepared_review_unit_hash": "150f663c93d6b2777d299cd7c30b0b62c0cf4b0096614d2380f1bafe55f49b36", "component_budget": { "node_count": 2, "item_count": 2, @@ -19253,18 +19151,18 @@ "aggressively_trimmed": false, "split": false, "analyst_result": { - "summary": "These are two separate 4-bit inference bugs: one Gemma4/audio-module failure from `torch.finfo()` on `uint8` weights, and one BitsAndBytesConfig regression where `llm_int8_skip_modules` drops the default `lm_head` exclusion and triggers an `AssertionError`. They share the broad theme of quantized inference but not the same underlying bug.", - "confidence": 0.96, - "canonical_issue_reason": "No canonical duplicate issue: the reports describe different failure modes and different code paths, despite both occurring during 4-bit inference.", + "summary": "The two issues both happen during 4-bit inference, but they describe different failure modes and likely different root causes: one is a torch.finfo() TypeError on uint8 audio-module weights, the other is an AssertionError caused by llm_int8_skip_modules dropping the default lm_head exclusion.", + "confidence": 0.95, + "canonical_issue_reason": "No single canonical duplicate stands out here; the issues target different code paths and fixes despite overlapping in 4-bit inference.", "canonical_pr_reason": null, - "best_issue_reason": "Issue 45674 is the better representative bug report because it identifies a specific configuration trigger, a clearer root cause in `BitsAndBytesConfig`, and a broader, self-contained impact surface.", + "best_issue_reason": "Issue 45674 is the clearer representative of the cluster because it states a specific configuration regression and an explicit failure condition, making it more actionable as the reference issue.", "best_pr_reason": null, "soft_edge_verdicts": [ { "left": "issue:45672", "right": "issue:45674", "accept": false, - "reason": "Reject: these are distinct bugs. One is a Gemma4/audio-module `torch.finfo()` TypeError on `uint8` weights; the other is a BitsAndBytesConfig `llm_int8_skip_modules`/`lm_head` exclusion regression causing an `AssertionError`. Shared 4-bit inference context is too broad for duplication." + "reason": "Reject: these are different bugs. 45672 is a dtype/torch.finfo failure in Gemma4 audio modules, while 45674 is a BitsAndBytesConfig skip-module regression causing an lm_head assertion." } ] }, @@ -19294,40 +19192,40 @@ "soft_pairs": [ "issue:45762|issue:45763" ], - "prepared_review_unit_hash": "9d222ccf9e236a9a2b737f7ef88bb89c4bbc8388379423409641f527ca0ad30b", + "prepared_review_unit_hash": "4385fa75ec3d2d25251c46a042418aea3159bdc1baf016b54ba0d3b36b4e1693", "component_budget": { "node_count": 2, "item_count": 2, "soft_pair_count": 1, - "serialized_chars": 1476, - "estimated_input_tokens": 369, - "estimated_eval_tokens": 994 + "serialized_chars": 1477, + "estimated_input_tokens": 370, + "estimated_eval_tokens": 996 }, "budget": { "node_count": 2, "item_count": 2, "soft_pair_count": 1, - "serialized_chars": 1476, - "estimated_input_tokens": 369, - "estimated_eval_tokens": 994 + "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": "These are two unrelated, likely spam/personal-name issues. They do not describe the same underlying bug or request, so they should not be clustered as duplicates.", + "summary": "The two closed issues do not look like duplicates of the same bug/change; they appear to be unrelated name-and-date entries with no shared context or problem description.", "confidence": 0.98, - "canonical_issue_reason": null, + "canonical_issue_reason": "No clear canonical issue: there is no evidence these reports describe the same underlying problem.", "canonical_pr_reason": null, - "best_issue_reason": null, + "best_issue_reason": "Neither issue is a good representative for a duplicate cluster because the titles/body provide no shared defect or change context.", "best_pr_reason": null, "soft_edge_verdicts": [ { "left": "issue:45762", "right": "issue:45763", "accept": false, - "reason": "Titles only share a surname pattern and both are just name/date strings; they do not describe the same issue or problem." + "reason": "Reject: the only similarity is the format of the titles; there is no description, target, or other evidence that they report the same issue." } ] }, @@ -19357,7 +19255,7 @@ "soft_pairs": [ "issue:45764|issue:45765" ], - "prepared_review_unit_hash": "cda98fe83d9f961afc716ecc47a2e04b482bf4b62d20056554d24aecd8146c53", + "prepared_review_unit_hash": "80709f98ec6ca8de142c8db9e251ceebe93e61b487ec287f92dde91304d90913", "component_budget": { "node_count": 2, "item_count": 2, @@ -19379,18 +19277,18 @@ "aggressively_trimmed": false, "split": false, "analyst_result": { - "summary": "The two issues are unrelated name/date stubs and do not describe the same underlying bug or request.", - "confidence": 0.97, + "summary": "These two issues are unrelated low-information records (different names and dates) and do not appear to describe the same bug or request.", + "confidence": 0.98, "canonical_issue_reason": null, "canonical_pr_reason": null, - "best_issue_reason": "No issue in this cluster is a good canonical representative; both are unrelated and look like noise or accidental submissions.", + "best_issue_reason": "Neither issue is a clear canonical choice; both are standalone and unrelated, with no substantive content linking them.", "best_pr_reason": null, "soft_edge_verdicts": [ { "left": "issue:45764", "right": "issue:45765", "accept": false, - "reason": "The titles reference different individuals and different dates; there is no shared bug, feature request, or code context." + "reason": "Only superficial similarity in the title pattern; the names and dates differ and there is no evidence of the same underlying issue." } ] }, @@ -19420,12 +19318,12 @@ "soft_pairs": [ "issue:45810|issue:46295" ], - "prepared_review_unit_hash": "8eb6c1a52eb6b117eaf57d0095f45637f08718275b74dccdbc7facb9f5b659dd", + "prepared_review_unit_hash": "e79bfe723f4cae4c9b1ec2f3c9506e5ef96acaba87423d0e6a4887245307ffab", "component_budget": { "node_count": 2, "item_count": 2, "soft_pair_count": 1, - "serialized_chars": 1542, + "serialized_chars": 1541, "estimated_input_tokens": 386, "estimated_eval_tokens": 1028 }, @@ -19433,7 +19331,7 @@ "node_count": 2, "item_count": 2, "soft_pair_count": 1, - "serialized_chars": 1542, + "serialized_chars": 1541, "estimated_input_tokens": 386, "estimated_eval_tokens": 1028 }, @@ -19442,18 +19340,18 @@ "aggressively_trimmed": false, "split": false, "analyst_result": { - "summary": "The two issues are not duplicates: one requests a new Qwen3_5 token-classification/value-model head, while the other asks for FP8 training support. They target different features and code paths.", - "confidence": 0.98, - "canonical_issue_reason": null, + "summary": "These are two unrelated open feature requests: one asks for a Qwen3_5 token-classification head, the other asks for FP8 training support. The similarity is superficial and they do not appear to describe the same underlying bug or change.", + "confidence": 0.96, + "canonical_issue_reason": "No canonical issue: the two items request different features in different areas of the stack (model head support vs FP8 training).", "canonical_pr_reason": null, - "best_issue_reason": "issue:45810 is the more concrete and actionable standalone request; issue:46295 is broader and unrelated to the other item.", + "best_issue_reason": "issue:45810 is the more concrete, implementation-oriented request; 46295 is a separate training feature and not a duplicate.", "best_pr_reason": null, "soft_edge_verdicts": [ { "left": "issue:45810", "right": "issue:46295", "accept": false, - "reason": "Reject: these are different requests with no shared underlying bug or change. One is a model-head/architecture feature request, the other is training precision support (FP8)." + "reason": "Different feature requests with no shared code path or explicit linkage; one is for Qwen3_5 token classification, the other for FP8 training support." } ] }, @@ -19493,91 +19391,91 @@ "issue:45871|issue:45873", "issue:45872|issue:45873" ], - "prepared_review_unit_hash": "2f54f7ac078b3d29f8c38d623679732334bd05291af323d92840cbcec093993b", + "prepared_review_unit_hash": "b6c98c744ebbe8c4977b34266aaf8ff9d83ef3b121278f7ac193b06943c568b7", "component_budget": { "node_count": 5, "item_count": 5, "soft_pair_count": 8, - "serialized_chars": 5724, - "estimated_input_tokens": 1431, - "estimated_eval_tokens": 3118 + "serialized_chars": 5727, + "estimated_input_tokens": 1432, + "estimated_eval_tokens": 3120 }, "budget": { "node_count": 5, "item_count": 5, "soft_pair_count": 8, - "serialized_chars": 5724, - "estimated_input_tokens": 1431, - "estimated_eval_tokens": 3118 + "serialized_chars": 5727, + "estimated_input_tokens": 1432, + "estimated_eval_tokens": 3120 }, "overflow_policy": "truncate_then_skip", "trimmed": false, "aggressively_trimmed": false, "split": false, "analyst_result": { - "summary": "The cluster is not a single duplicate set: it contains two exact duplicate issue pairs about different DataForge feature requests, plus one unrelated dataset/sample issue. Only the identical-title pairs should be merged.", - "confidence": 0.94, - "canonical_issue_reason": "No single canonical issue fits the whole cluster because the items split into separate topics. If forced to pick one representative, issue:45870 is the cleanest canonical for its exact duplicate pair with issue:45871.", + "summary": "The cluster looks like two likely duplicate issue pairs plus one separate issue: 45870/45871 share an identical title and body length, and 45872/45873 do as well, while 45869 is a different dataset-sample request. The metadata supports high similarity / likely duplication, but not a definitive exact-duplicate claim.", + "confidence": 0.86, + "canonical_issue_reason": "No single canonical issue for the whole cluster; the evidence supports two highly similar pairs (45870/45871 and 45872/45873) and one separate issue (45869), but the packet alone does not prove exact duplication.", "canonical_pr_reason": null, - "best_issue_reason": "Issue:45870 is the best representative of one true duplicate pair: it has an identical twin (issue:45871) and is the earlier copy. The rest of the cluster is a different feature request or unrelated sample-data post.", + "best_issue_reason": "45872 is the strongest representative of the more specific 'Financial time series + scraping service' thread, though its match with 45873 should be treated as highly similar rather than definitively exact from the provided metadata.", "best_pr_reason": null, "soft_edge_verdicts": [ { "left": "issue:45869", "right": "issue:45870", "accept": false, - "reason": "Different requests: a dataset bundle versus an example for financial time series prediction. Same broad domain, but not the same underlying issue." + "reason": "Different requests: a financial dataset bundle vs. a time-series prediction example. Related domain, but not the same underlying issue." }, { "left": "issue:45869", "right": "issue:45871", "accept": false, - "reason": "Different requests: a dataset bundle versus an example for financial time series prediction. Not the same change or bug." + "reason": "Different requests: a financial dataset bundle vs. a time-series prediction example. Not the same underlying issue." }, { "left": "issue:45870", "right": "issue:45871", "accept": true, - "reason": "Exact duplicate: same title and effectively the same financial time series prediction example request." + "reason": "Highly similar and likely the same request: identical title and identical body length strongly suggest a duplicate, though the metadata does not by itself prove exact duplication." }, { "left": "issue:45870", "right": "issue:45872", "accept": false, - "reason": "Related only at a high level (financial/DataForge), but one asks for an example and the other for a scraping service feature. Different underlying request." + "reason": "Different request scopes: a prediction example vs. a scraping-service integration request." }, { "left": "issue:45870", "right": "issue:45873", "accept": false, - "reason": "Related only at a high level (financial/DataForge), but one asks for an example and the other for a scraping service feature. Different underlying request." + "reason": "Different request scopes: a prediction example vs. a scraping-service integration request." }, { "left": "issue:45871", "right": "issue:45872", "accept": false, - "reason": "Different feature requests: example prediction workflow versus time-series-plus-scraping service." + "reason": "Different request scopes: a prediction example vs. a scraping-service integration request." }, { "left": "issue:45871", "right": "issue:45873", "accept": false, - "reason": "Different feature requests: example prediction workflow versus time-series-plus-scraping service." + "reason": "Different request scopes: a prediction example vs. a scraping-service integration request." }, { "left": "issue:45872", "right": "issue:45873", "accept": true, - "reason": "Exact duplicate: same title and same DataForge time-series + scraping service request." + "reason": "Highly similar and likely the same request: identical title and identical body length strongly suggest a duplicate, though the packet only supports a conservative likely-duplicate reading." } ] }, "evaluator_result": { - "accept": true, - "feedback": "Grounded overall: the cluster clearly splits into the dataset/sample issue and two same-title DataForge request pairs, and the non-duplicate rejections are conservative. The only slight stretch is calling the same-title pairs \u201cexact\u201d duplicates without body text, but the metadata supports them well enough here." + "accept": false, + "feedback": "The cluster summary is a bit too strong for the evidence in the packet. The packet shows identical titles, equal body lengths, and high soft-similarity for 45870/45871 and 45872/45873, but it does not directly establish they are exact duplicates. The conservative read is that those pairs are highly similar / likely duplicate, not definitively exact duplicate. Because the analyst\u2019s canonical and best-issue reasons rely on that stronger claim, I would reject and ask for softer wording grounded only in the provided metadata." }, "evaluator_used": true, - "retried": false, + "retried": true, "accepted_nontrivial_soft_edge": true, "error_kind": null, "error_message": null @@ -19601,46 +19499,49 @@ "soft_pairs": [ "issue:46439|issue:46441" ], - "prepared_review_unit_hash": "8497c5ed3add7fdd0c9b4c29732916c73adbef5b206bf56d55820a21385d8e3e", + "prepared_review_unit_hash": "132b15eda3023096b614e4f0a9f0766c61064c843b992670137effa48dc23f15", "component_budget": { "node_count": 2, "item_count": 2, "soft_pair_count": 1, - "serialized_chars": 1637, - "estimated_input_tokens": 410, - "estimated_eval_tokens": 1076 + "serialized_chars": 1636, + "estimated_input_tokens": 409, + "estimated_eval_tokens": 1074 }, "budget": { "node_count": 2, "item_count": 2, "soft_pair_count": 1, - "serialized_chars": 1637, - "estimated_input_tokens": 410, - "estimated_eval_tokens": 1076 + "serialized_chars": 1636, + "estimated_input_tokens": 409, + "estimated_eval_tokens": 1074 }, "overflow_policy": "truncate_then_skip", "trimmed": false, "aggressively_trimmed": false, "split": false, "analyst_result": { - "summary": "Both issues involve `cache_implementation=\"static\"` and linear-attention, but they describe different concrete failures: one is shape corruption during early initialization, the other is a Qwen3.5 `KeyError` on `linear_attention`. They may be related, but there isn\u2019t enough evidence to treat them as the same underlying bug.", - "confidence": 0.61, - "canonical_issue_reason": "Issue 46439 is the better canonical report because it describes the broader mechanism (`early_initialization` corrupting linear-attention layers) rather than a single model-specific crash.", + "summary": "These two issues are related to static-cache failures around linear attention, but the packet does not show enough shared evidence to confirm they are duplicates. The safer read is that they may be adjacent symptoms or the same area of breakage, not the same grounded bug report.", + "confidence": 0.58, + "canonical_issue_reason": "Issue #46439 is the better canonical issue because it describes the broader mechanism (`early_initialization` corrupting linear-attention layers) rather than one model-specific crash symptom.", "canonical_pr_reason": null, - "best_issue_reason": "Issue 46439 is the strongest representative of the cluster since it frames the likely root cause and failure mode more generally; 46441 looks like a specific manifestation or adjacent bug report.", + "best_issue_reason": "Issue #46439 is the strongest standalone tracker since it names the suspected root cause and the resulting shape corruption more directly, making it more useful for triage and eventual fixing.", "best_pr_reason": null, "soft_edge_verdicts": [ { "left": "issue:46439", "right": "issue:46441", "accept": false, - "reason": "Reject as a duplicate: both mention static cache and linear-attention, but the reported failures are different (`RuntimeError` from wrong shapes vs `KeyError: 'linear_attention'`) and there\u2019s no shared code/path evidence tying them to one concrete bug." + "reason": "The evidence is only soft similarity; there are no shared files, explicit targets, or body details here tying both reports to the same failure path. They may be related, but duplication is not confirmed from the provided packet." } ] }, - "evaluator_result": null, - "evaluator_used": false, - "retried": false, + "evaluator_result": { + "accept": false, + "feedback": "The duplicate verdict is too strong for the evidence in the packet. The pair only has soft similarity and no shared files/targets or body details here, so saying they are the same bug or the same failure path overstates what is grounded. A more conservative read would be that they may be related, but duplication is not confirmed from the provided evidence." + }, + "evaluator_used": true, + "retried": true, "accepted_nontrivial_soft_edge": false, "error_kind": null, "error_message": null @@ -19664,12 +19565,12 @@ "soft_pairs": [ "issue:46551|issue:46552" ], - "prepared_review_unit_hash": "ffa850371bba38da396a1b2225a2c51b916a9b6d0fba21fea6be5d8a739cf10f", + "prepared_review_unit_hash": "06093e7828b568d7eb96bd27305844ee7e0bd6443acb0eb2aae66066cd4cd3d1", "component_budget": { "node_count": 2, "item_count": 2, "soft_pair_count": 1, - "serialized_chars": 1707, + "serialized_chars": 1706, "estimated_input_tokens": 427, "estimated_eval_tokens": 1110 }, @@ -19677,7 +19578,7 @@ "node_count": 2, "item_count": 2, "soft_pair_count": 1, - "serialized_chars": 1707, + "serialized_chars": 1706, "estimated_input_tokens": 427, "estimated_eval_tokens": 1110 }, @@ -19686,27 +19587,27 @@ "aggressively_trimmed": false, "split": false, "analyst_result": { - "summary": "The cluster is a clear duplicate pair: both issues have the exact same title and appear to describe the same Wav2Vec2CTCTokenizer v5 special-token regression. The open issue is the better canonical target.", - "confidence": 0.99, - "canonical_issue_reason": "Issue #46552 is open, has inbound references, and matches #46551 exactly in title/content, making it the better canonical issue to keep tracking the bug.", + "summary": "The two issues appear to describe the same Wav2Vec2CTCTokenizer v5 regression, with identical titles and body lengths and only a soft-similarity signal supporting the match. The open issue is the better tracker, while the closed one looks like the older report in the pair.", + "confidence": 0.92, + "canonical_issue_reason": "issue:46552 is the better canonical issue because it is open and already has inbound references, so it is the best living tracker for the shared bug.", "canonical_pr_reason": null, - "best_issue_reason": "Issue #46552 is the best issue in the cluster because it is open and can serve as the active tracker for this identical report.", + "best_issue_reason": "issue:46552 is the strongest issue to keep for triage because it is open, has inbound references, and matches the same reported tokenizer regression.", "best_pr_reason": null, "soft_edge_verdicts": [ { "left": "issue:46551", "right": "issue:46552", "accept": true, - "reason": "Same title, same apparent description, same underlying tokenizer bug; these are duplicate reports of the exact issue." + "reason": "Identical title and body length, plus the soft-similarity signal, strongly suggest the same underlying tokenizer bug; however, this is a conservative same-bug match rather than evidence of an explicit duplicate relation." } ] }, "evaluator_result": { - "accept": true, - "feedback": "Grounded in the packet: the two issues have identical titles, identical body length, and soft-similarity/Jaccard support for duplication; choosing the open issue #46552 as the canonical tracker is conservative and well-supported. No overstatement beyond the provided evidence." + "accept": false, + "feedback": "The pair looks highly similar and likely refers to the same bug, but the verdict overstates the evidence by calling one a closed duplicate of the other. The packet only supports identical titles/body length, a soft-similarity signal, and that one issue is closed while the other is open with inbound references; it does not explicitly establish a duplicate relation. The canonical/open-tracker choice is reasonable, but the soft-edge reason should stay more conservative." }, "evaluator_used": true, - "retried": false, + "retried": true, "accepted_nontrivial_soft_edge": true, "error_kind": null, "error_message": null @@ -19714,7 +19615,7 @@ { "label": "pull_request", "component_index": 1, - "component_count": 39, + "component_count": 35, "review_unit_index": 1, "review_unit_count": 1, "status": "reviewed", @@ -19766,7 +19667,7 @@ "pull_request:44827|pull_request:46662", "pull_request:43251|pull_request:43254" ], - "prepared_review_unit_hash": "e55b8a403118ca3b1d7ea699260ad0c5252e708a220161cda0cee0660ff94d14", + "prepared_review_unit_hash": "e844037dac2181e9b2a01a8475522e5f330db793acbd1656c106e56918e52f6c", "component_budget": { "node_count": 15, "item_count": 15, @@ -19788,162 +19689,162 @@ "aggressively_trimmed": false, "split": false, "analyst_result": { - "summary": "Mostly a noisy similarity cluster: the items span unrelated model additions, tokenizer/processor tweaks, docs, export tests, and small bugfixes. The only true near-duplicate pair is the two loss_utils cross-entropy kwargs fixes (43251/43254).", - "confidence": 0.96, + "summary": "This cluster is mostly noisy: most PRs are unrelated model additions, docs, tests, or small subsystem fixes. The only true duplicate-like overlap is the pair of loss_utils PRs for issue 43240; everything else should stay separate.", + "confidence": 0.93, "canonical_issue_reason": null, - "canonical_pr_reason": "No single canonical PR for the whole cluster: most items are unrelated, and only 43251/43254 are actually the same underlying code-path fix.", + "canonical_pr_reason": "PR 43251 is the best canonical representative because it is the clearer issue-linked fix for 43240 and matches the same `loss_utils.py` code path as 43254, but with a more explicit fix-oriented scope and fuller review trail.", "best_issue_reason": null, - "best_pr_reason": "PR 43251 is the best representative of the only real duplicate-like change in the set: it directly fixes fixed_cross_entropy for the tracked issue 43240 and is the more explicit version of the two loss PRs.", + "best_pr_reason": "PR 43251 is the best overall PR candidate in this cluster because it is the strongest genuine overlap and most complete version of the shared `fixed_cross_entropy` change; the rest of the items are not duplicate-worthy.", "soft_edge_verdicts": [ { "left": "pull_request:39895", "right": "pull_request:43424", "accept": false, - "reason": "Different changes: VideoPrism model/doc addition vs ExecuTorch dynamic-shape test. Same broad area of testing/model infra, but not the same bug or mergeable fix." + "reason": "Unrelated PRs: one adds VideoPrism model support, the other adds Executorch dynamic-shape tests for decoder-only export." }, { "left": "pull_request:41855", "right": "pull_request:43424", "accept": false, - "reason": "Tokenizer missing-methods fix vs ExecuTorch export test; unrelated code paths and functionality." + "reason": "Different changes: Mistral tokenizer missing methods versus Executorch export testing." }, { "left": "pull_request:41855", "right": "pull_request:43823", "accept": false, - "reason": "Mistral tokenizer methods vs adding MobileLLM model support. Different model families and changes." + "reason": "Tokenizer helper additions are unrelated to a new MobileLLM model implementation." }, { "left": "pull_request:42668", "right": "pull_request:43823", "accept": false, - "reason": "Processor robustness refactor vs adding MobileLLM. Both touch model plumbing, but not the same underlying change." + "reason": "Both touch processor/model auto code, but they are for different models and different behaviors." }, { "left": "pull_request:43251", "right": "pull_request:43823", "accept": false, - "reason": "Loss function kwargs fix vs new MobileLLM model implementation; unrelated subsystems." + "reason": "Loss function keyword support is unrelated to adding a new MobileLLM model." }, { "left": "pull_request:43251", "right": "pull_request:44827", "accept": false, - "reason": "Cross-entropy kwargs fix vs Mistral4 test/model updates. No shared concrete bug or code path." + "reason": "Cross-entropy kwargs support is unrelated to Mistral4 test/model adjustments." }, { "left": "pull_request:43424", "right": "pull_request:44827", "accept": false, - "reason": "ExecuTorch dynamic-shape export test vs Mistral4 fixes; only both are tests/model infra, not duplicates." + "reason": "Executorch export tests and Mistral4 test fixes are different code paths and changes." }, { "left": "pull_request:43424", "right": "pull_request:43823", "accept": false, - "reason": "Dynamic-shape export test and MobileLLM model addition are unrelated changes." + "reason": "Executorch dynamic-shape export tests do not overlap with MobileLLM model addition." }, { "left": "pull_request:43823", "right": "pull_request:46662", "accept": false, - "reason": "Model addition for MobileLLM vs SmolVLM video padding fix; completely different code paths." + "reason": "Different model families and different bugs: MobileLLM addition versus SmolVLM video padding validation." }, { "left": "pull_request:43823", "right": "pull_request:45655", "accept": false, - "reason": "MobileLLM model addition vs config resolution order fix; unrelated scope." + "reason": "Model addition and config-resolution fix are unrelated changes." }, { "left": "pull_request:43823", "right": "pull_request:43995", "accept": false, - "reason": "MobileLLM model addition vs Falcon output-interface refactor; different models and problems." + "reason": "New MobileLLM model code does not match Falcon output-collection refactoring." }, { "left": "pull_request:43995", "right": "pull_request:46662", "accept": false, - "reason": "Falcon standardized output refactor vs SmolVLM video padding; unrelated changes." + "reason": "Falcon forward refactor and SmolVLM video padding are unrelated subsystems." }, { "left": "pull_request:43995", "right": "pull_request:45213", "accept": false, - "reason": "Falcon refactor vs model-creation skill/docs/meta PR; not the same bug or feature." + "reason": "Falcon model refactor is unrelated to the model-creation skill PR." }, { "left": "pull_request:43995", "right": "pull_request:44827", "accept": false, - "reason": "Falcon output-interface refactor vs Mistral4 test fixes; different model code paths." + "reason": "Falcon output-interface refactor and Mistral4 test fixes are not the same change." }, { "left": "pull_request:44827", "right": "pull_request:45170", "accept": false, - "reason": "Mistral4 test/model work vs layrnorm->layernorm renaming across CLIP-like models. Similar ML area, but not the same change." + "reason": "Mistral4 test fixes are unrelated to the layrnorm-to-layernorm rename conversion fix." }, { "left": "pull_request:44827", "right": "pull_request:46267", "accept": false, - "reason": "Mistral4 fixes vs chat response parsing overhaul; unrelated functionality." + "reason": "Mistral4 test/model work is unrelated to serve chat parsing updates." }, { "left": "pull_request:44827", "right": "pull_request:46048", "accept": false, - "reason": "Mistral4 tests/model work vs TDT loss kernel integration; different subsystems and fixes." + "reason": "Mistral4 fixes and TDT loss kernel integration are different features." }, { "left": "pull_request:44827", "right": "pull_request:45655", "accept": false, - "reason": "Mistral4 work vs config resolution ordering fix; unrelated." + "reason": "Mistral4 work is unrelated to config resolution ordering in `modeling_utils.py`." }, { "left": "pull_request:44827", "right": "pull_request:45213", "accept": false, - "reason": "Mistral4 fixes vs model-creation skill/meta PR; not the same underlying change." + "reason": "Mistral4 model/test changes do not duplicate the model-creation skill PR." }, { "left": "pull_request:43823", "right": "pull_request:45213", "accept": false, - "reason": "MobileLLM model addition vs meta skill/documentation PR; unrelated." + "reason": "MobileLLM model addition and the model-creation skill are not the same code change." }, { "left": "pull_request:45213", "right": "pull_request:46662", "accept": false, - "reason": "Meta model-creation skill PR vs SmolVLM video-padding fix; no shared bug or change." + "reason": "A model-creation skill draft is unrelated to SmolVLM video padding behavior." }, { "left": "pull_request:45655", "right": "pull_request:46662", "accept": false, - "reason": "Config resolution-order fix vs SmolVLM video padding; unrelated code paths." + "reason": "Config resolution ordering fix is unrelated to SmolVLM video padding." }, { "left": "pull_request:44827", "right": "pull_request:46662", "accept": false, - "reason": "Mistral4 fixes vs SmolVLM padding change; different models and problems." + "reason": "Mistral4 tests/model changes are unrelated to SmolVLM video padding." }, { "left": "pull_request:43251", "right": "pull_request:43254", "accept": true, - "reason": "Both modify src/transformers/loss/loss_utils.py to fix the same fixed_cross_entropy kwargs handling for issue 43240; these are mergeable into one PR." + "reason": "Same underlying loss-function change: both modify `fixed_cross_entropy` in `loss_utils.py` to pass through new `cross_entropy` kwargs for issue 43240, and they could plausibly be merged into one PR." } ] }, "evaluator_result": { "accept": true, - "feedback": "Grounded overall. The summary conservatively identifies the only mergeable near-duplicate pair as 43251/43254, which both touch src/transformers/loss/loss_utils.py and the same tracked issue 43240. The rejected soft edges are also conservative: the paired PRs are in clearly different subsystems or feature areas, so they do not look like single-bugfix duplicates." + "feedback": "The cluster summary is mostly grounded in the packet, and the only clearly mergeable same-bugfix overlap is the `loss_utils.py` pair for issue 43240 (`43251`/`43254`). The soft-edge false verdicts are conservative overall, and the accepted pair is well supported by the shared file and issue target. I did not see a clear overstatement that would require rejection." }, "evaluator_used": true, "retried": false, @@ -19954,7 +19855,7 @@ { "label": "pull_request", "component_index": 2, - "component_count": 39, + "component_count": 35, "review_unit_index": 1, "review_unit_count": 1, "status": "reviewed", @@ -19970,12 +19871,12 @@ "soft_pairs": [ "pull_request:40546|pull_request:42781" ], - "prepared_review_unit_hash": "f81c16f80f5ce60bf9fe98d9855c3e6502377eb24f0228c1805a247ef8ac439a", + "prepared_review_unit_hash": "b47d7109cd8571541e7cfa247865248a934fb17946159b699c57258e226a7083", "component_budget": { "node_count": 2, "item_count": 2, "soft_pair_count": 1, - "serialized_chars": 7225, + "serialized_chars": 7226, "estimated_input_tokens": 1807, "estimated_eval_tokens": 3870 }, @@ -19983,7 +19884,7 @@ "node_count": 2, "item_count": 2, "soft_pair_count": 1, - "serialized_chars": 7225, + "serialized_chars": 7226, "estimated_input_tokens": 1807, "estimated_eval_tokens": 3870 }, @@ -19992,18 +19893,18 @@ "aggressively_trimmed": false, "split": false, "analyst_result": { - "summary": "Both PRs are about VibeVoice, but they appear to be different deliverables: one is the base implementation, the other is a realtime variant/extension. They share several files because the second builds on the same model family, but they do not look like the same concrete change.", - "confidence": 0.91, + "summary": "Two open PRs both center on VibeVoice, but they are not the same change set: one implements the base model, the other adds a realtime variant and related tokenizer/generation work.", + "confidence": 0.94, "canonical_issue_reason": null, - "canonical_pr_reason": "PR 40546 is the broader foundational VibeVoice implementation, has much higher review activity, and is not a draft; PR 42781 is a later realtime follow-up rather than the same change.", + "canonical_pr_reason": "PR #40546 is the best canonical PR because it is the broader, earlier non-draft implementation of VibeVoice and appears to establish the core model support that later work can build on.", "best_issue_reason": null, - "best_pr_reason": "PR 40546 is the best representative of the cluster because it appears to establish the initial VibeVoice integration that the later realtime PR extends.", + "best_pr_reason": "PR #40546 is the strongest representative of the cluster since it is the main base VibeVoice implementation; #42781 looks like a follow-up feature branch rather than the same underlying PR.", "soft_edge_verdicts": [ { "left": "pull_request:40546", "right": "pull_request:42781", "accept": false, - "reason": "Reject: despite shared VibeVoice files, this looks like a base implementation versus a realtime extension, not one identical bugfix/change that could be merged as a single PR." + "reason": "Reject: both touch shared VibeVoice files, but #40546 implements the base model while #42781 adds a realtime variant plus acoustic tokenizer and generation changes. They are related, but not the same concrete bug/change and would not plausibly be merged as one PR." } ] }, @@ -20017,7 +19918,7 @@ { "label": "pull_request", "component_index": 3, - "component_count": 39, + "component_count": 35, "review_unit_index": 1, "review_unit_count": 1, "status": "reviewed", @@ -20033,12 +19934,12 @@ "soft_pairs": [ "pull_request:40755|pull_request:41224" ], - "prepared_review_unit_hash": "a01c0455c3690d37c839ca5bdac6e61506a76946d3adc389962c7bf60befdcff", + "prepared_review_unit_hash": "a06a7073da7c0fa39246e4d6bf706427980b0651b3eb65a0084b7c3f7a6469ac", "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 }, @@ -20046,7 +19947,7 @@ "node_count": 2, "item_count": 2, "soft_pair_count": 1, - "serialized_chars": 4674, + "serialized_chars": 4675, "estimated_input_tokens": 1169, "estimated_eval_tokens": 2594 }, @@ -20055,18 +19956,18 @@ "aggressively_trimmed": false, "split": false, "analyst_result": { - "summary": "These are two unrelated pull requests in different model families: TimesFM covariate forecasting vs. DINOv3 image classification support. They share no filenames, code paths, or concrete bug/change target, so they should not be clustered as duplicates.", - "confidence": 0.98, + "summary": "The cluster contains two unrelated PRs: one adds TimesFM covariate forecasting, the other adds DINOv3 image classification support. They do not describe the same change or code path, so they should not be merged as duplicates.", + "confidence": 0.99, "canonical_issue_reason": null, - "canonical_pr_reason": null, + "canonical_pr_reason": "Use PR #41224 as the cluster representative because it is the more concrete, self-contained feature addition and has substantial review activity; however, it is not a duplicate of #40755.", "best_issue_reason": null, - "best_pr_reason": "PR #41224 is the more focused and review-active change, but it is still a separate DINOv3 feature addition, not a canonical duplicate for the TimesFM PR.", + "best_pr_reason": "PR #41224 is the better standalone representative for this cluster because it is a clearly scoped model capability addition with focused files and discussion, while #40755 is a separate TimesFM feature.", "soft_edge_verdicts": [ { "left": "pull_request:40755", "right": "pull_request:41224", "accept": false, - "reason": "Different models and different functionality: TimesFM forecasting with covariates versus DINOv3 image-classification support. No shared code path or mergeable overlapping fix/change." + "reason": "Different models and different functionality: TimesFM covariate forecasting versus DINOv3 image classification. They do not share the same underlying bug or change and could not plausibly be merged into one PR." } ] }, @@ -20080,7 +19981,7 @@ { "label": "pull_request", "component_index": 4, - "component_count": 39, + "component_count": 35, "review_unit_index": 1, "review_unit_count": 1, "status": "reviewed", @@ -20143,78 +20044,7 @@ { "label": "pull_request", "component_index": 5, - "component_count": 39, - "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-41797-3", - "nodes": [ - "pull_request:41797", - "pull_request:45186", - "pull_request:45630" - ], - "soft_pairs": [ - "pull_request:41797|pull_request:45186", - "pull_request:45186|pull_request:45630" - ], - "prepared_review_unit_hash": "373ea9c3791a91c2ce818188a47a100c8eea678fdca97b30a412409d007dd1b4", - "component_budget": { - "node_count": 3, - "item_count": 3, - "soft_pair_count": 2, - "serialized_chars": 10436, - "estimated_input_tokens": 2609, - "estimated_eval_tokens": 5474 - }, - "budget": { - "node_count": 3, - "item_count": 3, - "soft_pair_count": 2, - "serialized_chars": 10436, - "estimated_input_tokens": 2609, - "estimated_eval_tokens": 5474 - }, - "overflow_policy": "truncate_then_skip", - "trimmed": false, - "aggressively_trimmed": false, - "split": false, - "analyst_result": { - "summary": "These are three separate model-addition PRs (DeepSeek OCR, Isaac, Kimi2-6), not duplicate fixes for the same underlying change.", - "confidence": 0.97, - "canonical_issue_reason": null, - "canonical_pr_reason": "No canonical PR: each PR introduces a different model integration with different code, docs, and tests.", - "best_issue_reason": null, - "best_pr_reason": "#45630 is the strongest representative of the set because it is the most expansive and recent integration, with the broadest supporting changes and test coverage, though it is not a duplicate of the others.", - "soft_edge_verdicts": [ - { - "left": "pull_request:41797", - "right": "pull_request:45186", - "accept": false, - "reason": "Different models and different implementation targets; they only overlap on generic Hugging Face model-registration plumbing, not the same bug or change." - }, - { - "left": "pull_request:45186", - "right": "pull_request:45630", - "accept": false, - "reason": "Both are new-model PRs, but for different models (Isaac vs Kimi2-6) with distinct code paths and artifacts, so they are 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": 6, - "component_count": 39, + "component_count": 35, "review_unit_index": 1, "review_unit_count": 1, "status": "reviewed", @@ -20230,7 +20060,7 @@ "soft_pairs": [ "pull_request:42975|pull_request:42976" ], - "prepared_review_unit_hash": "f4ab91d2863d6358eb50653eaf53e65a309d302fd2b6b65115f738c8fbc6b6f8", + "prepared_review_unit_hash": "afeac0341974deb4ad3902fe12381e1342022591444f11295f24d702bdff8e4c", "component_budget": { "node_count": 2, "item_count": 2, @@ -20252,18 +20082,18 @@ "aggressively_trimmed": false, "split": false, "analyst_result": { - "summary": "Two separate GitHub Actions maintenance PRs: one focused on Node 24 compatibility for workflow actions, the other on upgrading workflow/action versions more broadly, including docker/build-push-action. They overlap in area but do not look like the same underlying fix.", - "confidence": 0.91, + "summary": "Two open PRs both bulk-upgrade GitHub Actions workflows, but they target different action families: one is for Node 24 compatibility across general workflow actions, the other updates Docker build actions and a few related workflow steps. They are related maintenance, not the same underlying change.", + "confidence": 0.96, "canonical_issue_reason": null, - "canonical_pr_reason": "PR 42975 is the better canonical representative because it has the clearest concrete motivation (Node 24 compatibility) and a focused set of workflow action version bumps tied to that goal.", + "canonical_pr_reason": "PR 42975 is the better canonical representative because its title and diff clearly describe the broader workflow-action runtime compatibility upgrade, and its changes span several core GitHub Actions used across workflows.", "best_issue_reason": null, - "best_pr_reason": "PR 42975 is the best overall representative: it is the more specific, actionable workflow compatibility upgrade and better captures the cluster\u2019s shared maintenance theme than the broader version-bump PR.", + "best_pr_reason": "PR 42975 is the best PR in the cluster because it most directly captures the shared theme of workflow action version bumps and has the clearest, most general maintenance scope.", "soft_edge_verdicts": [ { "left": "pull_request:42975", "right": "pull_request:42976", "accept": false, - "reason": "Both are workflow dependency upgrade PRs, but they target different action families and likely different maintenance concerns: checkout/cache/upload-artifact Node 24 compatibility versus docker/build-push-action and other latest-version bumps. Same subsystem, not the same concrete change or bug fix." + "reason": "Do not merge: despite overlapping workflow files and similar upgrade wording, the PRs update different concrete actions (checkout/cache/upload-artifact vs docker/build-push-action and a few others). They do not fix the same code-path problem and would not plausibly be merged into one PR without conflating separate dependency bumps." } ] }, @@ -20276,8 +20106,8 @@ }, { "label": "pull_request", - "component_index": 7, - "component_count": 39, + "component_index": 6, + "component_count": 35, "review_unit_index": 1, "review_unit_count": 1, "status": "reviewed", @@ -20285,130 +20115,101 @@ "source": "llm", "cache_hit": false, "model": "gpt-5.4-mini?service_tier=flex", - "cluster_id": "cluster-42978-7", + "cluster_id": "cluster-42978-6", "nodes": [ "pull_request:42978", "pull_request:43448", "pull_request:43451", "pull_request:45490", "pull_request:46562", - "pull_request:46641", "pull_request:46658" ], "soft_pairs": [ "pull_request:42978|pull_request:43448", "pull_request:42978|pull_request:43451", - "pull_request:42978|pull_request:46641", "pull_request:42978|pull_request:45490", "pull_request:43448|pull_request:43451", - "pull_request:43448|pull_request:46641", "pull_request:43448|pull_request:45490", - "pull_request:43451|pull_request:46641", "pull_request:43451|pull_request:45490", "pull_request:45490|pull_request:46562", - "pull_request:45490|pull_request:46641", "pull_request:46562|pull_request:46658" ], - "prepared_review_unit_hash": "b5289b7ec47b729c3251064a01961be2034345ec040a1095fa18280598e6a69b", + "prepared_review_unit_hash": "5de7660b6a34267a6c684aa4a1fa8871a6e8d624d3b49592a8b4a68565cf3a90", "component_budget": { - "node_count": 7, - "item_count": 7, - "soft_pair_count": 12, - "serialized_chars": 23138, - "estimated_input_tokens": 5785, - "estimated_eval_tokens": 11826 + "node_count": 6, + "item_count": 6, + "soft_pair_count": 8, + "serialized_chars": 20309, + "estimated_input_tokens": 5078, + "estimated_eval_tokens": 10412 }, "budget": { - "node_count": 7, - "item_count": 7, - "soft_pair_count": 12, - "serialized_chars": 23138, - "estimated_input_tokens": 5785, - "estimated_eval_tokens": 11826 + "node_count": 6, + "item_count": 6, + "soft_pair_count": 8, + "serialized_chars": 20309, + "estimated_input_tokens": 5078, + "estimated_eval_tokens": 10412 }, "overflow_policy": "truncate_then_skip", "trimmed": false, "aggressively_trimmed": false, "split": false, "analyst_result": { - "summary": "These PRs are mostly unrelated model additions plus one tiny docs-only punctuation fix. The only true duplicate-style pair is the Step3p7 mini-model/demo PR that supports the main Step3p7 Flash PR; the rest share boilerplate files but not the same underlying change.", + "summary": "This cluster is mostly a set of unrelated model-addition PRs (ViT NEPA, Molmo, Molmo2, CTSM, Step3p7 Flash). The only closely related pair is the Step3p7 Flash PR and its WIP mini-model/support PR, but they still look like companion work rather than the same duplicate change.", "confidence": 0.96, "canonical_issue_reason": null, - "canonical_pr_reason": "No issue artifacts are present. For the duplicate pair, PR 46562 is the main Step3p7 Flash model addition; PR 46658 is a supporting mini-model/demo bundle for the same feature rather than a separate change.", + "canonical_pr_reason": "PR 46562 is the best canonical representative because it is the main Step3p7 Flash implementation PR and sits closest to the only near-related companion PR in the cluster.", "best_issue_reason": null, - "best_pr_reason": "PR 46562 is the best representative PR for the only real duplicate pair because it contains the primary model implementation and registry/docs/test wiring. PR 46658 is auxiliary scaffolding for that same feature.", + "best_pr_reason": "PR 43451 is the strongest standalone PR in the set: it has the broadest model-integration scope and the highest review activity, making it the most mature representative of the overall cluster theme.", "soft_edge_verdicts": [ { "left": "pull_request:42978", "right": "pull_request:43448", "accept": false, - "reason": "Different model additions: ViT NEPA vs Molmo. Shared auto/docs scaffolding is not enough." + "reason": "Different models (ViT NEPA vs Molmo); similar model-addition structure, but not the same change." }, { "left": "pull_request:42978", "right": "pull_request:43451", "accept": false, - "reason": "Different model additions: ViT NEPA vs Molmo2. Same repo boilerplate, different underlying feature." - }, - { - "left": "pull_request:42978", - "right": "pull_request:46641", - "accept": false, - "reason": "ViT NEPA model addition vs a one-line docs punctuation fix; unrelated changes." + "reason": "Different models (ViT NEPA vs Molmo2); shared framework files do not make this a duplicate." }, { "left": "pull_request:42978", "right": "pull_request:45490", "accept": false, - "reason": "Different model families: ViT NEPA vs CTSM. Overlap is only generic docs/auto-registration files." + "reason": "Different model additions (ViT NEPA vs CTSM), not the same underlying code change." }, { "left": "pull_request:43448", "right": "pull_request:43451", "accept": false, - "reason": "Molmo and Molmo2 are separate model integrations, not the same code change." - }, - { - "left": "pull_request:43448", - "right": "pull_request:46641", - "accept": false, - "reason": "Molmo model addition vs unrelated docs punctuation edit." + "reason": "Molmo and Molmo2 are related by name but are distinct model integrations with separate code and tests." }, { "left": "pull_request:43448", "right": "pull_request:45490", "accept": false, - "reason": "Molmo vs CTSM are distinct model additions despite shared registry/docs touchpoints." - }, - { - "left": "pull_request:43451", - "right": "pull_request:46641", - "accept": false, - "reason": "Molmo2 model addition vs unrelated docs punctuation edit." + "reason": "Molmo and CTSM are unrelated model additions." }, { "left": "pull_request:43451", "right": "pull_request:45490", "accept": false, - "reason": "Molmo2 vs CTSM are different model integrations, not duplicates." + "reason": "Molmo2 and CTSM are different model additions; same subsystem, but not duplicate work." }, { "left": "pull_request:45490", "right": "pull_request:46562", "accept": false, - "reason": "CTSM and Step3p7 Flash are different model additions; shared auto-mapping edits do not imply the same change." - }, - { - "left": "pull_request:45490", - "right": "pull_request:46641", - "accept": false, - "reason": "CTSM model addition vs unrelated docs punctuation edit." + "reason": "Both are model additions, but CTSM and Step3p7 Flash are distinct models and code paths." }, { "left": "pull_request:46562", "right": "pull_request:46658", - "accept": true, - "reason": "Both are about Step3p7 Flash: 46562 adds the model, and 46658 is a mini-model/demo bundle for the same feature and could plausibly be merged together." + "accept": false, + "reason": "46658 looks like WIP mini-model/assets and scripts for Step3p7 Flash, but not the same concrete implementation change as 46562." } ] }, @@ -20421,8 +20222,8 @@ }, { "label": "pull_request", - "component_index": 8, - "component_count": 39, + "component_index": 7, + "component_count": 35, "review_unit_index": 1, "review_unit_count": 1, "status": "reviewed", @@ -20457,118 +20258,118 @@ "pull_request:46380|pull_request:46539", "pull_request:46380|pull_request:46579" ], - "prepared_review_unit_hash": "67023dd68b492442a4b60bcfbf793cd64b37680ce71ee417a0e8d5ff6b14a4b2", + "prepared_review_unit_hash": "cbb600115a83942b08b9d1cc7c72709384bba80a2939d7a5c055652234c5f473", "component_budget": { "node_count": 8, "item_count": 8, "soft_pair_count": 14, - "serialized_chars": 23161, - "estimated_input_tokens": 5791, - "estimated_eval_tokens": 11838 + "serialized_chars": 23306, + "estimated_input_tokens": 5827, + "estimated_eval_tokens": 11910 }, "budget": { "node_count": 8, "item_count": 8, "soft_pair_count": 14, - "serialized_chars": 23161, - "estimated_input_tokens": 5791, - "estimated_eval_tokens": 11838 + "serialized_chars": 23306, + "estimated_input_tokens": 5827, + "estimated_eval_tokens": 11910 }, "overflow_policy": "truncate_then_skip", "trimmed": false, "aggressively_trimmed": false, "split": false, "analyst_result": { - "summary": "These PRs are mostly unrelated despite superficial similarity in titles like \"fix\". I do not see a true duplicate cluster; the overlap is mostly broad test/code cleanup or generic bug-fix wording.", - "confidence": 0.93, + "summary": "This cluster looks like a noisy similarity bucket, not a duplicate set: the PRs cover unrelated fixes in Whisper tests, tie-weights backward compatibility, model test placeholders, FP8 dequant logic, doctest cleanup, generation watermarking validation, CI workflow tweaks, and typo-only edits.", + "confidence": 0.95, "canonical_issue_reason": null, - "canonical_pr_reason": "PR #46579 is the best representative only in a loose sense: it is the broadest code cleanup/typo-fix change and the most generic among the set, whereas the others are narrow fixes in unrelated areas.", + "canonical_pr_reason": "No true canonical duplicate stands out; PR 46579 is the broadest and most central maintenance-style change in the cluster, but it does not share the same concrete bug/change with the others.", "best_issue_reason": null, - "best_pr_reason": "PR #46579 is the strongest standalone PR here because it spans many code files and appears to be a coherent cross-cutting cleanup, unlike the other narrowly scoped and unrelated fixes.", + "best_pr_reason": "PR 46579 is the best anchor for triage because it is the most broadly connected artifact in this otherwise heterogeneous cluster, though it is not a real duplicate of the others.", "soft_edge_verdicts": [ { "left": "pull_request:43291", "right": "pull_request:43665", "accept": false, - "reason": "Whisper tokenizer/test changes are unrelated to CLIP/ViT test edits; same vague \"fix\" theme only." + "reason": "Different areas and intents: Whisper tokenizer/test cleanup vs placeholder asserts in CLIP/ViT tests." }, { "left": "pull_request:43291", "right": "pull_request:44815", "accept": false, - "reason": "Whisper test/tokenizer work is unrelated to FP8 dequant/loading logic." + "reason": "Whisper decoding logic/tests are unrelated to FP8 dequant/model-loading fixes." }, { "left": "pull_request:43291", "right": "pull_request:46380", "accept": false, - "reason": "Whisper test/tokenizer work is unrelated to generation watermarking/config validation." + "reason": "Whisper test/tokenization change and generation watermarking validation are different code paths and bugs." }, { "left": "pull_request:43498", "right": "pull_request:46579", "accept": false, - "reason": "Backward-compatibility aliasing for tie_weights is not the same change as broad typo cleanup across many files." + "reason": "Backward-compatibility shim for tie_weights is unrelated to a broad typo-only cleanup." }, { "left": "pull_request:43665", "right": "pull_request:46539", "accept": false, - "reason": "Model test edits for CLIP/ViT are unrelated to Phimoe CI/integration-test infrastructure changes." + "reason": "Dummy/failing model-test edits are unrelated to Phimoe CI/workflow fixes." }, { "left": "pull_request:43665", "right": "pull_request:44815", "accept": false, - "reason": "Model test assertions are unrelated to dequantization/model-loading fixes." + "reason": "Model-test placeholder assertions do not match the FP8 dequantization/runtime fix in 44815." }, { "left": "pull_request:43665", "right": "pull_request:46380", "accept": false, - "reason": "CLIP/ViT test changes are unrelated to generation config watermarking validation." + "reason": "These touch different problems: test-file edits vs generation config/Whisper validation changes." }, { "left": "pull_request:43665", "right": "pull_request:46579", "accept": false, - "reason": "Placeholder test edits are not the same underlying change as broad code typo fixes." + "reason": "Test placeholder changes are not the same underlying change as code typo cleanup across many files." }, { "left": "pull_request:44815", "right": "pull_request:46539", "accept": false, - "reason": "FP8 dequant/model-loading changes are unrelated to Phimoe integration-test/CI adjustments." + "reason": "FP8 dequant/model-loading logic is unrelated to Phimoe integration test CI adjustments." }, { "left": "pull_request:44815", "right": "pull_request:46380", "accept": false, - "reason": "Dequantization kernel/loading fixes are unrelated to generation watermarking validation." + "reason": "Both are code changes, but they fix different concrete bugs in different subsystems." }, { "left": "pull_request:44815", "right": "pull_request:46579", "accept": false, - "reason": "Shared file overlap is incidental; one is an FP8 dequant fix, the other is a broad typo cleanup." + "reason": "FP8 kernel/loading fixes are not the same as general typo corrections." }, { "left": "pull_request:45114", "right": "pull_request:46579", "accept": false, - "reason": "Doctest documentation fixes are unrelated to code typo cleanup." + "reason": "Doctest/documentation edits do not share the same underlying change as code typo cleanup." }, { "left": "pull_request:46380", "right": "pull_request:46539", "accept": false, - "reason": "Generation watermarking validation changes are unrelated to Phimoe CI/integration-test fixes." + "reason": "Generation watermarking validation and Phimoe CI workflow changes are unrelated." }, { "left": "pull_request:46380", "right": "pull_request:46579", "accept": false, - "reason": "Generation config validation fixes are not the same as generic typo edits across code files." + "reason": "A generation/config runtime fix is not the same as broad typo-only edits across many files." } ] }, @@ -20581,8 +20382,8 @@ }, { "label": "pull_request", - "component_index": 9, - "component_count": 39, + "component_index": 8, + "component_count": 35, "review_unit_index": 1, "review_unit_count": 1, "status": "reviewed", @@ -20598,7 +20399,7 @@ "soft_pairs": [ "pull_request:43488|pull_request:43532" ], - "prepared_review_unit_hash": "f393f7b122c41e1d66375792477cda0cc7473db6ec82f8218788b74b4dfee810", + "prepared_review_unit_hash": "7416ef1d8bb197e54bafede1da61fada4615f5a143cec955c85edab5949297c7", "component_budget": { "node_count": 2, "item_count": 2, @@ -20620,18 +20421,18 @@ "aggressively_trimmed": false, "split": false, "analyst_result": { - "summary": "The two PRs are unrelated and should not be clustered as duplicates: one is a tiny setup/version-format bot check, the other is a broader workflow/test diff showcase. The only shared file is incidental.", - "confidence": 0.97, + "summary": "Two unrelated \"don't merge\" demo PRs with overlapping test-file touch but different code paths and scopes; they should not be deduplicated.", + "confidence": 0.63, "canonical_issue_reason": null, - "canonical_pr_reason": "No single canonical PR: 43488 is a setup/version-format check, while 43532 targets CI/workflow and testing utilities; they do not represent the same change.", + "canonical_pr_reason": "PR #43488 is the cleaner representative: its title and diff are narrowly centered on a repo-bot/setup.py format check, making it more self-contained than the broader workflow-heavy PR #43532.", "best_issue_reason": null, - "best_pr_reason": "43532 is the better representative only in the sense that it is the larger, more informative PR; however, it is not a good global representative for the cluster because the artifacts are not duplicates.", + "best_pr_reason": "PR #43488 is the best cluster representative because it is more focused and interpretable as a single change, whereas #43532 spans several workflows and utilities and is less canonical as a duplicate target.", "soft_edge_verdicts": [ { "left": "pull_request:43488", "right": "pull_request:43532", "accept": false, - "reason": "Reject: the PRs touch different code paths and purposes. 43488 changes setup.py/version/dependencies for a repo-bot check, while 43532 edits CI workflows, testing utilities, and tests. Shared test-file overlap is incidental, not the same underlying change." + "reason": "They only share a test filename; #43488 tweaks packaging/versioning and a vit test, while #43532 changes CI workflows, testing utilities, and notifications. They are not the same underlying fix/change." } ] }, @@ -20644,8 +20445,8 @@ }, { "label": "pull_request", - "component_index": 10, - "component_count": 39, + "component_index": 9, + "component_count": 35, "review_unit_index": 1, "review_unit_count": 1, "status": "reviewed", @@ -20661,7 +20462,7 @@ "soft_pairs": [ "pull_request:43612|pull_request:43613" ], - "prepared_review_unit_hash": "ca9e574c16cec23046f4e251c42d52ac3811b11e6bb5309d62fb516b136df412", + "prepared_review_unit_hash": "a63d71ae599348543106c7dff3fff3c3b73b0317823ded946d7839848b1e947e", "component_budget": { "node_count": 2, "item_count": 2, @@ -20683,18 +20484,18 @@ "aggressively_trimmed": false, "split": false, "analyst_result": { - "summary": "These are closely related pipeline-adding PRs, but they implement different features: promptable concept segmentation vs promptable visual segmentation, with different docs, pipeline modules, tests, and model integrations. They are not duplicates and should not be merged as one PR.", - "confidence": 0.96, + "summary": "Two open PRs add similarly named but distinct pipeline/features: one for promptable concept segmentation and one for promptable visual segmentation. They share boilerplate documentation/registry changes, but they do not appear to fix the same concrete code-path problem.", + "confidence": 0.91, "canonical_issue_reason": null, - "canonical_pr_reason": "PR #43613 is the better canonical PR only in the sense that it is the broader, more model-inclusive feature; however, it is not a duplicate of #43612 because it targets a different segmentation pipeline and different model family support.", + "canonical_pr_reason": null, "best_issue_reason": null, - "best_pr_reason": "PR #43613 is the strongest standalone PR for the cluster because it covers the broader promptable visual segmentation feature and includes more integration points and tests. PR #43612 is a separate, narrower promptable concept segmentation pipeline.", + "best_pr_reason": "PR #43613 is the stronger standalone change: it covers a broader, more integrated visual-segmentation stack (multiple model tests and a dedicated pipeline implementation), making it the more complete representative of this area.", "soft_edge_verdicts": [ { "left": "pull_request:43612", "right": "pull_request:43613", "accept": false, - "reason": "Different underlying changes: one adds PromptableConceptSegmentationPipeline, the other adds PromptableVisualSegmentationPipeline. They touch different docs, pipeline modules, tests, and model support, so they are related but not the same code-path fix/change." + "reason": "Different feature targets and underlying code paths: concept segmentation vs visual segmentation. The overlap is mostly shared pipeline/docs scaffolding, not the same concrete bug or change, so they should not be merged as duplicates." } ] }, @@ -20707,8 +20508,8 @@ }, { "label": "pull_request", - "component_index": 11, - "component_count": 39, + "component_index": 10, + "component_count": 35, "review_unit_index": 1, "review_unit_count": 1, "status": "reviewed", @@ -20724,7 +20525,7 @@ "soft_pairs": [ "pull_request:43651|pull_request:43663" ], - "prepared_review_unit_hash": "9b3e9f41e8e175ed039ffc69dd7f8ba76467ec1bd2a881c3b4bdb83e0f510ef2", + "prepared_review_unit_hash": "24adb5457d3e2df840c11edf572384006b5ad05262920e424efe11c078c0db6a", "component_budget": { "node_count": 2, "item_count": 2, @@ -20746,18 +20547,18 @@ "aggressively_trimmed": false, "split": false, "analyst_result": { - "summary": "Both PRs touch `trainer.py` and are about custom trainer extensibility, but they modify different code paths: one changes gradient-accumulation loss scaling, the other factors out signature-column filtering. They are not the same underlying change.", - "confidence": 0.94, + "summary": "Both pull requests touch `trainer.py` and improve trainer extensibility, but they fix different code paths: one changes gradient-accumulation loss scaling, the other extracts signature-column filtering into an overrideable helper. They are related in theme, not duplicates.", + "confidence": 0.96, "canonical_issue_reason": null, - "canonical_pr_reason": "No duplicate PR here; `#43651` is the more concrete training-step fix, but `#43663` is a separate refactor for column filtering.", + "canonical_pr_reason": "No canonical PR: 43651 and 43663 address separate trainer behaviors and would not be merged into one change.", "best_issue_reason": null, - "best_pr_reason": "`#43651` is the better representative of the cluster because it is a self-contained behavioral fix in `training_step`; `#43663` addresses a different trainer override hook.", + "best_pr_reason": "PR 43663 is a clean, self-contained trainer refactor, but it is only a representative of the cluster theme\u2014not a duplicate of 43651.", "soft_edge_verdicts": [ { "left": "pull_request:43651", "right": "pull_request:43663", "accept": false, - "reason": "Different trainer behaviors: loss scaling vs. signature-column filtering. Shared file and custom-trainer theme are not enough to treat them as the same fix." + "reason": "Different underlying changes in the same file: 43651 alters loss scaling during training, while 43663 adds an overrideable signature-column helper for dataset filtering. Shared subsystem alone is not enough to treat them as the same fix." } ] }, @@ -20770,8 +20571,8 @@ }, { "label": "pull_request", - "component_index": 12, - "component_count": 39, + "component_index": 11, + "component_count": 35, "review_unit_index": 1, "review_unit_count": 1, "status": "reviewed", @@ -20787,7 +20588,7 @@ "soft_pairs": [ "pull_request:43656|pull_request:43842" ], - "prepared_review_unit_hash": "ec05e31342ec1f204772faee19fd33f07b800f844da7e0f6aa0059f54c3116e0", + "prepared_review_unit_hash": "40e16068567901536c4c15a1c5ab165309060f08f2de6fb3c0ddde6e79fd555d", "component_budget": { "node_count": 2, "item_count": 2, @@ -20809,35 +20610,35 @@ "aggressively_trimmed": false, "split": false, "analyst_result": { - "summary": "Both PRs target the same import-time TypeAdapter NameError in `src/transformers/cli/serve.py`; they differ mainly in implementation style (`__future__`/typing cleanup vs `TYPE_CHECKING`-guarded import), not in the underlying code-path problem.", - "confidence": 0.79, + "summary": "The two PRs both touch `src/transformers/cli/serve.py` and both involve typing changes, but the packet does not show they fix the same concrete bug. PR 43842 is a targeted `TypeAdapter`/missing-pydantic guard, while 43656 is a broader `from __future__ import annotations` and annotation cleanup.", + "confidence": 0.81, "canonical_issue_reason": null, - "canonical_pr_reason": "PR #43842 is the tighter, more targeted fix for the same CLI `TypeAdapter` NameError, with less unrelated typing churn than #43656.", + "canonical_pr_reason": "PR 43842 is the better canonical PR because it addresses a specific runtime problem: guarding `TypeAdapter` when pydantic is absent and fixing the failing annotation site directly.", "best_issue_reason": null, - "best_pr_reason": "PR #43842 is the cleanest standalone representation of the underlying bug: it avoids importing `TypeAdapter` at runtime and directly addresses the optional-pydantic NameError.", + "best_pr_reason": "PR 43842 is the best representative for the cluster because it is the narrower, more concrete fix and the explicit issue target is specific; 43656 is broader typing churn and not clearly the same bug.", "soft_edge_verdicts": [ { "left": "pull_request:43656", "right": "pull_request:43842", - "accept": true, - "reason": "Same file and same concrete failure mode (`TypeAdapter` NameError during CLI serve import); the diffs are alternative implementations of the same fix rather than distinct changes." + "accept": false, + "reason": "Reject: the overlap is only in the same file and general typing/annotation changes. The packet does not establish the same underlying bug or a mergeable single fix, and the explicit issue targets differ." } ] }, "evaluator_result": { - "accept": true, - "feedback": "The summary is grounded in the packet: both PRs touch the same file and are explicitly about the same `TypeAdapter` CLI NameError, with different implementation approaches shown in the diffs. The pair also looks mergeable as alternative fixes for the same bug, so the soft-edge acceptance is reasonable." + "accept": false, + "feedback": "The cluster summary overstates the evidence. The packet only supports that both PRs make typing/annotation-related changes in `src/transformers/cli/serve.py`; it does not clearly show they fix the same bug, and the explicit issue targets differ (`43576` vs `43824`). PR 43842 is specifically about guarding `TypeAdapter` for missing pydantic, while 43656 appears to be a broader `from __future__ import annotations`/typing cleanup. The soft-pair acceptance is too strong without clearer same-bug linkage or mergeability evidence." }, "evaluator_used": true, - "retried": false, - "accepted_nontrivial_soft_edge": true, + "retried": true, + "accepted_nontrivial_soft_edge": false, "error_kind": null, "error_message": null }, { "label": "pull_request", - "component_index": 13, - "component_count": 39, + "component_index": 12, + "component_count": 35, "review_unit_index": 1, "review_unit_count": 1, "status": "reviewed", @@ -20902,8 +20703,8 @@ }, { "label": "pull_request", - "component_index": 14, - "component_count": 39, + "component_index": 13, + "component_count": 35, "review_unit_index": 1, "review_unit_count": 2, "status": "reviewed", @@ -20921,7 +20722,6 @@ "pull_request:44017", "pull_request:44018", "pull_request:44019", - "pull_request:44024", "pull_request:44044", "pull_request:44056", "pull_request:44059", @@ -20974,13 +20774,13 @@ "pull_request:44018|pull_request:44071", "pull_request:44059|pull_request:44722", "pull_request:44017|pull_request:44161", - "pull_request:44017|pull_request:44073", "pull_request:44017|pull_request:44722", + "pull_request:44017|pull_request:44073", "pull_request:44068|pull_request:44071", "pull_request:43997|pull_request:44073", "pull_request:44071|pull_request:44086", - "pull_request:44073|pull_request:44101", "pull_request:44017|pull_request:44071", + "pull_request:44073|pull_request:44101", "pull_request:44073|pull_request:44161", "pull_request:44056|pull_request:44101", "pull_request:44059|pull_request:44071", @@ -20996,9 +20796,10 @@ "pull_request:43997|pull_request:44161", "pull_request:44074|pull_request:44161", "pull_request:44018|pull_request:44161", + "pull_request:43997|pull_request:44071", "pull_request:44010|pull_request:44071", - "pull_request:44010|pull_request:44073", "pull_request:44018|pull_request:44129", + "pull_request:44010|pull_request:44073", "pull_request:43997|pull_request:44074", "pull_request:43997|pull_request:44129", "pull_request:44010|pull_request:44018", @@ -21013,300 +20814,335 @@ "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" + "pull_request:44044|pull_request:44154" ], - "prepared_review_unit_hash": "71bc40cec62cb0454aca0d468ce19b160b63b9582c3960181a5acb2f5e5753b8", + "prepared_review_unit_hash": "198d4dd5455eae6da978e1a512f28eac1adce076e9549999baf653435bb1d08e", "component_budget": { "node_count": 28, "item_count": 28, - "soft_pair_count": 77, - "serialized_chars": 84509, - "estimated_input_tokens": 21128, - "estimated_eval_tokens": 42512 + "soft_pair_count": 78, + "serialized_chars": 84902, + "estimated_input_tokens": 21226, + "estimated_eval_tokens": 42708 }, "budget": { - "node_count": 28, - "item_count": 28, + "node_count": 27, + "item_count": 27, "soft_pair_count": 72, - "serialized_chars": 76941, - "estimated_input_tokens": 19236, - "estimated_eval_tokens": 38728 + "serialized_chars": 75096, + "estimated_input_tokens": 18774, + "estimated_eval_tokens": 37804 }, "overflow_policy": "truncate_then_skip", "trimmed": false, "aggressively_trimmed": false, "split": true, "analyst_result": { - "summary": "Mostly independent per-model output-tracing refactors under the same umbrella issue. The only plausible duplicate-like overlap is the GPT-J/CodeGen line centered on 44722; the GPT-Neo pair and the rest are too model-specific to collapse into shared duplicates.", - "confidence": 0.88, + "summary": "This cluster is mostly a set of model-specific refactors around standardized output tracing. Most PRs share the same high-level theme but touch different model code paths, so they are not duplicates. The only strong duplicate-like pairs are the GPT-J iterations and the GPT-Neo iterations.", + "confidence": 0.78, "canonical_issue_reason": null, - "canonical_pr_reason": "Best anchor for the narrow GPT-J/CodeGen overlap: explicit #43979 linkage plus edits in both GPT-J and CodeGen files, making it the clearest representative of that small cluster.", + "canonical_pr_reason": "PR #44722 is the best representative because it is the most complete GPT-J/CodeGen output-tracing refactor in the set, targets the shared issue, and appears to subsume earlier GPT-J-specific iterations.", "best_issue_reason": null, - "best_pr_reason": "Most representative PR in the packet because it spans both GPT-J and CodeGen, carries the explicit issue link, and sits at the center of the only clearly related overlap.", + "best_pr_reason": "PR #44722 is the strongest single PR candidate for the cluster: it covers the same concrete refactor as the earlier GPT-J/CodeGen work and is the broadest, latest iteration among the close matches.", "soft_edge_verdicts": [ { "left": "pull_request:43996", "right": "pull_request:44085", "accept": false, - "reason": "Different model families and code paths; same broad tracing theme, but not the same concrete change." + "reason": "Different model families and code paths (CVT/FNet vs GPT-J); same theme, but not the same underlying change." }, { "left": "pull_request:43996", "right": "pull_request:44044", "accept": false, - "reason": "Distinct backbones; similar output-tracing refactor style only." + "reason": "CVT/FNet output-tracing refactor vs DeBERTa output-tracing refactor; separate model implementations." }, { "left": "pull_request:44066", "right": "pull_request:44085", - "accept": false, - "reason": "Same umbrella issue, but the file-level edits differ enough that this does not look like one mergeable PR." + "accept": true, + "reason": "Both are GPT-J output-tracing refactors in the same file/function; one is a narrower version of the same underlying change." }, { "left": "pull_request:44007", "right": "pull_request:44072", "accept": false, - "reason": "Different model code paths; same decorator migration theme but not the same change." + "reason": "RegNet and EfficientNet are different models with separate encoder implementations; same pattern, not the same bug/change." }, { "left": "pull_request:44072", "right": "pull_request:44722", "accept": false, - "reason": "EfficientNet vs GPT-J/CodeGen; unrelated concrete code paths." + "reason": "EfficientNet vs GPT-J/CodeGen are unrelated code paths, so they should not be merged as one duplicate PR." }, { "left": "pull_request:44066", "right": "pull_request:44072", "accept": false, - "reason": "GPT-J/CodeGen vs EfficientNet; only the high-level tracing initiative matches." + "reason": "GPT-J/CodeGen vs EfficientNet are different model refactors, not one concrete change." }, { "left": "pull_request:44013", "right": "pull_request:44044", "accept": false, - "reason": "MobileNetV2 and DeBERTa-v2 are unrelated implementation paths." + "reason": "MobileNetV2 and DeBERTa output tracing are separate model-specific refactors." }, { "left": "pull_request:44066", "right": "pull_request:44086", "accept": false, - "reason": "GPT-J/CodeGen vs MGP-STR; same umbrella pattern, different model code." + "reason": "GPT-J/CodeGen output tracing and MGP-STR output tracing are different model code paths." }, { "left": "pull_request:44066", "right": "pull_request:44071", "accept": false, - "reason": "GPT-J/CodeGen and MPT are different concrete refactors." + "reason": "GPT-J/CodeGen and MPT touch different architectures and different forward paths." }, { "left": "pull_request:44018", "right": "pull_request:44068", - "accept": false, - "reason": "GPT-Neo output-tracing refactors, but the packet does not support treating these as the same concrete change." + "accept": true, + "reason": "Both are GPT-Neo output-tracing refactors in the same model file, with overlapping test/code changes; they look like iterations of the same PR." }, { "left": "pull_request:44066", "right": "pull_request:44068", "accept": false, - "reason": "GPT-J/CodeGen vs GPT-Neo; similar topic, different model-specific changes." + "reason": "GPT-J/CodeGen and GPT-Neo are different model implementations, so these are not duplicates." }, { "left": "pull_request:44018", "right": "pull_request:44066", "accept": false, - "reason": "Different model families and files; not the same underlying bug/change." + "reason": "GPT-Neo vs GPT-J/CodeGen are separate model refactors even though they use the same standardized pattern." }, { "left": "pull_request:44019", "right": "pull_request:44722", "accept": false, - "reason": "ResNet vs GPT-J/CodeGen; unrelated concrete implementations." + "reason": "ResNet output tracing and GPT-J/CodeGen output tracing are unrelated model changes." }, { "left": "pull_request:44019", "right": "pull_request:44085", "accept": false, - "reason": "ResNet and GPT-J are different code paths." + "reason": "ResNet and GPT-J are different model families; same broad refactor theme is not enough to merge them." }, { "left": "pull_request:44086", "right": "pull_request:44722", "accept": false, - "reason": "MGP-STR vs GPT-J/CodeGen; same refactor pattern only." + "reason": "MGP-STR vs GPT-J/CodeGen are different code paths and should not be treated as the same change." }, { "left": "pull_request:44018", "right": "pull_request:44019", "accept": false, - "reason": "Different model backbones; not a duplicate change." + "reason": "GPT-Neo and ResNet are unrelated model-specific output-tracing refactors." }, { "left": "pull_request:44019", "right": "pull_request:44071", "accept": false, - "reason": "ResNet and MPT are separate concrete refactors." + "reason": "ResNet and MPT are different architectures with distinct implementations." }, { "left": "pull_request:44068", "right": "pull_request:44116", "accept": false, - "reason": "GPT-Neo vs Flaubert; same standardized-output theme, but different model code." + "reason": "GPT-Neo and Flaubert/XLM are different model code paths." }, { "left": "pull_request:44018", "right": "pull_request:44116", "accept": false, - "reason": "GPT-Neo and Flaubert are unrelated model implementations." + "reason": "GPT-Neo and Flaubert are separate refactors in unrelated models." }, { "left": "pull_request:44068", "right": "pull_request:44722", "accept": false, - "reason": "GPT-Neo vs GPT-J/CodeGen; not the same underlying change." + "reason": "GPT-Neo and GPT-J/CodeGen are different model implementations, so this is not one duplicate PR." }, { "left": "pull_request:44071", "right": "pull_request:44722", "accept": false, - "reason": "MPT and GPT-J/CodeGen are distinct code paths." + "reason": "MPT and GPT-J/CodeGen are different code paths despite similar output-tracing goals." }, { "left": "pull_request:44018", "right": "pull_request:44085", "accept": false, - "reason": "GPT-Neo vs GPT-J; same general tracing effort, but different model-specific refactors." + "reason": "GPT-Neo and GPT-J are different models; not the same underlying bug/change." + }, + { + "left": "pull_request:44018", + "right": "pull_request:44086", + "accept": false, + "reason": "GPT-Neo vs MGP-STR are unrelated model-specific refactors." + }, + { + "left": "pull_request:44068", + "right": "pull_request:44086", + "accept": false, + "reason": "GPT-Neo and MGP-STR do not share the same concrete code path." + }, + { + "left": "pull_request:44071", + "right": "pull_request:44085", + "accept": false, + "reason": "MPT and GPT-J are different model families." }, { "left": "pull_request:44059", "right": "pull_request:44086", "accept": false, - "reason": "GPT-2 and MGP-STR are unrelated concrete changes." + "reason": "GPT-2 and MGP-STR are distinct model implementations; same refactor pattern only." }, { "left": "pull_request:44085", "right": "pull_request:44086", "accept": false, - "reason": "GPT-J vs MGP-STR; no shared code path." + "reason": "GPT-J and MGP-STR are different model code paths." }, { "left": "pull_request:44098", "right": "pull_request:44101", "accept": false, - "reason": "ViLT and Flaubert/XLM are different model families." + "reason": "ViLT and Flaubert/XLM are separate model refactors." }, { - "left": "pull_request:44017", - "right": "pull_request:44161", + "left": "pull_request:44073", + "right": "pull_request:44101", + "accept": false, + "reason": "VisualBert and Flaubert are different models, so these are not duplicates." + }, + { + "left": "pull_request:44056", + "right": "pull_request:44101", + "accept": false, + "reason": "MPNet and Flaubert/XLM touch unrelated model implementations." + }, + { + "left": "pull_request:44059", + "right": "pull_request:44071", "accept": false, - "reason": "Segformer and LongT5 are unrelated implementations." + "reason": "GPT-2 and MPT are different architectures and different forward paths." }, { "left": "pull_request:43997", "right": "pull_request:44017", "accept": false, - "reason": "RegNet and Segformer are different model families." + "reason": "RegNet and Segformer are different model code paths." }, { "left": "pull_request:43997", "right": "pull_request:44056", "accept": false, - "reason": "RegNet and MPNet are unrelated concrete changes." + "reason": "RegNet and MPNet are unrelated model-specific changes." }, { "left": "pull_request:44071", - "right": "pull_request:44129", + "right": "pull_request:44076", "accept": false, - "reason": "MPT and SpeechT5 are different code paths." + "reason": "MPT and ImageGPT are different models with separate implementations." }, { - "left": "pull_request:43996", + "left": "pull_request:44074", + "right": "pull_request:44101", + "accept": false, + "reason": "TextNet and Flaubert/XLM are different model code paths." + }, + { + "left": "pull_request:44056", "right": "pull_request:44161", "accept": false, - "reason": "FNet/CVT and LongT5 are unrelated implementations." + "reason": "MPNet and LongT5 are unrelated refactors across different models." }, { - "left": "pull_request:43996", - "right": "pull_request:44017", + "left": "pull_request:43997", + "right": "pull_request:44074", "accept": false, - "reason": "FNet/CVT vs Segformer; only the high-level refactor theme matches." + "reason": "RegNet and TextNet are different architectures, so this is not one duplicate change." }, { "left": "pull_request:44010", - "right": "pull_request:44044", + "right": "pull_request:44071", "accept": false, - "reason": "SqueezeBert and DeBERTa-v2 are different code paths." + "reason": "SqueezeBert and MPT are separate model implementations." }, { - "left": "pull_request:44044", - "right": "pull_request:44074", + "left": "pull_request:44010", + "right": "pull_request:44073", "accept": false, - "reason": "DeBERTa-v2 and TextNet are separate model-specific refactors." + "reason": "SqueezeBert and VisualBert are different code paths." }, { - "left": "pull_request:44066", - "right": "pull_request:44722", - "accept": true, - "reason": "Same GPT-J/CodeGen output-tracing refactor family with overlapping file-level changes; plausible duplicate implementations." + "left": "pull_request:44074", + "right": "pull_request:44161", + "accept": false, + "reason": "TextNet and LongT5 are unrelated model-specific refactors." }, { - "left": "pull_request:44085", - "right": "pull_request:44722", - "accept": true, - "reason": "Same GPT-J output-tracing code path and decorator migration; the PRs look like overlapping variants of one change." + "left": "pull_request:43996", + "right": "pull_request:44007", + "accept": false, + "reason": "CVT/FNet and RegNet are different models; same output-tracing theme is too broad to merge." }, { - "left": "pull_request:44085", - "right": "pull_request:44129", + "left": "pull_request:43996", + "right": "pull_request:44072", "accept": false, - "reason": "GPT-J and SpeechT5 are unrelated model implementations." + "reason": "CVT/FNet and EfficientNet are separate model refactors." }, { "left": "pull_request:43996", - "right": "pull_request:44007", + "right": "pull_request:44017", "accept": false, - "reason": "FNet/CVT and RegNet/ResNet/RT-DETR are different concrete code paths." + "reason": "CVT/FNet and Segformer are different model code paths." }, { - "left": "pull_request:43996", + "left": "pull_request:44085", "right": "pull_request:44129", "accept": false, - "reason": "FNet/CVT vs SpeechT5; not the same change." + "reason": "GPT-J and SpeechT5 are different models and changes." + }, + { + "left": "pull_request:44066", + "right": "pull_request:44722", + "accept": true, + "reason": "Both are the same GPT-J/CodeGen standardized output-tracing refactor, with the same concrete code paths and issue target." }, { "left": "pull_request:44044", "right": "pull_request:44129", "accept": false, - "reason": "DeBERTa-v2 and SpeechT5 are unrelated models." + "reason": "DeBERTa and SpeechT5 are unrelated model implementations." }, { "left": "pull_request:44044", "right": "pull_request:44154", "accept": false, - "reason": "DeBERTa-v2 and ViLT are different concrete refactors." - }, - { - "left": "pull_request:44024", - "right": "pull_request:44076", - "accept": false, - "reason": "FocalNet and ImageGPT are distinct model-specific changes despite similar output-capture plumbing." + "reason": "DeBERTa and VITS are different model code paths, so they are not duplicates." } ] }, "evaluator_result": { - "accept": false, - "feedback": "The summary is a bit too strong for the packet. It overstates two duplicate subclusters: (1) the GPT-J/CodeGen set is not a full trio because 44066\u219444085 is marked false in the packet, even though both connect to 44722; and (2) the GPT-Neo pair 44018\u219444068 is also marked false in the packet. Those accepted soft edges look like overreach relative to the provided evidence. The rest of the false edges are conservative, but the positive cluster claims need to be toned down." + "accept": true, + "feedback": "The summary is grounded in the packet and stays conservative about cross-model pairs. The two accepted duplicate-like pairs (GPT-J iterations and GPT-Neo iterations) look mergeable into single PRs for the same refactor/bugfix, and the chosen canonical PR (44722) is supported by the broader GPT-J/CodeGen scope plus the shared issue target. No clear overstatement beyond the evidence provided." }, "evaluator_used": true, - "retried": true, - "accepted_nontrivial_soft_edge": false, + "retried": false, + "accepted_nontrivial_soft_edge": true, "error_kind": null, "error_message": null }, { "label": "pull_request", - "component_index": 14, - "component_count": 39, + "component_index": 13, + "component_count": 35, "review_unit_index": 2, "review_unit_count": 2, "status": "reviewed", @@ -21318,76 +21154,85 @@ "nodes": [ "pull_request:43997", "pull_request:44010", + "pull_request:44024", "pull_request:44056", "pull_request:44073", "pull_request:44074", + "pull_request:44076", "pull_request:44129" ], "soft_pairs": [ + "pull_request:44024|pull_request:44076", "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": "000bfa8be1162b6be794517cb9043f444ec15b52ec836a5375dde36e27135b83", + "prepared_review_unit_hash": "799805cb8a35ed347f3c6174e20f4fd04323c90029b598122686f635398a0892", "component_budget": { "node_count": 28, "item_count": 28, - "soft_pair_count": 77, - "serialized_chars": 84509, - "estimated_input_tokens": 21128, - "estimated_eval_tokens": 42512 + "soft_pair_count": 78, + "serialized_chars": 84902, + "estimated_input_tokens": 21226, + "estimated_eval_tokens": 42708 }, "budget": { - "node_count": 6, - "item_count": 6, - "soft_pair_count": 5, - "serialized_chars": 13068, - "estimated_input_tokens": 3267, - "estimated_eval_tokens": 6790 + "node_count": 8, + "item_count": 8, + "soft_pair_count": 6, + "serialized_chars": 17127, + "estimated_input_tokens": 4282, + "estimated_eval_tokens": 8820 }, "overflow_policy": "truncate_then_skip", "trimmed": false, "aggressively_trimmed": false, "split": true, "analyst_result": { - "summary": "These PRs are all model-specific refactors to migrate output tracing to the new standardized capture_outputs/can_return_tuple pattern, but they touch different architectures and code paths, so they are related by theme rather than duplicates.", - "confidence": 0.88, + "summary": "This cluster is a set of model-specific refactors to migrate different Transformers models to standardized output capture/tracing. They are thematically similar but touch different architectures and code paths, so they should not be treated as duplicates.", + "confidence": 0.92, "canonical_issue_reason": null, - "canonical_pr_reason": "PR 44129 is the best canonical representative because it has the clearest explicit issue linkage (43979) and a concrete, self-contained SpeechT5 output-tracing refactor.", + "canonical_pr_reason": "If one PR has to represent the cluster, 44129 is the most concrete and self-contained: it has an explicit issue target, a narrowly scoped SpeechT5 refactor, and clear standardized output-capture changes.", "best_issue_reason": null, - "best_pr_reason": "PR 44129 is the strongest standalone representative for the cluster because it is explicit about the target problem and has the most complete change description among the set.", + "best_pr_reason": "44129 is the strongest standalone PR in this set because it is explicitly tied to an issue and has a focused, model-specific implementation rather than a broad or ambiguous change.", "soft_edge_verdicts": [ + { + "left": "pull_request:44024", + "right": "pull_request:44076", + "accept": false, + "reason": "Same general migration theme, but FocalNet and ImageGPT are different models with different forward paths; not the same underlying change." + }, { "left": "pull_request:44073", "right": "pull_request:44074", "accept": false, - "reason": "Same refactor theme, but VisualBert and TextNet are different models with different forward/output code paths; not one mergeable duplicate PR." + "reason": "Both refactor output tracing, but VisualBert and TextNet are separate model implementations and these patches are not interchangeable." }, { "left": "pull_request:44056", "right": "pull_request:44073", "accept": false, - "reason": "Both migrate output tracing, but they apply to different model implementations (MPNet vs VisualBert) and are not the same concrete change." + "reason": "MPNet vs VisualBert: different code paths and model-specific tracing hooks, so this is not a duplicate fix/change." }, { "left": "pull_request:44056", "right": "pull_request:44074", "accept": false, - "reason": "Different model-specific refactors with no shared code path; they only share the generalized output-tracing migration pattern." + "reason": "MPNet and TextNet are distinct model families; they share a refactor pattern but not the same concrete bug or PR." }, { "left": "pull_request:43997", "right": "pull_request:44010", "accept": false, - "reason": "RegNet and SqueezeBert are unrelated model files; same broad subsystem, but not the same underlying bug or change." + "reason": "RegNet and SqueezeBert are unrelated implementations; these are separate standardized-output migrations, not one mergeable PR." }, { "left": "pull_request:44010", "right": "pull_request:44129", "accept": false, - "reason": "These are separate model migrations (SqueezeBert vs SpeechT5) and should not be merged just because both adopt standardized output capture." + "reason": "SqueezeBert and SpeechT5 change different model code paths; the explicit issue link on 44129 does not make them the same change." } ] }, @@ -21400,8 +21245,8 @@ }, { "label": "pull_request", - "component_index": 15, - "component_count": 39, + "component_index": 14, + "component_count": 35, "review_unit_index": 1, "review_unit_count": 1, "status": "reviewed", @@ -21467,274 +21312,268 @@ "pull_request:44028|pull_request:44030", "pull_request:44029|pull_request:44030" ], - "prepared_review_unit_hash": "b782142823a2550613e0935b86479d4d4e65bcd3c7f62cde48b814a96cbaf1ae", + "prepared_review_unit_hash": "a76970829c3bfb42dafbf0e139680ed56cf7c704d2b76a588ea24c395eeb4dea", "component_budget": { "node_count": 13, "item_count": 13, "soft_pair_count": 40, - "serialized_chars": 40176, - "estimated_input_tokens": 10044, - "estimated_eval_tokens": 20344 + "serialized_chars": 40180, + "estimated_input_tokens": 10045, + "estimated_eval_tokens": 20346 }, "budget": { "node_count": 13, "item_count": 13, "soft_pair_count": 40, - "serialized_chars": 40176, - "estimated_input_tokens": 10044, - "estimated_eval_tokens": 20344 + "serialized_chars": 40180, + "estimated_input_tokens": 10045, + "estimated_eval_tokens": 20346 }, "overflow_policy": "truncate_then_skip", "trimmed": false, "aggressively_trimmed": false, "split": false, "analyst_result": { - "summary": "These PRs are mostly model-specific output-tracing refactors. They share a template and utility changes, but they do not fix the same underlying bug or change the same concrete code path, so they should not be deduplicated.", + "summary": "These are not duplicates of one underlying fix; they are separate model-specific output-tracing refactors that share a common pattern but target different code paths.", "confidence": 0.96, "canonical_issue_reason": null, - "canonical_pr_reason": "No true duplicate PR exists here; if a single representative is needed, 44025 is the broadest example because it spans two related depth-anything variants.", + "canonical_pr_reason": "PR 44003 is the strongest representative: it applies the tracing refactor across two closely related Mamba-family implementations and shows the full change pattern.", "best_issue_reason": null, - "best_pr_reason": "44025 is the clearest cluster representative: it covers two closely related model files and reflects the shared output-tracing refactor pattern.", + "best_pr_reason": "PR 44003 is the best overall representative because it is the most substantive, self-contained example of the shared output-tracing refactor theme.", "soft_edge_verdicts": [ { "left": "pull_request:43998", "right": "pull_request:44002", "accept": false, - "reason": "Different model backbones and code paths; same refactor theme, not the same change." + "reason": "Different model families; same refactor motif only." }, { "left": "pull_request:43998", "right": "pull_request:44000", "accept": false, - "reason": "Different multimodal models; only the output-tracing pattern is shared." + "reason": "Unrelated architectures; not the same concrete change." }, { "left": "pull_request:43998", "right": "pull_request:44001", "accept": false, - "reason": "Separate model implementations with no shared bug or mergeable code path." + "reason": "Different code paths in different model stacks." }, { "left": "pull_request:43998", "right": "pull_request:44003", "accept": false, - "reason": "Different architectures; this is only a similar tracing refactor." + "reason": "Timm backbone vs Mamba/FalconMamba; separate fixes." }, { "left": "pull_request:43998", "right": "pull_request:44029", "accept": false, - "reason": "Unrelated model families; same utility pattern, not the same fix." + "reason": "Distinct architectures; only the output-tracing pattern matches." }, { "left": "pull_request:43999", "right": "pull_request:44001", "accept": false, - "reason": "Different models and files; both refactor outputs but do not overlap semantically." + "reason": "MobileNetV1 and UnivNet are unrelated model implementations." }, { "left": "pull_request:43999", "right": "pull_request:44002", "accept": false, - "reason": "Same refactor style, but distinct model code paths." + "reason": "Different subsystems; same broad refactor theme only." }, { "left": "pull_request:43999", "right": "pull_request:44025", "accept": false, - "reason": "Different model families; no shared underlying bug or change." + "reason": "Separate models with different forward paths." }, { "left": "pull_request:43999", "right": "pull_request:44026", "accept": false, - "reason": "Different encoder/decoder model wrappers; similar maintenance, not duplicate work." + "reason": "Different wrapper/model codepaths; not one bug fix." }, { "left": "pull_request:43999", "right": "pull_request:44027", "accept": false, - "reason": "Separate wrappers and code paths; do not look like one fix." + "reason": "MobileNetV1 vs speech encoder-decoder; unrelated changes." }, { "left": "pull_request:44000", "right": "pull_request:44001", "accept": false, - "reason": "Different model domains and implementations; only the refactor motif matches." + "reason": "Vision-text dual encoder and UnivNet fix different code." }, { "left": "pull_request:44000", "right": "pull_request:44002", "accept": false, - "reason": "Distinct model-specific tracing changes, not the same bug." + "reason": "Distinct model families; not mergeable as one PR." }, { "left": "pull_request:44000", "right": "pull_request:44003", "accept": false, - "reason": "Different model families and forward paths." + "reason": "Different architectures and output codepaths." }, { "left": "pull_request:44000", "right": "pull_request:44025", "accept": false, - "reason": "No shared concrete code path; just analogous output-tracing cleanup." + "reason": "Different models; same refactor pattern only." }, { "left": "pull_request:44001", "right": "pull_request:44002", "accept": false, - "reason": "Same high-level pattern, but separate models and changes." + "reason": "UnivNet and UperNet are unrelated fixes." }, { "left": "pull_request:44001", "right": "pull_request:44003", "accept": false, - "reason": "Different model code paths; not mergeable as one PR." + "reason": "Different model stacks and tracing implementations." }, { "left": "pull_request:44001", "right": "pull_request:44025", "accept": false, - "reason": "Separate model implementations with no common bug." + "reason": "Separate model families; not the same underlying change." }, { "left": "pull_request:44002", "right": "pull_request:44029", "accept": false, - "reason": "Different architecture families; no evidence of one underlying issue." + "reason": "UperNet and RWKV touch different code paths." }, { "left": "pull_request:44002", "right": "pull_request:44003", "accept": false, - "reason": "Different models and code paths; only superficially similar." + "reason": "Different architectures; only similarly named refactor." }, { "left": "pull_request:44003", "right": "pull_request:44028", "accept": false, - "reason": "Mamba-related refactor vs. SuperPoint refactor; not the same change." + "reason": "Mamba-family tracing vs SuperPoint output handling." }, { "left": "pull_request:44003", "right": "pull_request:44004", "accept": false, - "reason": "Both are refactors, but they target different model internals and problems." + "reason": "Mamba tracing refactor and CodeGen cache/signature cleanup differ." }, { "left": "pull_request:43998", "right": "pull_request:44004", "accept": false, - "reason": "Different models and forward implementations; same theme only." + "reason": "Timm backbone and CodeGen are unrelated model paths." }, { "left": "pull_request:44001", "right": "pull_request:44004", "accept": false, - "reason": "Different model families and code paths." + "reason": "UnivNet audio model vs CodeGen language model." }, { "left": "pull_request:44002", "right": "pull_request:44004", "accept": false, - "reason": "Unrelated model implementations; not the same underlying change." - }, - { - "left": "pull_request:44000", - "right": "pull_request:44004", - "accept": false, - "reason": "Similar output refactor theme, but separate model logic." + "reason": "Different subsystems; not the same bug or change." }, { "left": "pull_request:44025", "right": "pull_request:44026", "accept": false, - "reason": "Related wrapper patterns, but different model families and files." + "reason": "DepthAnything and VisionEncoderDecoder are separate wrappers." }, { "left": "pull_request:44025", "right": "pull_request:44027", "accept": false, - "reason": "Same utility refactor style, not the same code-path fix." + "reason": "Different model families and forward implementations." }, { "left": "pull_request:44025", "right": "pull_request:44028", "accept": false, - "reason": "Different models and outputs; no common bug or change." + "reason": "Depth estimation vs keypoint model; distinct codepaths." }, { "left": "pull_request:44025", "right": "pull_request:44029", "accept": false, - "reason": "Different model families; this is only a similar tracing refactor." + "reason": "DepthAnything and RWKV are unrelated." }, { "left": "pull_request:44025", "right": "pull_request:44030", "accept": false, - "reason": "Separate model-specific changes with no shared underlying issue." + "reason": "Different model families; same tracing pattern only." }, { "left": "pull_request:44026", "right": "pull_request:44029", "accept": false, - "reason": "Different wrappers and forward paths; not one mergeable fix." + "reason": "Vision encoder-decoder wrapper vs RWKV model." }, { "left": "pull_request:44026", "right": "pull_request:44027", "accept": false, - "reason": "Vision vs. speech encoder-decoder wrappers; analogous, not duplicate." + "reason": "Vision and speech wrappers are separate implementations." }, { "left": "pull_request:44026", "right": "pull_request:44028", "accept": false, - "reason": "Different model code paths and artifacts." + "reason": "Different model types; not one concrete fix." }, { "left": "pull_request:44026", "right": "pull_request:44030", "accept": false, - "reason": "Different models; only the output-tracing refactor pattern is shared." + "reason": "VisionEncoderDecoder and DPR are unrelated code paths." }, { "left": "pull_request:44027", "right": "pull_request:44028", "accept": false, - "reason": "Distinct models and implementations; no shared concrete fix." + "reason": "Speech encoder-decoder and SuperPoint are unrelated." }, { "left": "pull_request:44027", "right": "pull_request:44029", "accept": false, - "reason": "Separate model families with unrelated internals." + "reason": "Different model stacks; same general refactor only." }, { "left": "pull_request:44027", "right": "pull_request:44030", "accept": false, - "reason": "Different model-specific refactors, not the same bug." + "reason": "Speech encoder-decoder vs DPR are separate changes." }, { "left": "pull_request:44028", "right": "pull_request:44029", "accept": false, - "reason": "Different model architectures and forward paths." + "reason": "SuperPoint keypoint model vs RWKV language model." }, { "left": "pull_request:44028", "right": "pull_request:44030", "accept": false, - "reason": "SuperPoint vs. DPR are unrelated changes." + "reason": "Different architectures and output semantics." }, { "left": "pull_request:44029", "right": "pull_request:44030", "accept": false, - "reason": "Different model families and code paths; no duplicate evidence." + "reason": "RWKV and DPR are unrelated model implementations." } ] }, @@ -21747,8 +21586,8 @@ }, { "label": "pull_request", - "component_index": 16, - "component_count": 39, + "component_index": 15, + "component_count": 35, "review_unit_index": 1, "review_unit_count": 1, "status": "reviewed", @@ -21764,12 +21603,12 @@ "soft_pairs": [ "pull_request:44662|pull_request:45401" ], - "prepared_review_unit_hash": "73664c823c8e7c4ebd63832013921ee2caf6c33b2451dd082b0df56d263c21ad", + "prepared_review_unit_hash": "0dedb6745de8d58695b9be3e3215b3524d241e0965dec433532e028318117a5c", "component_budget": { "node_count": 2, "item_count": 2, "soft_pair_count": 1, - "serialized_chars": 6551, + "serialized_chars": 6552, "estimated_input_tokens": 1638, "estimated_eval_tokens": 3532 }, @@ -21777,7 +21616,7 @@ "node_count": 2, "item_count": 2, "soft_pair_count": 1, - "serialized_chars": 6551, + "serialized_chars": 6552, "estimated_input_tokens": 1638, "estimated_eval_tokens": 3532 }, @@ -21786,170 +21625,18 @@ "aggressively_trimmed": false, "split": false, "analyst_result": { - "summary": "The two PRs are both model-addition implementations, but they target different models and codepaths (PenguinVL vs Voxtral TTS), so they are not duplicates.", - "confidence": 0.95, + "summary": "These are two unrelated model-integration PRs that share only generated boilerplate and auto-registration files. They target different concrete models and do not appear to be mergeable into one change.", + "confidence": 0.97, "canonical_issue_reason": null, - "canonical_pr_reason": "PR #44662 is the stronger standalone representative: it is a full implementation PR with substantial review activity and a larger, more mature codebase addition.", + "canonical_pr_reason": null, "best_issue_reason": null, - "best_pr_reason": "PR #45401 is more self-contained and traceable because it explicitly targets an issue, but it is draft and narrower; #44662 is the better overall representative of the cluster's implementation work.", + "best_pr_reason": "PR 45401 is the cleaner representative if one must be chosen: it has an explicit issue target and a tighter, model-specific scope.", "soft_edge_verdicts": [ { "left": "pull_request:44662", "right": "pull_request:45401", "accept": false, - "reason": "Different models and different concrete features: PenguinVL implementation vs Voxtral TTS support. They share boilerplate auto-registration and docs patterns, but not 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": "pull_request", - "component_index": 17, - "component_count": 39, - "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-44958-7", - "nodes": [ - "pull_request:44958", - "pull_request:45415", - "pull_request:45425", - "pull_request:46635", - "pull_request:46647", - "pull_request:46648", - "pull_request:46649" - ], - "soft_pairs": [ - "pull_request:44958|pull_request:46647", - "pull_request:44958|pull_request:46635", - "pull_request:45415|pull_request:45425", - "pull_request:45415|pull_request:46647", - "pull_request:45415|pull_request:46648", - "pull_request:45415|pull_request:46649", - "pull_request:45415|pull_request:46635", - "pull_request:46635|pull_request:46649", - "pull_request:46635|pull_request:46647", - "pull_request:46635|pull_request:46648", - "pull_request:46647|pull_request:46648", - "pull_request:46647|pull_request:46649", - "pull_request:46648|pull_request:46649" - ], - "prepared_review_unit_hash": "7dc80deb006eacc0f2176d1f72e7799f510a5d2199110660541cd8fbcca9a83f", - "component_budget": { - "node_count": 7, - "item_count": 7, - "soft_pair_count": 13, - "serialized_chars": 19759, - "estimated_input_tokens": 4940, - "estimated_eval_tokens": 10136 - }, - "budget": { - "node_count": 7, - "item_count": 7, - "soft_pair_count": 13, - "serialized_chars": 19759, - "estimated_input_tokens": 4940, - "estimated_eval_tokens": 10136 - }, - "overflow_policy": "truncate_then_skip", - "trimmed": false, - "aggressively_trimmed": false, - "split": false, - "analyst_result": { - "summary": "This cluster is mostly heterogeneous maintenance work: one import-fix PR, one broad typing sweep, one typing-helper PR, and several unrelated typo-only PRs. I don\u2019t see any true duplicate PRs; the shared similarity is mostly superficial (same subsystem or style of cleanup).", - "confidence": 0.92, - "canonical_issue_reason": null, - "canonical_pr_reason": "PR 45425 is the most central and substantial maintenance change in the set, touching shared typing infrastructure and several core files; if one PR must anchor the cluster, it is the best representative.", - "best_issue_reason": null, - "best_pr_reason": "PR 45425 is the strongest global representative because it is the broadest, most cross-cutting change and has the most direct connection to the shared typing files seen across the cluster.", - "soft_edge_verdicts": [ - { - "left": "pull_request:44958", - "right": "pull_request:46635", - "accept": false, - "reason": "Different fixes: one is an import-guard change for PILImageResampling, the other is a one-line typo cleanup in generation configuration." - }, - { - "left": "pull_request:44958", - "right": "pull_request:46647", - "accept": false, - "reason": "Both are small maintenance edits, but they fix unrelated files and problems; no shared concrete bug or code path." - }, - { - "left": "pull_request:45415", - "right": "pull_request:45425", - "accept": false, - "reason": "Both are typing-related, but 45415 is a broad type-checking sweep while 45425 adds typing helpers and targeted model annotations; they are not the same change." - }, - { - "left": "pull_request:45415", - "right": "pull_request:46647", - "accept": false, - "reason": "A broad typing sweep versus a typo-fix PR in rope/processing text; same general theme, but not the same underlying change." - }, - { - "left": "pull_request:45415", - "right": "pull_request:46648", - "accept": false, - "reason": "The overlap is only in the broad typing/cleanup area; the actual edits target different files and fixes." - }, - { - "left": "pull_request:45415", - "right": "pull_request:46649", - "accept": false, - "reason": "These are separate maintenance efforts with different concrete edits; shared subsystem alone is not enough to merge them." - }, - { - "left": "pull_request:45415", - "right": "pull_request:46635", - "accept": false, - "reason": "A project-wide typing sweep and a one-line typo fix are unrelated changes." - }, - { - "left": "pull_request:46635", - "right": "pull_request:46649", - "accept": false, - "reason": "Both are typo fixes, but they correct different strings in different files and are not one concrete bug fix." - }, - { - "left": "pull_request:46635", - "right": "pull_request:46647", - "accept": false, - "reason": "Same cleanup style, but different typo targets and different code areas; not duplicates." - }, - { - "left": "pull_request:46635", - "right": "pull_request:46648", - "accept": false, - "reason": "Different typo fixes in different files; they could coexist but are not the same underlying change." - }, - { - "left": "pull_request:46647", - "right": "pull_request:46648", - "accept": false, - "reason": "They overlap on some typo edits, but each PR also fixes different typos and file sets; not a single concrete bug/change." - }, - { - "left": "pull_request:46647", - "right": "pull_request:46649", - "accept": false, - "reason": "Shared typo-fix flavor, but the specific corrections differ and the PRs are not mergeable as one change." - }, - { - "left": "pull_request:46648", - "right": "pull_request:46649", - "accept": false, - "reason": "Both are typo-only maintenance PRs, but they fix different typos in different files; same style, not same bug." + "reason": "Different underlying changes: PenguinVL implementation vs Voxtral TTS support. Overlapping docs/auto-registry files are incidental and not evidence of the same bug or feature." } ] }, @@ -21962,8 +21649,8 @@ }, { "label": "pull_request", - "component_index": 18, - "component_count": 39, + "component_index": 16, + "component_count": 35, "review_unit_index": 1, "review_unit_count": 1, "status": "reviewed", @@ -21979,12 +21666,12 @@ "soft_pairs": [ "pull_request:44979|pull_request:45363" ], - "prepared_review_unit_hash": "e29e6c049753c43291f03ea3fce118fc8e2a738c855bade8ce89bb278f6695da", + "prepared_review_unit_hash": "dd882918d669d9f8ee37194ff5d8e5b9ab1739227019d630753690fc8839694b", "component_budget": { "node_count": 2, "item_count": 2, "soft_pair_count": 1, - "serialized_chars": 4408, + "serialized_chars": 4405, "estimated_input_tokens": 1102, "estimated_eval_tokens": 2460 }, @@ -21992,7 +21679,7 @@ "node_count": 2, "item_count": 2, "soft_pair_count": 1, - "serialized_chars": 4408, + "serialized_chars": 4405, "estimated_input_tokens": 1102, "estimated_eval_tokens": 2460 }, @@ -22001,18 +21688,18 @@ "aggressively_trimmed": false, "split": false, "analyst_result": { - "summary": "These PRs are related but not duplicates: #44979 adds a general Module Fusion API, while #45363 builds a concrete n-to-1 kernel-fusion workflow on top of KernelConfig and integrations. They target different layers of the stack and would not be merged as one change.", - "confidence": 0.95, + "summary": "Two open PRs touch fusion-related code, but they implement different features: one adds a generic Module Fusion API, the other adds KernelConfig-based n-to-1 kernel fusion integration. They are related in theme but not the same underlying change.", + "confidence": 0.94, "canonical_issue_reason": null, - "canonical_pr_reason": "#45363 is the stronger representative for the cluster because it is the more concrete, user-facing change and covers the applied kernel-fusion path built on top of the fusion infrastructure.", + "canonical_pr_reason": "PR #45363 is the stronger representative of the cluster because it makes the broader end-to-end kernel-fusion integration change across core files, while #44979 is a narrower standalone API addition.", "best_issue_reason": null, - "best_pr_reason": "#45363 is the best standalone PR in this cluster because it describes a specific end-to-end feature (KernelConfig-driven n-to-1 fusion) rather than only the underlying API scaffolding.", + "best_pr_reason": "PR #45363 is the best single item to surface because it has the broader user-facing impact and spans the main integration points for kernel fusion, whereas #44979 is an isolated API/module addition.", "soft_edge_verdicts": [ { - "left": "Module Fusion API", - "right": "n-to-1 kernel fusion via `KernelConfig`", + "left": "pull_request:44979", + "right": "pull_request:45363", "accept": false, - "reason": "Different scope and code paths: #44979 introduces a generic fusion API, while #45363 wires that idea into kernel fusion with `KernelConfig`, integrations, and an example. Related, but not the same underlying change." + "reason": "Both are about fusion, but #44979 introduces a generic Module Fusion API in a new module, while #45363 implements KernelConfig-driven n-to-1 kernel fusion and related integrations. They do not appear to fix the same concrete code-path problem and would not plausibly be merged as one PR." } ] }, @@ -22025,8 +21712,8 @@ }, { "label": "pull_request", - "component_index": 19, - "component_count": 39, + "component_index": 17, + "component_count": 35, "review_unit_index": 1, "review_unit_count": 1, "status": "reviewed", @@ -22042,7 +21729,7 @@ "soft_pairs": [ "pull_request:45086|pull_request:45317" ], - "prepared_review_unit_hash": "ffe95192a57138678e3a4f339c5a389e45d024126f69afa997bbe9028ceba535", + "prepared_review_unit_hash": "896bf044af722b11ad0a6390d71c56e3f5368ca819b6a65ad16313f07a2ec665", "component_budget": { "node_count": 2, "item_count": 2, @@ -22064,18 +21751,18 @@ "aggressively_trimmed": false, "split": false, "analyst_result": { - "summary": "Both pull requests target the same Mistral regex patch in the same file for the same issue; 45317 is the more complete duplicate because it includes test coverage in addition to the fix.", + "summary": "Both pull requests implement the same fix for the Mistral regex patch AttributeError in the same file and target the same issue; 45317 is the fuller, test-backed version.", "confidence": 0.98, "canonical_issue_reason": null, - "canonical_pr_reason": "pull_request:45317 is the better canonical PR because it contains the same code fix as 45086 plus accompanying tests, making it the more complete representation of the change.", + "canonical_pr_reason": "PR 45317 is the better canonical PR because it contains the same code fix as 45086 plus test coverage, making it the more complete change for the issue.", "best_issue_reason": null, - "best_pr_reason": "pull_request:45317 is the strongest standalone PR for this bug: same concrete fix, same target issue, and it adds validation coverage.", + "best_pr_reason": "PR 45317 is the best representative of the cluster: same underlying bug fix, same target issue, and it adds validation/tests beyond the minimal patch.", "soft_edge_verdicts": [ { "left": "pull_request:45086", "right": "pull_request:45317", "accept": true, - "reason": "Same underlying AttributeError fix in _patch_mistral_regex on the same code path and same issue target; 45317 simply adds tests, so these are duplicate PRs that could be merged into one." + "reason": "Same concrete bug fix in tokenization_utils_tokenizers.py for _patch_mistral_regex; 45317 is essentially the expanded version of 45086 with tests." } ] }, @@ -22088,8 +21775,8 @@ }, { "label": "pull_request", - "component_index": 20, - "component_count": 39, + "component_index": 18, + "component_count": 35, "review_unit_index": 1, "review_unit_count": 1, "status": "reviewed", @@ -22109,52 +21796,115 @@ "pull_request:45569|pull_request:46565", "pull_request:46332|pull_request:46565" ], - "prepared_review_unit_hash": "7617f38c34b80835559ae6b1702233cab67dcda591fd697415b72969b856eebb", + "prepared_review_unit_hash": "2f3d6453a8989af9475bcbd3c729691affe279d51376388978f1798060a8a539", "component_budget": { "node_count": 4, "item_count": 4, "soft_pair_count": 3, - "serialized_chars": 13473, - "estimated_input_tokens": 3369, - "estimated_eval_tokens": 6994 + "serialized_chars": 13472, + "estimated_input_tokens": 3368, + "estimated_eval_tokens": 6992 + }, + "budget": { + "node_count": 4, + "item_count": 4, + "soft_pair_count": 3, + "serialized_chars": 13472, + "estimated_input_tokens": 3368, + "estimated_eval_tokens": 6992 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": false, + "analyst_result": { + "summary": "These PRs are all Nemotron-related, but they split into distinct change sets: a NemotronH inheritance/refactor, a NemotronH dense/sparse split, and two separate ASR streaming model additions. The overlap is mostly shared framework/docs plumbing rather than the same underlying bug or feature.", + "confidence": 0.92, + "canonical_issue_reason": null, + "canonical_pr_reason": "PR 45569 is the most central Nemotron-family architecture change in the cluster and the closest thing to a unifying representative, since it covers the core NemotronH refactor rather than a narrower follow-up or a separate ASR feature line.", + "best_issue_reason": null, + "best_pr_reason": "PR 45569 is the best overall representative for this cluster because it has the broadest impact on the core NemotronH implementation and ties most directly to the related NemotronH refactor PR; the ASR PRs are a different product area, not the same change.", + "soft_edge_verdicts": [ + { + "left": "pull_request:45115", + "right": "pull_request:45569", + "accept": false, + "reason": "Related NemotronH work, but not the same change: 45115 is a GraniteMoeHybrid inheritance/mapping refactor, while 45569 splits NemotronH into dense/sparse variants and updates docs/model registration. Shared files reflect the same subsystem, not duplicate code-path fixes." + }, + { + "left": "pull_request:45569", + "right": "pull_request:46565", + "accept": false, + "reason": "Different model families and goals: 45569 refactors NemotronH architecture, while 46565 adds Nemotron 3.5 ASR streaming. Similar auto/docs/loss files are shared infrastructure, not evidence of one mergeable PR." + }, + { + "left": "pull_request:46332", + "right": "pull_request:46565", + "accept": false, + "reason": "Both are ASR streaming PRs, but they target different concrete models (Nemotron ASR vs Nemotron 3.5 ASR). That makes them related follow-ups, not the same underlying feature or code-path fix." + } + ] + }, + "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": 35, + "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-45332-2", + "nodes": [ + "pull_request:45332", + "pull_request:45333" + ], + "soft_pairs": [ + "pull_request:45332|pull_request:45333" + ], + "prepared_review_unit_hash": "77ffd1df759b8457bfc82a0bbfbc30f73c5e037c633125312039b7557e58b351", + "component_budget": { + "node_count": 2, + "item_count": 2, + "soft_pair_count": 1, + "serialized_chars": 5466, + "estimated_input_tokens": 1367, + "estimated_eval_tokens": 2990 }, "budget": { - "node_count": 4, - "item_count": 4, - "soft_pair_count": 3, - "serialized_chars": 13473, - "estimated_input_tokens": 3369, - "estimated_eval_tokens": 6994 + "node_count": 2, + "item_count": 2, + "soft_pair_count": 1, + "serialized_chars": 5466, + "estimated_input_tokens": 1367, + "estimated_eval_tokens": 2990 }, "overflow_policy": "truncate_then_skip", "trimmed": false, "aggressively_trimmed": false, "split": false, "analyst_result": { - "summary": "This cluster actually contains two separate themes: Nemotron-H architecture/refactor work and Nemotron ASR streaming work. The overlap is mostly broad subsystem/shared-infra similarity, not the same concrete change.", - "confidence": 0.78, + "summary": "Two related heterogeneous-model feature PRs, but they cover different layers of the implementation: 45333 adds per-layer configuration plumbing, while 45332 adds model/runtime integration on top of that. They are not duplicates.", + "confidence": 0.94, "canonical_issue_reason": null, - "canonical_pr_reason": "No single PR cleanly represents the whole cluster, but 46565 is the best anchor for triage because it is the most complete and centrally connected to the ASR subgroup. The Nemotron-H PRs are a separate line of work.", + "canonical_pr_reason": "45333 is the more foundational, narrowly scoped PR: it introduces the per-layer config API and helper utilities that 45332 builds on. It is the better anchor if these are grouped by topic.", "best_issue_reason": null, - "best_pr_reason": "46565 is the broadest and most recent PR in the largest coherent subgroup, with the strongest overlap to 46332. It is the best representative PR for cluster review, though not a duplicate of the Nemotron-H work.", + "best_pr_reason": "45332 is the broader representative of the heterogeneous-model effort because it touches cache, masking, modeling, and tests; but it is still a separate follow-on change, not a duplicate of 45333.", "soft_edge_verdicts": [ { - "left": "pull_request:45115", - "right": "pull_request:45569", - "accept": false, - "reason": "Same model family, but not the same change: 45115 is a narrow conversion/inheritance refactor, while 45569 introduces the dense/sparse split and documentation/model wiring. Too broad to merge as one PR." - }, - { - "left": "pull_request:45569", - "right": "pull_request:46565", - "accept": false, - "reason": "Different features and model families: Nemotron-H architecture refactor versus Nemotron ASR streaming. Shared auto/doc plumbing is not enough to treat them as the same underlying change." - }, - { - "left": "pull_request:46332", - "right": "pull_request:46565", + "left": "pull_request:45332", + "right": "pull_request:45333", "accept": false, - "reason": "Both are ASR-related, but they target different concrete model variants (Nemotron ASR vs Nemotron 3.5 ASR). The overlap is mostly shared infrastructure and docs, not one fix/change that should be merged into a single PR." + "reason": "Related feature stack, but not the same concrete change. 45333 adds per-layer config support; 45332 adds heterogeneous model/runtime support that depends on that config plumbing. They could be sequential PRs, not one duplicate PR." } ] }, @@ -22167,8 +21917,8 @@ }, { "label": "pull_request", - "component_index": 21, - "component_count": 39, + "component_index": 20, + "component_count": 35, "review_unit_index": 1, "review_unit_count": 1, "status": "reviewed", @@ -22176,48 +21926,48 @@ "source": "llm", "cache_hit": false, "model": "gpt-5.4-mini?service_tier=flex", - "cluster_id": "cluster-45332-2", + "cluster_id": "cluster-45415-2", "nodes": [ - "pull_request:45332", - "pull_request:45333" + "pull_request:45415", + "pull_request:45425" ], "soft_pairs": [ - "pull_request:45332|pull_request:45333" + "pull_request:45415|pull_request:45425" ], - "prepared_review_unit_hash": "07126ee77d3a5b520f4209250940a609e9e08831b7dc24fce46bc1911b99e63d", + "prepared_review_unit_hash": "d65d51d795973219e9aa510f92cf2a9e8ba39d0c028085e69e0d52fa3202a52b", "component_budget": { "node_count": 2, "item_count": 2, "soft_pair_count": 1, - "serialized_chars": 5466, - "estimated_input_tokens": 1367, - "estimated_eval_tokens": 2990 + "serialized_chars": 5514, + "estimated_input_tokens": 1379, + "estimated_eval_tokens": 3014 }, "budget": { "node_count": 2, "item_count": 2, "soft_pair_count": 1, - "serialized_chars": 5466, - "estimated_input_tokens": 1367, - "estimated_eval_tokens": 2990 + "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": "These are related heterogeneous-modeling PRs, but they are not duplicates: one adds per-layer config infrastructure, the other adds model/cache/masking support that consumes it.", - "confidence": 0.89, + "summary": "Both PRs are typing-related, but they target different scopes: one is a broad repo-wide type-checking pass, the other is a narrower typing refactor centered on modeling_utils and shared typing helpers. They do not look like the same underlying change and should not be merged as duplicates.", + "confidence": 0.88, "canonical_issue_reason": null, - "canonical_pr_reason": "PR #45333 is the cleaner canonical representative because it introduces the foundational per-layer configuration API and helper plumbing; the larger PR #45332 depends on that groundwork and expands into modeling/runtime changes.", + "canonical_pr_reason": null, "best_issue_reason": null, - "best_pr_reason": "PR #45333 is the best single PR to keep as the cluster representative since its scope is narrower and more self-contained, while #45332 is a broader follow-on that spans multiple code paths.", + "best_pr_reason": "PR 45425 is the more focused change and reads like a distinct, self-contained typing cleanup, but it is not a duplicate of 45415.", "soft_edge_verdicts": [ { - "left": "pull_request:45332", - "right": "pull_request:45333", + "left": "pull_request:45415", + "right": "pull_request:45425", "accept": false, - "reason": "Related but not the same underlying change: #45333 implements heterogeneous config support, while #45332 implements heterogeneous model/runtime support using those configs. Shared files are supporting plumbing, not evidence of one mergeable PR for the same concrete fix." + "reason": "Different change intents and scopes: 45415 is a broad type-checking sweep across many transformers modules, while 45425 adds specific typing protocols/helpers and related modeling_utils fixes. Shared files alone are not enough to treat them as the same concrete PR." } ] }, @@ -22230,8 +21980,8 @@ }, { "label": "pull_request", - "component_index": 22, - "component_count": 39, + "component_index": 21, + "component_count": 35, "review_unit_index": 1, "review_unit_count": 1, "status": "reviewed", @@ -22247,40 +21997,40 @@ "soft_pairs": [ "pull_request:45513|pull_request:46084" ], - "prepared_review_unit_hash": "f53011b339c5e5bf5eafec27de267b3f5db4f7440cb4f187519dc913ccfc4474", + "prepared_review_unit_hash": "75a510b130d3913e763a9b65da0da2d93e7c73a5ccac50c461a4609f9b6de21d", "component_budget": { "node_count": 2, "item_count": 2, "soft_pair_count": 1, - "serialized_chars": 4949, - "estimated_input_tokens": 1238, - "estimated_eval_tokens": 2732 + "serialized_chars": 4948, + "estimated_input_tokens": 1237, + "estimated_eval_tokens": 2730 }, "budget": { "node_count": 2, "item_count": 2, "soft_pair_count": 1, - "serialized_chars": 4949, - "estimated_input_tokens": 1238, - "estimated_eval_tokens": 2732 + "serialized_chars": 4948, + "estimated_input_tokens": 1237, + "estimated_eval_tokens": 2730 }, "overflow_policy": "truncate_then_skip", "trimmed": false, "aggressively_trimmed": false, "split": false, "analyst_result": { - "summary": "These are different pull requests for different model families and different cache bugs: one fixes multi-token cached forward in Qwen3.5/GDN linear attention, the other fixes use_cache behavior for Mamba2 when seq_len > 1. They are not duplicates.", - "confidence": 0.97, + "summary": "These PRs are thematically similar (cache correctness with multi-token cached forward), but they fix different bugs in different model families and code paths, so they are not duplicates.", + "confidence": 0.95, "canonical_issue_reason": null, - "canonical_pr_reason": "PR #46084 is the cleaner representative: it has an explicit issue target, a narrowly scoped fix in one code path, and a focused test/update set.", + "canonical_pr_reason": "No canonical PR: 45513 and 46084 address separate cache bugs (Qwen3.5/Qwen3 Next/Olmo Hybrid linear attention vs. Mamba2 CUDA kernel cache handling), so they should not be merged or treated as the same change.", "best_issue_reason": null, - "best_pr_reason": "PR #46084 is the best standalone PR candidate because it is tightly scoped to a single concrete bug (use_cache with seq_len > 1) and directly documents the intended fix.", + "best_pr_reason": "46084 is the cleaner representative of its own bug: it is narrowly scoped, has an explicit issue target (#46032), and focuses on one concrete seq_len > 1 cache fix.", "soft_edge_verdicts": [ { "left": "pull_request:45513", "right": "pull_request:46084", "accept": false, - "reason": "Different underlying bugs in different architectures: Qwen3.5/GDN linear attention cached multi-token forward vs. Mamba2 use_cache with seq_len > 1. They are not plausibly one mergeable PR." + "reason": "Reject: both are about cached forward behavior, but 45513 changes GDN linear attention paths in Qwen3.5/Qwen3 Next/Olmo Hybrid, while 46084 fixes Mamba2 use_cache handling in cuda_kernels_forward. Different underlying bugs and not plausibly one merged PR." } ] }, @@ -22293,8 +22043,8 @@ }, { "label": "pull_request", - "component_index": 23, - "component_count": 39, + "component_index": 22, + "component_count": 35, "review_unit_index": 1, "review_unit_count": 1, "status": "reviewed", @@ -22310,7 +22060,7 @@ "soft_pairs": [ "pull_request:45548|pull_request:45662" ], - "prepared_review_unit_hash": "826e245a044a3194d6ee38c8e3979b343ad6afeb2ef2aaa03bdd39dbf03743fc", + "prepared_review_unit_hash": "2586c4914b4505a8f0081b862ab5f506e3c3c564f46177cb0ef07cb776448045", "component_budget": { "node_count": 2, "item_count": 2, @@ -22332,18 +22082,18 @@ "aggressively_trimmed": false, "split": false, "analyst_result": { - "summary": "Both PRs target expert-parallelism loading behavior, but they fix different concrete bugs: one is about DeepSpeed ZeRO-3 loading via accelerate launch, the other about FSDP2/rank-0 broadcast overwriting experts. They overlap in theme and one shared file, but are not the same change.", - "confidence": 0.93, + "summary": "Both PRs are about expert parallelism, but they fix different loading bugs in different distributed setups: one is config/serialization and EP detection for DeepSpeed ZeRO-3 via accelerate launch, the other fixes expert weights being overwritten during FSDP2/rank-0 broadcast.", + "confidence": 0.96, "canonical_issue_reason": null, - "canonical_pr_reason": "No single canonical PR: the fixes diverge by backend and failure mode (ZeRO-3/accelerate launch vs FSDP2 broadcast overwrite).", + "canonical_pr_reason": "PR 45662 is the more concrete representative of the cluster: it fixes a specific load-time overwrite path in model loading and distributed wrapping, with changes spread across the affected code paths.", "best_issue_reason": null, - "best_pr_reason": "45662 is the stronger representative of the shared EP-loading theme because it contains the more substantial, multi-file fix for a specific overwrite bug.", + "best_pr_reason": "PR 45548 is narrower and more focused on EP config persistence and the has_ep helper for a DeepSpeed/accelerate loading path; it is less representative of the broader distributed-loading bug set than 45662.", "soft_edge_verdicts": [ { "left": "pull_request:45548", "right": "pull_request:45662", "accept": false, - "reason": "Both are EP-related, but they address different underlying bugs and code paths: 45548 targets DeepSpeed ZeRO-3 loading via accelerate launch, while 45662 fixes FSDP2/rank-0 broadcast overwriting expert parameters. Shared subsystem alone is not enough to merge them." + "reason": "Different underlying bugs and code paths: 45548 adds distributed_config serialization and EP detection for DeepSpeed ZeRO-3/accelerate launch, while 45662 prevents expert parameters from being overwritten during FSDP2 rank-0 broadcast. They are related to EP but not the same fix." } ] }, @@ -22356,8 +22106,8 @@ }, { "label": "pull_request", - "component_index": 24, - "component_count": 39, + "component_index": 23, + "component_count": 35, "review_unit_index": 1, "review_unit_count": 1, "status": "reviewed", @@ -22373,7 +22123,7 @@ "soft_pairs": [ "pull_request:45639|pull_request:45645" ], - "prepared_review_unit_hash": "993fe4e7c8dcb3b12c7f2b8ad57a53f79b2eaa3fb6ffc266aecbaa4fdd4d15a2", + "prepared_review_unit_hash": "1d5c620b5fe7359090138557f763a48640224306236171d5df8e59afad4ba826", "component_budget": { "node_count": 2, "item_count": 2, @@ -22395,18 +22145,18 @@ "aggressively_trimmed": false, "split": false, "analyst_result": { - "summary": "Both PRs address the same xdist-related captured_info logging collision in testing_utils, with 45645 extending the fix to CI artifact collection and related cleanup. PR 45645 looks like the broader, more complete superset of 45639.", - "confidence": 0.84, + "summary": "Both PRs address the same xdist collision problem around `captured_info` debug artifacts; #45645 is a broader superset that also updates CI log collection, while #45639 is the narrower core fix.", + "confidence": 0.86, "canonical_issue_reason": null, - "canonical_pr_reason": "PR 45645 is the better canonical choice because it subsumes the core xdist-safe captured_info fix in testing_utils and also updates CI/workflow handling so the new per-worker logs are actually collected and preserved.", + "canonical_pr_reason": "PR #45645 is the best canonical representative because it contains the core xdist-safe artifact fix and the downstream CI changes needed to consume the new per-worker log files.", "best_issue_reason": null, - "best_pr_reason": "PR 45645 is the strongest representative of the cluster: it targets the same underlying bug as 45639 but covers the end-to-end artifact handling, not just the local logging write path.", + "best_pr_reason": "#45645 is the most suitable overall PR in the cluster since it resolves the core bug and also handles the CI/reporting side effects; #45639 is a narrower variant of the same fix.", "soft_edge_verdicts": [ { "left": "pull_request:45639", "right": "pull_request:45645", "accept": true, - "reason": "Both PRs fix the same captured_info/xdist collision problem in the same code path, and 45645 appears to be an expanded version of the same change rather than a separate fix." + "reason": "Same underlying bug/change: making patched testing debug logs xdist-safe for `captured_info` artifacts. #45645 extends the same fix with CI log aggregation changes, but the core code-path problem is shared." } ] }, @@ -22419,8 +22169,8 @@ }, { "label": "pull_request", - "component_index": 25, - "component_count": 39, + "component_index": 24, + "component_count": 35, "review_unit_index": 1, "review_unit_count": 1, "status": "reviewed", @@ -22439,7 +22189,7 @@ "pull_request:45670|pull_request:46401", "pull_request:46320|pull_request:46401" ], - "prepared_review_unit_hash": "9e14b7cdec10f5e8596a1a8c80a79fed838ad21f15882275de76b32d6f76933c", + "prepared_review_unit_hash": "625601330ad794042ce6abc550b07aa6404120b3947f80c0fd19f9a9ccf83445", "component_budget": { "node_count": 3, "item_count": 3, @@ -22461,30 +22211,30 @@ "aggressively_trimmed": false, "split": false, "analyst_result": { - "summary": "45670 is a separate AutoModel registration nit. 46320 and 46401 are both attn_implementation dispatch PRs, but they target different model families and code paths, so they are not duplicates of each other.", - "confidence": 0.95, + "summary": "These are three separate pull requests with only superficial similarity. Two PRs add attn_implementation support, but they target different model families and code paths; the third is an AutoModel registry change for a different model entirely.", + "confidence": 0.92, "canonical_issue_reason": null, - "canonical_pr_reason": "Choose 46401 as the best representative PR: it is the broadest attn_implementation dispatch change in this set, covering multiple BLIP-family models with tests/docs. It is not a duplicate of the other PRs, though.", + "canonical_pr_reason": null, "best_issue_reason": null, - "best_pr_reason": "46401 is the strongest single representative of the cluster theme because it addresses the same general feature area across several related models; 46320 is narrower and 45670 is unrelated.", + "best_pr_reason": null, "soft_edge_verdicts": [ { "left": "pull_request:45670", "right": "pull_request:46320", "accept": false, - "reason": "Different problems: 45670 adds an AutoModel mapping for glmasr/dia, while 46320 implements attention-dispatch support for RecurrentGemma." + "reason": "Unrelated changes: one adds GLM-ASR to AutoModel mappings, the other implements attn_implementation dispatch for RecurrentGemma. Different feature area and code path." }, { "left": "pull_request:45670", "right": "pull_request:46401", "accept": false, - "reason": "Different problems: 45670 is a model registry/nit change, while 46401 adds attn_implementation dispatch for BLIP-2/InstructBLIP family models." + "reason": "Unrelated changes: AutoModel mapping for glmasr versus attention backend dispatch for BLIP-2/InstructBLIP models. Not the same underlying bug or mergeable into one PR." }, { "left": "pull_request:46320", "right": "pull_request:46401", "accept": false, - "reason": "Same broad feature area, but different concrete code paths and model families; these could not plausibly be merged as one duplicate PR." + "reason": "Both are about attn_implementation support, but they fix different model-specific implementations (RecurrentGemma vs BLIP-2/InstructBLIP/InstructBLIPVideo). They are not the same concrete bug/change and should not be merged as duplicates." } ] }, @@ -22497,8 +22247,8 @@ }, { "label": "pull_request", - "component_index": 26, - "component_count": 39, + "component_index": 25, + "component_count": 35, "review_unit_index": 1, "review_unit_count": 1, "status": "reviewed", @@ -22514,7 +22264,7 @@ "soft_pairs": [ "pull_request:45720|pull_request:46259" ], - "prepared_review_unit_hash": "39fbe6dfd6a9f339c39e2a5f408fa3a472bb95d09e877aec198ac7a96e2f1239", + "prepared_review_unit_hash": "1467a267b1271f5365e316d09768243b7cb3a7c44a8edfe19effc74b18422542", "component_budget": { "node_count": 2, "item_count": 2, @@ -22536,18 +22286,18 @@ "aggressively_trimmed": false, "split": false, "analyst_result": { - "summary": "These are both linter-rule implementation PRs in transformers, but they target different rules and different code paths: TRF003 for model forward signatures versus TRF018 for model weight initialization. They are related at a broad subsystem level, not the same underlying change.", - "confidence": 0.94, + "summary": "These are two unrelated linter/chore PRs touching overlapping model files, but they implement different rules on different code paths: one changes forward return-tuple handling (TRF003) and the other changes weight-init patterns (TRF018).", + "confidence": 0.96, "canonical_issue_reason": null, - "canonical_pr_reason": "PR #45720 is the cleaner representative of the cluster: it has a focused, concrete fix shape (adding can_return_tuple / adjusting forward signatures) and a narrower scope than the broader multi-model TRF018 initialization rewrite.", + "canonical_pr_reason": null, "best_issue_reason": null, - "best_pr_reason": "PR #45720 is the best representative because it is more specific about the code-path change it makes and is easier to reason about as a standalone implementation PR.", + "best_pr_reason": "PR 45720 is the more coherent single change: it targets one concrete behavior across model forwards (return-tuple handling) and is easier to reason about than the broader weight-init refactor in 46259.", "soft_edge_verdicts": [ { "left": "pull_request:45720", "right": "pull_request:46259", "accept": false, - "reason": "Different lint rules and different mechanics: one changes forward return/signature handling, the other refactors _init_weights across many models. They are not the same bug or mergeable into one PR." + "reason": "Different underlying lint rules and fixes: TRF003 changes forward signatures/return behavior, while TRF018 changes initialization logic. Shared model filenames are incidental, not the same bug/change." } ] }, @@ -22560,8 +22310,8 @@ }, { "label": "pull_request", - "component_index": 27, - "component_count": 39, + "component_index": 26, + "component_count": 35, "review_unit_index": 1, "review_unit_count": 1, "status": "reviewed", @@ -22626,8 +22376,8 @@ }, { "label": "pull_request", - "component_index": 28, - "component_count": 39, + "component_index": 27, + "component_count": 35, "review_unit_index": 1, "review_unit_count": 1, "status": "reviewed", @@ -22643,12 +22393,12 @@ "soft_pairs": [ "pull_request:45932|pull_request:46100" ], - "prepared_review_unit_hash": "638df4a4bd51aa833edb8e97abf6c74c62e33a955473b8852eda25a3227e340d", + "prepared_review_unit_hash": "ec70548425334d4881988fc90a97d62c4d1279266017962e423fefefb44096ea", "component_budget": { "node_count": 2, "item_count": 2, "soft_pair_count": 1, - "serialized_chars": 5083, + "serialized_chars": 5084, "estimated_input_tokens": 1271, "estimated_eval_tokens": 2798 }, @@ -22656,7 +22406,7 @@ "node_count": 2, "item_count": 2, "soft_pair_count": 1, - "serialized_chars": 5083, + "serialized_chars": 5084, "estimated_input_tokens": 1271, "estimated_eval_tokens": 2798 }, @@ -22665,18 +22415,18 @@ "aggressively_trimmed": false, "split": false, "analyst_result": { - "summary": "Two open PRs are present, but they do not look like the same change: one is a broad model/code adjustment PR with a CI secret tweak, while the other is a test migration adding a new ViT test file and related test updates.", - "confidence": 0.93, + "summary": "These two pull requests are not duplicates: one is a broad CI/model maintenance PR with grafana workflow and model code changes, while the other is a ViT test-migration PR focused on test files. They do not appear to fix the same underlying bug or implement the same change.", + "confidence": 0.96, "canonical_issue_reason": null, - "canonical_pr_reason": "PR #45932 is the better canonical PR because it contains substantive model/code edits across multiple files, whereas #46100 is primarily a test migration with new/updated test files.", + "canonical_pr_reason": "No clear canonical PR for duplication: PR #45932 and PR #46100 address different areas and goals, so neither can serve as a duplicate representative for the other.", "best_issue_reason": null, - "best_pr_reason": "PR #45932 is the best representative of the cluster since it appears to address actual implementation behavior; #46100 is mostly test scaffolding and is less suitable as the primary duplicate target.", + "best_pr_reason": "PR #45932 is the broader, more substantive change set, but it is still unrelated to PR #46100; there is no single best PR as a duplicate canonicalization target.", "soft_edge_verdicts": [ { "left": "pull_request:45932", "right": "pull_request:46100", "accept": false, - "reason": "Reject: these PRs touch different kinds of work and do not appear to fix the same concrete code-path problem. #45932 changes model/config code and workflow secrets; #46100 migrates/adds tests. They would not plausibly be merged as one PR." + "reason": "Different changesets and purposes: #45932 updates CI/workflow and model code, while #46100 migrates ViT tests. They do not appear to fix the same concrete code-path problem." } ] }, @@ -22689,8 +22439,8 @@ }, { "label": "pull_request", - "component_index": 29, - "component_count": 39, + "component_index": 28, + "component_count": 35, "review_unit_index": 1, "review_unit_count": 1, "status": "reviewed", @@ -22709,12 +22459,12 @@ "pull_request:46073|pull_request:46196", "pull_request:46154|pull_request:46196" ], - "prepared_review_unit_hash": "4d23de808b5bf69ef4d040528ef2ebbe3a995319963c3332138342c44746eba9", + "prepared_review_unit_hash": "caa04397d8d773a43aa91a204e9bba7f9274282dcb8740c12c8fd8b592641e46", "component_budget": { "node_count": 3, "item_count": 3, "soft_pair_count": 3, - "serialized_chars": 8400, + "serialized_chars": 8398, "estimated_input_tokens": 2100, "estimated_eval_tokens": 4456 }, @@ -22722,7 +22472,7 @@ "node_count": 3, "item_count": 3, "soft_pair_count": 3, - "serialized_chars": 8400, + "serialized_chars": 8398, "estimated_input_tokens": 2100, "estimated_eval_tokens": 4456 }, @@ -22731,30 +22481,30 @@ "aggressively_trimmed": false, "split": false, "analyst_result": { - "summary": "All three PRs are about adding attention-implementation dispatch / SDPA support, but they target different model families and separate code paths, so they are related feature work rather than duplicates.", - "confidence": 0.96, + "summary": "These are three separate PRs adding SDPA/attn_implementation dispatch support to different model families (CTRL, T5, wav2vec2). They share a theme but touch different code paths and are not duplicates.", + "confidence": 0.97, "canonical_issue_reason": null, - "canonical_pr_reason": "No single PR is a true duplicate canonical here; if one representative is needed, PR 46154 is the clearest focused example of SDPA dispatch support with tests for one concrete model path.", + "canonical_pr_reason": null, "best_issue_reason": null, - "best_pr_reason": "PR 46154 is the best representative of the cluster because it is narrowly scoped, explicitly about SDPA dispatch, and includes test coverage, making it the cleanest exemplar of this change pattern.", + "best_pr_reason": "PR 46154 is the cleanest representative of the theme: a focused T5 SDPA dispatch change with accompanying tests, without the breadth of the wav2vec2 sweep.", "soft_edge_verdicts": [ { "left": "pull_request:46073", "right": "pull_request:46154", "accept": false, - "reason": "CTRL and T5 are different model implementations with distinct forward paths; same general feature, but not the same underlying bug/change and not mergeable into one PR." + "reason": "Different model implementations and code paths (CTRL vs T5); same broad feature, not the same fix and not something to merge into one PR." }, { "left": "pull_request:46073", "right": "pull_request:46196", "accept": false, - "reason": "CTRL and wav2vec2/audio models use different code paths and files; this is shared SDPA support work across model families, not a duplicate fix." + "reason": "CTRL attention dispatch and wav2vec2 attention dispatch are separate model-specific changes; no shared concrete bug/change." }, { "left": "pull_request:46154", "right": "pull_request:46196", "accept": false, - "reason": "T5 and wav2vec2 are unrelated model families; both add attention dispatch support, but they do not fix the same concrete problem in the same code path." + "reason": "T5 and wav2vec2 are unrelated model code paths; both add SDPA support but for distinct implementations, so not duplicates." } ] }, @@ -22767,8 +22517,8 @@ }, { "label": "pull_request", - "component_index": 30, - "component_count": 39, + "component_index": 29, + "component_count": 35, "review_unit_index": 1, "review_unit_count": 1, "status": "reviewed", @@ -22784,7 +22534,7 @@ "soft_pairs": [ "pull_request:46126|pull_request:46269" ], - "prepared_review_unit_hash": "9f98feff2ff22bc1cb7241195c4c4290b776f634329b2f06a3b74c28a402d419", + "prepared_review_unit_hash": "cc1f61506ec4ee537fe3117710f22b7d44c8105126b3438e35d819ab8005b06b", "component_budget": { "node_count": 2, "item_count": 2, @@ -22806,18 +22556,18 @@ "aggressively_trimmed": false, "split": false, "analyst_result": { - "summary": "These are not the same change set. PR #46126 is a focused EP/Trainer integration with RouterParallel changes, while PR #46269 is a broad distributed branch base touching many subsystems. The shared distributed filenames are not enough to treat them as duplicates.", - "confidence": 0.94, + "summary": "Two open PRs overlap on distributed/trainer files, but they are not the same concrete change: one is a focused EP + Trainer integration, the other is a broad distributed branch base touching many unrelated subsystems.", + "confidence": 0.88, "canonical_issue_reason": null, - "canonical_pr_reason": "PR #46126 is the more specific, self-contained change: it targets EP + Trainer integration and introduces concrete RouterParallel behavior. PR #46269 is a large umbrella branch with many distributed updates, so it is a poor canonical representative for this cluster.", + "canonical_pr_reason": "pull_request:46269 is the broader umbrella distributed branch-base PR; it spans the shared distributed files and substantially more, making it the best central representative of this cluster.", "best_issue_reason": null, - "best_pr_reason": "PR #46126 is the best representative because it has a clear, narrow purpose and the overlapping files are directly tied to its feature. PR #46269 is too broad and looks like a base/integration branch rather than a single fix.", + "best_pr_reason": "pull_request:46269 is the strongest representative of the cluster because it is the larger base-change PR and not just a narrow follow-on integration.", "soft_edge_verdicts": [ { "left": "pull_request:46126", "right": "pull_request:46269", "accept": false, - "reason": "Shared files are in the distributed stack, but the PRs do not appear to fix the same concrete bug or implement the same change. #46126 is a targeted EP/Trainer feature PR; #46269 is a broad distributed branch base with many unrelated modifications." + "reason": "They share some distributed filenames, but the scopes differ: 46126 is a focused EP/Trainer integration, while 46269 is a broad distributed branch base. They do not look like the same underlying fix and would not plausibly merge as one PR." } ] }, @@ -22830,8 +22580,8 @@ }, { "label": "pull_request", - "component_index": 31, - "component_count": 39, + "component_index": 30, + "component_count": 35, "review_unit_index": 1, "review_unit_count": 1, "status": "reviewed", @@ -22879,7 +22629,7 @@ "pull_request:46504|pull_request:46563", "pull_request:46432|pull_request:46563" ], - "prepared_review_unit_hash": "7cf3e4ff89c417dae834efd87ef178ae9dd7b3d8901ad9650935a178ec535533", + "prepared_review_unit_hash": "ff01194cad2dc6db6c64ed80d8b5bbc48268c1eeb703cda7baf06419930cceff", "component_budget": { "node_count": 9, "item_count": 9, @@ -22901,168 +22651,174 @@ "aggressively_trimmed": false, "split": false, "analyst_result": { - "summary": "Mostly unrelated workflow-chore PRs were grouped by title similarity. The only tight subcluster is the pr-ci-security-gate.yml edits, but those PRs make different changes (security checks vs permissions/approval gating vs comments), so I would not treat them as duplicates.", - "confidence": 0.95, + "summary": "These are mostly unrelated workflow-maintenance PRs that only look similar because of shared chore-style titles and YAML edits. The only close group is the pr-ci-security-gate.yml trio, but each PR changes a different part of that workflow, so they should not be merged as duplicates.", + "confidence": 0.91, "canonical_issue_reason": null, - "canonical_pr_reason": "pull_request:46554 is the most substantive and representative workflow-security change in the cluster; it adds the actual security checks, while the others are smaller or unrelated workflow tweaks.", + "canonical_pr_reason": "PR #46554 is the most substantive and central pr-ci-security-gate workflow change; the others are narrower follow-up tweaks or unrelated workflow edits.", "best_issue_reason": null, - "best_pr_reason": "pull_request:46554 is the best representative PR because it contains the concrete behavior change for the security gate and is more meaningful than the other single-file workflow maintenance PRs.", + "best_pr_reason": "PR #46554 is the best representative of the cluster because it contains the real security-gate logic change, while the rest are smaller or different workflow updates.", "soft_edge_verdicts": [ { "left": "pull_request:46138", "right": "pull_request:46554", "accept": false, - "reason": "Different workflows and different fixes; one updates stdin handling for self-comment CI, the other hardens the PR CI security gate." + "reason": "Different workflows and different fixes: self-comment CI stdin handling vs PR CI security-gate logic." }, { "left": "pull_request:46138", "right": "pull_request:46558", "accept": false, - "reason": "Unrelated workflows; no shared bug or change beyond both being YAML workflow edits." + "reason": "Different workflows and different concerns; no shared bug/change beyond both being YAML workflow edits." }, { "left": "pull_request:46138", "right": "pull_request:46563", "accept": false, - "reason": "Different workflows and purposes; comments about the security gate are not the same change as self-comment CI quoting." + "reason": "Different workflows; one updates self-comment CI, the other adjusts the PR CI security-gate reusable workflow note." }, { "left": "pull_request:46138", "right": "pull_request:46504", "accept": false, - "reason": "Both touch shell quoting in workflows, but they fix different scripts and different execution paths." + "reason": "Unrelated workflow changes: self-comment CI messaging vs build-ci-docker-images commit-message handling." }, { "left": "pull_request:46138", "right": "pull_request:46432", "accept": false, - "reason": "Completely different workflow files and goals; no common underlying bug." + "reason": "Different workflows and different code paths: self-comment CI vs ssh-runner hardening." }, { "left": "pull_request:46432", "right": "pull_request:46503", "accept": false, - "reason": "Different workflows and unrelated behavior changes; one is SSH runner setup, the other benchmark trigger logic." + "reason": "Different workflows: ssh-runner hardening vs benchmark_v2 trigger changes." }, { "left": "pull_request:46432", "right": "pull_request:46501", "accept": false, - "reason": "Same broad CI area, but different workflows and unrelated fixes." + "reason": "Both touch workflows, but the changes are unrelated: SSH runner secret handling vs anti-slop permissions." }, { "left": "pull_request:46432", "right": "pull_request:46502", "accept": false, - "reason": "Different workflow and different behavior; not the same concrete issue." + "reason": "Different workflow files and different behavior; ssh-runner setup is not the same as assign-reviewers checkout ref." }, { "left": "pull_request:46432", "right": "pull_request:46554", "accept": false, - "reason": "Different workflows and different problems; no duplicate code-path." + "reason": "Different workflows and different fixes: SSH runner hygiene vs PR CI security gate implementation." }, { "left": "pull_request:46432", "right": "pull_request:46558", "accept": false, - "reason": "No shared underlying bug; only superficial workflow-chore similarity." + "reason": "Different workflows and unrelated changes; no shared underlying bug." + }, + { + "left": "pull_request:46432", + "right": "pull_request:46563", + "accept": false, + "reason": "Different workflows and unrelated changes; ssh-runner secret cleanup is not the PR CI security-gate note update." }, { "left": "pull_request:46501", "right": "pull_request:46503", "accept": false, - "reason": "Both are workflow tweaks, but they affect different jobs and different conditions." + "reason": "Separate workflows with different purposes: anti-slop permissions vs benchmark trigger conditions." }, { "left": "pull_request:46501", "right": "pull_request:46502", "accept": false, - "reason": "Different workflow files and unrelated fixes." + "reason": "Different workflows and unrelated fixes; no common concrete bug." }, { "left": "pull_request:46501", "right": "pull_request:46504", "accept": false, - "reason": "Same general shell/YAML maintenance flavor, but different workflows and distinct changes." + "reason": "Unrelated workflow updates: anti-slop permissions vs build-ci-docker-images message parsing." }, { "left": "pull_request:46501", "right": "pull_request:46554", "accept": false, - "reason": "Different workflows and unrelated changes; not the same bug or PR." + "reason": "Different workflows and different concerns; not the same change." }, { "left": "pull_request:46502", "right": "pull_request:46503", "accept": false, - "reason": "Different workflow behaviors; one changes checkout base ref, the other dispatch conditions." + "reason": "Different workflows and different behavior changes; not duplicates." }, { "left": "pull_request:46502", "right": "pull_request:46504", "accept": false, - "reason": "Distinct workflows and fixes; no same underlying problem." + "reason": "Assign-reviewers checkout ref change is unrelated to build-ci-docker-images message handling." }, { "left": "pull_request:46502", "right": "pull_request:46554", "accept": false, - "reason": "Unrelated workflow files and different functionality." + "reason": "Different workflows and different code paths; no same underlying bug/change." }, { "left": "pull_request:46503", "right": "pull_request:46504", "accept": false, - "reason": "One changes benchmark triggers, the other changes shell handling in a build workflow." + "reason": "Different workflows: benchmark dispatch trigger vs build image commit-message parsing." }, { "left": "pull_request:46503", "right": "pull_request:46554", "accept": false, - "reason": "Different workflows and separate concerns." + "reason": "Different workflows and different operational concerns; not duplicates." }, { "left": "pull_request:46504", "right": "pull_request:46554", "accept": false, - "reason": "Both are workflow changes, but they fix different scripts and different code paths." + "reason": "Different workflows and fixes: build-ci-docker-images parsing vs PR CI security-gate logic." }, { "left": "pull_request:46504", "right": "pull_request:46558", "accept": false, - "reason": "No shared bug; only the same high-level repository workflow area." + "reason": "Different workflows and unrelated changes; both are generic chore updates only superficially similar." }, { "left": "pull_request:46554", "right": "pull_request:46558", "accept": false, - "reason": "Same file, but different changes: 46554 adds security checks, while 46558 changes permissions and approval gating." + "reason": "Same file, but different concrete changes: security-check logic versus permissions/approval gating." }, { "left": "pull_request:46554", "right": "pull_request:46563", "accept": false, - "reason": "46563 is comment-only guidance, not the same change as the security-check logic in 46554." + "reason": "Same file, but different concrete changes: security-check logic versus reusable-workflow commentary/annotation." }, { "left": "pull_request:46558", "right": "pull_request:46563", "accept": false, - "reason": "Different edits to the same workflow file, but not the same concrete fix." + "reason": "Same workflow file, but the edits address different parts of the job and are not the same bug/change." }, { "left": "pull_request:46504", "right": "pull_request:46563", "accept": false, - "reason": "Unrelated workflow changes; no duplicate bug or change." + "reason": "Different workflows and different fixes; no duplicate evidence beyond generic chore similarity." }, { "left": "pull_request:46432", "right": "pull_request:46563", "accept": false, - "reason": "Different workflow and unrelated purpose." + "reason": "Different workflows and unrelated changes; not the same underlying issue." } ] }, @@ -23075,8 +22831,8 @@ }, { "label": "pull_request", - "component_index": 32, - "component_count": 39, + "component_index": 31, + "component_count": 35, "review_unit_index": 1, "review_unit_count": 1, "status": "reviewed", @@ -23092,7 +22848,7 @@ "soft_pairs": [ "pull_request:46152|pull_request:46207" ], - "prepared_review_unit_hash": "259f32dc9e24c4c2e2bca97d477c52edc7bc8fed58b36703f08f9f6c49014632", + "prepared_review_unit_hash": "5341681ce81f1dd91db7614ce5d8dca6d323589e533298b0830c23cf33363d54", "component_budget": { "node_count": 2, "item_count": 2, @@ -23114,18 +22870,18 @@ "aggressively_trimmed": false, "split": false, "analyst_result": { - "summary": "These two PRs are related by a shared pattern of replacing file-open source introspection, but they fix different code paths in different modules and are not duplicates.", - "confidence": 0.93, + "summary": "Two PRs share a broad theme of replacing file-open-based source inspection, but they target different modules and code paths, so they are not duplicates.", + "confidence": 0.88, "canonical_issue_reason": null, - "canonical_pr_reason": "PR #46207 is the more concrete, self-contained fix: it updates _can_set_attn_implementation in modeling_utils.py and includes tests. PR #46152 is a separate refactor in import_utils.py for resolving package resources, so it is less representative of the cluster's main user-facing behavior.", + "canonical_pr_reason": "PR 46152 is the more general and central change: it replaces self-source file opening in the import-utils helper, with a broader infrastructure-oriented fix rather than a model-specific heuristic tweak.", "best_issue_reason": null, - "best_pr_reason": "PR #46207 is the stronger standalone candidate because it targets a specific bug-prone source-reading path and has accompanying tests, whereas #46152 is broader infrastructure work in a different file.", + "best_pr_reason": "PR 46152 is the best representative of the cluster because its scope is cleaner and broader, while PR 46207 is a narrower fix in a separate attention-implementation path.", "soft_edge_verdicts": [ { "left": "pull_request:46152", "right": "pull_request:46207", "accept": false, - "reason": "Do not merge: they address different functions in different modules (_resolve_traversable in import_utils.py vs _can_set_attn_implementation in modeling_utils.py). The shared theme is only replacing open() for source inspection, not the same underlying bug or code path." + "reason": "Same motivation, but not the same underlying bug/change: one rewrites source resolution in src/transformers/utils/import_utils.py, the other changes _can_set_attn_implementation in src/transformers/modeling_utils.py. Different code paths, different behaviors, not plausibly one merged PR." } ] }, @@ -23138,8 +22894,8 @@ }, { "label": "pull_request", - "component_index": 33, - "component_count": 39, + "component_index": 32, + "component_count": 35, "review_unit_index": 1, "review_unit_count": 1, "status": "reviewed", @@ -23157,7 +22913,7 @@ "pull_request:46436|pull_request:46474", "pull_request:46173|pull_request:46474" ], - "prepared_review_unit_hash": "53b65905e8aac336713937a04d1e34a14081b7230b0b37d854ebc8ad66c6a8ac", + "prepared_review_unit_hash": "03afe1d84421cabfe7d47addc353812dae48716212e2b9538554158c2840c089", "component_budget": { "node_count": 3, "item_count": 3, @@ -23179,30 +22935,30 @@ "aggressively_trimmed": false, "split": false, "analyst_result": { - "summary": "Two PRs are near-duplicates fixing the same StaticLayer.get_seq_length type-hint change; the cache max-len PR is a separate generation bug and should not be grouped with them.", - "confidence": 0.97, + "summary": "PRs 46173 and 46474 are the same StaticLayer.get_seq_length type-annotation fix in cache_utils.py and should be treated as duplicates. PR 46436 is unrelated; it changes static cache length selection in generation, not the return type annotation.", + "confidence": 0.98, "canonical_issue_reason": null, - "canonical_pr_reason": "PR #46173 is the best canonical representative: it is the earliest, directly targets issue #45987, and makes the exact StaticLayer.get_seq_length return-type update in src/transformers/cache_utils.py.", + "canonical_pr_reason": "PR 46474 is the cleanest representative of the duplicate pair: it makes the same cache_utils.py type-hint change as 46173, targets the same issue (#45987), and has the simpler, more direct title.", "best_issue_reason": null, - "best_pr_reason": "PR #46173 is the strongest overall representative for the cluster because it is the earliest and cleanest instance of the same concrete code change; PR #46474 is a later duplicate of the same type-hint fix, while PR #46436 addresses a different static-cache configuration bug.", + "best_pr_reason": "PR 46474 is the best cluster representative because it is the same concrete fix as 46173 with the same issue target and identical diff, while being more succinctly titled. 46436 addresses a different generation-cache behavior change and is not a candidate for canonical PR.", "soft_edge_verdicts": [ { "left": "pull_request:46436", "right": "pull_request:46474", "accept": false, - "reason": "Different bugs: #46436 changes static cache max-len handling in generation/utils.py, while #46474 only adjusts the StaticLayer.get_seq_length type hint in cache_utils.py. They share no issue target or code path and would not plausibly merge into one PR." + "reason": "Different bugs: 46436 changes max static cache length handling in generation/utils.py, while 46474 only updates the StaticLayer.get_seq_length type hint in cache_utils.py." }, { "left": "pull_request:46173", "right": "pull_request:46474", "accept": true, - "reason": "Same underlying change: both edit src/transformers/cache_utils.py to widen StaticLayer.get_seq_length to int | torch.Tensor and both target issue #45987. #46474 looks like a duplicate of #46173." + "reason": "Same underlying change: both modify cache_utils.py to widen StaticLayer.get_seq_length to int | torch.Tensor and both target issue 45987 with effectively identical diffs." } ] }, "evaluator_result": { "accept": true, - "feedback": "The summary is grounded in the packet: PR #46173 and PR #46474 make the same cache_utils.py type-hint change and share issue #45987, while PR #46436 touches generation/utils.py for a different cache-length configuration issue (#46424). The soft-edge verdicts are conservative and the mergeability judgment for the PR pair is well supported." + "feedback": "The summary is grounded in the packet. PRs 46173 and 46474 appear to be the same bugfix: both change StaticLayer.get_seq_length in src/transformers/cache_utils.py from int to int | torch.Tensor and share issue target 45987. The soft-edge rejection for 46436 vs 46474 is also conservative, since 46436 changes static cache length handling in generation/utils.py and tests, which is a different fix." }, "evaluator_used": true, "retried": false, @@ -23212,71 +22968,8 @@ }, { "label": "pull_request", - "component_index": 34, - "component_count": 39, - "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-46236-2", - "nodes": [ - "pull_request:46236", - "pull_request:46621" - ], - "soft_pairs": [ - "pull_request:46236|pull_request:46621" - ], - "prepared_review_unit_hash": "9d3f0d70fb9c4b89cabc17a817edab5d35e330effac72155f3ee825216468dbe", - "component_budget": { - "node_count": 2, - "item_count": 2, - "soft_pair_count": 1, - "serialized_chars": 5922, - "estimated_input_tokens": 1481, - "estimated_eval_tokens": 3218 - }, - "budget": { - "node_count": 2, - "item_count": 2, - "soft_pair_count": 1, - "serialized_chars": 5922, - "estimated_input_tokens": 1481, - "estimated_eval_tokens": 3218 - }, - "overflow_policy": "truncate_then_skip", - "trimmed": false, - "aggressively_trimmed": false, - "split": false, - "analyst_result": { - "summary": "These are unrelated documentation translation PRs: one adds a Bengali docs bundle, the other adds a Chinese kernels page. They do not look like the same underlying change.", - "confidence": 0.98, - "canonical_issue_reason": null, - "canonical_pr_reason": "No clear duplicate canonical PR here; if forced to anchor the cluster, PR 46236 is the broader Bengali i18n bundle and has an explicit issue link, while 46621 is a separate zh kernels translation.", - "best_issue_reason": null, - "best_pr_reason": "PR 46236 is the stronger representative only by scope and linkage, but it is not a duplicate of 46621.", - "soft_edge_verdicts": [ - { - "left": "pull_request:46236", - "right": "pull_request:46621", - "accept": false, - "reason": "Different languages and different content areas (Bengali docs bundle vs Chinese kernels page); not the same fix/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": 35, - "component_count": 39, + "component_index": 33, + "component_count": 35, "review_unit_index": 1, "review_unit_count": 1, "status": "reviewed", @@ -23292,12 +22985,12 @@ "soft_pairs": [ "pull_request:46310|pull_request:46546" ], - "prepared_review_unit_hash": "87738cd0d4ff5b4194ec3aa820de066223266cff726c170243256877b4628d97", + "prepared_review_unit_hash": "1c3688b10b1b69882cc6c78ec73505b1cd5cd21445d2469dcdd89410fe6fd89a", "component_budget": { "node_count": 2, "item_count": 2, "soft_pair_count": 1, - "serialized_chars": 4850, + "serialized_chars": 4851, "estimated_input_tokens": 1213, "estimated_eval_tokens": 2682 }, @@ -23305,7 +22998,7 @@ "node_count": 2, "item_count": 2, "soft_pair_count": 1, - "serialized_chars": 4850, + "serialized_chars": 4851, "estimated_input_tokens": 1213, "estimated_eval_tokens": 2682 }, @@ -23314,18 +23007,18 @@ "aggressively_trimmed": false, "split": false, "analyst_result": { - "summary": "Both PRs fix the same off-by-one boundary in sliding-window masking, but PR 46546 also extends the change to Gemma4 and adds tests, so it is broader than PR 46310 rather than a clean duplicate. I would not collapse them as the same PR.", - "confidence": 0.86, + "summary": "Both pull requests fix the same sliding-window boundary bug (`-dist <= right_window_size` vs `<`) in Moonshine streaming; 46546 is the broader, more complete version because it also applies the same fix to Gemma4 and adds a test.", + "confidence": 0.91, "canonical_issue_reason": null, - "canonical_pr_reason": "PR 46546 is the stronger representative because it fixes the Moonshine boundary bug and also covers the same issue in Gemma4 with tests, making it the more complete patch set.", + "canonical_pr_reason": "PR 46546 is the better canonical PR because it covers the Moonshine streaming fix plus the analogous Gemma4 change and includes test coverage, making it the more complete representative of this bug-fix pattern.", "best_issue_reason": null, - "best_pr_reason": "PR 46546 is the best overall PR because it includes the Moonshine fix plus an additional model fix and test coverage; PR 46310 is a narrower subset.", + "best_pr_reason": "PR 46546 is the best overall PR for this cluster: same concrete boundary fix as 46310, but broader scope and stronger validation.", "soft_edge_verdicts": [ { "left": "pull_request:46310", "right": "pull_request:46546", - "accept": false, - "reason": "Same boundary pattern, but not the same concrete change set: 46546 also modifies Gemma4 and adds tests, so it is broader than 46310 and should not be treated as a duplicate PR." + "accept": true, + "reason": "Same underlying bug and same code-path fix pattern in the Moonshine streaming mask boundary; 46546 is effectively a superset of 46310\u2019s change and they could plausibly have been combined into one PR." } ] }, @@ -23338,8 +23031,8 @@ }, { "label": "pull_request", - "component_index": 36, - "component_count": 39, + "component_index": 34, + "component_count": 35, "review_unit_index": 1, "review_unit_count": 1, "status": "reviewed", @@ -23355,12 +23048,12 @@ "soft_pairs": [ "pull_request:46376|pull_request:46417" ], - "prepared_review_unit_hash": "e715853421db7139d4edafdc8209a4fe0ce58bff842bd8a54ecb0518e4a78ff9", + "prepared_review_unit_hash": "c7a9e6879820f903909586bffa60cf1cdd831f9aced04773eba10ce61d311633", "component_budget": { "node_count": 2, "item_count": 2, "soft_pair_count": 1, - "serialized_chars": 5937, + "serialized_chars": 5938, "estimated_input_tokens": 1485, "estimated_eval_tokens": 3226 }, @@ -23368,7 +23061,7 @@ "node_count": 2, "item_count": 2, "soft_pair_count": 1, - "serialized_chars": 5937, + "serialized_chars": 5938, "estimated_input_tokens": 1485, "estimated_eval_tokens": 3226 }, @@ -23377,18 +23070,18 @@ "aggressively_trimmed": false, "split": false, "analyst_result": { - "summary": "These are unrelated pull requests: one adds multimodal support to continuous batching infrastructure, while the other adds the HunYuan VL model and its associated docs/tests. They share only a broad multimodal theme, not the same underlying code change.", + "summary": "The two pull requests are unrelated: one adds multimodal continuous batching support, while the other adds the HunYuan VL model with docs/tests. They should not be clustered as duplicates.", "confidence": 0.97, "canonical_issue_reason": null, - "canonical_pr_reason": "No clear canonical PR for duplication purposes; if one must be selected, PR #46417 is the more self-contained and specific change, but it is not a duplicate of #46376.", + "canonical_pr_reason": "PR 46417 is the better canonical representative because it is a self-contained model addition with substantial implementation, docs, and tests, and it is more complete/mature than the draft continuous-batching feature PR.", "best_issue_reason": null, - "best_pr_reason": "PR #46417 is the best standalone PR in this set because it represents a concrete feature addition with focused model/config/processor work and accompanying tests, whereas #46376 is a draft infrastructure change.", + "best_pr_reason": "PR 46417 is the best standalone PR in this cluster because it has a clear single feature scope and appears more ready for merge than the draft multimodal batching work.", "soft_edge_verdicts": [ { "left": "pull_request:46376", "right": "pull_request:46417", "accept": false, - "reason": "They address different problems: continuous batching multimodal support vs adding a specific HunYuan VL model. Shared multimodal context is too broad to treat as the same change, and they would not plausibly merge into one PR." + "reason": "Different underlying changes: 46376 targets continuous batching infrastructure for multimodal support, while 46417 introduces a new HunYuan VL model stack, documentation, and tests. They do not fix the same code-path problem and would not plausibly be merged into one PR." } ] }, @@ -23401,8 +23094,8 @@ }, { "label": "pull_request", - "component_index": 37, - "component_count": 39, + "component_index": 35, + "component_count": 35, "review_unit_index": 1, "review_unit_count": 1, "status": "reviewed", @@ -23418,12 +23111,12 @@ "soft_pairs": [ "pull_request:46603|pull_request:46604" ], - "prepared_review_unit_hash": "b89821189429bc17c4ddcd35ac75c61868305757a9ebded3589ee5aad7bad0dd", + "prepared_review_unit_hash": "f50cc3fb4253f30f4a9a7a7dc2a59e38e226d664e4af085cfab961fa44393cc7", "component_budget": { "node_count": 2, "item_count": 2, "soft_pair_count": 1, - "serialized_chars": 5495, + "serialized_chars": 5494, "estimated_input_tokens": 1374, "estimated_eval_tokens": 3004 }, @@ -23431,7 +23124,7 @@ "node_count": 2, "item_count": 2, "soft_pair_count": 1, - "serialized_chars": 5495, + "serialized_chars": 5494, "estimated_input_tokens": 1374, "estimated_eval_tokens": 3004 }, @@ -23440,18 +23133,18 @@ "aggressively_trimmed": false, "split": false, "analyst_result": { - "summary": "Two open Mistral PRs share package/test files but target different changes: one is a refactor moving MistralConverter into the integrations package, the other adds tekken tokenizer detection/conversion/save support. They are related, but not duplicates and should not be merged as the same PR.", - "confidence": 0.95, + "summary": "Two open Mistral-related PRs, but they are different changes: one refactors/moves the existing MistralConverter into a package, the other adds tekken tokenizer detection/conversion/save support. Shared files alone are not enough to treat them as duplicates.", + "confidence": 0.92, "canonical_issue_reason": null, - "canonical_pr_reason": "PR #46604 is the broader, more user-facing change in the Mistral tokenizer path; #46603 is a narrower refactor/move of an existing converter and does not represent the same underlying change.", + "canonical_pr_reason": "PR 46604 is the more substantial, feature-bearing change; PR 46603 is a narrower code move/refactor of MistralConverter.", "best_issue_reason": null, - "best_pr_reason": "PR #46604 is the best representative of the cluster because it contains the substantive tokenizer functionality work, while #46603 is a supporting code organization change.", + "best_pr_reason": "PR 46604 best represents the cluster because it introduces the main new tokenizer functionality, while PR 46603 is just organizational cleanup.", "soft_edge_verdicts": [ { "left": "pull_request:46603", "right": "pull_request:46604", "accept": false, - "reason": "They touch the same Mistral area and files, but the concrete changes differ: #46603 moves MistralConverter into the integrations package, while #46604 adds tekken tokenizer detection, conversion, and save utilities. Different code-path problems, so not duplicates." + "reason": "Both touch Mistral tokenizer code, but one is a refactor/move of MistralConverter and the other adds tekken detection/conversion/save utilities. They are different underlying changes and would not plausibly be merged as one PR." } ] }, @@ -23461,138 +23154,6 @@ "accepted_nontrivial_soft_edge": false, "error_kind": null, "error_message": null - }, - { - "label": "pull_request", - "component_index": 38, - "component_count": 39, - "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-46615-2", - "nodes": [ - "pull_request:46615", - "pull_request:46639" - ], - "soft_pairs": [ - "pull_request:46615|pull_request:46639" - ], - "prepared_review_unit_hash": "623cb87b3fb1e339140884de2d0b2fa181f4d4c8dc3edc0cba163a9b3726af9e", - "component_budget": { - "node_count": 2, - "item_count": 2, - "soft_pair_count": 1, - "serialized_chars": 4336, - "estimated_input_tokens": 1084, - "estimated_eval_tokens": 2424 - }, - "budget": { - "node_count": 2, - "item_count": 2, - "soft_pair_count": 1, - "serialized_chars": 4336, - "estimated_input_tokens": 1084, - "estimated_eval_tokens": 2424 - }, - "overflow_policy": "truncate_then_skip", - "trimmed": false, - "aggressively_trimmed": false, - "split": false, - "analyst_result": { - "summary": "Both PRs target the same generation/beam-search cache reordering bug in transformers and point to the same issue, with 46615 being the more complete fix and 46639 a narrower variant.", - "confidence": 0.92, - "canonical_issue_reason": "Issue 46612 is the shared explicit target for both PRs and appears to be the underlying bug report driving the work.", - "canonical_pr_reason": "PR 46615 is the stronger canonical PR because it implements the broader cache-reorder fix with helper logic and tests, covering non-standard cache types directly.", - "best_issue_reason": "Issue 46612 is the only issue in the cluster and is the central reference for the bug being fixed.", - "best_pr_reason": "PR 46615 is the best representative because it provides the more complete, test-backed implementation for the beam-search cache reorder problem.", - "soft_edge_verdicts": [ - { - "left": "pull_request:46615", - "right": "pull_request:46639", - "accept": true, - "reason": "Both PRs fix the same underlying beam-search cache reordering bug for non-standard cache types in generation/utils.py, share the same issue target 46612, and could plausibly be merged into one change set." - } - ] - }, - "evaluator_result": { - "accept": true, - "feedback": "The summary is grounded in the packet: both PRs explicitly target issue 46612, both modify generation/utils.py to address beam-search cache reordering, and 46615 is reasonably characterized as the more complete, test-backed fix. The soft-edge accept is conservative enough, and the pair looks mergeable as alternate implementations of the same bugfix." - }, - "evaluator_used": true, - "retried": false, - "accepted_nontrivial_soft_edge": true, - "error_kind": null, - "error_message": null - }, - { - "label": "pull_request", - "component_index": 39, - "component_count": 39, - "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-46616-2", - "nodes": [ - "pull_request:46616", - "pull_request:46638" - ], - "soft_pairs": [ - "pull_request:46616|pull_request:46638" - ], - "prepared_review_unit_hash": "488cccfc5f03b1e05ab9a04e4ff9c4566630de2406b6b1e39904bc4d88b8d779", - "component_budget": { - "node_count": 2, - "item_count": 2, - "soft_pair_count": 1, - "serialized_chars": 4653, - "estimated_input_tokens": 1164, - "estimated_eval_tokens": 2584 - }, - "budget": { - "node_count": 2, - "item_count": 2, - "soft_pair_count": 1, - "serialized_chars": 4653, - "estimated_input_tokens": 1164, - "estimated_eval_tokens": 2584 - }, - "overflow_policy": "truncate_then_skip", - "trimmed": false, - "aggressively_trimmed": false, - "split": false, - "analyst_result": { - "summary": "Two open PRs target the same Wav2Vec2 phonemizer language-state bug and the same issue, with nearly identical test coverage; 46638 is the better representative because it updates the cached language after backend reinitialization, matching the intended restore flow.", - "confidence": 0.87, - "canonical_issue_reason": null, - "canonical_pr_reason": "PR 46638 is the cleaner canonical PR: it fixes the same stale phonemizer_lang state bug but places the assignment after init_backend, which better matches restoring language state when switching back.", - "best_issue_reason": null, - "best_pr_reason": "PR 46638 most clearly and plausibly fixes the concrete code path in phonemize by restoring self.phonemizer_lang after backend switching; 46616 is the same bug theme but uses a less convincing assignment order.", - "soft_edge_verdicts": [ - { - "left": "pull_request:46616", - "right": "pull_request:46638", - "accept": true, - "reason": "Both PRs address the same stale phonemizer_lang bug in the same method, touch the same files, and are alternative implementations of one fix." - } - ] - }, - "evaluator_result": { - "accept": true, - "feedback": "The summary is grounded in the packet: both PRs explicitly target issue 46614, touch the same tokenizer and test file, and differ only in the ordering of the stale `phonemizer_lang` update relative to `init_backend()`. The soft-edge accept is conservative because these look like alternative implementations of the same bugfix and can reasonably be treated as one PR cluster." - }, - "evaluator_used": true, - "retried": false, - "accepted_nontrivial_soft_edge": true, - "error_kind": null, - "error_message": null } ] } diff --git a/analysis/current/manifest.json b/analysis/current/manifest.json index 4e67c699fa56ac119ecadadd6d99d34a064be690..5ffa1682e598053a84118ac8a98c074feaff7d0a 100644 --- a/analysis/current/manifest.json +++ b/analysis/current/manifest.json @@ -1,8 +1,8 @@ { - "analysis_id": "hybrid-model-20260615t120125z", + "analysis_id": "hybrid-model-20260615t180022z", "archived_artifacts": { - "hybrid": "snapshots/20260615T120125Z/analysis-runs/hybrid-model-20260615t120125z/analysis-report-hybrid.json", - "hybrid_reviews": "snapshots/20260615T120125Z/analysis-runs/hybrid-model-20260615t120125z/analysis-report-hybrid.llm-reviews.json" + "hybrid": "snapshots/20260615T180022Z/analysis-runs/hybrid-model-20260615t180022z/analysis-report-hybrid.json", + "hybrid_reviews": "snapshots/20260615T180022Z/analysis-runs/hybrid-model-20260615t180022z/analysis-report-hybrid.llm-reviews.json" }, "artifacts": { "hybrid": "analysis/current/analysis-report-hybrid.json", @@ -10,9 +10,9 @@ }, "channel": "canonical", "model": null, - "published_at": "2026-06-15T12:12:45Z", + "published_at": "2026-06-15T18:12:05Z", "repo": "huggingface/transformers", "schema_version": 1, - "snapshot_id": "20260615T120125Z", + "snapshot_id": "20260615T180022Z", "variant": "hybrid" } diff --git a/snapshots/20260615T180022Z/analysis-runs/hybrid-model-20260615t180022z/analysis-report-hybrid.json b/snapshots/20260615T180022Z/analysis-runs/hybrid-model-20260615t180022z/analysis-report-hybrid.json new file mode 100644 index 0000000000000000000000000000000000000000..4262aa64c0da6c0ed019af9382047837cfeba124 --- /dev/null +++ b/snapshots/20260615T180022Z/analysis-runs/hybrid-model-20260615t180022z/analysis-report-hybrid.json @@ -0,0 +1,2140 @@ +{ + "schema_version": "1.0", + "repo": "huggingface/transformers", + "snapshot_id": "20260615T180022Z", + "generated_at": "2026-06-15T18:12:03Z", + "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-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-41115-3", + "summary": "Cluster of 1 issues and 2 PRs centered on issue #41115.", + "status": "open", + "confidence": 0.55, + "canonical_issue_number": 41115, + "canonical_pr_number": 45613, + "issue_numbers": [ + 41115 + ], + "pr_numbers": [ + 41116, + 45613 + ], + "evidence_types": [ + "closing_reference" + ], + "pr_comparisons": [ + { + "left_pr_number": 41116, + "right_pr_number": 45613, + "code_similarity": 0.797, + "size_similarity": 0.885, + "file_overlap": 0.75, + "area_overlap": 0.831, + "patch_similarity": 0.693, + "shared_filenames": [ + "docs/source/en/_toctree.yml", + "docs/source/en/model_doc/minicpm3.md", + "src/transformers/models/auto/modeling_auto.py", + "src/transformers/models/minicpm3/__init__.py", + "src/transformers/models/minicpm3/configuration_minicpm3.py", + "src/transformers/models/minicpm3/modeling_minicpm3.py", + "src/transformers/models/minicpm3/modular_minicpm3.py", + "tests/models/minicpm3/__init__.py", + "tests/models/minicpm3/test_modeling_minicpm3.py" + ], + "shared_file_areas": [ + { + "filename": "docs/source/en/_toctree.yml", + "left_ranges": [ + [ + 713, + 720 + ] + ], + "right_ranges": [ + [ + 730, + 737 + ] + ] + }, + { + "filename": "docs/source/en/model_doc/minicpm3.md", + "left_ranges": [ + [ + 1, + 67 + ] + ], + "right_ranges": [ + [ + 1, + 45 + ] + ] + }, + { + "filename": "src/transformers/models/auto/modeling_auto.py", + "left_ranges": [ + [ + 285, + 291 + ], + [ + 700, + 706 + ], + [ + 1298, + 1304 + ] + ], + "right_ranges": [ + [ + 720, + 726 + ], + [ + 1331, + 1337 + ] + ] + }, + { + "filename": "src/transformers/models/minicpm3/__init__.py", + "left_ranges": [ + [ + 1, + 29 + ] + ], + "right_ranges": [ + [ + 1, + 29 + ] + ] + }, + { + "filename": "src/transformers/models/minicpm3/configuration_minicpm3.py", + "left_ranges": [ + [ + 1, + 141 + ] + ], + "right_ranges": [ + [ + 1, + 126 + ] + ] + }, + { + "filename": "src/transformers/models/minicpm3/modeling_minicpm3.py", + "left_ranges": [ + [ + 1, + 544 + ] + ], + "right_ranges": [ + [ + 1, + 522 + ] + ] + }, + { + "filename": "src/transformers/models/minicpm3/modular_minicpm3.py", + "left_ranges": [ + [ + 1, + 444 + ] + ], + "right_ranges": [ + [ + 1, + 342 + ] + ] + }, + { + "filename": "tests/models/minicpm3/test_modeling_minicpm3.py", + "left_ranges": [ + [ + 1, + 109 + ] + ], + "right_ranges": [ + [ + 1, + 136 + ] + ] + } + ] + } + ] + }, + { + "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-45987-3", + "summary": "Cluster of 1 issues and 2 PRs centered on issue #45987.", + "status": "open", + "confidence": 0.75, + "canonical_issue_number": 45987, + "canonical_pr_number": 46173, + "issue_numbers": [ + 45987 + ], + "pr_numbers": [ + 46173, + 46474 + ], + "evidence_types": [ + "closing_reference", + "shared_issue_target" + ], + "pr_comparisons": [ + { + "left_pr_number": 46173, + "right_pr_number": 46474, + "code_similarity": 0.65, + "size_similarity": 1.0, + "file_overlap": 1.0, + "area_overlap": 0.0, + "patch_similarity": 1.0, + "shared_filenames": [ + "src/transformers/cache_utils.py" + ], + "shared_file_areas": [ + { + "filename": "src/transformers/cache_utils.py", + "left_ranges": [ + [ + 69, + 75 + ], + [ + 375, + 381 + ] + ], + "right_ranges": [ + [ + 79, + 85 + ], + [ + 461, + 467 + ] + ] + } + ] + } + ] + }, + { + "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": 0.494, + "size_similarity": 0.619, + "file_overlap": 0.5, + "area_overlap": 0.434, + "patch_similarity": 0.451, + "shared_filenames": [ + "src/transformers/testing_utils.py", + "tests/utils/test_testing_utils.py" + ], + "shared_file_areas": [ + { + "filename": "src/transformers/testing_utils.py", + "left_ranges": [ + [ + 3595, + 3602 + ], + [ + 3819, + 3853 + ] + ], + "right_ranges": [ + [ + 3525, + 3558 + ], + [ + 3782, + 3788 + ] + ] + }, + { + "filename": "tests/utils/test_testing_utils.py", + "left_ranges": [ + [ + 1, + 86 + ] + ], + "right_ranges": [ + [ + 1, + 114 + ] + ] + } + ] + } + ] + }, + { + "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": 46546, + "issue_numbers": [ + 46305 + ], + "pr_numbers": [ + 46310, + 46546 + ], + "evidence_types": [ + "closing_reference", + "shared_issue_target", + "soft_similarity" + ], + "pr_comparisons": [ + { + "left_pr_number": 46310, + "right_pr_number": 46546, + "code_similarity": 0.566, + "size_similarity": 0.444, + "file_overlap": 0.5, + "area_overlap": 0.867, + "patch_similarity": 0.16, + "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": [ + [ + 355, + 361 + ] + ] + }, + { + "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-44018-2", + "summary": "Cluster of 2 related pull requests linked by soft_similarity.", + "status": "open", + "confidence": 0.5, + "canonical_issue_number": null, + "canonical_pr_number": 44068, + "issue_numbers": [], + "pr_numbers": [ + 44018, + 44068 + ], + "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 + ] + ] + } + ] + } + ] + } + ], + "duplicate_issues": [], + "duplicate_prs": [ + { + "cluster_id": "cluster-41115-3", + "canonical_pr_number": 45613, + "duplicate_pr_numbers": [ + 41116 + ], + "target_issue_number": 41115, + "reason": "PRs in cluster-41115-3 are treated as duplicates because they converge on issue #41115 with closing_reference evidence." + }, + { + "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-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-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, + "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-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-45987-3", + "canonical_pr_number": 46173, + "duplicate_pr_numbers": [ + 46474 + ], + "target_issue_number": 45987, + "reason": "PRs in cluster-45987-3 are treated as duplicates because they converge on issue #45987 with closing_reference, shared_issue_target evidence." + }, + { + "cluster_id": "cluster-46305-3", + "canonical_pr_number": 46546, + "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." + } + ], + "best_issue": { + "cluster_id": "cluster-41211-3", + "issue_number": 41211, + "reason": "Issue #41211 is the strongest global issue candidate because it is open, belongs to a cluster with 3 artifacts, and carries 6 discussion comments plus 6 inbound references.", + "score": 148.0 + }, + "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/20260615T180022Z/analysis-runs/hybrid-model-20260615t180022z/analysis-report-hybrid.llm-reviews.json b/snapshots/20260615T180022Z/analysis-runs/hybrid-model-20260615t180022z/analysis-report-hybrid.llm-reviews.json new file mode 100644 index 0000000000000000000000000000000000000000..abc8a3be0b81ff7da77775d2ebde954c9cc0753b --- /dev/null +++ b/snapshots/20260615T180022Z/analysis-runs/hybrid-model-20260615t180022z/analysis-report-hybrid.llm-reviews.json @@ -0,0 +1,23159 @@ +{ + "schema_version": "1.0", + "repo": "huggingface/transformers", + "snapshot_id": "20260615T180022Z", + "generated_at": "2026-06-15T18:12:03Z", + "model": "gpt-5.4-mini?service_tier=flex", + "reviews": [ + { + "label": "issue", + "component_index": 1, + "component_count": 21, + "review_unit_index": 1, + "review_unit_count": 56, + "status": "reviewed", + "reason": null, + "source": "llm", + "cache_hit": false, + "model": "gpt-5.4-mini?service_tier=flex", + "cluster_id": "cluster-10000-473", + "nodes": [ + "issue:24643", + "issue:30277", + "issue:34634", + "issue:34928", + "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:43827", + "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" + ], + "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:44509|issue:44512", + "issue:35545|issue:36010", + "issue:44918|issue:46165", + "issue:44805|issue:45823", + "issue:43638|issue:44805", + "issue:43638|issue:46165", + "issue:43596|issue:44805", + "issue:44805|issue:46294", + "issue:44805|issue:45161", + "issue:46165|issue:46468", + "issue:44560|issue:46165", + "issue:43596|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:44841|issue:46661", + "issue:34928|issue:44805", + "issue:24643|issue:30277", + "issue:45859|issue:45902", + "issue:43638|issue:45161", + "issue:34928|issue:36331", + "issue:43596|issue:45161", + "issue:43716|issue:45507", + "issue:45874|issue:46294", + "issue:44292|issue:45237", + "issue:39290|issue:45081", + "issue:45823|issue:45874", + "issue:45800|issue:46515", + "issue:43425|issue:44292", + "issue:34928|issue:41720", + "issue:43828|issue:45161", + "issue:43541|issue:43716", + "issue:43828|issue:44292", + "issue:43716|issue:46468", + "issue:43716|issue:45685", + "issue:44292|issue:44918", + "issue:43716|issue:44292", + "issue:45507|issue:45685", + "issue:44918|issue:45161", + "issue:43039|issue:44918", + "issue:41762|issue:43638", + "issue:43827|issue:44512", + "issue:43828|issue:45685", + "issue:44918|issue:45685", + "issue:43039|issue:43828", + "issue:30277|issue:36331", + "issue:34928|issue:36010", + "issue:45237|issue:46015", + "issue:43039|issue:44805" + ], + "prepared_review_unit_hash": "506a732315b3892b975a77fe12f29e851a066544837e3d2b6130d677329016d0", + "component_budget": { + "node_count": 473, + "item_count": 473, + "soft_pair_count": 1738, + "serialized_chars": 901437, + "estimated_input_tokens": 225360, + "estimated_eval_tokens": 450976 + }, + "budget": { + "node_count": 47, + "item_count": 47, + "soft_pair_count": 62, + "serialized_chars": 45730, + "estimated_input_tokens": 11433, + "estimated_eval_tokens": 23122 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": true, + "analyst_result": { + "summary": "This cluster is mostly a mix of unrelated bug reports. Only a couple of near-exact duplicate pairs stand out: the Bert/DeepSpeed ZeRO-3 IndexError reports and the torch 2.4.1 / transformers 5.8.0 incompatibility reports. No pull requests are present.", + "confidence": 0.71, + "canonical_issue_reason": "issue:45800 is the clearest and most complete standalone report of the torch 2.4.1 / transformers 5.8.0 incompatibility, making it the best representative issue in the cluster.", + "canonical_pr_reason": null, + "best_issue_reason": "issue:45800 is the best global representative: it is concrete, version-specific, and cleaner than the follow-up report it duplicates.", + "best_pr_reason": null, + "soft_edge_verdicts": [ + { + "left": "issue:44322", + "right": "issue:45464", + "accept": false, + "reason": "Different failures: missing config field in Qwen3.5 vs streaming chat/completions API error." + }, + { + "left": "issue:45237", + "right": "issue:45685", + "accept": false, + "reason": "Different models and hardware paths; GPT-OSS on AMD GPUs is unrelated to an MPS int histogram error." + }, + { + "left": "issue:43716", + "right": "issue:45237", + "accept": false, + "reason": "Different subsystems: multimodal dtype mismatch vs AMD GPU runtime failure." + }, + { + "left": "issue:43828", + "right": "issue:45237", + "accept": false, + "reason": "Autocast dtype mismatch in Phi-tiny-MoE is not the same bug as GPT-OSS on AMD GPUs." + }, + { + "left": "issue:45237", + "right": "issue:46468", + "accept": false, + "reason": "Unrelated model/runtime issues; Nemotron example breakage is not the same as GPT-OSS AMD GPU failure." + }, + { + "left": "issue:45237", + "right": "issue:45507", + "accept": false, + "reason": "Different concrete bugs: AMD GPU incompatibility vs invalid method call in GraniteMoEHybrid." + }, + { + "left": "issue:30277", + "right": "issue:35545", + "accept": false, + "reason": "Collective mismatch under ZeRO-3 is unrelated to ModernBERT ONNX export." + }, + { + "left": "issue:35545", + "right": "issue:42915", + "accept": false, + "reason": "ONNX export error and FineGrainedFP8 Qwen3Moe failure are different code paths." + }, + { + "left": "issue:35545", + "right": "issue:41720", + "accept": false, + "reason": "Export-time ONNX failure is not the same as Qwen3 auto device map cudaErrorAssert." + }, + { + "left": "issue:45901", + "right": "issue:46015", + "accept": false, + "reason": "Table QA crash vs outdated RAG examples are different issues." + }, + { + "left": "issue:44509", + "right": "issue:44512", + "accept": false, + "reason": "Both are docs cleanup, but they refer to different removed docs/topics and are not the same defect." + }, + { + "left": "issue:35545", + "right": "issue:36010", + "accept": false, + "reason": "Model export error is unrelated to the GenerationMixin import breakage." + }, + { + "left": "issue:44918", + "right": "issue:46165", + "accept": false, + "reason": "Different training problems: TRL SFT embedding unpacking vs resume-with-local-model remote code issue." + }, + { + "left": "issue:44805", + "right": "issue:45823", + "accept": false, + "reason": "Generic mask shape error and Gemma4 PLE device mismatch are different forward failures." + }, + { + "left": "issue:43638", + "right": "issue:44805", + "accept": false, + "reason": "Both are tensor/index errors, but the Bert ZeRO-3 initialization bug is a different code path from the mask-index mismatch." + }, + { + "left": "issue:43638", + "right": "issue:46165", + "accept": false, + "reason": "DeepSpeed ZeRO-3/Bert index error is unrelated to resuming training with remote code." + }, + { + "left": "issue:43596", + "right": "issue:44805", + "accept": false, + "reason": "Same broad error class, but not the same concrete bug or model path." + }, + { + "left": "issue:44805", + "right": "issue:46294", + "accept": false, + "reason": "Mask indexing failure is unrelated to Qwen3.5 video chat-template token mismatch." + }, + { + "left": "issue:44805", + "right": "issue:45161", + "accept": false, + "reason": "Different failures: mask shape mismatch vs GPT-OSS MoE tensor-parallel issue." + }, + { + "left": "issue:46165", + "right": "issue:46468", + "accept": false, + "reason": "Resume-training remote-code issue is unrelated to the broken Nemotron example." + }, + { + "left": "issue:44560", + "right": "issue:46165", + "accept": false, + "reason": "Video StopIteration bug is unrelated to resume-with-local-model remote code." + }, + { + "left": "issue:43596", + "right": "issue:46165", + "accept": false, + "reason": "Bert/ZeRO-3 IndexError and remote-code resume failure are different issues." + }, + { + "left": "issue:34634", + "right": "issue:35545", + "accept": false, + "reason": "BarkProcessor voice preset issue is unrelated to ModernBERT ONNX export." + }, + { + "left": "issue:30277", + "right": "issue:43638", + "accept": false, + "reason": "Both involve distributed training, but the collective mismatch and Bert ZeRO-3 IndexError are not the same bug." + }, + { + "left": "issue:43596", + "right": "issue:43638", + "accept": true, + "reason": "Very likely duplicate reports of the same DeepSpeed ZeRO-3/Bert IndexError on empty tensors during initialization." + }, + { + "left": "issue:43039", + "right": "issue:45161", + "accept": false, + "reason": "Liger Kernel cross_entropy being called is not the same as GPT-OSS MoE tensor-parallel failure." + }, + { + "left": "issue:43975", + "right": "issue:44457", + "accept": false, + "reason": "Detokenization bug and LoRA merge/save-load mismatch are unrelated." + }, + { + "left": "issue:44292", + "right": "issue:44457", + "accept": false, + "reason": "Qwen-3 NVFP4 runtime error and LoRA save/reload mismatch are different problems." + }, + { + "left": "issue:44841", + "right": "issue:46661", + "accept": false, + "reason": "Processor failure for Voxtral is unrelated to remote-code AutoModel config_class errors." + }, + { + "left": "issue:34928", + "right": "issue:44805", + "accept": false, + "reason": "FSDP activation-checkpointing recomputation mismatch is different from the mask-indexing error." + }, + { + "left": "issue:24643", + "right": "issue:30277", + "accept": false, + "reason": "Both mention DeepSpeed, but the weight-dimension training error and collective mismatch are different bugs." + }, + { + "left": "issue:45859", + "right": "issue:45902", + "accept": false, + "reason": "Related model/component, but one is a CPU-offload forward incompatibility and the other is a meta/to_empty uninitialized-buffer bug; not the same defect." + }, + { + "left": "issue:43638", + "right": "issue:45161", + "accept": false, + "reason": "DeepSpeed ZeRO-3/Bert index error is unrelated to GPT-OSS tensor-parallel behavior." + }, + { + "left": "issue:34928", + "right": "issue:36331", + "accept": false, + "reason": "Activation checkpointing/FSDP recomputation issue is unrelated to the custom trainer signature breakage." + }, + { + "left": "issue:43596", + "right": "issue:45161", + "accept": false, + "reason": "Bert ZeRO-3 index error is unrelated to GPT-OSS MoE tensor-parallel failure." + }, + { + "left": "issue:43716", + "right": "issue:45507", + "accept": false, + "reason": "Different model families and symptoms: multimodal dtype mismatch vs invalid method in GraniteMoEHybrid." + }, + { + "left": "issue:45874", + "right": "issue:46294", + "accept": false, + "reason": "Gemma4 reverse embedding lookup and Qwen3.5 video token mismatch are different model-specific bugs." + }, + { + "left": "issue:44292", + "right": "issue:45237", + "accept": false, + "reason": "Qwen NVFP4 runtime error and GPT-OSS AMD GPU failure are unrelated." + }, + { + "left": "issue:39290", + "right": "issue:45081", + "accept": false, + "reason": "Gemma3/vLLM missing config field and Mistral regex patch crash are different code paths." + }, + { + "left": "issue:45823", + "right": "issue:45874", + "accept": false, + "reason": "Device mismatch in Gemma4 PLE is unrelated to expensive reverse embedding lookup in Gemma4 E2B/E4B." + }, + { + "left": "issue:45800", + "right": "issue:46515", + "accept": true, + "reason": "Clear duplicate torch 2.4.1 incompatibility reports; same version-level issue with near-identical scope." + }, + { + "left": "issue:43425", + "right": "issue:44292", + "accept": false, + "reason": "Torch 2.10 compatibility issue and Qwen-3 NVFP4 runtime failure are different bugs." + }, + { + "left": "issue:34928", + "right": "issue:41720", + "accept": false, + "reason": "FSDP checkpointing mismatch is unrelated to Qwen3 auto device mapping cudaErrorAssert." + }, + { + "left": "issue:43828", + "right": "issue:45161", + "accept": false, + "reason": "Autocast dtype mismatch in Phi-tiny-MoE is not the same as GPT-OSS tensor-parallel failure." + }, + { + "left": "issue:43541", + "right": "issue:43716", + "accept": false, + "reason": "Mixtral grouped_mm tracing failure and Mistral-3 image-preprocessor dtype mismatch are different issues." + }, + { + "left": "issue:43828", + "right": "issue:44292", + "accept": false, + "reason": "Different model families and failure modes; not the same bug." + }, + { + "left": "issue:43716", + "right": "issue:46468", + "accept": false, + "reason": "Mistral-3 image dtype mismatch is unrelated to the Nemotron example bug." + }, + { + "left": "issue:43716", + "right": "issue:45685", + "accept": false, + "reason": "Different models and hardware paths; not the same underlying defect." + }, + { + "left": "issue:44292", + "right": "issue:44918", + "accept": false, + "reason": "Qwen NVFP4 runtime error and TRL embedding unpacking failure are unrelated." + }, + { + "left": "issue:43716", + "right": "issue:44292", + "accept": false, + "reason": "Different model-specific runtime errors with no concrete overlap." + }, + { + "left": "issue:45507", + "right": "issue:45685", + "accept": false, + "reason": "GraniteMoEHybrid invalid method and MPS histogram_int issue are unrelated." + }, + { + "left": "issue:44918", + "right": "issue:45161", + "accept": false, + "reason": "TRL SFT embedding unpacking issue is unrelated to GPT-OSS tensor-parallel failure." + }, + { + "left": "issue:43039", + "right": "issue:44918", + "accept": false, + "reason": "Liger Kernel cross_entropy call and Qwen3.5 embedding unpacking are different bugs." + }, + { + "left": "issue:41762", + "right": "issue:43638", + "accept": false, + "reason": "Gemma3 ZeRO-3 load failure and Bert ZeRO-3 index error are similar only superficially; not the same concrete issue." + }, + { + "left": "issue:43827", + "right": "issue:44512", + "accept": false, + "reason": "Both are docs regressions, but they cover different removed items and would not be one duplicate issue." + }, + { + "left": "issue:43828", + "right": "issue:45685", + "accept": false, + "reason": "Autocast dtype mismatch and MPS int histogram error are unrelated." + }, + { + "left": "issue:44918", + "right": "issue:45685", + "accept": false, + "reason": "TRL trainer embedding issue is unrelated to the MPS histogram bug." + }, + { + "left": "issue:43039", + "right": "issue:43828", + "accept": false, + "reason": "Different model/runtime problems; no shared concrete code-path." + }, + { + "left": "issue:30277", + "right": "issue:36331", + "accept": false, + "reason": "DeepSpeed collective mismatch and custom trainer argument error are unrelated." + }, + { + "left": "issue:34928", + "right": "issue:36010", + "accept": false, + "reason": "Activation checkpointing/FSDP issue is unrelated to the GenerationMixin import error." + }, + { + "left": "issue:45237", + "right": "issue:46015", + "accept": false, + "reason": "GPT-OSS AMD GPU failure and RAG docs example issue are unrelated." + }, + { + "left": "issue:43039", + "right": "issue:44805", + "accept": false, + "reason": "Liger Kernel cross_entropy call and mask shape mismatch are different problems." + } + ] + }, + "evaluator_result": { + "accept": true, + "feedback": "Grounded overall. The summary matches the packet: the cluster is mostly unrelated issues, with two conservative duplicate pairs accepted (43596/43638 and 45800/46515), and there are no PRs. The representative/canonical choice of issue:45800 is well supported by the titles and similarity signal." + }, + "evaluator_used": true, + "retried": false, + "accepted_nontrivial_soft_edge": true, + "error_kind": null, + "error_message": null + }, + { + "label": "issue", + "component_index": 1, + "component_count": 21, + "review_unit_index": 2, + "review_unit_count": 56, + "status": "reviewed", + "reason": null, + "source": "llm", + "cache_hit": false, + "model": "gpt-5.4-mini?service_tier=flex", + "cluster_id": "cluster-10000-473", + "nodes": [ + "issue:36010", + "issue:39290", + "issue:41093", + "issue:41720", + "issue:41762", + "issue:43039", + "issue:43296", + "issue:43366", + "issue:43404", + "issue:43425", + "issue:43531", + "issue:43541", + "issue:43572", + "issue:43596", + "issue:43638", + "issue:43792", + "issue:43828", + "issue:43854", + "issue:43866", + "issue:43872", + "issue:43901", + "issue:43975", + "issue:44292", + "issue:44387", + "issue:44512", + "issue:44560", + "issue:44589", + "issue:44841", + "issue:44863", + "issue:45005", + "issue:45020", + "issue:45070", + "issue:45084", + "issue:45237", + "issue:45362", + "issue:45399", + "issue:45464", + "issue:45685", + "issue:45800", + "issue:45812", + "issue:45823", + "issue:45901", + "issue:45920", + "issue:46015", + "issue:46294", + "issue:46304", + "issue:46402", + "issue:46661" + ], + "soft_pairs": [ + "issue:43866|issue:46402", + "issue:45800|issue:46304", + "issue:43296|issue:46304", + "issue:41762|issue:43596", + "issue:39290|issue:41720", + "issue:43901|issue:44512", + "issue:44589|issue:46304", + "issue:45237|issue:45901", + "issue:43039|issue:44560", + "issue:45823|issue:46294", + "issue:43975|issue:46015", + "issue:44841|issue:45823", + "issue:44863|issue:46015", + "issue:45020|issue:46304", + "issue:45812|issue:45920", + "issue:43866|issue:44863", + "issue:43872|issue:46304", + "issue:44292|issue:46015", + "issue:39290|issue:43531", + "issue:41093|issue:43638", + "issue:43792|issue:45901", + "issue:43854|issue:43866", + "issue:43425|issue:45070", + "issue:43975|issue:45901", + "issue:43596|issue:45800", + "issue:43792|issue:45070", + "issue:45362|issue:45464", + "issue:44292|issue:45901", + "issue:39290|issue:45823", + "issue:45464|issue:45685", + "issue:43541|issue:45685", + "issue:43541|issue:43828", + "issue:36010|issue:39290", + "issue:44841|issue:45084", + "issue:43404|issue:43866", + "issue:41093|issue:41762", + "issue:43296|issue:43366", + "issue:45399|issue:46661", + "issue:43572|issue:45070", + "issue:45800|issue:45823", + "issue:44387|issue:45005" + ], + "prepared_review_unit_hash": "217bb804c7e222b525afc7cbc17e697297d5b8a90f3b39d2614735db2149c3f7", + "component_budget": { + "node_count": 473, + "item_count": 473, + "soft_pair_count": 1738, + "serialized_chars": 901437, + "estimated_input_tokens": 225360, + "estimated_eval_tokens": 450976 + }, + "budget": { + "node_count": 48, + "item_count": 48, + "soft_pair_count": 41, + "serialized_chars": 39787, + "estimated_input_tokens": 9947, + "estimated_eval_tokens": 20150 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": true, + "analyst_result": { + "summary": "Mostly a grab bag of distinct Transformers issues; only the tokenizer regression reports look like the same underlying bug. A few loading/ZeRO-3 and sliding-window items are superficially similar, but not safe to dedupe strictly.", + "confidence": 0.79, + "canonical_issue_reason": "issue:45920 is the best umbrella: it is open, broad, and explicitly covers the same AutoTokenizer wrong-ID regression across multiple model families.", + "canonical_pr_reason": null, + "best_issue_reason": "issue:45920 has the clearest multi-model repro set and ongoing discussion, making it the strongest representative issue for the cluster.", + "best_pr_reason": null, + "soft_edge_verdicts": [ + { + "left": "issue:43866", + "right": "issue:46402", + "accept": false, + "reason": "Different models and failure modes: corrupted Ovis2 checkpoint vs missing text-encoder weight loading in Sam3Model." + }, + { + "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 specific USE_DISTRIBUTED=0 breakage in v5.9.0." + }, + { + "left": "issue:43296", + "right": "issue:46304", + "accept": false, + "reason": "Unrelated regressions: PaddleOCR-VL loading in vLLM vs PyTorch distributed support breakage." + }, + { + "left": "issue:41762", + "right": "issue:43596", + "accept": false, + "reason": "Both hit ZeRO-3 index errors, but they involve different models and different loading/training paths; too broad to treat as the same bug." + }, + { + "left": "issue:39290", + "right": "issue:41720", + "accept": false, + "reason": "Gemma3TextConfig missing a field for vLLM is unrelated to Qwen3 auto device mapping causing a CUDA assert." + }, + { + "left": "issue:43901", + "right": "issue:44512", + "accept": false, + "reason": "Both are docs issues, but they cover different doc text and different removed/changed behaviors." + }, + { + "left": "issue:44589", + "right": "issue:46304", + "accept": false, + "reason": "Float8 storage lookup failure is unrelated to distributed support regressions." + }, + { + "left": "issue:45237", + "right": "issue:45901", + "accept": false, + "reason": "AMD GPU support for GPT-OSS is unrelated to table-question-answering crashes." + }, + { + "left": "issue:43039", + "right": "issue:44560", + "accept": false, + "reason": "Liger Kernel cross-entropy routing is unrelated to a Qwen3-VL StopIteration video issue." + }, + { + "left": "issue:45823", + "right": "issue:46294", + "accept": false, + "reason": "Both touch multimodal/Qwen3.5 areas, but one is a device-map mismatch and the other is a chat-template video token mismatch." + }, + { + "left": "issue:43975", + "right": "issue:46015", + "accept": false, + "reason": "Model detokenization bug vs outdated RAG documentation are not the same underlying change." + }, + { + "left": "issue:44841", + "right": "issue:45823", + "accept": false, + "reason": "Processor failure for Voxtral-Mini and Gemma4 PLE device mismatch are different code paths and symptoms." + }, + { + "left": "issue:45020", + "right": "issue:46304", + "accept": false, + "reason": "Remote-code model breakage is not the same issue as PyTorch distributed support being disabled." + }, + { + "left": "issue:45812", + "right": "issue:45920", + "accept": true, + "reason": "Same tokenizer regression: both report wrong token IDs across multiple models after the v4\u2192v5 transition, with 45920 being the broader umbrella." + }, + { + "left": "issue:43866", + "right": "issue:44863", + "accept": false, + "reason": "Corrupted Ovis2 checkpoint and NemotronH checkpoint-loading failure are different model-specific problems." + }, + { + "left": "issue:43872", + "right": "issue:46304", + "accept": false, + "reason": "bitsandbytes Int8Params constructor incompatibility is unrelated to distributed build support." + }, + { + "left": "issue:44292", + "right": "issue:46015", + "accept": false, + "reason": "Qwen-3 NVFP4 runtime error is unrelated to RAG example docs." + }, + { + "left": "issue:39290", + "right": "issue:43531", + "accept": false, + "reason": "Both mention sliding_window, but they affect different models/configs and are not clearly the same defect." + }, + { + "left": "issue:41093", + "right": "issue:43638", + "accept": false, + "reason": "Both are IndexErrors, but the reported shapes and the zero-sized-tensor symptom point to different bugs." + }, + { + "left": "issue:43792", + "right": "issue:45901", + "accept": false, + "reason": "Whisper-large-v2 loading failure is unrelated to table-question-answering crashes." + }, + { + "left": "issue:43854", + "right": "issue:43866", + "accept": false, + "reason": "GLM-4.7-Flash load failure and Ovis2 checkpoint corruption are separate model-specific issues." + }, + { + "left": "issue:43425", + "right": "issue:45070", + "accept": false, + "reason": "Torch 2.10 compatibility and a pydantic PretrainedConfig field regression are unrelated." + }, + { + "left": "issue:43975", + "right": "issue:45901", + "accept": false, + "reason": "Detokenization regression and table QA crashes do not share a concrete bug or fix path." + }, + { + "left": "issue:43596", + "right": "issue:45800", + "accept": false, + "reason": "ZeRO-3 BertModel loading error is unrelated to a torch/transformers version incompatibility report." + }, + { + "left": "issue:43792", + "right": "issue:45070", + "accept": false, + "reason": "Whisper model loading and pydantic config-field breakage are unrelated." + }, + { + "left": "issue:45362", + "right": "issue:45464", + "accept": false, + "reason": "Both involve Qwen3.5 chat behavior, but one is a crash in transformers chat and the other is a streaming chat/completions API failure; not the same bug." + }, + { + "left": "issue:44292", + "right": "issue:45901", + "accept": false, + "reason": "Qwen-3 NVFP4 runtime error does not match table-question-answering task crashes." + }, + { + "left": "issue:39290", + "right": "issue:45823", + "accept": false, + "reason": "Gemma3/vLLM config error and Gemma4 device-map mismatch are different issues." + }, + { + "left": "issue:45464", + "right": "issue:45685", + "accept": false, + "reason": "Streaming inference failure for Qwen3.5-0.8B is unrelated to the MPS histogram int-type error." + }, + { + "left": "issue:43541", + "right": "issue:45685", + "accept": false, + "reason": "Different runtime backends and different ops: torch dynamo grouped_mm tracing vs MPS histogram implementation." + }, + { + "left": "issue:43541", + "right": "issue:43828", + "accept": false, + "reason": "Grouped_mm tracing failure in Mixtral is unrelated to an autocast dtype mismatch in Phi-tiny-MoE." + }, + { + "left": "issue:36010", + "right": "issue:39290", + "accept": false, + "reason": "ImportError for GenerationMixin is unrelated to the Gemma3TextConfig sliding-window regression." + }, + { + "left": "issue:44841", + "right": "issue:45084", + "accept": false, + "reason": "Voxtral processor failure and a template-compilation TypeError are distinct problems." + }, + { + "left": "issue:43404", + "right": "issue:43866", + "accept": false, + "reason": "Mistral3 tied-weights bug is unrelated to a corrupted Ovis2 checkpoint." + }, + { + "left": "issue:41093", + "right": "issue:41762", + "accept": false, + "reason": "Different IndexError signatures and different model/setup contexts; not a reliable duplicate." + }, + { + "left": "issue:43296", + "right": "issue:43366", + "accept": false, + "reason": "PaddleOCR-VL loading in vLLM and GGUF gpt-oss support are different model-support issues." + }, + { + "left": "issue:45399", + "right": "issue:46661", + "accept": false, + "reason": "flash-attn2 fallback gating and remote-code config_class errors are separate from each other." + }, + { + "left": "issue:43572", + "right": "issue:45070", + "accept": false, + "reason": "StableLmConfig pad_token_idx regression is unrelated to the pydantic PretrainedConfig field breakage." + }, + { + "left": "issue:45800", + "right": "issue:45823", + "accept": false, + "reason": "Torch/transformers version incompatibility and Gemma4 device-map forward mismatch are different bugs." + }, + { + "left": "issue:44387", + "right": "issue:45005", + "accept": false, + "reason": "Int4 OOM from higher CUDA reserved memory is not the same as tied-weights issues in translation models." + } + ] + }, + "evaluator_result": { + "accept": true, + "feedback": "The summary is grounded in the packet and stays conservative overall. The only accepted soft-edge pair, issue:45812 vs issue:45920, is reasonably supported by the titles as the same AutoTokenizer wrong-token-ID regression across different model families. The remaining pair rejections are appropriately cautious given the evidence available." + }, + "evaluator_used": true, + "retried": false, + "accepted_nontrivial_soft_edge": true, + "error_kind": null, + "error_message": null + }, + { + "label": "issue", + "component_index": 1, + "component_count": 21, + "review_unit_index": 3, + "review_unit_count": 56, + "status": "reviewed", + "reason": null, + "source": "llm", + "cache_hit": false, + "model": "gpt-5.4-mini?service_tier=flex", + "cluster_id": "cluster-10000-473", + "nodes": [ + "issue:39186", + "issue:40990", + "issue:41093", + "issue:42915", + "issue:43054", + "issue:43366", + "issue:43404", + "issue:43531", + "issue:43541", + "issue:43596", + "issue:43645", + "issue:43716", + "issue:43828", + "issue:43873", + "issue:43975", + "issue:43994", + "issue:44291", + "issue:44292", + "issue:44360", + "issue:44479", + "issue:44492", + "issue:44512", + "issue:44530", + "issue:44589", + "issue:44805", + "issue:44863", + "issue:44918", + "issue:44960", + "issue:45005", + "issue:45020", + "issue:45070", + "issue:45161", + "issue:45183", + "issue:45399", + "issue:45464", + "issue:45507", + "issue:45632", + "issue:45698", + "issue:45759", + "issue:45800", + "issue:45902", + "issue:46294", + "issue:46463", + "issue:46489", + "issue:46491", + "issue:46640", + "issue:46661" + ], + "soft_pairs": [ + "issue:44805|issue:44918", + "issue:43541|issue:45161", + "issue:43828|issue:45464", + "issue:43366|issue:45464", + "issue:44291|issue:45070", + "issue:43531|issue:45070", + "issue:43541|issue:44589", + "issue:43716|issue:46294", + "issue:45020|issue:45800", + "issue:41093|issue:43596", + "issue:44479|issue:46294", + "issue:43054|issue:43994", + "issue:45759|issue:45902", + "issue:46640|issue:46661", + "issue:39186|issue:44918", + "issue:43716|issue:44918", + "issue:40990|issue:44960", + "issue:45632|issue:45698", + "issue:43645|issue:45399", + "issue:43404|issue:45005", + "issue:42915|issue:45399", + "issue:44530|issue:45005", + "issue:43873|issue:45005", + "issue:44360|issue:44512", + "issue:44863|issue:45005", + "issue:45183|issue:46463", + "issue:44492|issue:44512", + "issue:44918|issue:45507", + "issue:46489|issue:46491", + "issue:43975|issue:44292", + "issue:44292|issue:45507" + ], + "prepared_review_unit_hash": "51b4819a636ad67399c8a1e195e1c425f6a90aa09eef5918c9101f5784fb7e69", + "component_budget": { + "node_count": 473, + "item_count": 473, + "soft_pair_count": 1738, + "serialized_chars": 901437, + "estimated_input_tokens": 225360, + "estimated_eval_tokens": 450976 + }, + "budget": { + "node_count": 47, + "item_count": 47, + "soft_pair_count": 31, + "serialized_chars": 36257, + "estimated_input_tokens": 9065, + "estimated_eval_tokens": 18386 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": true, + "analyst_result": { + "summary": "This cluster is mostly heterogeneous: many unrelated model-specific regressions, tokenizer issues, config/loading bugs, and docs/runtime complaints. Only a couple of soft links look like true duplicates, notably the Qwen video-input regression and the input-validation error-message bug.", + "confidence": 0.88, + "canonical_issue_reason": "No single issue cleanly represents the whole cluster because the items span multiple distinct bugs; if an umbrella representative is needed, issue 45020 is the broadest v5/remote-code regression, but it is still not a true canonical duplicate for most items.", + "canonical_pr_reason": null, + "best_issue_reason": "Issue 45020 is the closest high-level umbrella for the cluster\u2019s repeated v5/remote-code breakage theme, but the cluster remains too mixed for a strong canonical choice.", + "best_pr_reason": null, + "soft_edge_verdicts": [ + { + "left": "issue:44805", + "right": "issue:44918", + "accept": false, + "reason": "Both are indexing/shape-related, but they describe different failures in different code paths and models; not the same underlying bug." + }, + { + "left": "issue:43541", + "right": "issue:45161", + "accept": false, + "reason": "Different MoE/TP failures with different triggers; too broad a similarity to treat as one bug." + }, + { + "left": "issue:43828", + "right": "issue:45464", + "accept": false, + "reason": "Different models and symptoms: autocast dtype mismatch vs streaming chat/completions failure." + }, + { + "left": "issue:43366", + "right": "issue:45464", + "accept": false, + "reason": "GGUF gpt-oss support is unrelated to the Qwen3.5 streaming inference failure." + }, + { + "left": "issue:44291", + "right": "issue:45070", + "accept": false, + "reason": "Both involve v5 breaking changes, but one is init_empty_weights arg handling and the other is PretrainedConfig/pydantic behavior; not the same bug." + }, + { + "left": "issue:43531", + "right": "issue:45070", + "accept": false, + "reason": "Sliding-window Qwen3-MoE behavior is unrelated to the PretrainedConfig pydantic regression." + }, + { + "left": "issue:43541", + "right": "issue:44589", + "accept": false, + "reason": "Grouped_mm tracing failure and missing Float8 storage are distinct runtime issues." + }, + { + "left": "issue:43716", + "right": "issue:46294", + "accept": false, + "reason": "Both are multimodal dtype/input issues, but one is image preprocessor dtype mismatch and the other is video token/feature mismatch; not clearly the same bug." + }, + { + "left": "issue:45020", + "right": "issue:45800", + "accept": false, + "reason": "General remote_code breakage and torch/transformers version incompatibility are related only at a high level, not the same concrete defect." + }, + { + "left": "issue:41093", + "right": "issue:43596", + "accept": false, + "reason": "Both are IndexErrors, but they arise from different tensor/mask paths and different models." + }, + { + "left": "issue:44479", + "right": "issue:46294", + "accept": true, + "reason": "These look like the same Qwen video-input regression: both describe video feature/token mismatch failures in the Qwen3.5/Qwen video pipeline." + }, + { + "left": "issue:43054", + "right": "issue:43994", + "accept": false, + "reason": "Both involve Siglip2, but one is a quality/regression report and the other is an AutoModel/pipeline output bug; not enough evidence they are the same defect." + }, + { + "left": "issue:45759", + "right": "issue:45902", + "accept": false, + "reason": "Composite multimodal config handling and Qwen3_5Moe rotary embedding initialization are different code paths." + }, + { + "left": "issue:46640", + "right": "issue:46661", + "accept": false, + "reason": "Flash Attention 2 availability/fallback behavior is unrelated to remote-code model_class/config_class loading." + }, + { + "left": "issue:39186", + "right": "issue:44918", + "accept": false, + "reason": "FSDP 2-D weight error and Qwen3.5 embedding unpacking in TRL are distinct failures." + }, + { + "left": "issue:43716", + "right": "issue:44918", + "accept": false, + "reason": "Image-preprocessor dtype mismatch and TRL input-embedding unpacking are different bugs." + }, + { + "left": "issue:40990", + "right": "issue:44960", + "accept": false, + "reason": "These are unrelated model-quality/model-family issues." + }, + { + "left": "issue:45632", + "right": "issue:45698", + "accept": false, + "reason": "Both touch remote-code loading, but one is a leaf-dir cache collision and the other is wrong custom module selection after save_pretrained; not the same concrete bug." + }, + { + "left": "issue:43645", + "right": "issue:45399", + "accept": false, + "reason": "Jupyter custom-model initialization and Flash Attention fallback gating are unrelated." + }, + { + "left": "issue:42915", + "right": "issue:45399", + "accept": false, + "reason": "Qwen3-MoE FP8 failure and Flash Attention fallback checks are different issues." + }, + { + "left": "issue:44530", + "right": "issue:45005", + "accept": false, + "reason": "PagedAttentionCache linear_attention crash and tied-weights translation-model issues are unrelated." + }, + { + "left": "issue:43873", + "right": "issue:45005", + "accept": false, + "reason": "Quantization/offloading behavior and tied-weights translation-model issues do not match." + }, + { + "left": "issue:44360", + "right": "issue:44512", + "accept": false, + "reason": "A DSA indexer ReLU discussion and a removed-docs-command typo are unrelated." + }, + { + "left": "issue:44863", + "right": "issue:45005", + "accept": false, + "reason": "NemotronH checkpoint loading and translation tied-weights are different problems." + }, + { + "left": "issue:45183", + "right": "issue:46463", + "accept": true, + "reason": "These are the same validation bug family: incorrect input_ids/inputs_embeds XOR/error-message handling, with 46463 as a T5-specific manifestation." + }, + { + "left": "issue:44492", + "right": "issue:44512", + "accept": false, + "reason": "Cache-strategy typo and removed run-command docs are unrelated documentation issues." + }, + { + "left": "issue:44918", + "right": "issue:45507", + "accept": false, + "reason": "TRL embedding unpacking and GraniteMoEHybrid invalid method calls are different code paths." + }, + { + "left": "issue:46489", + "right": "issue:46491", + "accept": false, + "reason": "Both are tokenizer regressions, but for different models and different symptoms; not safe to merge as one bug." + }, + { + "left": "issue:43975", + "right": "issue:44292", + "accept": false, + "reason": "Tokenizer detokenization regression and init_empty_weights/_is_hf_initialized loading error are unrelated." + }, + { + "left": "issue:44292", + "right": "issue:45507", + "accept": false, + "reason": "Different model-loading/runtime failures with no clear shared underlying defect." + } + ] + }, + "evaluator_result": { + "accept": true, + "feedback": "The summary is grounded in the packet and stays appropriately conservative about the cluster\u2019s heterogeneity. The two accepted soft links are defensible as same-bug families, and the rejected pairs are phrased cautiously without overclaiming duplicates." + }, + "evaluator_used": true, + "retried": false, + "accepted_nontrivial_soft_edge": true, + "error_kind": null, + "error_message": null + }, + { + "label": "issue", + "component_index": 1, + "component_count": 21, + "review_unit_index": 4, + "review_unit_count": 56, + "status": "reviewed", + "reason": null, + "source": "llm", + "cache_hit": false, + "model": "gpt-5.4-mini?service_tier=flex", + "cluster_id": "cluster-10000-473", + "nodes": [ + "issue:24643", + "issue:28530", + "issue:36331", + "issue:38175", + "issue:39692", + "issue:41553", + "issue:41762", + "issue:42491", + "issue:42915", + "issue:42947", + "issue:43054", + "issue:43493", + "issue:43643", + "issue:43644", + "issue:43782", + "issue:43824", + "issue:43828", + "issue:43856", + "issue:43866", + "issue:43872", + "issue:43881", + "issue:43883", + "issue:43950", + "issue:43975", + "issue:44368", + "issue:44387", + "issue:44403", + "issue:44451", + "issue:44488", + "issue:44534", + "issue:44589", + "issue:44661", + "issue:44910", + "issue:44928", + "issue:44960", + "issue:44991", + "issue:45020", + "issue:45081", + "issue:45399", + "issue:45520", + "issue:45663", + "issue:45685", + "issue:45701", + "issue:45823", + "issue:45859", + "issue:45902", + "issue:45920", + "issue:46489" + ], + "soft_pairs": [ + "issue:43950|issue:44534", + "issue:39692|issue:43054", + "issue:43644|issue:43950", + "issue:36331|issue:41762", + "issue:43828|issue:44928", + "issue:43872|issue:44589", + "issue:45920|issue:46489", + "issue:43866|issue:44403", + "issue:43054|issue:45701", + "issue:43950|issue:45902", + "issue:38175|issue:43054", + "issue:41553|issue:43883", + "issue:44910|issue:45859", + "issue:42947|issue:44387", + "issue:43054|issue:43493", + "issue:45663|issue:45902", + "issue:41553|issue:43881", + "issue:44403|issue:44451", + "issue:43975|issue:45520", + "issue:24643|issue:28530", + "issue:44403|issue:44488", + "issue:42947|issue:43856", + "issue:43643|issue:45020", + "issue:41762|issue:42915", + "issue:44991|issue:45399", + "issue:43824|issue:43872", + "issue:45685|issue:45823", + "issue:43782|issue:44960", + "issue:43881|issue:43883", + "issue:45020|issue:45823", + "issue:44661|issue:45020", + "issue:45081|issue:45399", + "issue:42491|issue:44368", + "issue:43828|issue:45823" + ], + "prepared_review_unit_hash": "d7e4b42f388b7e6377b9a13e610a99b8af9bbae7e32bd31aff9ad06d23e67087", + "component_budget": { + "node_count": 473, + "item_count": 473, + "soft_pair_count": 1738, + "serialized_chars": 901437, + "estimated_input_tokens": 225360, + "estimated_eval_tokens": 450976 + }, + "budget": { + "node_count": 48, + "item_count": 48, + "soft_pair_count": 34, + "serialized_chars": 38114, + "estimated_input_tokens": 9529, + "estimated_eval_tokens": 19314 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": true, + "analyst_result": { + "summary": "Mostly unrelated loading/training regressions, with one clear non-persistent-buffer corruption cluster and one plausible SigLIP2 implementation-discrepancy pair. The 43950/45902 edge should be rejected: 45902 looks like a related materialization/uninitialized-memory symptom, but not the same regression as the broader from_pretrained() buffer corruption bug.", + "confidence": 0.77, + "canonical_issue_reason": "issue:43950 is the broadest and clearest representative of the transformers 5.x non-persistent-buffer corruption regression; it captures the shared root symptom better than the narrower follow-up reports.", + "canonical_pr_reason": null, + "best_issue_reason": "issue:43950 best represents the shared buffer-corruption defect and subsumes the closely matching junk-filled buffer reports without overextending to adjacent materialization bugs.", + "best_pr_reason": null, + "soft_edge_verdicts": [ + { + "left": "issue:43950", + "right": "issue:44534", + "accept": true, + "reason": "Same underlying transformers 5.x non-persistent-buffer corruption bug." + }, + { + "left": "issue:39692", + "right": "issue:43054", + "accept": false, + "reason": "Both mention SigLIP2, but one is a docs/example mismatch/quantization issue and the other is a model-quality complaint; not the same bug." + }, + { + "left": "issue:43644", + "right": "issue:43950", + "accept": true, + "reason": "Direct duplicate report of the non-persistent-buffer junk regression." + }, + { + "left": "issue:36331", + "right": "issue:41762", + "accept": false, + "reason": "Different failure modes and code paths: custom compute_loss arg handling vs Gemma3 loading under DeepSpeed ZeRO-3." + }, + { + "left": "issue:43828", + "right": "issue:44928", + "accept": false, + "reason": "Different execution bugs: autocast dtype mismatch vs 3D position_ids / flash-attn handling." + }, + { + "left": "issue:43872", + "right": "issue:44589", + "accept": false, + "reason": "Different low-level compatibility problems: bitsandbytes kwarg mismatch vs missing Float8 storage object." + }, + { + "left": "issue:45920", + "right": "issue:46489", + "accept": false, + "reason": "Related tokenizer-regression area, but not the same concrete defect." + }, + { + "left": "issue:43866", + "right": "issue:44403", + "accept": false, + "reason": "Unrelated issues: checkpoint corruption vs generic loading noise." + }, + { + "left": "issue:43054", + "right": "issue:45701", + "accept": false, + "reason": "SigLIP2 embedding-quality complaint is not the same as a version-wide tokenization change." + }, + { + "left": "issue:43950", + "right": "issue:45902", + "accept": false, + "reason": "Related symptom area, but 45902 is a more specific meta\u2192to_empty(cuda) materialization/uninitialized-memory problem, not clearly the same non-persistent-buffer corruption regression." + }, + { + "left": "issue:38175", + "right": "issue:43054", + "accept": false, + "reason": "Both concern SigLIP2, but zero probabilities and worse text embeddings are insufficiently specific to call the same bug." + }, + { + "left": "issue:41553", + "right": "issue:43883", + "accept": false, + "reason": "Different loading problems for different models; no shared concrete root cause is evident." + }, + { + "left": "issue:44910", + "right": "issue:45859", + "accept": false, + "reason": "Different Qwen3.5/MoE bugs: flash-attn illegal access with 3D position_ids vs CPU offload incompatibility in rotary embedding." + }, + { + "left": "issue:42947", + "right": "issue:44387", + "accept": false, + "reason": "Both are memory-related, but the mechanisms differ: gradient checkpointing ineffectiveness vs increased reserved memory under int4 quantization." + }, + { + "left": "issue:43054", + "right": "issue:43493", + "accept": true, + "reason": "Plausibly the same SigLIP2 implementation discrepancy: degraded text embeddings fit an HF-vs-JAX mismatch." + }, + { + "left": "issue:45663", + "right": "issue:45902", + "accept": false, + "reason": "Different bugs in different models/code paths: Gemma4 FSDP2 KeyError vs Qwen3_5Moe buffer materialization." + }, + { + "left": "issue:41553", + "right": "issue:43881", + "accept": false, + "reason": "Different loading failures for different models and causes." + }, + { + "left": "issue:44403", + "right": "issue:44451", + "accept": false, + "reason": "Generic loading noise vs inability to load a specific model; not the same defect." + }, + { + "left": "issue:43975", + "right": "issue:45520", + "accept": false, + "reason": "Tokenizer detokenization bug is unrelated to a Python 3.13 flash_attn import error." + }, + { + "left": "issue:24643", + "right": "issue:28530", + "accept": false, + "reason": "DeepSpeed training weight-shape runtime error is unrelated to early-stopping metric selection." + }, + { + "left": "issue:44403", + "right": "issue:44488", + "accept": false, + "reason": "Generic loading noise vs a specific model load failure; not the same issue." + }, + { + "left": "issue:42947", + "right": "issue:43856", + "accept": false, + "reason": "Both involve memory usage, but one is LoRA checkpointing behavior and the other is Qwen3 MoE training efficiency." + }, + { + "left": "issue:43643", + "right": "issue:45020", + "accept": false, + "reason": "Related to remote_code, but one is missing fields in AutoConfig and the other is a broader regression breaking remote_code models." + }, + { + "left": "issue:41762", + "right": "issue:42915", + "accept": false, + "reason": "Different model/load failures: Gemma3 with ZeRO-3 vs Qwen3 MoE with FineGrainedFP8Config." + }, + { + "left": "issue:44991", + "right": "issue:45399", + "accept": false, + "reason": "Tokenizer-loading failure for a specific model is unrelated to flash-attn fallback gating." + }, + { + "left": "issue:43824", + "right": "issue:43872", + "accept": false, + "reason": "Different compatibility issues: missing Qwen2_5_VL export vs bitsandbytes Int8Params kwarg mismatch." + }, + { + "left": "issue:45685", + "right": "issue:45823", + "accept": false, + "reason": "MPS histogram int support error is unrelated to Gemma4 device placement mismatch." + }, + { + "left": "issue:43782", + "right": "issue:44960", + "accept": false, + "reason": "Different models and different error classes: Qwen3VL weight_only loading error vs GLM5 loading." + }, + { + "left": "issue:43881", + "right": "issue:43883", + "accept": false, + "reason": "Unrelated loading failures for different models and root causes." + }, + { + "left": "issue:45020", + "right": "issue:45823", + "accept": false, + "reason": "Broad remote_code regression and Gemma4 device_map forward mismatch are separate issues." + }, + { + "left": "issue:44661", + "right": "issue:45020", + "accept": false, + "reason": "Tokenizer-mapping edge case is not the same bug as the remote_code regression." + }, + { + "left": "issue:45081", + "right": "issue:45399", + "accept": false, + "reason": "Mistral regex patch crash is unrelated to flash-attn fallback gating." + }, + { + "left": "issue:42491", + "right": "issue:44368", + "accept": false, + "reason": "Different Qwen3/Qwen3.5 LoRA-related problems: cross-version incompatibility vs tied-embeddings warning." + }, + { + "left": "issue:43828", + "right": "issue:45823", + "accept": false, + "reason": "Different model execution bugs: autocast dtype mismatch vs device-map placement mismatch." + } + ] + }, + "evaluator_result": { + "accept": false, + "feedback": "The cluster summary is mostly grounded, but one soft-edge verdict is too strong for the evidence in the packet: issue:43950 vs issue:45902. The titles support a possible related loading/materialization bug, but not clearly the same non-persistent-buffer corruption regression, so calling it a duplicate overstates the case. The SigLIP2 pair is more plausible, but the overall set is not conservative enough to accept as-is." + }, + "evaluator_used": true, + "retried": true, + "accepted_nontrivial_soft_edge": true, + "error_kind": null, + "error_message": null + }, + { + "label": "issue", + "component_index": 1, + "component_count": 21, + "review_unit_index": 5, + "review_unit_count": 56, + "status": "reviewed", + "reason": null, + "source": "llm", + "cache_hit": false, + "model": "gpt-5.4-mini?service_tier=flex", + "cluster_id": "cluster-10000-473", + "nodes": [ + "issue:24643", + "issue:30064", + "issue:36331", + "issue:41093", + "issue:41553", + "issue:42915", + "issue:42947", + "issue:43257", + "issue:43366", + "issue:43404", + "issue:43531", + "issue:43596", + "issue:43792", + "issue:43827", + "issue:43828", + "issue:43854", + "issue:43872", + "issue:43940", + "issue:43950", + "issue:44220", + "issue:44291", + "issue:44479", + "issue:44488", + "issue:44509", + "issue:44530", + "issue:44560", + "issue:44805", + "issue:44863", + "issue:44928", + "issue:44991", + "issue:45020", + "issue:45305", + "issue:45313", + "issue:45325", + "issue:45356", + "issue:45488", + "issue:45588", + "issue:45663", + "issue:45715", + "issue:45759", + "issue:45823", + "issue:45902", + "issue:45923", + "issue:46315", + "issue:46402", + "issue:46489", + "issue:46515", + "issue:46661" + ], + "soft_pairs": [ + "issue:45020|issue:45356", + "issue:45715|issue:46315", + "issue:45488|issue:46489", + "issue:43827|issue:44509", + "issue:43404|issue:43950", + "issue:45759|issue:46661", + "issue:42915|issue:43872", + "issue:43872|issue:44291", + "issue:45020|issue:46661", + "issue:43792|issue:44220", + "issue:43257|issue:45305", + "issue:42947|issue:45923", + "issue:41553|issue:44991", + "issue:36331|issue:41093", + "issue:43940|issue:45313", + "issue:43854|issue:44863", + "issue:44928|issue:45588", + "issue:44479|issue:44560", + "issue:41093|issue:44805", + "issue:41553|issue:44488", + "issue:24643|issue:30064", + "issue:44863|issue:46402", + "issue:44530|issue:44863", + "issue:43366|issue:43828", + "issue:43596|issue:46515", + "issue:43366|issue:43531", + "issue:45325|issue:45902", + "issue:45663|issue:45823" + ], + "prepared_review_unit_hash": "afac19b70ede5dd449c1a6a671a66e4c4a75f6143c2267567392667a65173197", + "component_budget": { + "node_count": 473, + "item_count": 473, + "soft_pair_count": 1738, + "serialized_chars": 901437, + "estimated_input_tokens": 225360, + "estimated_eval_tokens": 450976 + }, + "budget": { + "node_count": 48, + "item_count": 48, + "soft_pair_count": 28, + "serialized_chars": 36623, + "estimated_input_tokens": 9156, + "estimated_eval_tokens": 18568 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": true, + "analyst_result": { + "summary": "Mostly heterogeneous regressions and model-specific bugs. The strongest duplicate clusters are the v5 docs cleanup, the `_is_hf_initialized` loading regression, and the generic mask-shape IndexError pair. The DeepSeek/Llama tokenizer pair is related but not the same underlying bug from the packet alone.", + "confidence": 0.67, + "canonical_issue_reason": "Issue 44291 is the clearest representative because it states the concrete loading failure and root cause (`_is_hf_initialized` on init-empty loading), while 43872 is a narrower symptom of the same path.", + "canonical_pr_reason": null, + "best_issue_reason": "44291 is the best global issue anchor because it most precisely identifies a concrete code-path regression with a plausible single fix target; the rest of the cluster is mostly separate model-specific or docs-only problems.", + "best_pr_reason": null, + "soft_edge_verdicts": [ + { + "left": "issue:45020", + "right": "issue:45356", + "accept": false, + "reason": "Both are recent-version regressions, but they affect different areas: remote-code model loading vs tokenizer handling for Kimi-K2.5." + }, + { + "left": "issue:45715", + "right": "issue:46315", + "accept": false, + "reason": "Both are tokenizer performance issues, but they hit different APIs (`convert_ids_to_tokens` vs `convert_tokens_to_ids`) and do not read as the same bug." + }, + { + "left": "issue:45488", + "right": "issue:46489", + "accept": false, + "reason": "Both are tokenizer regressions, but the packet only gives titles and they point to different model families/failure modes; related, not clearly the same bug/root cause." + }, + { + "left": "issue:43827", + "right": "issue:44509", + "accept": true, + "reason": "Both are the same docs cleanup problem: stale v5 documentation still references removed pipeline tasks like summarization and translation." + }, + { + "left": "issue:43404", + "right": "issue:43950", + "accept": false, + "reason": "One is a weight-tying bug in Mistral3ForConditionalGeneration; the other is a from_pretrained buffer-corruption regression." + }, + { + "left": "issue:45759", + "right": "issue:46661", + "accept": false, + "reason": "These are different model-loading failures: composite config unwrapping vs remote-code `config_class` lookup." + }, + { + "left": "issue:42915", + "right": "issue:43872", + "accept": false, + "reason": "Qwen3Moe FP8 failure and the bitsandbytes `_is_hf_initialized` incompatibility are different issues affecting different paths." + }, + { + "left": "issue:43872", + "right": "issue:44291", + "accept": true, + "reason": "Same root cause: an unexpected `_is_hf_initialized` argument breaks loading/initialization, with 44291 describing the regression directly." + }, + { + "left": "issue:45020", + "right": "issue:46661", + "accept": false, + "reason": "Both involve remote-code models, but one is a broad regression claim and the other is a specific `config_class` attribute error; too broad to merge." + }, + { + "left": "issue:43792", + "right": "issue:44220", + "accept": false, + "reason": "The Whisper failure and `_torch_extract_fbank_features()` issue may be related, but the packet does not establish the same concrete bug path." + }, + { + "left": "issue:43257", + "right": "issue:45305", + "accept": false, + "reason": "One is a Qwen3 MoE weight-conversion/loading problem; the other is DeepSpeed gradient averaging with GAS." + }, + { + "left": "issue:42947", + "right": "issue:45923", + "accept": false, + "reason": "Gradient checkpointing being ineffective in PEFT LoRA is not the same as a missing `supports_gradient_checkpointing` flag on a remote-code variant." + }, + { + "left": "issue:41553", + "right": "issue:44991", + "accept": false, + "reason": "Both are tokenizer-loading complaints, but they involve different models and different failure modes." + }, + { + "left": "issue:36331", + "right": "issue:41093", + "accept": false, + "reason": "Unexpected `num_items_in_batch` in custom trainer loss and a tensor-mask shape mismatch are unrelated errors." + }, + { + "left": "issue:43940", + "right": "issue:45313", + "accept": false, + "reason": "Both are DeepSpeed weight-loading bugs, but one is Qwen3-Next missing params and the other is Qwen3.5 `language_model` loading." + }, + { + "left": "issue:43854", + "right": "issue:44863", + "accept": false, + "reason": "Unit-test model loading failure and NemotronH checkpoint loading failure are different model-specific problems." + }, + { + "left": "issue:44928", + "right": "issue:45588", + "accept": false, + "reason": "RLHF gradient explosion in Qwen3.5 and a flash-attention `s_aux=None` crash are unrelated." + }, + { + "left": "issue:44479", + "right": "issue:44560", + "accept": false, + "reason": "Both are Qwen/VL regressions, but one is a video-input regression and the other is a StopIteration on Qwen3-vl-embedding." + }, + { + "left": "issue:41093", + "right": "issue:44805", + "accept": true, + "reason": "Both describe the same generic index/mask length mismatch error with an off-by-one sized indexed tensor." + }, + { + "left": "issue:41553", + "right": "issue:44488", + "accept": false, + "reason": "Different tokenizer-loading issues for different models; one is an error message problem, the other is a load failure." + }, + { + "left": "issue:24643", + "right": "issue:30064", + "accept": false, + "reason": "Training with DeepSpeed and void segmentation-map processing are unrelated." + }, + { + "left": "issue:44863", + "right": "issue:46402", + "accept": false, + "reason": "NemotronH checkpoint loading and Sam3 text-encoder weight loading are different model-specific failures." + }, + { + "left": "issue:44530", + "right": "issue:44863", + "accept": false, + "reason": "PagedAttentionCache linear_attention crash and NemotronH checkpoint loading are unrelated." + }, + { + "left": "issue:43366", + "right": "issue:43828", + "accept": false, + "reason": "GGUF gpt-oss support and a torch autocast dtype mismatch in Phi-tiny-MoE are different issues." + }, + { + "left": "issue:43596", + "right": "issue:46515", + "accept": false, + "reason": "Zero-3/BertModel index-out-of-bounds and torch 2.4.1 incompatibility are not the same bug." + }, + { + "left": "issue:43366", + "right": "issue:43531", + "accept": false, + "reason": "GGUF architecture support and Qwen3-MoE sliding_window behavior are unrelated." + }, + { + "left": "issue:45325", + "right": "issue:45902", + "accept": false, + "reason": "Both touch rotary/position handling, but one is a rope-index scaling bug and the other is uninitialized `inv_freq` memory after materialization." + }, + { + "left": "issue:45663", + "right": "issue:45823", + "accept": false, + "reason": "Gemma4 FSDP2 shared_kv_states KeyError and Gemma4 device mismatch are separate forward-pass bugs." + } + ] + }, + "evaluator_result": { + "accept": false, + "feedback": "The cluster-level summary is mostly grounded, but one soft-edge accept overstates the evidence: `issue:45488` vs `issue:46489` are both tokenizer regressions, yet the packet only gives titles and they point to different model families/failure modes. From the packet alone, that is better described as related rather than the same bug/root cause. The other accepted pairs look reasonably conservative." + }, + "evaluator_used": true, + "retried": true, + "accepted_nontrivial_soft_edge": true, + "error_kind": null, + "error_message": null + }, + { + "label": "issue", + "component_index": 1, + "component_count": 21, + "review_unit_index": 6, + "review_unit_count": 56, + "status": "reviewed", + "reason": null, + "source": "llm", + "cache_hit": false, + "model": "gpt-5.4-mini?service_tier=flex", + "cluster_id": "cluster-10000-473", + "nodes": [ + "issue:36010", + "issue:38175", + "issue:39692", + "issue:41720", + "issue:42491", + "issue:42617", + "issue:42915", + "issue:43317", + "issue:43366", + "issue:43550", + "issue:43596", + "issue:43632", + "issue:43638", + "issue:43644", + "issue:43645", + "issue:43646", + "issue:43716", + "issue:43720", + "issue:43824", + "issue:43950", + "issue:43957", + "issue:43994", + "issue:44155", + "issue:44451", + "issue:44488", + "issue:44534", + "issue:44560", + "issue:44589", + "issue:44661", + "issue:44683", + "issue:44756", + "issue:44805", + "issue:44877", + "issue:44910", + "issue:44918", + "issue:44960", + "issue:44977", + "issue:45030", + "issue:45276", + "issue:45584", + "issue:45685", + "issue:45812", + "issue:45823", + "issue:45834", + "issue:46459", + "issue:46489", + "issue:46512", + "issue:46515" + ], + "soft_pairs": [ + "issue:43638|issue:46515", + "issue:43716|issue:44560", + "issue:44155|issue:45584", + "issue:43366|issue:44589", + "issue:43950|issue:45834", + "issue:43824|issue:43957", + "issue:42915|issue:43957", + "issue:44560|issue:45685", + "issue:43645|issue:43646", + "issue:43550|issue:44910", + "issue:44560|issue:44918", + "issue:44683|issue:46515", + "issue:42617|issue:43366", + "issue:36010|issue:42915", + "issue:45812|issue:46489", + "issue:45823|issue:46515", + "issue:44560|issue:44805", + "issue:44661|issue:45823", + "issue:38175|issue:43994", + "issue:43632|issue:45834", + "issue:44877|issue:45030", + "issue:43638|issue:44661", + "issue:43596|issue:44661", + "issue:43317|issue:44756", + "issue:44661|issue:44805", + "issue:38175|issue:39692", + "issue:42491|issue:43720", + "issue:43644|issue:44534", + "issue:41720|issue:42915", + "issue:42915|issue:44756", + "issue:42491|issue:45276", + "issue:44451|issue:44488", + "issue:46459|issue:46512", + "issue:44960|issue:44977" + ], + "prepared_review_unit_hash": "864186513e4275ee288d5726dc92703705a6154b26d368d7b98ea126d8303a92", + "component_budget": { + "node_count": 473, + "item_count": 473, + "soft_pair_count": 1738, + "serialized_chars": 901437, + "estimated_input_tokens": 225360, + "estimated_eval_tokens": 450976 + }, + "budget": { + "node_count": 48, + "item_count": 48, + "soft_pair_count": 34, + "serialized_chars": 37774, + "estimated_input_tokens": 9444, + "estimated_eval_tokens": 19144 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": true, + "analyst_result": { + "summary": "Mostly a loose collection of Transformers v5 regressions and model-specific inference/load bugs. Only a few soft edges look like true duplicates; most pairs are different code paths or different models. No PRs are present in this cluster.", + "confidence": 0.82, + "canonical_issue_reason": "issue:43646 is the broadest representative of the main theme here: Transformers v5 breaking custom model initialization/loading. It also subsumes the notebook-specific variant in issue:43645 better than any narrower report.", + "canonical_pr_reason": null, + "best_issue_reason": "issue:43646 is the best single anchor for the cluster because it captures the broadest and most reusable regression label among the mostly heterogeneous issues, even though the cluster is not a tight duplicate set.", + "best_pr_reason": null, + "soft_edge_verdicts": [ + { + "left": "issue:43638", + "right": "issue:46515", + "accept": false, + "reason": "Different bugs: zero3/BertModel init failure vs a generic torch 2.4.1 incompatibility." + }, + { + "left": "issue:43716", + "right": "issue:44560", + "accept": false, + "reason": "Different model paths: image preprocessor dtype mismatch vs Qwen3-VL video StopIteration." + }, + { + "left": "issue:44155", + "right": "issue:45584", + "accept": false, + "reason": "AudioFlamingo3 batch leakage and Whisper empty-transcription failure are unrelated." + }, + { + "left": "issue:43366", + "right": "issue:44589", + "accept": false, + "reason": "GGUF gpt-oss support request vs Float8 storage lookup error; not the same bug." + }, + { + "left": "issue:43950", + "right": "issue:45834", + "accept": false, + "reason": "Both are regressions, but non-persistent buffer corruption and Kosmos2.5 OCR index error are different failures." + }, + { + "left": "issue:43824", + "right": "issue:43957", + "accept": false, + "reason": "Import-path regression for Qwen2_5_VL vs generic torch.device(\"meta\") loading breakage." + }, + { + "left": "issue:42915", + "right": "issue:43957", + "accept": false, + "reason": "Qwen3MoE FP8 failure and meta-device loading failures are different code paths." + }, + { + "left": "issue:44560", + "right": "issue:45685", + "accept": false, + "reason": "Unrelated bugs: Qwen3-VL video StopIteration vs an MPS histogram/int error in MoE." + }, + { + "left": "issue:43645", + "right": "issue:43646", + "accept": true, + "reason": "Same underlying regression: custom model initialization breaks in Transformers v5; the notebook context is just a reproduction detail." + }, + { + "left": "issue:43550", + "right": "issue:44910", + "accept": false, + "reason": "Different model/runtime issues: Bamba torch.compile+SDPA vs Qwen3.5 Flash Attention packed-sequence misread." + }, + { + "left": "issue:44560", + "right": "issue:44918", + "accept": false, + "reason": "Different Qwen3-family failures with different symptoms and code paths." + }, + { + "left": "issue:44683", + "right": "issue:46515", + "accept": false, + "reason": "Compiled flex_attention on torch>=2.9 is unrelated to a torch 2.4.1 compatibility issue." + }, + { + "left": "issue:42617", + "right": "issue:43366", + "accept": false, + "reason": "3d_parallel.py runtime failure and GGUF gpt-oss support are unrelated." + }, + { + "left": "issue:36010", + "right": "issue:42915", + "accept": false, + "reason": "GenerationMixin import error is unrelated to Qwen3MoE FP8 loading failure." + }, + { + "left": "issue:45812", + "right": "issue:46489", + "accept": false, + "reason": "Both are tokenizer regressions, but they affect different model families and likely different code paths." + }, + { + "left": "issue:45823", + "right": "issue:46515", + "accept": false, + "reason": "Gemma4 device mismatch during forward is not the same as a generic torch version incompatibility." + }, + { + "left": "issue:44560", + "right": "issue:44805", + "accept": false, + "reason": "Video StopIteration and mask/tensor shape IndexError are different bugs." + }, + { + "left": "issue:44661", + "right": "issue:45823", + "accept": false, + "reason": "Tokenizer-mapping model-registration bug vs Gemma4 PLE device mismatch; not the same issue." + }, + { + "left": "issue:38175", + "right": "issue:43994", + "accept": true, + "reason": "Both report broken SigLIP2 base-patch16-224 outputs on the same model family, with the same underlying inference regression." + }, + { + "left": "issue:43632", + "right": "issue:45834", + "accept": false, + "reason": "_is_hf_initialized flag regression and Kosmos2.5 long-OCR index error are unrelated." + }, + { + "left": "issue:44877", + "right": "issue:45030", + "accept": false, + "reason": "Both mention config validation, but they are different model configs and not clearly the same concrete bug." + }, + { + "left": "issue:43638", + "right": "issue:44661", + "accept": false, + "reason": "Zero3/BertModel init failure is unrelated to the TOKENIZER_MAPPING_NAMES add-new-model-like bug." + }, + { + "left": "issue:43596", + "right": "issue:44661", + "accept": false, + "reason": "Zero3/BertModel init failure and tokenizer-mapping logic failure are different code paths." + }, + { + "left": "issue:43317", + "right": "issue:44756", + "accept": false, + "reason": "gpu+cpu offload loading failure and Strix Halo mmap OOM are different problems." + }, + { + "left": "issue:38175", + "right": "issue:39692", + "accept": false, + "reason": "Same model family, but the SigLIP2 output regression and the doc-example model/processor mismatch are not clearly the same bug." + }, + { + "left": "issue:42491", + "right": "issue:43720", + "accept": false, + "reason": "Qwen3_moe LoRA compatibility and BitNet packed-weight unpacking are unrelated." + }, + { + "left": "issue:43644", + "right": "issue:44534", + "accept": true, + "reason": "Exact same regression: Transformers v5 fills non-persistent buffers with junk; the titles are near-paraphrases." + }, + { + "left": "issue:41720", + "right": "issue:42915", + "accept": false, + "reason": "Qwen3 auto device mapping assert and Qwen3MoE FineGrainedFP8 failure are different runtime paths." + }, + { + "left": "issue:42915", + "right": "issue:44756", + "accept": false, + "reason": "Qwen3MoE FP8 failure is unrelated to Strix Halo mmap/OOM behavior." + }, + { + "left": "issue:42491", + "right": "issue:45276", + "accept": false, + "reason": "Qwen3_moe LoRA compatibility and Gemma4 resize_token_embeddings behavior are unrelated." + }, + { + "left": "issue:44451", + "right": "issue:44488", + "accept": false, + "reason": "Both are loading failures, but the titles point to different models and no shared concrete bug is visible." + }, + { + "left": "issue:46459", + "right": "issue:46512", + "accept": false, + "reason": "Both concern dtype handling, but one is checkpoint loading and the other is from_config/config-state synchronization." + }, + { + "left": "issue:44960", + "right": "issue:44977", + "accept": false, + "reason": "GLM5 and Qwen3.5 Flash Attention generation failures are unrelated." + } + ] + }, + "evaluator_result": { + "accept": true, + "feedback": "The cluster summary is grounded in the packet and stays conservative about the mostly loose/heterogeneous relationships. The accepted soft edges that are marked as duplicates are plausibly supported by the titles, especially the exact non-persistent-buffer regression pair and the shared SigLIP2 model-family pair. No PRs are present, so there is no mergeability concern to flag." + }, + "evaluator_used": true, + "retried": false, + "accepted_nontrivial_soft_edge": true, + "error_kind": null, + "error_message": null + }, + { + "label": "issue", + "component_index": 1, + "component_count": 21, + "review_unit_index": 7, + "review_unit_count": 56, + "status": "reviewed", + "reason": null, + "source": "llm", + "cache_hit": false, + "model": "gpt-5.4-mini?service_tier=flex", + "cluster_id": "cluster-10000-473", + "nodes": [ + "issue:36010", + "issue:39404", + "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:44387", + "issue:44402", + "issue:44488", + "issue:44756", + "issue:44910", + "issue:44960", + "issue:44991", + "issue:45276", + "issue:45292", + "issue:45313", + "issue:45335", + "issue:45357", + "issue:45362", + "issue:45406", + "issue:45488", + "issue:45507", + "issue:45588", + "issue:45663", + "issue:45704", + "issue:45759", + "issue:45812", + "issue:45859", + "issue:46050", + "issue:46468" + ], + "soft_pairs": [ + "issue:45313|issue:45357", + "issue:45507|issue:46468", + "issue:43856|issue:43873", + "issue:44488|issue:44991", + "issue:43644|issue:43645", + "issue:43708|issue:44075", + "issue:39692|issue:43994", + "issue:43856|issue:44387", + "issue:43854|issue:43931", + "issue:43257|issue:43854", + "issue:42491|issue:45292", + "issue:42491|issue:43931", + "issue:44960|issue:46050", + "issue:45488|issue:45812", + "issue:45362|issue:45406", + "issue:44164|issue:44291", + "issue:45276|issue:45335", + "issue:44910|issue:45588", + "issue:44402|issue:44488", + "issue:43425|issue:43957", + "issue:43873|issue:44387", + "issue:42947|issue:45704", + "issue:41720|issue:44155", + "issue:43404|issue:43454", + "issue:39692|issue:44960", + "issue:45663|issue:45759", + "issue:44291|issue:44756", + "issue:36010|issue:39404", + "issue:45663|issue:45859", + "issue:44387|issue:44756" + ], + "prepared_review_unit_hash": "bbdead1abf518d03a512a25ad4a0ec67571bf5239e236b70e177666a2597caea", + "component_budget": { + "node_count": 473, + "item_count": 473, + "soft_pair_count": 1738, + "serialized_chars": 901437, + "estimated_input_tokens": 225360, + "estimated_eval_tokens": 450976 + }, + "budget": { + "node_count": 47, + "item_count": 47, + "soft_pair_count": 30, + "serialized_chars": 36203, + "estimated_input_tokens": 9051, + "estimated_eval_tokens": 18358 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": true, + "analyst_result": { + "summary": "These items are mostly unrelated Transformers regressions that only share broad themes like loading, quantization, or model-specific compatibility. I do not see any true duplicate pairs among the soft edges. If forced to pick a representative issue, 45812 is the clearest standalone regression.", + "confidence": 0.79, + "canonical_issue_reason": "Issue 45812 is the strongest representative: it reports a clear, broad tokenizer regression with explicit impact across Granite models and a well-defined failure mode.", + "canonical_pr_reason": null, + "best_issue_reason": "45812 is the best standalone issue candidate because it is concrete, high-signal, and broader in scope than the other reports; it is not a duplicate of them.", + "best_pr_reason": null, + "soft_edge_verdicts": [ + { + "left": "issue:45313", + "right": "issue:45357", + "accept": false, + "reason": "Different bugs: one is DeepSpeed ZeRO-3 failing to load language_model weights, the other is save_pretrained saving wrong visual encoder keys." + }, + { + "left": "issue:45507", + "right": "issue:46468", + "accept": false, + "reason": "Different model families and failures: GraniteMoEHybrid invalid method call vs a broken Nemotron example." + }, + { + "left": "issue:43856", + "right": "issue:43873", + "accept": false, + "reason": "Both involve Qwen3/MoE resource behavior, but one is training memory inefficiency and the other is quantization/offloading not working; not the same code-path bug." + }, + { + "left": "issue:44488", + "right": "issue:44991", + "accept": false, + "reason": "Both are model-loading problems, but they affect different models and likely different tokenizer/config paths." + }, + { + "left": "issue:43644", + "right": "issue:43645", + "accept": false, + "reason": "Separate Transformers 5 regressions: non-persistent buffers getting junk vs Jupyter custom model initialization breaking." + }, + { + "left": "issue:43708", + "right": "issue:44075", + "accept": false, + "reason": "Trainer checkpoint-resume step \u0917\u0923ation issue vs SGD arguments being ignored; unrelated subsystems." + }, + { + "left": "issue:39692", + "right": "issue:43994", + "accept": false, + "reason": "Same family (SigLIP2) but different problems: a bad documentation example versus incorrect model outputs at runtime." + }, + { + "left": "issue:43856", + "right": "issue:44387", + "accept": false, + "reason": "Both mention memory/quantization, but one is MoE training memory use and the other is int4 reserved-memory OOM; different root causes." + }, + { + "left": "issue:43854", + "right": "issue:43931", + "accept": false, + "reason": "Different load failures for different models: GLM-4.7-Flash unit test loading vs Qwen3-VL weight-shape mismatch." + }, + { + "left": "issue:43257", + "right": "issue:43854", + "accept": false, + "reason": "Qwen3 MoE weight conversion under accelerate+deepspeed is unrelated to GLM-4.7-Flash loading in unit tests." + }, + { + "left": "issue:42491", + "right": "issue:45292", + "accept": false, + "reason": "LoRA/hf5 compatibility issue for Qwen3 MoE is unrelated to resize_token_embeddings not updating output embeddings." + }, + { + "left": "issue:42491", + "right": "issue:43931", + "accept": false, + "reason": "Different model and failure mode: Qwen3 MoE LoRA compatibility vs Qwen3-VL loading shape mismatch." + }, + { + "left": "issue:44960", + "right": "issue:46050", + "accept": false, + "reason": "GLM5 report versus sinq quantized save/load breakage; no shared concrete bug." + }, + { + "left": "issue:45488", + "right": "issue:45812", + "accept": false, + "reason": "Both are tokenizer regressions, but one swaps the pre-tokenizer for DeepSeek V3/R1 while the other produces wrong IDs for Granite models; different models and mechanisms." + }, + { + "left": "issue:45362", + "right": "issue:45406", + "accept": false, + "reason": "Qwen3.5 chat crash and Gemma4Processor missing _tokenizer are unrelated model/processor bugs." + }, + { + "left": "issue:44164", + "right": "issue:44291", + "accept": false, + "reason": "save/from_pretrained extra_state handling and init_empty_weights unexpected argument are separate serialization/init issues." + }, + { + "left": "issue:45276", + "right": "issue:45335", + "accept": false, + "reason": "Similar symptom around resize_token_embeddings, but different architectures and affected submodules; not enough evidence of one shared bug." + }, + { + "left": "issue:44910", + "right": "issue:45588", + "accept": false, + "reason": "Both touch flash attention, but one is a Qwen3.5 illegal-memory-access issue from packed-sequence interpretation and the other is an AttributeError on s_aux=None." + }, + { + "left": "issue:44402", + "right": "issue:44488", + "accept": false, + "reason": "Tokenizer vocab-size mismatch for one model versus a loading failure for another model; related theme but not the same bug." + }, + { + "left": "issue:43425", + "right": "issue:43957", + "accept": false, + "reason": "Torch 2.10 incompatibility and meta-device loading breakage are distinct compatibility regressions." + }, + { + "left": "issue:43873", + "right": "issue:44387", + "accept": false, + "reason": "Offloading not working with quantization versus increased reserved memory causing int4 OOM; different failure mechanisms." + }, + { + "left": "issue:42947", + "right": "issue:45704", + "accept": false, + "reason": "Gradient checkpointing with PEFT LoRA and a T5 apex RMSNorm memory leak are unrelated." + }, + { + "left": "issue:41720", + "right": "issue:44155", + "accept": false, + "reason": "Qwen3 auto device mapping CUDA assert and AudioFlamingo3 batched inference leakage are different bugs in different models." + }, + { + "left": "issue:43404", + "right": "issue:43454", + "accept": false, + "reason": "Both mention untied lm_head weights, but they are different model classes with separate implementations; no evidence they are the same concrete fix." + }, + { + "left": "issue:39692", + "right": "issue:44960", + "accept": false, + "reason": "SigLIP2 doc-example errors and GLM5 issues are unrelated." + }, + { + "left": "issue:45663", + "right": "issue:45759", + "accept": false, + "reason": "Gemma4 FSDP2 shared_kv_states rebuild bug is unrelated to Qwen 3.5 multimodal config unwrapping." + }, + { + "left": "issue:44291", + "right": "issue:44756", + "accept": false, + "reason": "Unexpected _is_hf_initialized argument under init_empty_weights versus mmap OOM on Strix Halo are different bugs." + }, + { + "left": "issue:36010", + "right": "issue:39404", + "accept": false, + "reason": "Importing GenerationMixin from transformers.generation and Whisper return_language pipeline regression are unrelated." + }, + { + "left": "issue:45663", + "right": "issue:45859", + "accept": false, + "reason": "Gemma4 FSDP2 attention-state bug versus Qwen3.5 CPU-offload incompatibility in rotary embedding are different code paths." + }, + { + "left": "issue:44387", + "right": "issue:44756", + "accept": false, + "reason": "Both are memory-related, but one is int4 quantization reserved-memory growth and the other is mmap-related OOM on Strix Halo." + } + ] + }, + "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": 21, + "review_unit_index": 8, + "review_unit_count": 56, + "status": "reviewed", + "reason": null, + "source": "llm", + "cache_hit": false, + "model": "gpt-5.4-mini?service_tier=flex", + "cluster_id": "cluster-10000-473", + "nodes": [ + "issue:39404", + "issue:42796", + "issue:42947", + "issue:43335", + "issue:43381", + "issue:43425", + "issue:43611", + "issue:43645", + "issue:43646", + "issue:43708", + "issue:43716", + "issue:43828", + "issue:43854", + "issue:43906", + "issue:43957", + "issue:44155", + "issue:44164", + "issue:44292", + "issue:44402", + "issue:44451", + "issue:44479", + "issue:44530", + "issue:44589", + "issue:44683", + "issue:44756", + "issue:44805", + "issue:44863", + "issue:44877", + "issue:44912", + "issue:44928", + "issue:45084", + "issue:45216", + "issue:45276", + "issue:45292", + "issue:45335", + "issue:45588", + "issue:45663", + "issue:45684", + "issue:45685", + "issue:45698", + "issue:45701", + "issue:45812", + "issue:45823", + "issue:46315", + "issue:46326", + "issue:46468", + "issue:46491", + "issue:46665" + ], + "soft_pairs": [ + "issue:44292|issue:44912", + "issue:44155|issue:45823", + "issue:43335|issue:45663", + "issue:42947|issue:43381", + "issue:43957|issue:44756", + "issue:45276|issue:45292", + "issue:44863|issue:46468", + "issue:45685|issue:46468", + "issue:44479|issue:44877", + "issue:44164|issue:44756", + "issue:43381|issue:44928", + "issue:43854|issue:45216", + "issue:45588|issue:45663", + "issue:45684|issue:45698", + "issue:42796|issue:46326", + "issue:45292|issue:45335", + "issue:43646|issue:43708", + "issue:43828|issue:43957", + "issue:44402|issue:44451", + "issue:43645|issue:44164", + "issue:44589|issue:44756", + "issue:43425|issue:43828", + "issue:45812|issue:46315", + "issue:46315|issue:46491", + "issue:39404|issue:45701", + "issue:43611|issue:43645", + "issue:39404|issue:46665", + "issue:44683|issue:45823", + "issue:44530|issue:45084", + "issue:43425|issue:45823", + "issue:43425|issue:43716", + "issue:44589|issue:45084", + "issue:43906|issue:44530", + "issue:44479|issue:44805" + ], + "prepared_review_unit_hash": "8f2621dbeb3ee7f764d906b2ffdd558945d8e070781631384aeb51e864b64707", + "component_budget": { + "node_count": 473, + "item_count": 473, + "soft_pair_count": 1738, + "serialized_chars": 901437, + "estimated_input_tokens": 225360, + "estimated_eval_tokens": 450976 + }, + "budget": { + "node_count": 48, + "item_count": 48, + "soft_pair_count": 34, + "serialized_chars": 37876, + "estimated_input_tokens": 9469, + "estimated_eval_tokens": 19194 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": true, + "analyst_result": { + "summary": "Most items are unrelated one-off bugs. The only clear duplicate cluster is the token-embedding resize regression around `resize_token_embeddings`; `issue:45292` is the best representative of that group.", + "confidence": 0.82, + "canonical_issue_reason": "`issue:45292` is the most general and central report in the resize-token-embeddings cluster, and the other matching reports are model-specific variants of the same symptom.", + "canonical_pr_reason": null, + "best_issue_reason": "`issue:45292` is the best global issue candidate because it describes the underlying bug at the highest level (`resize_token_embeddings` not updating output embeddings), making it the most reusable canonical anchor.", + "best_pr_reason": null, + "soft_edge_verdicts": [ + { + "left": "issue:44292", + "right": "issue:44912", + "accept": false, + "reason": "Both are quantization/loading problems, but they involve different models, quantization formats, and failure modes; not the same bug." + }, + { + "left": "issue:44155", + "right": "issue:45823", + "accept": false, + "reason": "AudioFlamingo track leakage during batched inference is unrelated to Gemma4 PLE device placement in forward." + }, + { + "left": "issue:43335", + "right": "issue:45663", + "accept": false, + "reason": "One is a SwitchTransformers config edge case; the other is a Gemma4 FSDP2 attention state bug. Different code paths." + }, + { + "left": "issue:42947", + "right": "issue:43381", + "accept": false, + "reason": "Both mention gradient checkpointing, but one is about LoRA/PEFT effectiveness and the other is an eval-mode restriction; not the same defect." + }, + { + "left": "issue:43957", + "right": "issue:44756", + "accept": false, + "reason": "Meta-device loading failures and Strix Halo mmap OOMs are different issues with different root causes." + }, + { + "left": "issue:45276", + "right": "issue:45292", + "accept": true, + "reason": "Both report `resize_token_embeddings` failing to update output/token embeddings; `45276` is a Gemma4-specific manifestation of the same bug." + }, + { + "left": "issue:44863", + "right": "issue:46468", + "accept": false, + "reason": "NemotronH checkpoint loading and a broken Nemotron example are related only by model family, not the same concrete bug." + }, + { + "left": "issue:45685", + "right": "issue:46468", + "accept": false, + "reason": "An MPS histogram kernel limitation in MoE is unrelated to the Nemotron example breakage." + }, + { + "left": "issue:44479", + "right": "issue:44877", + "accept": false, + "reason": "A Qwen VL video regression and a Granite speech config-loading issue are different model-specific failures." + }, + { + "left": "issue:44164", + "right": "issue:44756", + "accept": false, + "reason": "`extra_state` save/load handling is unrelated to disabling mmap on Strix Halo." + }, + { + "left": "issue:43381", + "right": "issue:44928", + "accept": false, + "reason": "Eval-mode gradient checkpointing and Qwen3.5 SDPA math-fallback NaNs are distinct bugs." + }, + { + "left": "issue:43854", + "right": "issue:45216", + "accept": false, + "reason": "GLM-4.7-Flash unit-test loading failure and Qwen3.5 save_pretrained regression are separate issues." + }, + { + "left": "issue:45588", + "right": "issue:45663", + "accept": false, + "reason": "Flash-attention sink handling and Gemma4 FSDP2 shared-kv state rebuilding are different code paths." + }, + { + "left": "issue:45684", + "right": "issue:45698", + "accept": false, + "reason": "Read-only permission propagation during save and wrong custom-module loading after save are not the same underlying bug." + }, + { + "left": "issue:42796", + "right": "issue:46326", + "accept": false, + "reason": "Tokenizer BPE-marker handling and `continue_final_message` behavior are related to tokenization but not the same defect." + }, + { + "left": "issue:45292", + "right": "issue:45335", + "accept": true, + "reason": "Both describe the same `resize_token_embeddings` regression failing to propagate resized embeddings to model-specific output/decoder embeddings." + }, + { + "left": "issue:43646", + "right": "issue:43708", + "accept": false, + "reason": "Custom model initialization in notebooks and checkpoint-resume `max_steps` calculation are unrelated." + }, + { + "left": "issue:43828", + "right": "issue:43957", + "accept": false, + "reason": "Autocast dtype mismatch and meta-device loading failures are different runtime problems." + }, + { + "left": "issue:44402", + "right": "issue:44451", + "accept": false, + "reason": "A vocab-size mismatch in one model and a separate ScandiBERT loading failure are not clearly the same bug." + }, + { + "left": "issue:43645", + "right": "issue:44164", + "accept": false, + "reason": "Notebook custom-model initialization and `extra_state` save/load handling are different issues." + }, + { + "left": "issue:44589", + "right": "issue:44756", + "accept": false, + "reason": "Missing Float8 storage support and Strix Halo mmap OOMs are unrelated." + }, + { + "left": "issue:43425", + "right": "issue:43828", + "accept": false, + "reason": "Torch 2.10 incompatibility is broader and different from an autocast dtype-mismatch bug." + }, + { + "left": "issue:45812", + "right": "issue:46315", + "accept": false, + "reason": "Wrong Granite token IDs and slow `convert_tokens_to_ids` are both tokenizer regressions but not the same bug." + }, + { + "left": "issue:46315", + "right": "issue:46491", + "accept": false, + "reason": "Tokenizer conversion slowness and CodeLlama leading-space round-trip corruption are different failures." + }, + { + "left": "issue:39404", + "right": "issue:45701", + "accept": false, + "reason": "Whisper pipeline `return_language` breakage and generic tokenization changes do not look like the same underlying issue." + }, + { + "left": "issue:43611", + "right": "issue:43645", + "accept": false, + "reason": "Both are about Transformers 5/custom models, but one is `base_model_prefix` loading and the other is notebook initialization; not one bug." + }, + { + "left": "issue:39404", + "right": "issue:46665", + "accept": false, + "reason": "Whisper pipeline behavior and `from_pretrained` inside Python classes are unrelated." + }, + { + "left": "issue:44683", + "right": "issue:45823", + "accept": false, + "reason": "Compiled flex_attention on new torch versions and Gemma4 device-mismatch during forward are different problems." + }, + { + "left": "issue:44530", + "right": "issue:45084", + "accept": false, + "reason": "PagedAttentionCache linear-attention group handling and a compiler node-type error are unrelated." + }, + { + "left": "issue:43425", + "right": "issue:45823", + "accept": false, + "reason": "Torch version compatibility and Gemma4 device placement are not the same bug." + }, + { + "left": "issue:43425", + "right": "issue:43716", + "accept": false, + "reason": "Torch compatibility and Mistral-3 image-preprocessor dtype mismatch are distinct issues." + }, + { + "left": "issue:44589", + "right": "issue:45084", + "accept": false, + "reason": "Float8 storage lookup failure and template-node compilation errors are separate code paths." + }, + { + "left": "issue:43906", + "right": "issue:44530", + "accept": false, + "reason": "An isolated reproduction of another issue and a PagedAttentionCache crash are not the same bug." + }, + { + "left": "issue:44479", + "right": "issue:44805", + "accept": false, + "reason": "A Qwen VL video regression and a mask-shape IndexError are different failure modes." + } + ] + }, + "evaluator_result": { + "accept": true, + "feedback": "The summary is grounded in the packet and stays conservative. The only clearly supported duplicate cluster is the resize_token_embeddings regression around issue:45292, and treating issue:45292 as the canonical issue is reasonable given the titles. The soft-edge acceptances for 45276\u219445292 and 45292\u219445335 are well supported by 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": 21, + "review_unit_index": 9, + "review_unit_count": 56, + "status": "reviewed", + "reason": null, + "source": "llm", + "cache_hit": false, + "model": "gpt-5.4-mini?service_tier=flex", + "cluster_id": "cluster-10000-473", + "nodes": [ + "issue:36032", + "issue:39404", + "issue:42796", + "issue:42915", + "issue:43299", + "issue:43404", + "issue:43425", + "issue:43452", + "issue:43582", + "issue:43596", + "issue:43611", + "issue:43632", + "issue:43643", + "issue:43644", + "issue:43716", + "issue:43856", + "issue:43901", + "issue:43906", + "issue:43931", + "issue:43994", + "issue:44079", + "issue:44164", + "issue:44265", + "issue:44462", + "issue:44479", + "issue:44488", + "issue:44509", + "issue:44530", + "issue:44661", + "issue:44671", + "issue:44683", + "issue:44737", + "issue:44756", + "issue:44898", + "issue:44912", + "issue:44918", + "issue:44987", + "issue:44991", + "issue:45081", + "issue:45084", + "issue:45759", + "issue:45859", + "issue:46050", + "issue:46285", + "issue:46326", + "issue:46402", + "issue:46489" + ], + "soft_pairs": [ + "issue:43901|issue:44509", + "issue:44265|issue:44898", + "issue:43425|issue:44918", + "issue:39404|issue:44912", + "issue:43994|issue:44079", + "issue:44671|issue:44991", + "issue:36032|issue:43452", + "issue:46050|issue:46326", + "issue:43299|issue:43931", + "issue:44164|issue:45759", + "issue:43906|issue:46326", + "issue:44991|issue:45081", + "issue:43856|issue:44756", + "issue:43596|issue:44683", + "issue:44164|issue:44479", + "issue:44488|issue:45081", + "issue:44462|issue:46489", + "issue:44488|issue:44987", + "issue:42796|issue:43404", + "issue:43611|issue:43644", + "issue:39404|issue:44671", + "issue:44737|issue:45859", + "issue:43716|issue:45084", + "issue:46285|issue:46402", + "issue:36032|issue:42915", + "issue:44661|issue:44683", + "issue:43582|issue:43632", + "issue:43643|issue:44530" + ], + "prepared_review_unit_hash": "127a26ba89caf1705a6bb8097aca4e64935f5b393590c7176af40964da550ae2", + "component_budget": { + "node_count": 473, + "item_count": 473, + "soft_pair_count": 1738, + "serialized_chars": 901437, + "estimated_input_tokens": 225360, + "estimated_eval_tokens": 450976 + }, + "budget": { + "node_count": 47, + "item_count": 47, + "soft_pair_count": 28, + "serialized_chars": 35589, + "estimated_input_tokens": 8898, + "estimated_eval_tokens": 18052 + }, + "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, especially tokenizer/model-loading breakages, plus some unrelated docs, quantization, and training/runtime issues. None of the soft pairs look like true duplicates; they are only loosely similar by subsystem or failure mode.", + "confidence": 0.78, + "canonical_issue_reason": "issue:43611 is the broadest and most central representative of the recurring v5 model-loading regression theme, covering a core load-path break that many of the other issues resemble only loosely.", + "canonical_pr_reason": null, + "best_issue_reason": "issue:43611 best captures the shared cluster theme of Transformers v5 breaking model/tokenizer loading in downstream models.", + "best_pr_reason": null, + "soft_edge_verdicts": [ + { + "left": "issue:43901", + "right": "issue:44509", + "accept": false, + "reason": "Both are docs regressions, but one is about `return_all_scores` in text classification and the other about removed pipeline tasks; different documentation gaps." + }, + { + "left": "issue:44265", + "right": "issue:44898", + "accept": false, + "reason": "Different failures: `torch.export` with `torch_compilable_check` versus Perceiver image classification with non-default resolution." + }, + { + "left": "issue:43425", + "right": "issue:44918", + "accept": false, + "reason": "Unrelated problems: Torch 2.10 compatibility versus Qwen3.5 embedding unpacking in TRL SFT trainer." + }, + { + "left": "issue:39404", + "right": "issue:44912", + "accept": false, + "reason": "Whisper pipeline `return_language` regression and MXFP4 quantization fallback are separate code paths." + }, + { + "left": "issue:43994", + "right": "issue:44079", + "accept": false, + "reason": "One is a SigLIP2 output-quality regression, the other is `ModelOutput` key assignment behavior; not the same bug." + }, + { + "left": "issue:44671", + "right": "issue:44991", + "accept": false, + "reason": "CamemBERT masked-LM prediction error and an EMBEDDIA tokenizer load failure are distinct tokenizer/model issues." + }, + { + "left": "issue:36032", + "right": "issue:43452", + "accept": false, + "reason": "T5 tokenizer `add_special_tokens` conflict and `gguf_file` loading failures are different tokenizer-loading bugs." + }, + { + "left": "issue:46050", + "right": "issue:46326", + "accept": false, + "reason": "Quantized model save/load corruption via `sinq` is unrelated to `continue_final_message` behavior for `mistral_common` tokenizers." + }, + { + "left": "issue:43299", + "right": "issue:43931", + "accept": false, + "reason": "Both involve Qwen3-VL loading, but one is a MoE regression and the other is a weight-shape mismatch on a specific model; not clearly the same fixable bug." + }, + { + "left": "issue:44164", + "right": "issue:45759", + "accept": false, + "reason": "`extra_state` save/load handling and `from_config` unwrapping of `text_config` are separate serialization/config issues." + }, + { + "left": "issue:43906", + "right": "issue:46326", + "accept": false, + "reason": "The isolated reproduction of issue 38071 is not the same as the mistral_common `continue_final_message` regression." + }, + { + "left": "issue:44991", + "right": "issue:45081", + "accept": false, + "reason": "Different tokenizer failures: EMBEDDIA/est-roberta loading versus `_patch_mistral_regex` crashing with missing `backend_tokenizer`." + }, + { + "left": "issue:43856", + "right": "issue:44756", + "accept": false, + "reason": "Training memory inefficiency on Qwen3 MoE and Strix Halo mmap OOM are unrelated resource issues." + }, + { + "left": "issue:43596", + "right": "issue:44683", + "accept": false, + "reason": "Deepspeed ZeRO-3/BertModel index error is unrelated to compiled `flex_attention` failures on newer torch." + }, + { + "left": "issue:46285", + "right": "issue:46402", + "accept": false, + "reason": "CLIP structure inconsistency and Sam3Model text-encoder weights not loading are different model API regressions." + }, + { + "left": "issue:36032", + "right": "issue:42915", + "accept": false, + "reason": "T5 tokenizer special-token conflict and Qwen3Moe FineGrainedFP8Config failure are unrelated model-loading problems." + }, + { + "left": "issue:44661", + "right": "issue:44683", + "accept": false, + "reason": "`add-new-model-like` failing in `TOKENIZER_MAPPING_NAMES` and `torch_compilable_check`/`flex_attention` are different internals." + }, + { + "left": "issue:43582", + "right": "issue:43632", + "accept": false, + "reason": "Apple Silicon `caching_allocator_warmup` TypeError and `_is_hf_initialized` flag breakage do not share the same code path." + }, + { + "left": "issue:43643", + "right": "issue:44530", + "accept": false, + "reason": "`trust_remote_code` config-field loss and Qwen3.5 `PagedAttentionCache` linear_attention crash 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": 21, + "review_unit_index": 10, + "review_unit_count": 56, + "status": "reviewed", + "reason": null, + "source": "llm", + "cache_hit": false, + "model": "gpt-5.4-mini?service_tier=flex", + "cluster_id": "cluster-10000-473", + "nodes": [ + "issue:28530", + "issue:34567", + "issue:36032", + "issue:40444", + "issue:40990", + "issue:42175", + "issue:42915", + "issue:43262", + "issue:43317", + "issue:43322", + "issue:43377", + "issue:43404", + "issue:43421", + "issue:43452", + "issue:43482", + "issue:43540", + "issue:43541", + "issue:43644", + "issue:43645", + "issue:43646", + "issue:43746", + "issue:43819", + "issue:43883", + "issue:43994", + "issue:44206", + "issue:44220", + "issue:44265", + "issue:44291", + "issue:44403", + "issue:44451", + "issue:44464", + "issue:44466", + "issue:44488", + "issue:44530", + "issue:44534", + "issue:44610", + "issue:44811", + "issue:44849", + "issue:44987", + "issue:44991", + "issue:45081", + "issue:45084", + "issue:45325", + "issue:45507", + "issue:45584", + "issue:45987", + "issue:46178", + "issue:46326" + ], + "soft_pairs": [ + "issue:40990|issue:43421", + "issue:43541|issue:45084", + "issue:44451|issue:44987", + "issue:42915|issue:44530", + "issue:34567|issue:40444", + "issue:42175|issue:46178", + "issue:43644|issue:43646", + "issue:44811|issue:45584", + "issue:44206|issue:44220", + "issue:43883|issue:44488", + "issue:40444|issue:46326", + "issue:36032|issue:43645", + "issue:43452|issue:43482", + "issue:43317|issue:43482", + "issue:45084|issue:45507", + "issue:44987|issue:44991", + "issue:43262|issue:43377", + "issue:43746|issue:46326", + "issue:44464|issue:45987", + "issue:42915|issue:45084", + "issue:43646|issue:44534", + "issue:43883|issue:44991", + "issue:44403|issue:45081", + "issue:43645|issue:44534", + "issue:43404|issue:44466", + "issue:43377|issue:43819", + "issue:43994|issue:44610", + "issue:28530|issue:34567", + "issue:44849|issue:45325", + "issue:44451|issue:45081", + "issue:44987|issue:45081", + "issue:42915|issue:44291", + "issue:43322|issue:43540", + "issue:44265|issue:44610" + ], + "prepared_review_unit_hash": "c4887b2e36ab5e5176d57f7c75abfeac0b7dd81beda10a7d22f796b2171aa300", + "component_budget": { + "node_count": 473, + "item_count": 473, + "soft_pair_count": 1738, + "serialized_chars": 901437, + "estimated_input_tokens": 225360, + "estimated_eval_tokens": 450976 + }, + "budget": { + "node_count": 48, + "item_count": 48, + "soft_pair_count": 34, + "serialized_chars": 38074, + "estimated_input_tokens": 9519, + "estimated_eval_tokens": 19294 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": true, + "analyst_result": { + "summary": "The cluster is mostly noisy and spans many unrelated bugs; several items are generic Transformers v5/model-loading regressions, but none of the soft pairs are strong enough to merge as duplicates.", + "confidence": 0.78, + "canonical_issue_reason": "issue:44987 is the most representative of the recurring v5 loading-regression theme in this set: it is a concrete, version-bounded model-load failure and is broader than the more niche audio/training/model-architecture reports.", + "canonical_pr_reason": null, + "best_issue_reason": "issue:44987 is the best cluster representative because it captures the common model/tokenizer loading regression pattern better than the other, more specialized issues.", + "best_pr_reason": null, + "soft_edge_verdicts": [ + { + "left": "issue:40990", + "right": "issue:43421", + "accept": false, + "reason": "Perplexity-quality complaint vs tokenizer post-processor update bug; different symptoms and code paths." + }, + { + "left": "issue:43541", + "right": "issue:45084", + "accept": false, + "reason": "Torch Dynamo grouped_mm tracing failure vs template-node compile error; unrelated failures." + }, + { + "left": "issue:44451", + "right": "issue:44987", + "accept": false, + "reason": "Both are load failures, but they target different models and no shared root cause is shown." + }, + { + "left": "issue:42915", + "right": "issue:44530", + "accept": false, + "reason": "Qwen3Moe FP8 config failure vs Qwen3.5 paged-attention linear_attention crash; different defects." + }, + { + "left": "issue:34567", + "right": "issue:40444", + "accept": false, + "reason": "Trainer token counter update bug vs multimodal iterable-dataset failure; unrelated." + }, + { + "left": "issue:42175", + "right": "issue:46178", + "accept": false, + "reason": "Packaging/backend inclusion issue vs backend-mapping error for tensorflow_text; distinct bugs." + }, + { + "left": "issue:43644", + "right": "issue:43646", + "accept": false, + "reason": "Both are v5 regressions, but non-persistent-buffer junk and custom model initialization are separate problems." + }, + { + "left": "issue:44811", + "right": "issue:45584", + "accept": false, + "reason": "Whisper batch_decode skip_special_tokens issue vs empty-transcription handling after align_special_tokens; different code paths." + }, + { + "left": "issue:44206", + "right": "issue:44220", + "accept": false, + "reason": "Both touch audio feature extraction, but one is a center-arg regression and the other is a lower-level fbank helper issue; not clearly the same bug." + }, + { + "left": "issue:43883", + "right": "issue:44488", + "accept": false, + "reason": "Missing all_tied_weights_keys attribute vs specific model-loading failure for cjvt/sleng-bert; unrelated." + }, + { + "left": "issue:40444", + "right": "issue:46326", + "accept": false, + "reason": "Multimodal iterable dataset with multiple images vs mistral_common continue_final_message tokenizer bug; unrelated." + }, + { + "left": "issue:36032", + "right": "issue:43645", + "accept": false, + "reason": "Tokenizer method-name conflict vs Jupyter/custom-model initialization regression; different defects." + }, + { + "left": "issue:43452", + "right": "issue:43482", + "accept": false, + "reason": "Both involve GGUF loading, but one is a generic API break and the other is a specific Qwen2.5-GGUF regression; not the same underlying bug." + }, + { + "left": "issue:43317", + "right": "issue:43482", + "accept": false, + "reason": "gpu+cpu offload dequantized-model loading failure vs GGUF loading failure; different failure modes." + }, + { + "left": "issue:45084", + "right": "issue:45507", + "accept": false, + "reason": "Template compilation error vs GraniteMoEHybrid invalid-method call; no shared code-path shown." + }, + { + "left": "issue:44987", + "right": "issue:44991", + "accept": false, + "reason": "Both are versioned load failures, but they involve different models/tokenizer paths and there is no evidence of one shared root cause." + }, + { + "left": "issue:43262", + "right": "issue:43377", + "accept": false, + "reason": "Audio processor sampling-rate default bug vs MIMI batched-vs-single output mismatch; distinct issues." + }, + { + "left": "issue:43746", + "right": "issue:46326", + "accept": false, + "reason": "PEFT checkpoint loading problem vs mistral_common continue_final_message bug; unrelated." + }, + { + "left": "issue:44464", + "right": "issue:45987", + "accept": false, + "reason": "Compiled-forward chunked-generation inconsistency vs StaticCache seq-length return-type bug; different problems." + }, + { + "left": "issue:42915", + "right": "issue:44291", + "accept": false, + "reason": "Qwen3Moe FP8 failure vs init_empty_weights unexpected-argument error; unrelated." + }, + { + "left": "issue:43646", + "right": "issue:44534", + "accept": false, + "reason": "Custom model initialization regression vs non-persistent buffer serialization junk; separate bugs." + }, + { + "left": "issue:43883", + "right": "issue:44991", + "accept": false, + "reason": "Tied-weights attribute error vs tokenizer load failure for a specific model; different issues." + }, + { + "left": "issue:44403", + "right": "issue:45081", + "accept": false, + "reason": "Generic loading noise vs Mistral regex patch crash; not the same defect." + }, + { + "left": "issue:43645", + "right": "issue:44534", + "accept": false, + "reason": "Notebook/custom-model initialization problem vs buffer-junk regression; distinct bugs." + }, + { + "left": "issue:43404", + "right": "issue:44466", + "accept": false, + "reason": "Both involve lm_head/tied weights, but one is missing tying in Mistral3 and the other is device-dependent serialization inconsistency; separate bugs." + }, + { + "left": "issue:43377", + "right": "issue:43819", + "accept": false, + "reason": "MIMI padding-mask bug vs DAC.from_latents/STE mismatch; unrelated audio-model issues." + }, + { + "left": "issue:43994", + "right": "issue:44610", + "accept": false, + "reason": "SigLIP2 nonsensical outputs vs OmDet-Turbo processor size mismatch; different root causes." + }, + { + "left": "issue:28530", + "right": "issue:34567", + "accept": false, + "reason": "Early-stopping metric selection issue vs num_input_tokens_seen not updating; unrelated trainer-state bugs." + }, + { + "left": "issue:44849", + "right": "issue:45325", + "accept": false, + "reason": "Qwen3.5 hidden-states bug vs Qwen2.5-VL rope-index scaling bug; different model internals." + }, + { + "left": "issue:44451", + "right": "issue:45081", + "accept": false, + "reason": "Model loading failure vs Mistral tokenizer regex-patch crash; different code paths and symptoms." + }, + { + "left": "issue:44987", + "right": "issue:45081", + "accept": false, + "reason": "Versioned model-loading regression vs tokenizer regex-patch crash; not the same underlying bug." + }, + { + "left": "issue:43322", + "right": "issue:43540", + "accept": false, + "reason": "Llava Next segfault vs Qwen3OmniMoe video-input ValueError; unrelated multimodal failures." + }, + { + "left": "issue:44265", + "right": "issue:44610", + "accept": false, + "reason": "torch.export export failure with torch_compilable_check vs processor/model image-size mismatch; separate 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": 21, + "review_unit_index": 11, + "review_unit_count": 56, + "status": "reviewed", + "reason": null, + "source": "llm", + "cache_hit": false, + "model": "gpt-5.4-mini?service_tier=flex", + "cluster_id": "cluster-10000-473", + "nodes": [ + "issue:33666", + "issue:35707", + "issue:36032", + "issue:40444", + "issue:40990", + "issue:41720", + "issue:42796", + "issue:42915", + "issue:43089", + "issue:43296", + "issue:43317", + "issue:43381", + "issue:43421", + "issue:43425", + "issue:43482", + "issue:43525", + "issue:43531", + "issue:43611", + "issue:43632", + "issue:43742", + "issue:43746", + "issue:43792", + "issue:43856", + "issue:44062", + "issue:44079", + "issue:44220", + "issue:44291", + "issue:44403", + "issue:44479", + "issue:44534", + "issue:44568", + "issue:44589", + "issue:44683", + "issue:44756", + "issue:44811", + "issue:44877", + "issue:44987", + "issue:44991", + "issue:45216", + "issue:45325", + "issue:45357", + "issue:45538", + "issue:45663", + "issue:45812", + "issue:45923", + "issue:45987", + "issue:46315", + "issue:46491" + ], + "soft_pairs": [ + "issue:43089|issue:45987", + "issue:36032|issue:43742", + "issue:42915|issue:44589", + "issue:44062|issue:44987", + "issue:45216|issue:45357", + "issue:44479|issue:45325", + "issue:44991|issue:46315", + "issue:43531|issue:44877", + "issue:36032|issue:41720", + "issue:44220|issue:44683", + "issue:43425|issue:44403", + "issue:43296|issue:43531", + "issue:43611|issue:44534", + "issue:40444|issue:45325", + "issue:43632|issue:44756", + "issue:45812|issue:46491", + "issue:44403|issue:44991", + "issue:42796|issue:43746", + "issue:43792|issue:44589", + "issue:33666|issue:40444", + "issue:43381|issue:45923", + "issue:44291|issue:44589", + "issue:44079|issue:45663", + "issue:40990|issue:44811", + "issue:43632|issue:44534", + "issue:43482|issue:43525", + "issue:35707|issue:40444", + "issue:43482|issue:45538", + "issue:43421|issue:44568", + "issue:43317|issue:43856" + ], + "prepared_review_unit_hash": "37903790f1f3cf30625d376a5605a3748698aa6f6e7b6d2ef6e0aad8061a976d", + "component_budget": { + "node_count": 473, + "item_count": 473, + "soft_pair_count": 1738, + "serialized_chars": 901437, + "estimated_input_tokens": 225360, + "estimated_eval_tokens": 450976 + }, + "budget": { + "node_count": 48, + "item_count": 48, + "soft_pair_count": 30, + "serialized_chars": 36549, + "estimated_input_tokens": 9138, + "estimated_eval_tokens": 18532 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": true, + "analyst_result": { + "summary": "The cluster is mostly heterogeneous: tokenizer regressions, Qwen/Qwen3 MoE issues, generation/cache bugs, and unrelated loading/config errors. I don\u2019t see any soft pair that clearly describes the same concrete bug/change, so I reject all soft edges. For a representative issue, 45812 is the strongest general regression example, but the cluster does not really have one true canonical bug.", + "confidence": 0.84, + "canonical_issue_reason": "issue:45812 is the clearest high-signal regression in the set and has the broadest, most explicit symptom description, but it is only a representative choice because the cluster is not actually cohesive.", + "canonical_pr_reason": null, + "best_issue_reason": "issue:45812 has the strongest standalone bug report quality: explicit regression, clear affected models, and a concrete incorrect-result symptom.", + "best_pr_reason": null, + "soft_edge_verdicts": [ + { + "left": "issue:43089", + "right": "issue:45987", + "accept": false, + "reason": "Generation throughput overhead and a StaticCache return-type bug are different problems; they do not look mergeable into one fix." + }, + { + "left": "issue:36032", + "right": "issue:43742", + "accept": false, + "reason": "Tokenizer method-name conflict vs MobileLLM loading KeyError are unrelated failure modes." + }, + { + "left": "issue:42915", + "right": "issue:44589", + "accept": false, + "reason": "Both mention FP8/floating-point issues, but one is a model/config failure and the other is a storage-type error; not the same code-path bug." + }, + { + "left": "issue:44062", + "right": "issue:44987", + "accept": false, + "reason": "AddedToken keyword-argument conflict vs loading a specific model are distinct tokenizer/load issues." + }, + { + "left": "issue:45216", + "right": "issue:45357", + "accept": false, + "reason": "Both concern Qwen3.5 save_pretrained, but one is a broad checkpoint corruption report and the other is specifically incorrect visual-encoder keys; related, not clearly identical." + }, + { + "left": "issue:44479", + "right": "issue:45325", + "accept": false, + "reason": "Video-input regression and still-image temporal position-id scaling are different multimodal bugs." + }, + { + "left": "issue:44991", + "right": "issue:46315", + "accept": false, + "reason": "Tokenizer loading failure vs slow convert_tokens_to_ids regression are different symptoms and likely different code paths." + }, + { + "left": "issue:43531", + "right": "issue:44877", + "accept": false, + "reason": "Qwen3-MoE sliding_window behavior and granite_speech strict-config loading are unrelated." + }, + { + "left": "issue:36032", + "right": "issue:41720", + "accept": false, + "reason": "Tokenizer attribute conflict vs Qwen3 auto-device-map cuda assert are unrelated." + }, + { + "left": "issue:44220", + "right": "issue:44683", + "accept": false, + "reason": "Audio feature extraction bug vs compiled flex_attention torch-version failure are unrelated." + }, + { + "left": "issue:43425", + "right": "issue:44403", + "accept": false, + "reason": "Torch 2.10 compatibility and loading noise are different issues." + }, + { + "left": "issue:43296", + "right": "issue:43531", + "accept": false, + "reason": "PaddleOCR-VL load failure and Qwen3-MoE sliding_window issue do not share the same bug." + }, + { + "left": "issue:43611", + "right": "issue:44534", + "accept": false, + "reason": "base_model_prefix loading regression and non-persistent buffer junk are different transformer-v5 regressions." + }, + { + "left": "issue:40444", + "right": "issue:45325", + "accept": false, + "reason": "Multi-image IterableDataset finetuning failure vs Qwen2.5-VL rope-index scaling bug are unrelated." + }, + { + "left": "issue:43632", + "right": "issue:44756", + "accept": false, + "reason": "_is_hf_initialized flag handling and mmap-on-Strix-Halo OOM are unrelated." + }, + { + "left": "issue:45812", + "right": "issue:46491", + "accept": false, + "reason": "Both are tokenizer regressions, but one is Granite token IDs and the other is CodeLlama space stripping; different model-specific bugs." + }, + { + "left": "issue:44403", + "right": "issue:44991", + "accept": false, + "reason": "Unnecessary loading noise and tokenizer-loading failure are not the same problem." + }, + { + "left": "issue:42796", + "right": "issue:43746", + "accept": false, + "reason": "BPE-marker handling and GraniteSpeech PEFT checkpoint loading are unrelated." + }, + { + "left": "issue:43792", + "right": "issue:44589", + "accept": false, + "reason": "Whisper runtime failure and Float8 storage lookup error are different issues." + }, + { + "left": "issue:33666", + "right": "issue:40444", + "accept": false, + "reason": "Multi-GPU training and multi-image iterable-dataset finetuning are not the same bug." + }, + { + "left": "issue:43381", + "right": "issue:45923", + "accept": false, + "reason": "Eval-mode gradient checkpointing limitation and missing trust_remote_code support flag are related only at a very high level, not the same concrete issue." + }, + { + "left": "issue:44291", + "right": "issue:44589", + "accept": false, + "reason": "Unexpected _is_hf_initialized argument and Float8 storage object errors are different transformer-v5 breakages." + }, + { + "left": "issue:44079", + "right": "issue:45663", + "accept": false, + "reason": "ModelOutput key assignment and Gemma4 FSDP2 shared_kv_states KeyError are unrelated." + }, + { + "left": "issue:40990", + "right": "issue:44811", + "accept": false, + "reason": "High perplexity and Whisper batch_decode skip_special_tokens behavior are unrelated." + }, + { + "left": "issue:43632", + "right": "issue:44534", + "accept": false, + "reason": "Different v5 regressions affecting different initialization/buffer behaviors." + }, + { + "left": "issue:43482", + "right": "issue:43525", + "accept": false, + "reason": "Qwen2.5-GGUF loading under transformers v5 and Llama4Config pad_token_id error are unrelated." + }, + { + "left": "issue:35707", + "right": "issue:40444", + "accept": false, + "reason": "Progressive generation with inputs_embeds/past_key_values and multi-image finetuning are different bugs." + }, + { + "left": "issue:43482", + "right": "issue:45538", + "accept": false, + "reason": "GGUF loading failure and CLIPTokenizer model_max_length are unrelated tokenizer/model-loading issues." + }, + { + "left": "issue:43421", + "right": "issue:44568", + "accept": false, + "reason": "Runtime special-token update behavior and mdeberta BOS/EOS add_special_tokens regression are different tokenizer bugs." + }, + { + "left": "issue:43317", + "right": "issue:43856", + "accept": false, + "reason": "gpu+cpu offload dequantized-model loading and Qwen3 MoE training memory usage are not the same failure." + } + ] + }, + "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": 21, + "review_unit_index": 12, + "review_unit_count": 56, + "status": "reviewed", + "reason": null, + "source": "llm", + "cache_hit": false, + "model": "gpt-5.4-mini?service_tier=flex", + "cluster_id": "cluster-10000-473", + "nodes": [ + "issue:40444", + "issue:41553", + "issue:41720", + "issue:42175", + "issue:42796", + "issue:42915", + "issue:43296", + "issue:43377", + "issue:43454", + "issue:43482", + "issue:43531", + "issue:43582", + "issue:43606", + "issue:43611", + "issue:43632", + "issue:43644", + "issue:43646", + "issue:43701", + "issue:43749", + "issue:43792", + "issue:43818", + "issue:43873", + "issue:43957", + "issue:43994", + "issue:44079", + "issue:44265", + "issue:44291", + "issue:44462", + "issue:44466", + "issue:44479", + "issue:44534", + "issue:44829", + "issue:44877", + "issue:44898", + "issue:45072", + "issue:45081", + "issue:45325", + "issue:45357", + "issue:45381", + "issue:45479", + "issue:45491", + "issue:45542", + "issue:45701", + "issue:45800", + "issue:45812", + "issue:45823", + "issue:45834", + "issue:46402" + ], + "soft_pairs": [ + "issue:41553|issue:43582", + "issue:42175|issue:45542", + "issue:43531|issue:44291", + "issue:44829|issue:45479", + "issue:44079|issue:44534", + "issue:44462|issue:45812", + "issue:41720|issue:45081", + "issue:43818|issue:46402", + "issue:43377|issue:45491", + "issue:43818|issue:45357", + "issue:43611|issue:43646", + "issue:43644|issue:44079", + "issue:44265|issue:45072", + "issue:43749|issue:44079", + "issue:42175|issue:43296", + "issue:40444|issue:45381", + "issue:45325|issue:45381", + "issue:43994|issue:44898", + "issue:43792|issue:43957", + "issue:43454|issue:44466", + "issue:45325|issue:45834", + "issue:42915|issue:43482", + "issue:43632|issue:45072", + "issue:42796|issue:45381", + "issue:43701|issue:45325", + "issue:44291|issue:44877", + "issue:43873|issue:45701", + "issue:43296|issue:45800", + "issue:44479|issue:45823", + "issue:43606|issue:44265" + ], + "prepared_review_unit_hash": "1d02e3e12caccb9d9fefc0886009890045f18b1e1203714df6a192d332fcd9c4", + "component_budget": { + "node_count": 473, + "item_count": 473, + "soft_pair_count": 1738, + "serialized_chars": 901437, + "estimated_input_tokens": 225360, + "estimated_eval_tokens": 450976 + }, + "budget": { + "node_count": 48, + "item_count": 48, + "soft_pair_count": 30, + "serialized_chars": 36848, + "estimated_input_tokens": 9212, + "estimated_eval_tokens": 18680 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": true, + "analyst_result": { + "summary": "The cluster is mostly a grab bag of unrelated issues, with one likely duplicate pair around Qwen2.5-VL video/vision position-id handling. Most other soft-similarity pairs are only loosely related by subsystem and should stay separate.", + "confidence": 0.49, + "canonical_issue_reason": "issue:44479 is the best umbrella issue: it describes the broad Qwen multimodal video-input regression across multiple model variants, making it the most useful canonical target for the Qwen position-id bug family.", + "canonical_pr_reason": null, + "best_issue_reason": "issue:44479 is the strongest single representative because it captures the broadest user-facing regression and can subsume narrower Qwen2.5-VL position-id reports.", + "best_pr_reason": null, + "soft_edge_verdicts": [ + { + "left": "issue:41553", + "right": "issue:43582", + "accept": false, + "reason": "Different failures: tokenizer loading error message vs Apple Silicon allocator TypeError." + }, + { + "left": "issue:42175", + "right": "issue:45542", + "accept": false, + "reason": "Both involve TF-related packaging, but one is missing TF in the torch extra and the other is a tensorboard-without-TF backend error." + }, + { + "left": "issue:43531", + "right": "issue:44291", + "accept": false, + "reason": "Unrelated: Qwen3-MoE sliding-window behavior vs init_empty_weights/_is_hf_initialized argument handling." + }, + { + "left": "issue:44829", + "right": "issue:45479", + "accept": false, + "reason": "Both are classification-training regressions, but the concrete bugs differ: flash_attention_3 vs num_labels=1 loss degeneracy." + }, + { + "left": "issue:44079", + "right": "issue:44534", + "accept": false, + "reason": "Different code paths: ModelOutput key assignment vs non-persistent buffer serialization." + }, + { + "left": "issue:44462", + "right": "issue:45812", + "accept": false, + "reason": "Both concern AutoTokenizer, but one ignores tokenizer.json while the other returns wrong Granite token IDs." + }, + { + "left": "issue:41720", + "right": "issue:45081", + "accept": false, + "reason": "Different subsystems and bugs: device-map cuda assert vs Mistral regex patch crashing in tokenizer loading." + }, + { + "left": "issue:43818", + "right": "issue:46402", + "accept": false, + "reason": "Different models and failures: Video-LLaVA weight/temporal-attention issue vs Sam3Model text-encoder weights not loading." + }, + { + "left": "issue:43377", + "right": "issue:45491", + "accept": false, + "reason": "Both are batching/padding problems, but the specific models and failure modes are different." + }, + { + "left": "issue:43818", + "right": "issue:45357", + "accept": false, + "reason": "Unrelated model families: Video-LLaVA architecture bug vs Qwen3.5 visual-encoder save_pretrained regression." + }, + { + "left": "issue:43611", + "right": "issue:43646", + "accept": false, + "reason": "Both are v5 initialization regressions, but not the same concrete bug or code path." + }, + { + "left": "issue:43644", + "right": "issue:44079", + "accept": false, + "reason": "Different bugs: non-persistent buffer junk vs ModelOutput key assignment." + }, + { + "left": "issue:44265", + "right": "issue:45072", + "accept": false, + "reason": "Unrelated: torch.export with torch_compilable_check vs bfloat16 dtype mismatch in inference." + }, + { + "left": "issue:43749", + "right": "issue:44079", + "accept": false, + "reason": "Different areas: FSDP CPU RAM efficient loading vs ModelOutput key assignment." + }, + { + "left": "issue:42175", + "right": "issue:43296", + "accept": false, + "reason": "Both mention backend/model loading, but one is TF packaging and the other is PaddleOCR-VL loading in vLLM." + }, + { + "left": "issue:40444", + "right": "issue:45381", + "accept": false, + "reason": "Both involve Qwen2.5-VL multimodal inputs, but one is finetuning with multiple images per prompt and the other is a video-position-id regression." + }, + { + "left": "issue:45325", + "right": "issue:45381", + "accept": true, + "reason": "Same underlying Qwen2.5-VL positional-indexing bug: both describe incorrect vision/temporal position_ids in the same multimodal code path." + }, + { + "left": "issue:43994", + "right": "issue:44898", + "accept": false, + "reason": "Different vision-model failures: SigLIP2 nonsensical outputs vs Perceiver resize/interpolate_pos_encoding behavior." + }, + { + "left": "issue:43792", + "right": "issue:43957", + "accept": false, + "reason": "Different causes: Whisper model runtime failure vs meta-device model loading regressions." + }, + { + "left": "issue:43454", + "right": "issue:44466", + "accept": false, + "reason": "Related to lm_head weights, but one is failing to tie weights and the other is inconsistent serialization." + }, + { + "left": "issue:45325", + "right": "issue:45834", + "accept": false, + "reason": "Both are OCR/vision-adjacent, but the bugs are different: Qwen2.5-VL position ids vs Kosmos2.5 long-input indexing." + }, + { + "left": "issue:42915", + "right": "issue:43482", + "accept": false, + "reason": "Different model/load paths: Qwen3Moe FP8 config vs Qwen2.5-GGUF loading under v5." + }, + { + "left": "issue:43632", + "right": "issue:45072", + "accept": false, + "reason": "Different regressions: _is_hf_initialized handling vs dtype mismatch in bfloat16 inference." + }, + { + "left": "issue:42796", + "right": "issue:45381", + "accept": false, + "reason": "Different tokenizer/model bugs: BPE marker handling vs Qwen2.5-VL video position IDs." + }, + { + "left": "issue:43701", + "right": "issue:45325", + "accept": false, + "reason": "Checkpoint resume key mismatch is unrelated to Qwen2.5-VL rope-index scaling." + }, + { + "left": "issue:44291", + "right": "issue:44877", + "accept": false, + "reason": "Both are loading/init issues, but they affect different configs and mechanisms." + }, + { + "left": "issue:43873", + "right": "issue:45701", + "accept": false, + "reason": "Different domains: quantization/offloading behavior vs tokenization changes across versions." + }, + { + "left": "issue:43296", + "right": "issue:45800", + "accept": false, + "reason": "Unrelated version incompatibilities: PaddleOCR-VL in vLLM vs torch 2.4.1/transformers 5.8.0." + }, + { + "left": "issue:44479", + "right": "issue:45823", + "accept": false, + "reason": "Different model families and symptoms: Qwen multimodal video regression vs Gemma4 device mismatch during forward." + }, + { + "left": "issue:43606", + "right": "issue:44265", + "accept": false, + "reason": "Different failures: CPU offload device mismatch vs torch.export/torch_compilable_check." + } + ] + }, + "evaluator_result": { + "accept": true, + "feedback": "The summary stays grounded in the packet and is suitably conservative: it identifies one plausible Qwen2.5-VL position-id duplicate family while keeping the rest of the soft-similarity pairs separate. The accepted pair (45325/45381) is reasonably supported by the titles as a closely related positional-indexing regression, and there are no PR-mergeability concerns here since these are issue pairs." + }, + "evaluator_used": true, + "retried": false, + "accepted_nontrivial_soft_edge": true, + "error_kind": null, + "error_message": null + }, + { + "label": "issue", + "component_index": 1, + "component_count": 21, + "review_unit_index": 13, + "review_unit_count": 56, + "status": "reviewed", + "reason": null, + "source": "llm", + "cache_hit": false, + "model": "gpt-5.4-mini?service_tier=flex", + "cluster_id": "cluster-10000-473", + "nodes": [ + "issue:34634", + "issue:35707", + "issue:39692", + "issue:40444", + "issue:40990", + "issue:41720", + "issue:42617", + "issue:42796", + "issue:42915", + "issue:42947", + "issue:43257", + "issue:43262", + "issue:43317", + "issue:43322", + "issue:43377", + "issue:43454", + "issue:43526", + "issue:43550", + "issue:43572", + "issue:43606", + "issue:43720", + "issue:43746", + "issue:43749", + "issue:43819", + "issue:43873", + "issue:43881", + "issue:43883", + "issue:44186", + "issue:44265", + "issue:44387", + "issue:44488", + "issue:44534", + "issue:44568", + "issue:44610", + "issue:44898", + "issue:44977", + "issue:44991", + "issue:45072", + "issue:45584", + "issue:45704", + "issue:45759", + "issue:45823", + "issue:45920", + "issue:45923", + "issue:46050", + "issue:46209", + "issue:46305", + "issue:46315" + ], + "soft_pairs": [ + "issue:43262|issue:43746", + "issue:40990|issue:45584", + "issue:34634|issue:35707", + "issue:35707|issue:41720", + "issue:43881|issue:44488", + "issue:43454|issue:43550", + "issue:43606|issue:43819", + "issue:39692|issue:40444", + "issue:44186|issue:44265", + "issue:41720|issue:45823", + "issue:46305|issue:46315", + "issue:43377|issue:44610", + "issue:42796|issue:46209", + "issue:43257|issue:45923", + "issue:44568|issue:45920", + "issue:43454|issue:45072", + "issue:43317|issue:43873", + "issue:43377|issue:44898", + "issue:42796|issue:44977", + "issue:43454|issue:43720", + "issue:40990|issue:42947", + "issue:42617|issue:42915", + "issue:43322|issue:46050", + "issue:43749|issue:44534", + "issue:43572|issue:45759", + "issue:41720|issue:43322", + "issue:43526|issue:43883", + "issue:44387|issue:45704", + "issue:41720|issue:44977", + "issue:43881|issue:44991" + ], + "prepared_review_unit_hash": "94aa4e5dc92ae9f99085bf7ed583ca708a6ef2b6a18788860a33c24fa68fbc37", + "component_budget": { + "node_count": 473, + "item_count": 473, + "soft_pair_count": 1738, + "serialized_chars": 901437, + "estimated_input_tokens": 225360, + "estimated_eval_tokens": 450976 + }, + "budget": { + "node_count": 48, + "item_count": 48, + "soft_pair_count": 30, + "serialized_chars": 36793, + "estimated_input_tokens": 9199, + "estimated_eval_tokens": 18654 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": true, + "analyst_result": { + "summary": "These items are broadly similar only at a subsystem level (tokenizers, loading, quantization, offload, multimodal), but the soft pairs describe different concrete failures and should not be merged as duplicates. No PRs are present in the cluster.", + "confidence": 0.9, + "canonical_issue_reason": "issue:43873 is the best representative because it is the most general, open bug report in the set and directly centers on a broad loading/offload/quantization failure mode, which makes it a reasonable anchor for the cluster.", + "canonical_pr_reason": null, + "best_issue_reason": "issue:43873 is the most suitable issue overall since it is open, broadly scoped, and closer than the others to the recurring theme of runtime/loading behavior under quantization/offload.", + "best_pr_reason": null, + "soft_edge_verdicts": [ + { + "left": "issue:43262", + "right": "issue:43746", + "accept": false, + "reason": "Different problems: audio processor chat-template sampling-rate default vs GraniteSpeech PEFT local-checkpoint loading failure." + }, + { + "left": "issue:40990", + "right": "issue:45584", + "accept": false, + "reason": "Unrelated failures: high perplexity in language-model evaluation vs Whisper empty-transcription generation after token alignment." + }, + { + "left": "issue:34634", + "right": "issue:35707", + "accept": false, + "reason": "Bark voice preset failure is unrelated to inputs_embeds/past_key_values generation behavior." + }, + { + "left": "issue:35707", + "right": "issue:41720", + "accept": false, + "reason": "Different code paths and symptoms: progressive generation with embeddings/cache vs Qwen3 device-map cuda assert." + }, + { + "left": "issue:43881", + "right": "issue:44488", + "accept": false, + "reason": "Both are loading issues, but one is glm-4v-9b load failure and the other is a general model-loading regression for cjvt/sleng-bert; no evidence of the same bug." + }, + { + "left": "issue:43454", + "right": "issue:43550", + "accept": false, + "reason": "AyaVision weight-tying bug and Bamba torch.compile/SDPA failure are separate model-specific issues." + }, + { + "left": "issue:43606", + "right": "issue:43819", + "accept": false, + "reason": "Both involve device/load behavior, but one is Bark CPU offload mismatch and the other is DAC.from_latents vs forward mismatch due to missing STE; not the same bug." + }, + { + "left": "issue:39692", + "right": "issue:40444", + "accept": false, + "reason": "SigLIP2 docs/quantization example errors are documentation/model-mismatch issues, while Qwen2.5-VL IterableDataset multi-image finetuning fails in training data handling." + }, + { + "left": "issue:44186", + "right": "issue:44265", + "accept": false, + "reason": "LayoutLMv2 tokenizer crashes on NER/padding versus torch.export failures with torch_compilable_check; different components and failure modes." + }, + { + "left": "issue:41720", + "right": "issue:45823", + "accept": false, + "reason": "Both are device-map related, but one is a Qwen3 cuda assert on A800 and the other is Gemma4 PLE forward device mismatch; not the same underlying bug." + }, + { + "left": "issue:46305", + "right": "issue:46315", + "accept": false, + "reason": "MoonshineStreaming off-by-one in sliding-window masking is unrelated to convert_tokens_to_ids performance regression." + }, + { + "left": "issue:43377", + "right": "issue:44610", + "accept": false, + "reason": "MIMI batching/padding-mask inconsistency is unrelated to OmDet-Turbo processor/model input-size mismatch." + }, + { + "left": "issue:42796", + "right": "issue:46209", + "accept": false, + "reason": "Tokenizer BPE-marker handling and FP8 B200 inference support are distinct problems in different layers." + }, + { + "left": "issue:43257", + "right": "issue:45923", + "accept": false, + "reason": "Qwen3 MoE loading/conversion with accelerate+deepspeed is unrelated to a missing gradient_checkpointing flag on a Nemotron trust_remote_code variant." + }, + { + "left": "issue:44568", + "right": "issue:45920", + "accept": false, + "reason": "mdeberta special-token insertion regression and broad AutoTokenizer wrong token IDs regression are both tokenizer issues but not the same concrete bug." + }, + { + "left": "issue:43454", + "right": "issue:45072", + "accept": false, + "reason": "AyaVision lm_head tying and bfloat16 dtype mismatches in SwitchTransformers/TimmWrapperModel are unrelated." + }, + { + "left": "issue:43317", + "right": "issue:43873", + "accept": false, + "reason": "Device-map dequantized-model loading on gpu+cpu offload overlaps only superficially with quantization offloading; the concrete failure described is different." + }, + { + "left": "issue:43377", + "right": "issue:44898", + "accept": false, + "reason": "MIMI encoder batching bug and Perceiver non-default resolution failure concern different models and tensor-shape logic." + }, + { + "left": "issue:42796", + "right": "issue:44977", + "accept": false, + "reason": "Tokenizer BPE marker handling is unrelated to Qwen3.5 flash-attention generation failure." + }, + { + "left": "issue:43454", + "right": "issue:43720", + "accept": false, + "reason": "AyaVision weight tying and BitNet packed-weight unpacking during accelerate loading are different model-loading defects." + }, + { + "left": "issue:40990", + "right": "issue:42947", + "accept": false, + "reason": "High perplexity evaluation issue is unrelated to gradient checkpointing being ineffective with PEFT LoRA." + }, + { + "left": "issue:42617", + "right": "issue:42915", + "accept": false, + "reason": "Failure to run 3d_parallel.py and Qwen3Moe FineGrainedFP8Config failure are unrelated." + }, + { + "left": "issue:43322", + "right": "issue:46050", + "accept": false, + "reason": "Llava Next segmentation fault and sinq quantized save/load breakage are different model/runtime bugs." + }, + { + "left": "issue:43749", + "right": "issue:44534", + "accept": false, + "reason": "FSDP CPU RAM efficient loading and non-persistent buffers filled with junk in Transformers v5 are separate core behaviors." + }, + { + "left": "issue:43572", + "right": "issue:45759", + "accept": false, + "reason": "StableLmConfig missing pad_token_idx and Qwen composite config text_config unwrapping are different configuration regressions." + }, + { + "left": "issue:41720", + "right": "issue:43322", + "accept": false, + "reason": "Qwen3 auto-device-map cuda assert is not the same as Llava Next segmentation fault during loading." + }, + { + "left": "issue:43526", + "right": "issue:43883", + "accept": false, + "reason": "BeitImageProcessorFast reduce_labels bug is unrelated to Molmo missing all_tied_weights_keys attribute." + }, + { + "left": "issue:44387", + "right": "issue:45704", + "accept": false, + "reason": "Int4 quantization CUDA reserved-memory OOM and T5 apex FusedRMSNorm memory leak are different performance/memory issues." + }, + { + "left": "issue:41720", + "right": "issue:44977", + "accept": false, + "reason": "A800 device-map cuda assert and Qwen3.5 flash-attention generation failure are separate runtime paths." + }, + { + "left": "issue:43881", + "right": "issue:44991", + "accept": false, + "reason": "glm-4v-9b loading failure is unrelated to tokenizer loading regression for EMBEDDIA/est-roberta." + } + ] + }, + "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": 21, + "review_unit_index": 14, + "review_unit_count": 56, + "status": "reviewed", + "reason": null, + "source": "llm", + "cache_hit": false, + "model": "gpt-5.4-mini?service_tier=flex", + "cluster_id": "cluster-10000-473", + "nodes": [ + "issue:26413", + "issue:28530", + "issue:33666", + "issue:36010", + "issue:36331", + "issue:40990", + "issue:41720", + "issue:42617", + "issue:42915", + "issue:43065", + "issue:43295", + "issue:43317", + "issue:43404", + "issue:43450", + "issue:43526", + "issue:43531", + "issue:43550", + "issue:43572", + "issue:43577", + "issue:43606", + "issue:43646", + "issue:43708", + "issue:43720", + "issue:43873", + "issue:43881", + "issue:44008", + "issue:44186", + "issue:44361", + "issue:44423", + "issue:44466", + "issue:44610", + "issue:44671", + "issue:44743", + "issue:44756", + "issue:44877", + "issue:44898", + "issue:44991", + "issue:45072", + "issue:45161", + "issue:45357", + "issue:45406", + "issue:45632", + "issue:45684", + "issue:45701", + "issue:45823", + "issue:46291", + "issue:46402", + "issue:46665" + ], + "soft_pairs": [ + "issue:36331|issue:42617", + "issue:44466|issue:45357", + "issue:26413|issue:28530", + "issue:36331|issue:42915", + "issue:44877|issue:44991", + "issue:43450|issue:45701", + "issue:40990|issue:45701", + "issue:44186|issue:45072", + "issue:43065|issue:46402", + "issue:43873|issue:44756", + "issue:43550|issue:43720", + "issue:41720|issue:42617", + "issue:43577|issue:43708", + "issue:43646|issue:46665", + "issue:43404|issue:45357", + "issue:43550|issue:43606", + "issue:43295|issue:45701", + "issue:44008|issue:45823", + "issue:44186|issue:44610", + "issue:44610|issue:44898", + "issue:42915|issue:43317", + "issue:44423|issue:45406", + "issue:36010|issue:42617", + "issue:45632|issue:45684", + "issue:41720|issue:45161", + "issue:42915|issue:46291", + "issue:28530|issue:33666", + "issue:43531|issue:43572", + "issue:44671|issue:44743", + "issue:36331|issue:41720", + "issue:44186|issue:44361", + "issue:43526|issue:43881" + ], + "prepared_review_unit_hash": "5e1851d74beccba98bd40af006f32a5f195f76ae1559a47aa9fb220b89a93805", + "component_budget": { + "node_count": 473, + "item_count": 473, + "soft_pair_count": 1738, + "serialized_chars": 901437, + "estimated_input_tokens": 225360, + "estimated_eval_tokens": 450976 + }, + "budget": { + "node_count": 48, + "item_count": 48, + "soft_pair_count": 32, + "serialized_chars": 37364, + "estimated_input_tokens": 9341, + "estimated_eval_tokens": 18938 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": true, + "analyst_result": { + "summary": "These issues are mostly unrelated and only share broad subsystem themes (trainer, loading, tokenizer, multimodal, quantization). I did not find any soft pair that looks like the same underlying bug or change, so all soft edges are rejected.", + "confidence": 0.88, + "canonical_issue_reason": "issue:43708 is the strongest standalone representative: it describes a specific Trainer resume_from_checkpoint regression with a clear repro and a concrete code path to fix.", + "canonical_pr_reason": null, + "best_issue_reason": "issue:43708 is the best overall issue candidate because it is detailed, reproducible, and narrowly scoped to one Trainer bug.", + "best_pr_reason": null, + "soft_edge_verdicts": [ + { + "left": "issue:36331", + "right": "issue:42617", + "accept": false, + "reason": "Different failures: trainer custom loss signature vs a 3d_parallel execution problem; only the general training context overlaps." + }, + { + "left": "issue:44466", + "right": "issue:45357", + "accept": false, + "reason": "Both concern serialization, but one is lm_head tying/device-dependent save behavior and the other is incorrect visual encoder keys in Qwen3.5." + }, + { + "left": "issue:26413", + "right": "issue:28530", + "accept": false, + "reason": "resume_from_checkpoint epoch iterator failure and early stopping metric resolution are separate Trainer bugs." + }, + { + "left": "issue:36331", + "right": "issue:42915", + "accept": false, + "reason": "CustomTrainer loss kwargs and Qwen3Moe FineGrainedFP8Config failure are unrelated code paths." + }, + { + "left": "issue:44877", + "right": "issue:44991", + "accept": false, + "reason": "Strict config loading for granite_speech is unrelated to tokenizer loading failure for EMBEDDIA/est-roberta." + }, + { + "left": "issue:43450", + "right": "issue:45701", + "accept": false, + "reason": "Video processor batch-shape bug is not the same as version-dependent tokenization changes." + }, + { + "left": "issue:40990", + "right": "issue:45701", + "accept": false, + "reason": "High perplexity on gpt-oss-20b is a model evaluation symptom, not the same underlying tokenizer-version regression." + }, + { + "left": "issue:44186", + "right": "issue:45072", + "accept": false, + "reason": "LayoutLMv2 tokenizer padding/NER crash and bfloat16 dtype mismatches are different bug classes." + }, + { + "left": "issue:43065", + "right": "issue:46402", + "accept": false, + "reason": "Sam3 PixelDecoder dummy conv and Sam3 text encoder weight loading are separate model issues." + }, + { + "left": "issue:43873", + "right": "issue:44756", + "accept": false, + "reason": "Quantization offloading behavior and mmap OOM on Strix Halo are not the same underlying defect." + }, + { + "left": "issue:43550", + "right": "issue:43720", + "accept": false, + "reason": "torch.compile with SDPA on Bamba-9B-v2 is unrelated to BitNet packed-weight loading in accelerate." + }, + { + "left": "issue:41720", + "right": "issue:42617", + "accept": false, + "reason": "Qwen3 device-map cuda assert and 3d_parallel execution failure do not describe the same bug." + }, + { + "left": "issue:43577", + "right": "issue:43708", + "accept": false, + "reason": "BLIP2 dtype propagation and Trainer max_steps on resume are unrelated." + }, + { + "left": "issue:43646", + "right": "issue:46665", + "accept": false, + "reason": "Custom model initialization breakage and from_pretrained behavior inside classes are too different to merge." + }, + { + "left": "issue:43404", + "right": "issue:45357", + "accept": false, + "reason": "lm_head weight tying in Mistral3ForConditionalGeneration is not the same as saving incorrect visual encoder keys." + }, + { + "left": "issue:43550", + "right": "issue:43606", + "accept": false, + "reason": "Bamba SDPA compile failure and Bark CPU offload device mismatch affect different models and code paths." + }, + { + "left": "issue:43295", + "right": "issue:45701", + "accept": false, + "reason": "Processor.tokenizer regression and tokenization changes across versions are not the same underlying issue." + }, + { + "left": "issue:44008", + "right": "issue:45823", + "accept": false, + "reason": "Gemma3n variable-name collision in forward and Gemma4 PLE device mismatch are distinct forward-path bugs." + }, + { + "left": "issue:44186", + "right": "issue:44610", + "accept": false, + "reason": "Tokenizer crash on batched NER inputs and OmDet-Turbo image-size mismatch are different failures." + }, + { + "left": "issue:44610", + "right": "issue:44898", + "accept": false, + "reason": "Processor producing the wrong resolution and Perceiver interpolation failing are not the same bug." + }, + { + "left": "issue:42915", + "right": "issue:43317", + "accept": false, + "reason": "FineGrainedFP8Config failure for Qwen3Moe is unrelated to device_map auto loading of dequantized models." + }, + { + "left": "issue:44423", + "right": "issue:45406", + "accept": false, + "reason": "Continuous-batching multimodal crash with str.to() and Gemma4Processor missing _tokenizer are separate serve issues." + }, + { + "left": "issue:36010", + "right": "issue:42617", + "accept": false, + "reason": "GenerationMixin import error and 3d_parallel runtime failure are unrelated." + }, + { + "left": "issue:45632", + "right": "issue:45684", + "accept": false, + "reason": "Cache-path collisions for trust_remote_code and read-only permissions propagation in save_pretrained are unrelated filesystem concerns." + }, + { + "left": "issue:41720", + "right": "issue:45161", + "accept": false, + "reason": "Auto device-mapping cuda assert and tensor-parallel-only GPT-OSS MoE failure are different distributed inference issues." + }, + { + "left": "issue:42915", + "right": "issue:46291", + "accept": false, + "reason": "Qwen3Moe FP8 failure and kernels v0.15 pipeline import failure do not share the same code path." + }, + { + "left": "issue:28530", + "right": "issue:33666", + "accept": false, + "reason": "Early stopping metric handling and Qwen2-VL multi-GPU training are unrelated trainer features." + }, + { + "left": "issue:43531", + "right": "issue:43572", + "accept": false, + "reason": "Qwen3-MoE sliding window behavior and StableLmConfig pad_token_idx are separate model/config issues." + }, + { + "left": "issue:44671", + "right": "issue:44743", + "accept": false, + "reason": "CamemBERT masked LM predictions and Qwen3.5 recurrent-state reset behavior are different model-specific bugs." + }, + { + "left": "issue:36331", + "right": "issue:41720", + "accept": false, + "reason": "Unexpected num_items_in_batch in compute_loss and Qwen3 auto device-map cuda assert are unrelated." + }, + { + "left": "issue:44186", + "right": "issue:44361", + "accept": false, + "reason": "Two tokenizer crashes, but on different models and different task/padding paths; not the same bug." + }, + { + "left": "issue:43526", + "right": "issue:43881", + "accept": false, + "reason": "BeitImageProcessorFast reduce_labels bug and glm-4v-9b 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": 21, + "review_unit_index": 15, + "review_unit_count": 56, + "status": "reviewed", + "reason": null, + "source": "llm", + "cache_hit": false, + "model": "gpt-5.4-mini?service_tier=flex", + "cluster_id": "cluster-10000-473", + "nodes": [ + "issue:28530", + "issue:29159", + "issue:30064", + "issue:33357", + "issue:34567", + "issue:34634", + "issue:35707", + "issue:36032", + "issue:41720", + "issue:42490", + "issue:42915", + "issue:43316", + "issue:43322", + "issue:43329", + "issue:43388", + "issue:43531", + "issue:43540", + "issue:43550", + "issue:43606", + "issue:43644", + "issue:43653", + "issue:43720", + "issue:43746", + "issue:43782", + "issue:43873", + "issue:44077", + "issue:44387", + "issue:44451", + "issue:44492", + "issue:44509", + "issue:44534", + "issue:44589", + "issue:44625", + "issue:44898", + "issue:44913", + "issue:44991", + "issue:45059", + "issue:45072", + "issue:45357", + "issue:45406", + "issue:45464", + "issue:45479", + "issue:45701", + "issue:45704", + "issue:45759", + "issue:45859", + "issue:46402", + "issue:46665" + ], + "soft_pairs": [ + "issue:42915|issue:43322", + "issue:43316|issue:44913", + "issue:28530|issue:30064", + "issue:43540|issue:43782", + "issue:33357|issue:35707", + "issue:43873|issue:45704", + "issue:43550|issue:43653", + "issue:43606|issue:43720", + "issue:30064|issue:34634", + "issue:29159|issue:36032", + "issue:44387|issue:45701", + "issue:44534|issue:46665", + "issue:44492|issue:44509", + "issue:43388|issue:45059", + "issue:41720|issue:43746", + "issue:43531|issue:45759", + "issue:42490|issue:43329", + "issue:43606|issue:43653", + "issue:45406|issue:45464", + "issue:34634|issue:41720", + "issue:34567|issue:35707", + "issue:45357|issue:46402", + "issue:44451|issue:44991", + "issue:43531|issue:44589", + "issue:45759|issue:45859", + "issue:44625|issue:45479", + "issue:44077|issue:44625", + "issue:43644|issue:46665", + "issue:44898|issue:45072" + ], + "prepared_review_unit_hash": "8e0080f59356c192b779f97c0005e7c708cd335bcca361ee552156221ba9a3fd", + "component_budget": { + "node_count": 473, + "item_count": 473, + "soft_pair_count": 1738, + "serialized_chars": 901437, + "estimated_input_tokens": 225360, + "estimated_eval_tokens": 450976 + }, + "budget": { + "node_count": 48, + "item_count": 48, + "soft_pair_count": 29, + "serialized_chars": 36450, + "estimated_input_tokens": 9113, + "estimated_eval_tokens": 18482 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": true, + "analyst_result": { + "summary": "The cluster is heterogeneous: the soft pairs span unrelated tokenizer, multimodal, config, quantization, loading, and docs issues. I would not merge any of the proposed soft edges as duplicates.", + "confidence": 0.84, + "canonical_issue_reason": "If a representative issue is needed, `issue:45701` is the broadest regression-style report with the most external references, but there is no true canonical duplicate for the cluster as a whole.", + "canonical_pr_reason": null, + "best_issue_reason": "`issue:45701` is the best overall representative among these issues because it is broad, user-facing, and comparatively well-connected, though it is still not a duplicate of the others.", + "best_pr_reason": null, + "soft_edge_verdicts": [ + { + "left": "issue:42915", + "right": "issue:43322", + "accept": false, + "reason": "Different failures: Qwen3Moe FineGrainedFP8Config vs Llava Next model load segfault." + }, + { + "left": "issue:43316", + "right": "issue:44913", + "accept": false, + "reason": "Both are config-related, but one is a Gemma3TextConfig API discrepancy and the other is GPTNeoX rotary_pct persistence on reload." + }, + { + "left": "issue:28530", + "right": "issue:30064", + "accept": false, + "reason": "Early-stopping metric selection and void segmentation-map processing are unrelated code paths." + }, + { + "left": "issue:43540", + "right": "issue:43782", + "accept": false, + "reason": "Different Qwen3 multimodal bugs: video input processing vs from_pretrained weight_only loading." + }, + { + "left": "issue:33357", + "right": "issue:35707", + "accept": false, + "reason": "MacOS CLIP bus error and progressive generation with inputs_embeds/past_key_values are unrelated." + }, + { + "left": "issue:43873", + "right": "issue:45704", + "accept": false, + "reason": "Quantization/offloading behavior and a T5 apex RMSNorm memory leak are distinct issues." + }, + { + "left": "issue:43550", + "right": "issue:43653", + "accept": false, + "reason": "torch.compile SDPA failure in Bamba is unrelated to BigBirdTokenizer special-token registration." + }, + { + "left": "issue:43606", + "right": "issue:43720", + "accept": false, + "reason": "One is a CPU offload device mismatch; the other is packed-weight unpacking during accelerate loading." + }, + { + "left": "issue:30064", + "right": "issue:34634", + "accept": false, + "reason": "Image-processor segmentation maps and BarkProcessor voice_preset handling do not share the same bug." + }, + { + "left": "issue:29159", + "right": "issue:36032", + "accept": false, + "reason": "Tokenizer slow/fast inconsistency is not the same as the T5 add_special_tokens name conflict." + }, + { + "left": "issue:44387", + "right": "issue:45701", + "accept": false, + "reason": "Int4 quantization memory growth and tokenization changes across versions are different regressions." + }, + { + "left": "issue:44534", + "right": "issue:46665", + "accept": false, + "reason": "Non-persistent buffer initialization junk and from_pretrained behavior inside Python classes are separate problems." + }, + { + "left": "issue:44492", + "right": "issue:44509", + "accept": false, + "reason": "A typo in cache strategies is unrelated to docs mentioning removed pipeline tasks." + }, + { + "left": "issue:43388", + "right": "issue:45059", + "accept": false, + "reason": "Metric gathering last-batch truncation and SAM3 text/bounding-box behavior are different code paths." + }, + { + "left": "issue:41720", + "right": "issue:43746", + "accept": false, + "reason": "Auto device-mapping cudaErrorAssert and PEFT local-checkpoint loading are not the same bug." + }, + { + "left": "issue:43531", + "right": "issue:45759", + "accept": false, + "reason": "Qwen3-MoE sliding_window and composite-config text_config unwrapping are distinct issues." + }, + { + "left": "issue:42490", + "right": "issue:43329", + "accept": false, + "reason": "Trailing-slash local model warning and undefined video-token helper names are unrelated." + }, + { + "left": "issue:43606", + "right": "issue:43653", + "accept": false, + "reason": "CPU offload device mismatch and tokenizer special-token registration are unrelated failures." + }, + { + "left": "issue:45406", + "right": "issue:45464", + "accept": false, + "reason": "Gemma4Processor missing _tokenizer and Qwen3.5 streaming chat/completions failure are different bugs." + }, + { + "left": "issue:34634", + "right": "issue:41720", + "accept": false, + "reason": "BarkProcessor voice_preset and Qwen3 auto device mapping failures do not overlap." + }, + { + "left": "issue:34567", + "right": "issue:35707", + "accept": false, + "reason": "TrainerState token counter update and generation with inputs_embeds/past_key_values are unrelated." + }, + { + "left": "issue:45357", + "right": "issue:46402", + "accept": false, + "reason": "Qwen3.5 visual-encoder save_pretrained keys and Sam3Model text-encoder loading are separate multimodal issues." + }, + { + "left": "issue:44451", + "right": "issue:44991", + "accept": false, + "reason": "Different model/tokenizer loading failures for ScandiBERT vs EMBEDDIA/est-roberta." + }, + { + "left": "issue:43531", + "right": "issue:44589", + "accept": false, + "reason": "Qwen3-MoE sliding-window handling and missing Float8 storage are unrelated." + }, + { + "left": "issue:45759", + "right": "issue:45859", + "accept": false, + "reason": "from_config text_config unwrapping and CPU-offload rotary embedding compatibility are not the same issue." + }, + { + "left": "issue:44625", + "right": "issue:45479", + "accept": false, + "reason": "num_labels propagation and single-label zero-loss classification are different bugs." + }, + { + "left": "issue:44077", + "right": "issue:44625", + "accept": false, + "reason": "Optional post_init policy for patchtsmixer is unrelated to Qwen3.5 num_labels propagation." + }, + { + "left": "issue:43644", + "right": "issue:46665", + "accept": false, + "reason": "Non-persistent buffer junk in Transformers 5.0 and from_pretrained behavior inside Python classes are distinct." + }, + { + "left": "issue:44898", + "right": "issue:45072", + "accept": false, + "reason": "Perceiver resolution handling and bfloat16 dtype mismatches affect different model 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": 21, + "review_unit_index": 16, + "review_unit_count": 56, + "status": "reviewed", + "reason": null, + "source": "llm", + "cache_hit": false, + "model": "gpt-5.4-mini?service_tier=flex", + "cluster_id": "cluster-10000-473", + "nodes": [ + "issue:30064", + "issue:36010", + "issue:39692", + "issue:41720", + "issue:42886", + "issue:43295", + "issue:43381", + "issue:43421", + "issue:43531", + "issue:43540", + "issue:43596", + "issue:43645", + "issue:43653", + "issue:43720", + "issue:43746", + "issue:43782", + "issue:43819", + "issue:43844", + "issue:43873", + "issue:43927", + "issue:43994", + "issue:44062", + "issue:44186", + "issue:44220", + "issue:44368", + "issue:44451", + "issue:44514", + "issue:44545", + "issue:44610", + "issue:44671", + "issue:44843", + "issue:44918", + "issue:44936", + "issue:44987", + "issue:45042", + "issue:45081", + "issue:45356", + "issue:45357", + "issue:45381", + "issue:45704", + "issue:45759", + "issue:45834", + "issue:45907", + "issue:45923", + "issue:46032", + "issue:46050", + "issue:46294", + "issue:46463" + ], + "soft_pairs": [ + "issue:44062|issue:44451", + "issue:44368|issue:45357", + "issue:44987|issue:45042", + "issue:43381|issue:43844", + "issue:39692|issue:44610", + "issue:44514|issue:44545", + "issue:44918|issue:44936", + "issue:43421|issue:43927", + "issue:45704|issue:45923", + "issue:43746|issue:46050", + "issue:41720|issue:45704", + "issue:43381|issue:45704", + "issue:43531|issue:43645", + "issue:42886|issue:44843", + "issue:46032|issue:46463", + "issue:45759|issue:45923", + "issue:43782|issue:45381", + "issue:43421|issue:43653", + "issue:39692|issue:43873", + "issue:43994|issue:44671", + "issue:43746|issue:43873", + "issue:43421|issue:44186", + "issue:30064|issue:36010", + "issue:45081|issue:45356", + "issue:43720|issue:43819", + "issue:43653|issue:43720", + "issue:45356|issue:45907", + "issue:43540|issue:46294", + "issue:43596|issue:44220", + "issue:43295|issue:45834" + ], + "prepared_review_unit_hash": "76a32d9a1ea60f3eab90c5cc8ecbe65f0d936fc959840a2128691ebb9795be45", + "component_budget": { + "node_count": 473, + "item_count": 473, + "soft_pair_count": 1738, + "serialized_chars": 901437, + "estimated_input_tokens": 225360, + "estimated_eval_tokens": 450976 + }, + "budget": { + "node_count": 48, + "item_count": 48, + "soft_pair_count": 30, + "serialized_chars": 37051, + "estimated_input_tokens": 9263, + "estimated_eval_tokens": 18782 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": true, + "analyst_result": { + "summary": "This cluster is highly heterogeneous: almost all pairs are different bugs in different models or subsystems. The only clear duplicate is the pair of Qwen2_5_VLProcessor.apply_chat_template reports with identical failure mode and wording.", + "confidence": 0.93, + "canonical_issue_reason": "No single canonical issue stands out because the cluster is mostly unrelated reports; the only true duplicate-looking pair is issue:44514/issue:44545.", + "canonical_pr_reason": "No PRs are present in this cluster.", + "best_issue_reason": "If a representative issue is needed, issue:44545 is the clearest duplicate target because it matches issue:44514 exactly and has the most directly stated failure scenario.", + "best_pr_reason": "No PRs are present, so there is no best PR candidate.", + "soft_edge_verdicts": [ + { + "left": "issue:44062", + "right": "issue:44451", + "accept": false, + "reason": "Both are tokenizer/model-loading failures, but the concrete errors and affected code paths differ." + }, + { + "left": "issue:44368", + "right": "issue:45357", + "accept": false, + "reason": "One is a tie_word_embeddings warning in fine-tuning; the other is incorrect visual encoder key saving. Different bugs." + }, + { + "left": "issue:44987", + "right": "issue:45042", + "accept": false, + "reason": "Loading a model repo and a PIL/torchvision backend requirement regression are unrelated." + }, + { + "left": "issue:43381", + "right": "issue:43844", + "accept": false, + "reason": "Gradient checkpointing in eval mode is a different problem from gradient explosion under ZeRO-3." + }, + { + "left": "issue:39692", + "right": "issue:44610", + "accept": false, + "reason": "Different multimodal examples and failures; no shared concrete bug." + }, + { + "left": "issue:44514", + "right": "issue:44545", + "accept": true, + "reason": "Same processor, same batched padding=False crash, same title and failure mode; clear duplicate." + }, + { + "left": "issue:44918", + "right": "issue:44936", + "accept": false, + "reason": "Both involve training/eval behavior, but they describe different failures and code paths." + }, + { + "left": "issue:43421", + "right": "issue:43927", + "accept": false, + "reason": "Runtime special-token post-processing and saved config token IDs are distinct tokenizer issues." + }, + { + "left": "issue:45704", + "right": "issue:45923", + "accept": false, + "reason": "Apex RMSNorm memory leak and missing gradient-checkpointing flag are unrelated." + }, + { + "left": "issue:43746", + "right": "issue:46050", + "accept": false, + "reason": "PEFT local checkpoint loading and quantized save/load breakage are different persistence bugs." + }, + { + "left": "issue:41720", + "right": "issue:45704", + "accept": false, + "reason": "Qwen3 device-mapping CUDA assert has no overlap with T5 RMSNorm memory leak." + }, + { + "left": "issue:43381", + "right": "issue:45704", + "accept": false, + "reason": "Different model families and different failure mechanisms." + }, + { + "left": "issue:43531", + "right": "issue:43645", + "accept": false, + "reason": "Sliding-window behavior in Qwen3-MoE is unrelated to Jupyter custom-model initialization." + }, + { + "left": "issue:42886", + "right": "issue:44843", + "accept": false, + "reason": "Both mention offline tokenizer loading, but one is a generic cache-load failure and the other is a specific _patch_mistral_regex/model_info regression." + }, + { + "left": "issue:46032", + "right": "issue:46463", + "accept": false, + "reason": "Incorrect use_cache results and an error-message bug in T5 are unrelated." + }, + { + "left": "issue:45759", + "right": "issue:45923", + "accept": false, + "reason": "Composite config unwrapping and a missing gradient-checkpointing flag are separate issues." + }, + { + "left": "issue:43782", + "right": "issue:45381", + "accept": false, + "reason": "Different Qwen model families and different multimodal failures." + }, + { + "left": "issue:43421", + "right": "issue:43653", + "accept": false, + "reason": "Both involve tokenizer special tokens, but one is about updating post-processors at runtime and the other is a BigBird-specific mask-token registration bug." + }, + { + "left": "issue:39692", + "right": "issue:43873", + "accept": false, + "reason": "A SigLIP2 documentation example bug is unrelated to quantization offloading." + }, + { + "left": "issue:43994", + "right": "issue:44671", + "accept": false, + "reason": "Incorrect outputs on two different models are not the same underlying bug." + }, + { + "left": "issue:43746", + "right": "issue:43873", + "accept": false, + "reason": "Checkpoint loading with PEFT and quantization offloading are different code paths." + }, + { + "left": "issue:43421", + "right": "issue:44186", + "accept": false, + "reason": "Runtime tokenizer post-processor updates do not match LayoutLMv2 NER/padding crashes." + }, + { + "left": "issue:30064", + "right": "issue:36010", + "accept": false, + "reason": "Image segmentation map processing and an ImportError for GenerationMixin are unrelated." + }, + { + "left": "issue:45081", + "right": "issue:45356", + "accept": false, + "reason": "Both are in the Mistral-regex/tokenizer area, but one is an AttributeError crash and the other is a Kimi-K2.5 regression about codec handling; not enough to treat as the same bug." + }, + { + "left": "issue:43720", + "right": "issue:43819", + "accept": false, + "reason": "Packed-weight loading and DAC latent reconstruction are unrelated." + }, + { + "left": "issue:43653", + "right": "issue:43720", + "accept": false, + "reason": "BigBird mask-token registration and BitNet packed-weight unpacking are different issues." + }, + { + "left": "issue:45356", + "right": "issue:45907", + "accept": false, + "reason": "Tokenizer regression and a hub API forwarding bug are unrelated." + }, + { + "left": "issue:43540", + "right": "issue:46294", + "accept": false, + "reason": "Both concern video inputs, but they affect different Qwen model families and describe different processor/template mismatches." + }, + { + "left": "issue:43596", + "right": "issue:44220", + "accept": false, + "reason": "DeepSpeed ZeRO-3/BertModel initialization and audio feature extraction are unrelated." + }, + { + "left": "issue:43295", + "right": "issue:45834", + "accept": false, + "reason": "Processor.tokenizer regression and Kosmos2.5 OCR index errors are different bugs." + } + ] + }, + "evaluator_result": { + "accept": true, + "feedback": "The cluster summary is well grounded in the packet: it correctly identifies the single obvious duplicate pair (issue:44514 / issue:44545) and treats the remaining soft pairs conservatively as non-duplicates. No PRs are present, so the PR-related fields are also consistent." + }, + "evaluator_used": true, + "retried": false, + "accepted_nontrivial_soft_edge": true, + "error_kind": null, + "error_message": null + }, + { + "label": "issue", + "component_index": 1, + "component_count": 21, + "review_unit_index": 17, + "review_unit_count": 56, + "status": "reviewed", + "reason": null, + "source": "llm", + "cache_hit": false, + "model": "gpt-5.4-mini?service_tier=flex", + "cluster_id": "cluster-10000-473", + "nodes": [ + "issue:30064", + "issue:36331", + "issue:39692", + "issue:41720", + "issue:43322", + "issue:43335", + "issue:43388", + "issue:43531", + "issue:43540", + "issue:43550", + "issue:43606", + "issue:43611", + "issue:43618", + "issue:43643", + "issue:43644", + "issue:43653", + "issue:43697", + "issue:43716", + "issue:43749", + "issue:43756", + "issue:43819", + "issue:43927", + "issue:43950", + "issue:44008", + "issue:44062", + "issue:44186", + "issue:44220", + "issue:44442", + "issue:44448", + "issue:44488", + "issue:44661", + "issue:44671", + "issue:44821", + "issue:44843", + "issue:44877", + "issue:44898", + "issue:44991", + "issue:45072", + "issue:45081", + "issue:45084", + "issue:45183", + "issue:45479", + "issue:45584", + "issue:45685", + "issue:45834", + "issue:45987", + "issue:46032", + "issue:46665" + ], + "soft_pairs": [ + "issue:43550|issue:45072", + "issue:43550|issue:43927", + "issue:43618|issue:43697", + "issue:44062|issue:44488", + "issue:43611|issue:46665", + "issue:44220|issue:44661", + "issue:44186|issue:44898", + "issue:43653|issue:43819", + "issue:43531|issue:43643", + "issue:43335|issue:43756", + "issue:43540|issue:44008", + "issue:43550|issue:43819", + "issue:30064|issue:36331", + "issue:43643|issue:44877", + "issue:41720|issue:44991", + "issue:44843|issue:45081", + "issue:44821|issue:45084", + "issue:39692|issue:45834", + "issue:43322|issue:45685", + "issue:44186|issue:44442", + "issue:43716|issue:44008", + "issue:44448|issue:44671", + "issue:43388|issue:46032", + "issue:45584|issue:45834", + "issue:43644|issue:43749", + "issue:45183|issue:45479", + "issue:43950|issue:45987", + "issue:43606|issue:45072" + ], + "prepared_review_unit_hash": "ddeea3d644ee6816b811ae698227f9e95bc42f69ebad7d671d95938f51d36d62", + "component_budget": { + "node_count": 473, + "item_count": 473, + "soft_pair_count": 1738, + "serialized_chars": 901437, + "estimated_input_tokens": 225360, + "estimated_eval_tokens": 450976 + }, + "budget": { + "node_count": 48, + "item_count": 48, + "soft_pair_count": 28, + "serialized_chars": 36312, + "estimated_input_tokens": 9078, + "estimated_eval_tokens": 18412 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": true, + "analyst_result": { + "summary": "This cluster is mostly a noisy mix of unrelated model-loading, tokenizer, and runtime regressions. The soft pairs share broad subsystem themes, but none look like the same concrete bug/change, so I would not merge any of them as duplicates.", + "confidence": 0.88, + "canonical_issue_reason": "issue:43611 is the best anchor because it is a broad, core loading regression in Transformers 5.0 and is more representative of the cluster\u2019s repeated from_pretrained/model-init complaints than the narrower, model-specific reports.", + "canonical_pr_reason": null, + "best_issue_reason": "issue:43611 is the strongest single issue to keep as the cluster representative: it is clearly scoped, impacts a common code path, and is more general than the highly model-specific failures here.", + "best_pr_reason": null, + "soft_edge_verdicts": [ + { + "left": "issue:43550", + "right": "issue:45072", + "accept": false, + "reason": "Different failures: torch.compile + SDPA in Bamba versus dtype mismatches in SwitchTransformers/TimmWrapper. Same broad runtime category, not the same bug." + }, + { + "left": "issue:43550", + "right": "issue:43927", + "accept": false, + "reason": "Bamba compile/SDPA crash and DiaConfig save/load token-ID loss are unrelated code paths and symptoms." + }, + { + "left": "issue:43618", + "right": "issue:43697", + "accept": false, + "reason": "CLIPOutput attentions regression and RTDetrV2 output drift are separate output-behavior issues in different models." + }, + { + "left": "issue:44062", + "right": "issue:44488", + "accept": false, + "reason": "Tokenizer AddedToken special-arg error is unrelated to loading cjvt/sleng-bert; no shared concrete bug." + }, + { + "left": "issue:43611", + "right": "issue:46665", + "accept": false, + "reason": "Both involve loading, but one is a specific base_model_prefix regression and the other is a generic from_pretrained-in-class behavior complaint." + }, + { + "left": "issue:44220", + "right": "issue:44661", + "accept": false, + "reason": "Audio fbank extraction and add-new-model-like / tokenizer mapping logic are different subsystems and failure modes." + }, + { + "left": "issue:44186", + "right": "issue:44898", + "accept": false, + "reason": "Tokenizer padding/truncation crash in LayoutLMv2 is unrelated to Perceiver image classification resolution handling." + }, + { + "left": "issue:43653", + "right": "issue:43819", + "accept": false, + "reason": "BigBirdTokenizer special-token registration and DAC.from_latents / STE mismatch are different bugs in different components." + }, + { + "left": "issue:43531", + "right": "issue:43643", + "accept": false, + "reason": "Qwen3-MoE sliding_window behavior and AutoConfig trust_remote_code field loss do not share the same underlying defect." + }, + { + "left": "issue:43335", + "right": "issue:43756", + "accept": false, + "reason": "Both are model-structure/configuration bugs, but they affect different models and different mechanisms, so they are not the same change." + }, + { + "left": "issue:43540", + "right": "issue:44008", + "accept": false, + "reason": "Video-input validation in Qwen3OmniMoe and Gemma3n variable-name collision are distinct multimodal forward-pass bugs." + }, + { + "left": "issue:43550", + "right": "issue:43819", + "accept": false, + "reason": "Bamba torch.compile/SDPA failure and DAC latents mismatch are unrelated; same broad audio/model area is not enough." + }, + { + "left": "issue:30064", + "right": "issue:36331", + "accept": false, + "reason": "Image segmentation-map processing and CustomTrainer.compute_loss signature breakage are unrelated." + }, + { + "left": "issue:43643", + "right": "issue:44877", + "accept": false, + "reason": "AutoConfig missing fields with trust_remote_code and strict config rejection of granite_speech are different config-loading problems." + }, + { + "left": "issue:41720", + "right": "issue:44991", + "accept": false, + "reason": "Qwen3 auto device mapping CUDA assert and tokenizer load failure for EMBEDDIA/est-roberta are unrelated failures." + }, + { + "left": "issue:44843", + "right": "issue:45081", + "accept": false, + "reason": "Both touch _patch_mistral_regex, but one is an offline model_info call problem and the other is an AttributeError from a different tokenizer object shape." + }, + { + "left": "issue:44821", + "right": "issue:45084", + "accept": false, + "reason": "AutoImageProcessor URL loading and template-node compilation errors are unrelated." + }, + { + "left": "issue:39692", + "right": "issue:45834", + "accept": false, + "reason": "SigLIP2 docs/example errors and Kosmos2.5 long-OCR index error are different issues with no shared concrete bug." + }, + { + "left": "issue:43322", + "right": "issue:45685", + "accept": false, + "reason": "Llava Next segfault on load and MPS histogram_mps integer error are unrelated runtime/device issues." + }, + { + "left": "issue:44186", + "right": "issue:44442", + "accept": false, + "reason": "LayoutLMv2Tokenizer NER/padding crash and FastSpeech2ConformerTokenizer auto-load failure are different tokenizer bugs." + }, + { + "left": "issue:43716", + "right": "issue:44008", + "accept": false, + "reason": "Mistral-3 image preprocessor dtype mismatch and Gemma 3n audio_mel_mask attribute collision are unrelated multimodal bugs." + }, + { + "left": "issue:44448", + "right": "issue:44671", + "accept": false, + "reason": "Both are v5 output regressions, but they affect different models and likely different code paths; not one underlying fix." + }, + { + "left": "issue:43388", + "right": "issue:46032", + "accept": false, + "reason": "Metric-gathering label truncation and Mamba2Mixer use_cache correctness are unrelated training/inference bugs." + }, + { + "left": "issue:45584", + "right": "issue:45834", + "accept": false, + "reason": "Whisper empty-transcription generation and Kosmos2.5 long-input index error are separate edge cases." + }, + { + "left": "issue:43644", + "right": "issue:43749", + "accept": false, + "reason": "Non-persistent buffer corruption during from_pretrained and FSDP CPU RAM efficient loading breakage are different regressions." + }, + { + "left": "issue:45183", + "right": "issue:45479", + "accept": false, + "reason": "Validation error messaging for input_ids/inputs_embeds and degenerate zero-loss classification behavior are not the same bug." + }, + { + "left": "issue:43950", + "right": "issue:45987", + "accept": false, + "reason": "Silent non-persistent buffer corruption and StaticCache.get_seq_length returning a shape-(1,) tensor are unrelated contract violations." + }, + { + "left": "issue:43606", + "right": "issue:45072", + "accept": false, + "reason": "CPU offload device mismatch in bark-small and bfloat16 dtype mismatches in SwitchTransformers/TimmWrapper are different problems." + } + ] + }, + "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": 21, + "review_unit_index": 18, + "review_unit_count": 56, + "status": "reviewed", + "reason": null, + "source": "llm", + "cache_hit": false, + "model": "gpt-5.4-mini?service_tier=flex", + "cluster_id": "cluster-10000-473", + "nodes": [ + "issue:28218", + "issue:29159", + "issue:34634", + "issue:36010", + "issue:36331", + "issue:42175", + "issue:43421", + "issue:43540", + "issue:43577", + "issue:43582", + "issue:43645", + "issue:43720", + "issue:43749", + "issue:43782", + "issue:43818", + "issue:43825", + "issue:43827", + "issue:43994", + "issue:44291", + "issue:44479", + "issue:44485", + "issue:44509", + "issue:44561", + "issue:44610", + "issue:44671", + "issue:44805", + "issue:44811", + "issue:44857", + "issue:44936", + "issue:44977", + "issue:44991", + "issue:45003", + "issue:45020", + "issue:45072", + "issue:45081", + "issue:45325", + "issue:45381", + "issue:45701", + "issue:45834", + "issue:45987", + "issue:46192", + "issue:46285", + "issue:46291", + "issue:46294", + "issue:46463", + "issue:46512", + "issue:46665" + ], + "soft_pairs": [ + "issue:43582|issue:46512", + "issue:44485|issue:44509", + "issue:34634|issue:36010", + "issue:44479|issue:45834", + "issue:43825|issue:43994", + "issue:43577|issue:43818", + "issue:43421|issue:44811", + "issue:43782|issue:44977", + "issue:44991|issue:46291", + "issue:46294|issue:46463", + "issue:44805|issue:44936", + "issue:43540|issue:45381", + "issue:44561|issue:45020", + "issue:44610|issue:45072", + "issue:43720|issue:45072", + "issue:28218|issue:29159", + "issue:43577|issue:43582", + "issue:43825|issue:43827", + "issue:45987|issue:46192", + "issue:43749|issue:44291", + "issue:43645|issue:46665", + "issue:44671|issue:45701", + "issue:42175|issue:45081", + "issue:34634|issue:36331", + "issue:43582|issue:44857", + "issue:45003|issue:46285", + "issue:43782|issue:45325" + ], + "prepared_review_unit_hash": "7269b861276ff9bec4523ea7cbcb7cdbd93154b9c640cad2e14a8fce356c73f5", + "component_budget": { + "node_count": 473, + "item_count": 473, + "soft_pair_count": 1738, + "serialized_chars": 901437, + "estimated_input_tokens": 225360, + "estimated_eval_tokens": 450976 + }, + "budget": { + "node_count": 47, + "item_count": 47, + "soft_pair_count": 27, + "serialized_chars": 35483, + "estimated_input_tokens": 8871, + "estimated_eval_tokens": 17998 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": true, + "analyst_result": { + "summary": "This cluster is mostly a mix of unrelated transformer bugs (tokenizers, video/VL, loading, dtype, docs, trainer). The only plausible duplicate pair is the tokenizer spacing/inconsistency reports; the rest do not look like the same underlying bug.", + "confidence": 0.74, + "canonical_issue_reason": "issue:29159 is the broader tokenizer-behavior report and can subsume the more specific added-token spacing symptom in issue:28218.", + "canonical_pr_reason": null, + "best_issue_reason": "issue:29159 is the best representative issue in the cluster because it is the clearest umbrella for the only apparent duplicate theme here; the rest of the issues are disparate.", + "best_pr_reason": null, + "soft_edge_verdicts": [ + { + "left": "issue:43582", + "right": "issue:46512", + "accept": false, + "reason": "Different failures: Apple Silicon allocator warmup TypeError vs stale config dtype after from_config." + }, + { + "left": "issue:44485", + "right": "issue:44509", + "accept": false, + "reason": "GLM-5 RoPE implementation discussion is unrelated to docs still mentioning removed pipelines." + }, + { + "left": "issue:34634", + "right": "issue:36010", + "accept": false, + "reason": "BarkProcessor voice_preset bug and a GenerationMixin import error are separate code paths." + }, + { + "left": "issue:44479", + "right": "issue:45834", + "accept": false, + "reason": "Both are vision/video-related, but they affect different models and different failure modes." + }, + { + "left": "issue:43825", + "right": "issue:43994", + "accept": false, + "reason": "One is an error-message/docs issue about removed pipeline tasks; the other is a model quality/inference bug." + }, + { + "left": "issue:43577", + "right": "issue:43818", + "accept": false, + "reason": "BLIP2 dtype loading and Video-LLaVA temporal-attention/weight-sharing are unrelated regressions." + }, + { + "left": "issue:43421", + "right": "issue:44811", + "accept": true, + "reason": "Both point to tokenizer/processor special-token handling being inconsistent; 29159 is the broader formulation of the same bug family." + }, + { + "left": "issue:43782", + "right": "issue:44977", + "accept": false, + "reason": "Qwen3VL weight_only loading error and Qwen3.5 flash-attention generation issues are different problems." + }, + { + "left": "issue:44991", + "right": "issue:46291", + "accept": false, + "reason": "Tokenizer loading for a specific model and a pipelines import failure from kernels v0.15 are different breakages." + }, + { + "left": "issue:46294", + "right": "issue:46463", + "accept": false, + "reason": "Qwen3.5 video chat-template mismatch is unrelated to T5 decoder-input validation messaging." + }, + { + "left": "issue:44805", + "right": "issue:44936", + "accept": false, + "reason": "Mask/index shape mismatch and trainer.evaluate() after train() are not the same bug." + }, + { + "left": "issue:43540", + "right": "issue:45381", + "accept": false, + "reason": "Both involve video inputs, but they target different models and likely different preprocessing/position-id logic." + }, + { + "left": "issue:44561", + "right": "issue:45020", + "accept": false, + "reason": "45020 is a broad umbrella about remote_code breakage; 44561 is a specific helper-removal regression, not clearly the same concrete bug." + }, + { + "left": "issue:44610", + "right": "issue:45072", + "accept": false, + "reason": "Processor image-size mismatch and bfloat16 dtype mismatches are unrelated." + }, + { + "left": "issue:43720", + "right": "issue:45072", + "accept": false, + "reason": "BitNet packed-weight loading and dtype mismatch inference failures are different issues." + }, + { + "left": "issue:28218", + "right": "issue:29159", + "accept": true, + "reason": "These describe the same tokenizer inconsistency around added tokens/spacing; 29159 is the broader duplicate umbrella." + }, + { + "left": "issue:43577", + "right": "issue:43582", + "accept": false, + "reason": "BLIP2 dtype handling and Apple Silicon allocator warmup TypeError are unrelated." + }, + { + "left": "issue:43825", + "right": "issue:43827", + "accept": false, + "reason": "Both mention removed pipeline tasks, but one is runtime messaging and the other is stale docs; not the same bug." + }, + { + "left": "issue:45987", + "right": "issue:46192", + "accept": false, + "reason": "StaticCache sequence-length contract and return_tensors being ignored are separate API bugs." + }, + { + "left": "issue:43749", + "right": "issue:44291", + "accept": false, + "reason": "FSDP CPU RAM efficient loading and init_empty_weights argument handling are different loading paths." + }, + { + "left": "issue:43645", + "right": "issue:46665", + "accept": false, + "reason": "Jupyter custom-model initialization and from_pretrained inside Python classes do not appear to be the same concrete issue." + }, + { + "left": "issue:44671", + "right": "issue:45701", + "accept": false, + "reason": "Masked LM prediction regression and tokenization changes are related only broadly, not the same underlying bug." + }, + { + "left": "issue:42175", + "right": "issue:45081", + "accept": false, + "reason": "Backend dependency packaging and Mistral regex patching are unrelated." + }, + { + "left": "issue:34634", + "right": "issue:36331", + "accept": false, + "reason": "BarkProcessor voice_preset and CustomTrainer.compute_loss signature breakage are unrelated." + }, + { + "left": "issue:43582", + "right": "issue:44857", + "accept": false, + "reason": "Allocator warmup on Apple Silicon and CUDA AMP loss crashes are unrelated runtime failures." + }, + { + "left": "issue:45003", + "right": "issue:46285", + "accept": false, + "reason": "Unsafe sys.modules access and CLIP text-model structure drift are different code changes." + }, + { + "left": "issue:43782", + "right": "issue:45325", + "accept": false, + "reason": "Qwen3VL weight_only loading and Qwen2.5-VL rope-index scaling are different bugs." + } + ] + }, + "evaluator_result": { + "accept": true, + "feedback": "The summary is grounded in the packet and stays conservative overall. The chosen canonical issue, issue:29159, reasonably serves as the broader umbrella for issue:28218\u2019s added-token spacing symptom, and the rejection calls on the other pairs do not overclaim shared bug identity from title-level evidence. The one accepted soft pair (issue:43421/issue:44811) is a bit broad, but still defensible as a tokenizer/processor special-token handling family given the available evidence." + }, + "evaluator_used": true, + "retried": false, + "accepted_nontrivial_soft_edge": true, + "error_kind": null, + "error_message": null + }, + { + "label": "issue", + "component_index": 1, + "component_count": 21, + "review_unit_index": 19, + "review_unit_count": 56, + "status": "reviewed", + "reason": null, + "source": "llm", + "cache_hit": false, + "model": "gpt-5.4-mini?service_tier=flex", + "cluster_id": "cluster-10000-473", + "nodes": [ + "issue:36010", + "issue:36331", + "issue:38175", + "issue:39692", + "issue:42175", + "issue:42490", + "issue:43257", + "issue:43262", + "issue:43329", + "issue:43452", + "issue:43526", + "issue:43531", + "issue:43550", + "issue:43638", + "issue:43645", + "issue:43653", + "issue:43697", + "issue:43720", + "issue:43742", + "issue:43746", + "issue:43819", + "issue:43854", + "issue:43906", + "issue:43927", + "issue:44060", + "issue:44368", + "issue:44387", + "issue:44554", + "issue:44589", + "issue:44610", + "issue:44617", + "issue:44811", + "issue:44912", + "issue:44936", + "issue:44977", + "issue:44991", + "issue:45020", + "issue:45042", + "issue:45081", + "issue:45084", + "issue:45372", + "issue:45406", + "issue:45538", + "issue:45600", + "issue:46209", + "issue:46291", + "issue:46326" + ], + "soft_pairs": [ + "issue:43653|issue:44912", + "issue:43262|issue:43819", + "issue:43526|issue:43697", + "issue:45081|issue:45406", + "issue:45372|issue:46326", + "issue:44589|issue:45042", + "issue:43329|issue:45600", + "issue:44060|issue:44368", + "issue:43742|issue:45020", + "issue:45084|issue:45538", + "issue:43526|issue:44554", + "issue:43550|issue:44912", + "issue:43906|issue:46209", + "issue:43452|issue:44991", + "issue:43638|issue:44936", + "issue:43257|issue:43329", + "issue:43819|issue:43927", + "issue:36010|issue:36331", + "issue:36010|issue:43645", + "issue:42175|issue:46291", + "issue:43531|issue:45042", + "issue:43720|issue:44610", + "issue:38175|issue:44610", + "issue:43746|issue:44811", + "issue:44387|issue:44617", + "issue:39692|issue:44912", + "issue:42490|issue:43257", + "issue:43854|issue:44977" + ], + "prepared_review_unit_hash": "2927d5b87f3b678ee75d52caf24414332d01799f95c1dea5bfeb42dcc35139f6", + "component_budget": { + "node_count": 473, + "item_count": 473, + "soft_pair_count": 1738, + "serialized_chars": 901437, + "estimated_input_tokens": 225360, + "estimated_eval_tokens": 450976 + }, + "budget": { + "node_count": 47, + "item_count": 47, + "soft_pair_count": 28, + "serialized_chars": 35879, + "estimated_input_tokens": 8970, + "estimated_eval_tokens": 18196 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": true, + "analyst_result": { + "summary": "This cluster is mostly a set of unrelated Transformers bugs that only share broad themes like loading, tokenizers, quantization, or model-family regressions. I would not merge any of the soft pairs; the best umbrella issue is the broad remote_code regression report, but even that is only a loose fit for the rest.", + "confidence": 0.87, + "canonical_issue_reason": "Issue 45020 is the broadest, most umbrella-like report in the set: it describes recent-version regressions breaking model loading via `remote_code`, which is the closest shared theme across several other reports. The rest are much more specific model/component bugs and do not form a true duplicate set.", + "canonical_pr_reason": null, + "best_issue_reason": "45020 is the best representative because it captures the highest-level regression pattern and is more general than the other issues, which are mostly narrow, model-specific failures.", + "best_pr_reason": null, + "soft_edge_verdicts": [ + { + "left": "issue:43653", + "right": "issue:44912", + "accept": false, + "reason": "BigBird tokenizer special-token decoding bug and MXFP4 quantization fallback are unrelated failures." + }, + { + "left": "issue:43262", + "right": "issue:43819", + "accept": false, + "reason": "Audio chat-template sampling-rate defaulting and DAC latent reconstruction mismatch are different code paths and symptoms." + }, + { + "left": "issue:43526", + "right": "issue:43697", + "accept": false, + "reason": "BEiT label reduction and RTDetrV2 output drift are distinct vision-model issues." + }, + { + "left": "issue:45081", + "right": "issue:45406", + "accept": false, + "reason": "Both involve tokenizer/processor loading, but one is a Mistral regex/backend_tokenizer crash and the other is a Gemma4Processor missing `_tokenizer`; different bugs." + }, + { + "left": "issue:45372", + "right": "issue:46326", + "accept": false, + "reason": "ReasoningEffort import failure and `continue_final_message` behavior are separate mistral_common-related problems, not the same underlying bug." + }, + { + "left": "issue:44589", + "right": "issue:45042", + "accept": false, + "reason": "Float8 storage lookup failure and PIL backend requiring torchvision are unrelated runtime/dependency issues." + }, + { + "left": "issue:43329", + "right": "issue:45600", + "accept": false, + "reason": "Undefined video-token helper variables and stale auto_mappings config references are different regressions." + }, + { + "left": "issue:44060", + "right": "issue:44368", + "accept": false, + "reason": "Both are warning-related on Qwen models, but one is a tied-weights warning bug and the other is a `tie_word_embeddings=False` warning during LoRA fine-tuning; not the same fix." + }, + { + "left": "issue:43742", + "right": "issue:45020", + "accept": false, + "reason": "A specific MobileLLM key error is not the same as the broader `remote_code` loading regression." + }, + { + "left": "issue:45084", + "right": "issue:45538", + "accept": false, + "reason": "Template compilation errors and CLIPTokenizer `model_max_length` behavior are unrelated." + }, + { + "left": "issue:43526", + "right": "issue:44554", + "accept": false, + "reason": "BEiT label processing and MPS attention correctness are different subsystems." + }, + { + "left": "issue:43550", + "right": "issue:44912", + "accept": false, + "reason": "Bamba SDPA/torch.compile failure and MXFP4 quantization fallback are separate issues." + }, + { + "left": "issue:43906", + "right": "issue:46209", + "accept": false, + "reason": "An isolated reproduction of a prior issue is not the same as the FP8 inference failure on B200." + }, + { + "left": "issue:43452", + "right": "issue:44991", + "accept": false, + "reason": "gguf_file handling for tokenizer/model loading and a tokenizer-loading regression for est-roberta are both loading issues, but they have different root causes and are not clearly the same bug." + }, + { + "left": "issue:43638", + "right": "issue:44936", + "accept": false, + "reason": "Zero3/non-pretrained Bert index error and `evaluate()` failing after `train()` are distinct trainer failures." + }, + { + "left": "issue:43257", + "right": "issue:43329", + "accept": false, + "reason": "Qwen3-MoE weight conversion with accelerate/deepspeed and multimodal token counting are unrelated." + }, + { + "left": "issue:43819", + "right": "issue:43927", + "accept": false, + "reason": "DAC latent forward mismatch and DiaConfig losing custom token IDs are different model/config bugs." + }, + { + "left": "issue:36010", + "right": "issue:36331", + "accept": false, + "reason": "`GenerationMixin` import failure and `compute_loss()` kwarg incompatibility are unrelated API breakages." + }, + { + "left": "issue:36010", + "right": "issue:43645", + "accept": false, + "reason": "Importing `GenerationMixin` and notebook custom-model initialization failures do not describe the same problem." + }, + { + "left": "issue:42175", + "right": "issue:46291", + "accept": false, + "reason": "Missing TensorFlow from the `.[torch]` extra and `transformers.pipelines` breaking on kernels v0.15 are separate dependency issues." + }, + { + "left": "issue:43531", + "right": "issue:45042", + "accept": false, + "reason": "Qwen3-MoE sliding-window handling and PIL image processor torchvision requirements are unrelated." + }, + { + "left": "issue:43720", + "right": "issue:44610", + "accept": false, + "reason": "BitNet packed-weight loading and OmDet-Turbo processor size mismatch are different code paths." + }, + { + "left": "issue:38175", + "right": "issue:44610", + "accept": false, + "reason": "SigLIP2 zero probabilities and OmDet-Turbo input-size mismatch are unrelated model bugs." + }, + { + "left": "issue:43746", + "right": "issue:44811", + "accept": false, + "reason": "GraniteSpeech PEFT local-checkpoint loading and Whisper batch_decode skipping special tokens are not the same issue." + }, + { + "left": "issue:44387", + "right": "issue:44617", + "accept": false, + "reason": "A Transformers 5.x int4 memory regression and Sam3Video OOM are different causes." + }, + { + "left": "issue:39692", + "right": "issue:44912", + "accept": false, + "reason": "SigLIP2 documentation example errors and MXFP4 quantization fallback are unrelated." + }, + { + "left": "issue:42490", + "right": "issue:43257", + "accept": false, + "reason": "Trailing-slash local model import warning and Qwen3 MoE weight conversion are unrelated." + }, + { + "left": "issue:43854", + "right": "issue:44977", + "accept": false, + "reason": "GLM-4.7-Flash loading in tests and Qwen3.5 flash-attention generation are different model-specific 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": 21, + "review_unit_index": 20, + "review_unit_count": 56, + "status": "reviewed", + "reason": null, + "source": "llm", + "cache_hit": false, + "model": "gpt-5.4-mini?service_tier=flex", + "cluster_id": "cluster-10000-473", + "nodes": [ + "issue:28282", + "issue:29159", + "issue:34567", + "issue:34689", + "issue:40990", + "issue:42915", + "issue:43257", + "issue:43262", + "issue:43329", + "issue:43404", + "issue:43425", + "issue:43450", + "issue:43454", + "issue:43525", + "issue:43577", + "issue:43606", + "issue:43643", + "issue:43650", + "issue:43696", + "issue:43697", + "issue:43742", + "issue:43749", + "issue:43906", + "issue:43927", + "issue:43931", + "issue:44291", + "issue:44479", + "issue:44568", + "issue:44756", + "issue:44811", + "issue:44821", + "issue:44912", + "issue:44987", + "issue:45042", + "issue:45127", + "issue:45137", + "issue:45356", + "issue:45357", + "issue:45405", + "issue:45459", + "issue:45538", + "issue:45759", + "issue:45803", + "issue:45859", + "issue:46291", + "issue:46421", + "issue:46424" + ], + "soft_pairs": [ + "issue:46421|issue:46424", + "issue:43262|issue:44811", + "issue:43643|issue:45759", + "issue:43257|issue:43696", + "issue:45459|issue:45803", + "issue:43329|issue:44479", + "issue:43650|issue:43906", + "issue:43257|issue:43931", + "issue:43329|issue:43450", + "issue:42915|issue:43742", + "issue:43606|issue:45859", + "issue:43749|issue:44756", + "issue:34567|issue:40990", + "issue:43454|issue:43577", + "issue:44291|issue:45042", + "issue:42915|issue:44821", + "issue:43525|issue:46291", + "issue:29159|issue:34567", + "issue:43927|issue:44912", + "issue:44568|issue:45356", + "issue:44987|issue:45538", + "issue:43697|issue:45357", + "issue:43404|issue:45127", + "issue:43425|issue:46291", + "issue:42915|issue:45538", + "issue:28282|issue:34689", + "issue:45137|issue:45405" + ], + "prepared_review_unit_hash": "23eed1b7ecc7cf81e573051068467c36f2e104f7e6dd2a7b05ca54b71425b410", + "component_budget": { + "node_count": 473, + "item_count": 473, + "soft_pair_count": 1738, + "serialized_chars": 901437, + "estimated_input_tokens": 225360, + "estimated_eval_tokens": 450976 + }, + "budget": { + "node_count": 47, + "item_count": 47, + "soft_pair_count": 27, + "serialized_chars": 35405, + "estimated_input_tokens": 8852, + "estimated_eval_tokens": 17960 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": true, + "analyst_result": { + "summary": "This is not a true duplicate cluster: the items span unrelated tokenizer, multimodal, loading, training, and cache bugs. None of the soft pairs look like the same underlying fixable change, so I would reject all of them; 46421 is only the least-bad anchor because it is current and broadly framed.", + "confidence": 0.81, + "canonical_issue_reason": "Issue 46421 is the best anchor only in a loose sense: it is current, open, and broadly scoped around cache behavior, whereas the rest are mostly unrelated one-off bugs.", + "canonical_pr_reason": null, + "best_issue_reason": "Issue 46421 is the most representative single issue here, but the set is too heterogeneous to treat as a real duplicate cluster.", + "best_pr_reason": null, + "soft_edge_verdicts": [ + { + "left": "issue:46421", + "right": "issue:46424", + "accept": false, + "reason": "Both are cache-related, but one is about chunked-prefill recompilation and the other about max_cache_length handling; different bugs and likely different fixes." + }, + { + "left": "issue:43262", + "right": "issue:44811", + "accept": false, + "reason": "Different subsystems and symptoms: audio chat-template sampling rate vs Whisper batch_decode special-token handling." + }, + { + "left": "issue:43643", + "right": "issue:45759", + "accept": false, + "reason": "Both involve config/model-loading logic, but one is missing fields from trust_remote_code and the other is composite config unwrapping; not the same bug." + }, + { + "left": "issue:43257", + "right": "issue:43696", + "accept": false, + "reason": "Qwen3 MoE weight conversion/loading bug vs GPT-oss OOM; unrelated failure modes." + }, + { + "left": "issue:45459", + "right": "issue:45803", + "accept": false, + "reason": "Both mention hiding errors, but they are different code paths and different root causes; not the same bug." + }, + { + "left": "issue:43329", + "right": "issue:44479", + "accept": false, + "reason": "Both are video-related, but one is an undefined-variable bug in multimodal token counting while the other is a broader v5.3 video-input regression across models." + }, + { + "left": "issue:43650", + "right": "issue:43906", + "accept": false, + "reason": "No shared concrete bug is visible; one is non-informative and the other is an isolated reproduction of a separate issue." + }, + { + "left": "issue:43257", + "right": "issue:43931", + "accept": false, + "reason": "Different Qwen3 problems: MoE weight conversion under accelerate/deepspeed vs Qwen3-VL weight-shape mismatch on load." + }, + { + "left": "issue:43329", + "right": "issue:43450", + "accept": false, + "reason": "Both touch video processing, but one is a token-count helper bug and the other is incorrect batched output shapes; different fixes." + }, + { + "left": "issue:42915", + "right": "issue:43742", + "accept": false, + "reason": "Unrelated model-loading failures with different models and error classes." + }, + { + "left": "issue:43606", + "right": "issue:45859", + "accept": false, + "reason": "Both involve CPU offload, but they are different models and different incompatibilities; not one duplicate bug." + }, + { + "left": "issue:43749", + "right": "issue:44756", + "accept": false, + "reason": "Different memory-related problems: FSDP RAM-efficient loading vs disabling mmap on Strix Halo." + }, + { + "left": "issue:34567", + "right": "issue:40990", + "accept": false, + "reason": "Training-state token counting bug vs high perplexity on a model/eval setup; no shared underlying issue." + }, + { + "left": "issue:43454", + "right": "issue:43577", + "accept": false, + "reason": "Weight-tying failure causing bad generation vs dtype propagation staying float32; unrelated issues." + }, + { + "left": "issue:44291", + "right": "issue:45042", + "accept": false, + "reason": "Init_empty_weights argument compatibility vs torchvision dependency requirement in PIL image processors; different code paths." + }, + { + "left": "issue:42915", + "right": "issue:44821", + "accept": false, + "reason": "Completely different areas: FP8 MoE loading vs loading AutoImageProcessor from URL." + }, + { + "left": "issue:43525", + "right": "issue:46291", + "accept": false, + "reason": "Llama4 config field missing vs pipelines import failing on kernels version/revision requirements; unrelated." + }, + { + "left": "issue:29159", + "right": "issue:34567", + "accept": false, + "reason": "Tokenizer slow/fast inconsistency vs TrainerState token-count updating; different components and symptoms." + }, + { + "left": "issue:43927", + "right": "issue:44912", + "accept": false, + "reason": "Saving/loading custom token IDs vs MXFP4 quantization fallback; not the same underlying bug." + }, + { + "left": "issue:44568", + "right": "issue:45356", + "accept": false, + "reason": "Both are tokenizer regressions, but they affect different models and appear to stem from different tokenizer logic." + }, + { + "left": "issue:43697", + "right": "issue:45357", + "accept": false, + "reason": "Detection-model output regression vs Qwen3.5 save_pretrained key serialization; unrelated." + }, + { + "left": "issue:43404", + "right": "issue:45127", + "accept": false, + "reason": "Weight tying is mentioned in both, but one is a missing tie in model init and the other is LoRA merge collapse with extended vocab; different bugs." + }, + { + "left": "issue:43425", + "right": "issue:46291", + "accept": false, + "reason": "Torch version incompatibility vs kernels dependency/version-revision import failure; unrelated." + }, + { + "left": "issue:42915", + "right": "issue:45538", + "accept": false, + "reason": "FP8 MoE load failure vs tokenizer model_max_length constant misuse; no shared code path." + }, + { + "left": "issue:28282", + "right": "issue:34689", + "accept": false, + "reason": "Different load-time dependency/model-loading problems with different affected models and environments." + }, + { + "left": "issue:45137", + "right": "issue:45405", + "accept": false, + "reason": "DeepSpeed ZeRO3 deque error vs an upper-bound version bump for PEFT; 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": 21, + "review_unit_index": 21, + "review_unit_count": 56, + "status": "reviewed", + "reason": null, + "source": "llm", + "cache_hit": false, + "model": "gpt-5.4-mini?service_tier=flex", + "cluster_id": "cluster-10000-473", + "nodes": [ + "issue:24643", + "issue:27957", + "issue:28282", + "issue:29127", + "issue:29159", + "issue:33357", + "issue:34567", + "issue:36331", + "issue:38175", + "issue:41720", + "issue:42175", + "issue:42915", + "issue:43257", + "issue:43317", + "issue:43452", + "issue:43493", + "issue:43540", + "issue:43577", + "issue:43582", + "issue:43742", + "issue:43749", + "issue:43756", + "issue:43824", + "issue:43994", + "issue:44062", + "issue:44291", + "issue:44479", + "issue:44488", + "issue:44560", + "issue:44568", + "issue:44589", + "issue:44737", + "issue:44792", + "issue:44821", + "issue:44849", + "issue:44898", + "issue:44964", + "issue:44987", + "issue:44991", + "issue:45072", + "issue:45081", + "issue:45381", + "issue:45488", + "issue:45538", + "issue:45701", + "issue:45907", + "issue:46463", + "issue:46512" + ], + "soft_pairs": [ + "issue:42175|issue:43824", + "issue:33357|issue:34567", + "issue:43577|issue:46512", + "issue:43452|issue:43582", + "issue:43317|issue:45538", + "issue:28282|issue:29127", + "issue:43742|issue:44291", + "issue:44062|issue:44991", + "issue:41720|issue:43540", + "issue:44849|issue:45701", + "issue:42915|issue:43749", + "issue:43540|issue:44560", + "issue:44479|issue:45381", + "issue:43540|issue:44479", + "issue:24643|issue:29159", + "issue:27957|issue:28282", + "issue:42175|issue:44488", + "issue:38175|issue:44898", + "issue:44568|issue:45488", + "issue:44987|issue:45907", + "issue:43493|issue:43994", + "issue:44291|issue:44821", + "issue:43257|issue:43756", + "issue:34567|issue:36331", + "issue:44737|issue:46463", + "issue:44792|issue:44964", + "issue:43577|issue:45072", + "issue:44062|issue:45081", + "issue:43742|issue:44589" + ], + "prepared_review_unit_hash": "c632403f8b32e856d89fb7b646fe00d4fda3b7e86d44c05ca308b1869d197435", + "component_budget": { + "node_count": 473, + "item_count": 473, + "soft_pair_count": 1738, + "serialized_chars": 901437, + "estimated_input_tokens": 225360, + "estimated_eval_tokens": 450976 + }, + "budget": { + "node_count": 48, + "item_count": 48, + "soft_pair_count": 29, + "serialized_chars": 36483, + "estimated_input_tokens": 9121, + "estimated_eval_tokens": 18498 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": true, + "analyst_result": { + "summary": "The cluster is fairly heterogeneous, with most items only loosely related by being model/tokenizer/load regressions. I found two clear duplicate-like subgroups: Qwen2.5-VL video regression (44479/45381) and SigLIP2 wrong outputs (43493/43994).", + "confidence": 0.61, + "canonical_issue_reason": "issue:44479 is the strongest duplicate anchor in the set: it describes a broad Qwen video-input regression and directly subsumes the narrower 45381 report.", + "canonical_pr_reason": null, + "best_issue_reason": "issue:44479 is the most actionable representative of the clearest same-bug subgroup in this cluster, with explicit regression scope and a direct narrower counterpart.", + "best_pr_reason": null, + "soft_edge_verdicts": [ + { + "left": "issue:42175", + "right": "issue:43824", + "accept": false, + "reason": "Packaging/backend availability and a model import failure are different bugs and code paths." + }, + { + "left": "issue:33357", + "right": "issue:34567", + "accept": false, + "reason": "A macOS bus error in CLIP loading is unrelated to TrainerState token accounting." + }, + { + "left": "issue:43577", + "right": "issue:46512", + "accept": false, + "reason": "Both involve dtype, but one is a BLIP2 load-time model dtype bug and the other is stale config metadata; not the same bug." + }, + { + "left": "issue:43452", + "right": "issue:43582", + "accept": false, + "reason": "GGUF/from_pretrained breakage and AppleSilicon allocator warmup are unrelated failures." + }, + { + "left": "issue:43317", + "right": "issue:45538", + "accept": false, + "reason": "Device-map/offload loading and CLIPTokenizer max-length handling are different issues." + }, + { + "left": "issue:28282", + "right": "issue:29127", + "accept": false, + "reason": "Missing PyTorch import and a LayoutLMv3 error-message clarity issue are not duplicates." + }, + { + "left": "issue:43742", + "right": "issue:44291", + "accept": false, + "reason": "Different load-time failures with different triggers; no clear shared code path." + }, + { + "left": "issue:44062", + "right": "issue:44991", + "accept": false, + "reason": "Both are tokenizer bugs, but the specific failure modes and affected code paths differ." + }, + { + "left": "issue:41720", + "right": "issue:43540", + "accept": false, + "reason": "Qwen3 auto device-mapping CUDA assert and Qwen3OmniMoe video input validation are distinct problems." + }, + { + "left": "issue:44849", + "right": "issue:45701", + "accept": false, + "reason": "A Qwen3.5 hidden-states bug is not the same as a generic tokenization regression report." + }, + { + "left": "issue:42915", + "right": "issue:43749", + "accept": false, + "reason": "FP8 config failure and FSDP CPU RAM efficient loading are separate distributed/runtime issues." + }, + { + "left": "issue:43540", + "right": "issue:44560", + "accept": false, + "reason": "Both are video-related, but they affect different Qwen models with different errors and no clear shared bug." + }, + { + "left": "issue:44479", + "right": "issue:45381", + "accept": true, + "reason": "45381 is a narrower report of the same Qwen2.5-VL video-input/vision_position_ids regression covered by 44479." + }, + { + "left": "issue:43540", + "right": "issue:44479", + "accept": false, + "reason": "Different Qwen video bugs in different model families; too little evidence they are the same defect." + }, + { + "left": "issue:24643", + "right": "issue:29159", + "accept": false, + "reason": "DeepSpeed weight-shape training failure and tokenizer inconsistency are unrelated." + }, + { + "left": "issue:27957", + "right": "issue:28282", + "accept": false, + "reason": "Flash Attention 2 on XLM-Roberta and missing PyTorch import are different issues." + }, + { + "left": "issue:42175", + "right": "issue:44488", + "accept": false, + "reason": "Backend packaging and a model-load failure are not the same underlying bug." + }, + { + "left": "issue:38175", + "right": "issue:44898", + "accept": false, + "reason": "SigLIP2 zero-probability output and Perceiver non-default resolution failure involve different models and code paths." + }, + { + "left": "issue:44568", + "right": "issue:45488", + "accept": false, + "reason": "Tokenizer special-token insertion and LlamaTokenizer pre-tokenizer override are distinct tokenizer regressions." + }, + { + "left": "issue:44987", + "right": "issue:45907", + "accept": false, + "reason": "A model load failure and a Hub API revision-forwarding bug are unrelated." + }, + { + "left": "issue:43493", + "right": "issue:43994", + "accept": true, + "reason": "Both report the same SigLIP2 implementation/output regression: one as a discrepancy vs JAX, the other as nonsensical outputs in AutoModel/pipeline." + }, + { + "left": "issue:44291", + "right": "issue:44821", + "accept": false, + "reason": "`init_empty_weights` argument handling and loading an image processor from URL are different failures." + }, + { + "left": "issue:43257", + "right": "issue:43756", + "accept": false, + "reason": "Qwen3 MoE deepspeed weight conversion and Smollm3 RoPE-layer dropping are unrelated model-specific bugs." + }, + { + "left": "issue:34567", + "right": "issue:36331", + "accept": false, + "reason": "TrainerState token counting and a CustomTrainer signature break are different trainer issues." + }, + { + "left": "issue:44737", + "right": "issue:46463", + "accept": false, + "reason": "XLNet CPU device placement and T5 decoder-input validation messaging are unrelated." + }, + { + "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 failure; not the same bug." + }, + { + "left": "issue:43577", + "right": "issue:45072", + "accept": false, + "reason": "These are separate dtype mismatch reports for different models/code paths." + }, + { + "left": "issue:44062", + "right": "issue:45081", + "accept": false, + "reason": "Different tokenizer failures: AddedToken construction vs Mistral regex patching." + }, + { + "left": "issue:43742", + "right": "issue:44589", + "accept": false, + "reason": "A MobileLLM key error and a Float8 storage lookup error are unrelated." + } + ] + }, + "evaluator_result": { + "accept": true, + "feedback": "The summary is grounded in the packet and stays conservative. The two accepted duplicate-like pairs are supported by the issue titles and the reported scopes, and the rejected soft edges are phrased cautiously without overclaiming. I don\u2019t see evidence of merge-incompatible pairings or unsupported certainty." + }, + "evaluator_used": true, + "retried": false, + "accepted_nontrivial_soft_edge": true, + "error_kind": null, + "error_message": null + }, + { + "label": "issue", + "component_index": 1, + "component_count": 21, + "review_unit_index": 22, + "review_unit_count": 56, + "status": "reviewed", + "reason": null, + "source": "llm", + "cache_hit": false, + "model": "gpt-5.4-mini?service_tier=flex", + "cluster_id": "cluster-10000-473", + "nodes": [ + "issue:24643", + "issue:28282", + "issue:29127", + "issue:32090", + "issue:35141", + "issue:41762", + "issue:43404", + "issue:43452", + "issue:43526", + "issue:43531", + "issue:43643", + "issue:43697", + "issue:43720", + "issue:43756", + "issue:43906", + "issue:44466", + "issue:44488", + "issue:44493", + "issue:44568", + "issue:44625", + "issue:44737", + "issue:44821", + "issue:44877", + "issue:44898", + "issue:44913", + "issue:44936", + "issue:45005", + "issue:45042", + "issue:45081", + "issue:45200", + "issue:45216", + "issue:45325", + "issue:45357", + "issue:45375", + "issue:45381", + "issue:45405", + "issue:45479", + "issue:45588", + "issue:45600", + "issue:45823", + "issue:45920", + "issue:45987", + "issue:46050", + "issue:46121", + "issue:46399", + "issue:46463", + "issue:46468", + "issue:46491" + ], + "soft_pairs": [ + "issue:43756|issue:46399", + "issue:44568|issue:46491", + "issue:45042|issue:45081", + "issue:43906|issue:46468", + "issue:43756|issue:45325", + "issue:44737|issue:45588", + "issue:45823|issue:46463", + "issue:45405|issue:45600", + "issue:43531|issue:44821", + "issue:43404|issue:45987", + "issue:43526|issue:45479", + "issue:43452|issue:44488", + "issue:45357|issue:45920", + "issue:43906|issue:45381", + "issue:44913|issue:46121", + "issue:44625|issue:44877", + "issue:43720|issue:44898", + "issue:43697|issue:45216", + "issue:24643|issue:29127", + "issue:43643|issue:45375", + "issue:28282|issue:32090", + "issue:45200|issue:46463", + "issue:28282|issue:35141", + "issue:41762|issue:44936", + "issue:44466|issue:45005", + "issue:44493|issue:46050" + ], + "prepared_review_unit_hash": "72bbae037fa57efdc430c9d54a454d2e265f72ac00781caa3ee803daafe35a06", + "component_budget": { + "node_count": 473, + "item_count": 473, + "soft_pair_count": 1738, + "serialized_chars": 901437, + "estimated_input_tokens": 225360, + "estimated_eval_tokens": 450976 + }, + "budget": { + "node_count": 48, + "item_count": 48, + "soft_pair_count": 26, + "serialized_chars": 36127, + "estimated_input_tokens": 9032, + "estimated_eval_tokens": 18320 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": true, + "analyst_result": { + "summary": "This cluster is a loose set of Transformer bug reports across unrelated subsystems (tokenizers, configs, training, vision, DeepSpeed, tied weights). None of the proposed soft pairs look like the same underlying bug/change.", + "confidence": 0.78, + "canonical_issue_reason": "issue:45920 is the broadest, most representative regression report here: it clearly describes a v5 AutoTokenizer break affecting multiple model families and is concrete enough to anchor the cluster.", + "canonical_pr_reason": null, + "best_issue_reason": "issue:45920 has the best combination of breadth, clear repro impact, and cross-model relevance, making it the strongest issue anchor among this mixed set.", + "best_pr_reason": null, + "soft_edge_verdicts": [ + { + "left": "issue:43756", + "right": "issue:46399", + "accept": false, + "reason": "Both mention RoPE/rotation-adjacent behavior, but one is about Smollm3 dropping layers and the other is a Qwen3.5 kernel-registration crash. Different models and different failure modes." + }, + { + "left": "issue:44568", + "right": "issue:46491", + "accept": false, + "reason": "Both are tokenizer regressions, but mdeberta-v3 missing add_special_tokens BOS/EOS is unrelated to CodeLlama stripping a leading space on round-trip." + }, + { + "left": "issue:45042", + "right": "issue:45081", + "accept": false, + "reason": "One is PIL image processors incorrectly requiring torchvision; the other is a Mistral tokenizer regex patch crash. Different code paths entirely." + }, + { + "left": "issue:43906", + "right": "issue:46468", + "accept": false, + "reason": "No evidence these are the same bug: one is a reproduction of an earlier issue, the other is a broken Nemotron example. Too generic to merge." + }, + { + "left": "issue:43756", + "right": "issue:45325", + "accept": false, + "reason": "Both involve RoPE/position-related logic, but they concern different models and distinct scaling/coverage bugs. Not the same underlying defect." + }, + { + "left": "issue:44737", + "right": "issue:45588", + "accept": false, + "reason": "XLNet CPU arange in relative position encoding is unrelated to flash_attention.py crashing on s_aux=None." + }, + { + "left": "issue:45823", + "right": "issue:46463", + "accept": false, + "reason": "Gemma4 PLE device mismatch during forward is unrelated to T5 decoder-input validation error messaging." + }, + { + "left": "issue:45405", + "right": "issue:45600", + "accept": false, + "reason": "A dependency-version bump and a stale auto_mappings reference are unrelated CI/config maintenance issues." + }, + { + "left": "issue:43531", + "right": "issue:44821", + "accept": false, + "reason": "Qwen3-MoE sliding_window behavior and AutoImageProcessor URL loading are different subsystems with no shared concrete bug." + }, + { + "left": "issue:43404", + "right": "issue:45987", + "accept": false, + "reason": "Tied lm_head weight serialization and StaticCache.get_seq_length returning a shape-(1,) tensor are unrelated issues." + }, + { + "left": "issue:43526", + "right": "issue:45479", + "accept": false, + "reason": "BEiTImageProcessorFast reduce_labels returning one label is unrelated to the zero-loss classification bug with num_labels=1." + }, + { + "left": "issue:43452", + "right": "issue:44488", + "accept": false, + "reason": "Both involve model/tokenizer loading failures, but one is about gguf_file handling and the other is a specific model load regression. Not the same bug." + }, + { + "left": "issue:45357", + "right": "issue:45920", + "accept": false, + "reason": "Both touch Qwen3.5, but one is about save_pretrained saving visual encoder keys and the other is AutoTokenizer producing wrong token IDs across many models." + }, + { + "left": "issue:43906", + "right": "issue:45381", + "accept": false, + "reason": "A reproduction of an earlier issue and a Qwen2.5-VL video position-id bug are different reports with no concrete overlap." + }, + { + "left": "issue:44913", + "right": "issue:46121", + "accept": false, + "reason": "GPTNeoX rotary_pct reloading default and convert_rope_params_to_dict failing on JSON-loaded lists are both rope-related, but they are distinct config-handling bugs." + }, + { + "left": "issue:44625", + "right": "issue:44877", + "accept": false, + "reason": "Qwen3.5 num_labels propagation and strict config blocking granite_speech loading are unrelated config issues." + }, + { + "left": "issue:43720", + "right": "issue:44898", + "accept": false, + "reason": "BitNet packed-weight unpacking during accelerate loading is unrelated to Perceiver image classification failing with interpolate_pos_encoding=True." + }, + { + "left": "issue:43697", + "right": "issue:45216", + "accept": false, + "reason": "RTDetrV2 output differences and Qwen3.5 save_pretrained corruption are separate regressions in different components." + }, + { + "left": "issue:24643", + "right": "issue:29127", + "accept": false, + "reason": "DeepSpeed training weight shape error and layoutlmv3 box-information error are entirely unrelated." + }, + { + "left": "issue:43643", + "right": "issue:45375", + "accept": false, + "reason": "Missing fields in AutoConfig with trust_remote_code and a specific Qwen3_5MoeVisionConfig field being dropped are not the same bug." + }, + { + "left": "issue:28282", + "right": "issue:32090", + "accept": false, + "reason": "PyTorch-not-found ImportError and Trainer _gpu_broadcast_one(NoneType) are unrelated environment/runtime failures." + }, + { + "left": "issue:45200", + "right": "issue:46463", + "accept": false, + "reason": "Gemma4 mm_token_type_ids defaulting to zeros and a T5 decoder-input validation error message are different issues." + }, + { + "left": "issue:44466", + "right": "issue:45005", + "accept": false, + "reason": "Both mention tied weights, but one is inconsistent serialization of lm_head.weight and the other is a broader translation-model tied-weights issue. Not enough to treat as one bug." + }, + { + "left": "issue:44493", + "right": "issue:46050", + "accept": false, + "reason": "Unexpected position-id key warnings and quantized-model save/load breakage 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": 21, + "review_unit_index": 23, + "review_unit_count": 56, + "status": "reviewed", + "reason": null, + "source": "llm", + "cache_hit": false, + "model": "gpt-5.4-mini?service_tier=flex", + "cluster_id": "cluster-10000-473", + "nodes": [ + "issue:29159", + "issue:33357", + "issue:38175", + "issue:42898", + "issue:43295", + "issue:43452", + "issue:43531", + "issue:43550", + "issue:43577", + "issue:43646", + "issue:43653", + "issue:43688", + "issue:43697", + "issue:43742", + "issue:43749", + "issue:43756", + "issue:43819", + "issue:43927", + "issue:44361", + "issue:44442", + "issue:44464", + "issue:44479", + "issue:44484", + "issue:44514", + "issue:44568", + "issue:44625", + "issue:44704", + "issue:44734", + "issue:44743", + "issue:44811", + "issue:44843", + "issue:44936", + "issue:44991", + "issue:45356", + "issue:45372", + "issue:45381", + "issue:45440", + "issue:45491", + "issue:45542", + "issue:45701", + "issue:45704", + "issue:45823", + "issue:46002", + "issue:46170", + "issue:46178", + "issue:46402", + "issue:46491", + "issue:46612" + ], + "soft_pairs": [ + "issue:43819|issue:44811", + "issue:44361|issue:44442", + "issue:43452|issue:44843", + "issue:43688|issue:43749", + "issue:38175|issue:43295", + "issue:43653|issue:44361", + "issue:45372|issue:46170", + "issue:44936|issue:45823", + "issue:45542|issue:46178", + "issue:42898|issue:46491", + "issue:43452|issue:43577", + "issue:43531|issue:43742", + "issue:44514|issue:45381", + "issue:43295|issue:45356", + "issue:43577|issue:43697", + "issue:43653|issue:44568", + "issue:43646|issue:43749", + "issue:44464|issue:44743", + "issue:44843|issue:44991", + "issue:29159|issue:33357", + "issue:43688|issue:45704", + "issue:44484|issue:44704", + "issue:43653|issue:43927", + "issue:43756|issue:46402", + "issue:43688|issue:45440", + "issue:43688|issue:45701", + "issue:43550|issue:44361", + "issue:44479|issue:44625", + "issue:45491|issue:46002", + "issue:44734|issue:46612" + ], + "prepared_review_unit_hash": "509cac7caf2dea4000d46198f8a3ffe2cba09800b896ef819a9a22c7a3198aaa", + "component_budget": { + "node_count": 473, + "item_count": 473, + "soft_pair_count": 1738, + "serialized_chars": 901437, + "estimated_input_tokens": 225360, + "estimated_eval_tokens": 450976 + }, + "budget": { + "node_count": 48, + "item_count": 48, + "soft_pair_count": 30, + "serialized_chars": 36849, + "estimated_input_tokens": 9213, + "estimated_eval_tokens": 18682 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": true, + "analyst_result": { + "summary": "This set is mostly a heterogeneous collection of unrelated bug reports; no soft pair looks like the same underlying issue, so I reject all candidate merges. The broadest tokenizer-regression report is the best available issue anchor, but the cluster itself is not cohesive.", + "confidence": 0.87, + "canonical_issue_reason": "issue:45701 is the broadest umbrella-style tokenizer regression report and the most plausible anchor among a mixed set.", + "canonical_pr_reason": null, + "best_issue_reason": "issue:45701 has the broadest scope and strongest tokenizer-regression framing, making it the best representative issue in this heterogeneous cluster.", + "best_pr_reason": null, + "soft_edge_verdicts": [ + { + "left": "issue:43819", + "right": "issue:44811", + "accept": false, + "reason": "Different models and APIs: DAC latent decode vs Whisper batch decode special-token handling." + }, + { + "left": "issue:44361", + "right": "issue:44442", + "accept": false, + "reason": "Both are tokenizer loading failures, but for different tokenizers and different error modes." + }, + { + "left": "issue:43452", + "right": "issue:44843", + "accept": false, + "reason": "Both touch tokenizer loading, but one is gguf loading and the other is offline hub regex patching." + }, + { + "left": "issue:43688", + "right": "issue:43749", + "accept": false, + "reason": "Unrelated subsystems: MoE auxiliary-loss normalization vs FSDP RAM-efficient loading." + }, + { + "left": "issue:38175", + "right": "issue:43295", + "accept": false, + "reason": "Different regression classes: zero probabilities in SigLIP2 vs processor/tokenizer regression in custom model code." + }, + { + "left": "issue:43653", + "right": "issue:44361", + "accept": false, + "reason": "Both are tokenizer bugs, but one is BigBird special-token registration and the other is MLuke task AttributeError." + }, + { + "left": "issue:45372", + "right": "issue:46170", + "accept": false, + "reason": "Completely different failures: Gemma processor import issue vs Windows/ROCm torch.distributed dependency." + }, + { + "left": "issue:44936", + "right": "issue:45823", + "accept": false, + "reason": "Training/eval state bug vs Gemma4 device-map mismatch; not the same code path." + }, + { + "left": "issue:45542", + "right": "issue:46178", + "accept": false, + "reason": "Both concern backend availability, but they target different backend-detection errors and different packages." + }, + { + "left": "issue:42898", + "right": "issue:46491", + "accept": false, + "reason": "Both are tokenization regressions, but one is a v5 cleanup behavior change and the other is a CodeLlama spacing round-trip bug." + }, + { + "left": "issue:43452", + "right": "issue:43577", + "accept": false, + "reason": "GGUF loading failure and Blip2 dtype propagation are unrelated issues." + }, + { + "left": "issue:43531", + "right": "issue:43742", + "accept": false, + "reason": "Different problems: Qwen3-MoE sliding_window behavior vs MobileLLM config loading key error." + }, + { + "left": "issue:44514", + "right": "issue:45381", + "accept": false, + "reason": "Same model family, but batching/padding crash and video vision_position_ids bug are distinct." + }, + { + "left": "issue:43295", + "right": "issue:45356", + "accept": false, + "reason": "Both are regressions, but one is custom processor/tokenizer access and the other is Kimi-K2.5 codec handling." + }, + { + "left": "issue:43577", + "right": "issue:43697", + "accept": false, + "reason": "Different defects: dtype loading for Blip2 vs divergent RTDetrV2 outputs in v5." + }, + { + "left": "issue:43653", + "right": "issue:44568", + "accept": false, + "reason": "Both involve special tokens, but the tokenizer classes and failure modes differ." + }, + { + "left": "issue:43646", + "right": "issue:43749", + "accept": false, + "reason": "Custom model initialization regression is unrelated to FSDP CPU RAM-efficient loading." + }, + { + "left": "issue:44464", + "right": "issue:44743", + "accept": false, + "reason": "Both involve generation/cache, but one is compiled chunked generation and the other is recurrent state reset." + }, + { + "left": "issue:44843", + "right": "issue:44991", + "accept": false, + "reason": "Different tokenizer-loading regressions with different triggers and models." + }, + { + "left": "issue:29159", + "right": "issue:33357", + "accept": false, + "reason": "Slow/fast tokenizer inconsistency is unrelated to the Mac bus error in CLIP loading." + }, + { + "left": "issue:43688", + "right": "issue:45704", + "accept": false, + "reason": "Aux-loss normalization and apex RMSNorm memory leak are unrelated." + }, + { + "left": "issue:44484", + "right": "issue:44704", + "accept": false, + "reason": "save_pretrained sharding policy question vs AutoProcessor kwargs propagation bug." + }, + { + "left": "issue:43653", + "right": "issue:43927", + "accept": false, + "reason": "Special-token registration bug vs config save/load losing custom token IDs." + }, + { + "left": "issue:43756", + "right": "issue:46402", + "accept": false, + "reason": "RoPE-layer truncation in Smollm3 is unrelated to Sam3 text encoder weights not loading." + }, + { + "left": "issue:43688", + "right": "issue:45440", + "accept": false, + "reason": "MoE loss normalization and DeepSeekV3 implementation divergence are different code-path problems." + }, + { + "left": "issue:43688", + "right": "issue:45701", + "accept": false, + "reason": "Auxiliary-loss normalization is unrelated to the broad tokenization regression report." + }, + { + "left": "issue:43550", + "right": "issue:44361", + "accept": false, + "reason": "torch.compile/SDPA model failure is unrelated to MLuke tokenizer task AttributeError." + }, + { + "left": "issue:44479", + "right": "issue:44625", + "accept": false, + "reason": "Video-input regression for Qwen models is distinct from num_labels propagation in Qwen3.5 config." + }, + { + "left": "issue:45491", + "right": "issue:46002", + "accept": false, + "reason": "Different NaN sources: sliding-window attention vs classification-head init/checkpointing." + }, + { + "left": "issue:44734", + "right": "issue:46612", + "accept": false, + "reason": "Both touch cache logic, but one is serve KV-continuation indexing and the other is beam-search cache reorder omission." + } + ] + }, + "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": 21, + "review_unit_index": 24, + "review_unit_count": 56, + "status": "reviewed", + "reason": null, + "source": "llm", + "cache_hit": false, + "model": "gpt-5.4-mini?service_tier=flex", + "cluster_id": "cluster-10000-473", + "nodes": [ + "issue:28218", + "issue:32090", + "issue:33357", + "issue:33666", + "issue:36010", + "issue:36331", + "issue:39692", + "issue:41720", + "issue:43262", + "issue:43278", + "issue:43295", + "issue:43452", + "issue:43493", + "issue:43504", + "issue:43525", + "issue:43550", + "issue:43572", + "issue:43577", + "issue:43606", + "issue:43643", + "issue:43688", + "issue:43720", + "issue:43742", + "issue:43825", + "issue:43844", + "issue:43927", + "issue:43994", + "issue:44361", + "issue:44451", + "issue:44464", + "issue:44466", + "issue:44514", + "issue:44554", + "issue:44561", + "issue:44610", + "issue:44625", + "issue:44821", + "issue:44849", + "issue:44964", + "issue:44977", + "issue:45290", + "issue:45405", + "issue:45584", + "issue:45823", + "issue:46050", + "issue:46192", + "issue:46209", + "issue:46365" + ], + "soft_pairs": [ + "issue:33666|issue:36010", + "issue:44464|issue:44977", + "issue:32090|issue:33666", + "issue:43742|issue:44821", + "issue:41720|issue:44464", + "issue:44466|issue:46050", + "issue:43504|issue:43720", + "issue:43504|issue:43550", + "issue:43643|issue:44561", + "issue:43577|issue:44554", + "issue:43504|issue:44610", + "issue:44514|issue:45290", + "issue:43720|issue:43927", + "issue:43606|issue:43927", + "issue:33666|issue:36331", + "issue:43262|issue:43295", + "issue:44554|issue:46365", + "issue:43278|issue:43688", + "issue:44977|issue:46209", + "issue:43720|issue:44361", + "issue:39692|issue:43493", + "issue:44964|issue:45405", + "issue:43452|issue:44451", + "issue:43525|issue:43572", + "issue:43493|issue:43825", + "issue:44625|issue:45584", + "issue:45823|issue:46050", + "issue:43994|issue:46192", + "issue:28218|issue:33357", + "issue:39692|issue:43295", + "issue:43278|issue:43844", + "issue:43493|issue:43844", + "issue:33357|issue:33666", + "issue:44849|issue:46209" + ], + "prepared_review_unit_hash": "ef71d6f2b42db92c42bbd9473241433b5e499c5c03cad997a162b38029f5028e", + "component_budget": { + "node_count": 473, + "item_count": 473, + "soft_pair_count": 1738, + "serialized_chars": 901437, + "estimated_input_tokens": 225360, + "estimated_eval_tokens": 450976 + }, + "budget": { + "node_count": 48, + "item_count": 48, + "soft_pair_count": 34, + "serialized_chars": 37961, + "estimated_input_tokens": 9491, + "estimated_eval_tokens": 19238 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": true, + "analyst_result": { + "summary": "This cluster is highly heterogeneous: the soft pairs point to different bugs across tokenization, model loading, training, generation, configs, and device/runtime issues. I reject every soft edge as a duplicate false positive.", + "confidence": 0.97, + "canonical_issue_reason": "No true canonical duplicate exists here because the items describe unrelated failures. If a representative must be named, issue 43295 is the broadest user-facing regression, but it still does not subsume the rest.", + "canonical_pr_reason": null, + "best_issue_reason": "Issue 43295 is the most suitable anchor only in the sense that it is a concrete, reproducible regression on a common API path; it is not a good dedupe hub for the other issues.", + "best_pr_reason": null, + "soft_edge_verdicts": [ + { + "left": "issue:33666", + "right": "issue:36010", + "accept": false, + "reason": "Different problems: Qwen2-VL multi-GPU training vs a GenerationMixin import error." + }, + { + "left": "issue:44464", + "right": "issue:44977", + "accept": false, + "reason": "Different failure modes: chunked generation with compiled forward vs Qwen3.5 flash-attention generation." + }, + { + "left": "issue:32090", + "right": "issue:33666", + "accept": false, + "reason": "Trainer GPU broadcast TypeError is unrelated to Qwen2-VL multi-GPU training." + }, + { + "left": "issue:43742", + "right": "issue:44821", + "accept": false, + "reason": "KeyError loading MobileLLM is unrelated to AutoImageProcessor-from-URL loading." + }, + { + "left": "issue:41720", + "right": "issue:44464", + "accept": false, + "reason": "CUDA assert under auto device mapping is not the same bug as compiled chunked generation inconsistency." + }, + { + "left": "issue:44466", + "right": "issue:46050", + "accept": false, + "reason": "lm_head serialization/tied-weights behavior differs from sinq save/load quantization breakage." + }, + { + "left": "issue:43504", + "right": "issue:43720", + "accept": false, + "reason": "Legacy-field BEiT load bug is unrelated to BitNet packed-weight loading during accelerate." + }, + { + "left": "issue:43504", + "right": "issue:43550", + "accept": false, + "reason": "BEiT pretrained-load failure and Bamba torch.compile SDPA failure are distinct issues." + }, + { + "left": "issue:43643", + "right": "issue:44561", + "accept": false, + "reason": "Missing fields from AutoConfig with trust_remote_code is not the same as removal of is_torch_fx_available breaking models." + }, + { + "left": "issue:43577", + "right": "issue:44554", + "accept": false, + "reason": "Blip2 dtype propagation bug is unrelated to the MPS attention correctness issue." + }, + { + "left": "issue:43504", + "right": "issue:44610", + "accept": false, + "reason": "BEiT legacy-field loading failure is unrelated to OmDet-Turbo processor/model size mismatch." + }, + { + "left": "issue:44514", + "right": "issue:45290", + "accept": false, + "reason": "Batched Qwen2.5-VL chat-template crash is a different code path from tool-call messages with no content." + }, + { + "left": "issue:43720", + "right": "issue:43927", + "accept": false, + "reason": "BitNet packed-weight loading and DiaConfig save/load token-ID loss are unrelated." + }, + { + "left": "issue:43606", + "right": "issue:43927", + "accept": false, + "reason": "CPU offload device mismatch in bark-small is unrelated to DiaConfig custom token IDs being lost." + }, + { + "left": "issue:33666", + "right": "issue:36331", + "accept": false, + "reason": "Qwen2-VL multi-GPU training and CustomTrainer.compute_loss signature mismatch are different issues." + }, + { + "left": "issue:43262", + "right": "issue:43295", + "accept": false, + "reason": "Audio processor sampling-rate defaulting is unrelated to processor.tokenizer access and image-tokenization regression." + }, + { + "left": "issue:44554", + "right": "issue:46365", + "accept": false, + "reason": "MPS attention correctness bug is unrelated to Trainer import failure caused by datasets." + }, + { + "left": "issue:43278", + "right": "issue:43688", + "accept": false, + "reason": "Evaluation dtype drift in embeddings is unrelated to auxiliary-loss normalization." + }, + { + "left": "issue:44977", + "right": "issue:46209", + "accept": false, + "reason": "Qwen3.5 flash-attention generation bug is unrelated to FP8 inference on B200." + }, + { + "left": "issue:43720", + "right": "issue:44361", + "accept": false, + "reason": "BitNet accelerate-loading bug is unrelated to MLukeTokenizer task-time AttributeError." + }, + { + "left": "issue:39692", + "right": "issue:43493", + "accept": false, + "reason": "SigLIP2 docs-example errors and SigLIP2 HF-vs-JAX discrepancy are different underlying problems." + }, + { + "left": "issue:44964", + "right": "issue:45405", + "accept": false, + "reason": "Phi-4 multimodal load failure is unrelated to the PEFT version bump issue." + }, + { + "left": "issue:43452", + "right": "issue:44451", + "accept": false, + "reason": "gguf_file regressions for tokenizer/model loading are not the same as loading ScandiBERT." + }, + { + "left": "issue:43525", + "right": "issue:43572", + "accept": false, + "reason": "Missing pad_token_id in Llama4Config and missing pad_token_idx in StableLmConfig are separate model/config bugs." + }, + { + "left": "issue:43493", + "right": "issue:43825", + "accept": false, + "reason": "SigLIP2 implementation discrepancy is unrelated to the pipeline translation-support error message." + }, + { + "left": "issue:44625", + "right": "issue:45584", + "accept": false, + "reason": "Qwen3.5 num_labels propagation and Whisper empty-transcription generation failure are distinct." + }, + { + "left": "issue:45823", + "right": "issue:46050", + "accept": false, + "reason": "Gemma4 device-mapping mismatch is unrelated to sinq quantization save/load breakage." + }, + { + "left": "issue:43994", + "right": "issue:46192", + "accept": false, + "reason": "SigLIP2 wrong outputs via AutoModel/pipeline is not the same as return_tensors being ignored with text_kwargs." + }, + { + "left": "issue:28218", + "right": "issue:33357", + "accept": false, + "reason": "Tokenizer extra-space behavior is unrelated to the MacOS bus error in CLIP loading." + }, + { + "left": "issue:39692", + "right": "issue:43295", + "accept": false, + "reason": "SigLIP2 docs-example errors do not match the processor.tokenizer regression in custom model code." + }, + { + "left": "issue:43278", + "right": "issue:43844", + "accept": false, + "reason": "Embedding dtype change at evaluation time is unrelated to the ZeRO-3 gradient growth issue." + }, + { + "left": "issue:43493", + "right": "issue:43844", + "accept": false, + "reason": "SigLIP2 HF-vs-JAX discrepancy is unrelated to abnormal gradients under DeepSpeed ZeRO-3." + }, + { + "left": "issue:33357", + "right": "issue:33666", + "accept": false, + "reason": "MacOS bus error in CLIP loading is unrelated to Qwen2-VL multi-GPU training." + }, + { + "left": "issue:44849", + "right": "issue:46209", + "accept": false, + "reason": "Qwen3.5 hidden-states bug is unrelated to FP8 inference failure on B200." + } + ] + }, + "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": 21, + "review_unit_index": 25, + "review_unit_count": 56, + "status": "reviewed", + "reason": null, + "source": "llm", + "cache_hit": false, + "model": "gpt-5.4-mini?service_tier=flex", + "cluster_id": "cluster-10000-473", + "nodes": [ + "issue:29942", + "issue:30333", + "issue:43257", + "issue:43278", + "issue:43299", + "issue:43334", + "issue:43504", + "issue:43540", + "issue:43572", + "issue:43575", + "issue:43606", + "issue:43646", + "issue:43653", + "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:44568", + "issue:44589", + "issue:44792", + "issue:44843", + "issue:44849", + "issue:44964", + "issue:44977", + "issue:45200", + "issue:45229", + "issue:45357", + "issue:45405", + "issue:45520", + "issue:45584", + "issue:45588", + "issue:45759", + "issue:45812", + "issue:45823", + "issue:46050", + "issue:46285", + "issue:46424", + "issue:46489" + ], + "soft_pairs": [ + "issue:43646|issue:44466", + "issue:44488|issue:44843", + "issue:44568|issue:46489", + "issue:43278|issue:43701", + "issue:43927|issue:44361", + "issue:43540|issue:44977", + "issue:44589|issue:45520", + "issue:43257|issue:43299", + "issue:43575|issue:45405", + "issue:44568|issue:45812", + "issue:44792|issue:45405", + "issue:43504|issue:43606", + "issue:44849|issue:45357", + "issue:43824|issue:44589", + "issue:43334|issue:43572", + "issue:43606|issue:44112", + "issue:43334|issue:45759", + "issue:44464|issue:46424", + "issue:43572|issue:44291", + "issue:43504|issue:43653", + "issue:44493|issue:46285", + "issue:45200|issue:45229", + "issue:44464|issue:45584", + "issue:45229|issue:45823", + "issue:43994|issue:44493", + "issue:43873|issue:46050", + "issue:43844|issue:44964", + "issue:45588|issue:46050", + "issue:29942|issue:30333" + ], + "prepared_review_unit_hash": "2d3067b1f4312f8720e40d3f067501e1b0ed0dab1291de0b210110601ceeb056", + "component_budget": { + "node_count": 473, + "item_count": 473, + "soft_pair_count": 1738, + "serialized_chars": 901437, + "estimated_input_tokens": 225360, + "estimated_eval_tokens": 450976 + }, + "budget": { + "node_count": 47, + "item_count": 47, + "soft_pair_count": 29, + "serialized_chars": 35932, + "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 mostly a loose collection of unrelated Transformers issues, with a few recurring themes around v5 regression, tokenizer/config loading, and Qwen/Gemma model-specific failures. None of the proposed soft pairs looks like the same underlying bug/change.", + "confidence": 0.41, + "canonical_issue_reason": "issue:43646 is the broadest, most general regression report in the set and best matches the common v5 loading/initialization failure theme.", + "canonical_pr_reason": null, + "best_issue_reason": "issue:43646 is the best single anchor because its scope is broad enough to represent the cluster\u2019s repeated model-loading breakages, unlike the many highly model-specific reports.", + "best_pr_reason": null, + "soft_edge_verdicts": [ + { + "left": "issue:43646", + "right": "issue:44466", + "accept": false, + "reason": "Generic initialization breakage vs. tied-weight serialization inconsistency; different failure modes and code paths." + }, + { + "left": "issue:44488", + "right": "issue:44843", + "accept": false, + "reason": "Specific model load failure vs. offline-mode hub call bug; not the same root cause." + }, + { + "left": "issue:44568", + "right": "issue:46489", + "accept": false, + "reason": "Both tokenizer bugs, but for different models and likely different fixes." + }, + { + "left": "issue:43278", + "right": "issue:43701", + "accept": false, + "reason": "BF16-to-FP32 dtype drift in eval vs. resume checkpoint key mismatch; unrelated bugs." + }, + { + "left": "issue:43927", + "right": "issue:44361", + "accept": false, + "reason": "Different tokenizer/config classes and different exceptions; no shared code-path evidence." + }, + { + "left": "issue:43540", + "right": "issue:44977", + "accept": false, + "reason": "Video input processing failure vs. flash-attention generation bug; distinct problems." + }, + { + "left": "issue:44589", + "right": "issue:45520", + "accept": false, + "reason": "Float8 storage deserialization error vs. flash_attn import key error; unrelated." + }, + { + "left": "issue:43257", + "right": "issue:43299", + "accept": false, + "reason": "Both involve Qwen MoE loading, but one is accelerate/deepspeed conversion and the other is Qwen3VL model loading; too different to merge." + }, + { + "left": "issue:43575", + "right": "issue:45405", + "accept": false, + "reason": "GPU OOM on tensor-parallel load vs. dependency version pin issue." + }, + { + "left": "issue:44568", + "right": "issue:45812", + "accept": false, + "reason": "Different tokenizer regressions affecting different model families; no same underlying bug." + }, + { + "left": "issue:44792", + "right": "issue:45405", + "accept": false, + "reason": "Janus test failure vs. PEFT version bump; unrelated." + }, + { + "left": "issue:43504", + "right": "issue:43606", + "accept": false, + "reason": "Legacy field loading bug vs. CPU offload device mismatch; unrelated." + }, + { + "left": "issue:44849", + "right": "issue:45357", + "accept": false, + "reason": "Forward/output_hidden_states bug vs. save_pretrained key regression; same model family but different code paths." + }, + { + "left": "issue:43824", + "right": "issue:44589", + "accept": false, + "reason": "Import error for a specific class vs. float8 storage lookup failure; unrelated." + }, + { + "left": "issue:43334", + "right": "issue:43572", + "accept": false, + "reason": "Both missing config fields, but for different models and different attributes; not the same bug." + }, + { + "left": "issue:43606", + "right": "issue:44112", + "accept": false, + "reason": "CPU offload mismatch in bark vs. a stale CI device override test in GraniteSpeech; different bugs." + }, + { + "left": "issue:43334", + "right": "issue:45759", + "accept": false, + "reason": "Both Qwen multimodal config issues, but one is missing pad token handling and the other is from_config unwrapping text_config." + }, + { + "left": "issue:44464", + "right": "issue:46424", + "accept": false, + "reason": "Compiled-forward generation inconsistency vs. cache-length handling for quantized cache; not the same issue." + }, + { + "left": "issue:43572", + "right": "issue:44291", + "accept": false, + "reason": "Config field rename bug vs. init_empty_weights argument incompatibility; unrelated." + }, + { + "left": "issue:43504", + "right": "issue:43653", + "accept": false, + "reason": "Legacy preset loading bug vs. tokenizer special-token registration bug; distinct fixes." + }, + { + "left": "issue:44493", + "right": "issue:46285", + "accept": false, + "reason": "Unexpected key regression across many models vs. CLIP class structure inconsistency; not one concrete bug." + }, + { + "left": "issue:45200", + "right": "issue:45229", + "accept": false, + "reason": "Missing default for mm_token_type_ids vs. multi-GPU CUDA OOM; different problems." + }, + { + "left": "issue:43994", + "right": "issue:44493", + "accept": false, + "reason": "Wrong SigLIP outputs vs. unexpected position-id keys; unrelated symptoms and paths." + }, + { + "left": "issue:43873", + "right": "issue:46050", + "accept": false, + "reason": "Quantization offloading behavior vs. quantized model save/load breakage; different failure modes." + }, + { + "left": "issue:43844", + "right": "issue:44964", + "accept": false, + "reason": "DeepSpeed gradient anomaly vs. Phi-4 multimodal load failure; unrelated." + }, + { + "left": "issue:45588", + "right": "issue:46050", + "accept": false, + "reason": "flash_attention sink-less model crash vs. sinq quantized save/load corruption; unrelated." + }, + { + "left": "issue:29942", + "right": "issue:30333", + "accept": false, + "reason": "Flash Attention 2 test failures vs. MLflow job-reporting issue; no shared 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": 21, + "review_unit_index": 26, + "review_unit_count": 56, + "status": "reviewed", + "reason": null, + "source": "llm", + "cache_hit": false, + "model": "gpt-5.4-mini?service_tier=flex", + "cluster_id": "cluster-10000-473", + "nodes": [ + "issue:15354", + "issue:28218", + "issue:29159", + "issue:32090", + "issue:34567", + "issue:37428", + "issue:38175", + "issue:39692", + "issue:42490", + "issue:43065", + "issue:43232", + "issue:43295", + "issue:43450", + "issue:43493", + "issue:43540", + "issue:43550", + "issue:43572", + "issue:43575", + "issue:43696", + "issue:43701", + "issue:43824", + "issue:43844", + "issue:43873", + "issue:43957", + "issue:44186", + "issue:44423", + "issue:44521", + "issue:44554", + "issue:44610", + "issue:44734", + "issue:44849", + "issue:44987", + "issue:45072", + "issue:45084", + "issue:45198", + "issue:45200", + "issue:45290", + "issue:45356", + "issue:45406", + "issue:45446", + "issue:45520", + "issue:45588", + "issue:45759", + "issue:45823", + "issue:46121", + "issue:46178", + "issue:46192", + "issue:46369" + ], + "soft_pairs": [ + "issue:42490|issue:43696", + "issue:29159|issue:32090", + "issue:43065|issue:46369", + "issue:45084|issue:45520", + "issue:28218|issue:34567", + "issue:44423|issue:44734", + "issue:44554|issue:45446", + "issue:15354|issue:38175", + "issue:45406|issue:46121", + "issue:43493|issue:43701", + "issue:45072|issue:45198", + "issue:37428|issue:43957", + "issue:43295|issue:43450", + "issue:43540|issue:45200", + "issue:45588|issue:45759", + "issue:44186|issue:46192", + "issue:45200|issue:45823", + "issue:45823|issue:46192", + "issue:43295|issue:44610", + "issue:43295|issue:43873", + "issue:43450|issue:43575", + "issue:45520|issue:46178", + "issue:43572|issue:44987", + "issue:43550|issue:43844", + "issue:43232|issue:44849", + "issue:44554|issue:45356", + "issue:37428|issue:43824", + "issue:39692|issue:46192", + "issue:43572|issue:43957", + "issue:44521|issue:45290", + "issue:42490|issue:43493" + ], + "prepared_review_unit_hash": "3bcac6422b86c61787028ef716d033e70417c43a33dc7b078f8a5e0d8e6bb6f3", + "component_budget": { + "node_count": 473, + "item_count": 473, + "soft_pair_count": 1738, + "serialized_chars": 901437, + "estimated_input_tokens": 225360, + "estimated_eval_tokens": 450976 + }, + "budget": { + "node_count": 48, + "item_count": 48, + "soft_pair_count": 31, + "serialized_chars": 37154, + "estimated_input_tokens": 9289, + "estimated_eval_tokens": 18834 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": true, + "analyst_result": { + "summary": "This cluster is a loose similarity bucket of mostly unrelated Transformers issues. The shared keywords (tokenizer, processor, serve, import, multimodal) are too broad to indicate duplicate bugs, so no soft edges should be merged.", + "confidence": 0.95, + "canonical_issue_reason": "Issue 44521 is the best anchor only because it is open, actively discussed, and has the most inbound references; however, the cluster is too heterogeneous to have a true canonical duplicate.", + "canonical_pr_reason": null, + "best_issue_reason": "Issue 44521 is the most representative high-signal report in the set due to its active discussion and broad multimodal chat-template impact, but it still does not unify the others as duplicates.", + "best_pr_reason": null, + "soft_edge_verdicts": [ + { + "left": "issue:42490", + "right": "issue:43696", + "accept": false, + "reason": "Local-model trailing-slash import warning vs GPT-oss OOM; unrelated failure modes." + }, + { + "left": "issue:29159", + "right": "issue:32090", + "accept": false, + "reason": "Tokenizer spacing inconsistency vs Trainer broadcast NoneType error; different components." + }, + { + "left": "issue:43065", + "right": "issue:46369", + "accept": false, + "reason": "Both mention SAM3, but one is a dummy Conv2d placeholder and the other is a memory overflow; not the same bug." + }, + { + "left": "issue:45084", + "right": "issue:45520", + "accept": false, + "reason": "Compile error on non-template nodes vs flash_attn import-key error; unrelated code paths." + }, + { + "left": "issue:28218", + "right": "issue:34567", + "accept": false, + "reason": "Added-token spacing bug vs TrainerState token counter not updating; different behavior." + }, + { + "left": "issue:44423", + "right": "issue:44734", + "accept": false, + "reason": "Both involve serve, but one is multimodal input conversion and the other is KV-cache continuation indexing." + }, + { + "left": "issue:44554", + "right": "issue:45446", + "accept": false, + "reason": "MPS attention correctness issue vs flex_attention AuxRequest version check; different subsystems." + }, + { + "left": "issue:15354", + "right": "issue:38175", + "accept": false, + "reason": "TorchScript export failure vs zero-probability output in another vision model; not the same underlying bug." + }, + { + "left": "issue:45406", + "right": "issue:46121", + "accept": false, + "reason": "Gemma4Processor _tokenizer crash vs rope-param JSON list type handling; unrelated." + }, + { + "left": "issue:43493", + "right": "issue:43701", + "accept": false, + "reason": "SigLIP2 implementation discrepancy vs resume checkpoint key mismatch; different root causes." + }, + { + "left": "issue:45072", + "right": "issue:45198", + "accept": false, + "reason": "bfloat16 dtype mismatch vs Wav2Vec2 save/tokenization failure; no shared code path." + }, + { + "left": "issue:37428", + "right": "issue:43957", + "accept": false, + "reason": "flash_attention_utils import error vs meta-device loading regression; unrelated." + }, + { + "left": "issue:43295", + "right": "issue:43450", + "accept": false, + "reason": "Processor/tokenizer regression vs batched video shape bug; same broad area, different issue." + }, + { + "left": "issue:43540", + "right": "issue:45200", + "accept": false, + "reason": "Video-processing ValueError vs mm_token_type_ids defaulting; different components." + }, + { + "left": "issue:45588", + "right": "issue:45759", + "accept": false, + "reason": "s_aux=None crash in flash_attention integration vs AutoModel config unwrapping; unrelated." + }, + { + "left": "issue:44186", + "right": "issue:46192", + "accept": false, + "reason": "LayoutLMv2Tokenizer crash on NER/padding vs return_tensors precedence; different tokenizer bugs." + }, + { + "left": "issue:45200", + "right": "issue:45823", + "accept": false, + "reason": "Missing mm_token_type_ids vs Gemma4 PLE device mismatch; unrelated." + }, + { + "left": "issue:45823", + "right": "issue:46192", + "accept": false, + "reason": "Device_map/forward device mismatch vs return_tensors being ignored; no common fix." + }, + { + "left": "issue:43295", + "right": "issue:44610", + "accept": false, + "reason": "Processor/tokenizer regression vs OmDet-Turbo input-size mismatch; different models and failures." + }, + { + "left": "issue:43295", + "right": "issue:43873", + "accept": false, + "reason": "Processor/tokenizer regression vs quantization offloading behavior; not the same bug." + }, + { + "left": "issue:43450", + "right": "issue:43575", + "accept": false, + "reason": "Video batch shape bug vs OOM when loading Qwen2-57B-A14B-Instruct; unrelated." + }, + { + "left": "issue:45520", + "right": "issue:46178", + "accept": false, + "reason": "flash_attn KeyError in import_utils vs tensorflow_text backend-mapping error; different import logic." + }, + { + "left": "issue:43572", + "right": "issue:44987", + "accept": false, + "reason": "StableLmConfig pad_token_idx regression vs loading physical-intelligence/fast on 5.1.0; different problems." + }, + { + "left": "issue:43550", + "right": "issue:43844", + "accept": false, + "reason": "torch.compile/SDPA inference failure vs ZeRO-3 gradient anomaly; unrelated training/inference paths." + }, + { + "left": "issue:43232", + "right": "issue:44849", + "accept": false, + "reason": "Generation kwargs update bug vs Qwen3.5 output_hidden_states bug; different code paths." + }, + { + "left": "issue:44554", + "right": "issue:45356", + "accept": false, + "reason": "MPS attention correctness vs Kimi tokenizer codec regression; not the same underlying issue." + }, + { + "left": "issue:37428", + "right": "issue:43824", + "accept": false, + "reason": "Different ImportErrors for different symbols and models; no shared fix." + }, + { + "left": "issue:39692", + "right": "issue:46192", + "accept": false, + "reason": "SigLIP2 docs-example errors vs return_tensors handling; unrelated." + }, + { + "left": "issue:43572", + "right": "issue:43957", + "accept": false, + "reason": "Config default regression vs meta-device model-loading breakage; different failure modes." + }, + { + "left": "issue:44521", + "right": "issue:45290", + "accept": false, + "reason": "Both use apply_chat_template, but one is multimodal assistant masks and the other is tool-call/no-content tokenization crash." + }, + { + "left": "issue:42490", + "right": "issue:43493", + "accept": false, + "reason": "Trailing-slash local model warning vs SigLIP2 implementation discrepancy; 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": 21, + "review_unit_index": 27, + "review_unit_count": 56, + "status": "reviewed", + "reason": null, + "source": "llm", + "cache_hit": false, + "model": "gpt-5.4-mini?service_tier=flex", + "cluster_id": "cluster-10000-473", + "nodes": [ + "issue:28218", + "issue:35141", + "issue:38175", + "issue:41720", + "issue:42915", + "issue:43262", + "issue:43278", + "issue:43334", + "issue:43335", + "issue:43388", + "issue:43450", + "issue:43452", + "issue:43493", + "issue:43525", + "issue:43526", + "issue:43550", + "issue:43653", + "issue:43819", + "issue:43824", + "issue:43827", + "issue:43856", + "issue:43901", + "issue:43950", + "issue:43975", + "issue:44112", + "issue:44206", + "issue:44442", + "issue:44466", + "issue:44479", + "issue:44561", + "issue:44610", + "issue:44792", + "issue:44857", + "issue:44898", + "issue:45083", + "issue:45127", + "issue:45198", + "issue:45200", + "issue:45325", + "issue:45381", + "issue:45412", + "issue:45563", + "issue:45584", + "issue:46192", + "issue:46209", + "issue:46234", + "issue:46285", + "issue:46512" + ], + "soft_pairs": [ + "issue:45381|issue:46209", + "issue:44466|issue:45127", + "issue:38175|issue:43493", + "issue:43653|issue:44112", + "issue:43278|issue:43856", + "issue:28218|issue:35141", + "issue:43452|issue:46512", + "issue:42915|issue:43824", + "issue:43950|issue:46285", + "issue:43334|issue:43525", + "issue:44898|issue:45412", + "issue:43262|issue:43450", + "issue:43824|issue:43975", + "issue:44442|issue:45198", + "issue:45200|issue:45325", + "issue:44479|issue:46192", + "issue:43388|issue:43526", + "issue:44792|issue:45563", + "issue:43819|issue:44112", + "issue:43827|issue:43901", + "issue:44610|issue:44857", + "issue:41720|issue:46209", + "issue:44561|issue:46234", + "issue:38175|issue:45412", + "issue:43335|issue:43550", + "issue:43819|issue:45083", + "issue:44206|issue:45584" + ], + "prepared_review_unit_hash": "d9db45a676e4b57703728969591b33178aed647fa2e69a02a4aebedbac6539e3", + "component_budget": { + "node_count": 473, + "item_count": 473, + "soft_pair_count": 1738, + "serialized_chars": 901437, + "estimated_input_tokens": 225360, + "estimated_eval_tokens": 450976 + }, + "budget": { + "node_count": 48, + "item_count": 48, + "soft_pair_count": 27, + "serialized_chars": 36198, + "estimated_input_tokens": 9050, + "estimated_eval_tokens": 18356 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": true, + "analyst_result": { + "summary": "Mostly unrelated bug reports; none of the soft pairs look like the same underlying bug/change with enough confidence to merge. The closest thematic overlaps are in model loading/serialization and Qwen multimodal regressions, but they still point to different code paths and model-specific failures.", + "confidence": 0.87, + "canonical_issue_reason": "issue:43950 is the strongest representative: it describes a clear, high-severity transformers 5.x regression in a core loading path (`from_pretrained`) with concrete corruption behavior and broad impact.", + "canonical_pr_reason": null, + "best_issue_reason": "issue:43950 is the best standalone issue in this set because it is specific, actionable, and affects a central API path rather than a model-specific edge case or test/doc inconsistency.", + "best_pr_reason": null, + "soft_edge_verdicts": [ + { + "left": "issue:45381", + "right": "issue:46209", + "accept": false, + "reason": "Different failures: Qwen2.5-VL video `vision_position_ids` regression vs FP8 inference crash on B200." + }, + { + "left": "issue:44466", + "right": "issue:45127", + "accept": false, + "reason": "Both touch tied weights, but one is device-dependent serialization and the other is LoRA merge with extended vocab; different bugs and fixes." + }, + { + "left": "issue:38175", + "right": "issue:43493", + "accept": false, + "reason": "Same model family, but zero probabilities and HF-vs-JAX discrepancy are not clearly the same concrete code-path issue." + }, + { + "left": "issue:43653", + "right": "issue:44112", + "accept": false, + "reason": "Tokenizer special-token registration bug vs GraniteSpeech CI device-override failure; unrelated." + }, + { + "left": "issue:43278", + "right": "issue:43856", + "accept": false, + "reason": "Embedding dtype regression vs MoE training memory usage; different symptoms and subsystems." + }, + { + "left": "issue:28218", + "right": "issue:35141", + "accept": false, + "reason": "Tokenizer spacing after added token is unrelated to embedding reinitialization after resize." + }, + { + "left": "issue:43452", + "right": "issue:46512", + "accept": false, + "reason": "gguf_file loading failure and stale `config.dtype` after `from_config` are separate loading-path issues." + }, + { + "left": "issue:42915", + "right": "issue:43824", + "accept": false, + "reason": "Qwen3Moe FP8 failure and Qwen2.5-VL import error are different model families and failure modes." + }, + { + "left": "issue:43950", + "right": "issue:46285", + "accept": false, + "reason": "Non-persistent buffer corruption in `from_pretrained` vs CLIP text-model structure inconsistency; not the same bug." + }, + { + "left": "issue:43334", + "right": "issue:43525", + "accept": false, + "reason": "Both mention missing `pad_token_id`, but they are model-specific config failures without evidence of one shared underlying fix." + }, + { + "left": "issue:44898", + "right": "issue:45412", + "accept": false, + "reason": "Perceiver interpolation failure and RT-DETR memory leak are unrelated." + }, + { + "left": "issue:43262", + "right": "issue:43450", + "accept": false, + "reason": "Audio chat-template sample-rate default bug vs batched video processor shape bug." + }, + { + "left": "issue:43824", + "right": "issue:43975", + "accept": false, + "reason": "Qwen2.5-VL import error and DeepSeek detokenization regression are unrelated." + }, + { + "left": "issue:44442", + "right": "issue:45198", + "accept": false, + "reason": "AutoTokenizer loading failure for FastSpeech2Conformer vs Wav2Vec2 save/tokenization failure; different code paths." + }, + { + "left": "issue:45200", + "right": "issue:45325", + "accept": false, + "reason": "Gemma 4 text-only fine-tuning default and Qwen2.5-VL rope-index scaling are unrelated multimodal bugs." + }, + { + "left": "issue:44479", + "right": "issue:46192", + "accept": false, + "reason": "Qwen video-input regression and `return_tensors` being ignored are distinct API issues." + }, + { + "left": "issue:43388", + "right": "issue:43526", + "accept": false, + "reason": "Metric gathering drops tuple labels; BeitImageProcessorFast reduces labels incorrectly. Similar area, different behavior." + }, + { + "left": "issue:44792", + "right": "issue:45563", + "accept": false, + "reason": "Janus image generation test failure and stale generate warning are unrelated." + }, + { + "left": "issue:43819", + "right": "issue:44112", + "accept": false, + "reason": "DAC latent mismatch and GraniteSpeech CI device override failure do not share the same code-path problem." + }, + { + "left": "issue:43827", + "right": "issue:43901", + "accept": false, + "reason": "Both are docs issues, but they concern different pipelines and different deprecated behaviors." + }, + { + "left": "issue:44610", + "right": "issue:44857", + "accept": false, + "reason": "OmDet-Turbo processor size mismatch vs LwDetrImageLoss AMP crash; unrelated." + }, + { + "left": "issue:41720", + "right": "issue:46209", + "accept": false, + "reason": "Auto device-mapping cuda assert and B200 FP8 inference failure are different hardware/runtime failures." + }, + { + "left": "issue:44561", + "right": "issue:46234", + "accept": false, + "reason": "Removal of one helper and absence of another are separate utility/API issues." + }, + { + "left": "issue:38175", + "right": "issue:45412", + "accept": false, + "reason": "SigLIP2 zero-probability issue and RT-DETR memory leak are unrelated." + }, + { + "left": "issue:43335", + "right": "issue:43550", + "accept": false, + "reason": "SwitchTransformers sparse-layer config bug vs Bamba torch.compile/SDPA failure; different subsystems." + }, + { + "left": "issue:43819", + "right": "issue:45083", + "accept": false, + "reason": "DAC.from_latents mismatch and qwen3_omni_moe feature-length helper behavior are unrelated." + }, + { + "left": "issue:44206", + "right": "issue:45584", + "accept": false, + "reason": "LaserFeatureExtractor unsupported arg and Whisper empty-transcription failure after token alignment 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": 21, + "review_unit_index": 28, + "review_unit_count": 56, + "status": "reviewed", + "reason": null, + "source": "llm", + "cache_hit": false, + "model": "gpt-5.4-mini?service_tier=flex", + "cluster_id": "cluster-10000-473", + "nodes": [ + "issue:28282", + "issue:29127", + "issue:29942", + "issue:32090", + "issue:33357", + "issue:34689", + "issue:35141", + "issue:37428", + "issue:39692", + "issue:43278", + "issue:43335", + "issue:43425", + "issue:43504", + "issue:43526", + "issue:43550", + "issue:43606", + "issue:43611", + "issue:43653", + "issue:43720", + "issue:43819", + "issue:43825", + "issue:43931", + "issue:44112", + "issue:44351", + "issue:44448", + "issue:44493", + "issue:44514", + "issue:44521", + "issue:44743", + "issue:44855", + "issue:44857", + "issue:45083", + "issue:45127", + "issue:45200", + "issue:45245", + "issue:45290", + "issue:45520", + "issue:45588", + "issue:45593", + "issue:45759", + "issue:45859", + "issue:46170", + "issue:46285", + "issue:46326", + "issue:46431", + "issue:46512", + "issue:46566" + ], + "soft_pairs": [ + "issue:45759|issue:46285", + "issue:44857|issue:45593", + "issue:45520|issue:46170", + "issue:43606|issue:45083", + "issue:37428|issue:43425", + "issue:29942|issue:32090", + "issue:43335|issue:43653", + "issue:43550|issue:44112", + "issue:28282|issue:33357", + "issue:44448|issue:44743", + "issue:43611|issue:43931", + "issue:44514|issue:44521", + "issue:45759|issue:46512", + "issue:43335|issue:43504", + "issue:44855|issue:45083", + "issue:46431|issue:46566", + "issue:43278|issue:43825", + "issue:29127|issue:32090", + "issue:37428|issue:44351", + "issue:45290|issue:46326", + "issue:43720|issue:44112", + "issue:43335|issue:45083", + "issue:43335|issue:44112", + "issue:45127|issue:45245", + "issue:45588|issue:45859", + "issue:43335|issue:45593", + "issue:45200|issue:45588", + "issue:34689|issue:35141", + "issue:39692|issue:43526", + "issue:43819|issue:44493" + ], + "prepared_review_unit_hash": "21a2e561d3be9232f2d6a8cb1f18a04605995afa96a7034ae5da1d5b3840134a", + "component_budget": { + "node_count": 473, + "item_count": 473, + "soft_pair_count": 1738, + "serialized_chars": 901437, + "estimated_input_tokens": 225360, + "estimated_eval_tokens": 450976 + }, + "budget": { + "node_count": 47, + "item_count": 47, + "soft_pair_count": 30, + "serialized_chars": 36526, + "estimated_input_tokens": 9132, + "estimated_eval_tokens": 18520 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": true, + "analyst_result": { + "summary": "No duplicate-worthy cluster emerged: the items are mostly unrelated issue reports, and the soft pairs share broad subsystems at most but not the same concrete bug.", + "confidence": 0.95, + "canonical_issue_reason": "No clear canonical issue: the cluster is heterogeneous, and the soft pairs differ in concrete failure mode even when they touch the same subsystem.", + "canonical_pr_reason": null, + "best_issue_reason": "No single issue is a strong global representative; the closest-looking pairs still describe different bugs, regressions, or error paths.", + "best_pr_reason": null, + "soft_edge_verdicts": [ + { + "left": "issue:45759", + "right": "issue:46285", + "accept": false, + "reason": "Both involve multimodal/composite model config loading, but one is AutoModelForCausalLM.from_config unwrapping text_config and the other is a CLIP text model structure regression." + }, + { + "left": "issue:44857", + "right": "issue:45593", + "accept": false, + "reason": "Both are vision/detection related, but one is an AMP float16 crash and the other is missing auxiliary losses when denoising is disabled." + }, + { + "left": "issue:45520", + "right": "issue:46170", + "accept": false, + "reason": "Both are import/platform errors, but they stem from different dependencies and environments: flash_attn on Python 3.13 vs torch.distributed on Windows ROCm." + }, + { + "left": "issue:43606", + "right": "issue:45083", + "accept": false, + "reason": "One is a CPU offload device-mismatch bug in bark-small; the other is a helper-length calculation issue in qwen3_omni_moe." + }, + { + "left": "issue:37428", + "right": "issue:43425", + "accept": false, + "reason": "Both mention torch/flash-attn compatibility, but one is a missing internal flash-attn import and the other is a broad torch 2.10 incompatibility report." + }, + { + "left": "issue:29942", + "right": "issue:32090", + "accept": false, + "reason": "One is failing Flash Attention 2 tests; the other is a Trainer broadcast TypeError on NoneType. Same area, different failure." + }, + { + "left": "issue:43335", + "right": "issue:43653", + "accept": false, + "reason": "One is a SwitchTransformers config bug creating sparse layers incorrectly; the other is a BigBirdTokenizer special-token registration issue." + }, + { + "left": "issue:43550", + "right": "issue:44112", + "accept": false, + "reason": "One is a Bamba torch.compile/SDPA failure; the other is a stale CI test in GraniteSpeech. Not the same code-path bug." + }, + { + "left": "issue:28282", + "right": "issue:33357", + "accept": false, + "reason": "One is a missing PyTorch ImportError; the other is a MacOS bus error loading a CLIP model. Different root causes." + }, + { + "left": "issue:44448", + "right": "issue:44743", + "accept": false, + "reason": "Both affect model behavior, but one is a Pegasus output regression and the other is recurrent state resetting in modular_qwen3_5." + }, + { + "left": "issue:43611", + "right": "issue:43931", + "accept": false, + "reason": "Both are model-loading failures, but one is about base_model_prefix handling and the other is a weight-shape mismatch for Qwen3-VL." + }, + { + "left": "issue:44514", + "right": "issue:44521", + "accept": false, + "reason": "Both touch apply_chat_template for multimodal inputs, but one crashes on batched padding=False and the other returns all-zero assistant masks." + }, + { + "left": "issue:45759", + "right": "issue:46512", + "accept": false, + "reason": "Both concern from_config/model initialization, but one is about unwrapping nested text_config and the other is stale config.dtype after dtype-aware construction." + }, + { + "left": "issue:43335", + "right": "issue:43504", + "accept": false, + "reason": "Both involve model/config setup, but one miscreates sparse layers and the other fails loading a BEiT preset due to a legacy field." + }, + { + "left": "issue:44855", + "right": "issue:45083", + "accept": false, + "reason": "One is a Python 3.13 import/torch.jit parsing failure; the other is an unexpected behavior bug in a feature-extraction helper." + }, + { + "left": "issue:46431", + "right": "issue:46566", + "accept": false, + "reason": "One is a missing chat template error in a processor; the other is weights \u10d3\u10d0\u10e0\u10e9ing on meta with device_map=auto. Different initialization paths." + }, + { + "left": "issue:43278", + "right": "issue:43825", + "accept": false, + "reason": "One is a dtype mismatch between training and evaluate; the other is a pipeline error message about unsupported translation tasks." + }, + { + "left": "issue:29127", + "right": "issue:32090", + "accept": false, + "reason": "One is about clarifying an error message in layoutlmv3; the other is a runtime NoneType broadcast failure." + }, + { + "left": "issue:37428", + "right": "issue:44351", + "accept": false, + "reason": "Both are import errors, but they involve different symbols and regressions: flash-attn support vs HybridCache availability." + }, + { + "left": "issue:45290", + "right": "issue:46326", + "accept": false, + "reason": "Both involve chat template/tokenization, but one crashes on assistant tool-call messages and the other is about continue_final_message for mistral_common tokenizers." + }, + { + "left": "issue:43720", + "right": "issue:44112", + "accept": false, + "reason": "One is BitNet accelerate-loading of packed weights; the other is a stale CI device-override test. No shared underlying bug." + }, + { + "left": "issue:43335", + "right": "issue:45083", + "accept": false, + "reason": "One is a SwitchTransformers config edge case; the other is a qwen3_omni_moe helper-length bug." + }, + { + "left": "issue:43335", + "right": "issue:44112", + "accept": false, + "reason": "One is a model construction bug and the other is a CI test failure; they do not share the same code-path." + }, + { + "left": "issue:45127", + "right": "issue:45245", + "accept": false, + "reason": "One is LoRA merge collapse with tied embeddings; the other is a categorical cardinality limit error. Different problems." + }, + { + "left": "issue:45588", + "right": "issue:45859", + "accept": false, + "reason": "Both are MoE/flash-related, but one is an AttributeError on s_aux=None and the other is a CPU-offload incompatibility in rotary embedding." + }, + { + "left": "issue:43335", + "right": "issue:45593", + "accept": false, + "reason": "One is a SwitchTransformers config bug; the other is a D-FINE loss-path issue. Same general domain only." + }, + { + "left": "issue:45200", + "right": "issue:45588", + "accept": false, + "reason": "One is about mm_token_type_ids defaults in Gemma 4; the other is a flash_attention sink-less model crash. Different mechanisms." + }, + { + "left": "issue:34689", + "right": "issue:35141", + "accept": false, + "reason": "One is a model-loading breakage for Llama 3.2 Vision; the other is output embeddings being reinitialized after resizing. Related to loading, not the same bug." + }, + { + "left": "issue:39692", + "right": "issue:43526", + "accept": false, + "reason": "One is documentation example errors; the other is a BeitImageProcessorFast reduce_labels bug. Different artifact types and failures." + }, + { + "left": "issue:43819", + "right": "issue:44493", + "accept": false, + "reason": "One is DAC.from_latents disagreeing with forward due to missing STE; the other is a position-id key warning across many models. Not the same 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": 21, + "review_unit_index": 29, + "review_unit_count": 56, + "status": "reviewed", + "reason": null, + "source": "llm", + "cache_hit": false, + "model": "gpt-5.4-mini?service_tier=flex", + "cluster_id": "cluster-10000-473", + "nodes": [ + "issue:10000", + "issue:15354", + "issue:26413", + "issue:27957", + "issue:28218", + "issue:29942", + "issue:31501", + "issue:32090", + "issue:33357", + "issue:34689", + "issue:35141", + "issue:38175", + "issue:39692", + "issue:40444", + "issue:42490", + "issue:43278", + "issue:43329", + "issue:43381", + "issue:43526", + "issue:43550", + "issue:43575", + "issue:43606", + "issue:43701", + "issue:43854", + "issue:44206", + "issue:44360", + "issue:44448", + "issue:44493", + "issue:44561", + "issue:44779", + "issue:44855", + "issue:44857", + "issue:44977", + "issue:45072", + "issue:45083", + "issue:45229", + "issue:45290", + "issue:45310", + "issue:45381", + "issue:45459", + "issue:45478", + "issue:45588", + "issue:45823", + "issue:45920", + "issue:46192", + "issue:46234", + "issue:46399", + "issue:46655" + ], + "soft_pairs": [ + "issue:26413|issue:31501", + "issue:45310|issue:45478", + "issue:42490|issue:43854", + "issue:45459|issue:46234", + "issue:43606|issue:44857", + "issue:44448|issue:44779", + "issue:39692|issue:44493", + "issue:44206|issue:45823", + "issue:28218|issue:32090", + "issue:44857|issue:45072", + "issue:44360|issue:44561", + "issue:43701|issue:45083", + "issue:38175|issue:44493", + "issue:32090|issue:33357", + "issue:42490|issue:43575", + "issue:40444|issue:43526", + "issue:29942|issue:35141", + "issue:43550|issue:45083", + "issue:45290|issue:45381", + "issue:44448|issue:44977", + "issue:27957|issue:34689", + "issue:45290|issue:46655", + "issue:45920|issue:46192", + "issue:44977|issue:45229", + "issue:43278|issue:43329", + "issue:10000|issue:15354", + "issue:44855|issue:46399", + "issue:43278|issue:43381", + "issue:44857|issue:45588" + ], + "prepared_review_unit_hash": "1dc8f5b4eab5b64fa12cddebdcb2feda9c1f651eb92bf86fc3e02c0db47bd615", + "component_budget": { + "node_count": 473, + "item_count": 473, + "soft_pair_count": 1738, + "serialized_chars": 901437, + "estimated_input_tokens": 225360, + "estimated_eval_tokens": 450976 + }, + "budget": { + "node_count": 48, + "item_count": 48, + "soft_pair_count": 29, + "serialized_chars": 36410, + "estimated_input_tokens": 9103, + "estimated_eval_tokens": 18462 + }, + "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 that only share broad subsystem keywords; one soft pair is a clear duplicate, but the rest are distinct bugs/regressions and should stay separate.", + "confidence": 0.95, + "canonical_issue_reason": "No single issue is a valid canonical for the cluster overall because the items describe different failures across tokenization, loading, training, multimodal processing, and dtype/runtime paths.", + "canonical_pr_reason": null, + "best_issue_reason": "No issue cleanly represents the whole cluster; the only near-duplicate pair is the duplicated Qwen3.5 MoE from_pretrained regression, which is too narrow to serve as a global representative.", + "best_pr_reason": null, + "soft_edge_verdicts": [ + { + "left": "issue:26413", + "right": "issue:31501", + "accept": false, + "reason": "Different problems: checkpoint resume/epoch iterator key handling vs training not starting on FineWeb. Same training area, but not the same bug." + }, + { + "left": "issue:45310", + "right": "issue:45478", + "accept": true, + "reason": "These appear to be the same Qwen3.5 MoE from_pretrained regression, with the same title and matching failure description." + }, + { + "left": "issue:42490", + "right": "issue:43854", + "accept": false, + "reason": "Different loading bugs: local path trailing-slash warning vs incorrect loading of a specific GLM-4.7 model in tests." + }, + { + "left": "issue:45459", + "right": "issue:46234", + "accept": false, + "reason": "Unrelated import-path issues: protobuf decode error masking tokenizer errors vs a missing import_utils helper function." + }, + { + "left": "issue:43606", + "right": "issue:44857", + "accept": false, + "reason": "Both involve runtime failures, but the concrete bugs differ: CPU offload device mismatch for bark-small vs float16 AMP crash in LwDetrImageLoss." + }, + { + "left": "issue:44448", + "right": "issue:44779", + "accept": false, + "reason": "Different regressions and code paths: Pegasus generation mismatch across v4/v5 vs DeepSeek tokenizer output changes in v5." + }, + { + "left": "issue:39692", + "right": "issue:44493", + "accept": false, + "reason": "Documentation example errors for SigLIP2 are not the same as the unexpected position-id key issue affecting models." + }, + { + "left": "issue:44206", + "right": "issue:45823", + "accept": false, + "reason": "Different failures: unsupported `center` arg in LasrFeatureExtractor vs Gemma4 PLE device mismatch during forward with `device_map=auto`." + }, + { + "left": "issue:28218", + "right": "issue:32090", + "accept": false, + "reason": "Tokenizer spacing bug and Trainer distributed broadcast NoneType error are unrelated." + }, + { + "left": "issue:44857", + "right": "issue:45072", + "accept": false, + "reason": "Both are dtype-related crashes, but they hit different models and code paths: AMP loss computation vs bfloat16 inference mismatches." + }, + { + "left": "issue:44360", + "right": "issue:44561", + "accept": false, + "reason": "DSA indexer missing ReLU is unrelated to the v5 removal of `is_torch_fx_available` breaking trust_remote_code models." + }, + { + "left": "issue:43701", + "right": "issue:45083", + "accept": false, + "reason": "Checkpoint key mismatch and multimodal feature-length helper behavior are different bugs." + }, + { + "left": "issue:38175", + "right": "issue:44493", + "accept": false, + "reason": "Unexpected zero probabilities in SigLIP2 are not the same as the unexpected position-id key regression." + }, + { + "left": "issue:32090", + "right": "issue:33357", + "accept": false, + "reason": "Trainer broadcast NoneType and MacOS bus error in CLIP loading are separate failures." + }, + { + "left": "issue:42490", + "right": "issue:43575", + "accept": false, + "reason": "Local-path import warning and tensor-parallel OOM for Qwen2-57B are unrelated." + }, + { + "left": "issue:40444", + "right": "issue:43526", + "accept": false, + "reason": "Fine-tuning a multimodal iterable dataset and BEiT label reduction are different processor/training bugs." + }, + { + "left": "issue:29942", + "right": "issue:35141", + "accept": false, + "reason": "Flash Attention 2 test failures do not match embedding reinitialization after resize in post_init." + }, + { + "left": "issue:43550", + "right": "issue:45083", + "accept": false, + "reason": "torch.compile/SDPA failure in Bamba is unrelated to qwen3_omni_moe feature-length helper behavior." + }, + { + "left": "issue:45290", + "right": "issue:45381", + "accept": false, + "reason": "Chat-template tool-call crash and wrong video vision_position_ids are both multimodal, but they are distinct code paths and bugs." + }, + { + "left": "issue:44448", + "right": "issue:44977", + "accept": false, + "reason": "Pegasus v4/v5 output difference and Qwen3.5 flash-attention generation failure are different model-specific regressions." + }, + { + "left": "issue:27957", + "right": "issue:34689", + "accept": false, + "reason": "Flash Attention 2 support issue and Llama 3.2 Vision model loading breakage are not the same problem." + }, + { + "left": "issue:45290", + "right": "issue:46655", + "accept": false, + "reason": "Tool-call chat templating crash and SmolVLM video padding calculation bug affect different layers of the stack." + }, + { + "left": "issue:45920", + "right": "issue:46192", + "accept": false, + "reason": "Wrong token IDs across many models is a tokenizer regression; silently ignored `return_tensors` is an argument-handling issue." + }, + { + "left": "issue:44977", + "right": "issue:45229", + "accept": false, + "reason": "Flash-attention generation bug in Qwen3.5 and Gemma4 multi-GPU CUDA OOM are different failure modes." + }, + { + "left": "issue:43278", + "right": "issue:43329", + "accept": false, + "reason": "Embedding dtype drift during evaluate is unrelated to the undefined-variable bug in multimodal token counting." + }, + { + "left": "issue:10000", + "right": "issue:15354", + "accept": false, + "reason": "These are entirely unrelated model/export issues with different architectures and failure mechanisms." + }, + { + "left": "issue:44855", + "right": "issue:46399", + "accept": false, + "reason": "Python 3.13 torch.jit parsing/IndentationError for DebertaV2 is unrelated to the Qwen3.5 missing-decorator registration crash." + }, + { + "left": "issue:43278", + "right": "issue:43381", + "accept": false, + "reason": "Eval-mode gradient-checkpointing restriction is a different bug from embedding dtype changing from BF16 to FP32." + }, + { + "left": "issue:44857", + "right": "issue:45588", + "accept": false, + "reason": "LwDetrImageLoss AMP crash and flash_attention.py s_aux=None AttributeError are separate code-path failures." + } + ] + }, + "evaluator_result": { + "accept": true, + "feedback": "The summary is grounded in the packet: the cluster is heterogeneous, and the only clearly duplicate soft pair is issue:45310/issue:45478, which the analyst identified conservatively. The remaining soft-edge verdicts are mostly cautious false negatives with reasons tied to distinct bug descriptions, and no claim appears overstated." + }, + "evaluator_used": true, + "retried": false, + "accepted_nontrivial_soft_edge": true, + "error_kind": null, + "error_message": null + }, + { + "label": "issue", + "component_index": 1, + "component_count": 21, + "review_unit_index": 30, + "review_unit_count": 56, + "status": "reviewed", + "reason": null, + "source": "llm", + "cache_hit": false, + "model": "gpt-5.4-mini?service_tier=flex", + "cluster_id": "cluster-10000-473", + "nodes": [ + "issue:25251", + "issue:29127", + "issue:32090", + "issue:33290", + "issue:33357", + "issue:34567", + "issue:37428", + "issue:39692", + "issue:41720", + "issue:42175", + "issue:42490", + "issue:42915", + "issue:42994", + "issue:43278", + "issue:43381", + "issue:43412", + "issue:43575", + "issue:43611", + "issue:43653", + "issue:43701", + "issue:43746", + "issue:43873", + "issue:43927", + "issue:43976", + "issue:44186", + "issue:44361", + "issue:44448", + "issue:44466", + "issue:44561", + "issue:44568", + "issue:44743", + "issue:44779", + "issue:44871", + "issue:44977", + "issue:45198", + "issue:45229", + "issue:45372", + "issue:45406", + "issue:45440", + "issue:45491", + "issue:45563", + "issue:45588", + "issue:45698", + "issue:45803", + "issue:45823", + "issue:46121", + "issue:46209" + ], + "soft_pairs": [ + "issue:43746|issue:46209", + "issue:43927|issue:45698", + "issue:37428|issue:42175", + "issue:44561|issue:44779", + "issue:43873|issue:45229", + "issue:32090|issue:33290", + "issue:25251|issue:32090", + "issue:41720|issue:45229", + "issue:43976|issue:45440", + "issue:44448|issue:44466", + "issue:42490|issue:43873", + "issue:44568|issue:45491", + "issue:45588|issue:45823", + "issue:43278|issue:43575", + "issue:44977|issue:45563", + "issue:33357|issue:39692", + "issue:45372|issue:45588", + "issue:29127|issue:34567", + "issue:45198|issue:45803", + "issue:43412|issue:45588", + "issue:43381|issue:43701", + "issue:43653|issue:44743", + "issue:44871|issue:44977", + "issue:43611|issue:46121", + "issue:44361|issue:45198", + "issue:44186|issue:45198", + "issue:45372|issue:45406", + "issue:42915|issue:42994" + ], + "prepared_review_unit_hash": "ed841b2f824d5aba87b4c8aa55643aed83cee2cc463c08eb81ba2637351b7330", + "component_budget": { + "node_count": 473, + "item_count": 473, + "soft_pair_count": 1738, + "serialized_chars": 901437, + "estimated_input_tokens": 225360, + "estimated_eval_tokens": 450976 + }, + "budget": { + "node_count": 47, + "item_count": 47, + "soft_pair_count": 28, + "serialized_chars": 35663, + "estimated_input_tokens": 8916, + "estimated_eval_tokens": 18088 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": true, + "analyst_result": { + "summary": "This cluster is heterogeneous: the soft-similarity pairs mostly link unrelated bugs across models, tokenizers, loading, quantization, and training. I would not merge any of the candidate soft edges.", + "confidence": 0.93, + "canonical_issue_reason": "No true duplicate stands out; if a representative is needed, issue 43873 is the most central/open discussion-heavy item, but it is not a clear canonical duplicate for the cluster.", + "canonical_pr_reason": null, + "best_issue_reason": "Issue 43873 is the broadest and most referenced open report in the set, so it is the best single representative issue, though the cluster itself is not a duplicate group.", + "best_pr_reason": null, + "soft_edge_verdicts": [ + { + "left": "issue:43746", + "right": "issue:46209", + "accept": false, + "reason": "Different failure modes: PEFT local checkpoint loading vs FP8/B200 inference on a specific GPU architecture." + }, + { + "left": "issue:43927", + "right": "issue:45698", + "accept": false, + "reason": "Both involve loading/saving, but one is custom token IDs being lost in config serialization, the other is the wrong custom module being loaded after save_pretrained." + }, + { + "left": "issue:37428", + "right": "issue:42175", + "accept": false, + "reason": "Unrelated: a flash-attention import error versus TensorFlow missing from a torch-only extra install." + }, + { + "left": "issue:44561", + "right": "issue:44779", + "accept": false, + "reason": "Different subsystems and symptoms: a removed utility breaking trust_remote_code imports versus tokenizer output regressions for DeepSeek." + }, + { + "left": "issue:43873", + "right": "issue:45229", + "accept": false, + "reason": "Both mention quantization/memory, but one is offloading not working and the other is multi-GPU CUDA OOM for Gemma4." + }, + { + "left": "issue:32090", + "right": "issue:33290", + "accept": false, + "reason": "Trainer NoneType broadcast error and Adafactor/DeepSpeed OOM are not the same underlying bug." + }, + { + "left": "issue:25251", + "right": "issue:32090", + "accept": false, + "reason": "Pipeline top_k nesting change is unrelated to the Trainer broadcast TypeError." + }, + { + "left": "issue:41720", + "right": "issue:45229", + "accept": false, + "reason": "Different models and root causes: Qwen3 auto device map assert vs Gemma4 multi-GPU OOM." + }, + { + "left": "issue:43976", + "right": "issue:45440", + "accept": false, + "reason": "Python version compatibility issue versus a DeepSeekV3 implementation divergence; not the same defect." + }, + { + "left": "issue:44448", + "right": "issue:44466", + "accept": false, + "reason": "Both are v5 regressions, but one is generation/output mismatch and the other is serialization of lm_head.weight." + }, + { + "left": "issue:42490", + "right": "issue:43873", + "accept": false, + "reason": "Local-path slash import warning is unrelated to quantization offloading behavior." + }, + { + "left": "issue:44568", + "right": "issue:45491", + "accept": false, + "reason": "Tokenizer special-token handling and Gemma3 sliding-window NaN embeddings are distinct bugs." + }, + { + "left": "issue:45588", + "right": "issue:45823", + "accept": false, + "reason": "Flash-attention s_aux None crash and Gemma4 PLE device mismatch are different code paths and fixes." + }, + { + "left": "issue:43278", + "right": "issue:43575", + "accept": false, + "reason": "Embedding dtype regression in eval is unrelated to TP loading OOM for Qwen2-57B." + }, + { + "left": "issue:44977", + "right": "issue:45563", + "accept": false, + "reason": "Flash-attention generation failure and a stale warning for num_return_sequences are not the same issue." + }, + { + "left": "issue:33357", + "right": "issue:39692", + "accept": false, + "reason": "MacOS CLIP bus error is unrelated to SigLIP2 documentation example errors." + }, + { + "left": "issue:45372", + "right": "issue:45588", + "accept": false, + "reason": "Processor import failure due to mistral_common dependency is unrelated to the flash_attention.py s_aux=None crash." + }, + { + "left": "issue:29127", + "right": "issue:34567", + "accept": false, + "reason": "LayoutLMv3 error-message clarity and TrainerState token counting are unrelated." + }, + { + "left": "issue:45198", + "right": "issue:45803", + "accept": false, + "reason": "wav2vec2 save/tokenization failure is unrelated to FuyuBatchFeature's bare except swallowing errors." + }, + { + "left": "issue:43412", + "right": "issue:45588", + "accept": false, + "reason": "Executorch export issues for gemma3n are a different path from flash-attention integration crashes." + }, + { + "left": "issue:43381", + "right": "issue:43701", + "accept": false, + "reason": "Gradient checkpointing in eval mode and resume_from_checkpoint key mismatch are separate training issues." + }, + { + "left": "issue:43653", + "right": "issue:44743", + "accept": false, + "reason": "Tokenizer special-token registration and recurrent state reset in Qwen3.5 are unrelated." + }, + { + "left": "issue:44871", + "right": "issue:44977", + "accept": false, + "reason": "Gemma eos_token_id config mismatch is unrelated to Qwen3.5 flash-attention generation failures." + }, + { + "left": "issue:43611", + "right": "issue:46121", + "accept": false, + "reason": "Base model prefix loading regression and rope-params dict conversion TypeError are different loading/config bugs." + }, + { + "left": "issue:44361", + "right": "issue:45198", + "accept": false, + "reason": "Two tokenizer-related bugs, but for different models and different failure points." + }, + { + "left": "issue:44186", + "right": "issue:45198", + "accept": false, + "reason": "LayoutLMv2 padding/NER crashes are unrelated to wav2vec2 save_pretrained/tokenization failures." + }, + { + "left": "issue:45372", + "right": "issue:45406", + "accept": false, + "reason": "Both involve Gemma4 processor loading, but one is an import dependency issue and the other is a missing _tokenizer attribute on serve; not the same bug." + }, + { + "left": "issue:42915", + "right": "issue:42994", + "accept": false, + "reason": "Quantization is the common theme, but one is FP8 inference failure for Qwen3Moe and the other is quantized model saving failure." + } + ] + }, + "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": 21, + "review_unit_index": 31, + "review_unit_count": 56, + "status": "reviewed", + "reason": null, + "source": "llm", + "cache_hit": false, + "model": "gpt-5.4-mini?service_tier=flex", + "cluster_id": "cluster-10000-473", + "nodes": [ + "issue:27957", + "issue:29127", + "issue:29942", + "issue:30064", + "issue:33290", + "issue:33357", + "issue:35141", + "issue:42490", + "issue:42915", + "issue:43065", + "issue:43066", + "issue:43381", + "issue:43421", + "issue:43526", + "issue:43575", + "issue:43653", + "issue:43701", + "issue:43756", + "issue:43824", + "issue:43931", + "issue:43994", + "issue:44186", + "issue:44206", + "issue:44265", + "issue:44279", + "issue:44387", + "issue:44442", + "issue:44462", + "issue:44479", + "issue:45072", + "issue:45198", + "issue:45245", + "issue:45250", + "issue:45372", + "issue:45412", + "issue:45440", + "issue:45459", + "issue:45491", + "issue:45563", + "issue:45593", + "issue:45712", + "issue:45834", + "issue:46121", + "issue:46369", + "issue:46399", + "issue:46640", + "issue:46655" + ], + "soft_pairs": [ + "issue:44206|issue:44479", + "issue:29942|issue:30064", + "issue:43065|issue:45412", + "issue:43065|issue:43994", + "issue:45459|issue:45712", + "issue:43756|issue:45440", + "issue:43653|issue:44442", + "issue:45245|issue:45563", + "issue:45372|issue:45459", + "issue:45412|issue:46369", + "issue:29127|issue:33357", + "issue:43994|issue:45412", + "issue:43066|issue:44462", + "issue:27957|issue:29942", + "issue:44265|issue:45198", + "issue:45440|issue:45593", + "issue:43575|issue:43701", + "issue:29127|issue:35141", + "issue:43421|issue:45491", + "issue:42490|issue:43381", + "issue:43526|issue:45834", + "issue:43575|issue:44387", + "issue:44442|issue:45072", + "issue:42915|issue:43931", + "issue:43526|issue:46655", + "issue:43526|issue:44186", + "issue:43994|issue:46369", + "issue:43824|issue:44279", + "issue:27957|issue:29127", + "issue:46121|issue:46399", + "issue:45250|issue:46640", + "issue:27957|issue:33290" + ], + "prepared_review_unit_hash": "e44ffe76a6a16014ecdf3475ecc5c6ba6e29ac9a5632a3c83b29fd7e6dde0428", + "component_budget": { + "node_count": 473, + "item_count": 473, + "soft_pair_count": 1738, + "serialized_chars": 901437, + "estimated_input_tokens": 225360, + "estimated_eval_tokens": 450976 + }, + "budget": { + "node_count": 47, + "item_count": 47, + "soft_pair_count": 32, + "serialized_chars": 36536, + "estimated_input_tokens": 9134, + "estimated_eval_tokens": 18524 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": true, + "analyst_result": { + "summary": "This cluster is heterogeneous: most items are distinct model-specific bugs, regressions, or feature requests. I found no soft-edge pair that looks like the same underlying bug/change, so none should be merged as duplicates.", + "confidence": 0.91, + "canonical_issue_reason": "No single canonical issue cleanly represents this mixed cluster. If one representative is needed, issue:44479 is the broadest concrete regression report, but it is not a duplicate of the others.", + "canonical_pr_reason": null, + "best_issue_reason": "issue:44479 is the best overall representative because it describes a concrete runtime regression with a clear failure mode and multiple affected model families; however, the cluster is too diverse to treat it as a true canonical for all items.", + "best_pr_reason": null, + "soft_edge_verdicts": [ + { + "left": "issue:44206", + "right": "issue:44479", + "accept": false, + "reason": "Different regressions in different code paths: LaserFeatureExtractor crash vs video-input regression for Qwen VL models." + }, + { + "left": "issue:29942", + "right": "issue:30064", + "accept": false, + "reason": "Flash Attention test failures and void segmentation map processing are unrelated subsystems and symptoms." + }, + { + "left": "issue:43065", + "right": "issue:45412", + "accept": false, + "reason": "Dummy Conv2d placeholder in Sam3PixelDecoder is unrelated to RT-DETR model memory leak on deletion." + }, + { + "left": "issue:43065", + "right": "issue:43994", + "accept": false, + "reason": "Sam3PixelDecoder dummy layer issue is not the same bug as Siglip2 producing nonsensical outputs." + }, + { + "left": "issue:45459", + "right": "issue:45712", + "accept": false, + "reason": "Tokenizer error masking with missing protobuf is unrelated to leftover dummy classes in dummy_pt_objects.py." + }, + { + "left": "issue:43756", + "right": "issue:45440", + "accept": false, + "reason": "Missing RoPE layers in Smollm3 and DeepseekV3MoE divergence are different model implementation problems." + }, + { + "left": "issue:43653", + "right": "issue:44442", + "accept": false, + "reason": "BigBirdTokenizer special-token registration and FastSpeech2ConformerTokenizer loading are different tokenizer bugs." + }, + { + "left": "issue:45245", + "right": "issue:45563", + "accept": false, + "reason": "Category-cardinality runtime error is unrelated to a stale generate() warning." + }, + { + "left": "issue:45372", + "right": "issue:45459", + "accept": false, + "reason": "Missing ReasoningEffort import and protobuf-related tokenizer error handling are distinct failures." + }, + { + "left": "issue:45412", + "right": "issue:46369", + "accept": false, + "reason": "RT-DETR garbage-collection memory leak is not the same as sam3-litetext memory overflow." + }, + { + "left": "issue:29127", + "right": "issue:33357", + "accept": false, + "reason": "LayoutLMv3 error-message clarity and a MacOS CLIP bus error are unrelated." + }, + { + "left": "issue:43994", + "right": "issue:45412", + "accept": false, + "reason": "Siglip2 wrong outputs and RT-DETR memory retention are different bugs." + }, + { + "left": "issue:43066", + "right": "issue:44462", + "accept": false, + "reason": "Wrong tokenizer decoder type in v5 is not the same as AutoTokenizer ignoring tokenizer.json." + }, + { + "left": "issue:27957", + "right": "issue:29942", + "accept": false, + "reason": "XLMRoberta Flash Attention 2 support and Flash Attention 2 test failures are not the same concrete bug." + }, + { + "left": "issue:44265", + "right": "issue:45198", + "accept": false, + "reason": "torch.export failure via torch_compilable_check is unrelated to Wav2Vec2 save_pretrained/tokenization failure." + }, + { + "left": "issue:45440", + "right": "issue:45593", + "accept": false, + "reason": "DeepseekV3MoE divergence and D-FINE auxiliary-loss behavior are unrelated model issues." + }, + { + "left": "issue:43575", + "right": "issue:43701", + "accept": false, + "reason": "Tensor-parallel OOM when loading Qwen2-57B and resume_from_checkpoint key mismatch are different problems." + }, + { + "left": "issue:29127", + "right": "issue:35141", + "accept": false, + "reason": "LayoutLMv3 error clarity and output embedding reinitialization after resizing are unrelated." + }, + { + "left": "issue:43421", + "right": "issue:45491", + "accept": false, + "reason": "Runtime special-token post-processor update bug and Gemma3 sliding-window NaNs are different issues." + }, + { + "left": "issue:42490", + "right": "issue:43381", + "accept": false, + "reason": "Trailing-slash import warning and gradient-checkpointing-in-eval-mode are unrelated." + }, + { + "left": "issue:43526", + "right": "issue:45834", + "accept": false, + "reason": "BeitImageProcessorFast reduce_labels bug and Kosmos2.5 OCR index error are different image-processing failures." + }, + { + "left": "issue:43575", + "right": "issue:44387", + "accept": false, + "reason": "Qwen2 tensor-parallel OOM and int4 quantization reserved-memory regression are not the same code-path problem." + }, + { + "left": "issue:44442", + "right": "issue:45072", + "accept": false, + "reason": "Tokenizer loading failure and bfloat16 dtype mismatch are unrelated." + }, + { + "left": "issue:42915", + "right": "issue:43931", + "accept": false, + "reason": "FineGrainedFP8Config failure for Qwen3Moe and Qwen3-VL weight-shape mismatch are distinct loading bugs." + }, + { + "left": "issue:43526", + "right": "issue:46655", + "accept": false, + "reason": "reduce_labels returning one label and SmolVLM video padding calculation are unrelated." + }, + { + "left": "issue:43526", + "right": "issue:44186", + "accept": false, + "reason": "BeitImageProcessorFast reduce_labels bug and LayoutLMv2Tokenizer NER padding/truncation crash are different tokenizer/processor issues." + }, + { + "left": "issue:43994", + "right": "issue:46369", + "accept": false, + "reason": "Siglip2 bad outputs and sam3-litetext memory overflow are unrelated." + }, + { + "left": "issue:43824", + "right": "issue:44279", + "accept": false, + "reason": "Missing Qwen2_5_VL import and a generic dependency issue are not the same concrete failure." + }, + { + "left": "issue:27957", + "right": "issue:29127", + "accept": false, + "reason": "Flash Attention 2 support for XLMRoberta and LayoutLMv3 error messaging are unrelated." + }, + { + "left": "issue:46121", + "right": "issue:46399", + "accept": false, + "reason": "Rope-parameter JSON list handling and missing decorator for apply_rotary_pos_emb registration are different bugs." + }, + { + "left": "issue:45250", + "right": "issue:46640", + "accept": false, + "reason": "Both mention Flash Attention 2, but one is a vague title and the other is a distinct issue; no shared concrete bug is evident." + }, + { + "left": "issue:27957", + "right": "issue:33290", + "accept": false, + "reason": "Flash Attention 2 model support and Adafactor-in-Deepspeed 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": 21, + "review_unit_index": 32, + "review_unit_count": 56, + "status": "reviewed", + "reason": null, + "source": "llm", + "cache_hit": false, + "model": "gpt-5.4-mini?service_tier=flex", + "cluster_id": "cluster-10000-473", + "nodes": [ + "issue:33290", + "issue:34567", + "issue:34689", + "issue:35141", + "issue:42175", + "issue:42490", + "issue:43232", + "issue:43262", + "issue:43316", + "issue:43334", + "issue:43381", + "issue:43388", + "issue:43493", + "issue:43525", + "issue:43531", + "issue:43611", + "issue:43646", + "issue:43720", + "issue:43723", + "issue:43761", + "issue:43827", + "issue:43873", + "issue:44062", + "issue:44077", + "issue:44079", + "issue:44188", + "issue:44206", + "issue:44442", + "issue:44464", + "issue:44479", + "issue:44743", + "issue:44811", + "issue:44871", + "issue:44898", + "issue:44977", + "issue:44987", + "issue:45003", + "issue:45245", + "issue:45440", + "issue:45479", + "issue:45491", + "issue:45588", + "issue:45600", + "issue:45859", + "issue:45907", + "issue:46178", + "issue:46234" + ], + "soft_pairs": [ + "issue:43334|issue:43531", + "issue:33290|issue:34689", + "issue:43761|issue:45859", + "issue:33290|issue:35141", + "issue:45588|issue:45600", + "issue:33290|issue:34567", + "issue:43262|issue:44206", + "issue:44977|issue:45245", + "issue:43381|issue:43873", + "issue:44464|issue:45491", + "issue:43646|issue:44077", + "issue:44743|issue:44811", + "issue:44442|issue:44898", + "issue:42490|issue:44743", + "issue:34689|issue:43611", + "issue:45003|issue:45440", + "issue:43720|issue:44442", + "issue:43723|issue:45479", + "issue:43316|issue:43525", + "issue:42175|issue:46234", + "issue:43493|issue:43827", + "issue:44077|issue:44479", + "issue:43761|issue:44079", + "issue:44062|issue:45907", + "issue:43873|issue:44871", + "issue:43761|issue:44188", + "issue:43232|issue:43388", + "issue:44987|issue:46178" + ], + "prepared_review_unit_hash": "88569fcddcd84f169ba15ec080bf52cd28d2530a6188e5946d6739f4a7f4e804", + "component_budget": { + "node_count": 473, + "item_count": 473, + "soft_pair_count": 1738, + "serialized_chars": 901437, + "estimated_input_tokens": 225360, + "estimated_eval_tokens": 450976 + }, + "budget": { + "node_count": 47, + "item_count": 47, + "soft_pair_count": 28, + "serialized_chars": 35703, + "estimated_input_tokens": 8926, + "estimated_eval_tokens": 18108 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": true, + "analyst_result": { + "summary": "This cluster is heterogeneous and does not look like a true duplicate group; the soft-similarity pairs mostly connect unrelated bugs that only share broad subsystem keywords.", + "confidence": 0.94, + "canonical_issue_reason": "No clear canonical issue: the items span unrelated regressions and API bugs across model loading, generation, tokenizers, configs, audio, quantization, and docs, so there is no safe duplicate anchor.", + "canonical_pr_reason": "No PRs are present in this cluster.", + "best_issue_reason": "No single issue is globally representative enough to serve as a cluster-wide anchor; the closest matches are still only broad thematic overlaps, not the same underlying defect.", + "best_pr_reason": "No PRs are present in this cluster.", + "soft_edge_verdicts": [ + { + "left": "issue:43334", + "right": "issue:43531", + "accept": false, + "reason": "Both mention Qwen3-family configs, but one is a missing `pad_token_id` on Qwen3-VL and the other is a `sliding_window` issue in Qwen3-MoE; different bugs and code paths." + }, + { + "left": "issue:33290", + "right": "issue:34689", + "accept": false, + "reason": "OOM during Adafactor+DeepSpeed training and Llama 3.2 Vision model loading failure are unrelated." + }, + { + "left": "issue:43761", + "right": "issue:45859", + "accept": false, + "reason": "CLIPVisionModel hidden-states regression and Qwen3_5Moe CPU offload incompatibility are different model families and failure modes." + }, + { + "left": "issue:33290", + "right": "issue:35141", + "accept": false, + "reason": "Training OOM with DeepSpeed is not the same as token-embedding resize causing output embedding reinitialization." + }, + { + "left": "issue:45588", + "right": "issue:45600", + "accept": false, + "reason": "Flash-attention `s_aux=None` crash and stale auto-mapping references to removed configs are separate issues." + }, + { + "left": "issue:33290", + "right": "issue:34567", + "accept": false, + "reason": "An optimizer OOM and `num_input_tokens_seen` not updating are unrelated trainer concerns." + }, + { + "left": "issue:43262", + "right": "issue:44206", + "accept": false, + "reason": "Both are audio-related, but one is chat template sampling-rate defaulting and the other is an unsupported `center` argument in a feature extractor." + }, + { + "left": "issue:44977", + "right": "issue:45245", + "accept": false, + "reason": "Qwen3.5 flash-attention generation regression and a category-count runtime limit are unrelated." + }, + { + "left": "issue:43381", + "right": "issue:43873", + "accept": false, + "reason": "Gradient checkpointing in eval mode and offloading with quantization do not describe the same defect." + }, + { + "left": "issue:44464", + "right": "issue:45491", + "accept": false, + "reason": "Both affect generation, but chunked generation with compiled forward and Gemma3 NaN embeddings under sliding-window/all-padding windows are distinct bugs." + }, + { + "left": "issue:43646", + "right": "issue:44077", + "accept": false, + "reason": "A broad v5 custom-init breakage is not the same as patchtsmixer-specific `post_init` policy enforcement." + }, + { + "left": "issue:44743", + "right": "issue:44811", + "accept": false, + "reason": "Recurrent-state reset in Qwen3.5-MoE and Whisper `batch_decode()` skipping tokens are unrelated." + }, + { + "left": "issue:44442", + "right": "issue:44898", + "accept": false, + "reason": "Tokenizer auto-load failure and Perceiver image-classification resize failure are different components and failure modes." + }, + { + "left": "issue:42490", + "right": "issue:44743", + "accept": false, + "reason": "A local-path trailing-slash import warning is not the same as Qwen3.5 recurrent state reset." + }, + { + "left": "issue:34689", + "right": "issue:43611", + "accept": false, + "reason": "Both are model-loading regressions, but they involve different root causes: a specific Llama vision breakage versus `base_model_prefix` handling in v5.0.0." + }, + { + "left": "issue:45003", + "right": "issue:45440", + "accept": false, + "reason": "Unsafe `sys.modules[]` access in modeling utils and DeepseekV3MoE divergence from the remote implementation are unrelated." + }, + { + "left": "issue:43720", + "right": "issue:44442", + "accept": false, + "reason": "Packed-weight unpacking during accelerate loading and tokenizer auto-loading are unrelated bugs." + }, + { + "left": "issue:43723", + "right": "issue:45479", + "accept": false, + "reason": "Tokenizer loading failure in v5 and zero-loss single-label classification are different subsystems." + }, + { + "left": "issue:43316", + "right": "issue:43525", + "accept": false, + "reason": "Both are config API complaints, but one is a general Gemma3TextConfig discrepancy and the other is a concrete missing `pad_token_id` on Llama4Config; not the same bug." + }, + { + "left": "issue:42175", + "right": "issue:46234", + "accept": false, + "reason": "Missing TensorFlow in a torch extra and a missing import_utils availability helper are unrelated packaging/import issues." + }, + { + "left": "issue:43493", + "right": "issue:43827", + "accept": false, + "reason": "A SigLIP2 implementation mismatch and docs still referencing removed pipelines are completely different kinds of issues." + }, + { + "left": "issue:44077", + "right": "issue:44479", + "accept": false, + "reason": "Patchtsmixer `post_init` policy and Qwen VL video-input regression are unrelated." + }, + { + "left": "issue:43761", + "right": "issue:44079", + "accept": false, + "reason": "Hidden-states not returned from CLIPVisionModel and incorrect `ModelOutput` key assignment are both output-related, but they are different layers and not the same defect." + }, + { + "left": "issue:44062", + "right": "issue:45907", + "accept": false, + "reason": "`AddedToken()` duplicate `special` argument and `list_repo_files` not forwarding `revision` are unrelated API bugs." + }, + { + "left": "issue:43873", + "right": "issue:44871", + "accept": false, + "reason": "Quantization/offloading behavior and Gemma-3 `eos_token_id` inconsistency do not share the same root cause." + }, + { + "left": "issue:43761", + "right": "issue:44188", + "accept": false, + "reason": "A CLIPVisionModel output regression and attention-kernel divergence under `torch.compile` are unrelated." + }, + { + "left": "issue:43232", + "right": "issue:43388", + "accept": false, + "reason": "Generation kwargs sync behavior and `gather_for_metrics()` dropping tuple labels are different distributed-training bugs." + }, + { + "left": "issue:44987", + "right": "issue:46178", + "accept": false, + "reason": "A model-loading regression for `physical-intelligence/fast` and a TensorFlow Text backend mapping error 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": 21, + "review_unit_index": 33, + "review_unit_count": 56, + "status": "reviewed", + "reason": null, + "source": "llm", + "cache_hit": false, + "model": "gpt-5.4-mini?service_tier=flex", + "cluster_id": "cluster-10000-473", + "nodes": [ + "issue:16998", + "issue:22355", + "issue:25251", + "issue:27957", + "issue:28282", + "issue:32090", + "issue:42175", + "issue:42617", + "issue:42907", + "issue:43295", + "issue:43316", + "issue:43452", + "issue:43550", + "issue:43575", + "issue:43606", + "issue:43643", + "issue:43646", + "issue:43701", + "issue:43716", + "issue:43742", + "issue:43749", + "issue:43825", + "issue:43931", + "issue:44077", + "issue:44162", + "issue:44190", + "issue:44442", + "issue:44464", + "issue:44496", + "issue:44534", + "issue:44610", + "issue:44792", + "issue:44855", + "issue:44871", + "issue:44913", + "issue:45137", + "issue:45216", + "issue:45584", + "issue:45698", + "issue:45759", + "issue:45803", + "issue:46050", + "issue:46234", + "issue:46291", + "issue:46365", + "issue:46369", + "issue:46421", + "issue:46506" + ], + "soft_pairs": [ + "issue:43749|issue:43931", + "issue:43931|issue:45216", + "issue:43316|issue:45759", + "issue:44855|issue:45137", + "issue:43643|issue:44913", + "issue:44442|issue:45803", + "issue:43701|issue:46506", + "issue:44190|issue:44496", + "issue:42907|issue:45698", + "issue:44464|issue:46421", + "issue:42617|issue:43575", + "issue:25251|issue:28282", + "issue:27957|issue:32090", + "issue:43550|issue:43701", + "issue:43575|issue:43742", + "issue:43606|issue:43701", + "issue:46234|issue:46291", + "issue:43295|issue:43825", + "issue:16998|issue:22355", + "issue:43295|issue:44871", + "issue:44871|issue:45584", + "issue:43316|issue:43716", + "issue:42175|issue:43316", + "issue:44077|issue:44534", + "issue:45803|issue:46365", + "issue:44871|issue:46050", + "issue:43646|issue:44913", + "issue:44610|issue:46369", + "issue:44162|issue:44792", + "issue:43452|issue:44496" + ], + "prepared_review_unit_hash": "69e809013a87b331cdb613f1dbe6fdc07b64c6f91c7ae5de984909a1a000205e", + "component_budget": { + "node_count": 473, + "item_count": 473, + "soft_pair_count": 1738, + "serialized_chars": 901437, + "estimated_input_tokens": 225360, + "estimated_eval_tokens": 450976 + }, + "budget": { + "node_count": 48, + "item_count": 48, + "soft_pair_count": 30, + "serialized_chars": 36481, + "estimated_input_tokens": 9121, + "estimated_eval_tokens": 18498 + }, + "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 that only share broad themes like loading, config serialization, or model execution. None of the proposed soft edges look like the same underlying bug/change, so I would not merge any of them.", + "confidence": 0.88, + "canonical_issue_reason": "issue:45698 is the broadest and most representative open issue in this set: it centers on incorrect module resolution during save_pretrained/from_pretrained, a core loading/serialization path that overlaps the general theme of several nearby reports.", + "canonical_pr_reason": null, + "best_issue_reason": "issue:45698 has the strongest triage value: open, high discussion, multiple inbound references, and a broad user-facing failure mode around custom-module loading after save/load.", + "best_pr_reason": null, + "soft_edge_verdicts": [ + { + "left": "issue:43749", + "right": "issue:43931", + "accept": false, + "reason": "Different failure modes: FSDP CPU RAM-efficient loading vs Qwen3-VL weight-shape mismatch during model load." + }, + { + "left": "issue:43931", + "right": "issue:45216", + "accept": false, + "reason": "Both involve Qwen family loading/saving, but one is a shape mismatch on load and the other is an incorrect checkpoint save_pretrained regression." + }, + { + "left": "issue:43316", + "right": "issue:45759", + "accept": false, + "reason": "Both touch config APIs, but Gemma3TextConfig inconsistency and Qwen multimodal text_config unwrapping are distinct bugs." + }, + { + "left": "issue:44855", + "right": "issue:45137", + "accept": false, + "reason": "Completely different subsystems: Python 3.13 JIT parsing/import issue vs DeepSpeed ZeRO3 deque error." + }, + { + "left": "issue:43643", + "right": "issue:44913", + "accept": false, + "reason": "Both are config reload issues, but one is trust_remote_code missing fields and the other is GPTNeoX rotary_pct not persisting." + }, + { + "left": "issue:44442", + "right": "issue:45803", + "accept": false, + "reason": "Tokenizer loading failure and a bare-except bug in Fuyu tensor conversion are unrelated." + }, + { + "left": "issue:43701", + "right": "issue:46506", + "accept": false, + "reason": "Checkpoint key mismatch on resume and CPU/CUDA output differences in PPDocLayoutV3 are not the same bug." + }, + { + "left": "issue:44190", + "right": "issue:44496", + "accept": false, + "reason": "Dataset loading in an example script is unrelated to unrecognized model/config loading." + }, + { + "left": "issue:42907", + "right": "issue:45698", + "accept": false, + "reason": "Both involve save/load paths, but dequantized model saving and wrong custom-module resolution are different root causes." + }, + { + "left": "issue:44464", + "right": "issue:46421", + "accept": false, + "reason": "Both mention chunked behavior, but one is incorrect generation output and the other is spurious recompilation from cache initialization." + }, + { + "left": "issue:42617", + "right": "issue:43575", + "accept": false, + "reason": "3D parallel script runtime issues and Qwen tp OOM are different execution problems." + }, + { + "left": "issue:25251", + "right": "issue:28282", + "accept": false, + "reason": "Pipeline top_k output-shape change is unrelated to missing PyTorch import errors." + }, + { + "left": "issue:27957", + "right": "issue:32090", + "accept": false, + "reason": "Flash Attention support for XLMRoberta and a Trainer GPU broadcast TypeError are unrelated." + }, + { + "left": "issue:43550", + "right": "issue:43701", + "accept": false, + "reason": "torch.compile/SDPA failure and resume_from_checkpoint key mismatch do not share a concrete code path." + }, + { + "left": "issue:43575", + "right": "issue:43742", + "accept": false, + "reason": "OOM while loading a large Qwen model and a key error loading MobileLLM are different load-time failures." + }, + { + "left": "issue:43606", + "right": "issue:43701", + "accept": false, + "reason": "CPU offload device mismatch is unrelated to checkpoint key mismatch on resume." + }, + { + "left": "issue:46234", + "right": "issue:46291", + "accept": false, + "reason": "Missing import_utils symbol and kernels version/revision validation are distinct dependency issues." + }, + { + "left": "issue:43295", + "right": "issue:43825", + "accept": false, + "reason": "Custom processor/tokenizer regression and pipeline task-support error messaging are different behaviors." + }, + { + "left": "issue:16998", + "right": "issue:22355", + "accept": false, + "reason": "model_max_length question for DeBERTa-V3 and missing transformers.onnx module are unrelated." + }, + { + "left": "issue:43295", + "right": "issue:44871", + "accept": false, + "reason": "Processor/tokenizer regression is unrelated to Gemma-3 eos_token_id inconsistency." + }, + { + "left": "issue:44871", + "right": "issue:45584", + "accept": false, + "reason": "Gemma-3 eos_token_id config mismatch and Whisper empty transcription after align_special_tokens are different issues." + }, + { + "left": "issue:43316", + "right": "issue:43716", + "accept": false, + "reason": "Gemma3TextConfig API discrepancy and Mistral-3 image preprocessor dtype mismatch are unrelated." + }, + { + "left": "issue:42175", + "right": "issue:43316", + "accept": false, + "reason": "Packaging backend omissions and config API discrepancies are not the same bug." + }, + { + "left": "issue:44077", + "right": "issue:44534", + "accept": false, + "reason": "Optional post_init enforcement for patchtsmixer is unrelated to non-persistent buffer initialization bugs." + }, + { + "left": "issue:45803", + "right": "issue:46365", + "accept": false, + "reason": "Bare except in Fuyu tensor conversion and Trainer import failure due to datasets are unrelated." + }, + { + "left": "issue:44871", + "right": "issue:46050", + "accept": false, + "reason": "Gemma-3 eos_token_id mismatch and quantized model save/load corruption are different serialization problems." + }, + { + "left": "issue:43646", + "right": "issue:44913", + "accept": false, + "reason": "Transformers 5 custom model init breakage and GPTNeoXConfig rotary_pct persistence are different config/model-init issues." + }, + { + "left": "issue:44610", + "right": "issue:46369", + "accept": false, + "reason": "OmDet-Turbo input-size mismatch and sam3-litetext memory overflow are unrelated model execution failures." + }, + { + "left": "issue:44162", + "right": "issue:44792", + "accept": false, + "reason": "ESM2 breakage and janus test failure are unrelated model-specific regressions." + }, + { + "left": "issue:43452", + "right": "issue:44496", + "accept": false, + "reason": "gguf_file breaking from_pretrained calls and unrecognized model config errors do not share the same root cause." + } + ] + }, + "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": 21, + "review_unit_index": 34, + "review_unit_count": 56, + "status": "reviewed", + "reason": null, + "source": "llm", + "cache_hit": false, + "model": "gpt-5.4-mini?service_tier=flex", + "cluster_id": "cluster-10000-473", + "nodes": [ + "issue:35141", + "issue:41950", + "issue:42490", + "issue:43065", + "issue:43352", + "issue:43388", + "issue:43404", + "issue:43441", + "issue:43475", + "issue:43504", + "issue:43525", + "issue:43531", + "issue:43540", + "issue:43572", + "issue:43650", + "issue:43673", + "issue:43688", + "issue:43761", + "issue:44190", + "issue:44479", + "issue:44514", + "issue:44534", + "issue:44617", + "issue:44734", + "issue:44743", + "issue:44792", + "issue:44855", + "issue:44913", + "issue:44964", + "issue:44977", + "issue:45127", + "issue:45200", + "issue:45406", + "issue:45542", + "issue:45584", + "issue:45588", + "issue:45600", + "issue:45632", + "issue:45698", + "issue:45803", + "issue:46170", + "issue:46178", + "issue:46202", + "issue:46234", + "issue:46424", + "issue:46617", + "issue:46640" + ], + "soft_pairs": [ + "issue:45632|issue:46424", + "issue:45803|issue:46170", + "issue:43650|issue:46202", + "issue:43404|issue:43441", + "issue:43065|issue:44617", + "issue:43540|issue:44514", + "issue:43761|issue:44479", + "issue:43572|issue:46178", + "issue:45127|issue:45200", + "issue:43388|issue:43688", + "issue:43352|issue:46640", + "issue:45542|issue:46234", + "issue:43504|issue:45698", + "issue:43673|issue:44743", + "issue:44792|issue:44977", + "issue:44792|issue:45600", + "issue:41950|issue:43475", + "issue:43525|issue:43531", + "issue:44734|issue:45406", + "issue:44190|issue:44964", + "issue:35141|issue:42490", + "issue:44534|issue:44913", + "issue:44617|issue:45584", + "issue:45632|issue:46617", + "issue:44855|issue:45588" + ], + "prepared_review_unit_hash": "58c96a3dcd1de070de2fabd73d3138c5f152577b7d0e870314747645ba00ee23", + "component_budget": { + "node_count": 473, + "item_count": 473, + "soft_pair_count": 1738, + "serialized_chars": 901437, + "estimated_input_tokens": 225360, + "estimated_eval_tokens": 450976 + }, + "budget": { + "node_count": 47, + "item_count": 47, + "soft_pair_count": 25, + "serialized_chars": 34668, + "estimated_input_tokens": 8667, + "estimated_eval_tokens": 17590 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": true, + "analyst_result": { + "summary": "The cluster is heterogeneous: most items are unrelated Transformers bugs across loading, cache, vision, flash-attention, and processor paths. None of the proposed soft pairs look like the same underlying change/bug closely enough to merge.", + "confidence": 0.94, + "canonical_issue_reason": "No single issue clearly acts as a true duplicate canonical here; the cluster is too mixed and the soft pairs are mostly broad similarity rather than the same bug.", + "canonical_pr_reason": null, + "best_issue_reason": "issue:45698 is the strongest standalone representative: it is detailed, active, has inbound references, and centers on a concrete loading regression rather than a narrow model-specific symptom.", + "best_pr_reason": null, + "soft_edge_verdicts": [ + { + "left": "issue:45632", + "right": "issue:46424", + "accept": false, + "reason": "Both involve cache/loading behavior, but one is a trust_remote_code path collision and the other is max_cache_length being ignored; different bugs." + }, + { + "left": "issue:45803", + "right": "issue:46170", + "accept": false, + "reason": "Unrelated failures: tensor conversion error handling vs Windows/ROCm torch.distributed import issues." + }, + { + "left": "issue:43650", + "right": "issue:46202", + "accept": false, + "reason": "No overlap beyond being user-facing bugs; one is a data request, the other is a Gemma tokenization roundtrip issue." + }, + { + "left": "issue:43404", + "right": "issue:43441", + "accept": false, + "reason": "Different concrete problems: weight tying in Mistral3 vs FlashAttention support in Ministral-3." + }, + { + "left": "issue:43065", + "right": "issue:44617", + "accept": false, + "reason": "Both mention SAM 3, but one is a dummy Conv2d implementation detail and the other is a CUDA OOM symptom." + }, + { + "left": "issue:43540", + "right": "issue:44514", + "accept": false, + "reason": "Both are multimodal input bugs, but they affect different models and different code paths (video processing vs batched chat template padding)." + }, + { + "left": "issue:43761", + "right": "issue:44479", + "accept": false, + "reason": "Different regression surfaces: CLIP hidden_states forwarding vs Qwen video-input generation regression." + }, + { + "left": "issue:43572", + "right": "issue:46178", + "accept": false, + "reason": "Unrelated: missing pad_token_idx in StableLmConfig vs an unknown tensorflow_text backend mapping error." + }, + { + "left": "issue:45127", + "right": "issue:45200", + "accept": false, + "reason": "One is LoRA/extended-vocabulary collapse with tied embeddings; the other is Gemma 4 mm_token_type_ids defaulting for text-only fine-tuning." + }, + { + "left": "issue:43388", + "right": "issue:43688", + "accept": false, + "reason": "Different training bugs: metric-gathering drops labels vs auxiliary-loss normalization in MoE models." + }, + { + "left": "issue:43352", + "right": "issue:46640", + "accept": false, + "reason": "Both mention Flash Attention 2, but one is a specific model not supporting it and the other is too vague to confirm the same issue." + }, + { + "left": "issue:45542", + "right": "issue:46234", + "accept": false, + "reason": "Different backend/import problems: TensorBoard without TensorFlow vs missing Falcon Mamba SSM availability helper." + }, + { + "left": "issue:43504", + "right": "issue:45698", + "accept": false, + "reason": "Both concern model loading, but one is a legacy preset field for BEiT and the other is selecting the wrong custom module after save_pretrained." + }, + { + "left": "issue:43673", + "right": "issue:44743", + "accept": false, + "reason": "Both involve cache state, but one is GenerationMixin cache missing in chunked prefill and the other is recurrent states resetting in modular_qwen3_5." + }, + { + "left": "issue:44792", + "right": "issue:44977", + "accept": false, + "reason": "Different models and symptoms: Janus image generation test failure vs Qwen3.5 flash-attention generation regression." + }, + { + "left": "issue:44792", + "right": "issue:45600", + "accept": false, + "reason": "No shared bug: Janus generation test failure vs stale auto_mappings entries for removed Sam3LiteText configs." + }, + { + "left": "issue:41950", + "right": "issue:43475", + "accept": false, + "reason": "Different SAM 3 paths: processor selection bug vs missing fpn_position_embeddings attribute on a video encoder output." + }, + { + "left": "issue:43525", + "right": "issue:43531", + "accept": false, + "reason": "Separate config regressions: missing pad_token_id in Llama4Config vs sliding_window behavior in Qwen3-MoE." + }, + { + "left": "issue:44734", + "right": "issue:45406", + "accept": false, + "reason": "Unrelated service/model issues: KV-cache continuation tensor indexing vs Gemma4Processor missing _tokenizer." + }, + { + "left": "issue:44190", + "right": "issue:44964", + "accept": false, + "reason": "Different loading failures: local dataset in example script vs a specific multimodal model failing to load on latest transformers." + }, + { + "left": "issue:35141", + "right": "issue:42490", + "accept": false, + "reason": "Different loading behaviors: embedding reinit after resize vs warning on local-model paths ending with a slash." + }, + { + "left": "issue:44534", + "right": "issue:44913", + "accept": false, + "reason": "One is non-persistent buffers being filled with junk; the other is GPTNeoXConfig rotary_pct not persisting on reload." + }, + { + "left": "issue:44617", + "right": "issue:45584", + "accept": false, + "reason": "Different symptoms and code paths: SAM 3 video OOM vs Whisper empty-transcription failure after align_special_tokens." + }, + { + "left": "issue:45632", + "right": "issue:46617", + "accept": false, + "reason": "Both are trust_remote_code cache-related, but one is a leaf-directory cache collision and the other is a FileNotFoundError from cache loading; not clearly the same bug." + }, + { + "left": "issue:44855", + "right": "issue:45588", + "accept": false, + "reason": "Completely different areas: Python 3.13 import parsing issue in DebertaV2 vs flash_attention.py crashing on s_aux=None." + } + ] + }, + "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": 21, + "review_unit_index": 35, + "review_unit_count": 56, + "status": "reviewed", + "reason": null, + "source": "llm", + "cache_hit": false, + "model": "gpt-5.4-mini?service_tier=flex", + "cluster_id": "cluster-10000-473", + "nodes": [ + "issue:25251", + "issue:27957", + "issue:39692", + "issue:41720", + "issue:43065", + "issue:43295", + "issue:43335", + "issue:43540", + "issue:43550", + "issue:43644", + "issue:43673", + "issue:43761", + "issue:43827", + "issue:43931", + "issue:43994", + "issue:44060", + "issue:44077", + "issue:44112", + "issue:44190", + "issue:44242", + "issue:44464", + "issue:44479", + "issue:44521", + "issue:44610", + "issue:44617", + "issue:44734", + "issue:44737", + "issue:44779", + "issue:44855", + "issue:44877", + "issue:44913", + "issue:44933", + "issue:44938", + "issue:45127", + "issue:45245", + "issue:45276", + "issue:45357", + "issue:45362", + "issue:45381", + "issue:45542", + "issue:45563", + "issue:45584", + "issue:45600", + "issue:45803", + "issue:46369", + "issue:46421", + "issue:46489", + "issue:46655" + ], + "soft_pairs": [ + "issue:43931|issue:44060", + "issue:43761|issue:44242", + "issue:43673|issue:46421", + "issue:43827|issue:43994", + "issue:45563|issue:45584", + "issue:45381|issue:46369", + "issue:44479|issue:44734", + "issue:43644|issue:44913", + "issue:25251|issue:27957", + "issue:44933|issue:45362", + "issue:43065|issue:44610", + "issue:41720|issue:45245", + "issue:43335|issue:44855", + "issue:43550|issue:44190", + "issue:44077|issue:44877", + "issue:44464|issue:45563", + "issue:45381|issue:46655", + "issue:43335|issue:44190", + "issue:44617|issue:46369", + "issue:45127|issue:45357", + "issue:44779|issue:46489", + "issue:39692|issue:44190", + "issue:44112|issue:45600", + "issue:44521|issue:45381", + "issue:43540|issue:44734", + "issue:44737|issue:45803", + "issue:43295|issue:45276", + "issue:44938|issue:45542", + "issue:45276|issue:45584" + ], + "prepared_review_unit_hash": "a504261e91e68be253ba349cfefc3ef20228e4445112de5f03c9f470a0bbb438", + "component_budget": { + "node_count": 473, + "item_count": 473, + "soft_pair_count": 1738, + "serialized_chars": 901437, + "estimated_input_tokens": 225360, + "estimated_eval_tokens": 450976 + }, + "budget": { + "node_count": 48, + "item_count": 48, + "soft_pair_count": 29, + "serialized_chars": 36346, + "estimated_input_tokens": 9087, + "estimated_eval_tokens": 18430 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": true, + "analyst_result": { + "summary": "This cluster is mostly a set of unrelated issue reports; nearly all soft edges are false positives. The only plausible duplicate/follow-up pair is the DeepSeek tokenizer v5 regression (44779/46489).", + "confidence": 0.89, + "canonical_issue_reason": "issue:44779 is the original, broader DeepSeek tokenizer v5 regression report and the clearest anchor for the only near-duplicate pair in this set.", + "canonical_pr_reason": null, + "best_issue_reason": "issue:44779 is the most representative issue here: concrete repro, broadest wording, and it anchors the only plausible duplicate relationship.", + "best_pr_reason": null, + "soft_edge_verdicts": [ + { + "left": "issue:43931", + "right": "issue:44060", + "accept": false, + "reason": "Different failure modes: Qwen3-VL weight-shape load error vs Qwen3-Next tied-weights warning." + }, + { + "left": "issue:43761", + "right": "issue:44242", + "accept": false, + "reason": "CLIPVision hidden_states regression and load-balancing-loss gating are unrelated code paths." + }, + { + "left": "issue:43673", + "right": "issue:46421", + "accept": false, + "reason": "Both mention chunked prefill/cache, but one is missing cache state and the other is recompilation from lazy init." + }, + { + "left": "issue:43827", + "right": "issue:43994", + "accept": false, + "reason": "Docs still referencing pipeline() is not the same bug as SigLIP2 producing bad outputs with AutoModel/pipeline." + }, + { + "left": "issue:45563", + "right": "issue:45584", + "accept": false, + "reason": "A stale generation warning is distinct from Whisper failing on empty transcription after align_special_tokens." + }, + { + "left": "issue:45381", + "right": "issue:46369", + "accept": false, + "reason": "Qwen2.5-VL video position-id regression and Sam3-LiteText memory overflow are different model-specific bugs." + }, + { + "left": "issue:44479", + "right": "issue:44734", + "accept": false, + "reason": "Video input regression and KV-cache continuation crash are separate failure points." + }, + { + "left": "issue:43644", + "right": "issue:44913", + "accept": false, + "reason": "Non-persistent buffer junk-fill and GPTNeoXConfig reload behavior are different issues." + }, + { + "left": "issue:25251", + "right": "issue:27957", + "accept": false, + "reason": "Pipeline output nesting and Flash Attention 2 support are unrelated." + }, + { + "left": "issue:44933", + "right": "issue:45362", + "accept": false, + "reason": "Missing import from image_utils is unrelated to Qwen3.5 chat crashes." + }, + { + "left": "issue:43065", + "right": "issue:44610", + "accept": false, + "reason": "Dummy Conv2d in Sam3PixelDecoder and OmDet-Turbo processor/model size mismatch are different bugs." + }, + { + "left": "issue:41720", + "right": "issue:45245", + "accept": false, + "reason": "CUDA assert on auto device mapping is unrelated to the 2^24 category limit error." + }, + { + "left": "issue:43335", + "right": "issue:44855", + "accept": false, + "reason": "SwitchTransformers sparse-layer construction and a Python 3.13 IndentationError are not the same problem." + }, + { + "left": "issue:43550", + "right": "issue:44190", + "accept": false, + "reason": "torch.compile+SDPA failure and local dataset loading are unrelated." + }, + { + "left": "issue:44077", + "right": "issue:44877", + "accept": false, + "reason": "Optional post_init on patchtsmixer and strict granite_speech config loading are separate config issues." + }, + { + "left": "issue:44464", + "right": "issue:45563", + "accept": false, + "reason": "Compiled-forward generation inconsistency is not the same as the stale num_return_sequences warning." + }, + { + "left": "issue:45381", + "right": "issue:46655", + "accept": false, + "reason": "Both are video bugs, but one is wrong vision_position_ids and the other is incorrect padding calculation." + }, + { + "left": "issue:43335", + "right": "issue:44190", + "accept": false, + "reason": "Sparse-layer config logic and a dataset-loading script failure are unrelated." + }, + { + "left": "issue:44617", + "right": "issue:46369", + "accept": false, + "reason": "OOM in Sam3Video and memory overflow in sam3-litetext are different symptoms on different variants." + }, + { + "left": "issue:45127", + "right": "issue:45357", + "accept": false, + "reason": "LoRA merge collapse with tied embeddings is unrelated to incorrect visual encoder keys on save_pretrained." + }, + { + "left": "issue:44779", + "right": "issue:46489", + "accept": true, + "reason": "Both describe the same DeepSeek tokenizer regression on v5+; 46489 reads like the remaining gap in the earlier fix rather than a separate bug." + }, + { + "left": "issue:39692", + "right": "issue:44190", + "accept": false, + "reason": "SigLIP2 documentation errors and a local dataset loading script issue are unrelated." + }, + { + "left": "issue:44112", + "right": "issue:45600", + "accept": false, + "reason": "Stale CI device-override test and removed Sam3LiteText auto_mapping references are different maintenance problems." + }, + { + "left": "issue:44521", + "right": "issue:45381", + "accept": false, + "reason": "All-zero assistant masks in apply_chat_template are unrelated to Qwen2.5-VL video position IDs." + }, + { + "left": "issue:43540", + "right": "issue:44734", + "accept": false, + "reason": "Qwen3OmniMoe video-input ValueError and serve KV-cache continuation indexing are unrelated." + }, + { + "left": "issue:44737", + "right": "issue:45803", + "accept": false, + "reason": "CPU arange/device placement bug in XLNet is unrelated to FuyuBatchFeature swallowing exceptions." + }, + { + "left": "issue:43295", + "right": "issue:45276", + "accept": false, + "reason": "processor.tokenizer regression and gemma4 embedding-resize propagation are different code paths." + }, + { + "left": "issue:44938", + "right": "issue:45542", + "accept": false, + "reason": "Python 3.14 import/load failure is unrelated to the undefined tf backend error." + }, + { + "left": "issue:45276", + "right": "issue:45584", + "accept": false, + "reason": "Embedding resize propagation and Whisper empty-transcription failure are unrelated." + } + ] + }, + "evaluator_result": { + "accept": true, + "feedback": "The summary is broadly grounded in the packet and stays conservative overall. The one accepted pair, 44779/46489, is supported by the titles as a plausible same-root DeepSeek tokenizer v5+ regression/follow-up. I don\u2019t see a clear overstatement strong enough to reject." + }, + "evaluator_used": true, + "retried": false, + "accepted_nontrivial_soft_edge": true, + "error_kind": null, + "error_message": null + }, + { + "label": "issue", + "component_index": 1, + "component_count": 21, + "review_unit_index": 36, + "review_unit_count": 56, + "status": "reviewed", + "reason": null, + "source": "llm", + "cache_hit": false, + "model": "gpt-5.4-mini?service_tier=flex", + "cluster_id": "cluster-10000-473", + "nodes": [ + "issue:27957", + "issue:32090", + "issue:35141", + "issue:36246", + "issue:38617", + "issue:41628", + "issue:41950", + "issue:42175", + "issue:42796", + "issue:43295", + "issue:43404", + "issue:43479", + "issue:43531", + "issue:43643", + "issue:43650", + "issue:43673", + "issue:43854", + "issue:43976", + "issue:44016", + "issue:44112", + "issue:44186", + "issue:44220", + "issue:44246", + "issue:44279", + "issue:44295", + "issue:44462", + "issue:44496", + "issue:44521", + "issue:44683", + "issue:44734", + "issue:44964", + "issue:44987", + "issue:45003", + "issue:45083", + "issue:45216", + "issue:45245", + "issue:45335", + "issue:45375", + "issue:45446", + "issue:45584", + "issue:45600", + "issue:45632", + "issue:45907", + "issue:46082", + "issue:46121", + "issue:46234", + "issue:46424", + "issue:46506" + ], + "soft_pairs": [ + "issue:44462|issue:45907", + "issue:45632|issue:46234", + "issue:44246|issue:45003", + "issue:46082|issue:46121", + "issue:42796|issue:44521", + "issue:45245|issue:45584", + "issue:44964|issue:45600", + "issue:42175|issue:44279", + "issue:44112|issue:44964", + "issue:43854|issue:44964", + "issue:43643|issue:44496", + "issue:27957|issue:35141", + "issue:36246|issue:41628", + "issue:44186|issue:44734", + "issue:43404|issue:45216", + "issue:36246|issue:38617", + "issue:44295|issue:46506", + "issue:43650|issue:44220", + "issue:41950|issue:44016", + "issue:32090|issue:35141", + "issue:44683|issue:45446", + "issue:43479|issue:45375", + "issue:43673|issue:46424", + "issue:43531|issue:45600", + "issue:44016|issue:44220", + "issue:45083|issue:46506", + "issue:43976|issue:44987", + "issue:43295|issue:45335" + ], + "prepared_review_unit_hash": "ed85dcf211b94fbb924d6f4814cf92888ea79b1d8b575ab229bf06f01cd5dc98", + "component_budget": { + "node_count": 473, + "item_count": 473, + "soft_pair_count": 1738, + "serialized_chars": 901437, + "estimated_input_tokens": 225360, + "estimated_eval_tokens": 450976 + }, + "budget": { + "node_count": 48, + "item_count": 48, + "soft_pair_count": 28, + "serialized_chars": 36079, + "estimated_input_tokens": 9020, + "estimated_eval_tokens": 18296 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": true, + "analyst_result": { + "summary": "The cluster is heterogeneous; the soft pairs connect different APIs, model families, or unrelated regressions, so none should be treated as duplicates.", + "confidence": 0.96, + "canonical_issue_reason": "No canonical issue stands out: these items are not one bug but many unrelated Transformers issues.", + "canonical_pr_reason": null, + "best_issue_reason": "No issue is a strong cluster representative because the candidates span unrelated failure modes and subsystems.", + "best_pr_reason": null, + "soft_edge_verdicts": [ + { + "left": "issue:44462", + "right": "issue:45907", + "accept": false, + "reason": "Tokenizer.json precedence vs revision forwarding are different APIs and failures." + }, + { + "left": "issue:45632", + "right": "issue:46234", + "accept": false, + "reason": "Local-model cache-path collision is unrelated to a missing import export." + }, + { + "left": "issue:44246", + "right": "issue:45003", + "accept": false, + "reason": "Slow import-time access to sys.modules is not the same bug as `_gpu_broadcast_one` rejecting None." + }, + { + "left": "issue:46082", + "right": "issue:46121", + "accept": false, + "reason": "Config validation for head_dim and JSON-loaded rope-key lists are separate errors." + }, + { + "left": "issue:42796", + "right": "issue:44521", + "accept": false, + "reason": "BPE marker handling and multimodal assistant masks are different tokenizer behaviors." + }, + { + "left": "issue:45245", + "right": "issue:45584", + "accept": false, + "reason": "Category-count overflow and Whisper empty-transcription handling are unrelated." + }, + { + "left": "issue:44964", + "right": "issue:45600", + "accept": false, + "reason": "Model-load breakage for Phi-4 multimodal and removed Sam3LiteText auto-mappings are different root causes." + }, + { + "left": "issue:42175", + "right": "issue:44279", + "accept": false, + "reason": "TensorFlow backend packaging and a generic dependency issue are not the same failure." + }, + { + "left": "issue:44112", + "right": "issue:44964", + "accept": false, + "reason": "GraniteSpeech CI test flake and Phi-4 multimodal loading are unrelated." + }, + { + "left": "issue:43854", + "right": "issue:44964", + "accept": false, + "reason": "GLM-4.7-Flash test loading and Phi-4 multimodal loading are different model-specific bugs." + }, + { + "left": "issue:43643", + "right": "issue:44496", + "accept": false, + "reason": "Missing config fields with trust_remote_code and unrecognized model_type are distinct config-loading issues." + }, + { + "left": "issue:27957", + "right": "issue:35141", + "accept": false, + "reason": "XLMRoberta Flash Attention 2 support and tie_word_embedding/post_init reinit are unrelated." + }, + { + "left": "issue:36246", + "right": "issue:41628", + "accept": false, + "reason": "Missing Qwen2_5_VLImageProcessor export and missing AutoImageProcessor export are separate import regressions." + }, + { + "left": "issue:44186", + "right": "issue:44734", + "accept": false, + "reason": "NER/batched-padding tokenizer crashes and KV-cache continuation tensor indexing are different code paths." + }, + { + "left": "issue:43404", + "right": "issue:45216", + "accept": false, + "reason": "Mistral3 lm_head tying and Qwen3.5 save_pretrained checkpoint corruption are different model bugs." + }, + { + "left": "issue:36246", + "right": "issue:38617", + "accept": false, + "reason": "Both are import errors, but they concern different missing symbols in different modules." + }, + { + "left": "issue:44295", + "right": "issue:46506", + "accept": false, + "reason": "position_ids buffer access and CPU/CUDA output divergence are unrelated." + }, + { + "left": "issue:43650", + "right": "issue:44220", + "accept": false, + "reason": "ADD THE DATA is not the same as an `_torch_extract_fbank_features()` bug." + }, + { + "left": "issue:41950", + "right": "issue:44016", + "accept": false, + "reason": "Video-classification pipeline image-processor lookup and a notebook syntax error are unrelated." + }, + { + "left": "issue:32090", + "right": "issue:35141", + "accept": false, + "reason": "Trainer NoneType broadcast and token-embedding resize reinitialization are different issues." + }, + { + "left": "issue:44683", + "right": "issue:45446", + "accept": false, + "reason": "Compiled flex_attention on torch>=2.9 and AuxRequest version gating are separate flex_attention regressions." + }, + { + "left": "issue:43479", + "right": "issue:45375", + "accept": false, + "reason": "Phi4MultimodalConfig default init and missing Qwen3_5MoeVisionConfig field are different config bugs." + }, + { + "left": "issue:43673", + "right": "issue:46424", + "accept": false, + "reason": "Generation cache missing in v5.0.0 chunked_prefill and max_cache_length only applying to quantized cache are different cache behaviors." + }, + { + "left": "issue:43531", + "right": "issue:45600", + "accept": false, + "reason": "Sliding-window behavior in Qwen3-MoE and removed Sam3LiteText auto-mappings are unrelated." + }, + { + "left": "issue:44016", + "right": "issue:44220", + "accept": false, + "reason": "Notebook syntax error and `_torch_extract_fbank_features()` are not the same problem." + }, + { + "left": "issue:45083", + "right": "issue:46506", + "accept": false, + "reason": "Helper output-length math in qwen3_omni_moe and CPU/CUDA differences are separate bugs." + }, + { + "left": "issue:43976", + "right": "issue:44987", + "accept": false, + "reason": "Python version support in Transformers 5.1.0 and loading a specific model are unrelated." + }, + { + "left": "issue:43295", + "right": "issue:45335", + "accept": false, + "reason": "Custom model processor.tokenizer access and resize_token_embeddings/decoder tying are different regressions." + } + ] + }, + "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": 21, + "review_unit_index": 37, + "review_unit_count": 56, + "status": "reviewed", + "reason": null, + "source": "llm", + "cache_hit": false, + "model": "gpt-5.4-mini?service_tier=flex", + "cluster_id": "cluster-10000-473", + "nodes": [ + "issue:39692", + "issue:41950", + "issue:42796", + "issue:42907", + "issue:43295", + "issue:43299", + "issue:43381", + "issue:43441", + "issue:43526", + "issue:43577", + "issue:43650", + "issue:43673", + "issue:43824", + "issue:43854", + "issue:43901", + "issue:43976", + "issue:44016", + "issue:44062", + "issue:44075", + "issue:44242", + "issue:44279", + "issue:44336", + "issue:44351", + "issue:44368", + "issue:44496", + "issue:44779", + "issue:44792", + "issue:44933", + "issue:45020", + "issue:45042", + "issue:45200", + "issue:45216", + "issue:45278", + "issue:45335", + "issue:45357", + "issue:45405", + "issue:45584", + "issue:45588", + "issue:45593", + "issue:45759", + "issue:45812", + "issue:46082", + "issue:46123", + "issue:46369", + "issue:46399", + "issue:46424", + "issue:46512", + "issue:46515" + ], + "soft_pairs": [ + "issue:43526|issue:43901", + "issue:44496|issue:45042", + "issue:42907|issue:43441", + "issue:43901|issue:44496", + "issue:43381|issue:45593", + "issue:43976|issue:44279", + "issue:43673|issue:44336", + "issue:43441|issue:45588", + "issue:46123|issue:46369", + "issue:44016|issue:44062", + "issue:44792|issue:46424", + "issue:43824|issue:44933", + "issue:44496|issue:45759", + "issue:43299|issue:45216", + "issue:44279|issue:46515", + "issue:45335|issue:45584", + "issue:46082|issue:46399", + "issue:44075|issue:44368", + "issue:44279|issue:45405", + "issue:42796|issue:43650", + "issue:43577|issue:45216", + "issue:39692|issue:43901", + "issue:43295|issue:44779", + "issue:44351|issue:45278", + "issue:43976|issue:45405", + "issue:43299|issue:45357", + "issue:44242|issue:45200", + "issue:46082|issue:46512", + "issue:44779|issue:45812", + "issue:41950|issue:45020", + "issue:43299|issue:43854" + ], + "prepared_review_unit_hash": "cf10573fdd752be7ab55a16742ffa2671e4dab251d8492b2c793594bd146eccf", + "component_budget": { + "node_count": 473, + "item_count": 473, + "soft_pair_count": 1738, + "serialized_chars": 901437, + "estimated_input_tokens": 225360, + "estimated_eval_tokens": 450976 + }, + "budget": { + "node_count": 48, + "item_count": 48, + "soft_pair_count": 31, + "serialized_chars": 36941, + "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 highly heterogeneous: most items are distinct model-, tokenizer-, cache-, or docs-related regressions, and none of the soft edges look like true duplicates. The broadest umbrella-style report is the generic upgrade/import regression issue, but it still does not unify the cluster into one underlying bug.", + "confidence": 0.78, + "canonical_issue_reason": "issue:45278 is the broadest umbrella report in the set: it frames a wide upgrade-induced import breakage rather than a narrow model-specific defect, so it is the best representative issue if one must be chosen.", + "canonical_pr_reason": null, + "best_issue_reason": "issue:45278 is the most useful global representative because it captures the widest class of version-upgrade regressions and could plausibly subsume multiple import-related symptoms, even though the cluster is not a true duplicate set.", + "best_pr_reason": null, + "soft_edge_verdicts": [ + { + "left": "issue:43526", + "right": "issue:43901", + "accept": false, + "reason": "Different bugs: BeitImageProcessorFast reduce_labels behavior vs a docs mismatch about return_all_scores." + }, + { + "left": "issue:44496", + "right": "issue:45042", + "accept": false, + "reason": "Unrecognized model/config loading error is unrelated to PIL backend image processors requiring torchvision." + }, + { + "left": "issue:42907", + "right": "issue:43441", + "accept": false, + "reason": "Saving dequantized Ministral/Devstral models is a different failure from Ministral flash-attention loading." + }, + { + "left": "issue:43901", + "right": "issue:44496", + "accept": false, + "reason": "Docs wording issue vs model-loading/config recognition failure; no shared underlying code path." + }, + { + "left": "issue:43381", + "right": "issue:45593", + "accept": false, + "reason": "Gradient checkpointing in eval mode and missing auxiliary losses with denoising off are separate training-path issues." + }, + { + "left": "issue:43976", + "right": "issue:44279", + "accept": false, + "reason": "Python version compatibility and a dependency issue are not the same concrete bug." + }, + { + "left": "issue:43673", + "right": "issue:44336", + "accept": false, + "reason": "Generation cache behavior and ANSI codes in loading_report are unrelated subsystems." + }, + { + "left": "issue:43441", + "right": "issue:45588", + "accept": false, + "reason": "Both mention flash attention, but one is a Ministral model-load failure and the other is an s_aux=None AttributeError for sink-less models." + }, + { + "left": "issue:46123", + "right": "issue:46369", + "accept": false, + "reason": "Partial-batch result dropping in MaskGenerationPipeline is unrelated to sam3-litetext memory overflow." + }, + { + "left": "issue:44016", + "right": "issue:44062", + "accept": false, + "reason": "Notebook syntax error and AddedToken keyword collision are distinct failures." + }, + { + "left": "issue:44792", + "right": "issue:46424", + "accept": false, + "reason": "A janus test failure is not the same as cache max length being honored only for quantized cache." + }, + { + "left": "issue:43824", + "right": "issue:44933", + "accept": false, + "reason": "Missing Qwen2.5-VL export and a nonexistent image_utils import are different import problems." + }, + { + "left": "issue:44496", + "right": "issue:45759", + "accept": false, + "reason": "Different model families and different mechanisms: unrecognized model config vs AutoModelForCausalLM not unwrapping text_config." + }, + { + "left": "issue:43299", + "right": "issue:45216", + "accept": false, + "reason": "Model loading failure and save_pretrained checkpoint corruption are separate lifecycle bugs." + }, + { + "left": "issue:44279", + "right": "issue:46515", + "accept": false, + "reason": "Generic dependency issue and torch 2.4.1 incompatibility are not the same root cause." + }, + { + "left": "issue:45335", + "right": "issue:45584", + "accept": false, + "reason": "T5Gemma embedding resize and Whisper empty-transcription generation are unrelated." + }, + { + "left": "issue:46082", + "right": "issue:46399", + "accept": false, + "reason": "LlamaConfig head_dim validation is unrelated to the missing decorator causing rotary-pos-emb registration failure." + }, + { + "left": "issue:44075", + "right": "issue:44368", + "accept": false, + "reason": "Unused SGD args and a Qwen warning about tie_word_embeddings are unrelated." + }, + { + "left": "issue:44279", + "right": "issue:45405", + "accept": false, + "reason": "A dependency issue is not the same as the MIN_PEFT_VERSION bump problem." + }, + { + "left": "issue:42796", + "right": "issue:43650", + "accept": false, + "reason": "Tokenizer BPE-marker handling has nothing in common with the placeholder 'ADD THE DATA' issue." + }, + { + "left": "issue:43577", + "right": "issue:45216", + "accept": false, + "reason": "BLIP2 dtype stale state and Qwen3.5 save_pretrained regression affect different models and code paths." + }, + { + "left": "issue:39692", + "right": "issue:43901", + "accept": false, + "reason": "SigLIP2 example errors and TextClassificationPipeline docs stale behavior are both docs-adjacent but not the same bug." + }, + { + "left": "issue:43295", + "right": "issue:44779", + "accept": false, + "reason": "Custom processor/tokenizer regression is different from Deepseek tokenizer output regression." + }, + { + "left": "issue:44351", + "right": "issue:45278", + "accept": false, + "reason": "These both involve imports, but one is a specific HybridCache missing export and the other is a broad upgrade regression; not the same concrete issue." + }, + { + "left": "issue:43976", + "right": "issue:45405", + "accept": false, + "reason": "Python support mismatch and unreleased PEFT minimum version are unrelated compatibility problems." + }, + { + "left": "issue:43299", + "right": "issue:45357", + "accept": false, + "reason": "Qwen3VL loading failure and incorrect visual encoder keys on save are distinct bugs." + }, + { + "left": "issue:44242", + "right": "issue:45200", + "accept": false, + "reason": "MoE load-balancing loss and Gemma mm_token_type_ids defaults are different model behaviors." + }, + { + "left": "issue:46082", + "right": "issue:46512", + "accept": false, + "reason": "Config head_dim validation is unrelated to from_config leaving model.config.dtype stale." + }, + { + "left": "issue:44779", + "right": "issue:45812", + "accept": false, + "reason": "Both are tokenizer regressions, but for different model families with different symptoms and likely fixes." + }, + { + "left": "issue:41950", + "right": "issue:45020", + "accept": false, + "reason": "Video-classification processor lookup and remote_code model breakage are separate subsystems." + }, + { + "left": "issue:43299", + "right": "issue:43854", + "accept": false, + "reason": "Qwen3VL loading regression and GLM-4.7-Flash unit-test loading failure are different model-specific 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": 21, + "review_unit_index": 38, + "review_unit_count": 56, + "status": "reviewed", + "reason": null, + "source": "llm", + "cache_hit": false, + "model": "gpt-5.4-mini?service_tier=flex", + "cluster_id": "cluster-10000-473", + "nodes": [ + "issue:33290", + "issue:38175", + "issue:41950", + "issue:42175", + "issue:42994", + "issue:43317", + "issue:43441", + "issue:43479", + "issue:43525", + "issue:43650", + "issue:43653", + "issue:43673", + "issue:43723", + "issue:43761", + "issue:43901", + "issue:43976", + "issue:43994", + "issue:44016", + "issue:44162", + "issue:44261", + "issue:44263", + "issue:44291", + "issue:44361", + "issue:44462", + "issue:44466", + "issue:44521", + "issue:44617", + "issue:44623", + "issue:44671", + "issue:44683", + "issue:44749", + "issue:44779", + "issue:45081", + "issue:45103", + "issue:45276", + "issue:45325", + "issue:45372", + "issue:45397", + "issue:45542", + "issue:45593", + "issue:45600", + "issue:45656", + "issue:45712", + "issue:45800", + "issue:45823", + "issue:45834", + "issue:45907", + "issue:46209" + ], + "soft_pairs": [ + "issue:38175|issue:44617", + "issue:43479|issue:43525", + "issue:43650|issue:46209", + "issue:44683|issue:45542", + "issue:41950|issue:45823", + "issue:44466|issue:45276", + "issue:44521|issue:45325", + "issue:43653|issue:45593", + "issue:45103|issue:45372", + "issue:44016|issue:44291", + "issue:44462|issue:45081", + "issue:43441|issue:45834", + "issue:33290|issue:45656", + "issue:44361|issue:45593", + "issue:42994|issue:43317", + "issue:44261|issue:44263", + "issue:45081|issue:45907", + "issue:44623|issue:44749", + "issue:45600|issue:45712", + "issue:44623|issue:45397", + "issue:42175|issue:43976", + "issue:43901|issue:43994", + "issue:44671|issue:44779", + "issue:43723|issue:44162", + "issue:43441|issue:44683", + "issue:45542|issue:45800", + "issue:43673|issue:43761" + ], + "prepared_review_unit_hash": "26ca815e1475368b261259f8c7da4932ee2da058a072c5b4ed6e7a3b9b793834", + "component_budget": { + "node_count": 473, + "item_count": 473, + "soft_pair_count": 1738, + "serialized_chars": 901437, + "estimated_input_tokens": 225360, + "estimated_eval_tokens": 450976 + }, + "budget": { + "node_count": 48, + "item_count": 48, + "soft_pair_count": 27, + "serialized_chars": 35783, + "estimated_input_tokens": 8946, + "estimated_eval_tokens": 18148 + }, + "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 bug reports across tokenizers, multimodal processors, model loading, quantization, deepspeed, and packaging. None of the soft-similarity pairs look like the same underlying bug, so I would not merge any of them as duplicates.", + "confidence": 0.54, + "canonical_issue_reason": null, + "canonical_pr_reason": null, + "best_issue_reason": "Issue 44521 is the best general representative if one must pick a hub: it is active, broadly scoped around multimodal chat template masking, and has the most inbound references among the candidates. It is still not a duplicate of the other issues.", + "best_pr_reason": null, + "soft_edge_verdicts": [ + { + "left": "issue:38175", + "right": "issue:44617", + "accept": false, + "reason": "SigLIP2 zero-probability outputs vs Sam3Video OOM are unrelated model-specific failures." + }, + { + "left": "issue:43479", + "right": "issue:43525", + "accept": false, + "reason": "Different models and different bugs: default multimodal config initialization vs missing pad_token_id on Llama4Config." + }, + { + "left": "issue:43650", + "right": "issue:46209", + "accept": false, + "reason": "Generic title only; one is an unspecified data request, the other is an FP8/B200 inference bug." + }, + { + "left": "issue:44683", + "right": "issue:45542", + "accept": false, + "reason": "Compiled flex_attention on torch>=2.9 is unrelated to a TensorFlow backend installation issue." + }, + { + "left": "issue:41950", + "right": "issue:45823", + "accept": false, + "reason": "Video pipeline processor lookup and Gemma4 device_map forward mismatch are different code paths." + }, + { + "left": "issue:44466", + "right": "issue:45276", + "accept": false, + "reason": "Serialization of tied lm_head weights and resize_token_embeddings not propagating to Gemma4 outputs are distinct bugs." + }, + { + "left": "issue:44521", + "right": "issue:45325", + "accept": false, + "reason": "Both are multimodal, but assistant_masks being all zero is not the same bug as Qwen2.5-VL rope index scaling." + }, + { + "left": "issue:43653", + "right": "issue:45593", + "accept": false, + "reason": "BigBirdTokenizer special-token registration is unrelated to D-FINE auxiliary loss behavior." + }, + { + "left": "issue:45103", + "right": "issue:45372", + "accept": false, + "reason": "auto_docstring annotation handling and Gemma4 processor loading/import errors are unrelated." + }, + { + "left": "issue:44016", + "right": "issue:44291", + "accept": false, + "reason": "Notebook syntax error and init_empty_weights argument mismatch do not share the same failure mode." + }, + { + "left": "issue:44462", + "right": "issue:45081", + "accept": false, + "reason": "AutoTokenizer ignoring tokenizer.json and Mistral regex patch crashing are separate tokenizer-loading problems." + }, + { + "left": "issue:43441", + "right": "issue:45834", + "accept": false, + "reason": "FlashAttention failure for Ministral-3 and Kosmos2.5 long OCR index error are unrelated." + }, + { + "left": "issue:33290", + "right": "issue:45656", + "accept": false, + "reason": "Deepspeed OOM with Adafactor and optimizer.step being called twice are different optimizer/deepspeed issues." + }, + { + "left": "issue:44361", + "right": "issue:45593", + "accept": false, + "reason": "MLukeTokenizer attribute errors and D-FINE auxiliary-loss handling are unrelated." + }, + { + "left": "issue:42994", + "right": "issue:43317", + "accept": false, + "reason": "Quantized model save failure and dequantized model load/offload failure are separate quantization pathways." + }, + { + "left": "issue:44261", + "right": "issue:44263", + "accept": false, + "reason": "Missing rms_norm_eps in MLA q_a_layernorm and torch.split return values in GlmMoeDsaIndexer are not the same bug." + }, + { + "left": "issue:45081", + "right": "issue:45907", + "accept": false, + "reason": "Mistral regex patch crash and list_repo_files missing revision forwarding are unrelated utility bugs." + }, + { + "left": "issue:44623", + "right": "issue:44749", + "accept": false, + "reason": "processor.save_pretrained missing files and a 10x slowdown after upgrade are not the same underlying issue." + }, + { + "left": "issue:45600", + "right": "issue:45712", + "accept": false, + "reason": "Both are CI/repo hygiene issues, but one is stale auto_mappings references and the other is leftover dummy classes leaking into dir()." + }, + { + "left": "issue:44623", + "right": "issue:45397", + "accept": false, + "reason": "Missing processor save files and Gemma-4 zero3 from_pretrained failures involve different loading/saving paths." + }, + { + "left": "issue:42175", + "right": "issue:43976", + "accept": false, + "reason": "TensorFlow backend installation metadata and Python 3.9/3.10 compatibility are unrelated packaging problems." + }, + { + "left": "issue:43901", + "right": "issue:43994", + "accept": false, + "reason": "Documentation mismatch for return_all_scores is not the same as incorrect SigLIP2 model/pipeline outputs." + }, + { + "left": "issue:44671", + "right": "issue:44779", + "accept": false, + "reason": "CamemBERT masked-LM regression and DeepSeek tokenizer regression are different components and failure modes." + }, + { + "left": "issue:43723", + "right": "issue:44162", + "accept": false, + "reason": "AutoTokenizer.from_pretrained loading regression and ESM2 broken workflow report are not the same bug." + }, + { + "left": "issue:43441", + "right": "issue:44683", + "accept": false, + "reason": "FlashAttention model failure and compiled flex_attention torch>=2.9 incompatibility are distinct attention-stack issues." + }, + { + "left": "issue:45542", + "right": "issue:45800", + "accept": false, + "reason": "Missing TensorFlow backend from extras and torch/transformers version incompatibility are different packaging/runtime problems." + }, + { + "left": "issue:43673", + "right": "issue:43761", + "accept": false, + "reason": "Generation cache missing in v5 chunked prefill and CLIPVisionModel hidden_states regression are unrelated model-specific 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": 21, + "review_unit_index": 39, + "review_unit_count": 56, + "status": "reviewed", + "reason": null, + "source": "llm", + "cache_hit": false, + "model": "gpt-5.4-mini?service_tier=flex", + "cluster_id": "cluster-10000-473", + "nodes": [ + "issue:36246", + "issue:41628", + "issue:41950", + "issue:42491", + "issue:42757", + "issue:42886", + "issue:43299", + "issue:43525", + "issue:43526", + "issue:43531", + "issue:43550", + "issue:43643", + "issue:43784", + "issue:43824", + "issue:43881", + "issue:43901", + "issue:43976", + "issue:44016", + "issue:44188", + "issue:44220", + "issue:44242", + "issue:44246", + "issue:44295", + "issue:44297", + "issue:44336", + "issue:44351", + "issue:44373", + "issue:44485", + "issue:44496", + "issue:44589", + "issue:44871", + "issue:45003", + "issue:45103", + "issue:45362", + "issue:45397", + "issue:45542", + "issue:45561", + "issue:45563", + "issue:45593", + "issue:45823", + "issue:46132", + "issue:46291", + "issue:46399", + "issue:46424", + "issue:46506", + "issue:46614", + "issue:46665" + ], + "soft_pairs": [ + "issue:43901|issue:44246", + "issue:46291|issue:46399", + "issue:44016|issue:45823", + "issue:43525|issue:43881", + "issue:43824|issue:43976", + "issue:43525|issue:43643", + "issue:44589|issue:45362", + "issue:44188|issue:44336", + "issue:44297|issue:45397", + "issue:44336|issue:44373", + "issue:43784|issue:45542", + "issue:41628|issue:43976", + "issue:44246|issue:44496", + "issue:46614|issue:46665", + "issue:45561|issue:45563", + "issue:45103|issue:46399", + "issue:44242|issue:45593", + "issue:42491|issue:43299", + "issue:44295|issue:45397", + "issue:43526|issue:45003", + "issue:43531|issue:45362", + "issue:44246|issue:44871", + "issue:44336|issue:44485", + "issue:43550|issue:46506", + "issue:42886|issue:46132", + "issue:44336|issue:46424", + "issue:41950|issue:44220", + "issue:36246|issue:42757", + "issue:36246|issue:44351" + ], + "prepared_review_unit_hash": "a07c6c4307f89afa1c83e6da20f9174780e6e453f2616a46cb0670c485df4881", + "component_budget": { + "node_count": 473, + "item_count": 473, + "soft_pair_count": 1738, + "serialized_chars": 901437, + "estimated_input_tokens": 225360, + "estimated_eval_tokens": 450976 + }, + "budget": { + "node_count": 47, + "item_count": 47, + "soft_pair_count": 29, + "serialized_chars": 35729, + "estimated_input_tokens": 8933, + "estimated_eval_tokens": 18122 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": true, + "analyst_result": { + "summary": "This cluster is heterogeneous: most items are unrelated import errors, model-loading regressions, docs problems, or performance/behavior bugs. No soft pair is a safe duplicate. The best representative issue is the Qwen2.5-VL import/export regression.", + "confidence": 0.88, + "canonical_issue_reason": "issue:43824 is the clearest and most concrete user-facing import regression, with stronger signal than the other heterogeneous issues.", + "canonical_pr_reason": null, + "best_issue_reason": "issue:43824 has the most directly actionable reproduction and the strongest external impact among the issues here.", + "best_pr_reason": null, + "soft_edge_verdicts": [ + { + "left": "issue:43901", + "right": "issue:44246", + "accept": false, + "reason": "Docs mismatch vs intermittent import-time slowdown; different problems." + }, + { + "left": "issue:46291", + "right": "issue:46399", + "accept": false, + "reason": "Both involve kernels, but one is a pipeline import failure and the other is a Qwen3.5 registration crash." + }, + { + "left": "issue:44016", + "right": "issue:45823", + "accept": false, + "reason": "Notebook syntax error vs Gemma4 device placement mismatch." + }, + { + "left": "issue:43525", + "right": "issue:43881", + "accept": false, + "reason": "Different model families and different loading failures." + }, + { + "left": "issue:43824", + "right": "issue:43976", + "accept": false, + "reason": "Specific missing export vs Python version compatibility breakage." + }, + { + "left": "issue:43525", + "right": "issue:43643", + "accept": false, + "reason": "Pad token attribute error vs missing config fields from trust_remote_code." + }, + { + "left": "issue:44589", + "right": "issue:45362", + "accept": false, + "reason": "Unrelated storage-object type error vs Qwen chat crash." + }, + { + "left": "issue:44188", + "right": "issue:44336", + "accept": false, + "reason": "torch.compile kernel divergence vs ANSI output formatting bug." + }, + { + "left": "issue:44297", + "right": "issue:45397", + "accept": false, + "reason": "Tokenizer save metadata mismatch vs ZeRO/from_pretrained failure." + }, + { + "left": "issue:44336", + "right": "issue:44373", + "accept": false, + "reason": "Terminal ANSI leakage vs wrong docstring." + }, + { + "left": "issue:43784", + "right": "issue:45542", + "accept": false, + "reason": "Different undefined-backend/import issues in different libraries." + }, + { + "left": "issue:41628", + "right": "issue:43976", + "accept": false, + "reason": "Missing import symbol vs Python-version support issue." + }, + { + "left": "issue:44246", + "right": "issue:44496", + "accept": false, + "reason": "Import performance complaint vs unrecognized model config." + }, + { + "left": "issue:46614", + "right": "issue:46665", + "accept": false, + "reason": "Stale phonemizer language state vs from_pretrained class behavior." + }, + { + "left": "issue:45561", + "right": "issue:45563", + "accept": false, + "reason": "Parallel test file race vs stale generation warning." + }, + { + "left": "issue:45103", + "right": "issue:46399", + "accept": false, + "reason": "Docstring helper crash vs missing decorator registration bug." + }, + { + "left": "issue:44242", + "right": "issue:45593", + "accept": false, + "reason": "Different missing-loss conditions in different models/features." + }, + { + "left": "issue:42491", + "right": "issue:43299", + "accept": false, + "reason": "Related Qwen MoE area, but distinct incompatibility/loading failures." + }, + { + "left": "issue:44295", + "right": "issue:45397", + "accept": false, + "reason": "Buffer access bug vs ZeRO/from_pretrained issue." + }, + { + "left": "issue:43526", + "right": "issue:45003", + "accept": false, + "reason": "BeitImageProcessor label reduction bug vs unsafe sys.modules access." + }, + { + "left": "issue:43531", + "right": "issue:45362", + "accept": false, + "reason": "Sliding-window bug vs chat runtime crash." + }, + { + "left": "issue:44246", + "right": "issue:44871", + "accept": false, + "reason": "Import slowdown vs Gemma eos_token_id inconsistency." + }, + { + "left": "issue:44336", + "right": "issue:44485", + "accept": false, + "reason": "ANSI code emission vs GLM-5 RoPE implementation discussion." + }, + { + "left": "issue:43550", + "right": "issue:46506", + "accept": false, + "reason": "Different model/runtime divergences in unrelated code paths." + }, + { + "left": "issue:42886", + "right": "issue:46132", + "accept": false, + "reason": "Offline cache loading bug vs AttentionInterface behavior change." + }, + { + "left": "issue:41950", + "right": "issue:44220", + "accept": false, + "reason": "Video pipeline choosing image processors vs audio feature extraction bug." + }, + { + "left": "issue:36246", + "right": "issue:42757", + "accept": false, + "reason": "Different missing imports from different packages." + }, + { + "left": "issue:36246", + "right": "issue:44351", + "accept": false, + "reason": "Different missing symbols; same broad symptom but not 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": 21, + "review_unit_index": 40, + "review_unit_count": 56, + "status": "reviewed", + "reason": null, + "source": "llm", + "cache_hit": false, + "model": "gpt-5.4-mini?service_tier=flex", + "cluster_id": "cluster-10000-473", + "nodes": [ + "issue:36246", + "issue:43257", + "issue:43262", + "issue:43352", + "issue:43475", + "issue:43540", + "issue:43596", + "issue:43673", + "issue:43901", + "issue:44038", + "issue:44060", + "issue:44077", + "issue:44162", + "issue:44242", + "issue:44246", + "issue:44295", + "issue:44466", + "issue:44479", + "issue:44496", + "issue:44561", + "issue:44734", + "issue:44749", + "issue:44779", + "issue:44857", + "issue:44877", + "issue:44987", + "issue:45003", + "issue:45137", + "issue:45200", + "issue:45250", + "issue:45292", + "issue:45335", + "issue:45375", + "issue:45412", + "issue:45459", + "issue:45561", + "issue:45584", + "issue:45600", + "issue:45701", + "issue:45759", + "issue:45859", + "issue:46123", + "issue:46170", + "issue:46369", + "issue:46399", + "issue:46424", + "issue:46506", + "issue:46655" + ], + "soft_pairs": [ + "issue:43540|issue:46655", + "issue:44779|issue:45292", + "issue:45859|issue:46399", + "issue:44734|issue:46424", + "issue:45003|issue:45561", + "issue:43901|issue:45003", + "issue:44479|issue:46655", + "issue:43257|issue:45137", + "issue:44242|issue:44857", + "issue:44496|issue:45003", + "issue:43475|issue:46369", + "issue:43673|issue:44561", + "issue:44295|issue:44749", + "issue:44162|issue:45200", + "issue:43262|issue:44246", + "issue:45375|issue:45759", + "issue:36246|issue:43596", + "issue:44877|issue:45375", + "issue:45292|issue:45584", + "issue:45137|issue:45600", + "issue:45292|issue:45701", + "issue:44561|issue:46399", + "issue:44038|issue:46506", + "issue:44466|issue:45335", + "issue:44987|issue:45459", + "issue:45412|issue:46123", + "issue:43352|issue:45250", + "issue:44479|issue:46170", + "issue:44060|issue:44077" + ], + "prepared_review_unit_hash": "3a053d07e4676a08851da581de4bf501ff76314c73f7e1e23662c1b2f87521ad", + "component_budget": { + "node_count": 473, + "item_count": 473, + "soft_pair_count": 1738, + "serialized_chars": 901437, + "estimated_input_tokens": 225360, + "estimated_eval_tokens": 450976 + }, + "budget": { + "node_count": 48, + "item_count": 48, + "soft_pair_count": 29, + "serialized_chars": 36471, + "estimated_input_tokens": 9118, + "estimated_eval_tokens": 18492 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": true, + "analyst_result": { + "summary": "This cluster is heterogeneous: the soft pairs mostly share broad keywords (video, cache, Qwen, DeepSpeed, Flash Attention) but describe different code paths, models, or symptoms. I do not see any true duplicate pair here.", + "confidence": 0.93, + "canonical_issue_reason": null, + "canonical_pr_reason": null, + "best_issue_reason": null, + "best_pr_reason": null, + "soft_edge_verdicts": [ + { + "left": "issue:43540", + "right": "issue:46655", + "accept": false, + "reason": "Both involve video inputs, but one is a Qwen3OmniMoe processing ValueError and the other is a SmolVLM padding bug; different models and likely different fixes." + }, + { + "left": "issue:44779", + "right": "issue:45292", + "accept": false, + "reason": "Tokenizer output regressions and resize_token_embeddings/output_embeddings mismatch are unrelated behaviors." + }, + { + "left": "issue:45859", + "right": "issue:46399", + "accept": false, + "reason": "Both are Qwen3.5-related, but one is CPU offload compatibility in rotary embedding and the other is a missing decorator causing kernel registration failure." + }, + { + "left": "issue:44734", + "right": "issue:46424", + "accept": false, + "reason": "KV-cache continuation crash and max_cache_length handling for quantized cache are separate cache issues." + }, + { + "left": "issue:45003", + "right": "issue:45561", + "accept": false, + "reason": "sys.modules access in modeling_utils and xdist race on captured_info.txt are unrelated." + }, + { + "left": "issue:43901", + "right": "issue:45003", + "accept": false, + "reason": "A docs mismatch for TextClassificationPipeline is not the same bug as unsafe sys.modules access." + }, + { + "left": "issue:44479", + "right": "issue:46655", + "accept": false, + "reason": "Both mention video regressions, but they affect different model families and different failures." + }, + { + "left": "issue:43257", + "right": "issue:45137", + "accept": false, + "reason": "Both involve DeepSpeed, but one is Qwen3 MOE weight conversion during loading and the other is a ZeRO3 deque IndexError." + }, + { + "left": "issue:44242", + "right": "issue:44857", + "accept": false, + "reason": "Load balancing loss gating and an AMP/CUDA crash in LwDetrImageLoss are different bugs." + }, + { + "left": "issue:44496", + "right": "issue:45003", + "accept": false, + "reason": "Unrecognized model config and sys.modules access are unrelated." + }, + { + "left": "issue:43475", + "right": "issue:46369", + "accept": false, + "reason": "SAM3 vision encoder missing an attribute and SAM3 lite text memory overflow are distinct failure modes." + }, + { + "left": "issue:43673", + "right": "issue:44561", + "accept": false, + "reason": "Chunked-prefill cache missing and removal of is_torch_fx_available breaking trust_remote_code models are unrelated." + }, + { + "left": "issue:44295", + "right": "issue:44749", + "accept": false, + "reason": "Position_ids buffer access error and a cross-version tokenization slowdown are not the same bug." + }, + { + "left": "issue:44162", + "right": "issue:45200", + "accept": false, + "reason": "ESM2 breakage and Gemma 4 mm_token_type_ids defaulting are unrelated model-specific issues." + }, + { + "left": "issue:43262", + "right": "issue:44246", + "accept": false, + "reason": "Audio chat-template sample-rate defaulting and slow import are unrelated." + }, + { + "left": "issue:45375", + "right": "issue:45759", + "accept": false, + "reason": "Missing config field silently dropped by strict and AutoModelForCausalLM text_config unwrapping are different multimodal config problems." + }, + { + "left": "issue:36246", + "right": "issue:43596", + "accept": false, + "reason": "An import error for Qwen2.5-VL image processor and a ZeRO3/BertModel index error are unrelated." + }, + { + "left": "issue:44877", + "right": "issue:45375", + "accept": false, + "reason": "Strict granite_speech config loading and Qwen3.5 Moe vision config field handling are separate config issues." + }, + { + "left": "issue:45292", + "right": "issue:45584", + "accept": false, + "reason": "Resize-token-embeddings behavior and Whisper empty transcription generation are different code paths." + }, + { + "left": "issue:45137", + "right": "issue:45600", + "accept": false, + "reason": "ZeRO3 deque failure and stale auto_mappings CI references are unrelated." + }, + { + "left": "issue:45292", + "right": "issue:45701", + "accept": false, + "reason": "Embedding resize behavior and tokenizer-version changes are not the same defect." + }, + { + "left": "issue:44561", + "right": "issue:46399", + "accept": false, + "reason": "trust_remote_code breakage from a removed helper and Qwen3.5 rotary-kernel registration failure are unrelated." + }, + { + "left": "issue:44038", + "right": "issue:46506", + "accept": false, + "reason": "Qwen3-VL-Moe regression and PPDocLayoutV3 CPU/CUDA divergence are separate model-specific bugs." + }, + { + "left": "issue:44466", + "right": "issue:45335", + "accept": false, + "reason": "Both touch tied/embedded weights, but one is device-dependent serialization and the other is resize_token_embeddings not updating decoder.embed_tokens." + }, + { + "left": "issue:44987", + "right": "issue:45459", + "accept": false, + "reason": "Loading physical-intelligence/fast and hidden tokenizer decode errors due to missing protobuf are unrelated." + }, + { + "left": "issue:45412", + "right": "issue:46123", + "accept": false, + "reason": "RT-DETR memory release and MaskGenerationPipeline partial-batch handling are unrelated." + }, + { + "left": "issue:43352", + "right": "issue:45250", + "accept": false, + "reason": "A specific model lacking Flash Attention 2.0 support is not the same as a generic Flash Attention 2.0 issue." + }, + { + "left": "issue:44479", + "right": "issue:46170", + "accept": false, + "reason": "Video-input regressions and Windows/ROCm torch.distributed import errors are unrelated." + }, + { + "left": "issue:44060", + "right": "issue:44077", + "accept": false, + "reason": "Incorrect tied-weights warning in Qwen3-Next and patchtsmixer post_init allowance 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": 21, + "review_unit_index": 41, + "review_unit_count": 56, + "status": "reviewed", + "reason": null, + "source": "llm", + "cache_hit": false, + "model": "gpt-5.4-mini?service_tier=flex", + "cluster_id": "cluster-10000-473", + "nodes": [ + "issue:42886", + "issue:42994", + "issue:43065", + "issue:43232", + "issue:43408", + "issue:43475", + "issue:43482", + "issue:43540", + "issue:43643", + "issue:43646", + "issue:43653", + "issue:43673", + "issue:43906", + "issue:43931", + "issue:44162", + "issue:44188", + "issue:44297", + "issue:44361", + "issue:44368", + "issue:44483", + "issue:44623", + "issue:44749", + "issue:44869", + "issue:44871", + "issue:44877", + "issue:44977", + "issue:44995", + "issue:45200", + "issue:45229", + "issue:45230", + "issue:45310", + "issue:45356", + "issue:45375", + "issue:45397", + "issue:45478", + "issue:45563", + "issue:45584", + "issue:45588", + "issue:45656", + "issue:45685", + "issue:45987", + "issue:46082", + "issue:46123", + "issue:46132", + "issue:46170", + "issue:46202", + "issue:46421", + "issue:46431" + ], + "soft_pairs": [ + "issue:44368|issue:44977", + "issue:44995|issue:45200", + "issue:45397|issue:45656", + "issue:43653|issue:44162", + "issue:43475|issue:43540", + "issue:42886|issue:45229", + "issue:44877|issue:46170", + "issue:44995|issue:45588", + "issue:44368|issue:45563", + "issue:43065|issue:43408", + "issue:43906|issue:46202", + "issue:43931|issue:44368", + "issue:44188|issue:46202", + "issue:45685|issue:46431", + "issue:44483|issue:46202", + "issue:45987|issue:46082", + "issue:45584|issue:46170", + "issue:44188|issue:46421", + "issue:43065|issue:46123", + "issue:44297|issue:44623", + "issue:44361|issue:44749", + "issue:44749|issue:45656", + "issue:44623|issue:45656", + "issue:44749|issue:45310", + "issue:43646|issue:44162", + "issue:44871|issue:46132", + "issue:43232|issue:43673", + "issue:42994|issue:43482", + "issue:45230|issue:45397", + "issue:44869|issue:45356", + "issue:45375|issue:45478", + "issue:43475|issue:43643" + ], + "prepared_review_unit_hash": "baec19de04e21490adacfef6a25f72ea332124119953b10f032f7394b3430d09", + "component_budget": { + "node_count": 473, + "item_count": 473, + "soft_pair_count": 1738, + "serialized_chars": 901437, + "estimated_input_tokens": 225360, + "estimated_eval_tokens": 450976 + }, + "budget": { + "node_count": 48, + "item_count": 48, + "soft_pair_count": 32, + "serialized_chars": 37218, + "estimated_input_tokens": 9305, + "estimated_eval_tokens": 18866 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": true, + "analyst_result": { + "summary": "This is a heterogeneous set of Transformers-related bug reports; the soft pairs are mostly related only at the subsystem or release level, not the same concrete bug. I found no safe duplicates among the provided soft edges.", + "confidence": 0.78, + "canonical_issue_reason": "Issue 43646 is the broadest upstream regression report in the set and the best representative umbrella for the surrounding Transformers 5.x breakage reports, even though it is not an exact duplicate of them.", + "canonical_pr_reason": null, + "best_issue_reason": "43646 is the most general issue here and the best single representative for the cluster because it covers broad custom-model initialization regressions rather than a narrow model-specific symptom.", + "best_pr_reason": null, + "soft_edge_verdicts": [ + { + "left": "issue:44368", + "right": "issue:44977", + "accept": false, + "reason": "Different problems: a warning about tie_word_embeddings on Qwen3.5 LoRA vs a flash-attention generation failure." + }, + { + "left": "issue:44995", + "right": "issue:45200", + "accept": false, + "reason": "Both touch model behavior, but one is a stale indexer cache crash and the other is a Gemma token-type defaulting issue." + }, + { + "left": "issue:45397", + "right": "issue:45656", + "accept": false, + "reason": "Gemma-4 ZeRO3 loading vs deepspeed optimizer stepping twice are unrelated code paths and failures." + }, + { + "left": "issue:43653", + "right": "issue:44162", + "accept": false, + "reason": "Tokenizer special-token registration and ESM2 breakage are different model/tokenizer issues." + }, + { + "left": "issue:43475", + "right": "issue:43540", + "accept": false, + "reason": "A missing SAM3 vision encoder attribute is not the same as a Qwen3OmniMoe video-input validation error." + }, + { + "left": "issue:42886", + "right": "issue:45229", + "accept": false, + "reason": "Offline tokenizer cache loading and Gemma4 multi-GPU OOM are clearly different bugs." + }, + { + "left": "issue:44877", + "right": "issue:46170", + "accept": false, + "reason": "Strict config loading for granite_speech and a Windows/ROCm torch.distributed import problem are unrelated." + }, + { + "left": "issue:44995", + "right": "issue:45588", + "accept": false, + "reason": "Stale indexer cache on second forward pass is not the same as flash_attention crashing on s_aux=None." + }, + { + "left": "issue:44368", + "right": "issue:45563", + "accept": false, + "reason": "A tie_word_embeddings warning is unrelated to a stale num_return_sequences warning in paged generate." + }, + { + "left": "issue:43065", + "right": "issue:43408", + "accept": false, + "reason": "One is a dummy Conv2d in Sam3PixelDecoder; the other is a model-type mismatch warning." + }, + { + "left": "issue:43906", + "right": "issue:46202", + "accept": false, + "reason": "An isolated reproduction of an older issue is not the same as Gemma pipeline roundtrip failure on \u2581." + }, + { + "left": "issue:43931", + "right": "issue:44368", + "accept": false, + "reason": "Qwen3-VL weight-shape mismatch and a Qwen3.5 tie_word_embeddings warning are different failures." + }, + { + "left": "issue:44188", + "right": "issue:46202", + "accept": false, + "reason": "torch.compile branching causing attention kernel divergence is unrelated to Gemma pipeline roundtrip behavior." + }, + { + "left": "issue:45685", + "right": "issue:46431", + "accept": false, + "reason": "MPS histogram support on Int and missing chat template errors are unrelated." + }, + { + "left": "issue:44483", + "right": "issue:46202", + "accept": false, + "reason": "A v5.3 chat/completions request rejection is not the same as a Gemma pipeline roundtrip tokenizer bug." + }, + { + "left": "issue:45987", + "right": "issue:46082", + "accept": false, + "reason": "StaticCache seq_length returning a tensor and LlamaConfig head_dim validation are different API/validation issues." + }, + { + "left": "issue:45584", + "right": "issue:46170", + "accept": false, + "reason": "Whisper empty-transcription decoding and Windows ROCm import failures do not share a code path." + }, + { + "left": "issue:44188", + "right": "issue:46421", + "accept": false, + "reason": "Both mention caching, but one is attention-kernel divergence under torch.compile and the other is lazy cache init causing recompilation." + }, + { + "left": "issue:43065", + "right": "issue:46123", + "accept": false, + "reason": "Sam3PixelDecoder dummy layer and MaskGenerationPipeline partial-batch handling are unrelated." + }, + { + "left": "issue:44297", + "right": "issue:44623", + "accept": false, + "reason": "Both are save_pretrained regressions, but one is tokenizer config metadata and the other is processor file emission." + }, + { + "left": "issue:44361", + "right": "issue:44749", + "accept": false, + "reason": "MLukeTokenizer AttributeError and a dataset filtering slowdown are unrelated bugs." + }, + { + "left": "issue:44749", + "right": "issue:45656", + "accept": false, + "reason": "A preprocessing performance regression and a deepspeed optimizer-step bug are not the same change." + }, + { + "left": "issue:44623", + "right": "issue:45656", + "accept": false, + "reason": "Processor.save_pretrained missing files and duplicated optimizer steps are completely different issues." + }, + { + "left": "issue:44749", + "right": "issue:45310", + "accept": false, + "reason": "Data filtering slowdown and Qwen3.5 MoE from_pretrained failure are unrelated." + }, + { + "left": "issue:43646", + "right": "issue:44162", + "accept": false, + "reason": "Both are broad regressions, but one is custom model initialization and the other is an ESM2-specific breakage." + }, + { + "left": "issue:44871", + "right": "issue:46132", + "accept": false, + "reason": "Gemma eos_token_id configuration mismatch and AttentionInterface.register behavior changes are different concerns." + }, + { + "left": "issue:43232", + "right": "issue:43673", + "accept": false, + "reason": "Both involve generation/cache, but one is sync_gpus update kwargs and the other is chunked_prefill cache availability." + }, + { + "left": "issue:42994", + "right": "issue:43482", + "accept": false, + "reason": "Quantized model saving failure and Qwen2.5-GGUF loading failure are different lifecycle operations." + }, + { + "left": "issue:45230", + "right": "issue:45397", + "accept": false, + "reason": "A placeholder bug report is not a duplicate of a specific Gemma-4 ZeRO3 from_pretrained failure." + }, + { + "left": "issue:44869", + "right": "issue:45356", + "accept": false, + "reason": "Whisper timestamp decoding on a trailing replacement character and Kimi-K2.5 tokenizer codec handling are related only at the tokenizer level, not the same bug." + }, + { + "left": "issue:45375", + "right": "issue:45478", + "accept": false, + "reason": "Same model family and loading context, but one is a missing config field dropped by @strict and the other is a generic from_pretrained failure; not the same concrete bug." + }, + { + "left": "issue:43475", + "right": "issue:43643", + "accept": false, + "reason": "A missing Sam3VisionEncoderOutput attribute and missing fields in AutoConfig with trust_remote_code 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": 21, + "review_unit_index": 42, + "review_unit_count": 56, + "status": "reviewed", + "reason": null, + "source": "llm", + "cache_hit": false, + "model": "gpt-5.4-mini?service_tier=flex", + "cluster_id": "cluster-10000-473", + "nodes": [ + "issue:28282", + "issue:30333", + "issue:30990", + "issue:33290", + "issue:36010", + "issue:41628", + "issue:42175", + "issue:42371", + "issue:42673", + "issue:42831", + "issue:42994", + "issue:43232", + "issue:43278", + "issue:43475", + "issue:43504", + "issue:43525", + "issue:43704", + "issue:43723", + "issue:43784", + "issue:43824", + "issue:43873", + "issue:43976", + "issue:44038", + "issue:44188", + "issue:44222", + "issue:44351", + "issue:44496", + "issue:44623", + "issue:44792", + "issue:44933", + "issue:44945", + "issue:44964", + "issue:45081", + "issue:45092", + "issue:45103", + "issue:45137", + "issue:45310", + "issue:45375", + "issue:45381", + "issue:45406", + "issue:45561", + "issue:45600", + "issue:45712", + "issue:45753", + "issue:46202", + "issue:46326", + "issue:46399" + ], + "soft_pairs": [ + "issue:44038|issue:44623", + "issue:28282|issue:30990", + "issue:41628|issue:44933", + "issue:42371|issue:45092", + "issue:45310|issue:45375", + "issue:42994|issue:43525", + "issue:43475|issue:45381", + "issue:44964|issue:45137", + "issue:43723|issue:46326", + "issue:43784|issue:46399", + "issue:45406|issue:46202", + "issue:43873|issue:44945", + "issue:43232|issue:44188", + "issue:41628|issue:44351", + "issue:44945|issue:45753", + "issue:45081|issue:45103", + "issue:42673|issue:43704", + "issue:42994|issue:43824", + "issue:44038|issue:45310", + "issue:44222|issue:45310", + "issue:42831|issue:44222", + "issue:36010|issue:41628", + "issue:43278|issue:45137", + "issue:43232|issue:44792", + "issue:42175|issue:44933", + "issue:44496|issue:45561", + "issue:43504|issue:45561", + "issue:45561|issue:45600", + "issue:43976|issue:44933", + "issue:30333|issue:33290", + "issue:42175|issue:45712" + ], + "prepared_review_unit_hash": "393e3dd196235966ccc7591282c14855b8d79f1dfcaff5bfcace92cff54fef99", + "component_budget": { + "node_count": 473, + "item_count": 473, + "soft_pair_count": 1738, + "serialized_chars": 901437, + "estimated_input_tokens": 225360, + "estimated_eval_tokens": 450976 + }, + "budget": { + "node_count": 47, + "item_count": 47, + "soft_pair_count": 31, + "serialized_chars": 36242, + "estimated_input_tokens": 9061, + "estimated_eval_tokens": 18378 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": true, + "analyst_result": { + "summary": "Mostly unrelated Transformers issues; the only clear duplicate pair is the Qwen3ForCausalLM VRAM leak reports (42673/43704). I corrected the canonical issue wording to avoid claiming it is open when the packet shows it is closed.", + "confidence": 0.95, + "canonical_issue_reason": "issue:42673 is the better canonical because it is the earlier, more detailed report for the same Qwen3ForCausalLM VRAM leak, and issue:43704 is an exact duplicate with the same title.", + "canonical_pr_reason": null, + "best_issue_reason": "issue:42673 is the best representative issue in this set because it is the clearest duplicate anchor for the one confirmed same-bug pair and has the richer discussion trail.", + "best_pr_reason": null, + "soft_edge_verdicts": [ + { + "left": "issue:44038", + "right": "issue:44623", + "accept": false, + "reason": "Different bugs: Qwen3-VL-Moe load/model compatibility vs missing files from processor.save_pretrained()." + }, + { + "left": "issue:28282", + "right": "issue:30990", + "accept": false, + "reason": "Different failure modes: missing PyTorch backend import vs Sentence Transformers hanging while loading." + }, + { + "left": "issue:41628", + "right": "issue:44933", + "accept": false, + "reason": "Both are import-related, but they concern different missing symbols in different modules." + }, + { + "left": "issue:42371", + "right": "issue:45092", + "accept": false, + "reason": "Unrelated issues: TF32 API guidance vs old InternVL2 checkpoints failing under v5 meta initialization." + }, + { + "left": "issue:45310", + "right": "issue:45375", + "accept": false, + "reason": "Related model family, but not the same concrete bug or fix set." + }, + { + "left": "issue:42994", + "right": "issue:43525", + "accept": false, + "reason": "Different problems: quantized model saving failure vs missing pad_token_id on Llama4Config." + }, + { + "left": "issue:43475", + "right": "issue:45381", + "accept": false, + "reason": "Different models and code paths: SAM 3 vision output attribute vs qwen2.5-vl video position IDs." + }, + { + "left": "issue:44964", + "right": "issue:45137", + "accept": false, + "reason": "Different subsystems: Phi-4 multimodal loading vs DeepSpeed ZeRO3 deque underflow." + }, + { + "left": "issue:43723", + "right": "issue:46326", + "accept": false, + "reason": "Both involve tokenizers, but the bugs are different concrete code paths." + }, + { + "left": "issue:43784", + "right": "issue:46399", + "accept": false, + "reason": "Different issues: sentence-transformers import NameError vs Qwen3.5 kernel registration crash." + }, + { + "left": "issue:45406", + "right": "issue:46202", + "accept": false, + "reason": "Different Gemma issues: missing _tokenizer on processor vs a pipeline roundtrip tokenization mismatch." + }, + { + "left": "issue:43873", + "right": "issue:44945", + "accept": false, + "reason": "Different behaviors: quantization/offloading not working as expected vs incorrect output under pipeline parallelism." + }, + { + "left": "issue:43232", + "right": "issue:44188", + "accept": false, + "reason": "Different generation/attention bugs with no shared concrete fix path." + }, + { + "left": "issue:41628", + "right": "issue:44351", + "accept": false, + "reason": "Different import errors in different APIs: AutoImageProcessor vs HybridCache." + }, + { + "left": "issue:44945", + "right": "issue:45753", + "accept": false, + "reason": "Different symptoms and code paths: pipeline parallelism output bug vs Qwen3_5 image infinite loop." + }, + { + "left": "issue:45081", + "right": "issue:45103", + "accept": false, + "reason": "Completely different crashes: Mistral regex patching vs auto_docstring with postponed annotations." + }, + { + "left": "issue:42673", + "right": "issue:43704", + "accept": true, + "reason": "Same underlying bug and identical title: Qwen3ForCausalLM VRAM leak across multiple dataloader threads." + }, + { + "left": "issue:42994", + "right": "issue:43824", + "accept": false, + "reason": "Different issues: quantized model saving failure vs missing Qwen2_5_VLForConditionalGeneration import." + }, + { + "left": "issue:44038", + "right": "issue:45310", + "accept": false, + "reason": "Same broad model family, but not the same concrete bug or fixable change set." + }, + { + "left": "issue:44222", + "right": "issue:45310", + "accept": false, + "reason": "Different bugs: FP8 save_pretrained/moe vs Qwen3.5 from_pretrained loading error." + }, + { + "left": "issue:42831", + "right": "issue:44222", + "accept": false, + "reason": "Different FP8-related issues: accuracy degradation vs save_pretrained failure." + }, + { + "left": "issue:36010", + "right": "issue:41628", + "accept": false, + "reason": "Both are import errors, but for different symbols and code paths." + }, + { + "left": "issue:43278", + "right": "issue:45137", + "accept": false, + "reason": "Different problems: dtype change in eval vs DeepSpeed ZeRO3 deque pop error." + }, + { + "left": "issue:43232", + "right": "issue:44792", + "accept": false, + "reason": "Different code paths: generation kwargs update after sync_gpus vs a janus image-generation test failure." + }, + { + "left": "issue:42175", + "right": "issue:44933", + "accept": false, + "reason": "Different bugs: dependency/backend packaging vs a nonexistent image_utils import." + }, + { + "left": "issue:44496", + "right": "issue:45561", + "accept": false, + "reason": "Unrelated: model config loading failure vs pytest-xdist race in testing utils." + }, + { + "left": "issue:43504", + "right": "issue:45561", + "accept": false, + "reason": "Different issues: legacy BEiT preset loading vs xdist captured_info.txt race." + }, + { + "left": "issue:45561", + "right": "issue:45600", + "accept": false, + "reason": "Different CI/test problems: xdist race vs stale auto_mappings references." + }, + { + "left": "issue:43976", + "right": "issue:44933", + "accept": false, + "reason": "Different bugs: Python version compatibility vs missing import from image_utils." + }, + { + "left": "issue:30333", + "right": "issue:33290", + "accept": false, + "reason": "Different failures: MLflow reporting after training-job failure vs OOM with Adafactor in DeepSpeed." + }, + { + "left": "issue:42175", + "right": "issue:45712", + "accept": false, + "reason": "Related packaging area, but different concrete bugs and fixes." + } + ] + }, + "evaluator_result": { + "accept": false, + "feedback": "Reject: the canonical_issue_reason overstates the packet by calling issue:42673 the 'original/open report' even though the packet shows it is closed. The rest is mostly conservative, and the 42673/43704 duplicate call looks grounded, but the state claim is factually incorrect." + }, + "evaluator_used": true, + "retried": true, + "accepted_nontrivial_soft_edge": true, + "error_kind": null, + "error_message": null + }, + { + "label": "issue", + "component_index": 1, + "component_count": 21, + "review_unit_index": 43, + "review_unit_count": 56, + "status": "reviewed", + "reason": null, + "source": "llm", + "cache_hit": false, + "model": "gpt-5.4-mini?service_tier=flex", + "cluster_id": "cluster-10000-473", + "nodes": [ + "issue:22355", + "issue:27957", + "issue:28282", + "issue:30333", + "issue:30990", + "issue:34689", + "issue:35141", + "issue:36331", + "issue:42222", + "issue:42831", + "issue:42898", + "issue:43116", + "issue:43295", + "issue:43377", + "issue:43479", + "issue:43723", + "issue:43784", + "issue:43824", + "issue:43874", + "issue:44016", + "issue:44060", + "issue:44075", + "issue:44079", + "issue:44190", + "issue:44222", + "issue:44242", + "issue:44295", + "issue:44297", + "issue:44568", + "issue:44623", + "issue:44734", + "issue:44779", + "issue:44869", + "issue:44933", + "issue:45092", + "issue:45103", + "issue:45125", + "issue:45200", + "issue:45310", + "issue:45356", + "issue:45405", + "issue:45656", + "issue:45712", + "issue:45812", + "issue:46077", + "issue:46123", + "issue:46231" + ], + "soft_pairs": [ + "issue:46077|issue:46123", + "issue:30990|issue:35141", + "issue:42898|issue:44568", + "issue:43377|issue:44242", + "issue:44623|issue:45310", + "issue:43479|issue:44869", + "issue:42222|issue:45200", + "issue:44016|issue:46231", + "issue:42898|issue:45812", + "issue:43874|issue:45092", + "issue:43723|issue:44568", + "issue:43479|issue:44079", + "issue:42898|issue:44779", + "issue:43784|issue:43824", + "issue:44734|issue:45103", + "issue:42898|issue:45356", + "issue:44295|issue:45656", + "issue:27957|issue:30333", + "issue:42898|issue:43295", + "issue:30333|issue:34689", + "issue:44060|issue:45125", + "issue:45405|issue:45712", + "issue:44075|issue:46231", + "issue:43116|issue:44190", + "issue:44222|issue:44297", + "issue:44933|issue:45712", + "issue:30333|issue:36331", + "issue:30333|issue:35141", + "issue:44297|issue:45656", + "issue:22355|issue:28282", + "issue:42831|issue:44623", + "issue:44222|issue:44623" + ], + "prepared_review_unit_hash": "e5f61587f3cdb5d5b123b6d67626d8b519fa8f47d68fe8e699c9dfbd9c688291", + "component_budget": { + "node_count": 473, + "item_count": 473, + "soft_pair_count": 1738, + "serialized_chars": 901437, + "estimated_input_tokens": 225360, + "estimated_eval_tokens": 450976 + }, + "budget": { + "node_count": 47, + "item_count": 47, + "soft_pair_count": 32, + "serialized_chars": 36920, + "estimated_input_tokens": 9230, + "estimated_eval_tokens": 18716 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": true, + "analyst_result": { + "summary": "This cluster is heterogeneous: most items are distinct regression reports across tokenization, model loading, training, serving, and docs, with no true duplicate pair standing out. The closest theme is tokenizer/v5 regressions, but even there the failures are model- or API-specific rather than the same bug.", + "confidence": 0.34, + "canonical_issue_reason": "issue:42898 is the broadest, most central report in the cluster: a core tokenizer behavior regression affecting v5 behavior generally, and it is the closest match to several other tokenizer-related issues.", + "canonical_pr_reason": null, + "best_issue_reason": "issue:42898 is the best single issue to anchor this cluster because it is a cross-cutting library regression rather than a niche model-specific or example-script-only problem.", + "best_pr_reason": null, + "soft_edge_verdicts": [ + { + "left": "issue:46077", + "right": "issue:46123", + "accept": false, + "reason": "Different bugs: Gemma4 attention-mask construction vs. a pipeline batching/last-chunk flag issue." + }, + { + "left": "issue:30990", + "right": "issue:35141", + "accept": false, + "reason": "Unrelated failure modes: a loading stall vs. embedding reinitialization after resize." + }, + { + "left": "issue:42898", + "right": "issue:44568", + "accept": false, + "reason": "Both are tokenizer regressions, but they affect different models/APIs and distinct code paths (`clean_up_tokenization_spaces` vs `add_special_tokens`)." + }, + { + "left": "issue:43377", + "right": "issue:44242", + "accept": false, + "reason": "Different components: MIMI padding-mask behavior vs. MoE load-balancing loss calculation." + }, + { + "left": "issue:44623", + "right": "issue:45310", + "accept": false, + "reason": "Different save/load problems: missing processor files vs. Qwen3.5 MoE `from_pretrained` failure." + }, + { + "left": "issue:43479", + "right": "issue:44869", + "accept": false, + "reason": "Different model classes and failure modes: config initialization vs. Whisper timestamp decode crash." + }, + { + "left": "issue:42222", + "right": "issue:45200", + "accept": false, + "reason": "Different model families and bugs: broken VitPose models vs. Gemma 4 token-type defaults." + }, + { + "left": "issue:44016", + "right": "issue:46231", + "accept": false, + "reason": "A notebook syntax error is not the same as a docs parameter-name mismatch." + }, + { + "left": "issue:42898", + "right": "issue:45812", + "accept": false, + "reason": "Both involve tokenizer regressions, but one is a global cleanup behavior change and the other is wrong token IDs for Granite models." + }, + { + "left": "issue:43874", + "right": "issue:45092", + "accept": false, + "reason": "Different areas: missing image-processor method vs. remote-code/meta-init incompatibility." + }, + { + "left": "issue:43723", + "right": "issue:44568", + "accept": false, + "reason": "Separate tokenizer issues with different symptoms: tokenizer loading vs. special-token insertion." + }, + { + "left": "issue:43479", + "right": "issue:44079", + "accept": false, + "reason": "Config-default construction and `ModelOutput` key handling are unrelated code paths." + }, + { + "left": "issue:42898", + "right": "issue:44779", + "accept": false, + "reason": "Both are tokenizer regressions, but they affect different models and different underlying tokenization logic." + }, + { + "left": "issue:43784", + "right": "issue:43824", + "accept": false, + "reason": "Different import errors: one is a missing symbol in sentence-transformers integration, the other is a missing Qwen2.5-VL export." + }, + { + "left": "issue:44734", + "right": "issue:45103", + "accept": false, + "reason": "Serving KV-cache indexing crash and auto-docstring annotation handling are unrelated." + }, + { + "left": "issue:42898", + "right": "issue:45356", + "accept": false, + "reason": "Both mention tokenizer regressions, but the Kimi issue is codec/regex-specific and not the same bug as `clean_up_tokenization_spaces` behavior." + }, + { + "left": "issue:44295", + "right": "issue:45656", + "accept": false, + "reason": "Different subsystems: buffer access on `position_ids` vs. optimizer stepping under DeepSpeed." + }, + { + "left": "issue:27957", + "right": "issue:30333", + "accept": false, + "reason": "Unrelated issues: Flash Attention 2 support for XLM-Roberta vs. MLflow reporting after failed training jobs." + }, + { + "left": "issue:42898", + "right": "issue:43295", + "accept": false, + "reason": "Tokenizer cleanup behavior and processor/tokenizer access regressions are different problems." + }, + { + "left": "issue:30333", + "right": "issue:34689", + "accept": false, + "reason": "MLflow job-tracking behavior is unrelated to model-loading breakage for Llama 3.2 Vision." + }, + { + "left": "issue:44060", + "right": "issue:45125", + "accept": false, + "reason": "Both are Qwen-related, but one is a tied-weights warning and the other is missing tensor-parallel plan support." + }, + { + "left": "issue:45405", + "right": "issue:45712", + "accept": false, + "reason": "A dependency-version bump and leftover dummy classes are unrelated maintenance issues." + }, + { + "left": "issue:44075", + "right": "issue:46231", + "accept": false, + "reason": "Optimizer arguments not being used is not the same as a documentation typo." + }, + { + "left": "issue:43116", + "right": "issue:44190", + "accept": false, + "reason": "Different example-script failures: empty multi-label results vs. local dataset loading." + }, + { + "left": "issue:44222", + "right": "issue:44297", + "accept": false, + "reason": "Different save-pretrained bugs affecting different object types: FP8 MoE vs. tokenizer metadata." + }, + { + "left": "issue:44933", + "right": "issue:45712", + "accept": false, + "reason": "Import exposure problems in `image_utils` and dummy class cleanup are not the same bug." + }, + { + "left": "issue:30333", + "right": "issue:36331", + "accept": false, + "reason": "MLflow callback/reporting and trainer loss-signature compatibility are different training issues." + }, + { + "left": "issue:30333", + "right": "issue:35141", + "accept": false, + "reason": "Failed-job reporting and embedding reinitialization are unrelated." + }, + { + "left": "issue:44297", + "right": "issue:45656", + "accept": false, + "reason": "Tokenizer save metadata mismatch and duplicate optimizer stepping are unrelated." + }, + { + "left": "issue:22355", + "right": "issue:28282", + "accept": false, + "reason": "Both are import errors, but one is a missing `transformers.onnx` module and the other is a missing PyTorch dependency check." + }, + { + "left": "issue:42831", + "right": "issue:44623", + "accept": false, + "reason": "An FP8 accuracy regression and missing processor files are different problem types." + }, + { + "left": "issue:44222", + "right": "issue:44623", + "accept": false, + "reason": "Both mention save_pretrained, but one is FP8/MoE serialization and the other is processor file persistence." + } + ] + }, + "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": 21, + "review_unit_index": 44, + "review_unit_count": 56, + "status": "reviewed", + "reason": null, + "source": "llm", + "cache_hit": false, + "model": "gpt-5.4-mini?service_tier=flex", + "cluster_id": "cluster-10000-473", + "nodes": [ + "issue:30990", + "issue:38175", + "issue:41628", + "issue:42757", + "issue:42796", + "issue:42831", + "issue:43122", + "issue:43232", + "issue:43475", + "issue:43479", + "issue:43644", + "issue:43723", + "issue:44190", + "issue:44206", + "issue:44222", + "issue:44261", + "issue:44263", + "issue:44279", + "issue:44295", + "issue:44297", + "issue:44360", + "issue:44373", + "issue:44448", + "issue:44484", + "issue:44485", + "issue:44623", + "issue:44734", + "issue:44871", + "issue:44977", + "issue:45071", + "issue:45092", + "issue:45103", + "issue:45245", + "issue:45310", + "issue:45357", + "issue:45381", + "issue:45834", + "issue:46082", + "issue:46132", + "issue:46209", + "issue:46305", + "issue:46328", + "issue:46421", + "issue:46424", + "issue:46431", + "issue:46655", + "issue:46665" + ], + "soft_pairs": [ + "issue:44448|issue:46132", + "issue:43479|issue:43644", + "issue:42796|issue:43723", + "issue:45245|issue:46132", + "issue:43479|issue:46665", + "issue:41628|issue:42757", + "issue:44297|issue:45310", + "issue:45071|issue:46082", + "issue:44373|issue:44485", + "issue:44263|issue:45092", + "issue:43723|issue:44977", + "issue:43122|issue:46209", + "issue:44222|issue:44484", + "issue:43232|issue:46424", + "issue:44206|issue:45103", + "issue:46209|issue:46431", + "issue:44190|issue:46082", + "issue:44734|issue:46421", + "issue:43232|issue:45834", + "issue:44295|issue:44623", + "issue:43475|issue:46305", + "issue:46305|issue:46655", + "issue:43723|issue:44871", + "issue:44263|issue:44360", + "issue:45381|issue:46305", + "issue:44261|issue:45092", + "issue:30990|issue:38175", + "issue:43644|issue:46305", + "issue:42831|issue:45310", + "issue:45092|issue:45357", + "issue:44279|issue:46328", + "issue:44263|issue:44373" + ], + "prepared_review_unit_hash": "e4eeb8bf5fe5d73a9acab4e9ea985a27da5d80aae39873509003366f29c11807", + "component_budget": { + "node_count": 473, + "item_count": 473, + "soft_pair_count": 1738, + "serialized_chars": 901437, + "estimated_input_tokens": 225360, + "estimated_eval_tokens": 450976 + }, + "budget": { + "node_count": 47, + "item_count": 47, + "soft_pair_count": 32, + "serialized_chars": 36407, + "estimated_input_tokens": 9102, + "estimated_eval_tokens": 18460 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": true, + "analyst_result": { + "summary": "This cluster looks like a grab-bag of unrelated Transformers bugs/regressions rather than one duplicate family. Based on titles alone, none of the soft pairs are safe duplicates; the closest umbrella report is the generic tokenizer-loading regression, but it is still too broad to serve as a true canonical duplicate.", + "confidence": 0.86, + "canonical_issue_reason": "Issue 43723 is the broadest generic loading-regression report in the set, so it is the best single representative if one must be chosen; however, the cluster is not actually a single duplicate thread.", + "canonical_pr_reason": null, + "best_issue_reason": "43723 is the most umbrella-like issue here because it covers a general v5 loading failure, but it still does not subsume the model-specific/import/config bugs in the cluster.", + "best_pr_reason": null, + "soft_edge_verdicts": [ + { + "left": "issue:44448", + "right": "issue:46132", + "accept": false, + "reason": "Different code paths: generation output mismatch vs attention registration behavior." + }, + { + "left": "issue:43479", + "right": "issue:43644", + "accept": false, + "reason": "Config initialization bug vs non-persistent buffer initialization bug; not the same defect." + }, + { + "left": "issue:42796", + "right": "issue:43723", + "accept": false, + "reason": "Model-specific tokenizer BPE handling vs generic AutoTokenizer loading failure." + }, + { + "left": "issue:45245", + "right": "issue:46132", + "accept": false, + "reason": "Category-cardinality runtime error is unrelated to AttentionInterface registration." + }, + { + "left": "issue:43479", + "right": "issue:46665", + "accept": false, + "reason": "Different from_pretrained/config construction problems with no shared code-path." + }, + { + "left": "issue:41628", + "right": "issue:42757", + "accept": false, + "reason": "Both are import failures, but for different symbols and different dependency paths." + }, + { + "left": "issue:44297", + "right": "issue:45310", + "accept": false, + "reason": "Tokenizer save metadata mismatch vs Qwen3.5 from_pretrained error; different bug." + }, + { + "left": "issue:45071", + "right": "issue:46082", + "accept": false, + "reason": "Type-checking regression is unrelated to LlamaConfig head_dim validation." + }, + { + "left": "issue:44373", + "right": "issue:44485", + "accept": false, + "reason": "Wrong docstring is not the same as a RoPE implementation bug." + }, + { + "left": "issue:44263", + "right": "issue:45092", + "accept": false, + "reason": "DSA split/indexing bug vs remote-code meta-init incompatibility; different failures." + }, + { + "left": "issue:43723", + "right": "issue:44977", + "accept": false, + "reason": "Tokenizer loading issue vs flash-attention generation bug; not the same underlying problem." + }, + { + "left": "issue:43122", + "right": "issue:46209", + "accept": false, + "reason": "Tokenization regression across versions is unrelated to FP8 inference on B200." + }, + { + "left": "issue:44222", + "right": "issue:44484", + "accept": false, + "reason": "FP8 save_pretrained serialization bug vs shard-size default behavior." + }, + { + "left": "issue:43232", + "right": "issue:46424", + "accept": false, + "reason": "Generation kwargs update bug vs cache-length configuration mismatch; related theme, different defect." + }, + { + "left": "issue:44206", + "right": "issue:45103", + "accept": false, + "reason": "Feature extractor argument handling vs auto_docstring annotation crash; separate subsystems." + }, + { + "left": "issue:46209", + "right": "issue:46431", + "accept": false, + "reason": "FP8 inference failure is unrelated to missing chat template support." + }, + { + "left": "issue:44190", + "right": "issue:46082", + "accept": false, + "reason": "Dataset loading script issue vs configuration validation issue." + }, + { + "left": "issue:44734", + "right": "issue:46421", + "accept": false, + "reason": "Serve KV-cache indexing crash vs lazy cache recompilation; both cache-related but different bugs." + }, + { + "left": "issue:43232", + "right": "issue:45834", + "accept": false, + "reason": "Generation kwargs bug is not the same as Kosmos2.5 OCR long-input index error." + }, + { + "left": "issue:44295", + "right": "issue:44623", + "accept": false, + "reason": "Buffer read issue vs missing files on processor save; different persistence paths." + }, + { + "left": "issue:43475", + "right": "issue:46305", + "accept": false, + "reason": "SAM3 video attribute error vs MoonshineStreaming mask off-by-one; unrelated models and code." + }, + { + "left": "issue:46305", + "right": "issue:46655", + "accept": false, + "reason": "Both concern video handling, but one is a sliding-window mask bug and the other is padding calculation in a different processor." + }, + { + "left": "issue:43723", + "right": "issue:44871", + "accept": false, + "reason": "Generic tokenizer loading regression vs eos_token_id config inconsistency." + }, + { + "left": "issue:44263", + "right": "issue:44360", + "accept": false, + "reason": "Same subsystem name, but torch.split return values and missing ReLU are distinct DSA-indexer bugs." + }, + { + "left": "issue:45381", + "right": "issue:46305", + "accept": false, + "reason": "Different models and failure modes: vision_position_ids vs sliding-window masking." + }, + { + "left": "issue:44261", + "right": "issue:45092", + "accept": false, + "reason": "Layernorm epsilon precision issue is unrelated to meta-initialization checkpoint compatibility." + }, + { + "left": "issue:30990", + "right": "issue:38175", + "accept": false, + "reason": "Sentence Transformers loading stall and zero-probability SigLIP issue are unrelated." + }, + { + "left": "issue:43644", + "right": "issue:46305", + "accept": false, + "reason": "Non-persistent buffer junk initialization is not the same as MoonshineStreaming mask logic." + }, + { + "left": "issue:42831", + "right": "issue:45310", + "accept": false, + "reason": "FP8 accuracy concern vs Qwen3.5 from_pretrained failure; different defect class." + }, + { + "left": "issue:45092", + "right": "issue:45357", + "accept": false, + "reason": "Different models and code paths: InternVL2 meta-init vs Qwen3.5 visual-key save_pretrained regression." + }, + { + "left": "issue:44263", + "right": "issue:44373", + "accept": false, + "reason": "DSA indexing behavior and a wrong docstring are not duplicates." + } + ] + }, + "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": 21, + "review_unit_index": 45, + "review_unit_count": 56, + "status": "reviewed", + "reason": null, + "source": "llm", + "cache_hit": false, + "model": "gpt-5.4-mini?service_tier=flex", + "cluster_id": "cluster-10000-473", + "nodes": [ + "issue:22355", + "issue:27957", + "issue:28218", + "issue:30990", + "issue:33357", + "issue:39401", + "issue:41628", + "issue:42371", + "issue:42757", + "issue:42796", + "issue:42831", + "issue:42886", + "issue:42907", + "issue:42994", + "issue:43066", + "issue:43122", + "issue:43232", + "issue:43408", + "issue:43550", + "issue:43576", + "issue:43646", + "issue:43756", + "issue:43784", + "issue:43824", + "issue:43976", + "issue:44038", + "issue:44060", + "issue:44117", + "issue:44188", + "issue:44261", + "issue:44263", + "issue:44295", + "issue:44297", + "issue:44373", + "issue:44483", + "issue:44568", + "issue:44704", + "issue:44829", + "issue:44945", + "issue:45278", + "issue:45406", + "issue:45593", + "issue:45753", + "issue:46002", + "issue:46050", + "issue:46132", + "issue:46369", + "issue:46402" + ], + "soft_pairs": [ + "issue:42907|issue:46050", + "issue:44829|issue:46002", + "issue:42831|issue:44038", + "issue:22355|issue:27957", + "issue:44483|issue:45406", + "issue:41628|issue:43824", + "issue:43408|issue:46402", + "issue:42831|issue:45753", + "issue:22355|issue:28218", + "issue:42757|issue:45278", + "issue:39401|issue:44568", + "issue:42886|issue:43066", + "issue:27957|issue:30990", + "issue:43232|issue:45593", + "issue:42886|issue:42994", + "issue:43408|issue:46369", + "issue:44188|issue:44373", + "issue:42796|issue:43122", + "issue:42371|issue:44263", + "issue:43646|issue:46002", + "issue:43550|issue:46132", + "issue:44261|issue:44373", + "issue:43576|issue:43784", + "issue:44295|issue:44297", + "issue:42907|issue:44945", + "issue:30990|issue:33357", + "issue:43756|issue:44060", + "issue:43976|issue:44945", + "issue:44117|issue:44704" + ], + "prepared_review_unit_hash": "94bfd6ba4b9b2579f68fafbf2b57b57e09a17f654bbaca6fe5b44e8e4e645b67", + "component_budget": { + "node_count": 473, + "item_count": 473, + "soft_pair_count": 1738, + "serialized_chars": 901437, + "estimated_input_tokens": 225360, + "estimated_eval_tokens": 450976 + }, + "budget": { + "node_count": 48, + "item_count": 48, + "soft_pair_count": 29, + "serialized_chars": 35882, + "estimated_input_tokens": 8971, + "estimated_eval_tokens": 18198 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": true, + "analyst_result": { + "summary": "These are mostly unrelated closed issues from the same repo; the soft matches are largely superficial (shared keywords like tokenizer, import, save/load, or sam3) rather than the same underlying bug. No PRs are present.", + "confidence": 0.81, + "canonical_issue_reason": "Issue #45278 is the broadest umbrella-style report in the set and best reflects the cluster\u2019s common pattern of version-related breakage and import regressions.", + "canonical_pr_reason": null, + "best_issue_reason": "#45278 is the most representative global anchor because it summarizes a wide class of failures after upgrade, making it the most useful issue to group near-related regressions around.", + "best_pr_reason": null, + "soft_edge_verdicts": [ + { + "left": "issue:42907", + "right": "issue:46050", + "accept": false, + "reason": "Both involve quantization/save behavior, but one is dequantized Ministral/Devstral saving and the other is sinq quantized round-tripping; different code paths and symptoms." + }, + { + "left": "issue:44829", + "right": "issue:46002", + "accept": false, + "reason": "Both mention bad training/NaNs, but one is flash_attention_3 classification degeneration and the other is init/checkout NaNs; not the same bug." + }, + { + "left": "issue:42831", + "right": "issue:44038", + "accept": false, + "reason": "FineGrainedFP8 accuracy loss and Qwen3-VL-Moe v5 breakage are unrelated model-specific issues." + }, + { + "left": "issue:22355", + "right": "issue:27957", + "accept": false, + "reason": "Module import failure for transformers.onnx is unrelated to XLMRoberta Flash Attention 2 behavior." + }, + { + "left": "issue:44483", + "right": "issue:45406", + "accept": false, + "reason": "Chat-completions request validation and Gemma4Processor attribute crash are different serving/API failures." + }, + { + "left": "issue:41628", + "right": "issue:43824", + "accept": false, + "reason": "Both are import errors, but for different symbols and different models; too broad to be the same defect." + }, + { + "left": "issue:43408", + "right": "issue:46402", + "accept": false, + "reason": "Both are SAM3-related, but one is a model-type warning and the other is missing text-encoder weights; not the same underlying issue." + }, + { + "left": "issue:42831", + "right": "issue:45753", + "accept": false, + "reason": "FineGrainedFP8 accuracy regression and Qwen3_5 infinite loop on one image are unrelated." + }, + { + "left": "issue:22355", + "right": "issue:28218", + "accept": false, + "reason": "ONNX import failure and added-token spacing regression are separate tokenizer/export problems." + }, + { + "left": "issue:42757", + "right": "issue:45278", + "accept": false, + "reason": "A specific is_offline_mode import error is not the same bug as a broad report of many import errors after upgrade." + }, + { + "left": "issue:39401", + "right": "issue:44568", + "accept": false, + "reason": "Wrong offset_mapping in Qwen3 and missing BOS/EOS insertion in mdeberta-v3 are different tokenizer behaviors." + }, + { + "left": "issue:42886", + "right": "issue:43066", + "accept": false, + "reason": "Offline cache-loading failure and wrong tokenizer decoder type are distinct tokenizer initialization problems." + }, + { + "left": "issue:27957", + "right": "issue:30990", + "accept": false, + "reason": "Flash Attention 2 on XLMRoberta and Sentence-Transformers loading hang are unrelated." + }, + { + "left": "issue:43232", + "right": "issue:45593", + "accept": false, + "reason": "Generation kwargs after sync_gpus and D-FINE auxiliary-loss handling do not share a code path." + }, + { + "left": "issue:42886", + "right": "issue:42994", + "accept": false, + "reason": "Tokenizer offline cache loading and quantized model saving are different persistence issues." + }, + { + "left": "issue:43408", + "right": "issue:46369", + "accept": false, + "reason": "SAM3 model-type mismatch warning and SAM3 litetext memory overflow are unrelated failures." + }, + { + "left": "issue:44188", + "right": "issue:44373", + "accept": false, + "reason": "Attention-kernel divergence and a docstring typo for position_ids are not the same problem." + }, + { + "left": "issue:42796", + "right": "issue:43122", + "accept": false, + "reason": "Ministral BPE-marker handling and version-to-version tokenization changes are too general to treat as one bug." + }, + { + "left": "issue:42371", + "right": "issue:44263", + "accept": false, + "reason": "TF32 settings guidance and GlmMoeDsaIndexer torch.split behavior are unrelated." + }, + { + "left": "issue:43646", + "right": "issue:46002", + "accept": false, + "reason": "Custom model initialization breakage and NaNs in classification heads are different regressions." + }, + { + "left": "issue:43550", + "right": "issue:46132", + "accept": false, + "reason": "torch.compile with SDPA on Bamba and AttentionInterface.register behavior changes are different attention issues." + }, + { + "left": "issue:44261", + "right": "issue:44373", + "accept": false, + "reason": "MLA q_a_layernorm precision/config bug and a docstring issue are not the same defect." + }, + { + "left": "issue:43576", + "right": "issue:43784", + "accept": false, + "reason": "Broken env command and a sentence-transformers import NameError are unrelated." + }, + { + "left": "issue:44295", + "right": "issue:44297", + "accept": false, + "reason": "Reading a registered position_ids buffer and tokenizer_class mismatch on save are separate serialization bugs." + }, + { + "left": "issue:42907", + "right": "issue:44945", + "accept": false, + "reason": "Quantized saving/dequantization and pipeline-parallelism output errors are unrelated." + }, + { + "left": "issue:30990", + "right": "issue:33357", + "accept": false, + "reason": "Loading hang in Sentence-Transformers and MacOS bus error on CLIP are different failures." + }, + { + "left": "issue:43756", + "right": "issue:44060", + "accept": false, + "reason": "Smollm3 RoPE-layer bug and Qwen3-Next tied-weights warning are unrelated model-definition issues." + }, + { + "left": "issue:43976", + "right": "issue:44945", + "accept": false, + "reason": "Python-version compatibility breakage and pipeline-parallelism output corruption do not match." + }, + { + "left": "issue:44117", + "right": "issue:44704", + "accept": false, + "reason": "TOKENIZER_MAPPING_NAMES returning None and AutoProcessor not forwarding kwargs are different from_pretrained/cached-file issues, but not 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": 21, + "review_unit_index": 46, + "review_unit_count": 56, + "status": "reviewed", + "reason": null, + "source": "llm", + "cache_hit": false, + "model": "gpt-5.4-mini?service_tier=flex", + "cluster_id": "cluster-10000-473", + "nodes": [ + "issue:22355", + "issue:29127", + "issue:29159", + "issue:36296", + "issue:38617", + "issue:39401", + "issue:41628", + "issue:42175", + "issue:42757", + "issue:42898", + "issue:42907", + "issue:43116", + "issue:43122", + "issue:43278", + "issue:43408", + "issue:43452", + "issue:43475", + "issue:43493", + "issue:43784", + "issue:43824", + "issue:43937", + "issue:44038", + "issue:44188", + "issue:44222", + "issue:44295", + "issue:44297", + "issue:44351", + "issue:44360", + "issue:44485", + "issue:44488", + "issue:44521", + "issue:44625", + "issue:44704", + "issue:44779", + "issue:44829", + "issue:44938", + "issue:44945", + "issue:44964", + "issue:45200", + "issue:45216", + "issue:45278", + "issue:45290", + "issue:45478", + "issue:45600", + "issue:45753", + "issue:46326", + "issue:46328", + "issue:46614" + ], + "soft_pairs": [ + "issue:39401|issue:44488", + "issue:41628|issue:45278", + "issue:43784|issue:45600", + "issue:42907|issue:44964", + "issue:42907|issue:43937", + "issue:44360|issue:44485", + "issue:44625|issue:44829", + "issue:39401|issue:44521", + "issue:44704|issue:45290", + "issue:44038|issue:44297", + "issue:43122|issue:46326", + "issue:43122|issue:44521", + "issue:43116|issue:45200", + "issue:43408|issue:43475", + "issue:44038|issue:44295", + "issue:39401|issue:46614", + "issue:42907|issue:43452", + "issue:44038|issue:44222", + "issue:22355|issue:29159", + "issue:45216|issue:45478", + "issue:43824|issue:44351", + "issue:43122|issue:44779", + "issue:42757|issue:44351", + "issue:36296|issue:43116", + "issue:43278|issue:44945", + "issue:42175|issue:44938", + "issue:22355|issue:29127", + "issue:43493|issue:45753", + "issue:44188|issue:46328", + "issue:38617|issue:45278", + "issue:42898|issue:46614", + "issue:43937|issue:44485" + ], + "prepared_review_unit_hash": "e06a0fd47d28f326f7ca8b2655c7b0b251adb3249b192ed068d914b5111976a4", + "component_budget": { + "node_count": 473, + "item_count": 473, + "soft_pair_count": 1738, + "serialized_chars": 901437, + "estimated_input_tokens": 225360, + "estimated_eval_tokens": 450976 + }, + "budget": { + "node_count": 48, + "item_count": 48, + "soft_pair_count": 32, + "serialized_chars": 36997, + "estimated_input_tokens": 9250, + "estimated_eval_tokens": 18756 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": true, + "analyst_result": { + "summary": "This set is a loose cluster of mostly unrelated regression reports (imports, tokenization, multimodal, save/load, and training bugs). I did not find any soft pair that looks like the same underlying issue, so all candidate links are rejected.", + "confidence": 0.84, + "canonical_issue_reason": "issue:43122 is the best representative because it is a high-signal tokenizer regression report and one of the more central version-change bugs in the set.", + "canonical_pr_reason": null, + "best_issue_reason": "issue:43122 is the most suitable issue to anchor this cluster as it captures the broadest recurring theme here: tokenizer behavior changing across versions.", + "best_pr_reason": null, + "soft_edge_verdicts": [ + { + "left": "issue:39401", + "right": "issue:44488", + "accept": false, + "reason": "Both are tokenizer-related, but one is an offset_mapping bug and the other is a model-loading failure; different concrete problems." + }, + { + "left": "issue:41628", + "right": "issue:45278", + "accept": false, + "reason": "One is a specific AutoImageProcessor import error, the other is a broad collection of import errors after upgrade; too general to treat as the same bug." + }, + { + "left": "issue:43784", + "right": "issue:45600", + "accept": false, + "reason": "Both involve imports, but one is a missing nn symbol at import time and the other is stale auto_mappings references; different code paths." + }, + { + "left": "issue:42907", + "right": "issue:44964", + "accept": false, + "reason": "These are separate model save/load failures for different models and APIs." + }, + { + "left": "issue:42907", + "right": "issue:43937", + "accept": false, + "reason": "Dequantized save_pretrained failure and GenerationConfig validation failure are unrelated." + }, + { + "left": "issue:44360", + "right": "issue:44485", + "accept": false, + "reason": "DSA indexer/ReLU discussion is unrelated to GLM-5 RoPE implementation." + }, + { + "left": "issue:44625", + "right": "issue:44829", + "accept": false, + "reason": "Different bug classes: config propagation vs degenerate training with flash attention 3." + }, + { + "left": "issue:39401", + "right": "issue:44521", + "accept": false, + "reason": "Both touch tokenization, but offset mapping and multimodal assistant mask generation are different failures." + }, + { + "left": "issue:44704", + "right": "issue:45290", + "accept": false, + "reason": "AutoProcessor kwarg forwarding and apply_chat_template crashes are different behaviors in different functions." + }, + { + "left": "issue:44038", + "right": "issue:44297", + "accept": false, + "reason": "A Qwen3-VL-Moe model bug is not the same as tokenizer_class metadata mismatch during save_pretrained." + }, + { + "left": "issue:43122", + "right": "issue:46326", + "accept": false, + "reason": "Both mention tokenizer behavior, but one is a general v5 regression and the other is a mistral_common continue_final_message bug." + }, + { + "left": "issue:43122", + "right": "issue:44521", + "accept": false, + "reason": "Version-regression tokenization differences are not the same as assistant_masks being all zero for multimodal inputs." + }, + { + "left": "issue:43116", + "right": "issue:45200", + "accept": false, + "reason": "Multi-label classification example output and mm_token_type_ids handling are unrelated." + }, + { + "left": "issue:43408", + "right": "issue:43475", + "accept": false, + "reason": "Model-type warning and missing fpn_position_embeddings attribute are different SAM3 video issues." + }, + { + "left": "issue:44038", + "right": "issue:44295", + "accept": false, + "reason": "Qwen3-VL-Moe behavior and position_ids buffer reading are unrelated." + }, + { + "left": "issue:39401", + "right": "issue:46614", + "accept": false, + "reason": "Wrong offset_mapping and stale phonemizer language backend are different tokenizer subsystems." + }, + { + "left": "issue:42907", + "right": "issue:43452", + "accept": false, + "reason": "save_pretrained on dequantized models and gguf_file load failures are distinct problems." + }, + { + "left": "issue:44038", + "right": "issue:44222", + "accept": false, + "reason": "MoE model bug and FP8 save_pretrained issue do not describe the same concrete failure." + }, + { + "left": "issue:22355", + "right": "issue:29159", + "accept": false, + "reason": "onnx module import failure and slow/fast tokenizer inconsistency are unrelated." + }, + { + "left": "issue:45216", + "right": "issue:45478", + "accept": false, + "reason": "Both involve Qwen3.5 and version 5.x, but one is save_pretrained corruption and the other is from_pretrained loading failure." + }, + { + "left": "issue:43824", + "right": "issue:44351", + "accept": false, + "reason": "Both are missing imports, but for different symbols and likely different export issues." + }, + { + "left": "issue:43122", + "right": "issue:44779", + "accept": false, + "reason": "Both are tokenizer regressions, but they are model-specific and do not look like the same code-path defect." + }, + { + "left": "issue:42757", + "right": "issue:44351", + "accept": false, + "reason": "Different missing imports from different packages/symbols; not the same underlying bug." + }, + { + "left": "issue:36296", + "right": "issue:43116", + "accept": false, + "reason": "Tensor parallel training bug and example script multi-label classification failure are unrelated." + }, + { + "left": "issue:43278", + "right": "issue:44945", + "accept": false, + "reason": "Embedding dtype drift in evaluate and incorrect pipeline-parallel output are different training/runtime issues." + }, + { + "left": "issue:42175", + "right": "issue:44938", + "accept": false, + "reason": "Backend packaging omission and Python 3.14 load failure are not the same bug." + }, + { + "left": "issue:22355", + "right": "issue:29127", + "accept": false, + "reason": "Transformers.onnx missing module and LayoutLMv3 error messaging are unrelated." + }, + { + "left": "issue:43493", + "right": "issue:45753", + "accept": false, + "reason": "SigLIP2 implementation discrepancy and Qwen3_5 infinite-loop image bug are different model-specific issues." + }, + { + "left": "issue:44188", + "right": "issue:46328", + "accept": false, + "reason": "torch.compile attention-kernel divergence is unrelated to SequenceFeatureExtractor.pad list conversion overhead." + }, + { + "left": "issue:38617", + "right": "issue:45278", + "accept": false, + "reason": "A specific layer_type_validation import error is not the same as the broader multi-import regression report." + }, + { + "left": "issue:42898", + "right": "issue:46614", + "accept": false, + "reason": "Tokenization cleanup behavior in v5 and phonemizer language-state leakage are separate issues." + }, + { + "left": "issue:43937", + "right": "issue:44485", + "accept": false, + "reason": "GenerationConfig validation for GLM-5 and GLM-5 RoPE implementation are different problems." + } + ] + }, + "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": 21, + "review_unit_index": 47, + "review_unit_count": 56, + "status": "reviewed", + "reason": null, + "source": "llm", + "cache_hit": false, + "model": "gpt-5.4-mini?service_tier=flex", + "cluster_id": "cluster-10000-473", + "nodes": [ + "issue:28348", + "issue:33453", + "issue:39401", + "issue:42222", + "issue:42371", + "issue:42673", + "issue:42796", + "issue:43064", + "issue:43066", + "issue:43116", + "issue:43441", + "issue:43493", + "issue:43611", + "issue:43673", + "issue:43825", + "issue:43827", + "issue:43874", + "issue:43906", + "issue:43937", + "issue:43976", + "issue:43992", + "issue:44038", + "issue:44075", + "issue:44117", + "issue:44291", + "issue:44355", + "issue:44380", + "issue:44625", + "issue:44704", + "issue:44779", + "issue:44936", + "issue:45071", + "issue:45072", + "issue:45083", + "issue:45356", + "issue:45357", + "issue:45405", + "issue:45446", + "issue:45478", + "issue:45563", + "issue:45588", + "issue:46002", + "issue:46077", + "issue:46231", + "issue:46328", + "issue:46614", + "issue:46640", + "issue:46655" + ], + "soft_pairs": [ + "issue:43066|issue:44779", + "issue:42673|issue:45072", + "issue:43976|issue:45446", + "issue:43116|issue:43493", + "issue:43611|issue:46002", + "issue:44355|issue:46231", + "issue:42371|issue:43874", + "issue:44038|issue:45478", + "issue:44625|issue:46002", + "issue:45357|issue:45478", + "issue:45083|issue:45446", + "issue:44075|issue:44936", + "issue:42222|issue:43441", + "issue:43116|issue:43825", + "issue:28348|issue:46640", + "issue:43992|issue:44704", + "issue:42673|issue:42796", + "issue:43064|issue:43673", + "issue:42673|issue:46655", + "issue:43827|issue:45446", + "issue:43937|issue:45563", + "issue:33453|issue:39401", + "issue:44380|issue:46077", + "issue:45356|issue:46614", + "issue:44117|issue:45588", + "issue:43673|issue:46328", + "issue:45071|issue:46614", + "issue:44291|issue:45071", + "issue:45405|issue:45446", + "issue:42673|issue:43906", + "issue:39401|issue:44779" + ], + "prepared_review_unit_hash": "79d39a53ab9124f2e6fff626aa086a589f126ab3e33d410f956c47c16ba85333", + "component_budget": { + "node_count": 473, + "item_count": 473, + "soft_pair_count": 1738, + "serialized_chars": 901437, + "estimated_input_tokens": 225360, + "estimated_eval_tokens": 450976 + }, + "budget": { + "node_count": 48, + "item_count": 48, + "soft_pair_count": 31, + "serialized_chars": 36793, + "estimated_input_tokens": 9199, + "estimated_eval_tokens": 18654 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": true, + "analyst_result": { + "summary": "This cluster is not a duplicate set: the items span many unrelated tokenizer, model-loading, training, attention, and documentation bugs. None of the proposed soft edges look like the same underlying bug or change.", + "confidence": 0.96, + "canonical_issue_reason": "No single canonical issue: the items are heterogeneous, and there is no clear duplicate family tying them together.", + "canonical_pr_reason": null, + "best_issue_reason": "issue:44779 is the clearest stand-alone bug report (a concrete tokenizer regression with a v4/v5 behavior difference), but it does not serve as a duplicate anchor for the rest of the set.", + "best_pr_reason": null, + "soft_edge_verdicts": [ + { + "left": "issue:43066", + "right": "issue:44779", + "accept": false, + "reason": "Both are tokenizer regressions, but one is a decoder-type loading bug and the other is incorrect DeepSeek tokenizer output; different failure modes." + }, + { + "left": "issue:42673", + "right": "issue:45072", + "accept": false, + "reason": "VRAM leak in multi-threaded Qwen3 inference vs dtype mismatch in bfloat16 inference; unrelated code paths." + }, + { + "left": "issue:43976", + "right": "issue:45446", + "accept": false, + "reason": "Python-version compatibility bug vs PyTorch version gate/import check bug; not the same issue." + }, + { + "left": "issue:43116", + "right": "issue:43493", + "accept": false, + "reason": "Example-script multi-label empties vs SigLIP2 implementation discrepancy; different models and symptoms." + }, + { + "left": "issue:43611", + "right": "issue:46002", + "accept": false, + "reason": "Model loading/base_model_prefix regression vs NaNs in classification heads; unrelated bugs." + }, + { + "left": "issue:44355", + "right": "issue:46231", + "accept": false, + "reason": "Compiled Python file errors vs documentation parameter-name mismatch; not the same underlying problem." + }, + { + "left": "issue:42371", + "right": "issue:43874", + "accept": false, + "reason": "TF32 behavior settings API issue vs missing image-processor method; unrelated." + }, + { + "left": "issue:44038", + "right": "issue:45478", + "accept": false, + "reason": "Both involve Qwen3 MoE, but they hit different model variants and different from_pretrained/FlashAttention failures." + }, + { + "left": "issue:44625", + "right": "issue:46002", + "accept": false, + "reason": "num_labels propagation bug vs NaNs on checkout/init; distinct failures." + }, + { + "left": "issue:45357", + "right": "issue:45478", + "accept": false, + "reason": "save_pretrained visual-key regression vs from_pretrained load error; separate save/load code paths." + }, + { + "left": "issue:45083", + "right": "issue:45446", + "accept": false, + "reason": "qwen3_omni_moe helper behavior vs flex_attention import version check; unrelated." + }, + { + "left": "issue:44075", + "right": "issue:44936", + "accept": false, + "reason": "Optimizer-argument handling bug vs train-then-evaluate lifecycle failure; not the same bug." + }, + { + "left": "issue:42222", + "right": "issue:43441", + "accept": false, + "reason": "VitPose model breakage vs Ministral FlashAttention failure; no concrete overlap." + }, + { + "left": "issue:43116", + "right": "issue:43825", + "accept": false, + "reason": "Multi-label example script returning empty results vs pipeline translation-support error message; different surfaces." + }, + { + "left": "issue:28348", + "right": "issue:46640", + "accept": false, + "reason": "FlashAttention support request for GPT-2 vs a general Flash Attention 2 issue; too broad and not the same bug." + }, + { + "left": "issue:43992", + "right": "issue:44704", + "accept": false, + "reason": "Missing embed token weights vs AutoProcessor kwargs forwarding; unrelated loading bugs." + }, + { + "left": "issue:42673", + "right": "issue:42796", + "accept": false, + "reason": "VRAM leak in multi-threaded inference vs tokenizer BPE-marker handling; different subsystems." + }, + { + "left": "issue:43064", + "right": "issue:43673", + "accept": false, + "reason": "FSDP/PEFT optimizer-state corruption vs missing GenerationMixin cache in chunked prefill; unrelated." + }, + { + "left": "issue:42673", + "right": "issue:46655", + "accept": false, + "reason": "VRAM leak vs video padding-calculation bug; not the same issue." + }, + { + "left": "issue:43827", + "right": "issue:45446", + "accept": false, + "reason": "Docs still referencing pipeline() vs PyTorch version check bug; unrelated." + }, + { + "left": "issue:43937", + "right": "issue:45563", + "accept": false, + "reason": "Invalid GenerationConfig vs stale warning for num_return_sequences; different behaviors." + }, + { + "left": "issue:33453", + "right": "issue:39401", + "accept": false, + "reason": "Both are tokenizer regressions, but one is loading regression and the other is wrong offset mapping; different concrete bugs." + }, + { + "left": "issue:44380", + "right": "issue:46077", + "accept": false, + "reason": "GPT-2 attention scaling ignored with SDPA/FlashAttention vs Gemma4 mask construction bug; different models and logic." + }, + { + "left": "issue:45356", + "right": "issue:46614", + "accept": false, + "reason": "Tokenizer codec/warning regression vs phonemizer backend state bug; unrelated tokenizer/audio issues." + }, + { + "left": "issue:44117", + "right": "issue:45588", + "accept": false, + "reason": "Tokenizer mapping None assumption vs flash_attention.py crash on s_aux=None; not the same code path." + }, + { + "left": "issue:43673", + "right": "issue:46328", + "accept": false, + "reason": "Missing cache in generation vs inefficient padding conversion in feature extractor; different kind of issue." + }, + { + "left": "issue:45071", + "right": "issue:46614", + "accept": false, + "reason": "PretrainedConfig type-checking regression vs stale phonemizer language backend; unrelated." + }, + { + "left": "issue:44291", + "right": "issue:45071", + "accept": false, + "reason": "init_empty_weights/_is_hf_initialized argument bug vs PretrainedConfig type-checking breakage; different model-loading failures." + }, + { + "left": "issue:45405", + "right": "issue:45446", + "accept": false, + "reason": "PEFT version pin problem vs flex_attention import version check; unrelated dependency/version issues." + }, + { + "left": "issue:42673", + "right": "issue:43906", + "accept": false, + "reason": "VRAM leak in Qwen3 training/inference threads vs an isolated reproduction of another issue; no evidence of 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": 21, + "review_unit_index": 48, + "review_unit_count": 56, + "status": "reviewed", + "reason": null, + "source": "llm", + "cache_hit": false, + "model": "gpt-5.4-mini?service_tier=flex", + "cluster_id": "cluster-10000-473", + "nodes": [ + "issue:38617", + "issue:39401", + "issue:41628", + "issue:42222", + "issue:42371", + "issue:42617", + "issue:43066", + "issue:43097", + "issue:43122", + "issue:43408", + "issue:43450", + "issue:43742", + "issue:43937", + "issue:43986", + "issue:44075", + "issue:44117", + "issue:44220", + "issue:44261", + "issue:44263", + "issue:44410", + "issue:44466", + "issue:44484", + "issue:44485", + "issue:44493", + "issue:44496", + "issue:44829", + "issue:44877", + "issue:44933", + "issue:44977", + "issue:45071", + "issue:45081", + "issue:45216", + "issue:45278", + "issue:45310", + "issue:45341", + "issue:45362", + "issue:45478", + "issue:45542", + "issue:45561", + "issue:45656", + "issue:46002", + "issue:46178", + "issue:46209", + "issue:46231", + "issue:46298", + "issue:46328", + "issue:46512", + "issue:46617" + ], + "soft_pairs": [ + "issue:44466|issue:46002", + "issue:43937|issue:44977", + "issue:43450|issue:43986", + "issue:44410|issue:45310", + "issue:45341|issue:45561", + "issue:44829|issue:45216", + "issue:45341|issue:46328", + "issue:45071|issue:46512", + "issue:44117|issue:45081", + "issue:38617|issue:41628", + "issue:43122|issue:46298", + "issue:42617|issue:45341", + "issue:42222|issue:46178", + "issue:42371|issue:44261", + "issue:45362|issue:46328", + "issue:43408|issue:43742", + "issue:44877|issue:46231", + "issue:42371|issue:46617", + "issue:44484|issue:45310", + "issue:39401|issue:46209", + "issue:43937|issue:44496", + "issue:44933|issue:45278", + "issue:44410|issue:45478", + "issue:38617|issue:44220", + "issue:44075|issue:45656", + "issue:43066|issue:43122", + "issue:42222|issue:45542", + "issue:44263|issue:44485", + "issue:43097|issue:44493" + ], + "prepared_review_unit_hash": "120f59e0dc5b7a59a8c833204301d50ffd9d400c96a65bba2f453fb613ecc27e", + "component_budget": { + "node_count": 473, + "item_count": 473, + "soft_pair_count": 1738, + "serialized_chars": 901437, + "estimated_input_tokens": 225360, + "estimated_eval_tokens": 450976 + }, + "budget": { + "node_count": 48, + "item_count": 48, + "soft_pair_count": 29, + "serialized_chars": 35985, + "estimated_input_tokens": 8997, + "estimated_eval_tokens": 18250 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": true, + "analyst_result": { + "summary": "No pair looks like the same underlying bug or change; most are only thematically similar (tokenizers, config, import errors, testing utils) but point to different failure modes. I\u2019d treat this cluster as noisy rather than duplicate-heavy.", + "confidence": 0.24, + "canonical_issue_reason": "Issue 45278 is the broadest umbrella in the set (\u2018many import errors after update\u2019) and is the closest thing to a cluster-wide regression report, though the items here are still mostly unrelated.", + "canonical_pr_reason": null, + "best_issue_reason": "Issue 45278 is the most general and representative issue for this noisy group; if one issue has to stand in for the cluster, it best captures the broad import/regression theme.", + "best_pr_reason": null, + "soft_edge_verdicts": [ + { + "left": "issue:44466", + "right": "issue:46002", + "accept": false, + "reason": "Different bugs: tied-weight serialization inconsistency vs NaNs during classification head init/checkpointing." + }, + { + "left": "issue:43937", + "right": "issue:44977", + "accept": false, + "reason": "Both affect generation, but one is an invalid GenerationConfig error and the other is a flash-attention behavior issue." + }, + { + "left": "issue:43450", + "right": "issue:43986", + "accept": false, + "reason": "Different video-processing failures: batched shape bug vs missing-torchvision crash in AutoProcessor loading." + }, + { + "left": "issue:44410", + "right": "issue:45310", + "accept": false, + "reason": "Different Qwen3.5-family problems: missing projections in layer 0 vs a from_pretrained MoE load error." + }, + { + "left": "issue:45341", + "right": "issue:45561", + "accept": false, + "reason": "Shared test-utils area, but one is a generic bug and the other is a specific xdist race on captured_info.txt." + }, + { + "left": "issue:44829", + "right": "issue:45216", + "accept": false, + "reason": "One is degenerate training with flash_attention_3; the other is incorrect save_pretrained checkpoint serialization." + }, + { + "left": "issue:45341", + "right": "issue:46328", + "accept": false, + "reason": "Testing-utils bug vs a SequenceFeatureExtractor.pad performance regression; not the same underlying issue." + }, + { + "left": "issue:45071", + "right": "issue:46512", + "accept": false, + "reason": "Type-checking regression in PretrainedConfig vs stale model.config.dtype after from_config(dtype=...)." + }, + { + "left": "issue:44117", + "right": "issue:45081", + "accept": false, + "reason": "Tokenizer loading assumptions vs a Mistral regex patch/backend_tokenizer AttributeError; different code paths." + }, + { + "left": "issue:38617", + "right": "issue:41628", + "accept": false, + "reason": "Both are import errors, but they concern different missing symbols and different API breakages." + }, + { + "left": "issue:43122", + "right": "issue:46298", + "accept": false, + "reason": "Tokenizer output regression vs a noisy logging message; unrelated changes." + }, + { + "left": "issue:42617", + "right": "issue:45341", + "accept": false, + "reason": "Script execution failure vs a testing_utils bug; no shared concrete defect." + }, + { + "left": "issue:42222", + "right": "issue:46178", + "accept": false, + "reason": "Vitpose model breakage and tensorflow_text backend-mapping failure are different loading issues." + }, + { + "left": "issue:42371", + "right": "issue:44261", + "accept": false, + "reason": "TF32 API guidance issue vs MLA q_a_layernorm precision/config bug; not the same change." + }, + { + "left": "issue:45362", + "right": "issue:46328", + "accept": false, + "reason": "Chat crash on a specific Qwen model vs a padding performance issue." + }, + { + "left": "issue:43408", + "right": "issue:43742", + "accept": false, + "reason": "Model-type mismatch warning vs a key error loading MobileLLM; different failures." + }, + { + "left": "issue:44877", + "right": "issue:46231", + "accept": false, + "reason": "Strict config loading issue vs documentation mismatch for set_training()." + }, + { + "left": "issue:42371", + "right": "issue:46617", + "accept": false, + "reason": "TF32 behavior notice vs cache FileNotFoundError for trust_remote_code; unrelated." + }, + { + "left": "issue:44484", + "right": "issue:45310", + "accept": false, + "reason": "Question about default shard size vs a Qwen3.5 MoE from_pretrained error." + }, + { + "left": "issue:39401", + "right": "issue:46209", + "accept": false, + "reason": "Tokenizer offset-mapping bug vs FP8 inference failure on B200; unrelated." + }, + { + "left": "issue:43937", + "right": "issue:44496", + "accept": false, + "reason": "Invalid GenerationConfig vs unrecognized model/model_type configuration." + }, + { + "left": "issue:44933", + "right": "issue:45278", + "accept": false, + "reason": "Both mention import problems, but one is a specific nonexistent import and the other is a broad post-update import regression." + }, + { + "left": "issue:44410", + "right": "issue:45478", + "accept": false, + "reason": "Different reports: missing projections in qwen3next layer 0 vs Qwen3.5 MoE from_pretrained error." + }, + { + "left": "issue:38617", + "right": "issue:44220", + "accept": false, + "reason": "ImportError for a missing symbol vs a feature-extraction bug in _torch_extract_fbank_features()." + }, + { + "left": "issue:44075", + "right": "issue:45656", + "accept": false, + "reason": "SGD arguments ignored vs optimizer.step being called twice with deepspeed; distinct optimizer bugs." + }, + { + "left": "issue:43066", + "right": "issue:43122", + "accept": false, + "reason": "Tokenizer decoder-type regression and tokenization-difference regression are related only broadly, not the same defect." + }, + { + "left": "issue:42222", + "right": "issue:45542", + "accept": false, + "reason": "Vitpose model breakage vs undefined TF backend due to tensorboard-only install; different subsystems." + }, + { + "left": "issue:44263", + "right": "issue:44485", + "accept": false, + "reason": "torch.split return-value handling vs GLM-5 RoPE implementation; different code paths." + }, + { + "left": "issue:43097", + "right": "issue:44493", + "accept": false, + "reason": "Removal of tie_embeddings_and_encoder_decoder vs unexpected key warnings for position id; not the same underlying change." + } + ] + }, + "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": 21, + "review_unit_index": 49, + "review_unit_count": 56, + "status": "reviewed", + "reason": null, + "source": "llm", + "cache_hit": false, + "model": "gpt-5.4-mini?service_tier=flex", + "cluster_id": "cluster-10000-473", + "nodes": [ + "issue:27957", + "issue:38617", + "issue:41628", + "issue:42175", + "issue:42222", + "issue:42757", + "issue:43012", + "issue:43064", + "issue:43097", + "issue:43352", + "issue:43441", + "issue:43452", + "issue:43531", + "issue:43577", + "issue:43673", + "issue:43874", + "issue:43931", + "issue:43986", + "issue:44038", + "issue:44117", + "issue:44261", + "issue:44263", + "issue:44315", + "issue:44360", + "issue:44410", + "issue:44483", + "issue:44484", + "issue:44485", + "issue:44829", + "issue:44933", + "issue:44987", + "issue:45003", + "issue:45071", + "issue:45341", + "issue:45372", + "issue:45375", + "issue:45440", + "issue:45464", + "issue:45478", + "issue:45542", + "issue:45588", + "issue:45753", + "issue:46123", + "issue:46172", + "issue:46231", + "issue:46369", + "issue:46655", + "issue:46664" + ], + "soft_pairs": [ + "issue:42175|issue:42222", + "issue:43673|issue:45341", + "issue:43874|issue:44261", + "issue:41628|issue:45071", + "issue:43012|issue:43577", + "issue:45071|issue:45542", + "issue:43931|issue:44410", + "issue:44263|issue:46172", + "issue:44360|issue:46172", + "issue:44484|issue:45478", + "issue:44483|issue:45464", + "issue:27957|issue:43352", + "issue:43931|issue:44038", + "issue:43441|issue:44829", + "issue:43097|issue:43452", + "issue:45003|issue:45341", + "issue:43986|issue:45341", + "issue:45753|issue:46369", + "issue:44933|issue:46231", + "issue:43874|issue:44263", + "issue:44261|issue:44485", + "issue:44117|issue:44987", + "issue:43673|issue:46123", + "issue:43531|issue:45341", + "issue:45440|issue:46172", + "issue:43064|issue:44315", + "issue:43097|issue:45588", + "issue:38617|issue:42757", + "issue:43986|issue:45372", + "issue:46655|issue:46664", + "issue:43986|issue:44484", + "issue:44410|issue:45375" + ], + "prepared_review_unit_hash": "061c7439529b6fe7d6cae3f020658cd04a9a6c829dccb087b52577f8ae33f314", + "component_budget": { + "node_count": 473, + "item_count": 473, + "soft_pair_count": 1738, + "serialized_chars": 901437, + "estimated_input_tokens": 225360, + "estimated_eval_tokens": 450976 + }, + "budget": { + "node_count": 48, + "item_count": 48, + "soft_pair_count": 32, + "serialized_chars": 37086, + "estimated_input_tokens": 9272, + "estimated_eval_tokens": 18800 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": true, + "analyst_result": { + "summary": "This cluster is a loose collection of Transformers v5-era regressions and import/model-loading bugs across many unrelated subsystems. I do not see any true duplicate pairs here; the similarities are mostly broad subsystem or version overlap, not the same concrete defect.", + "confidence": 0.74, + "canonical_issue_reason": "issue:43097 is the best umbrella representative: it describes a clear v5.0 regression with broad downstream impact and multiple references, making it the most central issue in this mixed regression cluster.", + "canonical_pr_reason": null, + "best_issue_reason": "issue:43097 is the strongest global representative for the cluster because it captures the widest version-breaking regression theme, even though most other items are separate bugs.", + "best_pr_reason": null, + "soft_edge_verdicts": [ + { + "left": "issue:42175", + "right": "issue:42222", + "accept": false, + "reason": "Packaging/backend dependency omission is a different problem from VitPose model breakage; same release context, not the same bug." + }, + { + "left": "issue:43673", + "right": "issue:45341", + "accept": false, + "reason": "Generation cache behavior during chunked prefill is unrelated to a testing_utils bug." + }, + { + "left": "issue:43874", + "right": "issue:44261", + "accept": false, + "reason": "Missing image-patch method in a processor and an MLA RMS norm epsilon issue are different model components and failure modes." + }, + { + "left": "issue:41628", + "right": "issue:45071", + "accept": false, + "reason": "AutoImageProcessor import failure and PretrainedConfig type-checking regression do not share the same code path." + }, + { + "left": "issue:43012", + "right": "issue:43577", + "accept": false, + "reason": "A PyTorch warning on bfloat16 compilation is not the same defect as model dtypes staying float32 after load." + }, + { + "left": "issue:45071", + "right": "issue:45542", + "accept": false, + "reason": "Config type-checking regression and TensorFlow backend installation/selection error are unrelated." + }, + { + "left": "issue:43931", + "right": "issue:44410", + "accept": false, + "reason": "Both are Qwen-related, but one is a weight-shape mismatch on a specific VL model and the other is missing projections in a different architecture." + }, + { + "left": "issue:44263", + "right": "issue:46172", + "accept": false, + "reason": "These are different tensor-operation bugs in different model code paths." + }, + { + "left": "issue:44360", + "right": "issue:46172", + "accept": false, + "reason": "A missing ReLU in the DSA indexer is not the same as the DeepSeek-v4 transpose/cast issue." + }, + { + "left": "issue:44484", + "right": "issue:45478", + "accept": false, + "reason": "save_pretrained shard-size behavior is unrelated to Qwen3.5 MoE from_pretrained loading errors." + }, + { + "left": "issue:44483", + "right": "issue:45464", + "accept": false, + "reason": "Both mention chat/completions and Qwen3.5, but one is a request/endpoint acceptance failure and the other is a streaming inference failure; not enough to treat as the same fix." + }, + { + "left": "issue:27957", + "right": "issue:43352", + "accept": false, + "reason": "Flash Attention 2 support for XLM-Roberta and NemotronH unsupported FA2 are model-specific feature gaps, not the same bug." + }, + { + "left": "issue:43931", + "right": "issue:44038", + "accept": false, + "reason": "Different Qwen3-VL/MoE loading problems with no clear shared concrete defect." + }, + { + "left": "issue:43441", + "right": "issue:44829", + "accept": false, + "reason": "FlashAttention failures in different models and with different symptoms (support failure vs degenerate training) are not the same bug." + }, + { + "left": "issue:43097", + "right": "issue:43452", + "accept": false, + "reason": "Removed tie_embeddings behavior and gguf_file loading breakage are unrelated." + }, + { + "left": "issue:45003", + "right": "issue:45341", + "accept": false, + "reason": "Unsafe sys.modules access and a testing_utils issue are different code paths and fixes." + }, + { + "left": "issue:43986", + "right": "issue:45341", + "accept": false, + "reason": "AutoProcessor/torchvision crash is unrelated to testing_utils." + }, + { + "left": "issue:45753", + "right": "issue:46369", + "accept": false, + "reason": "An infinite loop on a specific image and a memory overflow in another model are distinct failures." + }, + { + "left": "issue:44933", + "right": "issue:46231", + "accept": false, + "reason": "Missing image_utils import and a docs parameter-name mismatch are not the same issue." + }, + { + "left": "issue:43874", + "right": "issue:44263", + "accept": false, + "reason": "Different GLM subsystems and symptoms; not the same defect." + }, + { + "left": "issue:44261", + "right": "issue:44485", + "accept": false, + "reason": "RMS norm epsilon precision issue and GLM-5 RoPE implementation are separate model math problems." + }, + { + "left": "issue:44117", + "right": "issue:44987", + "accept": false, + "reason": "Tokenizer mapping returning None is too broad to equate with a specific model-load failure without shared implementation evidence." + }, + { + "left": "issue:43673", + "right": "issue:46123", + "accept": false, + "reason": "Generation cache missing during chunked prefill is unrelated to MaskGenerationPipeline partial-batch handling." + }, + { + "left": "issue:43531", + "right": "issue:45341", + "accept": false, + "reason": "Qwen3-MoE sliding_window behavior and testing_utils are unrelated." + }, + { + "left": "issue:45440", + "right": "issue:46172", + "accept": false, + "reason": "DeepseekV3MoE divergence from remote implementation and a DeepSeek-v4 tensor cast/transpose bug are not the same concrete problem." + }, + { + "left": "issue:43064", + "right": "issue:44315", + "accept": false, + "reason": "FSDP2/PEFT optimizer-state corruption and Liger Kernel application in model_init are different training bugs." + }, + { + "left": "issue:43097", + "right": "issue:45588", + "accept": false, + "reason": "A removed config utility and a flash_attention integration crash on s_aux=None are unrelated." + }, + { + "left": "issue:38617", + "right": "issue:42757", + "accept": false, + "reason": "Two import errors, but for different symbols and different modules; not the same underlying bug." + }, + { + "left": "issue:43986", + "right": "issue:45372", + "accept": false, + "reason": "Both are dependency/import-related, but one is torchvision absence during video processor loading and the other is a mistral_common ReasoningEffort import breakage." + }, + { + "left": "issue:46655", + "right": "issue:46664", + "accept": false, + "reason": "Both concern SmolVLM video processing, but padding math and num_frames semantics are separate bugs." + }, + { + "left": "issue:43986", + "right": "issue:44484", + "accept": false, + "reason": "Video processor import crash and save_pretrained shard sizing have no overlap." + }, + { + "left": "issue:44410", + "right": "issue:45375", + "accept": false, + "reason": "Both are Qwen3-family config issues, but missing projections and a silently dropped vision config field are different concrete 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": 21, + "review_unit_index": 50, + "review_unit_count": 56, + "status": "reviewed", + "reason": null, + "source": "llm", + "cache_hit": false, + "model": "gpt-5.4-mini?service_tier=flex", + "cluster_id": "cluster-10000-473", + "nodes": [ + "issue:36683", + "issue:38617", + "issue:42890", + "issue:42907", + "issue:43012", + "issue:43064", + "issue:43066", + "issue:43097", + "issue:43208", + "issue:43352", + "issue:43502", + "issue:43550", + "issue:43582", + "issue:43825", + "issue:43873", + "issue:43906", + "issue:43976", + "issue:43986", + "issue:43992", + "issue:44016", + "issue:44038", + "issue:44062", + "issue:44077", + "issue:44117", + "issue:44162", + "issue:44222", + "issue:44351", + "issue:44368", + "issue:44380", + "issue:44410", + "issue:44484", + "issue:44655", + "issue:44704", + "issue:44936", + "issue:44938", + "issue:45103", + "issue:45478", + "issue:45479", + "issue:45584", + "issue:45712", + "issue:45753", + "issue:46050", + "issue:46132", + "issue:46172", + "issue:46178", + "issue:46298", + "issue:46365", + "issue:46421" + ], + "soft_pairs": [ + "issue:43097|issue:46050", + "issue:44038|issue:44410", + "issue:42890|issue:46298", + "issue:38617|issue:44351", + "issue:44062|issue:44117", + "issue:43097|issue:43873", + "issue:44936|issue:46365", + "issue:43502|issue:43992", + "issue:44704|issue:45103", + "issue:44380|issue:46421", + "issue:43986|issue:44117", + "issue:44938|issue:46178", + "issue:43986|issue:44016", + "issue:43208|issue:43550", + "issue:43825|issue:44655", + "issue:42907|issue:44655", + "issue:43012|issue:43550", + "issue:42907|issue:45753", + "issue:45712|issue:46298", + "issue:43992|issue:44484", + "issue:43012|issue:44368", + "issue:43066|issue:43906", + "issue:43352|issue:44380", + "issue:43097|issue:46298", + "issue:43352|issue:43976", + "issue:44162|issue:46172", + "issue:44380|issue:46132", + "issue:45584|issue:46365", + "issue:44077|issue:46365", + "issue:44222|issue:45478", + "issue:44380|issue:45479", + "issue:36683|issue:43582", + "issue:43064|issue:44380" + ], + "prepared_review_unit_hash": "7b69aa90af0f8e5478dcc82ca58c3c7f970d91782a67bba0ee4a5bbc41f20a54", + "component_budget": { + "node_count": 473, + "item_count": 473, + "soft_pair_count": 1738, + "serialized_chars": 901437, + "estimated_input_tokens": 225360, + "estimated_eval_tokens": 450976 + }, + "budget": { + "node_count": 48, + "item_count": 48, + "soft_pair_count": 33, + "serialized_chars": 37394, + "estimated_input_tokens": 9349, + "estimated_eval_tokens": 18954 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": true, + "analyst_result": { + "summary": "This cluster is mostly heterogeneous: several unrelated Transformers regressions and API breakages share broad themes like v5, quantization, or loading, but the soft pairs do not look like true duplicates. I reject all soft edges.", + "confidence": 0.81, + "canonical_issue_reason": "issue:43097 is the broadest, most central v5 regression in the set and has downstream references, so it is the best anchor among these issue reports.", + "canonical_pr_reason": null, + "best_issue_reason": "issue:43097 is the clearest representative user-facing regression here; it is concrete, widely relevant, and better suited as a cluster anchor than the more niche reports.", + "best_pr_reason": null, + "soft_edge_verdicts": [ + { + "left": "issue:43097", + "right": "issue:46050", + "accept": false, + "reason": "Different bug classes: removed config field vs quantized model save/load corruption. Similar save/load wording is not enough." + }, + { + "left": "issue:44038", + "right": "issue:44410", + "accept": false, + "reason": "Both involve Qwen models, but one is a general Transformers 5.0 compatibility bug and the other is a specific missing-projection architecture issue." + }, + { + "left": "issue:42890", + "right": "issue:46298", + "accept": false, + "reason": "Flaky test seeding and noisy logging are unrelated." + }, + { + "left": "issue:38617", + "right": "issue:44351", + "accept": false, + "reason": "Different import failures from different symbols and code paths." + }, + { + "left": "issue:44062", + "right": "issue:44117", + "accept": false, + "reason": "Tokenizer construction with duplicate kwargs is distinct from TOKENIZER_MAPPING_NAMES returning None." + }, + { + "left": "issue:43097", + "right": "issue:43873", + "accept": false, + "reason": "Config-API removal regression and quantization/offloading behavior are separate issues." + }, + { + "left": "issue:44936", + "right": "issue:46365", + "accept": false, + "reason": "Trainer state corruption after train is unrelated to failing to import Trainer because of datasets." + }, + { + "left": "issue:43502", + "right": "issue:43992", + "accept": false, + "reason": "Local-files-only network leakage and missing encoder weights are different loading bugs." + }, + { + "left": "issue:44704", + "right": "issue:45103", + "accept": false, + "reason": "AutoProcessor kwarg passthrough and docstring annotation crashes affect different subsystems." + }, + { + "left": "issue:44380", + "right": "issue:46421", + "accept": false, + "reason": "Ignored attention scaling in GPT2 is unrelated to cache lazy initialization and recompilation." + }, + { + "left": "issue:43986", + "right": "issue:44117", + "accept": false, + "reason": "AutoProcessor/torchvision crash and tokenizer mapping None are unrelated." + }, + { + "left": "issue:44938", + "right": "issue:46178", + "accept": false, + "reason": "Python 3.14 load failure and missing backend registration are different compatibility problems." + }, + { + "left": "issue:43986", + "right": "issue:44016", + "accept": false, + "reason": "A video-processing import crash is not the same as a notebook syntax error." + }, + { + "left": "issue:43208", + "right": "issue:43550", + "accept": false, + "reason": "xLSTM training blockers and Bamba torch.compile/SDPA failure are different model-specific bugs." + }, + { + "left": "issue:43825", + "right": "issue:44655", + "accept": false, + "reason": "Incorrect pipeline error text is unrelated to saving Pipeline objects." + }, + { + "left": "issue:42907", + "right": "issue:44655", + "accept": false, + "reason": "Dequantized model save failure and pipeline save failure are separate save_pretrained problems." + }, + { + "left": "issue:43012", + "right": "issue:43550", + "accept": false, + "reason": "A PyTorch warning on bfloat16 compilation is not the same as a model-specific SDPA compile failure." + }, + { + "left": "issue:42907", + "right": "issue:45753", + "accept": false, + "reason": "Saving dequantized Ministrals/Devstrals and Qwen3_5 image looping are unrelated." + }, + { + "left": "issue:45712", + "right": "issue:46298", + "accept": false, + "reason": "Leftover dummy classes and a noisy log message are not the same bug." + }, + { + "left": "issue:43992", + "right": "issue:44484", + "accept": false, + "reason": "Missing embed token weights on load is unrelated to the default max shard size question." + }, + { + "left": "issue:43012", + "right": "issue:44368", + "accept": false, + "reason": "These are different warnings about different config/compile behaviors." + }, + { + "left": "issue:43066", + "right": "issue:43906", + "accept": false, + "reason": "Wrong tokenizer decoder type and an isolated reproduction of another issue are not the same code-path problem." + }, + { + "left": "issue:43352", + "right": "issue:44380", + "accept": false, + "reason": "Unsupported Flash Attention 2 for Nemotron is distinct from GPT2 attention scaling being ignored under SDPA/FlashAttention." + }, + { + "left": "issue:44162", + "right": "issue:46172", + "accept": false, + "reason": "ESM2 being broken and a DeepSeek tensor-transpose bug are unrelated model implementations." + }, + { + "left": "issue:44380", + "right": "issue:46132", + "accept": false, + "reason": "A backend-specific scaling bug and an AttentionInterface.register behavior change are different changes." + }, + { + "left": "issue:45584", + "right": "issue:46365", + "accept": false, + "reason": "Whisper empty-transcription failure is unrelated to importing Trainer." + }, + { + "left": "issue:44077", + "right": "issue:46365", + "accept": false, + "reason": "patchtsmixer post_init policy and Trainer import failure are unrelated." + }, + { + "left": "issue:44222", + "right": "issue:45478", + "accept": false, + "reason": "Both mention MoE save/load, but one is FP8 save_pretrained and the other is a from_pretrained error; too different to merge as one bug." + }, + { + "left": "issue:36683", + "right": "issue:43582", + "accept": false, + "reason": "Missing Gemma3Config vocab_size and an Apple Silicon allocator TypeError are unrelated." + }, + { + "left": "issue:43064", + "right": "issue:44380", + "accept": false, + "reason": "A distributed Trainer/FSDP2/PEFT optimizer-state bug is not the same as GPT2 attention scaling being ignored." + } + ] + }, + "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": 21, + "review_unit_index": 51, + "review_unit_count": 56, + "status": "reviewed", + "reason": null, + "source": "llm", + "cache_hit": false, + "model": "gpt-5.4-mini?service_tier=flex", + "cluster_id": "cluster-10000-473", + "nodes": [ + "issue:36010", + "issue:36683", + "issue:38617", + "issue:42175", + "issue:42490", + "issue:42757", + "issue:42898", + "issue:42994", + "issue:43012", + "issue:43066", + "issue:43208", + "issue:43299", + "issue:43352", + "issue:43475", + "issue:43502", + "issue:43525", + "issue:43526", + "issue:43611", + "issue:43618", + "issue:43756", + "issue:43761", + "issue:43824", + "issue:43867", + "issue:43901", + "issue:43937", + "issue:43976", + "issue:44038", + "issue:44077", + "issue:44230", + "issue:44380", + "issue:44483", + "issue:44485", + "issue:44514", + "issue:44625", + "issue:44655", + "issue:44683", + "issue:44938", + "issue:45216", + "issue:45276", + "issue:45290", + "issue:45600", + "issue:45698", + "issue:46077", + "issue:46082", + "issue:46172", + "issue:46298", + "issue:46365" + ], + "soft_pairs": [ + "issue:42175|issue:46365", + "issue:43208|issue:43526", + "issue:42490|issue:43012", + "issue:42757|issue:43824", + "issue:43012|issue:44683", + "issue:44483|issue:44514", + "issue:45600|issue:46298", + "issue:44038|issue:44230", + "issue:43618|issue:43761", + "issue:43618|issue:44380", + "issue:45698|issue:46172", + "issue:38617|issue:43502", + "issue:43756|issue:44485", + "issue:43299|issue:43867", + "issue:42994|issue:44655", + "issue:36683|issue:43475", + "issue:44483|issue:45290", + "issue:42898|issue:43066", + "issue:44625|issue:46365", + "issue:45276|issue:46077", + "issue:36010|issue:42757", + "issue:43352|issue:43937", + "issue:43901|issue:44655", + "issue:43611|issue:44655", + "issue:36683|issue:44077", + "issue:43976|issue:44938", + "issue:43867|issue:45216", + "issue:44625|issue:46077", + "issue:46077|issue:46082", + "issue:36683|issue:43525", + "issue:43618|issue:45216", + "issue:36683|issue:44625" + ], + "prepared_review_unit_hash": "8efdf36c0109b910d255e8cdb3a9c0a64f4d9882c1ebe6f54a61c0b4b8b818ed", + "component_budget": { + "node_count": 473, + "item_count": 473, + "soft_pair_count": 1738, + "serialized_chars": 901437, + "estimated_input_tokens": 225360, + "estimated_eval_tokens": 450976 + }, + "budget": { + "node_count": 47, + "item_count": 47, + "soft_pair_count": 32, + "serialized_chars": 36270, + "estimated_input_tokens": 9068, + "estimated_eval_tokens": 18392 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": true, + "analyst_result": { + "summary": "This cluster is mostly a false-positive similarity group: the items span unrelated import errors, config regressions, save/load bugs, backend warnings, and model-specific failures. None of the soft pairs looks like the same underlying bug or a safe PR merge candidate.", + "confidence": 0.86, + "canonical_issue_reason": null, + "canonical_pr_reason": null, + "best_issue_reason": "issue:45698 is the most discussion-heavy and reference-heavy item here, and it best represents the recurring save_pretrained/from_pretrained theme, even though the cluster is not a true duplicate set.", + "best_pr_reason": null, + "soft_edge_verdicts": [ + { + "left": "issue:42175", + "right": "issue:46365", + "accept": false, + "reason": "Different import/dependency failures: TensorFlow missing from a torch install vs Trainer import blocked by datasets." + }, + { + "left": "issue:43208", + "right": "issue:43526", + "accept": false, + "reason": "Unrelated model areas: xLSTM training bugs vs BeitImageProcessorFast reduce_labels behavior." + }, + { + "left": "issue:42490", + "right": "issue:43012", + "accept": false, + "reason": "Local-path import warning and bf16 compile warning are different code paths and symptoms." + }, + { + "left": "issue:42757", + "right": "issue:43824", + "accept": false, + "reason": "Both are import errors, but for different missing symbols and different fixes." + }, + { + "left": "issue:43012", + "right": "issue:44683", + "accept": false, + "reason": "Different torch-related problems: a precision warning vs a compiled flex_attention failure." + }, + { + "left": "issue:44483", + "right": "issue:44514", + "accept": false, + "reason": "Different failures in chat handling: `/v1/chat/completions` request rejection vs batched chat-template crash." + }, + { + "left": "issue:45600", + "right": "issue:46298", + "accept": false, + "reason": "CI break from removed Sam3LiteText configs is unrelated to a noisy logging message." + }, + { + "left": "issue:44038", + "right": "issue:44230", + "accept": false, + "reason": "Different model/features: Qwen3-VL-Moe loading bug vs fp8 support request for Qwen3-VL/Qwen3.5." + }, + { + "left": "issue:43618", + "right": "issue:43761", + "accept": false, + "reason": "Same model family, but different missing outputs: attentions not assigned vs hidden_states not returned." + }, + { + "left": "issue:43618", + "right": "issue:44380", + "accept": false, + "reason": "Different subsystems: CLIP output regression vs GPT2 attention scaling ignored under SDPA/FlashAttention." + }, + { + "left": "issue:45698", + "right": "issue:46172", + "accept": false, + "reason": "Save/load custom-module bug vs a tensor dtype/transpose issue in deepseek-v4." + }, + { + "left": "issue:38617", + "right": "issue:43502", + "accept": false, + "reason": "ImportError for layer_type_validation is unrelated to local_files_only still making API calls." + }, + { + "left": "issue:43756", + "right": "issue:44485", + "accept": false, + "reason": "Same rough topic (RoPE), but different models and no shared concrete failure mode." + }, + { + "left": "issue:43299", + "right": "issue:43867", + "accept": false, + "reason": "Both loading-related, but one is Qwen3VL-Moe breakage and the other is state_dict sorting causing a load error." + }, + { + "left": "issue:42994", + "right": "issue:44655", + "accept": false, + "reason": "Quantized model saving failure and pipeline save_pretrained failure are different serialization paths." + }, + { + "left": "issue:36683", + "right": "issue:43475", + "accept": false, + "reason": "Missing config attribute in Gemma3 vs missing output field in SAM3 Video; different models and bugs." + }, + { + "left": "issue:44483", + "right": "issue:45290", + "accept": false, + "reason": "Request validation at /v1/chat/completions is not the same as tokenizer chat-template crashes on tool-call messages." + }, + { + "left": "issue:42898", + "right": "issue:43066", + "accept": false, + "reason": "Both are v5 tokenizer changes, but one is clean_up_tokenization_spaces behavior and the other is decoder type selection." + }, + { + "left": "issue:44625", + "right": "issue:46365", + "accept": false, + "reason": "Qwen3.5 num_labels propagation is unrelated to Trainer import failing because of datasets." + }, + { + "left": "issue:45276", + "right": "issue:46077", + "accept": false, + "reason": "Gemma4 embedding resize bug and Gemma4 bidirectional-attention mask bug affect different code paths." + }, + { + "left": "issue:36010", + "right": "issue:42757", + "accept": false, + "reason": "Both are import errors, but they concern different removed/relocated symbols." + }, + { + "left": "issue:43352", + "right": "issue:43937", + "accept": false, + "reason": "Nemotron FlashAttention support error and GLM-5 GenerationConfig validation are unrelated model-specific issues." + }, + { + "left": "issue:43901", + "right": "issue:44655", + "accept": false, + "reason": "Docs mismatch for return_all_scores is not the same as pipeline serialization failure." + }, + { + "left": "issue:43611", + "right": "issue:44655", + "accept": false, + "reason": "Loading models with base_model_prefix and saving Pipeline objects are different stages and bugs." + }, + { + "left": "issue:36683", + "right": "issue:44077", + "accept": false, + "reason": "Gemma3 missing vocab_size and patchtsmixer post_init validation are unrelated config issues." + }, + { + "left": "issue:43976", + "right": "issue:44938", + "accept": false, + "reason": "Different Python compatibility failures: 3.9+ regression vs Python 3.14 load failure." + }, + { + "left": "issue:43867", + "right": "issue:45216", + "accept": false, + "reason": "Both involve load/save behavior, but one is state_dict ordering and the other is a Qwen3.5 checkpoint regression." + }, + { + "left": "issue:44625", + "right": "issue:46077", + "accept": false, + "reason": "Qwen3.5 label propagation and Gemma4 bidirectional attention are unrelated features." + }, + { + "left": "issue:46077", + "right": "issue:46082", + "accept": false, + "reason": "Gemma4 attention-mask behavior and LlamaConfig head_dim validation are different model/config bugs." + }, + { + "left": "issue:36683", + "right": "issue:43525", + "accept": false, + "reason": "Different missing config attributes in different models: vocab_size vs pad_token_id." + }, + { + "left": "issue:43618", + "right": "issue:45216", + "accept": false, + "reason": "CLIP output regression and Qwen3.5 save_pretrained checkpoint corruption are unrelated." + }, + { + "left": "issue:36683", + "right": "issue:44625", + "accept": false, + "reason": "Gemma3 config attribute loss and Qwen3.5 num_labels propagation are different config problems." + } + ] + }, + "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": 21, + "review_unit_index": 52, + "review_unit_count": 56, + "status": "reviewed", + "reason": null, + "source": "llm", + "cache_hit": false, + "model": "gpt-5.4-mini?service_tier=flex", + "cluster_id": "cluster-10000-473", + "nodes": [ + "issue:28282", + "issue:33453", + "issue:36296", + "issue:36683", + "issue:41669", + "issue:42890", + "issue:42898", + "issue:43010", + "issue:43064", + "issue:43502", + "issue:43575", + "issue:43618", + "issue:43723", + "issue:43867", + "issue:43874", + "issue:43976", + "issue:43992", + "issue:44062", + "issue:44075", + "issue:44077", + "issue:44112", + "issue:44222", + "issue:44246", + "issue:44355", + "issue:44360", + "issue:44492", + "issue:44556", + "issue:44734", + "issue:44855", + "issue:44857", + "issue:44938", + "issue:44964", + "issue:45071", + "issue:45125", + "issue:45278", + "issue:45310", + "issue:45357", + "issue:45375", + "issue:45440", + "issue:45478", + "issue:45561", + "issue:45600", + "issue:45850", + "issue:46202", + "issue:46285", + "issue:46365", + "issue:46617", + "issue:46664" + ], + "soft_pairs": [ + "issue:43992|issue:45310", + "issue:43064|issue:44492", + "issue:44075|issue:46365", + "issue:43010|issue:45278", + "issue:44355|issue:44855", + "issue:44938|issue:45071", + "issue:43867|issue:45375", + "issue:43867|issue:46202", + "issue:36683|issue:43064", + "issue:42890|issue:45561", + "issue:41669|issue:44246", + "issue:43010|issue:45850", + "issue:36296|issue:43575", + "issue:43010|issue:43976", + "issue:43618|issue:44077", + "issue:43867|issue:45357", + "issue:43992|issue:45478", + "issue:28282|issue:33453", + "issue:43618|issue:46285", + "issue:43618|issue:44857", + "issue:44360|issue:45440", + "issue:42890|issue:44964", + "issue:42890|issue:44112", + "issue:42898|issue:46664", + "issue:42890|issue:45600", + "issue:43874|issue:46617", + "issue:43502|issue:44062", + "issue:44222|issue:45125", + "issue:43723|issue:44556", + "issue:44734|issue:46617" + ], + "prepared_review_unit_hash": "4d6eca840a19d73fbe1b3b53ab595c8c94400f6bdf5520b5f76982de59d67260", + "component_budget": { + "node_count": 473, + "item_count": 473, + "soft_pair_count": 1738, + "serialized_chars": 901437, + "estimated_input_tokens": 225360, + "estimated_eval_tokens": 450976 + }, + "budget": { + "node_count": 48, + "item_count": 48, + "soft_pair_count": 30, + "serialized_chars": 36253, + "estimated_input_tokens": 9064, + "estimated_eval_tokens": 18384 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": true, + "analyst_result": { + "summary": "This cluster is a loose grouping of transformers v5 regression reports, mostly centered on import/loading failures and model/config compatibility breakages. The strongest representative issue is the broad upgrade-regression report, while the soft pairs are not the same underlying bug.", + "confidence": 0.82, + "canonical_issue_reason": "issue:45278 is the broadest and most representative report: it summarizes many import failures after upgrading from 4.57.0 to 5.5.0, matching the overall cluster theme better than narrower model-specific or API-specific reports.", + "canonical_pr_reason": null, + "best_issue_reason": "issue:45278 best captures the cluster\u2019s shared failure mode (post-v5 import/load regressions across models), making it a good umbrella issue for deduping related reports.", + "best_pr_reason": null, + "soft_edge_verdicts": [ + { + "left": "issue:43992", + "right": "issue:45310", + "accept": false, + "reason": "Both are Qwen/encoder loading problems, but they describe different concrete failures: missing embed_tokens.weight vs a Qwen3.5 Moe from_pretrained error. Too different to treat as the same bug." + }, + { + "left": "issue:43064", + "right": "issue:44492", + "accept": false, + "reason": "One is a trainer/FSDP/PEFT optimizer-state corruption bug; the other is a cache-strategy typo. Same broad v5 area, but not the same code path or fix." + }, + { + "left": "issue:44075", + "right": "issue:46365", + "accept": false, + "reason": "Optimizer arguments not being used is unrelated to importing Trainer being blocked by datasets. No shared underlying defect." + }, + { + "left": "issue:43010", + "right": "issue:45278", + "accept": false, + "reason": "Decorating cache/layer update methods with no_grad is a specific training/memory-safety change, while 45278 is a broad import-error regression report. Not the same issue." + }, + { + "left": "issue:44355", + "right": "issue:44855", + "accept": false, + "reason": "Compiled Python file errors and a Python 3.13 torch.jit parsing/IndentationError are different failure mechanisms." + }, + { + "left": "issue:44938", + "right": "issue:45071", + "accept": false, + "reason": "Both involve Python/version compatibility, but one is a load failure on Python 3.14 and the other is a PretrainedConfig type-checking regression in v5.4.0. Not the same bug." + }, + { + "left": "issue:43867", + "right": "issue:45375", + "accept": false, + "reason": "Model loading with sorted state_dict and a missing vision-config field are unrelated Qwen3.5 failures." + }, + { + "left": "issue:43867", + "right": "issue:46202", + "accept": false, + "reason": "State-dict sorting causing load failure and a Gemma pipeline roundtrip issue are distinct model/data-path problems." + }, + { + "left": "issue:36683", + "right": "issue:43064", + "accept": false, + "reason": "Gemma3Config missing vocab_size is a config-attribute bug; the other is optimizer-state mismatch under FSDP2/PEFT. Different layers of the stack." + }, + { + "left": "issue:42890", + "right": "issue:45561", + "accept": false, + "reason": "A missing set_seed in a flaky model integration test is unrelated to xdist workers racing on captured_info.txt in testing utilities." + }, + { + "left": "issue:41669", + "right": "issue:44246", + "accept": false, + "reason": "Removing import-* usage to reduce import time is a performance/refactor issue; import transformers taking long sometimes is a separate symptom and not the same concrete change." + }, + { + "left": "issue:43010", + "right": "issue:45850", + "accept": false, + "reason": "Both are broad complaints about v5 breakage, but one is a specific no_grad decoration request and the other is a generic 'too many models break' meta issue." + }, + { + "left": "issue:36296", + "right": "issue:43575", + "accept": false, + "reason": "Tensor-parallel training bug and OOM when loading Qwen2-57B-A14B-Instruct with tp are related only at a high level; the concrete causes are not established as the same." + }, + { + "left": "issue:43010", + "right": "issue:43976", + "accept": false, + "reason": "Cache update no_grad is a code-path-specific safety fix; Python 3.9/3.10 compatibility for Transformers 5.1.0 is a version-support issue." + }, + { + "left": "issue:43618", + "right": "issue:44077", + "accept": false, + "reason": "CLIPOutput attentions no longer assigned and PatchTSMixer post_init being optional are unrelated model-specific API regressions." + }, + { + "left": "issue:43867", + "right": "issue:45357", + "accept": false, + "reason": "A general load-model state_dict ordering bug and incorrect visual encoder keys in Qwen3.5 save_pretrained are different save/load defects." + }, + { + "left": "issue:43992", + "right": "issue:45478", + "accept": false, + "reason": "Missing embed_tokens.weight in UMT5Encoder and the Qwen3.5 Moe from_pretrained error affect different model families and different loading paths." + }, + { + "left": "issue:28282", + "right": "issue:33453", + "accept": false, + "reason": "AutoModel requiring PyTorch and tokenizer loading regression are both import/load issues, but they are distinct failure modes with different dependencies and code paths." + }, + { + "left": "issue:43618", + "right": "issue:46285", + "accept": false, + "reason": "CLIP attentions assignment and CLIP text-model structure divergence are both CLIP-related, but they are different API changes and fixes." + }, + { + "left": "issue:43618", + "right": "issue:44857", + "accept": false, + "reason": "CLIP attentions missing and LwDetrImageLoss AMP/CUDA crash are unrelated model/runtime bugs." + }, + { + "left": "issue:44360", + "right": "issue:45440", + "accept": false, + "reason": "The DSA indexer discussion and DeepseekV3MoE divergence describe completely different components and problems." + }, + { + "left": "issue:42890", + "right": "issue:44964", + "accept": false, + "reason": "A flaky SAM HQ test due to missing seeds is unrelated to loading Phi-4 multimodal with the latest transformers." + }, + { + "left": "issue:42890", + "right": "issue:44112", + "accept": false, + "reason": "Seed-related test flakiness and a GraniteSpeech stale device override CI failure are different test issues with different root causes." + }, + { + "left": "issue:42890", + "right": "issue:45600", + "accept": false, + "reason": "Missing set_seed in tests and auto_mappings referencing removed configs are unrelated." + }, + { + "left": "issue:43874", + "right": "issue:46617", + "accept": false, + "reason": "A missing image-patch helper in GLM46V and a trust_remote_code cache FileNotFoundError are different multimodal loading problems." + }, + { + "left": "issue:43502", + "right": "issue:44062", + "accept": false, + "reason": "local_files_only being ignored and AddedToken getting duplicate 'special' args are unrelated tokenizer/utilities bugs." + }, + { + "left": "issue:44222", + "right": "issue:45125", + "accept": false, + "reason": "FP8 save_pretrained for MoE and missing _tp_plan for tensor parallelism are separate model capability issues." + }, + { + "left": "issue:43723", + "right": "issue:44556", + "accept": false, + "reason": "Tokenizer loading in v5 and inability to reload a checkpoint after upgrading are both version-regression reports, but not the same concrete bug." + }, + { + "left": "issue:44734", + "right": "issue:46617", + "accept": false, + "reason": "KV-cache continuation crashing in transformers serve and a cache-loading FileNotFoundError for trust_remote_code models are different serving/loading 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": 21, + "review_unit_index": 53, + "review_unit_count": 56, + "status": "reviewed", + "reason": null, + "source": "llm", + "cache_hit": false, + "model": "gpt-5.4-mini?service_tier=flex", + "cluster_id": "cluster-10000-473", + "nodes": [ + "issue:27957", + "issue:30333", + "issue:30990", + "issue:31795", + "issue:33453", + "issue:34689", + "issue:36296", + "issue:38617", + "issue:41628", + "issue:42548", + "issue:42757", + "issue:43010", + "issue:43065", + "issue:43232", + "issue:43299", + "issue:43475", + "issue:43502", + "issue:43526", + "issue:43576", + "issue:43824", + "issue:43827", + "issue:43901", + "issue:44222", + "issue:44230", + "issue:44315", + "issue:44351", + "issue:44355", + "issue:44360", + "issue:44387", + "issue:44485", + "issue:44556", + "issue:44655", + "issue:44749", + "issue:44908", + "issue:44938", + "issue:44961", + "issue:45071", + "issue:45125", + "issue:45375", + "issue:45468", + "issue:45478", + "issue:46132", + "issue:46143", + "issue:46144", + "issue:46209", + "issue:46617", + "issue:46664" + ], + "soft_pairs": [ + "issue:44387|issue:44556", + "issue:43010|issue:43065", + "issue:43824|issue:46617", + "issue:43502|issue:44485", + "issue:43576|issue:46132", + "issue:45125|issue:45478", + "issue:30333|issue:33453", + "issue:44355|issue:44749", + "issue:42548|issue:44351", + "issue:45125|issue:45375", + "issue:43010|issue:43232", + "issue:43502|issue:44222", + "issue:30990|issue:31795", + "issue:43502|issue:44961", + "issue:43827|issue:44961", + "issue:30333|issue:36296", + "issue:43299|issue:44556", + "issue:44355|issue:45071", + "issue:44355|issue:44655", + "issue:33453|issue:34689", + "issue:27957|issue:33453", + "issue:42757|issue:43576", + "issue:43576|issue:45071", + "issue:44315|issue:44360", + "issue:43576|issue:43824", + "issue:31795|issue:38617", + "issue:44961|issue:45468", + "issue:44351|issue:46617", + "issue:44485|issue:44961", + "issue:43576|issue:44938", + "issue:44908|issue:44961", + "issue:44222|issue:44230", + "issue:36296|issue:43901", + "issue:46143|issue:46144", + "issue:36296|issue:43526", + "issue:43475|issue:46617", + "issue:41628|issue:42548", + "issue:44315|issue:46209", + "issue:44485|issue:46664" + ], + "prepared_review_unit_hash": "715ed05dbfb54fd33a6bb04f797c1bc70715353c08764b10c0b4f575828320a1", + "component_budget": { + "node_count": 473, + "item_count": 473, + "soft_pair_count": 1738, + "serialized_chars": 901437, + "estimated_input_tokens": 225360, + "estimated_eval_tokens": 450976 + }, + "budget": { + "node_count": 47, + "item_count": 47, + "soft_pair_count": 39, + "serialized_chars": 38114, + "estimated_input_tokens": 9529, + "estimated_eval_tokens": 19314 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": true, + "analyst_result": { + "summary": "This set is a heterogeneous mix of unrelated Transformers issues: imports, model loading, docs, schedulers, quantization, and model-specific regressions. There is no clear single duplicate cluster; issue 43576 is only the broadest hub among the soft links, not a true shared root cause.", + "confidence": 0.89, + "canonical_issue_reason": "No single issue is a real canonical duplicate target across the cluster. Issue 43576 is the closest hub only because it sits near several v5-regression reports, but its env-command failure is not the same root cause as the others.", + "canonical_pr_reason": null, + "best_issue_reason": "Issue 43576 is the best representative by connectivity and breadth, but it still does not unify the cluster into one underlying bug.", + "best_pr_reason": null, + "soft_edge_verdicts": [ + { + "left": "issue:44387", + "right": "issue:44556", + "accept": false, + "reason": "Both are v5-era failures, but one is int4/OOM memory behavior and the other is checkpoint reload compatibility; different bugs." + }, + { + "left": "issue:43010", + "right": "issue:43065", + "accept": false, + "reason": "Unrelated code paths: cache/layer update decoration vs a dummy Conv2d in Sam3PixelDecoder." + }, + { + "left": "issue:43824", + "right": "issue:46617", + "accept": false, + "reason": "ImportError for a specific model class vs FileNotFoundError when loading trust_remote_code from cache; not the same issue." + }, + { + "left": "issue:43502", + "right": "issue:44485", + "accept": false, + "reason": "local_files_only network leakage vs GLM-5 RoPE implementation; different subsystems and symptoms." + }, + { + "left": "issue:43576", + "right": "issue:46132", + "accept": false, + "reason": "Broken env command in v5 vs AttentionInterface.register behavior change; same broad release, different bug." + }, + { + "left": "issue:45125", + "right": "issue:45478", + "accept": false, + "reason": "Both involve Qwen3.5 MoE loading, but one is missing _tp_plan for tensor parallelism and the other is a broader from_pretrained error; not enough to merge." + }, + { + "left": "issue:30333", + "right": "issue:33453", + "accept": false, + "reason": "MLflow reporting on failed jobs vs tokenizer loading regression; unrelated failures." + }, + { + "left": "issue:44355", + "right": "issue:44749", + "accept": false, + "reason": "Compiled Python file errors vs slowdown after upgrading Transformers; distinct symptoms and code paths." + }, + { + "left": "issue:42548", + "right": "issue:44351", + "accept": false, + "reason": "Importing PreTrainedModel vs importing HybridCache; both import errors but for different symbols." + }, + { + "left": "issue:45125", + "right": "issue:45375", + "accept": false, + "reason": "Tensor-parallel plan missing for Qwen3_5Moe vs missing deepstack_visual_indexes in config; different model fields." + }, + { + "left": "issue:43010", + "right": "issue:43232", + "accept": false, + "reason": "Cache/Layer update no_grad issue vs generation kwargs update after sync_gpus; separate generation concerns." + }, + { + "left": "issue:43502", + "right": "issue:44222", + "accept": false, + "reason": "Offline/local-files-only network issue vs FP8 moe save_pretrained bug; different functionality." + }, + { + "left": "issue:30990", + "right": "issue:31795", + "accept": false, + "reason": "Loading gets stuck for Sentence Transformers vs documentation confusion in model.forward; unrelated." + }, + { + "left": "issue:43502", + "right": "issue:44961", + "accept": false, + "reason": "Network access despite local_files_only vs an unrelated issue titled racoon; no shared bug." + }, + { + "left": "issue:43827", + "right": "issue:44961", + "accept": false, + "reason": "Docs still referencing pipeline() vs unrelated racoon issue." + }, + { + "left": "issue:30333", + "right": "issue:36296", + "accept": false, + "reason": "Failed-job MLflow reporting vs tensor parallel training bug; different areas." + }, + { + "left": "issue:43299", + "right": "issue:44556", + "accept": false, + "reason": "Qwen3VL MoE model loading regression vs checkpoint reload incompatibility after version upgrade; not the same root cause." + }, + { + "left": "issue:44355", + "right": "issue:45071", + "accept": false, + "reason": "Compiled-file runtime errors vs PretrainedConfig type checking; unrelated." + }, + { + "left": "issue:44355", + "right": "issue:44655", + "accept": false, + "reason": "Compiled-file runtime errors vs pipeline save_pretrained failure; different problems." + }, + { + "left": "issue:33453", + "right": "issue:34689", + "accept": false, + "reason": "Tokenizer loading regression vs Llama 3.2 90B Vision model loading breakage; both loading-related but different models and failure modes." + }, + { + "left": "issue:27957", + "right": "issue:33453", + "accept": false, + "reason": "XLMRoberta Flash Attention 2 support vs tokenizer loading regression; unrelated." + }, + { + "left": "issue:42757", + "right": "issue:43576", + "accept": false, + "reason": "is_offline_mode import failure vs broken env command; different breakages." + }, + { + "left": "issue:43576", + "right": "issue:45071", + "accept": false, + "reason": "Broken env command vs PretrainedConfig type-checking regression; not the same bug." + }, + { + "left": "issue:44315", + "right": "issue:44360", + "accept": false, + "reason": "Liger Kernel not applied with model_init vs DSA indexer lacks ReLU; unrelated." + }, + { + "left": "issue:43576", + "right": "issue:43824", + "accept": false, + "reason": "Env command breakage vs Qwen2.5 VL import error; different commands and code paths." + }, + { + "left": "issue:31795", + "right": "issue:38617", + "accept": false, + "reason": "Docs confusion about forward args vs ImportError for layer_type_validation; documentation vs API export bug." + }, + { + "left": "issue:44961", + "right": "issue:45468", + "accept": false, + "reason": "Unrelated issue titled racoon vs Gemma-4 audio positional encoding bug." + }, + { + "left": "issue:44351", + "right": "issue:46617", + "accept": false, + "reason": "HybridCache import error vs cache FileNotFoundError for trust_remote_code; both cache/import themed but different failures." + }, + { + "left": "issue:44485", + "right": "issue:44961", + "accept": false, + "reason": "GLM-5 RoPE implementation vs unrelated racoon issue; no overlap." + }, + { + "left": "issue:43576", + "right": "issue:44938", + "accept": false, + "reason": "Broken env command vs Python 3.14 loading failure; different issues." + }, + { + "left": "issue:44908", + "right": "issue:44961", + "accept": false, + "reason": "Scheduler kwargs bug vs unrelated racoon issue." + }, + { + "left": "issue:44222", + "right": "issue:44230", + "accept": false, + "reason": "FP8 save_pretrained bug vs FP8 inference support request; related domain but not the same concrete bug." + }, + { + "left": "issue:36296", + "right": "issue:43901", + "accept": false, + "reason": "Tensor parallel training bug vs docs still mentioning return_all_scores; unrelated." + }, + { + "left": "issue:46143", + "right": "issue:46144", + "accept": false, + "reason": "RoFormer kwargs propagation bug vs RoFormer attention not using the attention interface; adjacent but distinct." + }, + { + "left": "issue:36296", + "right": "issue:43526", + "accept": false, + "reason": "Tensor parallel training bug vs BeitImageProcessorFast reduce_labels returning one label; unrelated." + }, + { + "left": "issue:43475", + "right": "issue:46617", + "accept": false, + "reason": "Sam3VisionEncoderOutput missing attribute vs trust_remote_code cache FileNotFoundError; different model/load paths." + }, + { + "left": "issue:41628", + "right": "issue:42548", + "accept": false, + "reason": "AutoImageProcessor import failure vs PreTrainedModel import failure; both imports but different missing symbols." + }, + { + "left": "issue:44315", + "right": "issue:46209", + "accept": false, + "reason": "Liger Kernel model_init issue vs FP8 inference on B200; unrelated implementation areas." + }, + { + "left": "issue:44485", + "right": "issue:46664", + "accept": false, + "reason": "GLM-5 RoPE bug vs SmolVLM num_frames meaning inconsistency; unrelated model-specific 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": 21, + "review_unit_index": 54, + "review_unit_count": 56, + "status": "reviewed", + "reason": null, + "source": "llm", + "cache_hit": false, + "model": "gpt-5.4-mini?service_tier=flex", + "cluster_id": "cluster-10000-473", + "nodes": [ + "issue:31795", + "issue:33453", + "issue:41669", + "issue:42371", + "issue:42890", + "issue:43352", + "issue:43408", + "issue:43525", + "issue:43572", + "issue:43576", + "issue:43673", + "issue:43827", + "issue:43976", + "issue:44038", + "issue:44077", + "issue:44230", + "issue:44246", + "issue:44273", + "issue:44295", + "issue:44297", + "issue:44315", + "issue:44351", + "issue:44360", + "issue:44393", + "issue:44492", + "issue:44556", + "issue:44623", + "issue:44704", + "issue:44829", + "issue:44861", + "issue:44908", + "issue:45003", + "issue:45071", + "issue:45092", + "issue:45125", + "issue:45230", + "issue:45278", + "issue:45341", + "issue:45362", + "issue:45381", + "issue:45478", + "issue:45542", + "issue:45753", + "issue:45850", + "issue:46095", + "issue:46218", + "issue:46328", + "issue:46664" + ], + "soft_pairs": [ + "issue:45362|issue:45850", + "issue:43673|issue:44492", + "issue:44623|issue:45230", + "issue:44360|issue:46664", + "issue:44273|issue:45542", + "issue:44038|issue:45125", + "issue:44273|issue:45003", + "issue:44393|issue:45753", + "issue:44393|issue:45071", + "issue:45381|issue:46664", + "issue:43576|issue:44556", + "issue:41669|issue:43673", + "issue:43525|issue:45850", + "issue:44230|issue:46095", + "issue:44556|issue:45092", + "issue:43572|issue:45850", + "issue:45278|issue:45850", + "issue:44492|issue:45850", + "issue:42371|issue:44492", + "issue:41669|issue:46218", + "issue:42890|issue:44908", + "issue:43408|issue:44315", + "issue:44246|issue:44273", + "issue:43352|issue:44393", + "issue:43576|issue:44908", + "issue:44704|issue:44908", + "issue:43976|issue:44393", + "issue:44908|issue:45071", + "issue:44297|issue:45230", + "issue:44295|issue:45230", + "issue:41669|issue:44351", + "issue:44492|issue:46328", + "issue:41669|issue:45341", + "issue:43976|issue:45850", + "issue:44393|issue:45478", + "issue:43976|issue:46218", + "issue:44077|issue:44861", + "issue:43827|issue:44908", + "issue:31795|issue:33453", + "issue:44315|issue:44829", + "issue:43576|issue:44273" + ], + "prepared_review_unit_hash": "76b7305fca47565e695b749088ace364ce46bf73268541b4f38b302d1eddc764", + "component_budget": { + "node_count": 473, + "item_count": 473, + "soft_pair_count": 1738, + "serialized_chars": 901437, + "estimated_input_tokens": 225360, + "estimated_eval_tokens": 450976 + }, + "budget": { + "node_count": 48, + "item_count": 48, + "soft_pair_count": 41, + "serialized_chars": 39432, + "estimated_input_tokens": 9858, + "estimated_eval_tokens": 19972 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": true, + "analyst_result": { + "summary": "The cluster is dominated by broad Transformers v5 regression complaints, but the individual issues mostly point to different concrete failures (model-specific, import/loading, config, processor, or docs). No soft pair looks like a safe duplicate/merge candidate for a PR.", + "confidence": 0.83, + "canonical_issue_reason": "issue:45850 is the broadest umbrella-style report in the set (\u201cSince 5.0 version it breaks too many models\u201d) and best matches the overall regression theme, even though it is not a true duplicate of the others.", + "canonical_pr_reason": null, + "best_issue_reason": "issue:45850 is the most central/general issue for the cluster because it captures the shared v5-breakage theme across multiple reports without being tied to one model or utility.", + "best_pr_reason": null, + "soft_edge_verdicts": [ + { + "left": "issue:45362", + "right": "issue:45850", + "accept": false, + "reason": "Same general v5 frustration, but one is a specific Qwen chat crash and the other is a broad umbrella complaint." + }, + { + "left": "issue:43673", + "right": "issue:44492", + "accept": false, + "reason": "Chunked-prefill cache regression vs. a cache-strategy typo; different bugs." + }, + { + "left": "issue:44623", + "right": "issue:45230", + "accept": false, + "reason": "Specific processor save_pretrained file loss vs. a generic bug report." + }, + { + "left": "issue:44360", + "right": "issue:46664", + "accept": false, + "reason": "Completely different subsystems: DSA/indexer vs. video processor frame semantics." + }, + { + "left": "issue:44273", + "right": "issue:45542", + "accept": false, + "reason": "Lazy loading/import mechanics vs. TensorFlow backend detection; no same code path." + }, + { + "left": "issue:44038", + "right": "issue:45125", + "accept": false, + "reason": "Both involve Qwen3-family models, but one is model support and the other is tensor-parallel metadata." + }, + { + "left": "issue:44273", + "right": "issue:45003", + "accept": false, + "reason": "Both are import/lazy-loading adjacent, but the symptoms and concrete failure modes differ." + }, + { + "left": "issue:44393", + "right": "issue:45753", + "accept": false, + "reason": "Both are Qwen3-VL image issues, but hallucination/bbox output is not the same bug as an infinite loop on one image." + }, + { + "left": "issue:44393", + "right": "issue:45071", + "accept": false, + "reason": "Model output bug vs. a v5 PretrainedConfig type-checking regression." + }, + { + "left": "issue:45381", + "right": "issue:46664", + "accept": false, + "reason": "Both are video-related, but they affect different models and different notions of position/frame handling." + }, + { + "left": "issue:43576", + "right": "issue:44556", + "accept": false, + "reason": "CLI/env command breakage vs. checkpoint reload compatibility; related to v5, but not the same bug." + }, + { + "left": "issue:41669", + "right": "issue:43673", + "accept": false, + "reason": "Import-* performance problem vs. cache-missing generation regression; different issues." + }, + { + "left": "issue:43525", + "right": "issue:45850", + "accept": false, + "reason": "Specific missing config attribute vs. a broad v5-breakage umbrella." + }, + { + "left": "issue:44230", + "right": "issue:46095", + "accept": false, + "reason": "FP8 support request for Qwen models vs. DeepSeek weight-conversion question; unrelated." + }, + { + "left": "issue:44556", + "right": "issue:45092", + "accept": false, + "reason": "Both concern loading older checkpoints in v5, but they target different checkpoint styles and failure causes." + }, + { + "left": "issue:43572", + "right": "issue:45850", + "accept": false, + "reason": "StableLm pad-token config regression is model-specific and not the same as the broad umbrella complaint." + }, + { + "left": "issue:45278", + "right": "issue:45850", + "accept": false, + "reason": "Both are broad v5 regression reports, but one is about import errors after upgrade and the other is a general claim that many models break." + }, + { + "left": "issue:44492", + "right": "issue:45850", + "accept": false, + "reason": "Cache typo is a localized docs/code issue, not the broad v5 breakage umbrella." + }, + { + "left": "issue:42371", + "right": "issue:44492", + "accept": false, + "reason": "TF32 API warning and cache-strategy typo are unrelated." + }, + { + "left": "issue:41669", + "right": "issue:46218", + "accept": false, + "reason": "Import-time performance regression vs. inability to import a specific Qwen module." + }, + { + "left": "issue:42890", + "right": "issue:44908", + "accept": false, + "reason": "Test flakiness due to missing seeds vs. scheduler kwargs being ignored; unrelated." + }, + { + "left": "issue:43408", + "right": "issue:44315", + "accept": false, + "reason": "Model-type mismatch warning vs. Liger Kernel not applied with model_init; different concerns." + }, + { + "left": "issue:44246", + "right": "issue:44273", + "accept": false, + "reason": "Import takes long sometimes vs. lazy loading malfunction; related area, but not the same concrete bug." + }, + { + "left": "issue:43352", + "right": "issue:44393", + "accept": false, + "reason": "Flash-Attention-2 support error vs. Qwen3-VL bounding-box output bug; different failures." + }, + { + "left": "issue:43576", + "right": "issue:44908", + "accept": false, + "reason": "Environment CLI breakage vs. scheduler argument handling; no shared bug." + }, + { + "left": "issue:44704", + "right": "issue:44908", + "accept": false, + "reason": "Processor cached_file kwargs forwarding vs. inverse_sqrt scheduler kwargs; unrelated." + }, + { + "left": "issue:43976", + "right": "issue:44393", + "accept": false, + "reason": "Python version support issue vs. multimodal bounding-box error." + }, + { + "left": "issue:44908", + "right": "issue:45071", + "accept": false, + "reason": "Scheduler kwarg bug vs. PretrainedConfig type checking regression." + }, + { + "left": "issue:44297", + "right": "issue:45230", + "accept": false, + "reason": "Tokenizer save_pretrained mismatch vs. generic bug report." + }, + { + "left": "issue:44295", + "right": "issue:45230", + "accept": false, + "reason": "Position_ids buffer read error vs. generic bug report." + }, + { + "left": "issue:41669", + "right": "issue:44351", + "accept": false, + "reason": "Import-* performance regression vs. missing HybridCache import; different import-related bugs." + }, + { + "left": "issue:44492", + "right": "issue:46328", + "accept": false, + "reason": "Cache-strategy typo vs. feature-extractor padding performance issue." + }, + { + "left": "issue:41669", + "right": "issue:45341", + "accept": false, + "reason": "Import-* slowdown vs. a small testing_utils bug." + }, + { + "left": "issue:43976", + "right": "issue:45850", + "accept": false, + "reason": "Python support regression vs. broad v5-breakage complaint." + }, + { + "left": "issue:44393", + "right": "issue:45478", + "accept": false, + "reason": "Different Qwen3-VL-family failures, but not the same concrete error path." + }, + { + "left": "issue:43976", + "right": "issue:46218", + "accept": false, + "reason": "Python version compatibility issue vs. missing module import." + }, + { + "left": "issue:44077", + "right": "issue:44861", + "accept": false, + "reason": "Optional post_init validation vs. tied-weights helper crash; unrelated." + }, + { + "left": "issue:43827", + "right": "issue:44908", + "accept": false, + "reason": "Docs still referencing removed pipeline() vs. scheduler kwargs bug." + }, + { + "left": "issue:31795", + "right": "issue:33453", + "accept": false, + "reason": "Documentation confusion vs. tokenizer loading regression; only loosely related to model API usage." + }, + { + "left": "issue:44315", + "right": "issue:44829", + "accept": false, + "reason": "Liger Kernel application in training setup vs. flash_attention_3 causing degenerate training; not the same defect." + }, + { + "left": "issue:43576", + "right": "issue:44273", + "accept": false, + "reason": "Broken env command vs. lazy loading malfunction; both are v5-era issues, but different 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": 21, + "review_unit_index": 55, + "review_unit_count": 56, + "status": "reviewed", + "reason": null, + "source": "llm", + "cache_hit": false, + "model": "gpt-5.4-mini?service_tier=flex", + "cluster_id": "cluster-10000-473", + "nodes": [ + "issue:28348", + "issue:30333", + "issue:31515", + "issue:31795", + "issue:33453", + "issue:34689", + "issue:36296", + "issue:38617", + "issue:41669", + "issue:42371", + "issue:42913", + "issue:43066", + "issue:43298", + "issue:43352", + "issue:43502", + "issue:43519", + "issue:43525", + "issue:43572", + "issue:43576", + "issue:43606", + "issue:43618", + "issue:43704", + "issue:44230", + "issue:44261", + "issue:44263", + "issue:44273", + "issue:44351", + "issue:44360", + "issue:44485", + "issue:44492", + "issue:44623", + "issue:44655", + "issue:44861", + "issue:44908", + "issue:44938", + "issue:44998", + "issue:45230", + "issue:45250", + "issue:45290", + "issue:45310", + "issue:45468", + "issue:45850", + "issue:46018", + "issue:46095", + "issue:46172", + "issue:46218", + "issue:46655" + ], + "soft_pairs": [ + "issue:43519|issue:46655", + "issue:44360|issue:45468", + "issue:44655|issue:44861", + "issue:45230|issue:45310", + "issue:44273|issue:44908", + "issue:43576|issue:46218", + "issue:43519|issue:44485", + "issue:43519|issue:44908", + "issue:41669|issue:43704", + "issue:46018|issue:46172", + "issue:31795|issue:34689", + "issue:31515|issue:31795", + "issue:43704|issue:44485", + "issue:44485|issue:45468", + "issue:45850|issue:46218", + "issue:43502|issue:43519", + "issue:42371|issue:43704", + "issue:43519|issue:44623", + "issue:44263|issue:45468", + "issue:43606|issue:45468", + "issue:43298|issue:46095", + "issue:44261|issue:45468", + "issue:43519|issue:43704", + "issue:44351|issue:46218", + "issue:43352|issue:46218", + "issue:42913|issue:43066", + "issue:28348|issue:45250", + "issue:44908|issue:44998", + "issue:44492|issue:44998", + "issue:43618|issue:44998", + "issue:44998|issue:45850", + "issue:44938|issue:44998", + "issue:44861|issue:45290", + "issue:43525|issue:44861", + "issue:31515|issue:38617", + "issue:43572|issue:44861", + "issue:44230|issue:44623", + "issue:28348|issue:43352", + "issue:30333|issue:31515", + "issue:31515|issue:36296", + "issue:44230|issue:45310", + "issue:31515|issue:33453", + "issue:43298|issue:44230" + ], + "prepared_review_unit_hash": "b68388c7da49a2c2946e4d2f710bb59c3b0fcdc90148cd029d542d7a13d42e82", + "component_budget": { + "node_count": 473, + "item_count": 473, + "soft_pair_count": 1738, + "serialized_chars": 901437, + "estimated_input_tokens": 225360, + "estimated_eval_tokens": 450976 + }, + "budget": { + "node_count": 47, + "item_count": 47, + "soft_pair_count": 43, + "serialized_chars": 39013, + "estimated_input_tokens": 9754, + "estimated_eval_tokens": 19764 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": true, + "analyst_result": { + "summary": "The soft pairs are mostly broad subsystem or version-related similarities, but none look like the same concrete bug/change closely enough to merge. The broadest umbrella issue is the general v5 breakage report.", + "confidence": 0.74, + "canonical_issue_reason": "issue:45850 is the strongest umbrella report: it broadly captures many regressions introduced by 5.0, making it the most plausible anchor for nearby duplicates.", + "canonical_pr_reason": null, + "best_issue_reason": "issue:45850 has the widest scope and is likely to attract related duplicate reports; the others are narrower, more specific, or unrelated.", + "best_pr_reason": null, + "soft_edge_verdicts": [ + { + "left": "issue:43519", + "right": "issue:46655", + "accept": false, + "reason": "Both are processor bugs, but they concern different models and different failure modes (timestamp math vs padding math)." + }, + { + "left": "issue:44360", + "right": "issue:45468", + "accept": false, + "reason": "Different model components and unrelated bugs: DSA indexer activation vs Gemma audio positional encoding config." + }, + { + "left": "issue:44655", + "right": "issue:44861", + "accept": false, + "reason": "Both relate to saving/loading internals, but one is pipeline save_pretrained and the other is tied-weights key handling; not the same bug." + }, + { + "left": "issue:45230", + "right": "issue:45310", + "accept": false, + "reason": "The first is an unspecified generic bug report; the second is a specific Qwen3.5 MoE from_pretrained regression." + }, + { + "left": "issue:44273", + "right": "issue:44908", + "accept": false, + "reason": "Lazy loading behavior and inverse_sqrt scheduler kwargs are unrelated code paths." + }, + { + "left": "issue:43576", + "right": "issue:46218", + "accept": false, + "reason": "One is a broken CLI/env command, the other is a missing model import; different problems." + }, + { + "left": "issue:43519", + "right": "issue:44485", + "accept": false, + "reason": "Different model bugs: Qwen3VL processor timestamp calculation vs GLM-5 RoPE implementation." + }, + { + "left": "issue:43519", + "right": "issue:44908", + "accept": false, + "reason": "Processor timestamp math and scheduler kwargs handling are unrelated." + }, + { + "left": "issue:41669", + "right": "issue:43704", + "accept": false, + "reason": "Import-star cleanup/perf regression and VRAM leak are different issues in different code paths." + }, + { + "left": "issue:46018", + "right": "issue:46172", + "accept": false, + "reason": "Both are DeepSeek-V4 related, but one is gating logic and the other is a tensor conversion/transposition bug." + }, + { + "left": "issue:31795", + "right": "issue:34689", + "accept": false, + "reason": "Documentation confusion vs an actual model-loading regression; not the same underlying bug." + }, + { + "left": "issue:31515", + "right": "issue:31795", + "accept": false, + "reason": "One is slow from_pretrained checkpoint loading; the other is confusing forward docs." + }, + { + "left": "issue:43704", + "right": "issue:44485", + "accept": false, + "reason": "VRAM leak in Qwen3ForCausalLM and GLM-5 RoPE implementation are unrelated." + }, + { + "left": "issue:44485", + "right": "issue:45468", + "accept": false, + "reason": "Both are model-specific implementation issues, but they affect different architectures and code paths." + }, + { + "left": "issue:45850", + "right": "issue:46218", + "accept": false, + "reason": "Broad v5 breakage is not the same concrete issue as a specific Qwen3_5ForCausalLM import failure." + }, + { + "left": "issue:43502", + "right": "issue:43519", + "accept": false, + "reason": "Local-files-only network leakage and video timestamp calculation are unrelated." + }, + { + "left": "issue:42371", + "right": "issue:43704", + "accept": false, + "reason": "TF32 configuration guidance and a Qwen3ForCausalLM VRAM leak do not match." + }, + { + "left": "issue:43519", + "right": "issue:44623", + "accept": false, + "reason": "Processor timestamp calculation and processor.save_pretrained missing files are different bugs." + }, + { + "left": "issue:44263", + "right": "issue:45468", + "accept": false, + "reason": "DSA indexer split behavior and Gemma audio positional encoding are unrelated." + }, + { + "left": "issue:43606", + "right": "issue:45468", + "accept": false, + "reason": "CPU offload device mismatch is unrelated to Gemma audio positional encoding." + }, + { + "left": "issue:43298", + "right": "issue:46095", + "accept": false, + "reason": "Continuous batching support for audio models and a DeepSeek-V4 weight conversion question are not the same change." + }, + { + "left": "issue:44261", + "right": "issue:45468", + "accept": false, + "reason": "Both mention model internals, but they are different architectures and different missing-configuration problems." + }, + { + "left": "issue:43519", + "right": "issue:43704", + "accept": false, + "reason": "Video processor timestamp math and multi-thread VRAM leak are unrelated." + }, + { + "left": "issue:44351", + "right": "issue:46218", + "accept": false, + "reason": "Both are import errors, but for different symbols and different root causes." + }, + { + "left": "issue:43352", + "right": "issue:46218", + "accept": false, + "reason": "Missing Flash Attention 2.0 support and a model import failure are different problems." + }, + { + "left": "issue:42913", + "right": "issue:43066", + "accept": false, + "reason": "Both involve tokenizer v5 behavior, but one is a broad behavior difference and the other is a specific decoder-type bug." + }, + { + "left": "issue:28348", + "right": "issue:45250", + "accept": false, + "reason": "Both mention Flash Attention 2.0, but one is GPT2LMHeadModel-specific while the other is a generic feature request." + }, + { + "left": "issue:44908", + "right": "issue:44998", + "accept": false, + "reason": "Scheduler kwargs handling and an unrelated placeholder issue are not the same bug." + }, + { + "left": "issue:44492", + "right": "issue:44998", + "accept": false, + "reason": "A cache-strategy typo is unrelated to the placeholder report." + }, + { + "left": "issue:43618", + "right": "issue:44998", + "accept": false, + "reason": "CLIPOutput attentions regression and the placeholder report are unrelated." + }, + { + "left": "issue:44998", + "right": "issue:45850", + "accept": false, + "reason": "The placeholder issue contains no concrete bug and is not the same as the broad v5 regression report." + }, + { + "left": "issue:44938", + "right": "issue:44998", + "accept": false, + "reason": "Python 3.14 loading failure is unrelated to the placeholder report." + }, + { + "left": "issue:44861", + "right": "issue:45290", + "accept": false, + "reason": "Tied-weight key handling and apply_chat_template tool-call crashes are different code paths." + }, + { + "left": "issue:43525", + "right": "issue:44861", + "accept": false, + "reason": "Missing pad_token_id in a config and tied-weight key handling are unrelated." + }, + { + "left": "issue:31515", + "right": "issue:38617", + "accept": false, + "reason": "Slow from_pretrained checkpoint loading and a missing import symbol are different failures." + }, + { + "left": "issue:43572", + "right": "issue:44861", + "accept": false, + "reason": "Missing pad_token_idx after v5 update and tied-weight key handling are unrelated." + }, + { + "left": "issue:44230", + "right": "issue:44623", + "accept": false, + "reason": "fp8 model support and processor.save_pretrained missing files are different issues." + }, + { + "left": "issue:28348", + "right": "issue:43352", + "accept": false, + "reason": "Both are Flash Attention 2.0 related, but they target different models and different levels of support." + }, + { + "left": "issue:30333", + "right": "issue:31515", + "accept": false, + "reason": "MLFlow reporting on failed training jobs and slow checkpoint loading are unrelated." + }, + { + "left": "issue:31515", + "right": "issue:36296", + "accept": false, + "reason": "Slow loading and tensor-parallel training bugs are different problems." + }, + { + "left": "issue:44230", + "right": "issue:45310", + "accept": false, + "reason": "fp8 inference support and a Qwen3.5 MoE from_pretrained regression are not the same bug." + }, + { + "left": "issue:31515", + "right": "issue:33453", + "accept": false, + "reason": "Slow checkpoint loading and tokenizer loading regression are only loosely related and not the same concrete issue." + }, + { + "left": "issue:43298", + "right": "issue:44230", + "accept": false, + "reason": "Audio continuous batching support and fp8 Qwen3-VL/Qwen3.5 MoE support are different feature requests." + } + ] + }, + "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": 21, + "review_unit_index": 56, + "review_unit_count": 56, + "status": "reviewed", + "reason": null, + "source": "llm", + "cache_hit": false, + "model": "gpt-5.4-mini?service_tier=flex", + "cluster_id": "cluster-10000-473", + "nodes": [ + "issue:37934", + "issue:42936", + "issue:43240", + "issue:43668", + "issue:43717", + "issue:43813", + "issue:44031", + "issue:46018", + "issue:46144" + ], + "soft_pairs": [ + "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:43240|issue:43668", + "issue:43668|issue:43813", + "issue:37934|issue:43668", + "issue:42936|issue:43668" + ], + "prepared_review_unit_hash": "8840a61ba2865c98265e7039292dc328c5e08137d248c3ae53764beaa7d50696", + "component_budget": { + "node_count": 473, + "item_count": 473, + "soft_pair_count": 1738, + "serialized_chars": 901437, + "estimated_input_tokens": 225360, + "estimated_eval_tokens": 450976 + }, + "budget": { + "node_count": 9, + "item_count": 9, + "soft_pair_count": 12, + "serialized_chars": 8610, + "estimated_input_tokens": 2153, + "estimated_eval_tokens": 4562 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": true, + "analyst_result": { + "summary": "These issues span unrelated models and bug classes (norms/config, tokenizers, loss plumbing, gated experts, attention implementation). None of the soft-similarity pairs look like the same underlying defect.", + "confidence": 0.97, + "canonical_issue_reason": null, + "canonical_pr_reason": null, + "best_issue_reason": null, + "best_pr_reason": null, + "soft_edge_verdicts": [ + { + "left": "issue:43717", + "right": "issue:46144", + "accept": false, + "reason": "Mamba/Mamba-2 init_weights behavior is unrelated to RoFormer attention not using the attention interface." + }, + { + "left": "issue:42936", + "right": "issue:46018", + "accept": false, + "reason": "Mask2Former ignore_value handling and DeepSeek-V4 shared expert gating are different models and different bugs." + }, + { + "left": "issue:43668", + "right": "issue:44031", + "accept": false, + "reason": "An incorrect type hint on ModernBERTConfig norm_eps is unrelated to tokenizer regex warning regressions." + }, + { + "left": "issue:42936", + "right": "issue:43240", + "accept": false, + "reason": "Unused ignore_value in Mask2Former is not the same as kwargs not being forwarded to a loss function." + }, + { + "left": "issue:43240", + "right": "issue:46144", + "accept": false, + "reason": "Loss-function kwargs plumbing and RoFormer attention interface usage are distinct code paths." + }, + { + "left": "issue:42936", + "right": "issue:44031", + "accept": false, + "reason": "Mask2Former ignore_value and tokenizer regex warnings have no common underlying bug." + }, + { + "left": "issue:42936", + "right": "issue:46144", + "accept": false, + "reason": "These concern different architectures and different failure modes." + }, + { + "left": "issue:42936", + "right": "issue:43813", + "accept": false, + "reason": "Mask2Former ignore_value usage is unrelated to a typo in a quantization operation name." + }, + { + "left": "issue:43240", + "right": "issue:43668", + "accept": false, + "reason": "Passing kwargs to loss calculation and fixing a config type hint are unrelated." + }, + { + "left": "issue:43668", + "right": "issue:43813", + "accept": false, + "reason": "ModernBERT norm_eps typing and quantization-operation naming are unrelated issues." + }, + { + "left": "issue:37934", + "right": "issue:43668", + "accept": false, + "reason": "A question about whether Llama4TextL2Norm should be RMSNorm is not the same as a ModernBERT type-hint bug." + }, + { + "left": "issue:42936", + "right": "issue:43668", + "accept": false, + "reason": "Mask2Former ignore_value handling and ModernBERT norm_eps typing do not describe 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": 2, + "component_count": 21, + "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": "c2f3895dc22c280d157a87d513f7db316e5ee48c1ced77cbabd6cb0ff06b4f59", + "component_budget": { + "node_count": 3, + "item_count": 3, + "soft_pair_count": 3, + "serialized_chars": 2876, + "estimated_input_tokens": 719, + "estimated_eval_tokens": 1694 + }, + "budget": { + "node_count": 3, + "item_count": 3, + "soft_pair_count": 3, + "serialized_chars": 2876, + "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 not duplicates: one asks for tokenizer truncation-from-beginning behavior, one asks for ReduceLROnPlateau-like scheduling, and one asks for Wav2Vec2 word boundary offsets.", + "confidence": 0.98, + "canonical_issue_reason": "Issue 10519 is the most concrete and self-contained request, but the cluster does not actually contain a single underlying bug/feature shared by all three issues.", + "canonical_pr_reason": null, + "best_issue_reason": "10519 is the best representative only in the sense that it is the clearest standalone feature request; however, it should not be treated as a true duplicate canonical for the other two issues.", + "best_pr_reason": null, + "soft_edge_verdicts": [ + { + "left": "issue:10519", + "right": "issue:11307", + "accept": false, + "reason": "Different feature areas: tokenizer truncation behavior vs. Wav2Vec2 word-offset extraction. No shared bug or code path." + }, + { + "left": "issue:10519", + "right": "issue:11005", + "accept": false, + "reason": "Unrelated requests: text truncation strategy vs. learning-rate scheduler functionality. They do not describe the same underlying change." + }, + { + "left": "issue:11005", + "right": "issue:11307", + "accept": false, + "reason": "Scheduler feature request and Wav2Vec2 timing-offset request are unrelated and would not be merged as one change." + } + ] + }, + "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": 21, + "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": "de205c5bb6c18949f77f68d042f8e2799c5396769ca3d7db97f6b7d4079d0e17", + "component_budget": { + "node_count": 3, + "item_count": 3, + "soft_pair_count": 2, + "serialized_chars": 2479, + "estimated_input_tokens": 620, + "estimated_eval_tokens": 1496 + }, + "budget": { + "node_count": 3, + "item_count": 3, + "soft_pair_count": 2, + "serialized_chars": 2479, + "estimated_input_tokens": 620, + "estimated_eval_tokens": 1496 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": false, + "analyst_result": { + "summary": "These look like three separate issue reports: a Tapas tokenization discrepancy, a private-model login/error-message problem, and a generic model loading failure. The only superficially related pair is the two model-loading issues, but their symptoms and likely root causes do not match closely enough to treat as duplicates.", + "confidence": 0.96, + "canonical_issue_reason": "No single canonical issue: 13244 is about tokenization behavior, while 15980 and 16618 concern model download/loading, and they do not describe the same concrete bug.", + "canonical_pr_reason": null, + "best_issue_reason": "No strong global representative; 16618 has the broadest title, but it is still not a clean stand-in for the other issues because it targets a different failure mode than 15980 and is unrelated to 13244.", + "best_pr_reason": null, + "soft_edge_verdicts": [ + { + "left": "issue:13244", + "right": "issue:16618", + "accept": false, + "reason": "Tokenization mismatch for Tapas vs. a generic model load failure are different code paths and different symptoms." + }, + { + "left": "issue:15980", + "right": "issue:16618", + "accept": false, + "reason": "Both mention model loading, but 15980 is specifically an unauthenticated private-model download/error-message issue, while 16618 is a broader bert-base-uncased load failure." + } + ] + }, + "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": 21, + "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": "88a26c04d2f184501ee74b46162da220d1ea524ac0bbc17c9723ca196e65b288", + "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": "These are two unrelated feature requests in the same broad area of model/runtime support, but they do not describe the same underlying change. One asks for Flash Attention 2 support across more architectures; the other asks for GGUF support across more architectures.", + "confidence": 0.98, + "canonical_issue_reason": "No single canonical duplicate issue: the titles and descriptions point to different feature work and different code paths (Flash Attention 2 vs GGUF support).", + "canonical_pr_reason": null, + "best_issue_reason": "If a representative issue must be chosen, issue #26350 is the better anchor because it has much higher discussion activity, but it is not a true canonical for #33260 since the requests are unrelated.", + "best_pr_reason": null, + "soft_edge_verdicts": [ + { + "left": "issue:26350", + "right": "issue:33260", + "accept": false, + "reason": "Reject: these are distinct feature requests with different underlying changes and target different functionality; they are not duplicates." + } + ] + }, + "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": 21, + "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-27088-8", + "nodes": [ + "issue:27088", + "issue:28936", + "issue:38435", + "issue:43834", + "issue:44365", + "issue:44370", + "issue:46222", + "issue:46226" + ], + "soft_pairs": [ + "issue:27088|issue:44370", + "issue:27088|issue:44365", + "issue:27088|issue:43834", + "issue:27088|issue:38435", + "issue:27088|issue:28936", + "issue:28936|issue:43834", + "issue:28936|issue:44365", + "issue:28936|issue:38435", + "issue:28936|issue:44370", + "issue:38435|issue:44370", + "issue:38435|issue:44365", + "issue:38435|issue:43834", + "issue:38435|issue:46222", + "issue:38435|issue:46226", + "issue:43834|issue:44370", + "issue:43834|issue:44365", + "issue:43834|issue:46222", + "issue:43834|issue:46226", + "issue:44365|issue:44370", + "issue:44365|issue:46222", + "issue:44365|issue:46226", + "issue:44370|issue:46222", + "issue:44370|issue:46226", + "issue:46222|issue:46226" + ], + "prepared_review_unit_hash": "51882bfe4676d7aa79680a550c151fe9ce2b226e5860f083fc83ea5446f18d7a", + "component_budget": { + "node_count": 8, + "item_count": 8, + "soft_pair_count": 24, + "serialized_chars": 13528, + "estimated_input_tokens": 3382, + "estimated_eval_tokens": 7020 + }, + "budget": { + "node_count": 8, + "item_count": 8, + "soft_pair_count": 24, + "serialized_chars": 13528, + "estimated_input_tokens": 3382, + "estimated_eval_tokens": 7020 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": false, + "analyst_result": { + "summary": "This cluster is mostly heterogeneous: several distinct translation-request issues for different languages, plus two small true-duplicate subclusters (the generic template issues and the identical Thai-titled issues).", + "confidence": 0.9, + "canonical_issue_reason": "Issue 43834 is the clearest representative of the repeated generic template subcluster ('[i18n-] Translating docs to '); the language-specific translation requests are separate items, not duplicates of it.", + "canonical_pr_reason": null, + "best_issue_reason": "43834 is the best issue representative because it is the most generic and centrally duplicated template form, whereas the concrete language requests are distinct translation tasks.", + "best_pr_reason": null, + "soft_edge_verdicts": [ + { + "left": "issue:27088", + "right": "issue:44370", + "accept": false, + "reason": "Same i18n theme, but Turkish vs. a generic template issue; different underlying request." + }, + { + "left": "issue:27088", + "right": "issue:44365", + "accept": false, + "reason": "Turkish request and generic template request are not the same issue." + }, + { + "left": "issue:27088", + "right": "issue:43834", + "accept": false, + "reason": "Concrete Turkish translation request is distinct from the generic placeholder template issue." + }, + { + "left": "issue:27088", + "right": "issue:38435", + "accept": false, + "reason": "Different target languages (Turkish vs Romanian); separate translation requests." + }, + { + "left": "issue:27088", + "right": "issue:28936", + "accept": false, + "reason": "Different target languages (Turkish vs Spanish); not the same underlying issue." + }, + { + "left": "issue:28936", + "right": "issue:43834", + "accept": false, + "reason": "Spanish request does not duplicate the generic template issue." + }, + { + "left": "issue:28936", + "right": "issue:44365", + "accept": false, + "reason": "Different language-specific request versus generic template; not a duplicate." + }, + { + "left": "issue:28936", + "right": "issue:38435", + "accept": false, + "reason": "Spanish vs Romanian translation requests are separate." + }, + { + "left": "issue:28936", + "right": "issue:44370", + "accept": false, + "reason": "Spanish request and generic template issue are not the same underlying task." + }, + { + "left": "issue:38435", + "right": "issue:44370", + "accept": false, + "reason": "Romanian request is distinct from the generic template issue." + }, + { + "left": "issue:38435", + "right": "issue:44365", + "accept": false, + "reason": "Romanian translation request is not the same as the generic template issue." + }, + { + "left": "issue:38435", + "right": "issue:43834", + "accept": false, + "reason": "Concrete Romanian request and generic template request are different issues." + }, + { + "left": "issue:38435", + "right": "issue:46222", + "accept": false, + "reason": "Different content and likely different request; not the same underlying issue." + }, + { + "left": "issue:38435", + "right": "issue:46226", + "accept": false, + "reason": "Romanian translation request is unrelated to the Thai-titled issue." + }, + { + "left": "issue:43834", + "right": "issue:44370", + "accept": true, + "reason": "Both are the same generic placeholder template issue with matching wording and no language-specific distinction." + }, + { + "left": "issue:43834", + "right": "issue:44365", + "accept": true, + "reason": "Same generic template title/body pattern; likely duplicate clones of the template issue." + }, + { + "left": "issue:43834", + "right": "issue:46222", + "accept": false, + "reason": "Generic translation template and the Thai-titled issue are not the same request." + }, + { + "left": "issue:43834", + "right": "issue:46226", + "accept": false, + "reason": "Generic template issue does not match the Thai-titled request." + }, + { + "left": "issue:44365", + "right": "issue:44370", + "accept": true, + "reason": "Identical generic placeholder translation issue; same underlying item." + }, + { + "left": "issue:44365", + "right": "issue:46222", + "accept": false, + "reason": "Generic template issue and Thai-titled issue are different requests." + }, + { + "left": "issue:44365", + "right": "issue:46226", + "accept": false, + "reason": "Generic template issue and Thai-titled issue are different requests." + }, + { + "left": "issue:44370", + "right": "issue:46222", + "accept": false, + "reason": "Generic template issue and Thai-titled issue are not duplicates." + }, + { + "left": "issue:44370", + "right": "issue:46226", + "accept": false, + "reason": "Generic template issue and Thai-titled issue are not the same underlying request." + }, + { + "left": "issue:46222", + "right": "issue:46226", + "accept": true, + "reason": "These are identical Thai-titled issues with matching structure, so they look like true duplicates." + } + ] + }, + "evaluator_result": { + "accept": true, + "feedback": "The summary stays within the packet evidence: it distinguishes the language-specific translation requests, groups the identical generic template issues together, and treats the Thai-titled pair as a duplicate subcluster. The soft-edge verdicts are conservative and do not overclaim beyond the visible titles/body similarities." + }, + "evaluator_used": true, + "retried": false, + "accepted_nontrivial_soft_edge": true, + "error_kind": null, + "error_message": null + }, + { + "label": "issue", + "component_index": 6, + "component_count": 21, + "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-36817-18", + "nodes": [ + "issue:36817", + "issue:36879", + "issue:37015", + "issue:41115", + "issue:41211", + "issue:42503", + "issue:42738", + "issue:43671", + "issue:43837", + "issue:43909", + "issue:44121", + "issue:44205", + "issue:45027", + "issue:45177", + "issue:45322", + "issue:46133", + "issue:46645", + "issue:46653" + ], + "soft_pairs": [ + "issue:36817|issue:41211", + "issue:36817|issue:43671", + "issue:36817|issue:36879", + "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:36879|issue:37015", + "issue:37015|issue:42503", + "issue:41115|issue:43837", + "issue:41115|issue:44121", + "issue:41115|issue:42503", + "issue:41115|issue:45027", + "issue:41211|issue:43671", + "issue:42503|issue:43837", + "issue:42503|issue:43671", + "issue:42503|issue:42738", + "issue:42738|issue:45027", + "issue:42738|issue:46653", + "issue:43671|issue:43837", + "issue:43671|issue:45027", + "issue:43671|issue:45177", + "issue:43671|issue:44121", + "issue:43671|issue:43909", + "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:43837|issue:45027", + "issue:37015|issue:45027", + "issue:43837|issue:45177", + "issue:45322|issue:46133", + "issue:42503|issue:46653", + "issue:45027|issue:46653" + ], + "prepared_review_unit_hash": "ab7f4abb56cb759d383c6bf4ccfc89e9440223d20b5054dcb4995e48eebba9f0", + "component_budget": { + "node_count": 18, + "item_count": 18, + "soft_pair_count": 40, + "serialized_chars": 24146, + "estimated_input_tokens": 6037, + "estimated_eval_tokens": 12330 + }, + "budget": { + "node_count": 18, + "item_count": 18, + "soft_pair_count": 40, + "serialized_chars": 24146, + "estimated_input_tokens": 6037, + "estimated_eval_tokens": 12330 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": false, + "analyst_result": { + "summary": "This cluster is a loose family of model-addition/support requests, but the artifacts are not true duplicates. They mostly share the same broad 'add support for model X' pattern rather than the same underlying bug or code change.", + "confidence": 0.87, + "canonical_issue_reason": "Issue 43837 is one of the most central and connected requests in the set, so it is the best single anchor if a representative issue is needed; however, it is still only a representative, not a duplicate of the others.", + "canonical_pr_reason": null, + "best_issue_reason": "43837 is a good cluster representative because it sits near the center of the similarity graph and is a concrete, specific support request rather than a generic placeholder.", + "best_pr_reason": null, + "soft_edge_verdicts": [ + { + "left": "issue:36817", + "right": "issue:41211", + "accept": false, + "reason": "Different models and different requested integrations; only the generic model-addition theme overlaps." + }, + { + "left": "issue:36817", + "right": "issue:43671", + "accept": false, + "reason": "EuroBert config support vs Qwen3 TTS support are unrelated changes." + }, + { + "left": "issue:36817", + "right": "issue:36879", + "accept": false, + "reason": "Different model families entirely; not the same bug or change." + }, + { + "left": "issue:36879", + "right": "issue:41211", + "accept": false, + "reason": "RF-DETR and DEIMv2 are distinct model requests with different code paths." + }, + { + "left": "issue:36879", + "right": "issue:43909", + "accept": false, + "reason": "Different model families and modalities; no common concrete fix." + }, + { + "left": "issue:36879", + "right": "issue:44121", + "accept": false, + "reason": "RF-DETR support and weight-sparse transformer support are unrelated." + }, + { + "left": "issue:36879", + "right": "issue:43837", + "accept": false, + "reason": "Different models and different feature requests; same broad 'add support' pattern only." + }, + { + "left": "issue:36879", + "right": "issue:43671", + "accept": false, + "reason": "RF-DETR vs Qwen3 TTS are unrelated additions." + }, + { + "left": "issue:37015", + "right": "issue:43837", + "accept": false, + "reason": "NeoBERT and Qwen3 ASR are distinct model integrations." + }, + { + "left": "issue:37015", + "right": "issue:44121", + "accept": false, + "reason": "Different model architectures and different implementation work." + }, + { + "left": "issue:37015", + "right": "issue:43909", + "accept": false, + "reason": "NeoBERT vs LFM2.5 Audio are unrelated requests." + }, + { + "left": "issue:36879", + "right": "issue:37015", + "accept": false, + "reason": "RF-DETR and NeoBERT are different model support tasks." + }, + { + "left": "issue:37015", + "right": "issue:42503", + "accept": false, + "reason": "NeoBERT and ModernVBERT are separate model requests, not one change." + }, + { + "left": "issue:41115", + "right": "issue:43837", + "accept": false, + "reason": "MiniCPM3 support and Qwen3 ASR support are different model integrations." + }, + { + "left": "issue:41115", + "right": "issue:44121", + "accept": false, + "reason": "MiniCPM3 and the weight-sparse transformer request are unrelated code changes." + }, + { + "left": "issue:41115", + "right": "issue:42503", + "accept": false, + "reason": "MiniCPM3 and ModernVBERT are distinct model additions." + }, + { + "left": "issue:41115", + "right": "issue:45027", + "accept": false, + "reason": "Different models and different target behavior; not mergeable as one fix." + }, + { + "left": "issue:41211", + "right": "issue:43671", + "accept": false, + "reason": "DEIMv2 and Qwen3 TTS are different model requests with no shared concrete bug." + }, + { + "left": "issue:42503", + "right": "issue:43837", + "accept": false, + "reason": "ModernVBERT models and Qwen3 ASR are unrelated support requests." + }, + { + "left": "issue:42503", + "right": "issue:43671", + "accept": false, + "reason": "ModernVBERT and Qwen3 TTS are different model additions." + }, + { + "left": "issue:42503", + "right": "issue:42738", + "accept": false, + "reason": "ModernVBERT models vs a generic BERT-like RoPE request are not the same underlying change." + }, + { + "left": "issue:42738", + "right": "issue:45027", + "accept": false, + "reason": "A generic RoPE-related BERT-like request and Voxtral TTS support are different features." + }, + { + "left": "issue:42738", + "right": "issue:46653", + "accept": false, + "reason": "Model architecture request vs unrelated personal-name issue; no overlap." + }, + { + "left": "issue:43671", + "right": "issue:43837", + "accept": false, + "reason": "Same Qwen3 family, but TTS and ASR are different modalities and implementation paths." + }, + { + "left": "issue:43671", + "right": "issue:45027", + "accept": false, + "reason": "Qwen3 TTS and Voxtral TTS are different model support requests, not one fix." + }, + { + "left": "issue:43671", + "right": "issue:45177", + "accept": false, + "reason": "Audio TTS support vs DeepSeek-OCR2 support are unrelated." + }, + { + "left": "issue:43671", + "right": "issue:44121", + "accept": false, + "reason": "Different model families and different requested functionality." + }, + { + "left": "issue:43671", + "right": "issue:43909", + "accept": false, + "reason": "Qwen3 TTS and LFM2.5 Audio are distinct audio-model requests." + }, + { + "left": "issue:43837", + "right": "issue:43909", + "accept": false, + "reason": "Qwen3 ASR and LFM2.5 Audio are different model integrations." + }, + { + "left": "issue:43837", + "right": "issue:45322", + "accept": false, + "reason": "ASR support and EUPE support are unrelated model requests." + }, + { + "left": "issue:43837", + "right": "issue:44121", + "accept": false, + "reason": "Qwen3 ASR and weight-sparse transformer support are different changes." + }, + { + "left": "issue:43909", + "right": "issue:44205", + "accept": false, + "reason": "Audio model support and SAM3-LiteText support do not share the same underlying fix." + }, + { + "left": "issue:43909", + "right": "issue:46645", + "accept": false, + "reason": "LFM2.5 Audio support and Granite SWA+sink option are different code paths." + }, + { + "left": "issue:44121", + "right": "issue:45322", + "accept": false, + "reason": "Weight-sparse transformer support and EUPE support are unrelated model requests." + }, + { + "left": "issue:43837", + "right": "issue:45027", + "accept": false, + "reason": "Qwen3 ASR and Voxtral TTS are separate audio-model support requests." + }, + { + "left": "issue:37015", + "right": "issue:45027", + "accept": false, + "reason": "NeoBERT support and Voxtral TTS support are unrelated model additions." + }, + { + "left": "issue:43837", + "right": "issue:45177", + "accept": false, + "reason": "Qwen3 ASR and DeepSeek-OCR2 are different model families and different code paths." + }, + { + "left": "issue:45322", + "right": "issue:46133", + "accept": false, + "reason": "EUPE and TIPSv2 are different model requests, even though both are vision-language related." + }, + { + "left": "issue:42503", + "right": "issue:46653", + "accept": false, + "reason": "ModernVBERT models and the unrelated name issue do not match any shared bug or change." + }, + { + "left": "issue:45027", + "right": "issue:46653", + "accept": false, + "reason": "Voxtral TTS support and the unrelated name issue are not duplicates." + } + ] + }, + "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": 21, + "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": "5f76c80171f44e59f657b934e611e41bf2184a087e0037235417077562cfeea3", + "component_budget": { + "node_count": 2, + "item_count": 2, + "soft_pair_count": 1, + "serialized_chars": 1506, + "estimated_input_tokens": 377, + "estimated_eval_tokens": 1010 + }, + "budget": { + "node_count": 2, + "item_count": 2, + "soft_pair_count": 1, + "serialized_chars": 1506, + "estimated_input_tokens": 377, + "estimated_eval_tokens": 1010 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": false, + "analyst_result": { + "summary": "The two issues are unrelated: one is a docs request to add Pruna as an optimization framework, and the other is a Vision RoPE refactor. They do not describe the same bug or change.", + "confidence": 0.99, + "canonical_issue_reason": null, + "canonical_pr_reason": null, + "best_issue_reason": null, + "best_pr_reason": null, + "soft_edge_verdicts": [ + { + "left": "issue:38740", + "right": "issue:46443", + "accept": false, + "reason": "Different topics and scope: docs/framework listing versus a Vision RoPE code refactor. No shared target, filename, or underlying problem." + } + ] + }, + "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": 21, + "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": "140262fdef78d1e5db59209b6debbeb0823075d65aa42ccf513eddf0f92141c4", + "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 about Set Block Decoding, the other requests DeepSeek v3.2 support. They describe different problem areas with no shared targets or filenames.", + "confidence": 0.97, + "canonical_issue_reason": "No canonical issue for deduping here; the issues are unrelated and should remain separate.", + "canonical_pr_reason": null, + "best_issue_reason": "If a single representative is needed, issue 43037 is the clearer product request, but it is not the same underlying issue as 41084.", + "best_pr_reason": null, + "soft_edge_verdicts": [ + { + "left": "issue:41084", + "right": "issue:43037", + "accept": false, + "reason": "Different topics: decoding behavior vs model compatibility/support request; no evidence they describe 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": 9, + "component_count": 21, + "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": "2824152c316bd9502a8162fedf90f47900dbe18ee8d3fa7c30204c61ce6f599b", + "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": "The two issues are only loosely related by the Trainer/custom trainer theme, but they describe different problems: one is about `remove_unused_column` customization, the other about FLOPS estimation for custom trainers. They should not be merged as duplicates.", + "confidence": 0.88, + "canonical_issue_reason": "Issue #43598 is the broader, more representative Trainer customizability request, while #43600 is a separate FLOPS-estimation concern.", + "canonical_pr_reason": null, + "best_issue_reason": "#43598 is the better cluster representative because it centers on Trainer behavior customization rather than a narrower metrics-estimation detail.", + "best_pr_reason": null, + "soft_edge_verdicts": [ + { + "left": "issue:43598", + "right": "issue:43600", + "accept": false, + "reason": "They share the custom-trainer context, but the underlying problems are different: unused-column handling vs FLOPS estimation. They would not 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": 10, + "component_count": 21, + "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": "6a8c1bd73ce3609c2d8ba9683efa94068f155bb1f43113d62c63888f7ab9d455", + "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 two issues are related only at a very high level (both about adding Qwen model support), but they target different model families and different APIs/tasks, so they should not be treated as duplicates.", + "confidence": 0.97, + "canonical_issue_reason": "No single canonical issue: #43684 is about registering Qwen3-Omni with AutoModel/AutoModelForConditionalGeneration, while #44405 requests AutoModelForSequenceClassification support for Qwen3.5 (Qwen3_5Config). Different models and different heads.", + "canonical_pr_reason": null, + "best_issue_reason": "#43684 is the broader infrastructure-style request, but it is not a duplicate of #44405; the two issues should remain separate.", + "best_pr_reason": null, + "soft_edge_verdicts": [ + { + "left": "issue:43684", + "right": "issue:44405", + "accept": false, + "reason": "Different underlying change: model registration for Qwen3-Omni vs sequence-classification support for Qwen3.5. They do not look mergeable into one fix." + } + ] + }, + "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": 21, + "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": "c7076139e2be8500a39b0ac5d4eb87a7a43b06f0ddc9f3c2ccd7b4f95df5a8da", + "component_budget": { + "node_count": 4, + "item_count": 4, + "soft_pair_count": 6, + "serialized_chars": 4263, + "estimated_input_tokens": 1066, + "estimated_eval_tokens": 2388 + }, + "budget": { + "node_count": 4, + "item_count": 4, + "soft_pair_count": 6, + "serialized_chars": 4263, + "estimated_input_tokens": 1066, + "estimated_eval_tokens": 2388 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": false, + "analyst_result": { + "summary": "Three issues share the same near-empty title \"huggingface\" and are likely duplicates; the \"Fett\" issue appears unrelated and should stay separate.", + "confidence": 0.95, + "canonical_issue_reason": "Issue 43845 is the earliest of the three nearly identical \"huggingface\" reports, so it is the best canonical issue among that duplicate set.", + "canonical_pr_reason": null, + "best_issue_reason": "Among all issues in the cluster, 43845 is the strongest representative of the duplicate huggingface reports; 43835 does not match the same topic.", + "best_pr_reason": null, + "soft_edge_verdicts": [ + { + "left": "issue:43835", + "right": "issue:43845", + "accept": false, + "reason": "The titles do not indicate the same underlying problem; \"Fett\" is unrelated to the repeated \"huggingface\" reports." + }, + { + "left": "issue:43835", + "right": "issue:43846", + "accept": false, + "reason": "No evidence that \"Fett\" refers to the same bug or request as the repeated \"huggingface\" issues." + }, + { + "left": "issue:43835", + "right": "issue:43859", + "accept": false, + "reason": "Different title/topic; this does not look like the same underlying issue." + }, + { + "left": "issue:43845", + "right": "issue:43846", + "accept": true, + "reason": "Same title, same vague topic, and close creation times strongly suggest duplicate reports of the same issue." + }, + { + "left": "issue:43845", + "right": "issue:43859", + "accept": true, + "reason": "These are indistinguishable from the available data: same title and no differing details." + }, + { + "left": "issue:43846", + "right": "issue:43859", + "accept": true, + "reason": "Same short title and no contradicting evidence; likely the same underlying report." + } + ] + }, + "evaluator_result": { + "accept": true, + "feedback": "The assessment is grounded in the packet: 43835 has a different title/topic, while 43845/43846/43859 all share the same title and high soft-similarity. The canonical choice of the earliest huggingface issue is conservative enough. No soft-edge verdict appears to overstate evidence materially." + }, + "evaluator_used": true, + "retried": false, + "accepted_nontrivial_soft_edge": true, + "error_kind": null, + "error_message": null + }, + { + "label": "issue", + "component_index": 12, + "component_count": 21, + "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-44371-5", + "nodes": [ + "issue:44371", + "issue:44418", + "issue:44572", + "issue:44573", + "issue:44937" + ], + "soft_pairs": [ + "issue:44371|issue:44572", + "issue:44371|issue:44573", + "issue:44371|issue:44418", + "issue:44371|issue:44937", + "issue:44418|issue:44572", + "issue:44418|issue:44573", + "issue:44572|issue:44573", + "issue:44572|issue:44937", + "issue:44573|issue:44937" + ], + "prepared_review_unit_hash": "2d97391a820649c875fb00aca473091a743ede9975242c1b062d7a244ee9af3a", + "component_budget": { + "node_count": 5, + "item_count": 5, + "soft_pair_count": 9, + "serialized_chars": 5785, + "estimated_input_tokens": 1447, + "estimated_eval_tokens": 3150 + }, + "budget": { + "node_count": 5, + "item_count": 5, + "soft_pair_count": 9, + "serialized_chars": 5785, + "estimated_input_tokens": 1447, + "estimated_eval_tokens": 3150 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": false, + "analyst_result": { + "summary": "All five items are low-information closed issues, but their titles point to different things: generic spam placeholders, a documentation suggestion, and a \u201cGoogle LLC\u201d promo-style post. There is no convincing duplicate cluster here.", + "confidence": 0.95, + "canonical_issue_reason": "No strong canonical issue exists because the titles do not describe the same underlying bug or request; if a placeholder must be chosen, issue 44371 is the most generic and earliest spam-like entry.", + "canonical_pr_reason": null, + "best_issue_reason": "None of these issues is a good global representative for a duplicate cluster, since they are not clearly about the same problem and lack meaningful bodies or targets.", + "best_pr_reason": null, + "soft_edge_verdicts": [ + { + "left": "issue:44371", + "right": "issue:44572", + "accept": false, + "reason": "Both are spam-like placeholders, but there is no evidence they refer to the same underlying issue or change; titles alone are identical generic noise." + }, + { + "left": "issue:44371", + "right": "issue:44573", + "accept": false, + "reason": "Same as above: identical spam placeholder titles are not enough to conclude duplicate intent or the same concrete topic." + }, + { + "left": "issue:44371", + "right": "issue:44418", + "accept": false, + "reason": "One is a generic spam placeholder and the other is a documentation enhancement suggestion; these are different issue types and topics." + }, + { + "left": "issue:44371", + "right": "issue:44937", + "accept": false, + "reason": "Generic spam placeholder versus a specific \u201cGoogle LLC\u201d promotional title; no shared underlying bug or request is evident." + }, + { + "left": "issue:44418", + "right": "issue:44572", + "accept": false, + "reason": "Documentation enhancement suggestion does not match a generic spam placeholder; no shared concrete issue is visible." + }, + { + "left": "issue:44418", + "right": "issue:44573", + "accept": false, + "reason": "Documentation enhancement suggestion and spam placeholder are unrelated, with no evidence of the same bug or change." + }, + { + "left": "issue:44572", + "right": "issue:44573", + "accept": false, + "reason": "Both are spam placeholders, but that is too generic to establish they are the same underlying issue; no concrete duplicate evidence." + }, + { + "left": "issue:44572", + "right": "issue:44937", + "accept": false, + "reason": "Spam placeholder and \u201cCheck out \u2018Google LLC\u2019\u201d look like different low-quality posts, not the same issue." + }, + { + "left": "issue:44573", + "right": "issue:44937", + "accept": false, + "reason": "Spam placeholder versus promotional-sounding post; no indication they describe the same underlying problem." + } + ] + }, + "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": 21, + "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": "569aecf18080f78e9e46867fd7d8efa3d744d3e8b253c476e492aed9d3ebb83b", + "component_budget": { + "node_count": 2, + "item_count": 2, + "soft_pair_count": 1, + "serialized_chars": 1645, + "estimated_input_tokens": 412, + "estimated_eval_tokens": 1080 + }, + "budget": { + "node_count": 2, + "item_count": 2, + "soft_pair_count": 1, + "serialized_chars": 1645, + "estimated_input_tokens": 412, + "estimated_eval_tokens": 1080 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": false, + "analyst_result": { + "summary": "These two closed issues appear to be duplicates of the same energy-efficiency proposal, with identical titles and highly similar content. Issue #45145 is the better canonical thread because it was opened first and has an inbound reference.", + "confidence": 0.98, + "canonical_issue_reason": "#45145 is the earliest report and already has an inbound reference, so it is the best canonical issue for this duplicated topic.", + "canonical_pr_reason": null, + "best_issue_reason": "#45145 is the strongest single issue to keep because it predates #45151 and appears to be the original discussion thread for the same request.", + "best_pr_reason": null, + "soft_edge_verdicts": [ + { + "left": "issue:45145", + "right": "issue:45151", + "accept": true, + "reason": "Same title and highly similar issue content; they look like duplicate reports of the same underlying energy-reduction proposal." + } + ] + }, + "evaluator_result": { + "accept": true, + "feedback": "The analyst\u2019s conclusion is grounded in the packet: the issues share the exact same title, have soft similarity evidence, and #45145 is earlier with one inbound reference. The duplicate/canonical-issue reasoning is conservative enough here, and I don\u2019t see an overstatement beyond the available evidence." + }, + "evaluator_used": true, + "retried": false, + "accepted_nontrivial_soft_edge": true, + "error_kind": null, + "error_message": null + }, + { + "label": "issue", + "component_index": 14, + "component_count": 21, + "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": 15, + "component_count": 21, + "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-45672-2", + "nodes": [ + "issue:45672", + "issue:45674" + ], + "soft_pairs": [ + "issue:45672|issue:45674" + ], + "prepared_review_unit_hash": "150f663c93d6b2777d299cd7c30b0b62c0cf4b0096614d2380f1bafe55f49b36", + "component_budget": { + "node_count": 2, + "item_count": 2, + "soft_pair_count": 1, + "serialized_chars": 1672, + "estimated_input_tokens": 418, + "estimated_eval_tokens": 1092 + }, + "budget": { + "node_count": 2, + "item_count": 2, + "soft_pair_count": 1, + "serialized_chars": 1672, + "estimated_input_tokens": 418, + "estimated_eval_tokens": 1092 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": false, + "analyst_result": { + "summary": "The two issues both happen during 4-bit inference, but they describe different failure modes and likely different root causes: one is a torch.finfo() TypeError on uint8 audio-module weights, the other is an AssertionError caused by llm_int8_skip_modules dropping the default lm_head exclusion.", + "confidence": 0.95, + "canonical_issue_reason": "No single canonical duplicate stands out here; the issues target different code paths and fixes despite overlapping in 4-bit inference.", + "canonical_pr_reason": null, + "best_issue_reason": "Issue 45674 is the clearer representative of the cluster because it states a specific configuration regression and an explicit failure condition, making it more actionable as the reference issue.", + "best_pr_reason": null, + "soft_edge_verdicts": [ + { + "left": "issue:45672", + "right": "issue:45674", + "accept": false, + "reason": "Reject: these are different bugs. 45672 is a dtype/torch.finfo failure in Gemma4 audio modules, while 45674 is a BitsAndBytesConfig skip-module regression causing an lm_head assertion." + } + ] + }, + "evaluator_result": null, + "evaluator_used": false, + "retried": false, + "accepted_nontrivial_soft_edge": false, + "error_kind": null, + "error_message": null + }, + { + "label": "issue", + "component_index": 16, + "component_count": 21, + "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": "4385fa75ec3d2d25251c46a042418aea3159bdc1baf016b54ba0d3b36b4e1693", + "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 two closed issues do not look like duplicates of the same bug/change; they appear to be unrelated name-and-date entries with no shared context or problem description.", + "confidence": 0.98, + "canonical_issue_reason": "No clear canonical issue: there is no evidence these reports describe the same underlying problem.", + "canonical_pr_reason": null, + "best_issue_reason": "Neither issue is a good representative for a duplicate cluster because the titles/body provide no shared defect or change context.", + "best_pr_reason": null, + "soft_edge_verdicts": [ + { + "left": "issue:45762", + "right": "issue:45763", + "accept": false, + "reason": "Reject: the only similarity is the format of the titles; there is no description, target, or other evidence that they report the same issue." + } + ] + }, + "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": 21, + "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": "80709f98ec6ca8de142c8db9e251ceebe93e61b487ec287f92dde91304d90913", + "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": "These two issues are unrelated low-information records (different names and dates) and do not appear to describe the same bug or request.", + "confidence": 0.98, + "canonical_issue_reason": null, + "canonical_pr_reason": null, + "best_issue_reason": "Neither issue is a clear canonical choice; both are standalone and unrelated, with no substantive content linking them.", + "best_pr_reason": null, + "soft_edge_verdicts": [ + { + "left": "issue:45764", + "right": "issue:45765", + "accept": false, + "reason": "Only superficial similarity in the title pattern; the names and dates differ and there is no evidence of 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": 18, + "component_count": 21, + "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": "e79bfe723f4cae4c9b1ec2f3c9506e5ef96acaba87423d0e6a4887245307ffab", + "component_budget": { + "node_count": 2, + "item_count": 2, + "soft_pair_count": 1, + "serialized_chars": 1541, + "estimated_input_tokens": 386, + "estimated_eval_tokens": 1028 + }, + "budget": { + "node_count": 2, + "item_count": 2, + "soft_pair_count": 1, + "serialized_chars": 1541, + "estimated_input_tokens": 386, + "estimated_eval_tokens": 1028 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": false, + "analyst_result": { + "summary": "These are two unrelated open feature requests: one asks for a Qwen3_5 token-classification head, the other asks for FP8 training support. The similarity is superficial and they do not appear to describe the same underlying bug or change.", + "confidence": 0.96, + "canonical_issue_reason": "No canonical issue: the two items request different features in different areas of the stack (model head support vs FP8 training).", + "canonical_pr_reason": null, + "best_issue_reason": "issue:45810 is the more concrete, implementation-oriented request; 46295 is a separate training feature and not a duplicate.", + "best_pr_reason": null, + "soft_edge_verdicts": [ + { + "left": "issue:45810", + "right": "issue:46295", + "accept": false, + "reason": "Different feature requests with no shared code path or explicit linkage; one is for Qwen3_5 token classification, the other for FP8 training support." + } + ] + }, + "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": 21, + "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": "b6c98c744ebbe8c4977b34266aaf8ff9d83ef3b121278f7ac193b06943c568b7", + "component_budget": { + "node_count": 5, + "item_count": 5, + "soft_pair_count": 8, + "serialized_chars": 5727, + "estimated_input_tokens": 1432, + "estimated_eval_tokens": 3120 + }, + "budget": { + "node_count": 5, + "item_count": 5, + "soft_pair_count": 8, + "serialized_chars": 5727, + "estimated_input_tokens": 1432, + "estimated_eval_tokens": 3120 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": false, + "analyst_result": { + "summary": "The cluster looks like two likely duplicate issue pairs plus one separate issue: 45870/45871 share an identical title and body length, and 45872/45873 do as well, while 45869 is a different dataset-sample request. The metadata supports high similarity / likely duplication, but not a definitive exact-duplicate claim.", + "confidence": 0.86, + "canonical_issue_reason": "No single canonical issue for the whole cluster; the evidence supports two highly similar pairs (45870/45871 and 45872/45873) and one separate issue (45869), but the packet alone does not prove exact duplication.", + "canonical_pr_reason": null, + "best_issue_reason": "45872 is the strongest representative of the more specific 'Financial time series + scraping service' thread, though its match with 45873 should be treated as highly similar rather than definitively exact from the provided metadata.", + "best_pr_reason": null, + "soft_edge_verdicts": [ + { + "left": "issue:45869", + "right": "issue:45870", + "accept": false, + "reason": "Different requests: a financial dataset bundle vs. a time-series prediction example. Related domain, but not the same underlying issue." + }, + { + "left": "issue:45869", + "right": "issue:45871", + "accept": false, + "reason": "Different requests: a financial dataset bundle vs. a time-series prediction example. Not the same underlying issue." + }, + { + "left": "issue:45870", + "right": "issue:45871", + "accept": true, + "reason": "Highly similar and likely the same request: identical title and identical body length strongly suggest a duplicate, though the metadata does not by itself prove exact duplication." + }, + { + "left": "issue:45870", + "right": "issue:45872", + "accept": false, + "reason": "Different request scopes: a prediction example vs. a scraping-service integration request." + }, + { + "left": "issue:45870", + "right": "issue:45873", + "accept": false, + "reason": "Different request scopes: a prediction example vs. a scraping-service integration request." + }, + { + "left": "issue:45871", + "right": "issue:45872", + "accept": false, + "reason": "Different request scopes: a prediction example vs. a scraping-service integration request." + }, + { + "left": "issue:45871", + "right": "issue:45873", + "accept": false, + "reason": "Different request scopes: a prediction example vs. a scraping-service integration request." + }, + { + "left": "issue:45872", + "right": "issue:45873", + "accept": true, + "reason": "Highly similar and likely the same request: identical title and identical body length strongly suggest a duplicate, though the packet only supports a conservative likely-duplicate reading." + } + ] + }, + "evaluator_result": { + "accept": false, + "feedback": "The cluster summary is a bit too strong for the evidence in the packet. The packet shows identical titles, equal body lengths, and high soft-similarity for 45870/45871 and 45872/45873, but it does not directly establish they are exact duplicates. The conservative read is that those pairs are highly similar / likely duplicate, not definitively exact duplicate. Because the analyst\u2019s canonical and best-issue reasons rely on that stronger claim, I would reject and ask for softer wording grounded only in the provided metadata." + }, + "evaluator_used": true, + "retried": true, + "accepted_nontrivial_soft_edge": true, + "error_kind": null, + "error_message": null + }, + { + "label": "issue", + "component_index": 20, + "component_count": 21, + "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": "132b15eda3023096b614e4f0a9f0766c61064c843b992670137effa48dc23f15", + "component_budget": { + "node_count": 2, + "item_count": 2, + "soft_pair_count": 1, + "serialized_chars": 1636, + "estimated_input_tokens": 409, + "estimated_eval_tokens": 1074 + }, + "budget": { + "node_count": 2, + "item_count": 2, + "soft_pair_count": 1, + "serialized_chars": 1636, + "estimated_input_tokens": 409, + "estimated_eval_tokens": 1074 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": false, + "analyst_result": { + "summary": "These two issues are related to static-cache failures around linear attention, but the packet does not show enough shared evidence to confirm they are duplicates. The safer read is that they may be adjacent symptoms or the same area of breakage, not the same grounded bug report.", + "confidence": 0.58, + "canonical_issue_reason": "Issue #46439 is the better canonical issue because it describes the broader mechanism (`early_initialization` corrupting linear-attention layers) rather than one model-specific crash symptom.", + "canonical_pr_reason": null, + "best_issue_reason": "Issue #46439 is the strongest standalone tracker since it names the suspected root cause and the resulting shape corruption more directly, making it more useful for triage and eventual fixing.", + "best_pr_reason": null, + "soft_edge_verdicts": [ + { + "left": "issue:46439", + "right": "issue:46441", + "accept": false, + "reason": "The evidence is only soft similarity; there are no shared files, explicit targets, or body details here tying both reports to the same failure path. They may be related, but duplication is not confirmed from the provided packet." + } + ] + }, + "evaluator_result": { + "accept": false, + "feedback": "The duplicate verdict is too strong for the evidence in the packet. The pair only has soft similarity and no shared files/targets or body details here, so saying they are the same bug or the same failure path overstates what is grounded. A more conservative read would be that they may be related, but duplication is not confirmed from the provided evidence." + }, + "evaluator_used": true, + "retried": true, + "accepted_nontrivial_soft_edge": false, + "error_kind": null, + "error_message": null + }, + { + "label": "issue", + "component_index": 21, + "component_count": 21, + "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": "06093e7828b568d7eb96bd27305844ee7e0bd6443acb0eb2aae66066cd4cd3d1", + "component_budget": { + "node_count": 2, + "item_count": 2, + "soft_pair_count": 1, + "serialized_chars": 1706, + "estimated_input_tokens": 427, + "estimated_eval_tokens": 1110 + }, + "budget": { + "node_count": 2, + "item_count": 2, + "soft_pair_count": 1, + "serialized_chars": 1706, + "estimated_input_tokens": 427, + "estimated_eval_tokens": 1110 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": false, + "analyst_result": { + "summary": "The two issues appear to describe the same Wav2Vec2CTCTokenizer v5 regression, with identical titles and body lengths and only a soft-similarity signal supporting the match. The open issue is the better tracker, while the closed one looks like the older report in the pair.", + "confidence": 0.92, + "canonical_issue_reason": "issue:46552 is the better canonical issue because it is open and already has inbound references, so it is the best living tracker for the shared bug.", + "canonical_pr_reason": null, + "best_issue_reason": "issue:46552 is the strongest issue to keep for triage because it is open, has inbound references, and matches the same reported tokenizer regression.", + "best_pr_reason": null, + "soft_edge_verdicts": [ + { + "left": "issue:46551", + "right": "issue:46552", + "accept": true, + "reason": "Identical title and body length, plus the soft-similarity signal, strongly suggest the same underlying tokenizer bug; however, this is a conservative same-bug match rather than evidence of an explicit duplicate relation." + } + ] + }, + "evaluator_result": { + "accept": false, + "feedback": "The pair looks highly similar and likely refers to the same bug, but the verdict overstates the evidence by calling one a closed duplicate of the other. The packet only supports identical titles/body length, a soft-similarity signal, and that one issue is closed while the other is open with inbound references; it does not explicitly establish a duplicate relation. The canonical/open-tracker choice is reasonable, but the soft-edge reason should stay 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": 35, + "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-39895-15", + "nodes": [ + "pull_request:39895", + "pull_request:41855", + "pull_request:42668", + "pull_request:43251", + "pull_request:43254", + "pull_request:43424", + "pull_request:43823", + "pull_request:43995", + "pull_request:44827", + "pull_request:45170", + "pull_request:45213", + "pull_request:45655", + "pull_request:46048", + "pull_request:46267", + "pull_request:46662" + ], + "soft_pairs": [ + "pull_request:39895|pull_request:43424", + "pull_request:41855|pull_request:43424", + "pull_request:41855|pull_request:43823", + "pull_request:42668|pull_request:43823", + "pull_request:43251|pull_request:43823", + "pull_request:43251|pull_request:44827", + "pull_request:43424|pull_request:44827", + "pull_request:43424|pull_request:43823", + "pull_request:43823|pull_request:46662", + "pull_request:43823|pull_request:45655", + "pull_request:43823|pull_request:43995", + "pull_request:43995|pull_request:46662", + "pull_request:43995|pull_request:45213", + "pull_request:43995|pull_request:44827", + "pull_request:44827|pull_request:45170", + "pull_request:44827|pull_request:46267", + "pull_request:44827|pull_request:46048", + "pull_request:44827|pull_request:45655", + "pull_request:44827|pull_request:45213", + "pull_request:43823|pull_request:45213", + "pull_request:45213|pull_request:46662", + "pull_request:45655|pull_request:46662", + "pull_request:44827|pull_request:46662", + "pull_request:43251|pull_request:43254" + ], + "prepared_review_unit_hash": "e844037dac2181e9b2a01a8475522e5f330db793acbd1656c106e56918e52f6c", + "component_budget": { + "node_count": 15, + "item_count": 15, + "soft_pair_count": 24, + "serialized_chars": 41942, + "estimated_input_tokens": 10486, + "estimated_eval_tokens": 21228 + }, + "budget": { + "node_count": 15, + "item_count": 15, + "soft_pair_count": 24, + "serialized_chars": 41942, + "estimated_input_tokens": 10486, + "estimated_eval_tokens": 21228 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": false, + "analyst_result": { + "summary": "This cluster is mostly noisy: most PRs are unrelated model additions, docs, tests, or small subsystem fixes. The only true duplicate-like overlap is the pair of loss_utils PRs for issue 43240; everything else should stay separate.", + "confidence": 0.93, + "canonical_issue_reason": null, + "canonical_pr_reason": "PR 43251 is the best canonical representative because it is the clearer issue-linked fix for 43240 and matches the same `loss_utils.py` code path as 43254, but with a more explicit fix-oriented scope and fuller review trail.", + "best_issue_reason": null, + "best_pr_reason": "PR 43251 is the best overall PR candidate in this cluster because it is the strongest genuine overlap and most complete version of the shared `fixed_cross_entropy` change; the rest of the items are not duplicate-worthy.", + "soft_edge_verdicts": [ + { + "left": "pull_request:39895", + "right": "pull_request:43424", + "accept": false, + "reason": "Unrelated PRs: one adds VideoPrism model support, the other adds Executorch dynamic-shape tests for decoder-only export." + }, + { + "left": "pull_request:41855", + "right": "pull_request:43424", + "accept": false, + "reason": "Different changes: Mistral tokenizer missing methods versus Executorch export testing." + }, + { + "left": "pull_request:41855", + "right": "pull_request:43823", + "accept": false, + "reason": "Tokenizer helper additions are unrelated to a new MobileLLM model implementation." + }, + { + "left": "pull_request:42668", + "right": "pull_request:43823", + "accept": false, + "reason": "Both touch processor/model auto code, but they are for different models and different behaviors." + }, + { + "left": "pull_request:43251", + "right": "pull_request:43823", + "accept": false, + "reason": "Loss function keyword support is unrelated to adding a new MobileLLM model." + }, + { + "left": "pull_request:43251", + "right": "pull_request:44827", + "accept": false, + "reason": "Cross-entropy kwargs support is unrelated to Mistral4 test/model adjustments." + }, + { + "left": "pull_request:43424", + "right": "pull_request:44827", + "accept": false, + "reason": "Executorch export tests and Mistral4 test fixes are different code paths and changes." + }, + { + "left": "pull_request:43424", + "right": "pull_request:43823", + "accept": false, + "reason": "Executorch dynamic-shape export tests do not overlap with MobileLLM model addition." + }, + { + "left": "pull_request:43823", + "right": "pull_request:46662", + "accept": false, + "reason": "Different model families and different bugs: MobileLLM addition versus SmolVLM video padding validation." + }, + { + "left": "pull_request:43823", + "right": "pull_request:45655", + "accept": false, + "reason": "Model addition and config-resolution fix are unrelated changes." + }, + { + "left": "pull_request:43823", + "right": "pull_request:43995", + "accept": false, + "reason": "New MobileLLM model code does not match Falcon output-collection refactoring." + }, + { + "left": "pull_request:43995", + "right": "pull_request:46662", + "accept": false, + "reason": "Falcon forward refactor and SmolVLM video padding are unrelated subsystems." + }, + { + "left": "pull_request:43995", + "right": "pull_request:45213", + "accept": false, + "reason": "Falcon model refactor is unrelated to the model-creation skill PR." + }, + { + "left": "pull_request:43995", + "right": "pull_request:44827", + "accept": false, + "reason": "Falcon output-interface refactor and Mistral4 test fixes are not the same change." + }, + { + "left": "pull_request:44827", + "right": "pull_request:45170", + "accept": false, + "reason": "Mistral4 test fixes are unrelated to the layrnorm-to-layernorm rename conversion fix." + }, + { + "left": "pull_request:44827", + "right": "pull_request:46267", + "accept": false, + "reason": "Mistral4 test/model work is unrelated to serve chat parsing updates." + }, + { + "left": "pull_request:44827", + "right": "pull_request:46048", + "accept": false, + "reason": "Mistral4 fixes and TDT loss kernel integration are different features." + }, + { + "left": "pull_request:44827", + "right": "pull_request:45655", + "accept": false, + "reason": "Mistral4 work is unrelated to config resolution ordering in `modeling_utils.py`." + }, + { + "left": "pull_request:44827", + "right": "pull_request:45213", + "accept": false, + "reason": "Mistral4 model/test changes do not duplicate the model-creation skill PR." + }, + { + "left": "pull_request:43823", + "right": "pull_request:45213", + "accept": false, + "reason": "MobileLLM model addition and the model-creation skill are not the same code change." + }, + { + "left": "pull_request:45213", + "right": "pull_request:46662", + "accept": false, + "reason": "A model-creation skill draft is unrelated to SmolVLM video padding behavior." + }, + { + "left": "pull_request:45655", + "right": "pull_request:46662", + "accept": false, + "reason": "Config resolution ordering fix is unrelated to SmolVLM video padding." + }, + { + "left": "pull_request:44827", + "right": "pull_request:46662", + "accept": false, + "reason": "Mistral4 tests/model changes are unrelated to SmolVLM video padding." + }, + { + "left": "pull_request:43251", + "right": "pull_request:43254", + "accept": true, + "reason": "Same underlying loss-function change: both modify `fixed_cross_entropy` in `loss_utils.py` to pass through new `cross_entropy` kwargs for issue 43240, and they could plausibly be merged into one PR." + } + ] + }, + "evaluator_result": { + "accept": true, + "feedback": "The cluster summary is mostly grounded in the packet, and the only clearly mergeable same-bugfix overlap is the `loss_utils.py` pair for issue 43240 (`43251`/`43254`). The soft-edge false verdicts are conservative overall, and the accepted pair is well supported by the shared file and issue target. I did not see a clear overstatement that would require rejection." + }, + "evaluator_used": true, + "retried": false, + "accepted_nontrivial_soft_edge": true, + "error_kind": null, + "error_message": null + }, + { + "label": "pull_request", + "component_index": 2, + "component_count": 35, + "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-40546-2", + "nodes": [ + "pull_request:40546", + "pull_request:42781" + ], + "soft_pairs": [ + "pull_request:40546|pull_request:42781" + ], + "prepared_review_unit_hash": "b47d7109cd8571541e7cfa247865248a934fb17946159b699c57258e226a7083", + "component_budget": { + "node_count": 2, + "item_count": 2, + "soft_pair_count": 1, + "serialized_chars": 7226, + "estimated_input_tokens": 1807, + "estimated_eval_tokens": 3870 + }, + "budget": { + "node_count": 2, + "item_count": 2, + "soft_pair_count": 1, + "serialized_chars": 7226, + "estimated_input_tokens": 1807, + "estimated_eval_tokens": 3870 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": false, + "analyst_result": { + "summary": "Two open PRs both center on VibeVoice, but they are not the same change set: one implements the base model, the other adds a realtime variant and related tokenizer/generation work.", + "confidence": 0.94, + "canonical_issue_reason": null, + "canonical_pr_reason": "PR #40546 is the best canonical PR because it is the broader, earlier non-draft implementation of VibeVoice and appears to establish the core model support that later work can build on.", + "best_issue_reason": null, + "best_pr_reason": "PR #40546 is the strongest representative of the cluster since it is the main base VibeVoice implementation; #42781 looks like a follow-up feature branch rather than the same underlying PR.", + "soft_edge_verdicts": [ + { + "left": "pull_request:40546", + "right": "pull_request:42781", + "accept": false, + "reason": "Reject: both touch shared VibeVoice files, but #40546 implements the base model while #42781 adds a realtime variant plus acoustic tokenizer and generation changes. They are related, but not the same concrete bug/change 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": 3, + "component_count": 35, + "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": "a06a7073da7c0fa39246e4d6bf706427980b0651b3eb65a0084b7c3f7a6469ac", + "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 cluster contains two unrelated PRs: one adds TimesFM covariate forecasting, the other adds DINOv3 image classification support. They do not describe the same change or code path, so they should not be merged as duplicates.", + "confidence": 0.99, + "canonical_issue_reason": null, + "canonical_pr_reason": "Use PR #41224 as the cluster representative because it is the more concrete, self-contained feature addition and has substantial review activity; however, it is not a duplicate of #40755.", + "best_issue_reason": null, + "best_pr_reason": "PR #41224 is the better standalone representative for this cluster because it is a clearly scoped model capability addition with focused files and discussion, while #40755 is a separate TimesFM feature.", + "soft_edge_verdicts": [ + { + "left": "pull_request:40755", + "right": "pull_request:41224", + "accept": false, + "reason": "Different models and different functionality: TimesFM covariate forecasting versus DINOv3 image classification. They do not share the same underlying bug or 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": 4, + "component_count": 35, + "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": 5, + "component_count": 35, + "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": "afeac0341974deb4ad3902fe12381e1342022591444f11295f24d702bdff8e4c", + "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": "Two open PRs both bulk-upgrade GitHub Actions workflows, but they target different action families: one is for Node 24 compatibility across general workflow actions, the other updates Docker build actions and a few related workflow steps. They are related maintenance, not the same underlying change.", + "confidence": 0.96, + "canonical_issue_reason": null, + "canonical_pr_reason": "PR 42975 is the better canonical representative because its title and diff clearly describe the broader workflow-action runtime compatibility upgrade, and its changes span several core GitHub Actions used across workflows.", + "best_issue_reason": null, + "best_pr_reason": "PR 42975 is the best PR in the cluster because it most directly captures the shared theme of workflow action version bumps and has the clearest, most general maintenance scope.", + "soft_edge_verdicts": [ + { + "left": "pull_request:42975", + "right": "pull_request:42976", + "accept": false, + "reason": "Do not merge: despite overlapping workflow files and similar upgrade wording, the PRs update different concrete actions (checkout/cache/upload-artifact vs docker/build-push-action and a few others). They do not fix the same code-path problem and would not plausibly be merged into one PR without conflating separate dependency bumps." + } + ] + }, + "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": 35, + "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-42978-6", + "nodes": [ + "pull_request:42978", + "pull_request:43448", + "pull_request:43451", + "pull_request:45490", + "pull_request:46562", + "pull_request:46658" + ], + "soft_pairs": [ + "pull_request:42978|pull_request:43448", + "pull_request:42978|pull_request:43451", + "pull_request:42978|pull_request:45490", + "pull_request:43448|pull_request:43451", + "pull_request:43448|pull_request:45490", + "pull_request:43451|pull_request:45490", + "pull_request:45490|pull_request:46562", + "pull_request:46562|pull_request:46658" + ], + "prepared_review_unit_hash": "5de7660b6a34267a6c684aa4a1fa8871a6e8d624d3b49592a8b4a68565cf3a90", + "component_budget": { + "node_count": 6, + "item_count": 6, + "soft_pair_count": 8, + "serialized_chars": 20309, + "estimated_input_tokens": 5078, + "estimated_eval_tokens": 10412 + }, + "budget": { + "node_count": 6, + "item_count": 6, + "soft_pair_count": 8, + "serialized_chars": 20309, + "estimated_input_tokens": 5078, + "estimated_eval_tokens": 10412 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": false, + "analyst_result": { + "summary": "This cluster is mostly a set of unrelated model-addition PRs (ViT NEPA, Molmo, Molmo2, CTSM, Step3p7 Flash). The only closely related pair is the Step3p7 Flash PR and its WIP mini-model/support PR, but they still look like companion work rather than the same duplicate change.", + "confidence": 0.96, + "canonical_issue_reason": null, + "canonical_pr_reason": "PR 46562 is the best canonical representative because it is the main Step3p7 Flash implementation PR and sits closest to the only near-related companion PR in the cluster.", + "best_issue_reason": null, + "best_pr_reason": "PR 43451 is the strongest standalone PR in the set: it has the broadest model-integration scope and the highest review activity, making it the most mature representative of the overall cluster theme.", + "soft_edge_verdicts": [ + { + "left": "pull_request:42978", + "right": "pull_request:43448", + "accept": false, + "reason": "Different models (ViT NEPA vs Molmo); similar model-addition structure, but not the same change." + }, + { + "left": "pull_request:42978", + "right": "pull_request:43451", + "accept": false, + "reason": "Different models (ViT NEPA vs Molmo2); shared framework files do not make this a duplicate." + }, + { + "left": "pull_request:42978", + "right": "pull_request:45490", + "accept": false, + "reason": "Different model additions (ViT NEPA vs CTSM), not the same underlying code change." + }, + { + "left": "pull_request:43448", + "right": "pull_request:43451", + "accept": false, + "reason": "Molmo and Molmo2 are related by name but are distinct model integrations with separate code and tests." + }, + { + "left": "pull_request:43448", + "right": "pull_request:45490", + "accept": false, + "reason": "Molmo and CTSM are unrelated model additions." + }, + { + "left": "pull_request:43451", + "right": "pull_request:45490", + "accept": false, + "reason": "Molmo2 and CTSM are different model additions; same subsystem, but not duplicate work." + }, + { + "left": "pull_request:45490", + "right": "pull_request:46562", + "accept": false, + "reason": "Both are model additions, but CTSM and Step3p7 Flash are distinct models and code paths." + }, + { + "left": "pull_request:46562", + "right": "pull_request:46658", + "accept": false, + "reason": "46658 looks like WIP mini-model/assets and scripts for Step3p7 Flash, but not the same concrete implementation change as 46562." + } + ] + }, + "evaluator_result": null, + "evaluator_used": false, + "retried": false, + "accepted_nontrivial_soft_edge": false, + "error_kind": null, + "error_message": null + }, + { + "label": "pull_request", + "component_index": 7, + "component_count": 35, + "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:43498", + "pull_request:43665", + "pull_request:44815", + "pull_request:45114", + "pull_request:46380", + "pull_request:46539", + "pull_request:46579" + ], + "soft_pairs": [ + "pull_request:43291|pull_request:43665", + "pull_request:43291|pull_request:44815", + "pull_request:43291|pull_request:46380", + "pull_request:43498|pull_request:46579", + "pull_request:43665|pull_request:46539", + "pull_request:43665|pull_request:44815", + "pull_request:43665|pull_request:46380", + "pull_request:43665|pull_request:46579", + "pull_request:44815|pull_request:46539", + "pull_request:44815|pull_request:46380", + "pull_request:44815|pull_request:46579", + "pull_request:45114|pull_request:46579", + "pull_request:46380|pull_request:46539", + "pull_request:46380|pull_request:46579" + ], + "prepared_review_unit_hash": "cbb600115a83942b08b9d1cc7c72709384bba80a2939d7a5c055652234c5f473", + "component_budget": { + "node_count": 8, + "item_count": 8, + "soft_pair_count": 14, + "serialized_chars": 23306, + "estimated_input_tokens": 5827, + "estimated_eval_tokens": 11910 + }, + "budget": { + "node_count": 8, + "item_count": 8, + "soft_pair_count": 14, + "serialized_chars": 23306, + "estimated_input_tokens": 5827, + "estimated_eval_tokens": 11910 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": false, + "analyst_result": { + "summary": "This cluster looks like a noisy similarity bucket, not a duplicate set: the PRs cover unrelated fixes in Whisper tests, tie-weights backward compatibility, model test placeholders, FP8 dequant logic, doctest cleanup, generation watermarking validation, CI workflow tweaks, and typo-only edits.", + "confidence": 0.95, + "canonical_issue_reason": null, + "canonical_pr_reason": "No true canonical duplicate stands out; PR 46579 is the broadest and most central maintenance-style change in the cluster, but it does not share the same concrete bug/change with the others.", + "best_issue_reason": null, + "best_pr_reason": "PR 46579 is the best anchor for triage because it is the most broadly connected artifact in this otherwise heterogeneous cluster, though it is not a real duplicate of the others.", + "soft_edge_verdicts": [ + { + "left": "pull_request:43291", + "right": "pull_request:43665", + "accept": false, + "reason": "Different areas and intents: Whisper tokenizer/test cleanup vs placeholder asserts in CLIP/ViT tests." + }, + { + "left": "pull_request:43291", + "right": "pull_request:44815", + "accept": false, + "reason": "Whisper decoding logic/tests are unrelated to FP8 dequant/model-loading fixes." + }, + { + "left": "pull_request:43291", + "right": "pull_request:46380", + "accept": false, + "reason": "Whisper test/tokenization change and generation watermarking validation are different code paths and bugs." + }, + { + "left": "pull_request:43498", + "right": "pull_request:46579", + "accept": false, + "reason": "Backward-compatibility shim for tie_weights is unrelated to a broad typo-only cleanup." + }, + { + "left": "pull_request:43665", + "right": "pull_request:46539", + "accept": false, + "reason": "Dummy/failing model-test edits are unrelated to Phimoe CI/workflow fixes." + }, + { + "left": "pull_request:43665", + "right": "pull_request:44815", + "accept": false, + "reason": "Model-test placeholder assertions do not match the FP8 dequantization/runtime fix in 44815." + }, + { + "left": "pull_request:43665", + "right": "pull_request:46380", + "accept": false, + "reason": "These touch different problems: test-file edits vs generation config/Whisper validation changes." + }, + { + "left": "pull_request:43665", + "right": "pull_request:46579", + "accept": false, + "reason": "Test placeholder changes are not the same underlying change as code typo cleanup across many files." + }, + { + "left": "pull_request:44815", + "right": "pull_request:46539", + "accept": false, + "reason": "FP8 dequant/model-loading logic is unrelated to Phimoe integration test CI adjustments." + }, + { + "left": "pull_request:44815", + "right": "pull_request:46380", + "accept": false, + "reason": "Both are code changes, but they fix different concrete bugs in different subsystems." + }, + { + "left": "pull_request:44815", + "right": "pull_request:46579", + "accept": false, + "reason": "FP8 kernel/loading fixes are not the same as general typo corrections." + }, + { + "left": "pull_request:45114", + "right": "pull_request:46579", + "accept": false, + "reason": "Doctest/documentation edits do not share the same underlying change as code typo cleanup." + }, + { + "left": "pull_request:46380", + "right": "pull_request:46539", + "accept": false, + "reason": "Generation watermarking validation and Phimoe CI workflow changes are unrelated." + }, + { + "left": "pull_request:46380", + "right": "pull_request:46579", + "accept": false, + "reason": "A generation/config runtime fix is not the same as broad typo-only edits across many 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": 8, + "component_count": 35, + "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-43488-2", + "nodes": [ + "pull_request:43488", + "pull_request:43532" + ], + "soft_pairs": [ + "pull_request:43488|pull_request:43532" + ], + "prepared_review_unit_hash": "7416ef1d8bb197e54bafede1da61fada4615f5a143cec955c85edab5949297c7", + "component_budget": { + "node_count": 2, + "item_count": 2, + "soft_pair_count": 1, + "serialized_chars": 4563, + "estimated_input_tokens": 1141, + "estimated_eval_tokens": 2538 + }, + "budget": { + "node_count": 2, + "item_count": 2, + "soft_pair_count": 1, + "serialized_chars": 4563, + "estimated_input_tokens": 1141, + "estimated_eval_tokens": 2538 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": false, + "analyst_result": { + "summary": "Two unrelated \"don't merge\" demo PRs with overlapping test-file touch but different code paths and scopes; they should not be deduplicated.", + "confidence": 0.63, + "canonical_issue_reason": null, + "canonical_pr_reason": "PR #43488 is the cleaner representative: its title and diff are narrowly centered on a repo-bot/setup.py format check, making it more self-contained than the broader workflow-heavy PR #43532.", + "best_issue_reason": null, + "best_pr_reason": "PR #43488 is the best cluster representative because it is more focused and interpretable as a single change, whereas #43532 spans several workflows and utilities and is less canonical as a duplicate target.", + "soft_edge_verdicts": [ + { + "left": "pull_request:43488", + "right": "pull_request:43532", + "accept": false, + "reason": "They only share a test filename; #43488 tweaks packaging/versioning and a vit test, while #43532 changes CI workflows, testing utilities, and notifications. They are not the same underlying fix/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": 35, + "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": "a63d71ae599348543106c7dff3fff3c3b73b0317823ded946d7839848b1e947e", + "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": "Two open PRs add similarly named but distinct pipeline/features: one for promptable concept segmentation and one for promptable visual segmentation. They share boilerplate documentation/registry changes, but they do not appear to fix the same concrete code-path problem.", + "confidence": 0.91, + "canonical_issue_reason": null, + "canonical_pr_reason": null, + "best_issue_reason": null, + "best_pr_reason": "PR #43613 is the stronger standalone change: it covers a broader, more integrated visual-segmentation stack (multiple model tests and a dedicated pipeline implementation), making it the more complete representative of this area.", + "soft_edge_verdicts": [ + { + "left": "pull_request:43612", + "right": "pull_request:43613", + "accept": false, + "reason": "Different feature targets and underlying code paths: concept segmentation vs visual segmentation. The overlap is mostly shared pipeline/docs scaffolding, not the same concrete bug or change, so they should not be merged 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": 10, + "component_count": 35, + "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": "24adb5457d3e2df840c11edf572384006b5ad05262920e424efe11c078c0db6a", + "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 pull requests touch `trainer.py` and improve trainer extensibility, but they fix different code paths: one changes gradient-accumulation loss scaling, the other extracts signature-column filtering into an overrideable helper. They are related in theme, not duplicates.", + "confidence": 0.96, + "canonical_issue_reason": null, + "canonical_pr_reason": "No canonical PR: 43651 and 43663 address separate trainer behaviors and would not be merged into one change.", + "best_issue_reason": null, + "best_pr_reason": "PR 43663 is a clean, self-contained trainer refactor, but it is only a representative of the cluster theme\u2014not a duplicate of 43651.", + "soft_edge_verdicts": [ + { + "left": "pull_request:43651", + "right": "pull_request:43663", + "accept": false, + "reason": "Different underlying changes in the same file: 43651 alters loss scaling during training, while 43663 adds an overrideable signature-column helper for dataset filtering. Shared subsystem alone is not enough to treat them as the same fix." + } + ] + }, + "evaluator_result": null, + "evaluator_used": false, + "retried": false, + "accepted_nontrivial_soft_edge": false, + "error_kind": null, + "error_message": null + }, + { + "label": "pull_request", + "component_index": 11, + "component_count": 35, + "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": "40e16068567901536c4c15a1c5ab165309060f08f2de6fb3c0ddde6e79fd555d", + "component_budget": { + "node_count": 2, + "item_count": 2, + "soft_pair_count": 1, + "serialized_chars": 3664, + "estimated_input_tokens": 916, + "estimated_eval_tokens": 2088 + }, + "budget": { + "node_count": 2, + "item_count": 2, + "soft_pair_count": 1, + "serialized_chars": 3664, + "estimated_input_tokens": 916, + "estimated_eval_tokens": 2088 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": false, + "analyst_result": { + "summary": "The two PRs both touch `src/transformers/cli/serve.py` and both involve typing changes, but the packet does not show they fix the same concrete bug. PR 43842 is a targeted `TypeAdapter`/missing-pydantic guard, while 43656 is a broader `from __future__ import annotations` and annotation cleanup.", + "confidence": 0.81, + "canonical_issue_reason": null, + "canonical_pr_reason": "PR 43842 is the better canonical PR because it addresses a specific runtime problem: guarding `TypeAdapter` when pydantic is absent and fixing the failing annotation site directly.", + "best_issue_reason": null, + "best_pr_reason": "PR 43842 is the best representative for the cluster because it is the narrower, more concrete fix and the explicit issue target is specific; 43656 is broader typing churn and not clearly the same bug.", + "soft_edge_verdicts": [ + { + "left": "pull_request:43656", + "right": "pull_request:43842", + "accept": false, + "reason": "Reject: the overlap is only in the same file and general typing/annotation changes. The packet does not establish the same underlying bug or a mergeable single fix, and the explicit issue targets differ." + } + ] + }, + "evaluator_result": { + "accept": false, + "feedback": "The cluster summary overstates the evidence. The packet only supports that both PRs make typing/annotation-related changes in `src/transformers/cli/serve.py`; it does not clearly show they fix the same bug, and the explicit issue targets differ (`43576` vs `43824`). PR 43842 is specifically about guarding `TypeAdapter` for missing pydantic, while 43656 appears to be a broader `from __future__ import annotations`/typing cleanup. The soft-pair acceptance is too strong without clearer same-bug linkage or mergeability evidence." + }, + "evaluator_used": true, + "retried": true, + "accepted_nontrivial_soft_edge": false, + "error_kind": null, + "error_message": null + }, + { + "label": "pull_request", + "component_index": 12, + "component_count": 35, + "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": 13, + "component_count": 35, + "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: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:44019|pull_request:44085", + "pull_request:44086|pull_request:44722", + "pull_request:44018|pull_request:44019", + "pull_request:44019|pull_request:44071", + "pull_request:44068|pull_request:44116", + "pull_request:44018|pull_request:44116", + "pull_request:44068|pull_request:44722", + "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:44059|pull_request:44086", + "pull_request:44085|pull_request:44086", + "pull_request:44098|pull_request:44101", + "pull_request:44018|pull_request:44071", + "pull_request:44059|pull_request:44722", + "pull_request:44017|pull_request:44161", + "pull_request:44017|pull_request:44722", + "pull_request:44017|pull_request:44073", + "pull_request:44068|pull_request:44071", + "pull_request:43997|pull_request:44073", + "pull_request:44071|pull_request:44086", + "pull_request:44017|pull_request:44071", + "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:44068", + "pull_request:44101|pull_request:44161", + "pull_request:43997|pull_request:44017", + "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:43997|pull_request:44071", + "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:43997|pull_request:44129", + "pull_request:44010|pull_request:44018", + "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" + ], + "prepared_review_unit_hash": "198d4dd5455eae6da978e1a512f28eac1adce076e9549999baf653435bb1d08e", + "component_budget": { + "node_count": 28, + "item_count": 28, + "soft_pair_count": 78, + "serialized_chars": 84902, + "estimated_input_tokens": 21226, + "estimated_eval_tokens": 42708 + }, + "budget": { + "node_count": 27, + "item_count": 27, + "soft_pair_count": 72, + "serialized_chars": 75096, + "estimated_input_tokens": 18774, + "estimated_eval_tokens": 37804 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": true, + "analyst_result": { + "summary": "This cluster is mostly a set of model-specific refactors around standardized output tracing. Most PRs share the same high-level theme but touch different model code paths, so they are not duplicates. The only strong duplicate-like pairs are the GPT-J iterations and the GPT-Neo iterations.", + "confidence": 0.78, + "canonical_issue_reason": null, + "canonical_pr_reason": "PR #44722 is the best representative because it is the most complete GPT-J/CodeGen output-tracing refactor in the set, targets the shared issue, and appears to subsume earlier GPT-J-specific iterations.", + "best_issue_reason": null, + "best_pr_reason": "PR #44722 is the strongest single PR candidate for the cluster: it covers the same concrete refactor as the earlier GPT-J/CodeGen work and is the broadest, latest iteration among the close matches.", + "soft_edge_verdicts": [ + { + "left": "pull_request:43996", + "right": "pull_request:44085", + "accept": false, + "reason": "Different model families and code paths (CVT/FNet vs GPT-J); same theme, but not the same underlying change." + }, + { + "left": "pull_request:43996", + "right": "pull_request:44044", + "accept": false, + "reason": "CVT/FNet output-tracing refactor vs DeBERTa output-tracing refactor; separate model implementations." + }, + { + "left": "pull_request:44066", + "right": "pull_request:44085", + "accept": true, + "reason": "Both are GPT-J output-tracing refactors in the same file/function; one is a narrower version of the same underlying change." + }, + { + "left": "pull_request:44007", + "right": "pull_request:44072", + "accept": false, + "reason": "RegNet and EfficientNet are different models with separate encoder implementations; same pattern, not the same bug/change." + }, + { + "left": "pull_request:44072", + "right": "pull_request:44722", + "accept": false, + "reason": "EfficientNet vs GPT-J/CodeGen are unrelated code paths, so they should not be merged as one duplicate PR." + }, + { + "left": "pull_request:44066", + "right": "pull_request:44072", + "accept": false, + "reason": "GPT-J/CodeGen vs EfficientNet are different model refactors, not one concrete change." + }, + { + "left": "pull_request:44013", + "right": "pull_request:44044", + "accept": false, + "reason": "MobileNetV2 and DeBERTa output tracing are separate model-specific refactors." + }, + { + "left": "pull_request:44066", + "right": "pull_request:44086", + "accept": false, + "reason": "GPT-J/CodeGen output tracing and MGP-STR output tracing are different model code paths." + }, + { + "left": "pull_request:44066", + "right": "pull_request:44071", + "accept": false, + "reason": "GPT-J/CodeGen and MPT touch different architectures and different forward paths." + }, + { + "left": "pull_request:44018", + "right": "pull_request:44068", + "accept": true, + "reason": "Both are GPT-Neo output-tracing refactors in the same model file, with overlapping test/code changes; they look like iterations of the same PR." + }, + { + "left": "pull_request:44066", + "right": "pull_request:44068", + "accept": false, + "reason": "GPT-J/CodeGen and GPT-Neo are different model implementations, so these are not duplicates." + }, + { + "left": "pull_request:44018", + "right": "pull_request:44066", + "accept": false, + "reason": "GPT-Neo vs GPT-J/CodeGen are separate model refactors even though they use the same standardized pattern." + }, + { + "left": "pull_request:44019", + "right": "pull_request:44722", + "accept": false, + "reason": "ResNet output tracing and GPT-J/CodeGen output tracing are unrelated model changes." + }, + { + "left": "pull_request:44019", + "right": "pull_request:44085", + "accept": false, + "reason": "ResNet and GPT-J are different model families; same broad refactor theme is not enough to merge them." + }, + { + "left": "pull_request:44086", + "right": "pull_request:44722", + "accept": false, + "reason": "MGP-STR vs GPT-J/CodeGen are different code paths and should not be treated as the same change." + }, + { + "left": "pull_request:44018", + "right": "pull_request:44019", + "accept": false, + "reason": "GPT-Neo and ResNet are unrelated model-specific output-tracing refactors." + }, + { + "left": "pull_request:44019", + "right": "pull_request:44071", + "accept": false, + "reason": "ResNet and MPT are different architectures with distinct implementations." + }, + { + "left": "pull_request:44068", + "right": "pull_request:44116", + "accept": false, + "reason": "GPT-Neo and Flaubert/XLM are different model code paths." + }, + { + "left": "pull_request:44018", + "right": "pull_request:44116", + "accept": false, + "reason": "GPT-Neo and Flaubert are separate refactors in unrelated models." + }, + { + "left": "pull_request:44068", + "right": "pull_request:44722", + "accept": false, + "reason": "GPT-Neo and GPT-J/CodeGen are different model implementations, so this is not one duplicate PR." + }, + { + "left": "pull_request:44071", + "right": "pull_request:44722", + "accept": false, + "reason": "MPT and GPT-J/CodeGen are different code paths despite similar output-tracing goals." + }, + { + "left": "pull_request:44018", + "right": "pull_request:44085", + "accept": false, + "reason": "GPT-Neo and GPT-J are different models; not the same underlying bug/change." + }, + { + "left": "pull_request:44018", + "right": "pull_request:44086", + "accept": false, + "reason": "GPT-Neo vs MGP-STR are unrelated model-specific refactors." + }, + { + "left": "pull_request:44068", + "right": "pull_request:44086", + "accept": false, + "reason": "GPT-Neo and MGP-STR do not share the same concrete code path." + }, + { + "left": "pull_request:44071", + "right": "pull_request:44085", + "accept": false, + "reason": "MPT and GPT-J are different model families." + }, + { + "left": "pull_request:44059", + "right": "pull_request:44086", + "accept": false, + "reason": "GPT-2 and MGP-STR are distinct model implementations; same refactor pattern only." + }, + { + "left": "pull_request:44085", + "right": "pull_request:44086", + "accept": false, + "reason": "GPT-J and MGP-STR are different model code paths." + }, + { + "left": "pull_request:44098", + "right": "pull_request:44101", + "accept": false, + "reason": "ViLT and Flaubert/XLM are separate model refactors." + }, + { + "left": "pull_request:44073", + "right": "pull_request:44101", + "accept": false, + "reason": "VisualBert and Flaubert are different models, so these are not duplicates." + }, + { + "left": "pull_request:44056", + "right": "pull_request:44101", + "accept": false, + "reason": "MPNet and Flaubert/XLM touch unrelated model implementations." + }, + { + "left": "pull_request:44059", + "right": "pull_request:44071", + "accept": false, + "reason": "GPT-2 and MPT are different architectures and different forward paths." + }, + { + "left": "pull_request:43997", + "right": "pull_request:44017", + "accept": false, + "reason": "RegNet and Segformer are different model code paths." + }, + { + "left": "pull_request:43997", + "right": "pull_request:44056", + "accept": false, + "reason": "RegNet and MPNet are unrelated model-specific changes." + }, + { + "left": "pull_request:44071", + "right": "pull_request:44076", + "accept": false, + "reason": "MPT and ImageGPT are different models with separate implementations." + }, + { + "left": "pull_request:44074", + "right": "pull_request:44101", + "accept": false, + "reason": "TextNet and Flaubert/XLM are different model code paths." + }, + { + "left": "pull_request:44056", + "right": "pull_request:44161", + "accept": false, + "reason": "MPNet and LongT5 are unrelated refactors across different models." + }, + { + "left": "pull_request:43997", + "right": "pull_request:44074", + "accept": false, + "reason": "RegNet and TextNet are different architectures, so this is not one duplicate change." + }, + { + "left": "pull_request:44010", + "right": "pull_request:44071", + "accept": false, + "reason": "SqueezeBert and MPT are separate model implementations." + }, + { + "left": "pull_request:44010", + "right": "pull_request:44073", + "accept": false, + "reason": "SqueezeBert and VisualBert are different code paths." + }, + { + "left": "pull_request:44074", + "right": "pull_request:44161", + "accept": false, + "reason": "TextNet and LongT5 are unrelated model-specific refactors." + }, + { + "left": "pull_request:43996", + "right": "pull_request:44007", + "accept": false, + "reason": "CVT/FNet and RegNet are different models; same output-tracing theme is too broad to merge." + }, + { + "left": "pull_request:43996", + "right": "pull_request:44072", + "accept": false, + "reason": "CVT/FNet and EfficientNet are separate model refactors." + }, + { + "left": "pull_request:43996", + "right": "pull_request:44017", + "accept": false, + "reason": "CVT/FNet and Segformer are different model code paths." + }, + { + "left": "pull_request:44085", + "right": "pull_request:44129", + "accept": false, + "reason": "GPT-J and SpeechT5 are different models and changes." + }, + { + "left": "pull_request:44066", + "right": "pull_request:44722", + "accept": true, + "reason": "Both are the same GPT-J/CodeGen standardized output-tracing refactor, with the same concrete code paths and issue target." + }, + { + "left": "pull_request:44044", + "right": "pull_request:44129", + "accept": false, + "reason": "DeBERTa and SpeechT5 are unrelated model implementations." + }, + { + "left": "pull_request:44044", + "right": "pull_request:44154", + "accept": false, + "reason": "DeBERTa and VITS are different model code paths, so they are not duplicates." + } + ] + }, + "evaluator_result": { + "accept": true, + "feedback": "The summary is grounded in the packet and stays conservative about cross-model pairs. The two accepted duplicate-like pairs (GPT-J iterations and GPT-Neo iterations) look mergeable into single PRs for the same refactor/bugfix, and the chosen canonical PR (44722) is supported by the broader GPT-J/CodeGen scope plus the shared issue target. No clear overstatement beyond the evidence provided." + }, + "evaluator_used": true, + "retried": false, + "accepted_nontrivial_soft_edge": true, + "error_kind": null, + "error_message": null + }, + { + "label": "pull_request", + "component_index": 13, + "component_count": 35, + "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:44024", + "pull_request:44056", + "pull_request:44073", + "pull_request:44074", + "pull_request:44076", + "pull_request:44129" + ], + "soft_pairs": [ + "pull_request:44024|pull_request:44076", + "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": "799805cb8a35ed347f3c6174e20f4fd04323c90029b598122686f635398a0892", + "component_budget": { + "node_count": 28, + "item_count": 28, + "soft_pair_count": 78, + "serialized_chars": 84902, + "estimated_input_tokens": 21226, + "estimated_eval_tokens": 42708 + }, + "budget": { + "node_count": 8, + "item_count": 8, + "soft_pair_count": 6, + "serialized_chars": 17127, + "estimated_input_tokens": 4282, + "estimated_eval_tokens": 8820 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": true, + "analyst_result": { + "summary": "This cluster is a set of model-specific refactors to migrate different Transformers models to standardized output capture/tracing. They are thematically similar but touch different architectures and code paths, so they should not be treated as duplicates.", + "confidence": 0.92, + "canonical_issue_reason": null, + "canonical_pr_reason": "If one PR has to represent the cluster, 44129 is the most concrete and self-contained: it has an explicit issue target, a narrowly scoped SpeechT5 refactor, and clear standardized output-capture changes.", + "best_issue_reason": null, + "best_pr_reason": "44129 is the strongest standalone PR in this set because it is explicitly tied to an issue and has a focused, model-specific implementation rather than a broad or ambiguous change.", + "soft_edge_verdicts": [ + { + "left": "pull_request:44024", + "right": "pull_request:44076", + "accept": false, + "reason": "Same general migration theme, but FocalNet and ImageGPT are different models with different forward paths; not the same underlying change." + }, + { + "left": "pull_request:44073", + "right": "pull_request:44074", + "accept": false, + "reason": "Both refactor output tracing, but VisualBert and TextNet are separate model implementations and these patches are not interchangeable." + }, + { + "left": "pull_request:44056", + "right": "pull_request:44073", + "accept": false, + "reason": "MPNet vs VisualBert: different code paths and model-specific tracing hooks, so this is not a duplicate fix/change." + }, + { + "left": "pull_request:44056", + "right": "pull_request:44074", + "accept": false, + "reason": "MPNet and TextNet are distinct model families; they share a refactor pattern but not the same concrete bug or PR." + }, + { + "left": "pull_request:43997", + "right": "pull_request:44010", + "accept": false, + "reason": "RegNet and SqueezeBert are unrelated implementations; these are separate standardized-output migrations, not one mergeable PR." + }, + { + "left": "pull_request:44010", + "right": "pull_request:44129", + "accept": false, + "reason": "SqueezeBert and SpeechT5 change different model code paths; the explicit issue link on 44129 does not make them 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": 14, + "component_count": 35, + "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:44028", + "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": "a76970829c3bfb42dafbf0e139680ed56cf7c704d2b76a588ea24c395eeb4dea", + "component_budget": { + "node_count": 13, + "item_count": 13, + "soft_pair_count": 40, + "serialized_chars": 40180, + "estimated_input_tokens": 10045, + "estimated_eval_tokens": 20346 + }, + "budget": { + "node_count": 13, + "item_count": 13, + "soft_pair_count": 40, + "serialized_chars": 40180, + "estimated_input_tokens": 10045, + "estimated_eval_tokens": 20346 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": false, + "analyst_result": { + "summary": "These are not duplicates of one underlying fix; they are separate model-specific output-tracing refactors that share a common pattern but target different code paths.", + "confidence": 0.96, + "canonical_issue_reason": null, + "canonical_pr_reason": "PR 44003 is the strongest representative: it applies the tracing refactor across two closely related Mamba-family implementations and shows the full change pattern.", + "best_issue_reason": null, + "best_pr_reason": "PR 44003 is the best overall representative because it is the most substantive, self-contained example of the shared output-tracing refactor theme.", + "soft_edge_verdicts": [ + { + "left": "pull_request:43998", + "right": "pull_request:44002", + "accept": false, + "reason": "Different model families; same refactor motif only." + }, + { + "left": "pull_request:43998", + "right": "pull_request:44000", + "accept": false, + "reason": "Unrelated architectures; not the same concrete change." + }, + { + "left": "pull_request:43998", + "right": "pull_request:44001", + "accept": false, + "reason": "Different code paths in different model stacks." + }, + { + "left": "pull_request:43998", + "right": "pull_request:44003", + "accept": false, + "reason": "Timm backbone vs Mamba/FalconMamba; separate fixes." + }, + { + "left": "pull_request:43998", + "right": "pull_request:44029", + "accept": false, + "reason": "Distinct architectures; only the output-tracing pattern matches." + }, + { + "left": "pull_request:43999", + "right": "pull_request:44001", + "accept": false, + "reason": "MobileNetV1 and UnivNet are unrelated model implementations." + }, + { + "left": "pull_request:43999", + "right": "pull_request:44002", + "accept": false, + "reason": "Different subsystems; same broad refactor theme only." + }, + { + "left": "pull_request:43999", + "right": "pull_request:44025", + "accept": false, + "reason": "Separate models with different forward paths." + }, + { + "left": "pull_request:43999", + "right": "pull_request:44026", + "accept": false, + "reason": "Different wrapper/model codepaths; not one bug fix." + }, + { + "left": "pull_request:43999", + "right": "pull_request:44027", + "accept": false, + "reason": "MobileNetV1 vs speech encoder-decoder; unrelated changes." + }, + { + "left": "pull_request:44000", + "right": "pull_request:44001", + "accept": false, + "reason": "Vision-text dual encoder and UnivNet fix different code." + }, + { + "left": "pull_request:44000", + "right": "pull_request:44002", + "accept": false, + "reason": "Distinct model families; not mergeable as one PR." + }, + { + "left": "pull_request:44000", + "right": "pull_request:44003", + "accept": false, + "reason": "Different architectures and output codepaths." + }, + { + "left": "pull_request:44000", + "right": "pull_request:44025", + "accept": false, + "reason": "Different models; same refactor pattern only." + }, + { + "left": "pull_request:44001", + "right": "pull_request:44002", + "accept": false, + "reason": "UnivNet and UperNet are unrelated fixes." + }, + { + "left": "pull_request:44001", + "right": "pull_request:44003", + "accept": false, + "reason": "Different model stacks and tracing implementations." + }, + { + "left": "pull_request:44001", + "right": "pull_request:44025", + "accept": false, + "reason": "Separate model families; not the same underlying change." + }, + { + "left": "pull_request:44002", + "right": "pull_request:44029", + "accept": false, + "reason": "UperNet and RWKV touch different code paths." + }, + { + "left": "pull_request:44002", + "right": "pull_request:44003", + "accept": false, + "reason": "Different architectures; only similarly named refactor." + }, + { + "left": "pull_request:44003", + "right": "pull_request:44028", + "accept": false, + "reason": "Mamba-family tracing vs SuperPoint output handling." + }, + { + "left": "pull_request:44003", + "right": "pull_request:44004", + "accept": false, + "reason": "Mamba tracing refactor and CodeGen cache/signature cleanup differ." + }, + { + "left": "pull_request:43998", + "right": "pull_request:44004", + "accept": false, + "reason": "Timm backbone and CodeGen are unrelated model paths." + }, + { + "left": "pull_request:44001", + "right": "pull_request:44004", + "accept": false, + "reason": "UnivNet audio model vs CodeGen language model." + }, + { + "left": "pull_request:44002", + "right": "pull_request:44004", + "accept": false, + "reason": "Different subsystems; not the same bug or change." + }, + { + "left": "pull_request:44025", + "right": "pull_request:44026", + "accept": false, + "reason": "DepthAnything and VisionEncoderDecoder are separate wrappers." + }, + { + "left": "pull_request:44025", + "right": "pull_request:44027", + "accept": false, + "reason": "Different model families and forward implementations." + }, + { + "left": "pull_request:44025", + "right": "pull_request:44028", + "accept": false, + "reason": "Depth estimation vs keypoint model; distinct codepaths." + }, + { + "left": "pull_request:44025", + "right": "pull_request:44029", + "accept": false, + "reason": "DepthAnything and RWKV are unrelated." + }, + { + "left": "pull_request:44025", + "right": "pull_request:44030", + "accept": false, + "reason": "Different model families; same tracing pattern only." + }, + { + "left": "pull_request:44026", + "right": "pull_request:44029", + "accept": false, + "reason": "Vision encoder-decoder wrapper vs RWKV model." + }, + { + "left": "pull_request:44026", + "right": "pull_request:44027", + "accept": false, + "reason": "Vision and speech wrappers are separate implementations." + }, + { + "left": "pull_request:44026", + "right": "pull_request:44028", + "accept": false, + "reason": "Different model types; not one concrete fix." + }, + { + "left": "pull_request:44026", + "right": "pull_request:44030", + "accept": false, + "reason": "VisionEncoderDecoder and DPR are unrelated code paths." + }, + { + "left": "pull_request:44027", + "right": "pull_request:44028", + "accept": false, + "reason": "Speech encoder-decoder and SuperPoint are unrelated." + }, + { + "left": "pull_request:44027", + "right": "pull_request:44029", + "accept": false, + "reason": "Different model stacks; same general refactor only." + }, + { + "left": "pull_request:44027", + "right": "pull_request:44030", + "accept": false, + "reason": "Speech encoder-decoder vs DPR are separate changes." + }, + { + "left": "pull_request:44028", + "right": "pull_request:44029", + "accept": false, + "reason": "SuperPoint keypoint model vs RWKV language model." + }, + { + "left": "pull_request:44028", + "right": "pull_request:44030", + "accept": false, + "reason": "Different architectures and output semantics." + }, + { + "left": "pull_request:44029", + "right": "pull_request:44030", + "accept": false, + "reason": "RWKV and DPR are unrelated model implementations." + } + ] + }, + "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": 35, + "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": "0dedb6745de8d58695b9be3e3215b3524d241e0965dec433532e028318117a5c", + "component_budget": { + "node_count": 2, + "item_count": 2, + "soft_pair_count": 1, + "serialized_chars": 6552, + "estimated_input_tokens": 1638, + "estimated_eval_tokens": 3532 + }, + "budget": { + "node_count": 2, + "item_count": 2, + "soft_pair_count": 1, + "serialized_chars": 6552, + "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-integration PRs that share only generated boilerplate and auto-registration files. They target different concrete models and do not appear to be mergeable into one change.", + "confidence": 0.97, + "canonical_issue_reason": null, + "canonical_pr_reason": null, + "best_issue_reason": null, + "best_pr_reason": "PR 45401 is the cleaner representative if one must be chosen: it has an explicit issue target and a tighter, model-specific scope.", + "soft_edge_verdicts": [ + { + "left": "pull_request:44662", + "right": "pull_request:45401", + "accept": false, + "reason": "Different underlying changes: PenguinVL implementation vs Voxtral TTS support. Overlapping docs/auto-registry files are incidental and not evidence of the same bug or feature." + } + ] + }, + "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": 35, + "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": "dd882918d669d9f8ee37194ff5d8e5b9ab1739227019d630753690fc8839694b", + "component_budget": { + "node_count": 2, + "item_count": 2, + "soft_pair_count": 1, + "serialized_chars": 4405, + "estimated_input_tokens": 1102, + "estimated_eval_tokens": 2460 + }, + "budget": { + "node_count": 2, + "item_count": 2, + "soft_pair_count": 1, + "serialized_chars": 4405, + "estimated_input_tokens": 1102, + "estimated_eval_tokens": 2460 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": false, + "analyst_result": { + "summary": "Two open PRs touch fusion-related code, but they implement different features: one adds a generic Module Fusion API, the other adds KernelConfig-based n-to-1 kernel fusion integration. They are related in theme but not the same underlying change.", + "confidence": 0.94, + "canonical_issue_reason": null, + "canonical_pr_reason": "PR #45363 is the stronger representative of the cluster because it makes the broader end-to-end kernel-fusion integration change across core files, while #44979 is a narrower standalone API addition.", + "best_issue_reason": null, + "best_pr_reason": "PR #45363 is the best single item to surface because it has the broader user-facing impact and spans the main integration points for kernel fusion, whereas #44979 is an isolated API/module addition.", + "soft_edge_verdicts": [ + { + "left": "pull_request:44979", + "right": "pull_request:45363", + "accept": false, + "reason": "Both are about fusion, but #44979 introduces a generic Module Fusion API in a new module, while #45363 implements KernelConfig-driven n-to-1 kernel fusion and related integrations. They do not appear to fix the same concrete code-path problem 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": 17, + "component_count": 35, + "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": "896bf044af722b11ad0a6390d71c56e3f5368ca819b6a65ad16313f07a2ec665", + "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 implement the same fix for the Mistral regex patch AttributeError in the same file and target the same issue; 45317 is the fuller, test-backed version.", + "confidence": 0.98, + "canonical_issue_reason": null, + "canonical_pr_reason": "PR 45317 is the better canonical PR because it contains the same code fix as 45086 plus test coverage, making it the more complete change for the issue.", + "best_issue_reason": null, + "best_pr_reason": "PR 45317 is the best representative of the cluster: same underlying bug fix, same target issue, and it adds validation/tests beyond the minimal patch.", + "soft_edge_verdicts": [ + { + "left": "pull_request:45086", + "right": "pull_request:45317", + "accept": true, + "reason": "Same concrete bug fix in tokenization_utils_tokenizers.py for _patch_mistral_regex; 45317 is essentially the expanded version of 45086 with tests." + } + ] + }, + "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": 35, + "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-45115-4", + "nodes": [ + "pull_request:45115", + "pull_request:45569", + "pull_request:46332", + "pull_request:46565" + ], + "soft_pairs": [ + "pull_request:45115|pull_request:45569", + "pull_request:45569|pull_request:46565", + "pull_request:46332|pull_request:46565" + ], + "prepared_review_unit_hash": "2f3d6453a8989af9475bcbd3c729691affe279d51376388978f1798060a8a539", + "component_budget": { + "node_count": 4, + "item_count": 4, + "soft_pair_count": 3, + "serialized_chars": 13472, + "estimated_input_tokens": 3368, + "estimated_eval_tokens": 6992 + }, + "budget": { + "node_count": 4, + "item_count": 4, + "soft_pair_count": 3, + "serialized_chars": 13472, + "estimated_input_tokens": 3368, + "estimated_eval_tokens": 6992 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": false, + "analyst_result": { + "summary": "These PRs are all Nemotron-related, but they split into distinct change sets: a NemotronH inheritance/refactor, a NemotronH dense/sparse split, and two separate ASR streaming model additions. The overlap is mostly shared framework/docs plumbing rather than the same underlying bug or feature.", + "confidence": 0.92, + "canonical_issue_reason": null, + "canonical_pr_reason": "PR 45569 is the most central Nemotron-family architecture change in the cluster and the closest thing to a unifying representative, since it covers the core NemotronH refactor rather than a narrower follow-up or a separate ASR feature line.", + "best_issue_reason": null, + "best_pr_reason": "PR 45569 is the best overall representative for this cluster because it has the broadest impact on the core NemotronH implementation and ties most directly to the related NemotronH refactor PR; the ASR PRs are a different product area, not the same change.", + "soft_edge_verdicts": [ + { + "left": "pull_request:45115", + "right": "pull_request:45569", + "accept": false, + "reason": "Related NemotronH work, but not the same change: 45115 is a GraniteMoeHybrid inheritance/mapping refactor, while 45569 splits NemotronH into dense/sparse variants and updates docs/model registration. Shared files reflect the same subsystem, not duplicate code-path fixes." + }, + { + "left": "pull_request:45569", + "right": "pull_request:46565", + "accept": false, + "reason": "Different model families and goals: 45569 refactors NemotronH architecture, while 46565 adds Nemotron 3.5 ASR streaming. Similar auto/docs/loss files are shared infrastructure, not evidence of one mergeable PR." + }, + { + "left": "pull_request:46332", + "right": "pull_request:46565", + "accept": false, + "reason": "Both are ASR streaming PRs, but they target different concrete models (Nemotron ASR vs Nemotron 3.5 ASR). That makes them related follow-ups, not the same underlying feature or code-path fix." + } + ] + }, + "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": 35, + "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-45332-2", + "nodes": [ + "pull_request:45332", + "pull_request:45333" + ], + "soft_pairs": [ + "pull_request:45332|pull_request:45333" + ], + "prepared_review_unit_hash": "77ffd1df759b8457bfc82a0bbfbc30f73c5e037c633125312039b7557e58b351", + "component_budget": { + "node_count": 2, + "item_count": 2, + "soft_pair_count": 1, + "serialized_chars": 5466, + "estimated_input_tokens": 1367, + "estimated_eval_tokens": 2990 + }, + "budget": { + "node_count": 2, + "item_count": 2, + "soft_pair_count": 1, + "serialized_chars": 5466, + "estimated_input_tokens": 1367, + "estimated_eval_tokens": 2990 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": false, + "analyst_result": { + "summary": "Two related heterogeneous-model feature PRs, but they cover different layers of the implementation: 45333 adds per-layer configuration plumbing, while 45332 adds model/runtime integration on top of that. They are not duplicates.", + "confidence": 0.94, + "canonical_issue_reason": null, + "canonical_pr_reason": "45333 is the more foundational, narrowly scoped PR: it introduces the per-layer config API and helper utilities that 45332 builds on. It is the better anchor if these are grouped by topic.", + "best_issue_reason": null, + "best_pr_reason": "45332 is the broader representative of the heterogeneous-model effort because it touches cache, masking, modeling, and tests; but it is still a separate follow-on change, not a duplicate of 45333.", + "soft_edge_verdicts": [ + { + "left": "pull_request:45332", + "right": "pull_request:45333", + "accept": false, + "reason": "Related feature stack, but not the same concrete change. 45333 adds per-layer config support; 45332 adds heterogeneous model/runtime support that depends on that config plumbing. They could be sequential PRs, not one 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": 20, + "component_count": 35, + "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": "d65d51d795973219e9aa510f92cf2a9e8ba39d0c028085e69e0d52fa3202a52b", + "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 PRs are typing-related, but they target different scopes: one is a broad repo-wide type-checking pass, the other is a narrower typing refactor centered on modeling_utils and shared typing helpers. They do not look like the same underlying change and should not be merged as duplicates.", + "confidence": 0.88, + "canonical_issue_reason": null, + "canonical_pr_reason": null, + "best_issue_reason": null, + "best_pr_reason": "PR 45425 is the more focused change and reads like a distinct, self-contained typing cleanup, but it is not a duplicate of 45415.", + "soft_edge_verdicts": [ + { + "left": "pull_request:45415", + "right": "pull_request:45425", + "accept": false, + "reason": "Different change intents and scopes: 45415 is a broad type-checking sweep across many transformers modules, while 45425 adds specific typing protocols/helpers and related modeling_utils fixes. Shared files alone are not enough to treat them as the same concrete 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": 21, + "component_count": 35, + "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-45513-2", + "nodes": [ + "pull_request:45513", + "pull_request:46084" + ], + "soft_pairs": [ + "pull_request:45513|pull_request:46084" + ], + "prepared_review_unit_hash": "75a510b130d3913e763a9b65da0da2d93e7c73a5ccac50c461a4609f9b6de21d", + "component_budget": { + "node_count": 2, + "item_count": 2, + "soft_pair_count": 1, + "serialized_chars": 4948, + "estimated_input_tokens": 1237, + "estimated_eval_tokens": 2730 + }, + "budget": { + "node_count": 2, + "item_count": 2, + "soft_pair_count": 1, + "serialized_chars": 4948, + "estimated_input_tokens": 1237, + "estimated_eval_tokens": 2730 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": false, + "analyst_result": { + "summary": "These PRs are thematically similar (cache correctness with multi-token cached forward), but they fix different bugs in different model families and code paths, so they are not duplicates.", + "confidence": 0.95, + "canonical_issue_reason": null, + "canonical_pr_reason": "No canonical PR: 45513 and 46084 address separate cache bugs (Qwen3.5/Qwen3 Next/Olmo Hybrid linear attention vs. Mamba2 CUDA kernel cache handling), so they should not be merged or treated as the same change.", + "best_issue_reason": null, + "best_pr_reason": "46084 is the cleaner representative of its own bug: it is narrowly scoped, has an explicit issue target (#46032), and focuses on one concrete seq_len > 1 cache fix.", + "soft_edge_verdicts": [ + { + "left": "pull_request:45513", + "right": "pull_request:46084", + "accept": false, + "reason": "Reject: both are about cached forward behavior, but 45513 changes GDN linear attention paths in Qwen3.5/Qwen3 Next/Olmo Hybrid, while 46084 fixes Mamba2 use_cache handling in cuda_kernels_forward. Different underlying bugs 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": 22, + "component_count": 35, + "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": "2586c4914b4505a8f0081b862ab5f506e3c3c564f46177cb0ef07cb776448045", + "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": "Both PRs are about expert parallelism, but they fix different loading bugs in different distributed setups: one is config/serialization and EP detection for DeepSpeed ZeRO-3 via accelerate launch, the other fixes expert weights being overwritten during FSDP2/rank-0 broadcast.", + "confidence": 0.96, + "canonical_issue_reason": null, + "canonical_pr_reason": "PR 45662 is the more concrete representative of the cluster: it fixes a specific load-time overwrite path in model loading and distributed wrapping, with changes spread across the affected code paths.", + "best_issue_reason": null, + "best_pr_reason": "PR 45548 is narrower and more focused on EP config persistence and the has_ep helper for a DeepSpeed/accelerate loading path; it is less representative of the broader distributed-loading bug set than 45662.", + "soft_edge_verdicts": [ + { + "left": "pull_request:45548", + "right": "pull_request:45662", + "accept": false, + "reason": "Different underlying bugs and code paths: 45548 adds distributed_config serialization and EP detection for DeepSpeed ZeRO-3/accelerate launch, while 45662 prevents expert parameters from being overwritten during FSDP2 rank-0 broadcast. They are related to EP but not the same fix." + } + ] + }, + "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": 35, + "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": "1d5c620b5fe7359090138557f763a48640224306236171d5df8e59afad4ba826", + "component_budget": { + "node_count": 2, + "item_count": 2, + "soft_pair_count": 1, + "serialized_chars": 4590, + "estimated_input_tokens": 1148, + "estimated_eval_tokens": 2552 + }, + "budget": { + "node_count": 2, + "item_count": 2, + "soft_pair_count": 1, + "serialized_chars": 4590, + "estimated_input_tokens": 1148, + "estimated_eval_tokens": 2552 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": false, + "analyst_result": { + "summary": "Both PRs address the same xdist collision problem around `captured_info` debug artifacts; #45645 is a broader superset that also updates CI log collection, while #45639 is the narrower core fix.", + "confidence": 0.86, + "canonical_issue_reason": null, + "canonical_pr_reason": "PR #45645 is the best canonical representative because it contains the core xdist-safe artifact fix and the downstream CI changes needed to consume the new per-worker log files.", + "best_issue_reason": null, + "best_pr_reason": "#45645 is the most suitable overall PR in the cluster since it resolves the core bug and also handles the CI/reporting side effects; #45639 is a narrower variant of the same fix.", + "soft_edge_verdicts": [ + { + "left": "pull_request:45639", + "right": "pull_request:45645", + "accept": true, + "reason": "Same underlying bug/change: making patched testing debug logs xdist-safe for `captured_info` artifacts. #45645 extends the same fix with CI log aggregation changes, but the core code-path problem is shared." + } + ] + }, + "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": 35, + "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-45670-3", + "nodes": [ + "pull_request:45670", + "pull_request:46320", + "pull_request:46401" + ], + "soft_pairs": [ + "pull_request:45670|pull_request:46320", + "pull_request:45670|pull_request:46401", + "pull_request:46320|pull_request:46401" + ], + "prepared_review_unit_hash": "625601330ad794042ce6abc550b07aa6404120b3947f80c0fd19f9a9ccf83445", + "component_budget": { + "node_count": 3, + "item_count": 3, + "soft_pair_count": 3, + "serialized_chars": 7525, + "estimated_input_tokens": 1882, + "estimated_eval_tokens": 4020 + }, + "budget": { + "node_count": 3, + "item_count": 3, + "soft_pair_count": 3, + "serialized_chars": 7525, + "estimated_input_tokens": 1882, + "estimated_eval_tokens": 4020 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": false, + "analyst_result": { + "summary": "These are three separate pull requests with only superficial similarity. Two PRs add attn_implementation support, but they target different model families and code paths; the third is an AutoModel registry change for a different model entirely.", + "confidence": 0.92, + "canonical_issue_reason": null, + "canonical_pr_reason": null, + "best_issue_reason": null, + "best_pr_reason": null, + "soft_edge_verdicts": [ + { + "left": "pull_request:45670", + "right": "pull_request:46320", + "accept": false, + "reason": "Unrelated changes: one adds GLM-ASR to AutoModel mappings, the other implements attn_implementation dispatch for RecurrentGemma. Different feature area and code path." + }, + { + "left": "pull_request:45670", + "right": "pull_request:46401", + "accept": false, + "reason": "Unrelated changes: AutoModel mapping for glmasr versus attention backend dispatch for BLIP-2/InstructBLIP models. Not the same underlying bug or mergeable into one PR." + }, + { + "left": "pull_request:46320", + "right": "pull_request:46401", + "accept": false, + "reason": "Both are about attn_implementation support, but they fix different model-specific implementations (RecurrentGemma vs BLIP-2/InstructBLIP/InstructBLIPVideo). They are not the same concrete bug/change and should not be merged 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": 25, + "component_count": 35, + "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-45720-2", + "nodes": [ + "pull_request:45720", + "pull_request:46259" + ], + "soft_pairs": [ + "pull_request:45720|pull_request:46259" + ], + "prepared_review_unit_hash": "1467a267b1271f5365e316d09768243b7cb3a7c44a8edfe19effc74b18422542", + "component_budget": { + "node_count": 2, + "item_count": 2, + "soft_pair_count": 1, + "serialized_chars": 7248, + "estimated_input_tokens": 1812, + "estimated_eval_tokens": 3880 + }, + "budget": { + "node_count": 2, + "item_count": 2, + "soft_pair_count": 1, + "serialized_chars": 7248, + "estimated_input_tokens": 1812, + "estimated_eval_tokens": 3880 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": false, + "analyst_result": { + "summary": "These are two unrelated linter/chore PRs touching overlapping model files, but they implement different rules on different code paths: one changes forward return-tuple handling (TRF003) and the other changes weight-init patterns (TRF018).", + "confidence": 0.96, + "canonical_issue_reason": null, + "canonical_pr_reason": null, + "best_issue_reason": null, + "best_pr_reason": "PR 45720 is the more coherent single change: it targets one concrete behavior across model forwards (return-tuple handling) and is easier to reason about than the broader weight-init refactor in 46259.", + "soft_edge_verdicts": [ + { + "left": "pull_request:45720", + "right": "pull_request:46259", + "accept": false, + "reason": "Different underlying lint rules and fixes: TRF003 changes forward signatures/return behavior, while TRF018 changes initialization logic. Shared model filenames are incidental, 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": "pull_request", + "component_index": 26, + "component_count": 35, + "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": 27, + "component_count": 35, + "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-45932-2", + "nodes": [ + "pull_request:45932", + "pull_request:46100" + ], + "soft_pairs": [ + "pull_request:45932|pull_request:46100" + ], + "prepared_review_unit_hash": "ec70548425334d4881988fc90a97d62c4d1279266017962e423fefefb44096ea", + "component_budget": { + "node_count": 2, + "item_count": 2, + "soft_pair_count": 1, + "serialized_chars": 5084, + "estimated_input_tokens": 1271, + "estimated_eval_tokens": 2798 + }, + "budget": { + "node_count": 2, + "item_count": 2, + "soft_pair_count": 1, + "serialized_chars": 5084, + "estimated_input_tokens": 1271, + "estimated_eval_tokens": 2798 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": false, + "analyst_result": { + "summary": "These two pull requests are not duplicates: one is a broad CI/model maintenance PR with grafana workflow and model code changes, while the other is a ViT test-migration PR focused on test files. They do not appear to fix the same underlying bug or implement the same change.", + "confidence": 0.96, + "canonical_issue_reason": null, + "canonical_pr_reason": "No clear canonical PR for duplication: PR #45932 and PR #46100 address different areas and goals, so neither can serve as a duplicate representative for the other.", + "best_issue_reason": null, + "best_pr_reason": "PR #45932 is the broader, more substantive change set, but it is still unrelated to PR #46100; there is no single best PR as a duplicate canonicalization target.", + "soft_edge_verdicts": [ + { + "left": "pull_request:45932", + "right": "pull_request:46100", + "accept": false, + "reason": "Different changesets and purposes: #45932 updates CI/workflow and model code, while #46100 migrates ViT tests. They do not appear to fix the same concrete code-path 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": 28, + "component_count": 35, + "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-46073-3", + "nodes": [ + "pull_request:46073", + "pull_request:46154", + "pull_request:46196" + ], + "soft_pairs": [ + "pull_request:46073|pull_request:46154", + "pull_request:46073|pull_request:46196", + "pull_request:46154|pull_request:46196" + ], + "prepared_review_unit_hash": "caa04397d8d773a43aa91a204e9bba7f9274282dcb8740c12c8fd8b592641e46", + "component_budget": { + "node_count": 3, + "item_count": 3, + "soft_pair_count": 3, + "serialized_chars": 8398, + "estimated_input_tokens": 2100, + "estimated_eval_tokens": 4456 + }, + "budget": { + "node_count": 3, + "item_count": 3, + "soft_pair_count": 3, + "serialized_chars": 8398, + "estimated_input_tokens": 2100, + "estimated_eval_tokens": 4456 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": false, + "analyst_result": { + "summary": "These are three separate PRs adding SDPA/attn_implementation dispatch support to different model families (CTRL, T5, wav2vec2). They share a theme but touch different code paths and are not duplicates.", + "confidence": 0.97, + "canonical_issue_reason": null, + "canonical_pr_reason": null, + "best_issue_reason": null, + "best_pr_reason": "PR 46154 is the cleanest representative of the theme: a focused T5 SDPA dispatch change with accompanying tests, without the breadth of the wav2vec2 sweep.", + "soft_edge_verdicts": [ + { + "left": "pull_request:46073", + "right": "pull_request:46154", + "accept": false, + "reason": "Different model implementations and code paths (CTRL vs T5); same broad feature, not the same fix and not something to merge into one PR." + }, + { + "left": "pull_request:46073", + "right": "pull_request:46196", + "accept": false, + "reason": "CTRL attention dispatch and wav2vec2 attention dispatch are separate model-specific changes; no shared concrete bug/change." + }, + { + "left": "pull_request:46154", + "right": "pull_request:46196", + "accept": false, + "reason": "T5 and wav2vec2 are unrelated model code paths; both add SDPA support but for distinct implementations, so 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": 29, + "component_count": 35, + "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": "cc1f61506ec4ee537fe3117710f22b7d44c8105126b3438e35d819ab8005b06b", + "component_budget": { + "node_count": 2, + "item_count": 2, + "soft_pair_count": 1, + "serialized_chars": 5572, + "estimated_input_tokens": 1393, + "estimated_eval_tokens": 3042 + }, + "budget": { + "node_count": 2, + "item_count": 2, + "soft_pair_count": 1, + "serialized_chars": 5572, + "estimated_input_tokens": 1393, + "estimated_eval_tokens": 3042 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": false, + "analyst_result": { + "summary": "Two open PRs overlap on distributed/trainer files, but they are not the same concrete change: one is a focused EP + Trainer integration, the other is a broad distributed branch base touching many unrelated subsystems.", + "confidence": 0.88, + "canonical_issue_reason": null, + "canonical_pr_reason": "pull_request:46269 is the broader umbrella distributed branch-base PR; it spans the shared distributed files and substantially more, making it the best central representative of this cluster.", + "best_issue_reason": null, + "best_pr_reason": "pull_request:46269 is the strongest representative of the cluster because it is the larger base-change PR and not just a narrow follow-on integration.", + "soft_edge_verdicts": [ + { + "left": "pull_request:46126", + "right": "pull_request:46269", + "accept": false, + "reason": "They share some distributed filenames, but the scopes differ: 46126 is a focused EP/Trainer integration, while 46269 is a broad distributed branch base. They do not look like the same underlying fix and would not plausibly merge 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": 30, + "component_count": 35, + "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-46138-9", + "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" + ], + "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:46504", + "pull_request:46138|pull_request:46432", + "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:46558|pull_request:46563", + "pull_request:46504|pull_request:46563", + "pull_request:46432|pull_request:46563" + ], + "prepared_review_unit_hash": "ff01194cad2dc6db6c64ed80d8b5bbc48268c1eeb703cda7baf06419930cceff", + "component_budget": { + "node_count": 9, + "item_count": 9, + "soft_pair_count": 26, + "serialized_chars": 23357, + "estimated_input_tokens": 5840, + "estimated_eval_tokens": 11936 + }, + "budget": { + "node_count": 9, + "item_count": 9, + "soft_pair_count": 26, + "serialized_chars": 23357, + "estimated_input_tokens": 5840, + "estimated_eval_tokens": 11936 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": false, + "analyst_result": { + "summary": "These are mostly unrelated workflow-maintenance PRs that only look similar because of shared chore-style titles and YAML edits. The only close group is the pr-ci-security-gate.yml trio, but each PR changes a different part of that workflow, so they should not be merged as duplicates.", + "confidence": 0.91, + "canonical_issue_reason": null, + "canonical_pr_reason": "PR #46554 is the most substantive and central pr-ci-security-gate workflow change; the others are narrower follow-up tweaks or unrelated workflow edits.", + "best_issue_reason": null, + "best_pr_reason": "PR #46554 is the best representative of the cluster because it contains the real security-gate logic change, while the rest are smaller or different workflow updates.", + "soft_edge_verdicts": [ + { + "left": "pull_request:46138", + "right": "pull_request:46554", + "accept": false, + "reason": "Different workflows and different fixes: self-comment CI stdin handling vs PR CI security-gate logic." + }, + { + "left": "pull_request:46138", + "right": "pull_request:46558", + "accept": false, + "reason": "Different workflows and different concerns; no shared bug/change beyond both being YAML workflow edits." + }, + { + "left": "pull_request:46138", + "right": "pull_request:46563", + "accept": false, + "reason": "Different workflows; one updates self-comment CI, the other adjusts the PR CI security-gate reusable workflow note." + }, + { + "left": "pull_request:46138", + "right": "pull_request:46504", + "accept": false, + "reason": "Unrelated workflow changes: self-comment CI messaging vs build-ci-docker-images commit-message handling." + }, + { + "left": "pull_request:46138", + "right": "pull_request:46432", + "accept": false, + "reason": "Different workflows and different code paths: self-comment CI vs ssh-runner hardening." + }, + { + "left": "pull_request:46432", + "right": "pull_request:46503", + "accept": false, + "reason": "Different workflows: ssh-runner hardening vs benchmark_v2 trigger changes." + }, + { + "left": "pull_request:46432", + "right": "pull_request:46501", + "accept": false, + "reason": "Both touch workflows, but the changes are unrelated: SSH runner secret handling vs anti-slop permissions." + }, + { + "left": "pull_request:46432", + "right": "pull_request:46502", + "accept": false, + "reason": "Different workflow files and different behavior; ssh-runner setup is not the same as assign-reviewers checkout ref." + }, + { + "left": "pull_request:46432", + "right": "pull_request:46554", + "accept": false, + "reason": "Different workflows and different fixes: SSH runner hygiene vs PR CI security gate implementation." + }, + { + "left": "pull_request:46432", + "right": "pull_request:46558", + "accept": false, + "reason": "Different workflows and unrelated changes; no shared underlying bug." + }, + { + "left": "pull_request:46432", + "right": "pull_request:46563", + "accept": false, + "reason": "Different workflows and unrelated changes; ssh-runner secret cleanup is not the PR CI security-gate note update." + }, + { + "left": "pull_request:46501", + "right": "pull_request:46503", + "accept": false, + "reason": "Separate workflows with different purposes: anti-slop permissions vs benchmark trigger conditions." + }, + { + "left": "pull_request:46501", + "right": "pull_request:46502", + "accept": false, + "reason": "Different workflows and unrelated fixes; no common concrete bug." + }, + { + "left": "pull_request:46501", + "right": "pull_request:46504", + "accept": false, + "reason": "Unrelated workflow updates: anti-slop permissions vs build-ci-docker-images message parsing." + }, + { + "left": "pull_request:46501", + "right": "pull_request:46554", + "accept": false, + "reason": "Different workflows and different concerns; not the same change." + }, + { + "left": "pull_request:46502", + "right": "pull_request:46503", + "accept": false, + "reason": "Different workflows and different behavior changes; not duplicates." + }, + { + "left": "pull_request:46502", + "right": "pull_request:46504", + "accept": false, + "reason": "Assign-reviewers checkout ref change is unrelated to build-ci-docker-images message handling." + }, + { + "left": "pull_request:46502", + "right": "pull_request:46554", + "accept": false, + "reason": "Different workflows and different code paths; no same underlying bug/change." + }, + { + "left": "pull_request:46503", + "right": "pull_request:46504", + "accept": false, + "reason": "Different workflows: benchmark dispatch trigger vs build image commit-message parsing." + }, + { + "left": "pull_request:46503", + "right": "pull_request:46554", + "accept": false, + "reason": "Different workflows and different operational concerns; not duplicates." + }, + { + "left": "pull_request:46504", + "right": "pull_request:46554", + "accept": false, + "reason": "Different workflows and fixes: build-ci-docker-images parsing vs PR CI security-gate logic." + }, + { + "left": "pull_request:46504", + "right": "pull_request:46558", + "accept": false, + "reason": "Different workflows and unrelated changes; both are generic chore updates only superficially similar." + }, + { + "left": "pull_request:46554", + "right": "pull_request:46558", + "accept": false, + "reason": "Same file, but different concrete changes: security-check logic versus permissions/approval gating." + }, + { + "left": "pull_request:46554", + "right": "pull_request:46563", + "accept": false, + "reason": "Same file, but different concrete changes: security-check logic versus reusable-workflow commentary/annotation." + }, + { + "left": "pull_request:46558", + "right": "pull_request:46563", + "accept": false, + "reason": "Same workflow file, but the edits address different parts of the job and are not the same bug/change." + }, + { + "left": "pull_request:46504", + "right": "pull_request:46563", + "accept": false, + "reason": "Different workflows and different fixes; no duplicate evidence beyond generic chore similarity." + }, + { + "left": "pull_request:46432", + "right": "pull_request:46563", + "accept": false, + "reason": "Different workflows and unrelated changes; 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": "pull_request", + "component_index": 31, + "component_count": 35, + "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-46152-2", + "nodes": [ + "pull_request:46152", + "pull_request:46207" + ], + "soft_pairs": [ + "pull_request:46152|pull_request:46207" + ], + "prepared_review_unit_hash": "5341681ce81f1dd91db7614ce5d8dca6d323589e533298b0830c23cf33363d54", + "component_budget": { + "node_count": 2, + "item_count": 2, + "soft_pair_count": 1, + "serialized_chars": 4276, + "estimated_input_tokens": 1069, + "estimated_eval_tokens": 2394 + }, + "budget": { + "node_count": 2, + "item_count": 2, + "soft_pair_count": 1, + "serialized_chars": 4276, + "estimated_input_tokens": 1069, + "estimated_eval_tokens": 2394 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": false, + "analyst_result": { + "summary": "Two PRs share a broad theme of replacing file-open-based source inspection, but they target different modules and code paths, so they are not duplicates.", + "confidence": 0.88, + "canonical_issue_reason": null, + "canonical_pr_reason": "PR 46152 is the more general and central change: it replaces self-source file opening in the import-utils helper, with a broader infrastructure-oriented fix rather than a model-specific heuristic tweak.", + "best_issue_reason": null, + "best_pr_reason": "PR 46152 is the best representative of the cluster because its scope is cleaner and broader, while PR 46207 is a narrower fix in a separate attention-implementation path.", + "soft_edge_verdicts": [ + { + "left": "pull_request:46152", + "right": "pull_request:46207", + "accept": false, + "reason": "Same motivation, but not the same underlying bug/change: one rewrites source resolution in src/transformers/utils/import_utils.py, the other changes _can_set_attn_implementation in src/transformers/modeling_utils.py. Different code paths, different behaviors, 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": 32, + "component_count": 35, + "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-46173-3", + "nodes": [ + "pull_request:46173", + "pull_request:46436", + "pull_request:46474" + ], + "soft_pairs": [ + "pull_request:46436|pull_request:46474", + "pull_request:46173|pull_request:46474" + ], + "prepared_review_unit_hash": "03afe1d84421cabfe7d47addc353812dae48716212e2b9538554158c2840c089", + "component_budget": { + "node_count": 3, + "item_count": 3, + "soft_pair_count": 2, + "serialized_chars": 5921, + "estimated_input_tokens": 1481, + "estimated_eval_tokens": 3218 + }, + "budget": { + "node_count": 3, + "item_count": 3, + "soft_pair_count": 2, + "serialized_chars": 5921, + "estimated_input_tokens": 1481, + "estimated_eval_tokens": 3218 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": false, + "analyst_result": { + "summary": "PRs 46173 and 46474 are the same StaticLayer.get_seq_length type-annotation fix in cache_utils.py and should be treated as duplicates. PR 46436 is unrelated; it changes static cache length selection in generation, not the return type annotation.", + "confidence": 0.98, + "canonical_issue_reason": null, + "canonical_pr_reason": "PR 46474 is the cleanest representative of the duplicate pair: it makes the same cache_utils.py type-hint change as 46173, targets the same issue (#45987), and has the simpler, more direct title.", + "best_issue_reason": null, + "best_pr_reason": "PR 46474 is the best cluster representative because it is the same concrete fix as 46173 with the same issue target and identical diff, while being more succinctly titled. 46436 addresses a different generation-cache behavior change and is not a candidate for canonical PR.", + "soft_edge_verdicts": [ + { + "left": "pull_request:46436", + "right": "pull_request:46474", + "accept": false, + "reason": "Different bugs: 46436 changes max static cache length handling in generation/utils.py, while 46474 only updates the StaticLayer.get_seq_length type hint in cache_utils.py." + }, + { + "left": "pull_request:46173", + "right": "pull_request:46474", + "accept": true, + "reason": "Same underlying change: both modify cache_utils.py to widen StaticLayer.get_seq_length to int | torch.Tensor and both target issue 45987 with effectively identical diffs." + } + ] + }, + "evaluator_result": { + "accept": true, + "feedback": "The summary is grounded in the packet. PRs 46173 and 46474 appear to be the same bugfix: both change StaticLayer.get_seq_length in src/transformers/cache_utils.py from int to int | torch.Tensor and share issue target 45987. The soft-edge rejection for 46436 vs 46474 is also conservative, since 46436 changes static cache length handling in generation/utils.py and tests, which is a different fix." + }, + "evaluator_used": true, + "retried": false, + "accepted_nontrivial_soft_edge": true, + "error_kind": null, + "error_message": null + }, + { + "label": "pull_request", + "component_index": 33, + "component_count": 35, + "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:46546" + ], + "soft_pairs": [ + "pull_request:46310|pull_request:46546" + ], + "prepared_review_unit_hash": "1c3688b10b1b69882cc6c78ec73505b1cd5cd21445d2469dcdd89410fe6fd89a", + "component_budget": { + "node_count": 2, + "item_count": 2, + "soft_pair_count": 1, + "serialized_chars": 4851, + "estimated_input_tokens": 1213, + "estimated_eval_tokens": 2682 + }, + "budget": { + "node_count": 2, + "item_count": 2, + "soft_pair_count": 1, + "serialized_chars": 4851, + "estimated_input_tokens": 1213, + "estimated_eval_tokens": 2682 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": false, + "analyst_result": { + "summary": "Both pull requests fix the same sliding-window boundary bug (`-dist <= right_window_size` vs `<`) in Moonshine streaming; 46546 is the broader, more complete version because it also applies the same fix to Gemma4 and adds a test.", + "confidence": 0.91, + "canonical_issue_reason": null, + "canonical_pr_reason": "PR 46546 is the better canonical PR because it covers the Moonshine streaming fix plus the analogous Gemma4 change and includes test coverage, making it the more complete representative of this bug-fix pattern.", + "best_issue_reason": null, + "best_pr_reason": "PR 46546 is the best overall PR for this cluster: same concrete boundary fix as 46310, but broader scope and stronger validation.", + "soft_edge_verdicts": [ + { + "left": "pull_request:46310", + "right": "pull_request:46546", + "accept": true, + "reason": "Same underlying bug and same code-path fix pattern in the Moonshine streaming mask boundary; 46546 is effectively a superset of 46310\u2019s change and they could plausibly have been combined 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": 34, + "component_count": 35, + "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-46376-2", + "nodes": [ + "pull_request:46376", + "pull_request:46417" + ], + "soft_pairs": [ + "pull_request:46376|pull_request:46417" + ], + "prepared_review_unit_hash": "c7a9e6879820f903909586bffa60cf1cdd831f9aced04773eba10ce61d311633", + "component_budget": { + "node_count": 2, + "item_count": 2, + "soft_pair_count": 1, + "serialized_chars": 5938, + "estimated_input_tokens": 1485, + "estimated_eval_tokens": 3226 + }, + "budget": { + "node_count": 2, + "item_count": 2, + "soft_pair_count": 1, + "serialized_chars": 5938, + "estimated_input_tokens": 1485, + "estimated_eval_tokens": 3226 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": false, + "analyst_result": { + "summary": "The two pull requests are unrelated: one adds multimodal continuous batching support, while the other adds the HunYuan VL model with docs/tests. They should not be clustered as duplicates.", + "confidence": 0.97, + "canonical_issue_reason": null, + "canonical_pr_reason": "PR 46417 is the better canonical representative because it is a self-contained model addition with substantial implementation, docs, and tests, and it is more complete/mature than the draft continuous-batching feature PR.", + "best_issue_reason": null, + "best_pr_reason": "PR 46417 is the best standalone PR in this cluster because it has a clear single feature scope and appears more ready for merge than the draft multimodal batching work.", + "soft_edge_verdicts": [ + { + "left": "pull_request:46376", + "right": "pull_request:46417", + "accept": false, + "reason": "Different underlying changes: 46376 targets continuous batching infrastructure for multimodal support, while 46417 introduces a new HunYuan VL model stack, documentation, and tests. They do not fix the same code-path problem and would 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": 35, + "component_count": 35, + "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-46603-2", + "nodes": [ + "pull_request:46603", + "pull_request:46604" + ], + "soft_pairs": [ + "pull_request:46603|pull_request:46604" + ], + "prepared_review_unit_hash": "f50cc3fb4253f30f4a9a7a7dc2a59e38e226d664e4af085cfab961fa44393cc7", + "component_budget": { + "node_count": 2, + "item_count": 2, + "soft_pair_count": 1, + "serialized_chars": 5494, + "estimated_input_tokens": 1374, + "estimated_eval_tokens": 3004 + }, + "budget": { + "node_count": 2, + "item_count": 2, + "soft_pair_count": 1, + "serialized_chars": 5494, + "estimated_input_tokens": 1374, + "estimated_eval_tokens": 3004 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": false, + "analyst_result": { + "summary": "Two open Mistral-related PRs, but they are different changes: one refactors/moves the existing MistralConverter into a package, the other adds tekken tokenizer detection/conversion/save support. Shared files alone are not enough to treat them as duplicates.", + "confidence": 0.92, + "canonical_issue_reason": null, + "canonical_pr_reason": "PR 46604 is the more substantial, feature-bearing change; PR 46603 is a narrower code move/refactor of MistralConverter.", + "best_issue_reason": null, + "best_pr_reason": "PR 46604 best represents the cluster because it introduces the main new tokenizer functionality, while PR 46603 is just organizational cleanup.", + "soft_edge_verdicts": [ + { + "left": "pull_request:46603", + "right": "pull_request:46604", + "accept": false, + "reason": "Both touch Mistral tokenizer code, but one is a refactor/move of MistralConverter and the other adds tekken detection/conversion/save utilities. They are different underlying changes 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 + } + ] +} diff --git a/snapshots/20260615T180022Z/analysis-runs/hybrid-model-20260615t180022z/manifest.json b/snapshots/20260615T180022Z/analysis-runs/hybrid-model-20260615t180022z/manifest.json new file mode 100644 index 0000000000000000000000000000000000000000..673ad73f31f5de1f98bd260900dc727104ff524b --- /dev/null +++ b/snapshots/20260615T180022Z/analysis-runs/hybrid-model-20260615t180022z/manifest.json @@ -0,0 +1,14 @@ +{ + "analysis_id": "hybrid-model-20260615t180022z", + "artifacts": { + "hybrid": "snapshots/20260615T180022Z/analysis-runs/hybrid-model-20260615t180022z/analysis-report-hybrid.json", + "hybrid_reviews": "snapshots/20260615T180022Z/analysis-runs/hybrid-model-20260615t180022z/analysis-report-hybrid.llm-reviews.json" + }, + "channel": "canonical", + "model": null, + "published_at": "2026-06-15T18:12:05Z", + "repo": "huggingface/transformers", + "schema_version": 1, + "snapshot_id": "20260615T180022Z", + "variant": "hybrid" +} diff --git a/snapshots/20260615T180022Z/manifest.json b/snapshots/20260615T180022Z/manifest.json index 919be1e623dce61bee99e6aa55d7e49d0ddf16a7..77df2225a7bcbed6bf3e48b1e60a6ed646af6dba 100644 --- a/snapshots/20260615T180022Z/manifest.json +++ b/snapshots/20260615T180022Z/manifest.json @@ -32,6 +32,24 @@ "timeline_events": 2421 }, "extracted_at": "2026-06-15T18:00:22Z", + "published_analysis": { + "canonical_analysis_id": "hybrid-model-20260615t180022z", + "runs": { + "hybrid-model-20260615t180022z": { + "analysis_id": "hybrid-model-20260615t180022z", + "artifacts": { + "hybrid": "snapshots/20260615T180022Z/analysis-runs/hybrid-model-20260615t180022z/analysis-report-hybrid.json", + "hybrid_reviews": "snapshots/20260615T180022Z/analysis-runs/hybrid-model-20260615t180022z/analysis-report-hybrid.llm-reviews.json" + }, + "channel": "canonical", + "manifest_path": "snapshots/20260615T180022Z/analysis-runs/hybrid-model-20260615t180022z/manifest.json", + "model": null, + "published_at": "2026-06-15T18:12:05Z", + "variant": "hybrid" + } + }, + "schema_version": 1 + }, "repo": "huggingface/transformers", "snapshot_id": "20260615T180022Z", "watermark": {