diff --git a/analysis/current/analysis-report-hybrid.json b/analysis/current/analysis-report-hybrid.json index eb16bef0203264cb800f6da9100384a07294bb53..667dfe84a5731abfff1f848f6939f9b75f6eab97 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": "20260708T060015Z", - "generated_at": "2026-07-08T06:12:58Z", + "snapshot_id": "20260708T120111Z", + "generated_at": "2026-07-08T12:14:07Z", "evidence_quality": "full", "llm_enrichment": true, "meta_bugs": [ @@ -1050,8 +1050,8 @@ ] }, { - "cluster_id": "cluster-43451-4", - "summary": "Cluster of 4 related pull requests linked by soft_similarity.", + "cluster_id": "cluster-43451-6", + "summary": "Cluster of 6 related pull requests linked by soft_similarity.", "status": "open", "confidence": 0.5, "canonical_issue_number": null, @@ -1059,9 +1059,11 @@ "issue_numbers": [], "pr_numbers": [ 43451, - 46196, - 47092, - 47096 + 46912, + 47035, + 47096, + 47163, + 47166 ], "evidence_types": [ "soft_similarity" @@ -1069,23 +1071,23 @@ "pr_comparisons": [ { "left_pr_number": 43451, - "right_pr_number": 46196, - "code_similarity": 0.172, - "size_similarity": 0.767, + "right_pr_number": 46912, + "code_similarity": 0.108, + "size_similarity": 0.441, "file_overlap": 0.0, "area_overlap": 0.0, - "patch_similarity": 0.128, + "patch_similarity": 0.134, "shared_filenames": [], "shared_file_areas": [] }, { "left_pr_number": 43451, - "right_pr_number": 47092, + "right_pr_number": 47035, "code_similarity": 0.001, "size_similarity": 0.002, "file_overlap": 0.0, "area_overlap": 0.0, - "patch_similarity": 0.003, + "patch_similarity": 0.001, "shared_filenames": [], "shared_file_areas": [] }, @@ -1101,35 +1103,134 @@ "shared_file_areas": [] }, { - "left_pr_number": 46196, - "right_pr_number": 47092, - "code_similarity": 0.002, - "size_similarity": 0.002, + "left_pr_number": 43451, + "right_pr_number": 47163, + "code_similarity": 0.004, + "size_similarity": 0.005, "file_overlap": 0.0, "area_overlap": 0.0, - "patch_similarity": 0.011, + "patch_similarity": 0.02, "shared_filenames": [], "shared_file_areas": [] }, { - "left_pr_number": 46196, - "right_pr_number": 47096, - "code_similarity": 0.045, - "size_similarity": 0.129, + "left_pr_number": 43451, + "right_pr_number": 47166, + "code_similarity": 0.004, + "size_similarity": 0.011, "file_overlap": 0.0, "area_overlap": 0.0, - "patch_similarity": 0.129, + "patch_similarity": 0.013, "shared_filenames": [], "shared_file_areas": [] }, { - "left_pr_number": 47092, + "left_pr_number": 46912, + "right_pr_number": 47035, + "code_similarity": 0.001, + "size_similarity": 0.004, + "file_overlap": 0.0, + "area_overlap": 0.0, + "patch_similarity": 0.002, + "shared_filenames": [], + "shared_file_areas": [] + }, + { + "left_pr_number": 46912, "right_pr_number": 47096, + "code_similarity": 0.06, + "size_similarity": 0.224, + "file_overlap": 0.0, + "area_overlap": 0.0, + "patch_similarity": 0.1, + "shared_filenames": [], + "shared_file_areas": [] + }, + { + "left_pr_number": 46912, + "right_pr_number": 47163, "code_similarity": 0.005, + "size_similarity": 0.01, + "file_overlap": 0.0, + "area_overlap": 0.0, + "patch_similarity": 0.022, + "shared_filenames": [], + "shared_file_areas": [] + }, + { + "left_pr_number": 46912, + "right_pr_number": 47166, + "code_similarity": 0.008, + "size_similarity": 0.025, + "file_overlap": 0.0, + "area_overlap": 0.0, + "patch_similarity": 0.017, + "shared_filenames": [], + "shared_file_areas": [] + }, + { + "left_pr_number": 47035, + "right_pr_number": 47096, + "code_similarity": 0.004, "size_similarity": 0.019, "file_overlap": 0.0, "area_overlap": 0.0, - "patch_similarity": 0.011, + "patch_similarity": 0.004, + "shared_filenames": [], + "shared_file_areas": [] + }, + { + "left_pr_number": 47035, + "right_pr_number": 47163, + "code_similarity": 0.083, + "size_similarity": 0.414, + "file_overlap": 0.0, + "area_overlap": 0.0, + "patch_similarity": 0.0, + "shared_filenames": [], + "shared_file_areas": [] + }, + { + "left_pr_number": 47035, + "right_pr_number": 47166, + "code_similarity": 0.039, + "size_similarity": 0.171, + "file_overlap": 0.0, + "area_overlap": 0.0, + "patch_similarity": 0.033, + "shared_filenames": [], + "shared_file_areas": [] + }, + { + "left_pr_number": 47096, + "right_pr_number": 47163, + "code_similarity": 0.013, + "size_similarity": 0.046, + "file_overlap": 0.0, + "area_overlap": 0.0, + "patch_similarity": 0.023, + "shared_filenames": [], + "shared_file_areas": [] + }, + { + "left_pr_number": 47096, + "right_pr_number": 47166, + "code_similarity": 0.025, + "size_similarity": 0.112, + "file_overlap": 0.0, + "area_overlap": 0.0, + "patch_similarity": 0.016, + "shared_filenames": [], + "shared_file_areas": [] + }, + { + "left_pr_number": 47163, + "right_pr_number": 47166, + "code_similarity": 0.089, + "size_similarity": 0.414, + "file_overlap": 0.0, + "area_overlap": 0.0, + "patch_similarity": 0.043, "shared_filenames": [], "shared_file_areas": [] } @@ -1473,6 +1574,70 @@ } ] }, + { + "cluster_id": "cluster-44779-5", + "summary": "Cluster of 3 issues and 2 PRs centered on issue #45488.", + "status": "open", + "confidence": 0.6, + "canonical_issue_number": 45488, + "canonical_pr_number": 47017, + "issue_numbers": [ + 44779, + 45488, + 46489 + ], + "pr_numbers": [ + 46580, + 47017 + ], + "evidence_types": [ + "closing_reference", + "soft_similarity" + ], + "pr_comparisons": [ + { + "left_pr_number": 46580, + "right_pr_number": 47017, + "code_similarity": 0.241, + "size_similarity": 0.352, + "file_overlap": 0.5, + "area_overlap": 0.0, + "patch_similarity": 0.136, + "shared_filenames": [ + "src/transformers/models/auto/tokenization_auto.py" + ], + "shared_file_areas": [ + { + "filename": "src/transformers/models/auto/tokenization_auto.py", + "left_ranges": [ + [ + 361, + 367 + ], + [ + 399, + 421 + ], + [ + 724, + 739 + ] + ], + "right_ranges": [ + [ + 428, + 475 + ], + [ + 807, + 825 + ] + ] + } + ] + } + ] + }, { "cluster_id": "cluster-43240-3", "summary": "Cluster of 1 issues and 2 PRs centered on issue #43240.", @@ -2152,56 +2317,65 @@ ] }, { - "cluster_id": "cluster-43698-3", - "summary": "Cluster of 1 issues and 2 PRs centered on issue #43698.", + "cluster_id": "cluster-46305-3", + "summary": "Cluster of 1 issues and 2 PRs centered on issue #46305.", "status": "open", - "confidence": 0.75, - "canonical_issue_number": 43698, - "canonical_pr_number": 43779, + "confidence": 0.8, + "canonical_issue_number": 46305, + "canonical_pr_number": 46310, "issue_numbers": [ - 43698 + 46305 ], "pr_numbers": [ - 43779, - 43816 + 46310, + 47162 ], "evidence_types": [ "closing_reference", - "shared_issue_target" + "shared_issue_target", + "soft_similarity" ], "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, + "left_pr_number": 46310, + "right_pr_number": 47162, + "code_similarity": 0.577, + "size_similarity": 0.338, + "file_overlap": 0.5, + "area_overlap": 1.0, + "patch_similarity": 0.062, "shared_filenames": [ - "src/transformers/integrations/integration_utils.py" + "src/transformers/models/moonshine_streaming/modeling_moonshine_streaming.py", + "src/transformers/models/moonshine_streaming/modular_moonshine_streaming.py" ], "shared_file_areas": [ { - "filename": "src/transformers/integrations/integration_utils.py", + "filename": "src/transformers/models/moonshine_streaming/modeling_moonshine_streaming.py", "left_ranges": [ [ - 2227, - 2241 - ], - [ - 2303, - 2309 + 356, + 362 ] ], "right_ranges": [ [ - 2278, - 2291 - ], + 356, + 362 + ] + ] + }, + { + "filename": "src/transformers/models/moonshine_streaming/modular_moonshine_streaming.py", + "left_ranges": [ [ - 2309, - 2322 + 269, + 275 + ] + ], + "right_ranges": [ + [ + 269, + 275 ] ] } @@ -2210,101 +2384,56 @@ ] }, { - "cluster_id": "cluster-46836-2", - "summary": "Cluster of 2 related pull requests linked by soft_similarity.", + "cluster_id": "cluster-43698-3", + "summary": "Cluster of 1 issues and 2 PRs centered on issue #43698.", "status": "open", - "confidence": 0.5, - "canonical_issue_number": null, - "canonical_pr_number": 46836, - "issue_numbers": [], + "confidence": 0.75, + "canonical_issue_number": 43698, + "canonical_pr_number": 43779, + "issue_numbers": [ + 43698 + ], "pr_numbers": [ - 46836, - 47143 + 43779, + 43816 ], "evidence_types": [ - "soft_similarity" + "closing_reference", + "shared_issue_target" ], "pr_comparisons": [ { - "left_pr_number": 46836, - "right_pr_number": 47143, - "code_similarity": 0.169, - "size_similarity": 0.48, - "file_overlap": 0.111, - "area_overlap": 0.0, - "patch_similarity": 0.266, + "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/models/__init__.py", - "src/transformers/models/auto/auto_mappings.py", - "src/transformers/models/auto/modeling_auto.py" + "src/transformers/integrations/integration_utils.py" ], "shared_file_areas": [ { - "filename": "src/transformers/models/__init__.py", - "left_ranges": [ - [ - 452, - 458 - ] - ], - "right_ranges": [ - [ - 305, - 311 - ] - ] - }, - { - "filename": "src/transformers/models/auto/auto_mappings.py", - "left_ranges": [ - [ - 628, - 638 - ], - [ - 905, - 914 - ], - [ - 1112, - 1118 - ], - [ - 1238, - 1244 - ] - ], - "right_ranges": [ - [ - 412, - 418 - ], - [ - 975, - 981 - ] - ] - }, - { - "filename": "src/transformers/models/auto/modeling_auto.py", + "filename": "src/transformers/integrations/integration_utils.py", "left_ranges": [ [ - 501, - 507 + 2227, + 2241 ], [ - 1107, - 1113 + 2303, + 2309 ] ], "right_ranges": [ [ - 348, - 354 + 2278, + 2291 ], [ - 1960, - 1966 + 2309, + 2322 ] ] } @@ -2589,6 +2718,15 @@ } ], "duplicate_issues": [ + { + "cluster_id": "cluster-44779-5", + "canonical_issue_number": 45488, + "duplicate_issue_numbers": [ + 44779, + 46489 + ], + "reason": "Issues in cluster-44779-5 are treated as duplicates because they share closing_reference, soft_similarity evidence." + }, { "cluster_id": "cluster-46894-5", "canonical_issue_number": 46894, @@ -2618,15 +2756,17 @@ "reason": "PRs in cluster-43240-3 are treated as duplicates because they converge on issue #43240 with closing_reference, shared_issue_target evidence." }, { - "cluster_id": "cluster-43451-4", + "cluster_id": "cluster-43451-6", "canonical_pr_number": 43451, "duplicate_pr_numbers": [ - 46196, - 47092, - 47096 + 46912, + 47035, + 47096, + 47163, + 47166 ], "target_issue_number": null, - "reason": "PRs in cluster-43451-4 are treated as duplicates because they share soft_similarity evidence." + "reason": "PRs in cluster-43451-6 are treated as duplicates because they share soft_similarity evidence." }, { "cluster_id": "cluster-43698-3", @@ -2663,6 +2803,15 @@ "target_issue_number": null, "reason": "PRs in cluster-44018-2 are treated as duplicates because they share soft_similarity evidence." }, + { + "cluster_id": "cluster-44779-5", + "canonical_pr_number": 47017, + "duplicate_pr_numbers": [ + 46580 + ], + "target_issue_number": 45488, + "reason": "PRs in cluster-44779-5 are treated as duplicates because they converge on issue #45488 with closing_reference, soft_similarity evidence." + }, { "cluster_id": "cluster-45081-3", "canonical_pr_number": 45317, @@ -2699,6 +2848,15 @@ "target_issue_number": 46159, "reason": "PRs in cluster-46159-3 are treated as duplicates because they converge on issue #46159 with closing_reference evidence." }, + { + "cluster_id": "cluster-46305-3", + "canonical_pr_number": 46310, + "duplicate_pr_numbers": [ + 47162 + ], + "target_issue_number": 46305, + "reason": "PRs in cluster-46305-3 are treated as duplicates because they converge on issue #46305 with closing_reference, shared_issue_target, soft_similarity evidence." + }, { "cluster_id": "cluster-46693-3", "canonical_pr_number": 47134, @@ -2708,15 +2866,6 @@ "target_issue_number": 46693, "reason": "PRs in cluster-46693-3 are treated as duplicates because they converge on issue #46693 with closing_reference evidence." }, - { - "cluster_id": "cluster-46836-2", - "canonical_pr_number": 46836, - "duplicate_pr_numbers": [ - 47143 - ], - "target_issue_number": null, - "reason": "PRs in cluster-46836-2 are treated as duplicates because they share soft_similarity evidence." - }, { "cluster_id": "cluster-46869-3", "canonical_pr_number": 46886, @@ -2738,10 +2887,10 @@ } ], "best_issue": { - "cluster_id": "cluster-46894-5", - "issue_number": 46894, - "reason": "Issue #46894 is the strongest global issue candidate because it is open, belongs to a cluster with 5 artifacts, and carries 1 discussion comments plus 9 inbound references.", - "score": 161.0 + "cluster_id": "cluster-44779-5", + "issue_number": 45488, + "reason": "Issue #45488 is the strongest global issue candidate because it is open, belongs to a cluster with 5 artifacts, and carries 16 discussion comments plus 4 inbound references.", + "score": 186.0 }, "best_pr": { "cluster_id": "cluster-43979-11", diff --git a/analysis/current/analysis-report-hybrid.llm-reviews.json b/analysis/current/analysis-report-hybrid.llm-reviews.json index f84e6dcd1b759736293cf5c3c8ebb047287a138c..f8a5bab6c1418b5e5bc1e65e1738b1521fefd571 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": "20260708T060015Z", - "generated_at": "2026-07-08T06:12:58Z", + "snapshot_id": "20260708T120111Z", + "generated_at": "2026-07-08T12:14:07Z", "model": "gpt-5.4-mini?service_tier=flex", "reviews": [ { @@ -131,334 +131,340 @@ "issue:45237|issue:46015", "issue:43039|issue:44805" ], - "prepared_review_unit_hash": "659f863ec8c91cf3f47c0673ab1ea1c903f76d3d98935f689a7d811842787cc2", + "prepared_review_unit_hash": "97619858853a9afd1a3f06970dcc6263fa045271b03d45966a259ee3c741e7ea", "component_budget": { "node_count": 514, "item_count": 514, "soft_pair_count": 1891, - "serialized_chars": 980926, - "estimated_input_tokens": 245232, - "estimated_eval_tokens": 490720 + "serialized_chars": 980895, + "estimated_input_tokens": 245224, + "estimated_eval_tokens": 490704 }, "budget": { "node_count": 48, "item_count": 48, "soft_pair_count": 62, - "serialized_chars": 46282, - "estimated_input_tokens": 11571, - "estimated_eval_tokens": 23398 + "serialized_chars": 46274, + "estimated_input_tokens": 11569, + "estimated_eval_tokens": 23394 }, "overflow_policy": "truncate_then_skip", "trimmed": false, "aggressively_trimmed": false, "split": true, "analyst_result": { - "summary": "Mostly a heterogeneous cluster of transformer bugs and docs regressions. The clearest duplicate pair is the DeepSpeed ZeRO-3/BERT IndexError reports (43596/43638), and there is a plausible torch 2.4.1 compatibility duplicate pair (45800/46515). Most other pairs share only a broad subsystem and are not the same underlying bug.", - "confidence": 0.76, - "canonical_issue_reason": "Issue 43596 is the clearest representative of a true duplicate group: 43638 reports the same DeepSpeed ZeRO-3/BERT IndexError with nearly the same failure mode.", + "summary": "This cluster is mostly a set of unrelated issues grouped by broad text similarity. The only clear duplicate families are the DeepSpeed ZeRO-3 \"index 0 is out of bounds\" reports, a torch-2.4.1 compatibility pair, and a Qwen3_5MoeTextRotaryEmbedding offload/materialization pair.", + "confidence": 0.74, + "canonical_issue_reason": "Issue 43596 is the most general representative of the DeepSpeed ZeRO-3 index-out-of-bounds family; 43638 and 41762 read like model-specific variants of the same zero-sized tensor/init bug.", "canonical_pr_reason": null, - "best_issue_reason": "43596 is the strongest issue to keep as the cluster representative because it has the most concrete duplicate evidence and a clear, reproducible bug description.", + "best_issue_reason": "43596 is the best global issue anchor because it describes the shared failure mode most cleanly and broadly, while the other accepted duplicates are narrower reproductions of the same ZeRO-3 init problem.", "best_pr_reason": null, "soft_edge_verdicts": [ { "left": "issue:44322", "right": "issue:45464", "accept": false, - "reason": "Qwen3.5 config attribute failure vs streaming chat/completions failure; different code paths and symptoms." + "reason": "Different Qwen3.5 failures: missing config field vs streaming chat/completions failure." }, { "left": "issue:45237", "right": "issue:45685", "accept": false, - "reason": "Different bugs: GPT-OSS device-map failure on AMD GPUs versus an MPS histogram op limitation." + "reason": "Unrelated hardware/runtime problems: AMD GPU support vs MPS histogram op support." }, { "left": "issue:43716", "right": "issue:45237", "accept": false, - "reason": "Both involve model/runtime mismatches, but one is a Mistral-3 dtype issue and the other is a GPT-OSS AMD GPU failure." + "reason": "Different bug classes and models; dtype mismatch in preprocessing is not the AMD GPU failure." }, { "left": "issue:43828", "right": "issue:45237", "accept": false, - "reason": "Autocast dtype mismatch for Phi-tiny-MoE is not the same bug as GPT-OSS failing on AMD GPUs." + "reason": "Autocast dtype mismatch on Phi MoE is unrelated to GPT-OSS AMD GPU runtime failure." }, { "left": "issue:45237", "right": "issue:46468", "accept": false, - "reason": "Unrelated failures: GPT-OSS on AMD GPUs versus a broken Nemotron example." + "reason": "Different models and symptoms; Nemotron example breakage is not the GPT-OSS AMD GPU issue." }, { "left": "issue:45237", "right": "issue:45507", "accept": false, - "reason": "GPT-OSS device-map/GPU issue and GraniteMoEHybrid invalid-method issue are different concrete problems." + "reason": "Different concrete failures: AMD GPU runtime issue vs GraniteMoEHybrid invalid-method call." }, { "left": "issue:30277", "right": "issue:35545", "accept": false, - "reason": "DeepSpeed ZeRO-3 collective mismatch on Jamba is unrelated to ModernBERT ONNX export failure." + "reason": "DeepSpeed collective mismatch and ONNX export error are unrelated." }, { "left": "issue:35545", "right": "issue:42915", "accept": false, - "reason": "ONNX export error and Qwen3Moe FP8 config failure are distinct issues." + "reason": "ONNX export failure and FineGrainedFP8 failure do not share the same code-path." }, { "left": "issue:35545", "right": "issue:41720", "accept": false, - "reason": "ModernBERT export failure is unrelated to Qwen3 auto device mapping CUDA assert." + "reason": "ONNX export error vs Qwen3 auto-device-map CUDA assert are different issues." }, { "left": "issue:45901", "right": "issue:46015", "accept": false, - "reason": "Table QA runtime crash and outdated RAG docs are different issue types and different problems." + "reason": "Both are docs/examples-related at a high level, but they target different content and fixes." }, { "left": "issue:35545", "right": "issue:36010", "accept": false, - "reason": "Export-time ModernBERT failure does not match the GenerationMixin import regression." + "reason": "Export-to-ONNX and GenerationMixin import errors are unrelated." }, { "left": "issue:44509", "right": "issue:44512", "accept": false, - "reason": "Both are docs cleanups for v5 removals, but they target different removed references and are not the same change." + "reason": "Both are v5 docs cleanup items, but they reference different removed features and likely different edits." }, { "left": "issue:44918", "right": "issue:46165", "accept": false, - "reason": "TRL SFT embedding unpacking bug is unrelated to resuming training with trust_remote_code." + "reason": "TRl SFT embedding unpacking and model resume/trust_remote_code are different training issues." }, { "left": "issue:46165", "right": "issue:46869", "accept": false, - "reason": "Resume-training/remote-code issue is different from YOLO multi-GPU device-map auto failure." + "reason": "Remote-code resume bug and YOLO multi-GPU device-map bug are unrelated." }, { "left": "issue:44805", "right": "issue:45823", "accept": false, - "reason": "Mask shape mismatch and Gemma4 device mismatch are different runtime errors." + "reason": "Mask-shape mismatch and Gemma4 device mismatch are different forward-path failures." }, { "left": "issue:43638", "right": "issue:44805", "accept": false, - "reason": "Both are IndexErrors, but one is ZeRO-3/BERT loading and the other is a mask/tensor shape mismatch; not the same bug." + "reason": "Zero-3 index-out-of-bounds and mask/tensor shape mismatch are different failures." }, { "left": "issue:43638", "right": "issue:46165", "accept": false, - "reason": "ZeRO-3/BERT IndexError is unrelated to the model-resume/trust_remote_code issue." + "reason": "DeepSpeed ZeRO-3 init bug and model resume/trust_remote_code bug are unrelated." }, { "left": "issue:43596", "right": "issue:44805", "accept": false, - "reason": "Different failure modes: ZeRO-3/BERT loading IndexError versus mask shape mismatch." + "reason": "Different symptom and subsystem; this is not the same Zero-3 init bug." }, { "left": "issue:44805", "right": "issue:46294", "accept": false, - "reason": "Mask shape mismatch is unrelated to Qwen3.5 video token/feature mismatch." + "reason": "Mask/index tensor mismatch is unrelated to Qwen3.5 video token/feature mismatch." }, { "left": "issue:44805", "right": "issue:45161", "accept": false, - "reason": "Generic tensor shape mismatch and GPT-OSS MoE tensor-parallel issue are not the same underlying bug." + "reason": "Mask-shape error and GPT-OSS TP-only failure are different bugs." + }, + { + "left": "issue:46165", + "right": "issue:46468", + "accept": false, + "reason": "Different areas: resuming remote-code training vs Nemotron example breakage." + }, + { + "left": "issue:44560", + "right": "issue:46165", + "accept": false, + "reason": "Video chat-template failure and remote-code resume failure are unrelated." }, { "left": "issue:34634", "right": "issue:35545", "accept": false, - "reason": "BarkProcessor voice_preset issue is unrelated to ModernBERT ONNX export." + "reason": "BarkProcessor voice_preset bug is unrelated to ModernBERT ONNX export." }, { "left": "issue:30277", "right": "issue:43638", "accept": false, - "reason": "Both involve distributed training, but the errors and models differ (collective mismatch vs ZeRO-3/BERT IndexError)." + "reason": "Collective mismatch under ZeRO-3 is a different issue from the index-out-of-bounds init bug." }, { "left": "issue:43596", "right": "issue:43638", "accept": true, - "reason": "Same DeepSpeed ZeRO-3/BERT IndexError during loading/training; essentially the same concrete bug." + "reason": "Same DeepSpeed ZeRO-3 'index 0 is out of bounds' failure; BertModel vs non-pretrained Bert is just a repro variant." }, { "left": "issue:43039", "right": "issue:45161", "accept": false, - "reason": "Liger Kernel cross_entropy call issue is unrelated to GPT-OSS tensor-parallel behavior." + "reason": "Liger Kernel cross_entropy being called is unrelated to GPT-OSS tensor-parallel failure." }, { "left": "issue:43975", "right": "issue:44457", "accept": false, - "reason": "Incorrect detokenization and LoRA merge/save mismatch are different problems." + "reason": "Wrong detokenization and LoRA merge/save mismatch are different issues." }, { "left": "issue:44292", "right": "issue:44457", "accept": false, - "reason": "Qwen-3-8B-NVFP4 runtime failure is unrelated to LoRA weight merge/save inconsistency." + "reason": "Qwen-3 NVFP4 runtime failure is unrelated to LoRA merge/save behavior." }, { "left": "issue:44841", "right": "issue:46661", "accept": false, - "reason": "Processor failure for Voxtral is unrelated to remote-code model_class/config_class import behavior." + "reason": "Processor failure for Voxtral is unrelated to remote-code model_class/config_class import handling." }, { "left": "issue:43716", "right": "issue:46869", "accept": false, - "reason": "Mistral-3 preprocessing dtype mismatch is unrelated to YOLO multi-GPU device-map auto failure." + "reason": "Mistral-3 dtype mismatch and YOLO multi-GPU device-map failure are different issues." }, { "left": "issue:24643", "right": "issue:30277", "accept": false, - "reason": "DeepSpeed training weight-shape error and collective mismatch are different distributed-training bugs." + "reason": "Both involve DeepSpeed, but the concrete errors and code paths differ." }, { "left": "issue:34928", "right": "issue:44805", "accept": false, - "reason": "FSDP activation-checkpointing recompute mismatch is not the same as a mask shape IndexError." + "reason": "Activation-checkpointing/FSDP recompute mismatch is not the same as a mask-shape index error." }, { "left": "issue:45685", "right": "issue:46869", "accept": false, - "reason": "MPS histogram operator limitation is unrelated to YOLO multi-GPU device-map auto." + "reason": "MPS histogram limitation and YOLO multi-GPU device-map failure are unrelated." }, { "left": "issue:45859", "right": "issue:45902", - "accept": false, - "reason": "Both mention Qwen3_5MoeTextRotaryEmbedding, but one is CPU offload compatibility and the other is uninitialized-memory materialization; not clearly the same bug." + "accept": true, + "reason": "Same Qwen3_5MoeTextRotaryEmbedding buffer/offload bug; one report is a forward/offload symptom and the other is a materialization symptom." }, { "left": "issue:34928", "right": "issue:36331", "accept": false, - "reason": "Activation-checkpointing/FSDP recompute mismatch is unrelated to the CustomTrainer compute_loss signature regression." + "reason": "FSDP activation-checkpointing mismatch and CustomTrainer API breakage are unrelated." }, { "left": "issue:43638", "right": "issue:45161", "accept": false, - "reason": "ZeRO-3/BERT IndexError and GPT-OSS only-TP issue are unrelated." + "reason": "DeepSpeed ZeRO-3 init bug and GPT-OSS TP-only failure are unrelated." }, { "left": "issue:43596", "right": "issue:45161", "accept": false, - "reason": "ZeRO-3/BERT loading IndexError is unrelated to GPT-OSS tensor-parallel behavior." + "reason": "DeepSpeed ZeRO-3 index error and GPT-OSS tensor-parallel bug are unrelated." }, { "left": "issue:43716", - "right": "issue:45507", + "right": "issue:44292", "accept": false, - "reason": "Different models and failures: Mistral-3 dtype mismatch vs GraniteMoEHybrid invalid method." + "reason": "Different models and failure modes; dtype mismatch vs NVFP4 runtime error." }, { "left": "issue:44292", "right": "issue:45237", "accept": false, - "reason": "Qwen-3-8B-NVFP4 runtime failure and GPT-OSS on AMD GPUs are different issues." + "reason": "Qwen-3 NVFP4 failure and GPT-OSS AMD GPU failure are different issues." }, { "left": "issue:45874", "right": "issue:46294", "accept": false, - "reason": "Gemma4 inputs_embeds reverse-lookup issue is unrelated to Qwen3.5 video chat-template mismatch." + "reason": "Gemma4 reverse-embedding lookup and Qwen3.5 video token mismatch are unrelated." }, { "left": "issue:39290", "right": "issue:45081", "accept": false, - "reason": "Gemma3TextConfig sliding_window_pattern regression is unrelated to the Mistral tokenizer regex crash." + "reason": "Gemma3 config attribute error under vLLM and Mistral regex patch crash are different bugs." }, { "left": "issue:45800", "right": "issue:46515", "accept": true, - "reason": "Both describe the same torch 2.4.1 compatibility regression; the second title is just a more generic restatement." + "reason": "Both report the same torch 2.4.1 incompatibility; 45800 just names transformers 5.8.0 explicitly." }, { "left": "issue:45823", "right": "issue:45874", "accept": false, - "reason": "Gemma4 device mismatch in forward is unrelated to Gemma4 reverse embedding lookup cost." + "reason": "Gemma4 device mismatch and Gemma4 reverse-embedding lookup are different forward-path problems." }, { "left": "issue:34928", "right": "issue:41720", "accept": false, - "reason": "FSDP activation-checkpointing mismatch and Qwen3 auto-device-map CUDA assert are different bugs." + "reason": "FSDP checkpoint recompute mismatch is unrelated to Qwen3 auto-device-map CUDA assert." }, { - "left": "issue:43827", - "right": "issue:44512", + "left": "issue:43425", + "right": "issue:44292", "accept": false, - "reason": "Both are docs regressions, but one is about removed pipeline tasks and the other about the removed transformers run command." + "reason": "Generic torch incompatibility and Qwen-3 NVFP4 failure are not the same bug." }, { "left": "issue:43828", - "right": "issue:45685", - "accept": false, - "reason": "Phi-tiny-MoE autocast dtype mismatch is not the same as the MPS histogram int limitation." - }, - { - "left": "issue:43039", - "right": "issue:44918", + "right": "issue:45161", "accept": false, - "reason": "Liger Kernel cross_entropy being called and Qwen3.5 embedding unpacking under TRL are different issues." + "reason": "Phi autocast dtype mismatch and GPT-OSS TP-only failure are different issues." }, { - "left": "issue:30277", - "right": "issue:36331", + "left": "issue:43541", + "right": "issue:43716", "accept": false, - "reason": "Distributed collective mismatch and trainer compute_loss kwarg mismatch are unrelated." + "reason": "Mixtral grouped_mm tracing error is unrelated to Mistral-3 preprocessing dtype mismatch." }, { - "left": "issue:34928", - "right": "issue:36010", + "left": "issue:43828", + "right": "issue:44292", "accept": false, - "reason": "FSDP checkpointing mismatch is unrelated to the GenerationMixin import error." + "reason": "Different models and failure modes: autocast dtype mismatch vs NVFP4 runtime failure." }, { - "left": "issue:45237", - "right": "issue:46015", + "left": "issue:43716", + "right": "issue:45685", "accept": false, - "reason": "GPT-OSS GPU failure and outdated RAG docs are unrelated." + "reason": "Mistral-3 dtype mismatch and MPS histogram limitation are unrelated." }, { - "left": "issue:43039", - "right": "issue:44805", + "left": "issue:45507", + "right": "issue:45685", "accept": false, - "reason": "Liger Kernel cross_entropy invocation and generic mask shape mismatch are not the same bug." + "reason": "GraniteMoEHybrid invalid-method bug and MPS histogram limitation are unrelated." } ] }, "evaluator_result": { "accept": true, - "feedback": "The cluster summary is grounded in the packet, and the conservative duplicate call on 43596/43638 is well supported by the titles and descriptions. The 45800/46515 pair is presented as only plausible, which is appropriately cautious. I did not see any soft-edge verdicts that overstate the evidence." + "feedback": "The summary is grounded in the packet and stays conservative. The three accepted duplicate families are supported by the titles, and the rejected soft edges are mostly cautious/non-overstated. I don\u2019t see an evidence stretch that would justify rejection." }, "evaluator_used": true, "retried": false, @@ -572,292 +578,292 @@ "issue:43572|issue:45070", "issue:45800|issue:45823" ], - "prepared_review_unit_hash": "96fa9a6bbbd16c2a6aa5c003ddd9c7dc41d3589cf26fc3cbb34154d5a659ef7e", + "prepared_review_unit_hash": "cd8bdc78eca6880ffc2eb8bb39920fcfc8a659b0cebe9a4f89b04adaedf4098b", "component_budget": { "node_count": 514, "item_count": 514, "soft_pair_count": 1891, - "serialized_chars": 980926, - "estimated_input_tokens": 245232, - "estimated_eval_tokens": 490720 + "serialized_chars": 980895, + "estimated_input_tokens": 245224, + "estimated_eval_tokens": 490704 }, "budget": { "node_count": 48, "item_count": 48, "soft_pair_count": 42, - "serialized_chars": 40047, - "estimated_input_tokens": 10012, - "estimated_eval_tokens": 20280 + "serialized_chars": 40038, + "estimated_input_tokens": 10010, + "estimated_eval_tokens": 20276 }, "overflow_policy": "truncate_then_skip", "trimmed": false, "aggressively_trimmed": false, "split": true, "analyst_result": { - "summary": "This cluster is mostly a mix of unrelated Transformers regressions and model-loading/runtime compatibility bugs. The only clear duplicate-style pairs are the ZeRO-3 loading IndexError reports and the v5 AutoTokenizer token-ID regression reports.", - "confidence": 0.72, - "canonical_issue_reason": "issue:45020 is the best umbrella candidate: it is the broadest, most central regression report and naturally covers several nearby v5 / model-loading breakages, especially remote-code compatibility.", + "summary": "This cluster is mostly a grab-bag of unrelated model-loading/runtime regressions. The only clear duplicate pair is the AutoTokenizer v5 regression across multiple model families; most other soft pairs are merely similar symptom classes and should stay separate.", + "confidence": 0.62, + "canonical_issue_reason": "Issue 45920 is the broader, more representative report: it names multiple affected model families and frames the problem as a cross-model v5 AutoTokenizer regression, whereas 45812 is Granite-specific.", "canonical_pr_reason": null, - "best_issue_reason": "issue:45020 is the strongest global representative because it captures a broad, user-facing Transformers regression pattern rather than a single model-specific symptom.", + "best_issue_reason": "45920 is the best issue to anchor the cluster because it is the most general and likely to capture the underlying tokenizer regression with the widest impact.", "best_pr_reason": null, "soft_edge_verdicts": [ { "left": "issue:45800", "right": "issue:46304", "accept": false, - "reason": "Both are version-compatibility complaints, but they describe different failure modes and environments, not the same concrete bug." + "reason": "Both are version-compatibility complaints, but one is a torch/transformers incompatibility and the other is a specific USE_DISTRIBUTED=0 breakage; not the same concrete bug." }, { "left": "issue:43296", "right": "issue:46304", "accept": false, - "reason": "vLLM/Gemma3 loading failure vs PyTorch USE_DISTRIBUTED=0 breakage are different code paths and not the same underlying issue." + "reason": "Different failures in different stacks: PaddleOCR-VL loading in vLLM vs distributed support regression in PyTorch support." }, { "left": "issue:43866", "right": "issue:46402", "accept": false, - "reason": "Different models and different loading failures; one is a corrupted checkpoint report, the other is missing text-encoder weights." + "reason": "Both are model-load failures, but one is a corrupted checkpoint claim and the other is missing text-encoder weights; different underlying problems." }, { "left": "issue:41762", "right": "issue:43596", - "accept": true, - "reason": "Both report the same ZeRO-3 loading IndexError pattern: index 0 out of bounds with size 0 during model initialization/loading." + "accept": false, + "reason": "Both mention Zero-3 index errors, but they involve different models and loading paths; too little evidence they are the same bug." }, { "left": "issue:39290", "right": "issue:41720", "accept": false, - "reason": "Gemma3 sliding_window_pattern error with vLLM and Qwen3 auto device-map cuda assert are unrelated failures." + "reason": "Gemma3 config attribute error in vLLM and Qwen3 auto device-map cuda assert are unrelated runtime failures." }, { "left": "issue:43901", "right": "issue:44512", "accept": false, - "reason": "Both are documentation-related, but they concern different docs changes and not the same underlying bug or change." + "reason": "Both are docs regressions, but they concern different deprecated docs behaviors and are not the same issue." }, { "left": "issue:44589", "right": "issue:46304", "accept": false, - "reason": "Float8 storage lookup failure is a separate runtime issue from the distributed-build compatibility problem." + "reason": "Float8 storage lookup failure is unrelated to the PyTorch distributed support regression." }, { "left": "issue:45237", "right": "issue:45901", "accept": false, - "reason": "AMD GPU execution failure and table-question-answering crash are unrelated." + "reason": "AMD GPU execution failure and table-question-answering crash are different code paths and symptoms." }, { "left": "issue:43039", "right": "issue:44560", "accept": false, - "reason": "Liger Kernel cross_entropy behavior and Qwen3-vl video StopIteration are different subsystems and symptoms." + "reason": "Liger Kernel cross-entropy invocation and Qwen3-vl video StopIteration are unrelated issues." }, { "left": "issue:45823", "right": "issue:46294", "accept": false, - "reason": "Both are model-specific forward/template failures, but they are not the same bug or even the same modality." + "reason": "Both involve device/input handling, but one is a forward device mismatch and the other is video token/feature mismatch in chat templating; not the same bug." }, { "left": "issue:43975", "right": "issue:46015", "accept": false, - "reason": "Incorrect detokenization and outdated RAG examples are unrelated." + "reason": "Tokenizer detokenization bug and outdated RAG docs are unrelated." }, { "left": "issue:44841", "right": "issue:45823", "accept": false, - "reason": "Processor failure for Voxtral and Gemma4 PLE device mismatch are different issues." + "reason": "Voxtral processor failure and Gemma4 device mismatch do not look like the same underlying problem." }, { "left": "issue:43866", "right": "issue:46772", "accept": false, - "reason": "Checkpoint corruption is not the same as NVFP4/ModelOpt compatibility." + "reason": "Checkpoint corruption and NVFP4/model-opt compatibility are different failure modes." }, { "left": "issue:44863", "right": "issue:46015", "accept": false, - "reason": "NemotronH checkpoint loading and outdated RAG docs are unrelated." + "reason": "NemotronH checkpoint-loading issue and RAG docs are unrelated." }, { "left": "issue:45020", "right": "issue:46304", "accept": false, - "reason": "Remote-code breakage and USE_DISTRIBUTED=0 compatibility are both broad regressions, but they are different concrete bugs." + "reason": "Remote-code model breakage and PyTorch distributed support regression affect different mechanisms." }, { "left": "issue:45812", "right": "issue:45920", "accept": true, - "reason": "Same v5 AutoTokenizer regression family: wrong token IDs across multiple models, with the same underlying tokenizer code path." + "reason": "Same underlying tokenizer regression: both report wrong token IDs in v5 across Granite/other models, and one is clearly the broader umbrella report." }, { "left": "issue:43872", "right": "issue:46304", "accept": false, - "reason": "bitsandbytes Int8Params incompatibility is a different dependency issue from the PyTorch distributed support breakage." + "reason": "bitsandbytes Int8Params constructor error is unrelated to the torch distributed compatibility issue." }, { "left": "issue:44292", "right": "issue:46015", "accept": false, - "reason": "Qwen-3-8B-NVFP4 runtime failure and RAG docs are unrelated." + "reason": "NVFP4 inference failure and outdated RAG docs are unrelated." }, { "left": "issue:43866", "right": "issue:44863", "accept": false, - "reason": "A corrupted checkpoint report and a model implementation not loading checkpoints are not the same bug." + "reason": "Both are loading problems, but they concern different model families and different failure descriptions." }, { "left": "issue:39290", "right": "issue:43531", "accept": false, - "reason": "Both mention sliding window / Qwen3-related behavior, but one is a Gemma3/vLLM config attribute error and the other is a Qwen3-MoE sliding_window issue." + "reason": "Both mention sliding-window behavior, but the concrete failures differ too much: Gemma3 config attribute error vs Qwen3-MoE sliding_window issue." }, { "left": "issue:41093", "right": "issue:43638", "accept": false, - "reason": "Both are IndexError reports, but one is a mask-length mismatch and the other is a size-0 access during ZeRO-3 loading; not the same failure." + "reason": "Shape mismatch in attention mask indexing is not the same as the Zero-3 empty-tensor index error." }, { "left": "issue:43792", "right": "issue:45901", "accept": false, - "reason": "Whisper loading/runtime failure and table-question-answering crash are unrelated." + "reason": "Whisper loading/running failure and table QA crash are unrelated." }, { "left": "issue:43854", "right": "issue:43866", "accept": false, - "reason": "GLM-4.7 unit-test loading failure and Ovis2 checkpoint corruption are different model-specific problems." + "reason": "Unit-test loading failure for GLM-4.7-Flash and a corrupted Ovis2 checkpoint are different issues." }, { "left": "issue:43425", "right": "issue:45070", "accept": false, - "reason": "Torch 2.10 incompatibility and a pydantic PretrainedConfig regression are separate compatibility issues." + "reason": "Torch version incompatibility and a PretrainedConfig/pydantic regression are not the same concrete bug." }, { "left": "issue:43975", "right": "issue:45901", "accept": false, - "reason": "Incorrect detokenization and table QA crash do not share the same underlying bug." + "reason": "DeepSeek detokenization bug and table-question-answering crash are unrelated." }, { "left": "issue:43792", "right": "issue:45070", "accept": false, - "reason": "Whisper loading failure and the PretrainedConfig pydantic regression are unrelated." + "reason": "Whisper inference failure and pydantic config regression do not share the same code path." }, { "left": "issue:45362", "right": "issue:45464", "accept": false, - "reason": "Both are Qwen3.5 chat-related, but one is a crash in transformers chat and the other is a streaming inference API failure; not the same bug." + "reason": "Both involve Qwen3.5 chat, but one is a generic chat crash and the other is a streaming chat/completions API failure; not clearly the same bug." }, { "left": "issue:43596", "right": "issue:45800", "accept": false, - "reason": "They are both loading/compatibility failures, but the reported errors and environments are different." + "reason": "Zero-3 Bert loading failure and torch/transformers version incompatibility are different classes of problems." }, { "left": "issue:43541", "right": "issue:46869", "accept": false, - "reason": "Grouped_mm tracing errors and YOLO multi-GPU device-map issues are different runtime paths." + "reason": "Mixtral grouped_mm tracing failure and YOLO multi-GPU device-map failure are unrelated." }, { "left": "issue:44292", "right": "issue:45901", "accept": false, - "reason": "NVFP4 model execution and table QA crash are unrelated." + "reason": "NVFP4 model execution failure and table QA crash are unrelated." }, { "left": "issue:39290", "right": "issue:45823", "accept": false, - "reason": "Gemma3/vLLM config attribute error and Gemma4 device mismatch are different model/version problems." + "reason": "Gemma3 config/sliding-window failure and Gemma4 device mismatch are different bugs despite both being model-specific runtime issues." }, { "left": "issue:45464", "right": "issue:45685", "accept": false, - "reason": "Qwen3.5 streaming inference failure and MPS histogram not implemented are unrelated." + "reason": "Chat/completions streaming failure and MPS histogram unsupported-for-Int error are unrelated." }, { "left": "issue:36010", "right": "issue:39290", "accept": false, - "reason": "Importing GenerationMixin and Gemma3TextConfig sliding_window_pattern are different API breakages." + "reason": "GenerationMixin import error and Gemma3 config attribute error are unrelated." }, { "left": "issue:43541", "right": "issue:45685", "accept": false, - "reason": "Torch dynamo grouped_mm tracing error and MPS histogram implementation error are unrelated." + "reason": "Grouped_mm tracing failure and MPS histogram type error do not look like the same bug." }, { "left": "issue:43541", "right": "issue:43828", "accept": false, - "reason": "Grouped_mm tracing error and autocast dtype mismatch are different model execution bugs." + "reason": "Grouped_mm tracing failure and autocast dtype mismatch are different runtime issues." }, { "left": "issue:44841", "right": "issue:45084", "accept": false, - "reason": "Voxtral processor failure and non-template-node compilation error do not share the same code path." + "reason": "Voxtral processor failure and template compilation error are unrelated." }, { "left": "issue:45507", "right": "issue:46869", "accept": false, - "reason": "GraniteMoEHybrid invalid method call and YOLO multi-GPU auto device-map failure are not the same bug." + "reason": "GraniteMoEHybrid invalid method call and YOLO multi-GPU device-map failure are different model-specific bugs." }, { "left": "issue:43404", "right": "issue:43866", "accept": false, - "reason": "Weight-tying bug in Mistral3 and Ovis2 checkpoint corruption are different issues." + "reason": "Mistral3 lm_head tying bug and Ovis2 checkpoint corruption are unrelated model issues." }, { "left": "issue:41093", "right": "issue:41762", "accept": false, - "reason": "Both are IndexError reports, but one is a mask shape mismatch and the other is ZeRO-3 size-0 loading; not the same underlying defect." + "reason": "Both are tensor/indexing failures, but the exact symptoms and contexts differ enough that they do not look like the same bug." }, { "left": "issue:45399", "right": "issue:46661", "accept": false, - "reason": "Fallback blocking for flash-attn2 and remote-code model_class.config_class errors are different compatibility problems." + "reason": "Flash-attn2 fallback logic and remote-code config_class attribute errors are separate code paths." }, { "left": "issue:43572", "right": "issue:45070", "accept": false, - "reason": "Missing pad_token_idx and a pydantic PretrainedConfig regression are unrelated." + "reason": "Missing pad_token_idx in StableLmConfig and PretrainedConfig pydantic regression are unrelated configuration issues." }, { "left": "issue:45800", "right": "issue:45823", "accept": false, - "reason": "Torch/PyTorch compatibility and Gemma4 device-map mismatch are different issues." + "reason": "Torch/transformers compatibility and Gemma4 device mismatch are different failures." } ] }, "evaluator_result": { "accept": true, - "feedback": "Grounded and conservative overall. The umbrella choice of issue:45020 is defensible as the broadest regression report in the set, and the summary does not overclaim duplication. The two accepted soft-edge pairs (41762/43596 and 45812/45920) are reasonable same-bug families, while the other pair verdicts stay appropriately cautious." + "feedback": "The cluster summary is grounded in the packet and stays conservative. The accepted pair 45812/45920 is well supported by the titles: both describe the same AutoTokenizer v5 token-ID regression, with 45920 clearly the broader umbrella report. The other soft-edge rejections are appropriately cautious given only symptom-level similarity. No mergeability issues apply here since this is an issue-clustering judgment, not a PR-pair merge decision." }, "evaluator_used": true, "retried": false, @@ -932,8 +938,8 @@ "issue:43541|issue:45161", "issue:43828|issue:45464", "issue:43366|issue:45464", - "issue:44387|issue:45005", "issue:44291|issue:45070", + "issue:44387|issue:45005", "issue:43531|issue:45070", "issue:41093|issue:43596", "issue:43716|issue:46294", @@ -941,8 +947,8 @@ "issue:44479|issue:46294", "issue:43054|issue:43994", "issue:46468|issue:46869", - "issue:45759|issue:45902", "issue:43596|issue:46165", + "issue:45759|issue:45902", "issue:46640|issue:46661", "issue:43716|issue:44918", "issue:39186|issue:44918", @@ -959,229 +965,229 @@ "issue:45183|issue:46463", "issue:44863|issue:45005" ], - "prepared_review_unit_hash": "27f5f4b226568e380405d2ed4b68bdb2e621c10dcaabf3fa3cf231bed1d27cf5", + "prepared_review_unit_hash": "529d0e998713d73792452d9ca998fe2d0065aa9d10920213375f62ece4522669", "component_budget": { "node_count": 514, "item_count": 514, "soft_pair_count": 1891, - "serialized_chars": 980926, - "estimated_input_tokens": 245232, - "estimated_eval_tokens": 490720 + "serialized_chars": 980895, + "estimated_input_tokens": 245224, + "estimated_eval_tokens": 490704 }, "budget": { "node_count": 47, "item_count": 47, "soft_pair_count": 31, - "serialized_chars": 36246, - "estimated_input_tokens": 9062, - "estimated_eval_tokens": 18380 + "serialized_chars": 36244, + "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 issue set, with one clear validation/error-message cluster and one additional soft duplicate pair around the same mask/tensor shape-mismatch IndexError title pattern. No PRs are present.", - "confidence": 0.81, - "canonical_issue_reason": "issue:45183 is the best canonical issue for the validation/error-message subcluster because it states the underlying XOR input-validation problem broadly, and issue:46463 is a concrete T5-family manifestation of that bug.", + "summary": "The cluster is mostly heterogeneous: several unrelated bug reports across model families, quantization, multimodal inputs, remote code, and validation. A few soft edges look like plausible duplicates, but there is no single global duplicate group or canonical issue/PR for the whole cluster.", + "confidence": 0.79, + "canonical_issue_reason": null, "canonical_pr_reason": null, - "best_issue_reason": "issue:45183 is the strongest representative issue in the packet because it describes the root validation bug at a higher level than the narrower model-specific reports.", - "best_pr_reason": null, + "best_issue_reason": "No clear global best issue; the items split into multiple unrelated bug families, so picking one as canonical would be misleading.", + "best_pr_reason": "No PRs are present in this cluster.", "soft_edge_verdicts": [ { "left": "issue:43296", "right": "issue:43366", "accept": false, - "reason": "Different problems: PaddleOCR-VL loading/version incompatibility vs GGUF gpt-oss architecture support." + "reason": "Different failures: PaddleOCR-VL load error vs GGUF gpt-oss architecture support." }, { "left": "issue:44805", "right": "issue:44918", "accept": false, - "reason": "They both involve IndexError-style failures, but the titles point to different concrete contexts and do not justify merging as the same bug." + "reason": "Different code paths: mask/tensor shape mismatch vs Qwen3.5 embedding unpacking with SFT trainer." }, { "left": "issue:43541", "right": "issue:45161", "accept": false, - "reason": "Both mention MoE, but one is a torch-dynamo grouped_mm tracing failure and the other is tensor-parallel support; not the same bug." + "reason": "Different model/runtime problems: torch dynamo grouped_mm tracing vs TP support for GPT-OSS MoE." }, { "left": "issue:43828", "right": "issue:45464", "accept": false, - "reason": "Autocast dtype mismatch and streaming chat/completions inference failure are different failure modes." + "reason": "One is an autocast dtype mismatch in Phi-tiny-MoE; the other is a Qwen3.5 streaming chat/completions failure." }, { "left": "issue:43366", "right": "issue:45464", "accept": false, - "reason": "Unsupported GGUF architecture support is unrelated to Qwen3.5 streaming inference." + "reason": "GGUF gpt-oss support and Qwen3.5 streaming inference are unrelated issues." }, { - "left": "issue:44387", - "right": "issue:45005", + "left": "issue:44291", + "right": "issue:45070", "accept": false, - "reason": "Int4 quantization OOM is not the same as tied-weights behavior on translation models." + "reason": "Different breakages: init_empty_weights argument handling vs PretrainedConfig/pydantic field changes." }, { - "left": "issue:44291", - "right": "issue:45070", + "left": "issue:44387", + "right": "issue:45005", "accept": false, - "reason": "init_empty_weights argument handling and pydantic PretrainedConfig field regression are separate issues." + "reason": "Quantization OOM under v5.x is not the same bug as tied-weights handling on translation models." }, { "left": "issue:43531", "right": "issue:45070", "accept": false, - "reason": "Sliding-window behavior in Qwen3-MoE is unrelated to the pydantic model field breakage." + "reason": "Qwen3-MoE sliding_window behavior is unrelated to the PretrainedConfig pydantic regression." }, { "left": "issue:41093", "right": "issue:43596", "accept": false, - "reason": "They are both shape/index errors, but the available titles only suggest different model/setup contexts; not enough to treat as the same bug." + "reason": "Both are index errors, but they affect different model/setup paths and do not look like the same bug." }, { "left": "issue:43716", "right": "issue:46294", "accept": false, - "reason": "Image-preprocessor dtype mismatch is not the same as video-feature/token mismatch in chat templating." + "reason": "Different multimodal mismatches: image preprocessor dtype vs video feature/token mismatch." }, { "left": "issue:45020", "right": "issue:45800", "accept": false, - "reason": "Remote-code model breakage and torch/transformers version incompatibility are different failure modes." + "reason": "Remote-code model breakage and torch/transformers version incompatibility are distinct issues." }, { "left": "issue:44479", "right": "issue:46294", - "accept": false, - "reason": "Related video-input area, but the broader v5.3.0 regression and the specific apply_chat_template token mismatch are not clearly the same concrete bug." + "accept": true, + "reason": "Both target the Qwen3.5/Qwen-VL video-input path and describe the same kind of video token/feature regression." }, { "left": "issue:43054", "right": "issue:43994", - "accept": false, - "reason": "Both involve siglip2 quality, but one is degraded text embeddings and the other is bad AutoModel/pipeline outputs." + "accept": true, + "reason": "Same Siglip2 family with closely related bad-output symptoms; both point to the model/pipeline producing degraded or nonsensical embeddings." }, { "left": "issue:46468", "right": "issue:46869", "accept": false, - "reason": "Official Nemotron example failure is unrelated to YOLO multi-GPU device-map issues." + "reason": "Official Nemotron example breakage is unrelated to YOLO multi-GPU device-map failure." }, { - "left": "issue:45759", - "right": "issue:45902", + "left": "issue:43596", + "right": "issue:46165", "accept": false, - "reason": "Composite config unwrapping and uninitialized-memory rotary embedding issues are different bugs." + "reason": "Deepspeed ZeRO3/BertModel index failure is unrelated to resuming remote-code training." }, { - "left": "issue:43596", - "right": "issue:46165", + "left": "issue:45759", + "right": "issue:45902", "accept": false, - "reason": "Different root causes: ZeRO-3 initialization index error vs resume/trust_remote_code loading issue." + "reason": "Composite multimodal config unwrapping and uninitialized rotary embedding memory are different bugs." }, { "left": "issue:46640", "right": "issue:46661", "accept": false, - "reason": "Flash Attention 2 integration and remote-code config_class attribute errors are unrelated." + "reason": "Flash Attention 2 availability issue is unrelated to remote-code model_class/config_class loading." }, { "left": "issue:43716", "right": "issue:44918", "accept": false, - "reason": "Different modalities and code paths: image preprocessing dtype mismatch vs SFT trainer embedding unpacking." + "reason": "Image preprocessor dtype mismatch and SFT embedding unpacking are different multimodal/training issues." }, { "left": "issue:39186", "right": "issue:44918", "accept": false, - "reason": "FSDP weight-shape runtime error is unrelated to the Qwen3.5/TRl embedding unpacking failure." + "reason": "FSDP weight-shape RuntimeError is unrelated to Qwen3.5 embedding unpacking." }, { "left": "issue:43645", "right": "issue:45399", "accept": false, - "reason": "Custom-model notebook initialization and flash-attn fallback gating are separate regressions." + "reason": "Jupyter custom-model initialization and flash-attn2 fallback gating are not the same bug." }, { "left": "issue:40990", "right": "issue:44960", "accept": false, - "reason": "High perplexity on gpt-oss and a GLM5 issue do not look like the same bug." + "reason": "High perplexity on gpt-oss-20b and GLM5 are unrelated model-quality reports." }, { "left": "issue:45632", "right": "issue:45698", - "accept": false, - "reason": "Both touch remote code/custom modules, but one is a cache-path collision and the other is wrong-module loading after save_pretrained." + "accept": true, + "reason": "Both concern remote-code/custom-module resolution picking the wrong cached or loaded module for local/saved models." }, { "left": "issue:42915", "right": "issue:45399", "accept": false, - "reason": "FineGrainedFP8Config failure and flash-attn fallback blocking are unrelated." + "reason": "Qwen3-MoE FP8 config failure is unrelated to flash-attn2 fallback checks." }, { "left": "issue:43404", "right": "issue:45005", "accept": false, - "reason": "Tied lm_head weights in Mistral3 is not the same as translation-model tied-weights issues in v5." + "reason": "Mistral3 tied lm_head weights and translation-model tied-weights regressions are different model families and failure modes." }, { "left": "issue:44360", "right": "issue:44512", "accept": false, - "reason": "A DSA indexer bug and a documentation typo are clearly unrelated." + "reason": "DSA indexer behavior and docs cleanup are unrelated." }, { "left": "issue:43873", "right": "issue:45005", "accept": false, - "reason": "Quantization/offloading memory behavior is unrelated to tied weights on translation models." + "reason": "Quantization offloading and tied-weights handling do not describe the same underlying bug." }, { "left": "issue:44492", "right": "issue:44512", "accept": false, - "reason": "Cache strategy typo and removed run-command docs are different documentation issues." + "reason": "A cache-strategy typo is not the same as outdated CLI docs." }, { "left": "issue:44530", "right": "issue:45005", "accept": false, - "reason": "PagedAttentionCache linear_attention crash is unrelated to tied-weights translation model regressions." + "reason": "PagedAttentionCache linear_attention crash is unrelated to tied-weights regressions." }, { "left": "issue:44918", "right": "issue:45507", "accept": false, - "reason": "Qwen3.5 embedding unpacking and GraniteMoEHybrid invalid-method calls are different model bugs." + "reason": "Qwen3.5 embedding unpacking and GraniteMoEHybrid invalid method calls are different code paths." }, { "left": "issue:45183", "right": "issue:46463", - "accept": true, - "reason": "Same validation/error-message bug: 46463 is a T5-family-specific manifestation of the broader XOR input validation problem." + "accept": false, + "reason": "Broader input_ids/inputs_embeds XOR validation issue is not the same as T5 decoder-input error messaging." }, { "left": "issue:44863", "right": "issue:45005", "accept": false, - "reason": "NemotronH checkpoint loading is unrelated to translation-model tied weights." + "reason": "NemotronH checkpoint-loading failure is unrelated to tied-weights issues in translation models." } ] }, "evaluator_result": { - "accept": false, - "feedback": "Mostly grounded, but at least one soft-edge verdict is too strong for the evidence available. The rejection of issue:41093 vs issue:44805 is not well supported by the packet: their titles are near-identical mask/tensor shape-mismatch IndexErrors, so saying they are definitely different code paths/symptoms overstates what can be inferred from titles alone. Because the evaluator requires conservative soft-edge calls, I can\u2019t accept this as-is." + "accept": true, + "feedback": "The summary is grounded in the packet and stays conservative about cluster heterogeneity. The accepted soft edges are plausible duplicates and do not overstate certainty, and there are no PR pairs to assess." }, "evaluator_used": true, - "retried": true, + "retried": false, "accepted_nontrivial_soft_edge": true, "error_kind": null, "error_message": null @@ -1273,27 +1279,27 @@ "issue:43975|issue:45520", "issue:44403|issue:44488", "issue:43643|issue:45020", - "issue:42947|issue:43856", "issue:41762|issue:42915", + "issue:42947|issue:43856", "issue:44991|issue:45399", "issue:43824|issue:43872", "issue:45685|issue:45823", "issue:43782|issue:44960" ], - "prepared_review_unit_hash": "f192c0583c79d9d96e90120289545070d24f02502ce4690f9369c69a177a904e", + "prepared_review_unit_hash": "5bdf928a82e2c4909bb620bdb7ff250696276a1886fdaa829b47c1327a2b9093", "component_budget": { "node_count": 514, "item_count": 514, "soft_pair_count": 1891, - "serialized_chars": 980926, - "estimated_input_tokens": 245232, - "estimated_eval_tokens": 490720 + "serialized_chars": 980895, + "estimated_input_tokens": 245224, + "estimated_eval_tokens": 490704 }, "budget": { "node_count": 48, "item_count": 48, "soft_pair_count": 30, - "serialized_chars": 36904, + "serialized_chars": 36903, "estimated_input_tokens": 9226, "estimated_eval_tokens": 18708 }, @@ -1302,198 +1308,198 @@ "aggressively_trimmed": false, "split": true, "analyst_result": { - "summary": "This cluster is mostly a grab-bag of unrelated Transformers issues, but it does contain a small tight duplicate set around the v5 non-persistent buffer regression and another around StaticCache.get_seq_length(). Most other soft pairs are only broadly related by model family or subsystem and should not be merged.", - "confidence": 0.8, - "canonical_issue_reason": "Issue #43950 is the clearest representative of the non-persistent buffer corruption regression and has the strongest near-duplicate neighborhood (#43644, #44534, #45902).", + "summary": "This cluster is mostly a collection of unrelated Transformers bug reports. I only found three true duplicate groups: non-persistent buffer corruption on load, the StaticCache.get_seq_length() return-type bug, and the Gemma4 quantized vision/audio dtype-cast bug. The rest are only loosely similar by subsystem or model family, not the same underlying issue.", + "confidence": 0.87, + "canonical_issue_reason": "issue:43950 is the clearest and most complete report of the non-persistent-buffer corruption regression, and it has the strongest duplicate evidence from closely matching follow-up reports.", "canonical_pr_reason": null, - "best_issue_reason": "#43950 best captures the core bug theme in this cluster and is the most useful anchor for the duplicate subcluster.", + "best_issue_reason": "issue:43950 is the best overall issue representative here because it precisely describes a concrete regression and is the center of one of the true duplicate groups.", "best_pr_reason": null, "soft_edge_verdicts": [ { "left": "issue:43975", "right": "issue:44292", "accept": false, - "reason": "Different models and different failures; tokenizer bug vs Qwen NVFP4 load/runtime error." + "reason": "Both are error reports, but they describe different failures on different models/code paths (DeepSeek tokenizer detokenization vs Qwen NVFP4 loading)." }, { "left": "issue:46489", "right": "issue:46491", "accept": false, - "reason": "Both are tokenizer regressions, but they affect different models and describe different symptoms." + "reason": "Both are tokenizer regressions, but the symptoms and affected models differ; no clear evidence they come from the same concrete bug." }, { "left": "issue:39692", "right": "issue:43054", "accept": false, - "reason": "Same SigLIP2 family, but one is a docs/example failure and the other is a quality/discrepancy report." + "reason": "One is a broken docs example/model-processor mismatch and quantization failure; the other is degraded SigLIP2 text embeddings. Different problems." }, { "left": "issue:43950", "right": "issue:44534", "accept": true, - "reason": "Same underlying v5 regression: non-persistent buffers are corrupted/fill with junk during load." + "reason": "Same underlying regression: non-persistent buffers are being mishandled during from_pretrained/load in Transformers v5." }, { "left": "issue:36331", "right": "issue:41762", "accept": false, - "reason": "CustomTrainer compute_loss kwarg mismatch vs Gemma3 DeepSpeed ZeRO-3 loading error; unrelated bugs." + "reason": "Unrelated failures: a Trainer custom loss signature change vs a DeepSpeed ZeRO-3 model-loading index error." }, { "left": "issue:43644", "right": "issue:43950", "accept": true, - "reason": "Both report the same non-persistent buffer corruption regression in Transformers 5.x." + "reason": "Both report the same non-persistent-buffer corruption bug in Transformers 5.x." }, { "left": "issue:43872", "right": "issue:44589", "accept": false, - "reason": "Different quantization/storage incompatibilities; not the same concrete bug." + "reason": "Both are runtime/type errors, but one is bitsandbytes initialization incompatibility and the other is a float8 storage lookup issue." }, { "left": "issue:46899", "right": "issue:46932", "accept": true, - "reason": "Same Gemma4 quantized vision/audio embedder dtype-cast bug; #46932 is the fix-side duplicate." + "reason": "These describe the same Gemma4 quantized vision/audio embedder dtype-casting bug; 46932 explicitly says it is the same as another report." }, { "left": "issue:43828", "right": "issue:44928", "accept": false, - "reason": "Different Qwen3.5 failures: autocast dtype mismatch vs flash-attn crash from 3D position_ids handling." + "reason": "Different Qwen/MoE/autocast issues with different symptoms and code paths; not the same bug." }, { "left": "issue:43054", "right": "issue:45701", "accept": false, - "reason": "One is a model-quality complaint, the other is a broad tokenization-version regression; too different." + "reason": "One is a model quality discrepancy, the other a tokenizer-version behavior change. Too broad to merge as duplicates." }, { "left": "issue:45920", "right": "issue:46489", "accept": false, - "reason": "Broad tokenizer regression vs a specific DeepSeek-Coder symptom; not enough evidence of one concrete bug." + "reason": "Both concern tokenizer regressions in v5, but they affect different models and report different failure modes." }, { "left": "issue:38175", "right": "issue:43054", "accept": false, - "reason": "Both mention SigLIP2, but the reported failures are different enough to treat separately." + "reason": "Both mention SigLIP2, but one is zero probabilities and the other is worse text embeddings; not the same concrete defect." }, { "left": "issue:45987", "right": "issue:46858", "accept": true, - "reason": "Same StaticCache.get_seq_length contract bug; one report is shape-(1,) tensor, the other device-tensor arithmetic breakage." + "reason": "Same cache API bug: get_seq_length() returns a tensor instead of a plain int, causing downstream static-cache arithmetic issues." }, { "left": "issue:43950", "right": "issue:45902", - "accept": true, - "reason": "Same non-persistent buffer/materialization regression; #45902 is a model-specific manifestation." + "accept": false, + "reason": "One is about buffer corruption on load; the other is uninitialized memory after meta/to_empty materialization. Related area, different bug." }, { "left": "issue:43054", "right": "issue:43493", "accept": false, - "reason": "Same model family, but one is a quality complaint and the other is a HF-vs-JAX implementation discrepancy." + "reason": "Both involve SigLIP2, but one is poor embedding quality and the other is an implementation discrepancy vs JAX. Not the same issue." }, { "left": "issue:44910", "right": "issue:45859", "accept": false, - "reason": "Different code paths: flash-attn packed-sequence crash vs CPU offload incompatibility in rotary embedding." + "reason": "Flash-attention packed-sequence misinterpretation vs CPU-offload incompatibility in rotary embedding are distinct problems." }, { "left": "issue:42947", "right": "issue:44387", "accept": false, - "reason": "Different problems: gradient checkpointing ineffective vs int4 quantization memory/OOM regression." + "reason": "Gradient checkpointing/LoRA behavior and int4 quantization memory OOM are separate training issues." }, { "left": "issue:45663", "right": "issue:45902", "accept": false, - "reason": "Different model and failure mode; Gemma4 FSDP2 key error is unrelated to Qwen3_5Moe buffer init." + "reason": "Different Gemma/Qwen bugs: FSDP2 shared_kv_states rebuilding vs uninitialized inv_freq after materialization." }, { "left": "issue:44403", "right": "issue:44451", "accept": false, - "reason": "Generic load noise vs a specific model load failure; not the same bug." + "reason": "One is a noisy load message, the other is a hard model-loading failure. No evidence of the same root cause." }, { "left": "issue:41553", "right": "issue:43883", "accept": false, - "reason": "Bad AutoTokenizer error message vs missing all_tied_weights_keys attribute; unrelated." + "reason": "AutoTokenizer error handling for Voxtral and Molmo weight-tying metadata are unrelated." }, { "left": "issue:24643", "right": "issue:28530", "accept": false, - "reason": "DeepSpeed training weight-shape error vs early-stopping metric lookup; unrelated issues." + "reason": "DeepSpeed training weight-shape error and early-stopping metric selection are unrelated." }, { "left": "issue:43975", "right": "issue:45520", "accept": false, - "reason": "Tokenizer detokenization bug vs import_utils KeyError on Python 3.13; different layers entirely." + "reason": "DeepSeek detokenization regression and Python 3.13 flash_attn import error are unrelated." }, { "left": "issue:44403", "right": "issue:44488", "accept": false, - "reason": "A generic loading-noise report vs a specific model loading failure; not the same underlying bug." + "reason": "Both mention model loading, but there is no concrete evidence they share one bug beyond broad similarity." }, { "left": "issue:43643", "right": "issue:45020", "accept": false, - "reason": "Both involve remote_code, but #45020 is too broad to confirm the same concrete bug from the title alone." + "reason": "45020 is a broad remote_code regression bucket; 43643 is a specific missing-fields symptom. Too broad to call duplicates confidently." }, { - "left": "issue:42947", - "right": "issue:43856", + "left": "issue:41762", + "right": "issue:42915", "accept": false, - "reason": "LoRA checkpointing issue vs Qwen3 MoE memory issue; same broad training area, different bug." + "reason": "Gemma3 ZeRO-3 loading error and Qwen3 MoE FP8 failure are different models and failure modes." }, { - "left": "issue:41762", - "right": "issue:42915", + "left": "issue:42947", + "right": "issue:43856", "accept": false, - "reason": "Different models and exceptions; Gemma3 ZeRO-3 load error vs Qwen3 MoE FP8 failure." + "reason": "LoRA checkpointing inefficiency and Qwen3 MoE memory usage are separate performance issues." }, { "left": "issue:44991", "right": "issue:45399", "accept": false, - "reason": "Tokenizer loading failure vs flash-attn fallback gating; unrelated." + "reason": "Tokenizer loading failure and flash-attn2 fallback checks are unrelated subsystems." }, { "left": "issue:43824", "right": "issue:43872", "accept": false, - "reason": "ImportError for a missing model class vs bitsandbytes Int8Params kwarg incompatibility." + "reason": "Different incompatibilities: missing Qwen2_5_VL export vs bitsandbytes Int8Params keyword mismatch." }, { "left": "issue:45685", "right": "issue:45823", "accept": false, - "reason": "MPS histogram int error vs Gemma4 PLE device mismatch; not the same bug." + "reason": "MPS histogram/int issue in MoE and Gemma4 device-mapping mismatch are unrelated." }, { "left": "issue:43782", "right": "issue:44960", "accept": false, - "reason": "Qwen3VL weight_only load error vs GLM5 issue; unrelated." + "reason": "Different model families and different failures; no evidence of the same underlying bug." } ] }, "evaluator_result": { "accept": true, - "feedback": "Grounded and conservative overall. The canonical issue choice (#43950) is supported by the issue titles and the nearby duplicate-like pairs around non-persistent buffers. The summary appropriately labels most of the cluster as unrelated and does not overclaim. The accepted soft-edge pairs are all plausibly the same concrete bug or a fix-side duplicate, and the rejections are phrased cautiously." + "feedback": "The summary is grounded in the packet and the true-duplicate calls are conservative. The three accepted groups are well-supported by the titles/details: the non-persistent-buffer regression around `from_pretrained`, the `StaticCache.get_seq_length()` return-type bug, and the Gemma4 quantized vision/audio dtype-cast bug. The remaining pairs are appropriately treated as similar-but-distinct rather than duplicates." }, "evaluator_used": true, "retried": false, @@ -1568,8 +1574,8 @@ "issue:45020|issue:45823", "issue:44661|issue:45020", "issue:43828|issue:45823", - "issue:45020|issue:45356", "issue:42491|issue:44368", + "issue:45020|issue:45356", "issue:43881|issue:43883", "issue:45715|issue:46315", "issue:43872|issue:46869", @@ -1582,30 +1588,30 @@ "issue:46894|issue:46897", "issue:43792|issue:44220", "issue:43257|issue:45305", - "issue:42947|issue:45923", "issue:36331|issue:41093", + "issue:42947|issue:45923", "issue:43940|issue:45313", "issue:43854|issue:44863", - "issue:44928|issue:45588", "issue:24643|issue:30064", "issue:44479|issue:44560", + "issue:44928|issue:45588", "issue:41093|issue:44805", "issue:41553|issue:44991" ], - "prepared_review_unit_hash": "d563786ca38cd85aa8b56ba5c9c1a598dfd4c884210a10a3bb99be33ad052d3e", + "prepared_review_unit_hash": "361d85e7f0e4234114f8ded4236e537081af363a6722b16aa037c7e0af9d96d8", "component_budget": { "node_count": 514, "item_count": 514, "soft_pair_count": 1891, - "serialized_chars": 980926, - "estimated_input_tokens": 245232, - "estimated_eval_tokens": 490720 + "serialized_chars": 980895, + "estimated_input_tokens": 245224, + "estimated_eval_tokens": 490704 }, "budget": { "node_count": 47, "item_count": 47, "soft_pair_count": 28, - "serialized_chars": 36179, + "serialized_chars": 36178, "estimated_input_tokens": 9045, "estimated_eval_tokens": 18346 }, @@ -1614,186 +1620,186 @@ "aggressively_trimmed": false, "split": true, "analyst_result": { - "summary": "This cluster is mostly a set of unrelated issues, with a few clear duplicate-like pairs around docs cleanup, DeepSpeed/Qwen3 weight loading, video-input regressions, and a training-loss double-shift bug.", - "confidence": 0.56, - "canonical_issue_reason": "Issue 45020 is the broadest, most representative bug report in the set: a recent-version regression affecting remote_code model loading, which is a central infrastructure-style issue rather than a narrow model-specific symptom.", + "summary": "Most pairs are distinct model- or subsystem-specific regressions, but a few are clear duplicates: the v5 docs pipeline reference issue, the `_is_hf_initialized` loading regression, the DeepSeek tokenizer regression, and the training-loss double-shift reports. The broad remote-code regression report is the best umbrella issue for the overall cluster.", + "confidence": 0.74, + "canonical_issue_reason": "Issue 45020 is the broadest, most central report of the recurring v5 compatibility regressions around model/tokenizer loading, so it best represents the cluster.", "canonical_pr_reason": null, - "best_issue_reason": "Issue 45020 is the best global issue candidate because it captures the widest user-facing regression scope and is more general than the other model- or docs-specific reports.", + "best_issue_reason": "45020 is the strongest umbrella issue for the shared regression theme; the other issues are narrower manifestations or adjacent edge cases.", "best_pr_reason": null, "soft_edge_verdicts": [ { "left": "issue:45081", "right": "issue:45399", "accept": false, - "reason": "Both touch attention-related code, but one is a tokenizer regex crash and the other is a flash-attn fallback gating issue; different failures and code paths." + "reason": "Different bugs: a Mistral regex patch crash vs. a flash-attn fallback gating problem." }, { "left": "issue:41553", "right": "issue:43881", "accept": false, - "reason": "Different models and different symptoms: bad AutoTokenizer error messaging for Voxtral vs a GLM-4v loading failure." + "reason": "Both are loading failures, but they concern different models and different failure modes." }, { "left": "issue:45020", "right": "issue:45823", "accept": false, - "reason": "Remote-code loading regressions and device-map forward mismatches are different bugs." + "reason": "Remote-code loading regression vs. Gemma4 device-map mismatch are separate code paths." }, { "left": "issue:44661", "right": "issue:45020", "accept": false, - "reason": "Tokenizer-mapping edge case vs broad remote_code regression; related area, but not the same concrete bug." + "reason": "Tokenizer-mapping edge case vs. broad remote-code regression; not the same bug." }, { "left": "issue:43828", "right": "issue:45823", "accept": false, - "reason": "Both are runtime mismatches, but dtype autocast failure and device placement failure are distinct code paths." + "reason": "Different runtime failures: dtype mismatch under autocast vs. device placement mismatch." }, { - "left": "issue:45020", - "right": "issue:45356", + "left": "issue:42491", + "right": "issue:44368", "accept": false, - "reason": "General remote_code breakage and a specific Kimi tokenizer codec regression are not the same issue." + "reason": "Different LoRA/tie-weights behaviors; one is a compatibility break, the other is a warning about config flags." }, { - "left": "issue:42491", - "right": "issue:44368", + "left": "issue:45020", + "right": "issue:45356", "accept": false, - "reason": "Both involve Qwen3/LoRA, but one is a loading incompatibility and the other is a warning about tie_word_embeddings; not the same bug." + "reason": "Remote-code model loading breakage and a tokenizer codec regression are unrelated." }, { "left": "issue:43881", "right": "issue:43883", "accept": false, - "reason": "Different model classes and different missing attributes; no shared concrete failure." + "reason": "Different model-loading failures with different missing pieces in the implementation." }, { "left": "issue:45715", "right": "issue:46315", "accept": false, - "reason": "Both are tokenizer performance complaints, but one is convert_ids_to_tokens and the other is convert_tokens_to_ids; not the same bug." + "reason": "Both are tokenizer performance regressions, but they hit different methods and likely different hot paths." }, { "left": "issue:43872", "right": "issue:46869", "accept": false, - "reason": "bitsandbytes constructor incompatibility vs YOLO multi-GPU device-map failure are unrelated." + "reason": "Unexpected kwarg in parameter initialization vs. YOLO multi-GPU device-map failure are unrelated." }, { "left": "issue:45759", "right": "issue:46661", "accept": false, - "reason": "Composite-config unwrapping issue vs remote-code model_class/config_class AttributeError are different problems." + "reason": "Different AutoModel API problems: config unwrapping vs. missing config_class on remote-code models." }, { "left": "issue:43872", "right": "issue:44291", - "accept": false, - "reason": "Both mention _is_hf_initialized, but one is bitsandbytes Int8Params incompatibility and the other is init_empty_weights model loading; not the same concrete path." + "accept": true, + "reason": "Same `_is_hf_initialized` keyword-argument regression during weight initialization/loading; 43872 is a bitsandbytes manifestation." }, { "left": "issue:43827", "right": "issue:44509", "accept": true, - "reason": "Both report docs still referencing removed pipeline tasks after v5; same docs cleanup regression." + "reason": "Same docs regression: stale v5 pipeline-task references remain in the documentation." }, { "left": "issue:45488", "right": "issue:46489", - "accept": false, - "reason": "Both concern tokenizer behavior in DeepSeek-family models, but one is a tokenizer pre-tokenizer override bug and the other is an output regression from a different fix gap." + "accept": true, + "reason": "Same tokenizer regression in v5 around preserving tokenizer.json pre-tokenizers; both break DeepSeek-family tokenization." }, { "left": "issue:45020", "right": "issue:46661", "accept": false, - "reason": "Both involve remote-code models, but one is a broad version regression and the other is a specific missing config_class attribute." + "reason": "Remote-code regressions and missing config_class are related only broadly, not the same concrete bug." }, { "left": "issue:43404", "right": "issue:43950", "accept": false, - "reason": "lm_head tying and non-persistent buffer corruption are unrelated model-state bugs." + "reason": "Both are tied to weight handling, but one is a tied-weight bug and the other is a non-persistent buffer regression." }, { "left": "issue:46894", "right": "issue:46897", "accept": true, - "reason": "These are the same training-loss double-shift bug pattern reported on different model families; likely one underlying loss-shift issue." + "reason": "Same training-loss double-shift bug pattern; only the affected model family differs." }, { "left": "issue:43792", "right": "issue:44220", "accept": false, - "reason": "Whisper failing to run and a specific fbank feature-extraction issue may be related, but the reports do not identify the same concrete bug clearly enough." + "reason": "Whisper runtime failure and _torch_extract_fbank_features are too generic to treat as the same concrete bug." }, { "left": "issue:43257", "right": "issue:45305", "accept": false, - "reason": "Both mention DeepSpeed and Qwen3, but one is missing weight conversion and the other is incorrect gradient averaging; distinct failures." + "reason": "DeepSpeed weight-conversion failure vs. gradient averaging bug are different training issues." }, { - "left": "issue:42947", - "right": "issue:45923", + "left": "issue:36331", + "right": "issue:41093", "accept": false, - "reason": "Gradient checkpointing ineffective in PEFT LoRA and a missing supports_gradient_checkpointing flag are different bugs." + "reason": "Unexpected compute_loss kwarg and mask shape mismatch are unrelated failures." }, { - "left": "issue:36331", - "right": "issue:41093", + "left": "issue:42947", + "right": "issue:45923", "accept": false, - "reason": "CustomTrainer signature mismatch and mask/tensor shape mismatch are unrelated." + "reason": "Both involve gradient checkpointing, but one is ineffective LoRA checkpointing and the other is a missing capability flag." }, { "left": "issue:43940", "right": "issue:45313", - "accept": true, - "reason": "Both are DeepSpeed ZeRO-3 weight-loading failures on Qwen3-family models; same underlying loading path appears broken." + "accept": false, + "reason": "Both are ZeRO-3 loading failures, but the model/module specifics differ and the bugs do not clearly collapse to one fix." }, { "left": "issue:43854", "right": "issue:44863", "accept": false, - "reason": "Different model-loading failures for unrelated architectures." - }, - { - "left": "issue:44928", - "right": "issue:45588", - "accept": false, - "reason": "RLHF gradient explosion from position_ids/SDPA behavior is unrelated to a flash-attention AttributeError." + "reason": "Different model-loading failures for different architectures." }, { "left": "issue:24643", "right": "issue:30064", "accept": false, - "reason": "DeepSpeed training weight-shape error and segmentation-map preprocessing failure are unrelated." + "reason": "DeepSpeed training error and segmentation-map processing are unrelated." }, { "left": "issue:44479", "right": "issue:44560", - "accept": true, - "reason": "Both describe video-input regressions in the Qwen3/Qwen3.5 visual model family in v5.3.x, pointing to the same video-processing path." + "accept": false, + "reason": "Both are video-input regressions, but the affected models and failure modes differ." + }, + { + "left": "issue:44928", + "right": "issue:45588", + "accept": false, + "reason": "Gradient-explosion RLHF bug vs. flash-attention integration crash are separate problems." }, { "left": "issue:41093", "right": "issue:44805", "accept": false, - "reason": "Same error text pattern, but they are separate reports and no shared concrete bug is established from the titles alone." + "reason": "Same error shape family, but not enough evidence they are the same masking bug." }, { "left": "issue:41553", "right": "issue:44991", "accept": false, - "reason": "Bad AutoTokenizer messaging for Voxtral and tokenizer-loading failure for est-roberta are different model-specific issues." + "reason": "Tokenizer loading failures for different models; the error origins are different." } ] }, "evaluator_result": { "accept": true, - "feedback": "The summary is grounded in the packet, and the soft-edge verdicts are mostly conservative. The accepted pairs are reasonably supported by the titles and descriptions (docs cleanup, Qwen3-family ZeRO-3 loading failures, training-loss double-shift, and video regressions), while the rejections avoid overclaiming equivalence across distinct failure modes." + "feedback": "The cluster summary is grounded in the packet, and the accepted soft-edge pairs are all conservative duplicates/near-duplicates based on the issue titles and descriptions. The rejected pairs are also framed cautiously as different bugs or only loosely related. The choice of 45020 as the umbrella issue is defensible as a broad remote-code regression anchor." }, "evaluator_used": true, "retried": false, @@ -1866,8 +1872,8 @@ "soft_pairs": [ "issue:43716|issue:46468", "issue:44530|issue:44863", - "issue:44863|issue:46402", "issue:43366|issue:43828", + "issue:44863|issue:46402", "issue:43366|issue:43531", "issue:43596|issue:46515", "issue:45325|issue:45902", @@ -1885,8 +1891,8 @@ "issue:44560|issue:45685", "issue:44910|issue:46858", "issue:36010|issue:42915", - "issue:43550|issue:44910", "issue:44560|issue:44918", + "issue:43550|issue:44910", "issue:44683|issue:46515", "issue:45812|issue:46489", "issue:45823|issue:46515", @@ -1894,25 +1900,25 @@ "issue:44560|issue:44805", "issue:38175|issue:43994", "issue:44877|issue:45030", - "issue:45672|issue:46899", "issue:43638|issue:44661", "issue:43596|issue:44661", + "issue:45672|issue:46899", "issue:42491|issue:43720" ], - "prepared_review_unit_hash": "60d26409202691979d72d0ab26fc52f17b8f660bbf5c8b6531630c2a0040d1e4", + "prepared_review_unit_hash": "f141435a4385435e2a62d615d06d759ac6f0ad30f2d0ae0f340c9528e060e128", "component_budget": { "node_count": 514, "item_count": 514, "soft_pair_count": 1891, - "serialized_chars": 980926, - "estimated_input_tokens": 245232, - "estimated_eval_tokens": 490720 + "serialized_chars": 980895, + "estimated_input_tokens": 245224, + "estimated_eval_tokens": 490704 }, "budget": { "node_count": 48, "item_count": 48, "soft_pair_count": 34, - "serialized_chars": 38045, + "serialized_chars": 38047, "estimated_input_tokens": 9512, "estimated_eval_tokens": 19280 }, @@ -1921,222 +1927,222 @@ "aggressively_trimmed": false, "split": true, "analyst_result": { - "summary": "This cluster is mostly a grab-bag of distinct Transformers bug reports. I only accepted a few soft edges that look like true duplicate reports of the same underlying regression; the rest are too model- or codepath-specific to merge safely.", - "confidence": 0.66, - "canonical_issue_reason": "issue:43646 is the broadest representative of the strongest duplicate family here: it describes the general custom model initialization regression, while issue:43645 is a notebook-specific variant of the same breakage.", + "summary": "The set is mostly unrelated model/regression reports with two clear duplicate-like pairs: custom model initialization in Transformers 5.0 and bad SigLIP2 inference output.", + "confidence": 0.84, + "canonical_issue_reason": "issue:43646 is the best canonical issue because it states the shared Transformers 5.0 custom-model initialization regression in the broadest, cleanest form; issue:43645 is just a notebook-specific variant.", "canonical_pr_reason": null, - "best_issue_reason": "issue:43646 is the best overall issue anchor because it is the most general and reusable statement of the custom-model-init regression, making it the clearest canonical representative among the mixed reports.", + "best_issue_reason": "issue:43646 is the strongest representative of the cluster: concise, broad, and clearly about the underlying regression rather than an environment-specific symptom.", "best_pr_reason": null, "soft_edge_verdicts": [ { "left": "issue:43716", "right": "issue:46468", "accept": false, - "reason": "Different models and failures: Mistral-3 image-preprocessor dtype mismatch vs Nemotron example breakage." + "reason": "Different bugs and code paths: image preprocessor/model dtype mismatch vs a broken Nemotron example." }, { "left": "issue:44530", "right": "issue:44863", "accept": false, - "reason": "Both are loading-related, but one is a Qwen3.5 cache/attention crash and the other is NemotronH checkpoint loading." + "reason": "Both are loading/runtime failures, but they affect different models and mechanisms (PagedAttentionCache vs NemotronH checkpoint loading)." }, { - "left": "issue:44863", - "right": "issue:46402", + "left": "issue:43366", + "right": "issue:43828", "accept": false, - "reason": "Different model families and symptoms: NemotronH checkpoint loading vs Sam3Model text-encoder weights missing." + "reason": "Unrelated symptoms: GGUF architecture support vs autocast dtype mismatch on Phi-tiny-MoE." }, { - "left": "issue:43366", - "right": "issue:43828", + "left": "issue:44863", + "right": "issue:46402", "accept": false, - "reason": "Unrelated bugs: GGUF gpt-oss support vs Phi-tiny-MoE autocast dtype mismatch." + "reason": "Different model-loading failures for different models; no shared concrete bug." }, { "left": "issue:43366", "right": "issue:43531", "accept": false, - "reason": "Different requests: GGUF architecture support vs a Qwen3-MoE sliding_window behavior issue." + "reason": "Same general Qwen/GGUF space, but different underlying issues: model architecture support vs sliding-window behavior." }, { "left": "issue:43596", "right": "issue:46515", "accept": false, - "reason": "IndexError in BertModel under ZeRO-3 is unrelated to a generic torch 2.4.1 incompatibility report." + "reason": "The first is a specific Bert/deepspeed zero3 IndexError; the second is a broad torch version incompatibility report." }, { "left": "issue:45325", "right": "issue:45902", "accept": false, - "reason": "Different Qwen variants and different bugs: rope-index scaling vs uninitialized rotary frequency after meta materialization." + "reason": "Different mechanisms: Qwen2.5-VL rope scaling for still images vs Qwen3_5Moe rotary embedding materialization bug." }, { "left": "issue:41553", "right": "issue:44488", "accept": false, - "reason": "Both are tokenizer/loading complaints, but they target different models and different failure modes." + "reason": "Both are load failures, but for different tokenizer/model targets and likely different root causes." }, { "left": "issue:43366", "right": "issue:44589", "accept": false, - "reason": "No shared codepath: GGUF support request vs missing Float8 storage object error." + "reason": "GGUF architecture support and a Float8 storage lookup error are unrelated." }, { "left": "issue:43638", "right": "issue:46515", "accept": false, - "reason": "ZeRO-3/BertModel IndexError is not the same issue as a torch version incompatibility report." + "reason": "A Bert/deepspeed zero3 index error is not the same as a generic torch compatibility issue." }, { "left": "issue:44155", "right": "issue:45584", "accept": false, - "reason": "Different subsystems: AudioFlamingo3 batched inference leakage vs Whisper empty-transcription generation failure." + "reason": "Different product areas and failures: AudioFlamingo3 batching leak vs Whisper empty-transcription generation." }, { "left": "issue:45663", "right": "issue:45823", "accept": false, - "reason": "Same model family, but different codepaths: FSDP2 shared_kv_state rebuilding vs device_map auto device mismatch." + "reason": "Same model family, but training FSDP2 KeyError and inference device-map mismatch are different bugs." }, { "left": "issue:42915", "right": "issue:43957", "accept": false, - "reason": "Qwen3-MoE FP8 config failure is unrelated to meta-device loading regressions." + "reason": "Qwen3Moe FP8 failure and meta-device loading regression are different code-path problems." }, { "left": "issue:43824", "right": "issue:43957", "accept": false, - "reason": "ImportError for Qwen2.5-VL class is a different problem from meta-device loading breakage." + "reason": "Qwen2.5-VL import error is unrelated to the generic meta-device loading bug." }, { "left": "issue:43716", "right": "issue:44560", "accept": false, - "reason": "Different models and failures: Mistral-3 dtype mismatch vs Qwen3-vl-embedding StopIteration." + "reason": "Mistral-3 dtype mismatch and Qwen3-vl video StopIteration are different failures." }, { "left": "issue:43950", "right": "issue:45834", "accept": false, - "reason": "Non-persistent buffer corruption from from_pretrained is unrelated to Kosmos2.5 long-ocr indexing." + "reason": "Strict config validation for granite_speech and Kosmos2.5 OCR index error are unrelated." }, { "left": "issue:42617", "right": "issue:43366", "accept": false, - "reason": "3d_parallel.py execution failure and GGUF gpt-oss support are unrelated issues." + "reason": "Not the same bug: running 3d_parallel.py vs GGUF gpt-oss architecture support." }, { "left": "issue:43645", "right": "issue:43646", "accept": true, - "reason": "Same underlying regression: custom model initialization broke in Transformers 5.0.0; 43645 is the Jupyter-notebook-specific manifestation of 43646." + "reason": "Same underlying Transformers 5.0 custom-model initialization regression; the notebook wording is just a narrower symptom." }, { "left": "issue:44560", "right": "issue:45685", "accept": false, - "reason": "Different models and causes: Qwen3-vl StopIteration vs MPS histogram not implemented for Int." + "reason": "Qwen3-vl video StopIteration and an MPS histogram-not-implemented error are unrelated." }, { "left": "issue:44910", "right": "issue:46858", "accept": false, - "reason": "Both involve sequence/cache handling, but one is a flash-attention packed-sequence misread and the other is a device-tensor slice-arithmetic bug." + "reason": "Both involve cache/attention, but one is a flash-attention memory access bug and the other is a static-cache seq_length arithmetic issue." }, { "left": "issue:36010", "right": "issue:42915", "accept": false, - "reason": "ImportError for GenerationMixin and Qwen3-MoE FP8 failures are unrelated." + "reason": "ImportError for GenerationMixin and Qwen3Moe FP8 failure are unrelated." }, { - "left": "issue:43550", - "right": "issue:44910", + "left": "issue:44560", + "right": "issue:44918", "accept": false, - "reason": "Different models and failure modes: Bamba-9B-v2 torch.compile/SDPA vs Qwen3.5 flash-attention position_ids handling." + "reason": "Different symptoms and code paths: video StopIteration vs TRL SFT unpacking failure." }, { - "left": "issue:44560", - "right": "issue:44918", + "left": "issue:43550", + "right": "issue:44910", "accept": false, - "reason": "Both touch Qwen3.5/VL-ish flows, but one is a video StopIteration error and the other is SFT unpacking embeddings." + "reason": "Bamba torch.compile+SDPA and Qwen3.5 flash-attention memory access are different model/runtime bugs." }, { "left": "issue:44683", "right": "issue:46515", "accept": false, - "reason": "Compiled flex_attention on torch >= 2.9 is a separate compatibility bug from a torch 2.4.1 incompatibility report." + "reason": "Compiled flex_attention on torch>=2.9 is not the same as a broad torch 2.4.1 incompatibility report." }, { "left": "issue:45812", "right": "issue:46489", "accept": false, - "reason": "Different tokenizer regressions affecting different model families: Granite vs DeepSeek-Coder." + "reason": "Both are tokenizer regressions, but they affect different model families and likely different tokenizer code paths." }, { "left": "issue:45823", "right": "issue:46515", "accept": false, - "reason": "Gemma4 device mismatch under device_map=auto is unrelated to a generic torch version incompatibility." + "reason": "Gemma4 device-map mismatch and a general torch version incompatibility are not the same issue." }, { "left": "issue:44661", "right": "issue:45823", "accept": false, - "reason": "Tokenizer-mapping/new-model registration failure is unrelated to Gemma4 PLE device placement." + "reason": "Tokenizer-mapping failure and Gemma4 device mismatch are unrelated." }, { "left": "issue:44560", "right": "issue:44805", "accept": false, - "reason": "Different model families and distinct shape errors; no shared underlying bug." + "reason": "Qwen3-vl video StopIteration does not match the generic mask/tensor shape IndexError." }, { "left": "issue:38175", "right": "issue:43994", "accept": true, - "reason": "Same SigLIP2 regression on the same model: one reports zero probabilities, the other nonsensical outputs via AutoModel/pipeline." + "reason": "Same model and same symptom family: incorrect SigLIP2 inference outputs, with one phrased as zero probabilities and the other as nonsensical pipeline/AutoModel results." }, { "left": "issue:44877", "right": "issue:45030", "accept": false, - "reason": "Strict config validation for granite_speech and glm4v config validation are different config-schema issues." - }, - { - "left": "issue:45672", - "right": "issue:46899", - "accept": true, - "reason": "Same Gemma4 4-bit multimodal quantization bug family; 46899 broadens the unsafe dtype-cast problem first reported in 45672." + "reason": "Different config-validation problems affecting different model types." }, { "left": "issue:43638", "right": "issue:44661", "accept": false, - "reason": "ZeRO-3/BertModel initialization crash is unrelated to the TOKENIZER_MAPPING_NAMES add-new-model-like failure." + "reason": "Bert/deepspeed zero3 index error and tokenizer-mapping changes are unrelated." }, { "left": "issue:43596", "right": "issue:44661", "accept": false, - "reason": "Different root causes: BertModel ZeRO-3 IndexError vs tokenizer-mapping registration logic." + "reason": "A zero3/Bert IndexError is not the same as the add-new-model-like tokenizer mapping bug." + }, + { + "left": "issue:45672", + "right": "issue:46899", + "accept": false, + "reason": "Both are Gemma4 quantization-related, but they describe different dtype/weight-loading failures." }, { "left": "issue:42491", "right": "issue:43720", "accept": false, - "reason": "Both mention Qwen3-MoE/loading, but one is a LoRA compatibility issue and the other is BitNet packed-weight loading during accelerate init." + "reason": "LoRA/Qwen3_Moe HF4.x vs HF5.x incompatibility and BitNet packed-weight loading are different problems." } ] }, "evaluator_result": { "accept": true, - "feedback": "The clustering is conservative and grounded in the packet. The canonical issue choice of issue:43646 is well-supported by the title/body overlap with issue:43645, and the accepted soft edges (38175\u219443994, 43645\u219443646, 45672\u219446899) look like plausible same-bug variants rather than overreaches. The rejected pairs are also explained in a model- and symptom-specific way without overstating equivalence." + "feedback": "Grounded overall: the canonical issue choice is supported by the titles, and the soft-edge decisions stay conservative. The two accepted pairs are plausibly duplicate-like from the packet, while the rest are reasonably rejected as different model/code-path failures. No mergeability issue for PR pairs since none are present." }, "evaluator_used": true, "retried": false, @@ -2222,15 +2228,15 @@ "issue:45313|issue:45357", "issue:45507|issue:46468", "issue:43644|issue:43645", - "issue:39692|issue:43994", "issue:43257|issue:43854", + "issue:39692|issue:43994", "issue:44488|issue:44991", "issue:43856|issue:43873", "issue:42491|issue:45292", "issue:43854|issue:43931", - "issue:43856|issue:44387", "issue:42491|issue:45335", "issue:43708|issue:44075", + "issue:43856|issue:44387", "issue:44960|issue:46050", "issue:44164|issue:44291", "issue:46897|issue:46901", @@ -2238,20 +2244,20 @@ "issue:45488|issue:45812", "issue:45362|issue:45406" ], - "prepared_review_unit_hash": "c0ac081f089111e3c6e12bb404a58d820d1e689543b62f34d86eb82f28baf209", + "prepared_review_unit_hash": "6e4b0f4ad9cf603261136deb4743bffe64af41f4146ace0a5f9715010cab4d90", "component_budget": { "node_count": 514, "item_count": 514, "soft_pair_count": 1891, - "serialized_chars": 980926, - "estimated_input_tokens": 245232, - "estimated_eval_tokens": 490720 + "serialized_chars": 980895, + "estimated_input_tokens": 245224, + "estimated_eval_tokens": 490704 }, "budget": { "node_count": 48, "item_count": 48, "soft_pair_count": 30, - "serialized_chars": 36521, + "serialized_chars": 36522, "estimated_input_tokens": 9131, "estimated_eval_tokens": 18518 }, @@ -2260,198 +2266,198 @@ "aggressively_trimmed": false, "split": true, "analyst_result": { - "summary": "This set is mostly unrelated issues with a few true duplicate pairs. The clearest duplicates are the Transformers v5 non-persistent-buffer regression and the SigLIP2 model/processor regression; most other soft matches are only broad symptom similarity and should stay separate.", - "confidence": 0.78, - "canonical_issue_reason": "issue:43644 is the cleanest canonical anchor: it describes a concrete Transformers v5 regression directly, and issue:44534 reads like a near-verbatim repeat of the same bug.", + "summary": "This cluster is mostly heterogeneous; only one pair looks like a true duplicate. The clearest canonical issue is the generic Transformers v5 non-persistent-buffer regression.", + "confidence": 0.43, + "canonical_issue_reason": "issue:43644 is the broader, root-cause phrased report and exactly matches issue:44534\u2019s symptom; it is the better anchor than the more symptom-focused duplicate.", "canonical_pr_reason": null, - "best_issue_reason": "issue:43644 is the best representative issue overall because it is specific, actionable, and has the strongest duplicate evidence among the candidates.", + "best_issue_reason": "issue:43644 is the best representative issue because it describes the underlying v5 regression in a general way and already aligns with the only clear duplicate pair.", "best_pr_reason": null, "soft_edge_verdicts": [ { "left": "issue:43632", "right": "issue:45834", "accept": false, - "reason": "Different subsystems and failures: `_is_hf_initialized` regression vs Kosmos2.5 OCR index error." + "reason": "Different failures in unrelated model paths; same version mention is not enough to treat them as one bug." }, { "left": "issue:44661", "right": "issue:44805", "accept": false, - "reason": "Unrelated bugs: tokenizer-mapping registration failure vs a mask/tensor shape mismatch." + "reason": "One is a tokenizer-mapping edge case, the other is a mask/tensor shape error; no shared concrete code path is evident." }, { "left": "issue:43317", "right": "issue:44756", "accept": false, - "reason": "Both are memory/loading problems, but one is dequantized offload loading and the other is mmap on Strix Halo OOM." + "reason": "GPU/CPU offload loading failure vs mmap OOM on Strix Halo are different issues and fixes." }, { "left": "issue:38175", "right": "issue:39692", "accept": false, - "reason": "Same model family, but different reported problems: zero probabilities vs doc-example/model-processor mismatch and quantization issues." + "reason": "Both are about SigLIP2, but one is zero probabilities and the other is docs/quantization mistakes; not the same underlying bug." }, { "left": "issue:43644", "right": "issue:44534", "accept": true, - "reason": "Same Transformers v5 non-persistent-buffer corruption regression; the titles are effectively describing the same bug." + "reason": "Same Transformers v5 non-persistent-buffer junk regression; the titles describe the same underlying bug." }, { "left": "issue:41720", "right": "issue:42915", "accept": false, - "reason": "Different Qwen3 failures: auto device mapping assert vs FineGrainedFP8Config behavior." + "reason": "Both involve Qwen3/MoE, but one is device-mapping CUDA assert and the other is FineGrainedFP8 loading failure." }, { "left": "issue:42915", "right": "issue:44756", "accept": false, - "reason": "Unrelated: Qwen3 MoE FP8 failure vs mmap/OOM workaround on a different platform." + "reason": "Qwen3Moe FP8 failure is unrelated to mmap/OOM behavior on Strix Halo." }, { "left": "issue:42491", "right": "issue:45276", "accept": false, - "reason": "Both involve Qwen3/Gemma-style model loading, but the concrete bugs differ: LoRA compatibility vs resize_token_embeddings propagation." + "reason": "LoRA loading on hf4.x vs resize_token_embeddings propagation in Gemma4 are unrelated bugs." }, { "left": "issue:44451", "right": "issue:44488", "accept": false, - "reason": "Both are tokenizer-loading complaints, but they concern different model repos and no shared concrete failure is shown." + "reason": "Both are model loading complaints, but they name different models and do not clearly point to the same code-path defect." }, { "left": "issue:44960", "right": "issue:44977", "accept": false, - "reason": "Different models and failure modes: GLM5 issue vs Qwen3.5 flash-attention generation bug." + "reason": "GLM5 and Qwen3.5 flash-attention failures are separate model-specific problems." }, { "left": "issue:44863", "right": "issue:46846", "accept": false, - "reason": "Different problems: checkpoint loading failure vs missing tensor-parallel plan entries causing OOM/shape errors." + "reason": "NemotronH checkpoint loading vs Qwen3.5 tensor-parallel plan omissions are different issues." }, { "left": "issue:46459", "right": "issue:46512", "accept": false, - "reason": "Both mention dtype, but one is loading composite checkpoints and the other is from_config dtype bookkeeping." + "reason": "Both mention dtype, but one is composite checkpoint loading and the other is from_config bookkeeping; not the same bug." }, { "left": "issue:45313", "right": "issue:45357", "accept": false, - "reason": "Different Qwen3.5 regressions: ZeRO-3 weight loading vs save_pretrained key serialization." + "reason": "DeepSpeed ZeRO-3 weight loading and incorrect save_pretrained visual keys are different failure modes." }, { "left": "issue:45507", "right": "issue:46468", "accept": false, - "reason": "Different model-specific errors with no clear shared code path." + "reason": "Invalid method call in GraniteMoEHybrid is unrelated to the broken Nemotron example." }, { "left": "issue:43644", "right": "issue:43645", "accept": false, - "reason": "Related to Transformers v5 notebook/custom-model behavior, but one is buffer initialization and the other is Jupyter custom-model initialization." - }, - { - "left": "issue:39692", - "right": "issue:43994", - "accept": true, - "reason": "Same SigLIP2 base-patch16-224 regression: both report nonsense outputs from AutoModel/pipeline and point to the same model/processor mismatch." + "reason": "Related Transformers v5 regressions, but one is junk in buffers and the other is notebook custom-model initialization; not the same concrete bug." }, { "left": "issue:43257", "right": "issue:43854", "accept": false, - "reason": "Different models and bugs: Qwen3 MoE conversion with accelerate+deepspeed vs GLM-4.7-Flash loading in tests." + "reason": "Qwen3 MOE weight conversion with accelerate+deepspeed and GLM-4.7 Flash test loading are unrelated." + }, + { + "left": "issue:39692", + "right": "issue:43994", + "accept": false, + "reason": "Both involve SigLIP2, but one is a bad docs example while the other is a runtime wrong-output regression." }, { "left": "issue:44488", "right": "issue:44991", "accept": false, - "reason": "Both are tokenizer-loading failures, but the affected models and concrete breakages are different." + "reason": "Tokenizer loading failures for different models are too vague to merge as one concrete bug." }, { "left": "issue:43856", "right": "issue:43873", "accept": false, - "reason": "Different issues: Qwen3 MoE training memory usage vs quantization offloading behavior." + "reason": "Training-memory inefficiency and quantized offloading behavior are distinct problems." }, { "left": "issue:42491", "right": "issue:45292", "accept": false, - "reason": "Different symptoms and scopes: LoRA compatibility for qwen3_moe vs a resize_token_embeddings propagation bug." + "reason": "LoRA/model-version compatibility is unrelated to resize_token_embeddings not updating output embeddings." }, { "left": "issue:43854", "right": "issue:43931", "accept": false, - "reason": "Different model-loading failures for different model families." - }, - { - "left": "issue:43856", - "right": "issue:44387", - "accept": false, - "reason": "Both are memory-related, but one is training memory efficiency and the other is int4 quantization reserved-memory OOM." + "reason": "Different model load errors: unit-test load failure vs weight shape mismatch." }, { "left": "issue:42491", "right": "issue:45335", "accept": false, - "reason": "Different issues: LoRA/qwen3_moe compatibility vs t5gemma embedding resize propagation." + "reason": "Different model families and different affected embedding surfaces; not one shared bug." }, { "left": "issue:43708", "right": "issue:44075", "accept": false, - "reason": "Trainer checkpoint-resume math is unrelated to SGD optimizer arguments being ignored." + "reason": "Trainer checkpoint-resume math and SGD args not being used are unrelated." + }, + { + "left": "issue:43856", + "right": "issue:44387", + "accept": false, + "reason": "Memory inefficiency during Qwen3 MoE training is not the same as int4 quantization reserved-memory OOM regression." }, { "left": "issue:44960", "right": "issue:46050", "accept": false, - "reason": "Different model and lifecycle bug: GLM5 issue vs quantized save/load breakage." + "reason": "GLM5 and sinq save/load breakage are unrelated model lifecycle issues." }, { "left": "issue:44164", "right": "issue:44291", "accept": false, - "reason": "Both concern loading/saving internals, but extra_state handling and `_is_hf_initialized` argument errors are distinct bugs." + "reason": "Both concern save/load internals, but extra_state handling and unexpected _is_hf_initialized arguments are separate defects." }, { "left": "issue:46897", "right": "issue:46901", "accept": false, - "reason": "Same broad 'double-shift' pattern, but they are model-specific training bugs and not clearly the same code path." + "reason": "Both mention a double-shift loss bug, but they are model-specific reports with no evidence of the same shared code path." }, { "left": "issue:45276", "right": "issue:45335", "accept": false, - "reason": "Similar resize_token_embeddings symptom, but different model implementations and different embedding targets make this too broad to merge." + "reason": "Both mention resize_token_embeddings, but they affect different model implementations and embedding targets; too weak to treat as one bug." }, { "left": "issue:45488", "right": "issue:45812", "accept": false, - "reason": "Both are tokenizer regressions, but they affect different model families and show different concrete failures." + "reason": "Tokenizer-pretokenizer override for DeepSeek models is not the same as wrong token IDs for Granite models." }, { "left": "issue:45362", "right": "issue:45406", "accept": false, - "reason": "Different model/application failures: Qwen3.5 chat crash vs Gemma4 processor attribute error." + "reason": "Qwen3.5 chat crash and Gemma4Processor missing _tokenizer are different failures." } ] }, "evaluator_result": { "accept": true, - "feedback": "The summary is grounded in the packet and the verdicts are conservative overall. The two accepted duplicate pairs are strongly supported by title-level evidence (43644/44534 exact same Transformers v5 buffer bug, and 39692/43994 same SigLIP2 regression family), while the other soft matches are appropriately kept separate despite superficial similarity." + "feedback": "Grounded overall and appropriately conservative. The single accepted pair (issue:43644 / issue:44534) is clearly duplicative from the packet, and the other rejections are generally well-justified by distinct titles/symptoms rather than superficial similarity. No evidence of overstatement in the canonical issue choice." }, "evaluator_used": true, "retried": false, @@ -2522,15 +2528,15 @@ "issue:46829" ], "soft_pairs": [ - "issue:44910|issue:45588", "issue:43425|issue:43957", + "issue:44910|issue:45588", "issue:44402|issue:44488", "issue:39692|issue:44960", "issue:41720|issue:44155", "issue:44291|issue:44756", "issue:45663|issue:45759", - "issue:42947|issue:45704", "issue:43404|issue:43454", + "issue:42947|issue:45704", "issue:43873|issue:44387", "issue:44079|issue:46829", "issue:45663|issue:45859", @@ -2555,118 +2561,118 @@ "issue:42175|issue:43957", "issue:44589|issue:44756" ], - "prepared_review_unit_hash": "ef754ba92cdff1f4d822865b5c4ff97a7b3d19c34344daa888b261aa09fdf325", + "prepared_review_unit_hash": "8225e142a5660d84f6bc4b7818e878d32b97e7988271ae4943710ff94fcb32d6", "component_budget": { "node_count": 514, "item_count": 514, "soft_pair_count": 1891, - "serialized_chars": 980926, - "estimated_input_tokens": 245232, - "estimated_eval_tokens": 490720 + "serialized_chars": 980895, + "estimated_input_tokens": 245224, + "estimated_eval_tokens": 490704 }, "budget": { "node_count": 48, "item_count": 48, "soft_pair_count": 32, - "serialized_chars": 37383, - "estimated_input_tokens": 9346, - "estimated_eval_tokens": 18948 + "serialized_chars": 37385, + "estimated_input_tokens": 9347, + "estimated_eval_tokens": 18950 }, "overflow_policy": "truncate_then_skip", "trimmed": false, "aggressively_trimmed": false, "split": true, "analyst_result": { - "summary": "The cluster is mostly a set of unrelated issues that happen to share broad subsystem keywords. The only strong duplicate pair is the Gemma4 resize_token_embeddings reports; the rest should stay separate.", - "confidence": 0.93, - "canonical_issue_reason": "Issue 45276 is the most complete and specific report for the resize_token_embeddings bug, covering Gemma4 and both embed_tokens_per_layer/output_embeddings propagation failures.", + "summary": "This cluster is mostly a loose mix of unrelated bug reports. The only clear duplicate chain is the resize_token_embeddings regression across generic and model-specific reports; the rest are too model- or subsystem-specific to merge as the same underlying issue.", + "confidence": 0.79, + "canonical_issue_reason": "issue:45292 is the broadest, most model-agnostic description of the shared resize_token_embeddings regression, while issue:45276 and issue:45335 are specific manifestations of the same missing propagation to downstream embeddings.", "canonical_pr_reason": null, - "best_issue_reason": "45276 is the best canonical issue because it most clearly describes the underlying bug and the affected Gemma4-specific code path.", + "best_issue_reason": "issue:45292 is the best umbrella issue because it states the core failure concisely and can absorb the more specific Gemma4 and t5gemma variants as duplicates.", "best_pr_reason": null, "soft_edge_verdicts": [ { - "left": "issue:44910", - "right": "issue:45588", + "left": "issue:43425", + "right": "issue:43957", "accept": false, - "reason": "Both involve flash_attention, but one is a Qwen3.5 position_ids/packed-sequence crash and the other is an s_aux=None AttributeError for sink-less models; different failure modes." + "reason": "Both involve model loading/runtime failures, but one is a Torch 2.10 compatibility complaint and the other is a meta-device loading regression; different code paths and likely different fixes." }, { - "left": "issue:43425", - "right": "issue:43957", + "left": "issue:44910", + "right": "issue:45588", "accept": false, - "reason": "Torch version incompatibility and meta-device loading failures are separate issues with different triggers and code paths." + "reason": "Both touch flash attention, but one is a Qwen3.5 packed-sequence/position_ids bug and the other is an s_aux=None crash in sink-less models; not the same underlying issue." }, { "left": "issue:44402", "right": "issue:44488", "accept": false, - "reason": "Both mention model loading/tokenizer problems, but one is about tokenizer vocab bounds and the other is a specific model load failure; not enough evidence of the same bug." + "reason": "Both mention model loading, but the titles point to different models and likely different tokenizer/vocab problems; too little evidence they share one bug." }, { "left": "issue:39692", "right": "issue:44960", "accept": false, - "reason": "A SigLIP2 documentation example bug and a GLM5 model issue are unrelated." + "reason": "One is a SigLIP2 documentation/example error report and the other is a GLM5 issue; unrelated." }, { "left": "issue:41720", "right": "issue:44155", "accept": false, - "reason": "Qwen3 device-mapping CUDA assert and AudioFlamingo3 batched inference token leakage are unrelated bugs." + "reason": "Qwen3 device-mapping cudaErrorAssert and AudioFlamingo3 batched inference leakage are distinct failures in different models." }, { "left": "issue:44291", "right": "issue:44756", "accept": false, - "reason": "Unexpected init_empty_weights argument handling and mmap-related OOM on Strix Halo are different problems." + "reason": "Unexpected _is_hf_initialized argument handling and Strix Halo mmap OOM are unrelated initialization vs memory-management problems." }, { "left": "issue:45663", "right": "issue:45759", "accept": false, - "reason": "Gemma4 FSDP2 shared_kv_states KeyError and Qwen multimodal text_config unwrapping are distinct implementation issues." + "reason": "Gemma4 FSDP2 shared_kv_states KeyError and Qwen 3.5/3.6 text_config unwrapping are different model/config code paths." }, { - "left": "issue:42947", - "right": "issue:45704", + "left": "issue:43404", + "right": "issue:43454", "accept": false, - "reason": "Gradient checkpointing with LoRA and an apex RMSNorm memory leak are unrelated." + "reason": "Both are about lm_head tying, but they target different model families and different integration paths; not enough to treat as the same concrete bug." }, { - "left": "issue:43404", - "right": "issue:43454", + "left": "issue:42947", + "right": "issue:45704", "accept": false, - "reason": "Both concern lm_head tying, but they affect different model families and there is no clear shared code-path defect." + "reason": "Gradient checkpointing/LoRA and apex.FusedRMSNorm memory leak are unrelated." }, { "left": "issue:43873", "right": "issue:44387", "accept": false, - "reason": "Both are quantization/memory complaints, but one is offloading behavior and the other is increased reserved memory leading to OOM; too broad to merge." + "reason": "Both mention quantization/OOM, but one is offloading behavior and the other is increased reserved memory under int4; too broad to merge." }, { "left": "issue:44079", "right": "issue:46829", "accept": false, - "reason": "These are separate output-object field bugs in different classes; they are not the same underlying issue." + "reason": "ModelOutput key assignment bug and VideoPrism output field mix-up are different manifestations with different affected classes." }, { "left": "issue:45663", "right": "issue:45859", "accept": false, - "reason": "Gemma4 FSDP2 KeyError and Qwen3_5Moe CPU-offload rotary embedding incompatibility are unrelated." + "reason": "Different models and different failures: Gemma4 FSDP2 shared state vs Qwen3_5Moe CPU offload incompatibility." }, { "left": "issue:44292", "right": "issue:44912", "accept": false, - "reason": "Different quantization formats, different models, and different symptoms; not the same bug." + "reason": "Different quantization formats and symptoms: NVFP4 runtime error vs MXFP4 load fallback to bf16." }, { "left": "issue:42947", "right": "issue:43381", "accept": false, - "reason": "One is LoRA checkpointing inefficiency, the other is checkpointing disallowed in eval mode; different problems." + "reason": "Both involve gradient checkpointing, but one is ineffective with PEFT LoRA and the other is an eval-mode restriction; not the same bug." }, { "left": "issue:36010", @@ -2678,109 +2684,109 @@ "left": "issue:43957", "right": "issue:44756", "accept": false, - "reason": "Meta-device model loading breakage and disabling mmap for OOM are not the same underlying issue." + "reason": "Meta-device loading regressions and Strix Halo mmap OOM are different issues." }, { "left": "issue:44155", "right": "issue:45823", "accept": false, - "reason": "AudioFlamingo3 batched inference leakage and Gemma4 device-map mismatch are different bugs." + "reason": "AudioFlamingo3 batch leakage and Gemma4 device_map mismatch are different model-specific forward-path bugs." }, { "left": "issue:44387", "right": "issue:44756", "accept": false, - "reason": "Both involve memory, but int4 quantization OOM and Strix Halo mmap OOM are different root causes." + "reason": "Both are memory-related, but one is int4 reserved-memory growth and the other is a platform-specific mmap OOM workaround; not the same root cause." }, { "left": "issue:43335", "right": "issue:45663", "accept": false, - "reason": "Sparse-layer creation bug in SwitchTransformers and Gemma4 FSDP2 attention state rebuild are unrelated." + "reason": "SwitchTransformersConfig sparse-layer construction and Gemma4 FSDP2 shared_kv_states are unrelated." }, { "left": "issue:44863", "right": "issue:46468", "accept": false, - "reason": "NemotronH checkpoint loading and a broken official Nemotron example may be related by model family, but the reports describe different failures and are not clearly the same bug." + "reason": "NemotronH checkpoint loading and the official Nemotron example being broken are related in theme but not clearly the same underlying bug." }, { "left": "issue:45276", "right": "issue:45292", "accept": true, - "reason": "Both report the same resize_token_embeddings propagation bug; 45276 is the more detailed Gemma4-specific version and 45292 is the generic duplicate." + "reason": "These describe the same resize_token_embeddings regression; 45276 is a model-specific instance and 45292 states the generic bug." }, { "left": "issue:44479", "right": "issue:44877", "accept": false, - "reason": "Video-input regression across Qwen models and strict Granite Speech config loading are unrelated." + "reason": "Video-input regression for Qwen VL models and strict granite_speech config loading are different problems." }, { "left": "issue:43381", "right": "issue:44928", "accept": false, - "reason": "Eval-mode gradient checkpointing restriction and Qwen3.5 RLHF NaN explosion are different issues." + "reason": "Eval-mode gradient checkpointing restriction and Qwen3.5 RLHF NaN/SDPA fallback are unrelated." }, { "left": "issue:44164", "right": "issue:44756", "accept": false, - "reason": "save/from_pretrained extra_state handling and mmap-induced OOM are unrelated." + "reason": "save/from_pretrained extra_state handling and Strix Halo mmap OOM are unrelated." }, { "left": "issue:42796", "right": "issue:46326", "accept": false, - "reason": "Tokenizer BPE marker handling and continue_final_message behavior are both tokenizer-related but not clearly the same concrete defect." + "reason": "Tokenizer BPE marker handling and continue_final_message behavior are both mistral_common-adjacent but not the same concrete bug." }, { "left": "issue:45588", "right": "issue:45663", "accept": false, - "reason": "Flash-attention s_aux=None crash and Gemma4 shared_kv_states KeyError are different code paths." + "reason": "Flash-attention s_aux=None crash and Gemma4 FSDP2 shared_kv_states KeyError are distinct implementation failures." }, { "left": "issue:45684", "right": "issue:45698", "accept": false, - "reason": "Custom-model file permission propagation and wrong-module loading after save_pretrained are separate save/load issues." + "reason": "One is read-only permission propagation during save_pretrained; the other is loading the wrong custom module after save_pretrained. Related area, different bugs." }, { "left": "issue:43854", "right": "issue:45216", "accept": false, - "reason": "GLM-4.7-Flash unit-test loading and Qwen3.5 checkpoint save regression are unrelated." + "reason": "GLM-4.7-Flash test loading issue and Qwen3.5 save_pretrained regression are different model-specific problems." }, { "left": "issue:45292", "right": "issue:45335", - "accept": false, - "reason": "Both mention resize_token_embeddings, but one is about output_embeddings while the other is decoder.embed_tokens in t5gemma; likely separate model-specific bugs." + "accept": true, + "reason": "Both report the same resize_token_embeddings bug; 45335 is just the t5gemma-specific symptom of the generic failure described in 45292." }, { "left": "issue:43645", "right": "issue:44164", "accept": false, - "reason": "Jupyter custom-model initialization and extra_state save/load handling are different problems." + "reason": "Jupyter custom-model initialization breakage and extra_state save/from_pretrained handling are different serialization/loading issues." }, { "left": "issue:42175", "right": "issue:43957", "accept": false, - "reason": "Missing TensorFlow backend in pip extras and meta-device loading failures are unrelated." + "reason": "Missing TensorFlow backend in a torch extra install and meta-device loading regressions are unrelated." }, { "left": "issue:44589", "right": "issue:44756", "accept": false, - "reason": "Float8 storage type lookup failure and Strix Halo mmap OOM are different issues." + "reason": "Float8 storage lookup failure and Strix Halo mmap OOM are different low-level/runtime problems." } ] }, "evaluator_result": { "accept": true, - "feedback": "Grounded in the packet: the only clearly supported duplicate is issue:45276 with issue:45292, and the rest of the soft-edge rejections stay conservative and consistent with the issue titles. The summary does not overclaim beyond the available evidence." + "feedback": "The summary is grounded in the packet and stays conservative on the soft-edge pairs. The resize_token_embeddings chain (45292 with 45276/45335) is well supported by the titles, and the other refusals avoid overclaiming shared root cause from only thematic similarity." }, "evaluator_used": true, "retried": false, @@ -2805,6 +2811,7 @@ "issue:39404", "issue:43425", "issue:43452", + "issue:43596", "issue:43611", "issue:43645", "issue:43646", @@ -2856,13 +2863,13 @@ "issue:43425|issue:43828", "issue:46315|issue:46491", "issue:45812|issue:46315", - "issue:39404|issue:45701", "issue:44530|issue:45084", + "issue:39404|issue:45701", "issue:43611|issue:43645", "issue:43425|issue:46869", "issue:44683|issue:45823", - "issue:39404|issue:46665", "issue:43425|issue:43716", + "issue:39404|issue:46665", "issue:43425|issue:45823", "issue:44589|issue:45084", "issue:44479|issue:44805", @@ -2874,234 +2881,244 @@ "issue:44265|issue:44898", "issue:46050|issue:46326", "issue:43906|issue:46326", - "issue:39404|issue:44912", "issue:44991|issue:45081", + "issue:39404|issue:44912", "issue:44671|issue:44991", "issue:45823|issue:46899", "issue:45479|issue:46894", + "issue:44164|issue:46821", "issue:43856|issue:44756", - "issue:44164|issue:46821" + "issue:43596|issue:44683" ], - "prepared_review_unit_hash": "19e6b55df5cf18fa62ae310a062d4384b1e30775c4a79eff74f3ce8d9c91733b", + "prepared_review_unit_hash": "6cfbee9f681fcec4e1bb7683f453e3fe026f8254bf75686f5125b12412ca7eb2", "component_budget": { "node_count": 514, "item_count": 514, "soft_pair_count": 1891, - "serialized_chars": 980926, - "estimated_input_tokens": 245232, - "estimated_eval_tokens": 490720 + "serialized_chars": 980895, + "estimated_input_tokens": 245224, + "estimated_eval_tokens": 490704 }, "budget": { - "node_count": 47, - "item_count": 47, - "soft_pair_count": 31, - "serialized_chars": 36644, - "estimated_input_tokens": 9161, - "estimated_eval_tokens": 18578 + "node_count": 48, + "item_count": 48, + "soft_pair_count": 32, + "serialized_chars": 37541, + "estimated_input_tokens": 9386, + "estimated_eval_tokens": 19028 }, "overflow_policy": "truncate_then_skip", "trimmed": false, "aggressively_trimmed": false, "split": true, "analyst_result": { - "summary": "This cluster is a loose mix of Transformers v5 regressions and unrelated model/tooling bugs. The items mostly share only broad subsystem overlap (tokenizers, model loading, quantization, training), not the same underlying defect, so no soft edge should be merged.", - "confidence": 0.84, - "canonical_issue_reason": "issue:45812 is the strongest representative: it is a concrete, high-impact v4\u2192v5 tokenizer regression affecting an entire model family, with a clear symptom (wrong token IDs) and a well-defined code path.", + "summary": "The cluster is heterogeneous and I would reject all soft edges. The closest-looking pair, 44402/44451, still lacks evidence of the same underlying bug: 44402 reports a concrete tokenizer invariant violation (`pad`/`unk` outside `vocab_size`), while 44451 only reports that ScandiBERT cannot be loaded.", + "confidence": 0.94, + "canonical_issue_reason": "issue:44402 is the clearest standalone root-cause report because it identifies a specific tokenizer/vocab invariant violation rather than only a downstream loading symptom.", "canonical_pr_reason": null, - "best_issue_reason": "issue:45812 is the best single issue to anchor the cluster because it is specific, broad in user impact, and more representative of the tokenizer-related regression theme than the narrower or more incidental issues.", + "best_issue_reason": "issue:44402 is the most actionable issue in the set: it gives the concrete failure mode and is easiest to map to a fix.", "best_pr_reason": null, "soft_edge_verdicts": [ { "left": "issue:44402", "right": "issue:44451", "accept": false, - "reason": "Both are tokenizer-loading failures, but they involve different models and different failure modes (vocab-size mismatch vs cannot load tokenizer). Same subsystem, not the same bug." + "reason": "Both are tokenizer/load failures, but the evidence is not enough to say they share the same underlying bug; 44402 is a specific vocab-size invariant violation, while 44451 only reports a ScandiBERT load failure." }, { "left": "issue:43646", "right": "issue:43708", "accept": false, - "reason": "Custom model initialization in notebooks is unrelated to Trainer resume/max_steps accounting; different code paths and symptoms." + "reason": "Different bugs: custom model initialization in notebooks versus Trainer resume/max_steps miscalculation." }, { "left": "issue:46894", "right": "issue:46901", "accept": false, - "reason": "Both mention a training-loss double-shift pattern, but they target different models and no evidence shows the same underlying implementation bug or mergeable fix." + "reason": "Same broad double-shift pattern, but the reports point to different models and code paths; not enough evidence they are the same bug." }, { "left": "issue:43425", "right": "issue:43828", "accept": false, - "reason": "Torch version incompatibility is a dependency-level issue; the Phi-tiny-MoE autocast error is a model-specific dtype mismatch. Not the same defect." + "reason": "One is general Torch 2.10 incompatibility; the other is a specific autocast dtype mismatch on Phi-tiny-MoE. Not the same failure." }, { "left": "issue:46315", "right": "issue:46491", "accept": false, - "reason": "One is a performance regression in convert_tokens_to_ids, the other is a CodeLlama round-trip tokenization correctness bug." + "reason": "Different tokenizer regressions: performance slowdown in `convert_tokens_to_ids` versus an encode/decode spacing regression." }, { "left": "issue:45812", "right": "issue:46315", "accept": false, - "reason": "Granite token IDs are wrong; this is a correctness regression, while the other issue is about slow conversion for slow tokenizers with many added tokens." + "reason": "Wrong token IDs for Granite models is a functional tokenizer correctness bug, not the same as `convert_tokens_to_ids` slowness." }, { - "left": "issue:39404", - "right": "issue:45701", + "left": "issue:44530", + "right": "issue:45084", "accept": false, - "reason": "Whisper return_language pipeline behavior and a generic version-changes-the-tokenization report are different failures with no shared concrete bug." + "reason": "Different code paths and symptoms: Qwen3.5 PagedAttentionCache crash versus a template compilation error." }, { - "left": "issue:44530", - "right": "issue:45084", + "left": "issue:39404", + "right": "issue:45701", "accept": false, - "reason": "PagedAttentionCache's invalid group type crash and a template-node compilation TypeError are unrelated runtime/compiler failures." + "reason": "Whisper pipeline `return_language` behavior and a general tokenization-change report are unrelated." }, { "left": "issue:43611", "right": "issue:43645", "accept": false, - "reason": "Both mention Transformers 5.0.0, but one is base_model_prefix loading and the other is custom model init in Jupyter notebooks; different root causes." + "reason": "Both involve custom models, but one is about `base_model_prefix` loading and the other is notebook-defined model initialization." }, { "left": "issue:43425", "right": "issue:46869", "accept": false, - "reason": "Torch 2.10 incompatibility is not the same as a YOLO device_map/multi-GPU forwarding failure." + "reason": "General Torch version incompatibility and YOLO multi-GPU device-map failure are different bugs." }, { "left": "issue:44683", "right": "issue:45823", "accept": false, - "reason": "Compiled flex_attention on torch >=2.9 and Gemma4 device mismatch are separate issues affecting different subsystems." + "reason": "Compiled flex_attention on newer torch and Gemma4 device-mismatch during forward are unrelated failures." }, { - "left": "issue:39404", - "right": "issue:46665", + "left": "issue:43425", + "right": "issue:43716", "accept": false, - "reason": "Whisper return_language regression and from_pretrained behavior inside Python classes are distinct pipeline vs object-construction bugs." + "reason": "Torch 2.10 incompatibility is not the same as a Mistral-3 image-preprocessor/model dtype mismatch." }, { - "left": "issue:43425", - "right": "issue:43716", + "left": "issue:39404", + "right": "issue:46665", "accept": false, - "reason": "General torch version incompatibility is unrelated to Mistral-3 image preprocessor/model dtype mismatch." + "reason": "Whisper pipeline language-return regression and `from_pretrained` inside Python classes are different issues." }, { "left": "issue:43425", "right": "issue:45823", "accept": false, - "reason": "Torch compatibility and Gemma4 device_map placement are different problems." + "reason": "A broad Torch incompatibility report does not match the specific Gemma4 device-map bug." }, { "left": "issue:44589", "right": "issue:45084", "accept": false, - "reason": "Float8 storage lookup failure and template compilation failure are unrelated low-level errors." + "reason": "Missing Float8 storage type and non-template-node compilation failure are unrelated." }, { "left": "issue:44479", "right": "issue:44805", "accept": false, - "reason": "A video-input regression for Qwen models is not the same as a generic mask shape IndexError." + "reason": "Video-input regression in Qwen models and a mask-shape IndexError are different bugs." }, { "left": "issue:43901", "right": "issue:44509", "accept": false, - "reason": "Both are docs-related, but they refer to different removed/changed pipeline APIs and do not describe the same documentation bug." + "reason": "Both are documentation-related, but they concern different deprecated pipeline tasks and not the same change." }, { "left": "issue:43994", "right": "issue:44079", "accept": false, - "reason": "Siglip2 nonsensical outputs and ModelOutput key assignment are different issues at different layers." + "reason": "Incorrect Siglip2 outputs and `ModelOutput` key assignment logic are different problems." }, { "left": "issue:43906", "right": "issue:44530", "accept": false, - "reason": "An isolated reproduction of a prior issue is not the same as the Qwen3.5 PagedAttentionCache crash." + "reason": "An isolated reproduction of another issue is not the same as the Qwen3.5 PagedAttention crash itself." }, { "left": "issue:44164", "right": "issue:45759", "accept": false, - "reason": "save/from_pretrained extra_state handling and composite-config unwrapping are separate model persistence vs config-handling issues." + "reason": "Serialization of `extra_state` and unwrapping `text_config` in multimodal configs are separate bugs." }, { "left": "issue:36032", "right": "issue:43452", "accept": false, - "reason": "T5Tokenizer add_special_tokens conflict and gguf_file loading breakage are unrelated tokenizer/model loading failures." + "reason": "T5 tokenizer method-name conflict and gguf loading breakage are unrelated tokenizer/model-loading issues." }, { "left": "issue:44265", "right": "issue:44898", "accept": false, - "reason": "torch.export with torch_compilable_check and Perceiver non-default-resolution failure are different export/vision bugs." + "reason": "`torch.export` failure with `torch_compilable_check` is different from the Perceiver interpolation failure." }, { "left": "issue:46050", "right": "issue:46326", "accept": false, - "reason": "Quantized save/load breakage and mistral_common continue_final_message behavior do not share a concrete code path." + "reason": "Quantized save/load breakage is unrelated to the `continue_final_message` bug on mistral_common tokenizers." }, { "left": "issue:43906", "right": "issue:46326", "accept": false, - "reason": "Reproduction of an earlier issue and a tokenizer continue_final_message bug are unrelated." - }, - { - "left": "issue:39404", - "right": "issue:44912", - "accept": false, - "reason": "Whisper pipeline return_language and MXFP4 quantization fallback are different regressions." + "reason": "A reproduction issue and a tokenizer continuation bug are not the same underlying defect." }, { "left": "issue:44991", "right": "issue:45081", "accept": false, - "reason": "Both involve tokenizers, but one is a model-specific tokenizer loading regression and the other is a Mistral regex patch crash; not the same underlying bug." + "reason": "Both are tokenizer-loading regressions, but they involve different tokenizers and distinct failure mechanisms." }, { - "left": "issue:44671", - "right": "issue:44991", + "left": "issue:39404", + "right": "issue:44912", + "accept": false, + "reason": "Whisper pipeline behavior and MXFP4 quantization fallback are unrelated." + }, + { + "left": "issue:44671", + "right": "issue:44991", "accept": false, - "reason": "CamemBERT masked-LM prediction errors and est-roberta tokenizer loading failure are unrelated." + "reason": "CamemBERT masked-LM prediction regression and tokenizer load failure are different bugs." }, { "left": "issue:45823", "right": "issue:46899", "accept": false, - "reason": "Both concern Gemma4, but one is a device placement mismatch and the other is a 4-bit quantization dtype-cast problem in multimodal embedders." + "reason": "Gemma4 device mismatch during forward is different from 4-bit multimodal dtype-cast breakage." }, { "left": "issue:45479", "right": "issue:46894", "accept": false, - "reason": "Single-label classification zero-loss and ASR training-loss double-shift are different training bugs with different model families." + "reason": "Sequence-classification zero-loss and ASR training-loss double shift are different failure modes." + }, + { + "left": "issue:44164", + "right": "issue:46821", + "accept": false, + "reason": "`extra_state` save/load handling and a TimesFM `window_size` attribute error are unrelated." }, { "left": "issue:43856", "right": "issue:44756", "accept": false, - "reason": "Qwen3 MoE memory usage and disabling mmap on Strix Halo are unrelated OOM/memory-management issues." + "reason": "Qwen3 MoE memory inefficiency and Strix Halo mmap OOM are different memory/performance problems." }, { - "left": "issue:44164", - "right": "issue:46821", + "left": "issue:43596", + "right": "issue:44683", "accept": false, - "reason": "extra_state persistence and TimesFM window_size AttributeError are different API bugs." + "reason": "Deepspeed Zero3 BertModel indexing failure is unrelated to compiled flex_attention on torch >= 2.9." } ] }, - "evaluator_result": null, - "evaluator_used": false, - "retried": false, + "evaluator_result": { + "accept": false, + "feedback": "The set is mostly conservative, but the 44402/44451 soft-edge is overstated. In the packet, 44402 is a specific tokenizer invariant violation (`pad`/`unk` outside `vocab_size`), while 44451 only says ScandiBERT cannot be loaded; that is not enough evidence to call them the same underlying bug from the provided material. The rest of the pairwise rejections look appropriately cautious." + }, + "evaluator_used": true, + "retried": true, "accepted_nontrivial_soft_edge": false, "error_kind": null, "error_message": null @@ -3135,7 +3152,6 @@ "issue:43482", "issue:43541", "issue:43582", - "issue:43596", "issue:43611", "issue:43632", "issue:43643", @@ -3155,6 +3171,7 @@ "issue:44737", "issue:44811", "issue:44987", + "issue:44991", "issue:45081", "issue:45084", "issue:45479", @@ -3170,15 +3187,14 @@ ], "soft_pairs": [ "issue:43299|issue:43931", - "issue:43596|issue:44683", - "issue:43632|issue:47050", "issue:45479|issue:46897", "issue:44488|issue:44987", + "issue:43632|issue:47050", "issue:44488|issue:45081", "issue:43611|issue:43644", "issue:36032|issue:42915", - "issue:42796|issue:43404", "issue:44462|issue:46489", + "issue:42796|issue:43404", "issue:43716|issue:45084", "issue:46285|issue:46402", "issue:44661|issue:44683", @@ -3196,211 +3212,218 @@ "issue:43582|issue:43632", "issue:36032|issue:43645", "issue:44206|issue:44220", + "issue:43452|issue:43482", "issue:44811|issue:45584", - "issue:43452|issue:43482" + "issue:44987|issue:44991" ], - "prepared_review_unit_hash": "54fec704b84bba89a1ade55779f3e0e4830c6fc1cfcee732692a939f6759fdd3", + "prepared_review_unit_hash": "06d8079f1636a34ba2cbd5db9588a2f6729967b8d35a9216e7f1583db1557d7a", "component_budget": { "node_count": 514, "item_count": 514, "soft_pair_count": 1891, - "serialized_chars": 980926, - "estimated_input_tokens": 245232, - "estimated_eval_tokens": 490720 + "serialized_chars": 980895, + "estimated_input_tokens": 245224, + "estimated_eval_tokens": 490704 }, "budget": { "node_count": 48, "item_count": 48, "soft_pair_count": 29, - "serialized_chars": 36554, - "estimated_input_tokens": 9139, - "estimated_eval_tokens": 18534 + "serialized_chars": 36516, + "estimated_input_tokens": 9129, + "estimated_eval_tokens": 18514 }, "overflow_policy": "truncate_then_skip", "trimmed": false, "aggressively_trimmed": false, "split": true, "analyst_result": { - "summary": "Mostly a heterogeneous set of unrelated closed issues, with a few small duplicate subclusters around GGUF loading, custom model initialization, audio feature extraction, and Qwen3-VL loading regressions.", - "confidence": 0.62, - "canonical_issue_reason": "Issue #43452 is the broadest representative of one real duplicate subcluster: a general GGUF/from_pretrained loading breakage, with #43482 looking like a specific model instance of the same failure.", + "summary": "Most pairs are separate regressions affecting different models/features. I\u2019d use the broad v5 custom-model initialization regression as the issue canonical, and only one soft edge looks like the same audio fbank extraction bug.", + "confidence": 0.69, + "canonical_issue_reason": "issue:43646 is the broadest umbrella among the cluster\u2019s v5 initialization/loading regressions and is the most plausible duplicate sink for nearby custom-model breakage reports.", "canonical_pr_reason": null, - "best_issue_reason": "#43452 is the best umbrella issue among the accepted duplicate candidates because it describes the loading regression at the API level rather than a single model variant.", + "best_issue_reason": "issue:43646 is the most generally phrased and centrally applicable issue in the set, making it the best global duplicate target for related model-initialization regressions.", "best_pr_reason": null, "soft_edge_verdicts": [ { "left": "issue:43299", "right": "issue:43931", - "accept": true, - "reason": "Both point to the same Qwen3-VL MoE model-loading regression; #43931 looks like a concrete failure mode of the broader #43299 bug." - }, - { - "left": "issue:43596", - "right": "issue:44683", "accept": false, - "reason": "Different code paths and failures: deepspeed/BertModel init vs torch 2.9 flex_attention compilation." - }, - { - "left": "issue:43632", - "right": "issue:47050", - "accept": false, - "reason": "Unrelated regressions: an HF init flag issue versus a Canine ONNX dtype mismatch in attention." + "reason": "Both are Qwen-related loading errors, but one is a Qwen3Moe FP8 config failure and the other is a Qwen3-VL shape mismatch; different concrete bugs." }, { "left": "issue:45479", "right": "issue:46897", "accept": false, - "reason": "Both are training-loss bugs, but the underlying mechanisms differ: degenerate classification loss vs sequence-shift error." + "reason": "One is a sequence-classification zero-loss bug, the other is a training-loss double-shift issue in Florence2; the failure modes and code paths differ." }, { "left": "issue:44488", "right": "issue:44987", "accept": false, - "reason": "Both are loading failures, but the model families and described symptoms differ; not enough evidence of one shared bug." + "reason": "Both mention model loading, but one is about tokenizer.json being ignored and the other is a specific transformers>=5.1.0 loading failure; not clearly the same bug." + }, + { + "left": "issue:43632", + "right": "issue:47050", + "accept": false, + "reason": "These are unrelated: `_is_hf_initialized` regression versus an ONNX fp16 attention dtype mismatch." }, { "left": "issue:44488", "right": "issue:45081", "accept": false, - "reason": "Generic model loading failure vs a specific Mistral regex patch crash; different paths." + "reason": "Both are tokenizer-loading regressions, but the Mistral regex crash is a different specific path than ignoring tokenizer.json." }, { "left": "issue:43611", "right": "issue:43644", "accept": false, - "reason": "Base-model-prefix loading regression and non-persistent buffer initialization are separate problems." + "reason": "Both are v5 breakages, but one is base_model_prefix loading and the other is non-persistent buffer initialization; different defects." }, { "left": "issue:36032", "right": "issue:42915", "accept": false, - "reason": "Tokenizer method-name conflict in T5 is unrelated to Qwen3Moe failing with FineGrainedFP8Config." + "reason": "T5 tokenizer method-name conflict and Qwen3Moe FP8 loading are unrelated." }, { - "left": "issue:42796", - "right": "issue:43404", + "left": "issue:44462", + "right": "issue:46489", "accept": false, - "reason": "Tokenizer BPE-marker handling and lm_head weight tying are different Mistral3 issues." + "reason": "Both are tokenizer issues, but one is repository tokenizer.json selection and the other is DeepSeek-Coder output correctness after a prior fix; not the same bug." }, { - "left": "issue:44462", - "right": "issue:46489", + "left": "issue:42796", + "right": "issue:43404", "accept": false, - "reason": "Both involve tokenizer behavior, but one is generic tokenizer.json loading and the other is a model-specific output regression." + "reason": "Tokenizer BPE-marker handling and Mistral3 lm_head weight tying are different problems." }, { "left": "issue:43716", "right": "issue:45084", "accept": false, - "reason": "Different subsystems: image-preprocessor dtype mismatch vs a template compilation error." + "reason": "Image preprocessor dtype mismatch and template compilation failure are unrelated." }, { "left": "issue:46285", "right": "issue:46402", "accept": false, - "reason": "CLIP text-model structure inconsistency and Sam3Model weight-loading are unrelated." + "reason": "CLIP structure inconsistency and Sam3Model text-encoder weight loading are different model-specific regressions." }, { "left": "issue:44661", "right": "issue:44683", "accept": false, - "reason": "Tokenizer-mapping utility failure is unrelated to the flex_attention compile bug." + "reason": "Model registration logic and torch>=2.9 flex_attention compilation are unrelated." }, { "left": "issue:32090", "right": "issue:39404", "accept": false, - "reason": "Trainer GPU broadcast NoneType error is not the same as Whisper return_language pipeline behavior." + "reason": "Trainer GPU broadcast NoneType error and Whisper return_language regression are unrelated." }, { "left": "issue:43643", "right": "issue:44530", "accept": false, - "reason": "AutoConfig trust_remote_code missing fields is unrelated to the Qwen3.5 PagedAttentionCache crash." + "reason": "Missing fields from AutoConfig with trust_remote_code and PagedAttentionCache invalid group type are different failures." }, { "left": "issue:44451", "right": "issue:44987", "accept": false, - "reason": "Both are loading problems, but the titles point to different models and different failure causes." + "reason": "Both are loading failures, but they concern different repos and likely different root causes." + }, + { + "left": "issue:43541", + "right": "issue:45084", + "accept": false, + "reason": "Mixtral grouped_mm tracing failure and non-template-node compilation error are not the same code-path bug." }, { "left": "issue:42915", "right": "issue:44530", "accept": false, - "reason": "Both are Qwen-family issues, but one is FP8 config loading and the other is a cache/group-type runtime crash." + "reason": "Both involve Qwen-family models, but one is FineGrainedFP8Config loading and the other is cache group-type handling; distinct issues." }, { "left": "issue:44737", "right": "issue:45859", "accept": false, - "reason": "XLNet CPU arange placement bug is unrelated to Qwen3_5Moe CPU-offload rotary embedding behavior." + "reason": "XLNet CPU tensor creation and Qwen3_5MoE CPU offload incompatibility are unrelated." }, { "left": "issue:40990", "right": "issue:43421", "accept": false, - "reason": "Perplexity regression and runtime post-processor updates are unrelated issues." + "reason": "Perplexity regression on gpt-oss and runtime post-processor updates for special tokens are different problems." }, { "left": "issue:34567", "right": "issue:40444", "accept": false, - "reason": "TrainerState token counting and multimodal iterable-dataset failures are unrelated." + "reason": "TrainerState token counting and multi-image iterable-dataset finetuning failures are unrelated." }, { "left": "issue:43644", "right": "issue:43646", - "accept": true, - "reason": "Both describe the same v5 custom-model initialization regression; #43644 appears to be one manifestation of #43646's broader issue." + "accept": false, + "reason": "Both are v5/custom-model regressions, but junk-filled non-persistent buffers is not the same bug as custom model initialization breaking." }, { "left": "issue:42175", "right": "issue:46178", "accept": false, - "reason": "Missing tensorflow in torch extras and an invalid BACKENDS_MAPPING entry are different packaging/backend issues." + "reason": "Backend packaging omission and BACKENDS_MAPPING validation are different dependency/backend issues." }, { "left": "issue:40444", "right": "issue:46326", "accept": false, - "reason": "Multimodal dataset loading and mistral_common continue_final_message are different code paths." + "reason": "Multi-image Qwen2.5-VL finetuning and mistral_common continue_final_message behavior are unrelated." }, { "left": "issue:43582", "right": "issue:43632", "accept": false, - "reason": "Apple Silicon warmup TypeError and the _is_hf_initialized regression are unrelated." + "reason": "Apple Silicon allocator warmup TypeError and `_is_hf_initialized` regression are unrelated." }, { "left": "issue:36032", "right": "issue:43645", "accept": false, - "reason": "T5 tokenizer method conflict and Jupyter custom-model initialization are different bugs." + "reason": "T5 tokenizer method conflict and Jupyter custom model initialization breakage are different issues." }, { "left": "issue:44206", "right": "issue:44220", "accept": true, - "reason": "These look like the same audio feature-extraction bug; #44206 is a specific crash caused by the shared _torch_extract_fbank_features path." + "reason": "Both point to the same audio fbank extraction path: 44206 names the unsupported center-arg crash in `_torch_extract_fbank_features`, and 44220 is the generic report on that function." + }, + { + "left": "issue:43452", + "right": "issue:43482", + "accept": false, + "reason": "gguf_file loading regressions and Qwen2.5-GGUF v5 loading failure are related only broadly; the specific failure is not clearly the same." }, { "left": "issue:44811", "right": "issue:45584", "accept": false, - "reason": "Both involve Whisper special tokens, but one is decode behavior and the other is generation failure after align_special_tokens." + "reason": "Whisper batch_decode skip_special_tokens and empty-transcription generation after align_special_tokens are separate tokenizer/generation bugs." }, { - "left": "issue:43452", - "right": "issue:43482", - "accept": true, - "reason": "Same GGUF-loading regression: #43452 is the generic from_pretrained breakage and #43482 is a specific Qwen2.5-GGUF instance." + "left": "issue:44987", + "right": "issue:44991", + "accept": false, + "reason": "Both are tokenization/loading regressions on different repositories, but the titles indicate different concrete breakages." } ] }, "evaluator_result": { "accept": true, - "feedback": "The summary is grounded in the packet: it stays at the title level, identifies only small duplicate-like clusters, and the accepted pairs are plausible matches rather than overclaims. No clear evidence of overstatement or non-mergeable PR pairing issues appears here." + "feedback": "The summary is grounded in the packet, and the one accepted soft edge (44206/44220) is well supported by the titles. The rejections are conservative overall and do not overclaim identity beyond what the packet shows. The canonical issue choice is plausible given the nearby v5 custom-model initialization regressions, without asserting stronger evidence than available." }, "evaluator_used": true, "retried": false, @@ -3428,10 +3451,12 @@ "issue:43208", "issue:43262", "issue:43317", + "issue:43322", "issue:43377", "issue:43404", "issue:43482", "issue:43531", + "issue:43540", "issue:43632", "issue:43645", "issue:43646", @@ -3456,7 +3481,6 @@ "issue:44849", "issue:44877", "issue:44987", - "issue:44991", "issue:45081", "issue:45084", "issue:45325", @@ -3470,14 +3494,13 @@ "issue:47013" ], "soft_pairs": [ - "issue:44987|issue:44991", "issue:45084|issue:45507", "issue:43317|issue:43482", "issue:43746|issue:46326", "issue:43262|issue:43377", "issue:43208|issue:47013", - "issue:44403|issue:45081", "issue:43792|issue:46869", + "issue:44403|issue:45081", "issue:42915|issue:45084", "issue:43645|issue:46661", "issue:43646|issue:44534", @@ -3496,205 +3519,206 @@ "issue:44008|issue:46899", "issue:42915|issue:44589", "issue:44265|issue:44610", - "issue:43632|issue:46821", - "issue:43531|issue:44877" + "issue:43531|issue:44877", + "issue:43322|issue:43540", + "issue:43632|issue:46821" ], - "prepared_review_unit_hash": "2bdf64bab5a58403d618207888c413484c80bc1d30c5239bcad729565a13e1db", + "prepared_review_unit_hash": "61c3cffde65db7c3f5f796b02cee37a466bb24d8dacf1e180ce15a494c1940da", "component_budget": { "node_count": 514, "item_count": 514, "soft_pair_count": 1891, - "serialized_chars": 980926, - "estimated_input_tokens": 245232, - "estimated_eval_tokens": 490720 + "serialized_chars": 980895, + "estimated_input_tokens": 245224, + "estimated_eval_tokens": 490704 }, "budget": { - "node_count": 47, - "item_count": 47, + "node_count": 48, + "item_count": 48, "soft_pair_count": 28, - "serialized_chars": 35717, - "estimated_input_tokens": 8930, - "estimated_eval_tokens": 18116 + "serialized_chars": 36236, + "estimated_input_tokens": 9059, + "estimated_eval_tokens": 18374 }, "overflow_policy": "truncate_then_skip", "trimmed": false, "aggressively_trimmed": false, "split": true, "analyst_result": { - "summary": "This cluster is mostly a loose collection of Transformers bug reports with only superficial similarity in a few v5/custom-model-loading cases. I found no soft pair that looks like the same underlying bug/change; the closest items still appear to be distinct failures in different code paths or models.", - "confidence": 0.63, - "canonical_issue_reason": "issue:43632 is the broadest/root-regression report here: it describes the v5 `_is_hf_initialized` breakage that plausibly underlies several nearby custom-model initialization complaints.", + "summary": "This cluster is a loose set of unrelated issue reports that only weakly resemble one another through broad \"Transformers v5 regression\" wording. They cover many distinct models, loaders, and runtime paths, so none of the soft pairs look like safe duplicates.", + "confidence": 0.86, + "canonical_issue_reason": "Issue #43646 is the broadest representative of the cluster\u2019s main recurring theme: Transformers v5 breaking model initialization/loading. It is more general than the model-specific reports, so it works best as the cluster\u2019s issue anchor.", "canonical_pr_reason": null, - "best_issue_reason": "issue:43632 is the best representative because it captures the most general and reusable failure mode in the set, rather than a model-specific symptom.", + "best_issue_reason": "#43646 is the most central and general issue in this set, but it should be treated as a representative umbrella report rather than a true duplicate target for the others.", "best_pr_reason": null, "soft_edge_verdicts": [ - { - "left": "issue:44987", - "right": "issue:44991", - "accept": false, - "reason": "Both are load failures after v5, but they affect different models/tokenizer paths and don\u2019t describe the same bug." - }, { "left": "issue:45084", "right": "issue:45507", "accept": false, - "reason": "Different model/code-path failures; no evidence they share a concrete root cause." + "reason": "Different problems: template compilation error vs GraniteMoEHybrid calling an invalid method." }, { "left": "issue:43317", "right": "issue:43482", "accept": false, - "reason": "Both involve model loading, but one is dequantized device_map offload and the other is GGUF/v5 loading; distinct issues." + "reason": "Both are loading failures, but one is device_map/offload dequantization and the other is Qwen2.5-GGUF/Transformers v5 compatibility." }, { "left": "issue:43746", "right": "issue:46326", "accept": false, - "reason": "Unrelated bugs: PEFT local checkpoint loading vs tokenizer chat-template continuation behavior." + "reason": "PEFT local checkpoint loading for GraniteSpeech is unrelated to the mistral_common continue_final_message tokenizer bug." }, { "left": "issue:43262", "right": "issue:43377", "accept": false, - "reason": "Different subsystems and symptoms: audio sampling-rate defaulting vs MIMI batch padding-mask mismatch." + "reason": "Audio chat-template sampling-rate mismatch is unrelated to MIMI encoder batch padding-mask behavior." }, { "left": "issue:43208", "right": "issue:47013", "accept": false, - "reason": "Same broad xLSTM area, but one is a bundle of three training bugs and the other is a specific train-mode unpacking crash." + "reason": "Both concern xLSTM, but one reports multiple small-model training bugs and the other is a specific return_last_states unpacking crash." }, { - "left": "issue:44403", - "right": "issue:45081", + "left": "issue:43792", + "right": "issue:46869", "accept": false, - "reason": "Noise when loading a transformer is not the same as the Mistral regex patch crash." + "reason": "Whisper model runtime failure and YOLO multi-GPU device_map_auto failure are different code paths and model families." }, { - "left": "issue:43792", - "right": "issue:46869", + "left": "issue:44403", + "right": "issue:45081", "accept": false, - "reason": "Different models and failures; both are loading-related only at a very high level." + "reason": "Unnecessary loading noise is not the same bug as Mistral regex patch crashing on tokenizer backend access." }, { "left": "issue:42915", "right": "issue:45084", "accept": false, - "reason": "Qwen3MoE FP8 config failure is unrelated to the template-node compilation error." + "reason": "Qwen3MoE FP8 config failure and template-node compilation failure are unrelated." }, { "left": "issue:43645", "right": "issue:46661", "accept": false, - "reason": "Both concern custom/remote-code model initialization, but the reported failures and triggers differ enough that they are not the same bug." + "reason": "Notebook custom-model initialization breakage and remote-code config_class AttributeError are different initialization failures." }, { "left": "issue:43646", "right": "issue:44534", "accept": false, - "reason": "Custom model initialization breakage and non-persistent buffer serialization are separate regressions." + "reason": "Generic custom model initialization breakage is distinct from non-persistent buffer corruption/junk values." }, { "left": "issue:43377", "right": "issue:43819", "accept": false, - "reason": "Both compare forward/other-path outputs, but they involve different models and mechanisms." + "reason": "MIMI padding-mask discrepancy and DAC from_latents/forward mismatch are different model-specific numerical bugs." }, { "left": "issue:44849", "right": "issue:45325", "accept": false, - "reason": "Different Qwen variants and different bugs: hidden-states output vs rope index scaling." + "reason": "Qwen3.5 hidden_states bug and Qwen2.5-VL rope_index scaling bug affect different model variants and different internals." }, { "left": "issue:43883", "right": "issue:44488", "accept": false, - "reason": "Molmo tied-weights attribute error and BERT loading failure are unrelated." + "reason": "Missing all_tied_weights_keys on MolmoForCausalLM is unrelated to ScandiBERT loading failures." }, { "left": "issue:44987", "right": "issue:45081", "accept": false, - "reason": "Tokenizer loading failure for one model is not the same as the Mistral regex backend crash." + "reason": "Physical-intelligence/fast loading regression and Mistral regex patch crash are different loaders and failures." }, { "left": "issue:28530", "right": "issue:34567", "accept": false, - "reason": "Training early-stopping metric selection and token-count tracking are unrelated training bugs." + "reason": "Early stopping metric selection and num_input_tokens_seen not updating are unrelated Trainer behaviors." }, { "left": "issue:43404", "right": "issue:44466", "accept": false, - "reason": "Both mention tied lm_head weights, but one is a model-specific tying bug and the other is device-dependent serialization behavior." + "reason": "lm_head weight tying in Mistral3ForConditionalGeneration is unrelated to device-dependent lm_head serialization." }, { "left": "issue:42915", "right": "issue:44291", "accept": false, - "reason": "Both are v5-related, but one is FP8 config handling and the other is init_empty_weights/_is_hf_initialized argument handling." + "reason": "Qwen3Moe FineGrainedFP8Config failure and init_empty_weights _is_hf_initialized argument issue are different v5 regressions." }, { "left": "issue:43994", "right": "issue:44610", "accept": false, - "reason": "Different models and different failure modes: nonsensical outputs vs wrong processor image size." + "reason": "SigLIP2 nonsensical outputs with AutoModel/pipeline are not the same as OmDet-Turbo processor/model input-size mismatch." }, { "left": "issue:36032", "right": "issue:43742", "accept": false, - "reason": "Tokenizer special-token constructor conflict and a key error loading a model are not the same issue." + "reason": "T5Tokenizer add_special_tokens collision and MobileLLM-125M key error are separate tokenizer/model loading issues." }, { "left": "issue:44451", "right": "issue:45081", "accept": false, - "reason": "Different tokenizer/model loading failures with different triggers." + "reason": "ScandiBERT loading failure is unrelated to the Mistral tokenizer regex backend crash." }, { "left": "issue:44464", "right": "issue:45987", "accept": false, - "reason": "Chunked generation inconsistency and StaticCache return-type mismatch are separate cache/generation bugs." + "reason": "Chunked generation inconsistency with compiled forward and StaticCache.get_seq_length returning tensor shape are different generation/cache bugs." }, { "left": "issue:44062", "right": "issue:44987", "accept": false, - "reason": "Both involve tokenizers, but the AddedToken constructor error is distinct from the est-roberta loading failure." + "reason": "AddedToken special-arg duplication and physical-intelligence/fast loading regression do not share a concrete code path." }, { "left": "issue:44008", "right": "issue:46899", "accept": false, - "reason": "Different Gemma model variants and different bugs: variable-name collision vs unsafe dtype cast in embedders." + "reason": "Gemma 3n variable-name collision in forward is unrelated to Gemma4 unsafe dtype casting in multimodal embedders." }, { "left": "issue:42915", "right": "issue:44589", "accept": false, - "reason": "FP8 config failure and missing Float8 storage type are related only thematically, not the same bug." + "reason": "Qwen3Moe FP8 config failure and missing Float8 storage type are different quantization/storage issues." }, { "left": "issue:44265", "right": "issue:44610", "accept": false, - "reason": "torch.export/checking failure and processor image-size mismatch are unrelated." + "reason": "torch.export/torch_compilable_check failure and OmDet-Turbo image-size mismatch are unrelated." }, { - "left": "issue:43632", - "right": "issue:46821", + "left": "issue:43531", + "right": "issue:44877", "accept": false, - "reason": "The `_is_hf_initialized` regression is unrelated to the TimesFM window_size attribute error." + "reason": "Qwen3-MoE sliding_window behavior and strict config blocking granite_speech loading are different model/config issues." }, { - "left": "issue:43531", - "right": "issue:44877", + "left": "issue:43322", + "right": "issue:43540", + "accept": false, + "reason": "Llava Next loading segfault and Qwen3OmniMoe video input ValueError are different multimodal failures." + }, + { + "left": "issue:43632", + "right": "issue:46821", "accept": false, - "reason": "Qwen3-MoE sliding-window behavior and granite_speech config strictness are unrelated." + "reason": "_is_hf_initialized flag breakage and TimesFM window_size AttributeError are unrelated Transformers v5 regressions." } ] }, @@ -3724,17 +3748,16 @@ "issue:40444", "issue:40990", "issue:41720", + "issue:42175", "issue:42796", "issue:43089", "issue:43296", - "issue:43322", "issue:43381", "issue:43421", "issue:43425", "issue:43482", "issue:43525", "issue:43531", - "issue:43540", "issue:43611", "issue:43632", "issue:43716", @@ -3757,6 +3780,8 @@ "issue:45216", "issue:45325", "issue:45357", + "issue:45538", + "issue:45542", "issue:45663", "issue:45812", "issue:45923", @@ -3767,13 +3792,12 @@ "issue:46899" ], "soft_pairs": [ - "issue:43322|issue:43540", "issue:36032|issue:41720", "issue:44479|issue:45325", "issue:43089|issue:45987", "issue:43425|issue:44403", - "issue:43296|issue:43531", "issue:45216|issue:45357", + "issue:43296|issue:43531", "issue:40444|issue:45325", "issue:43611|issue:44534", "issue:44991|issue:46315", @@ -3784,227 +3808,235 @@ "issue:43381|issue:45923", "issue:42796|issue:43746", "issue:45812|issue:46491", - "issue:43632|issue:44756", "issue:43792|issue:44589", + "issue:43632|issue:44756", "issue:33666|issue:40444", "issue:44291|issue:44589", - "issue:42796|issue:46710", "issue:40990|issue:44811", + "issue:42796|issue:46710", "issue:43482|issue:43525", "issue:35707|issue:40444", "issue:44079|issue:45663", "issue:43531|issue:44291", "issue:43421|issue:44568", - "issue:41720|issue:45081" + "issue:41720|issue:45081", + "issue:42175|issue:45542", + "issue:43482|issue:45538" ], - "prepared_review_unit_hash": "8d3658e5247acd281ffffb0efa34268ad2f491cc7f8e4472097e82f695b0bf04", + "prepared_review_unit_hash": "284a006b07eebc40daf31b92a73027887225eafc3991fe7069aab113069a27fd", "component_budget": { "node_count": 514, "item_count": 514, "soft_pair_count": 1891, - "serialized_chars": 980926, - "estimated_input_tokens": 245232, - "estimated_eval_tokens": 490720 + "serialized_chars": 980895, + "estimated_input_tokens": 245224, + "estimated_eval_tokens": 490704 }, "budget": { - "node_count": 47, - "item_count": 47, - "soft_pair_count": 29, - "serialized_chars": 35916, - "estimated_input_tokens": 8979, - "estimated_eval_tokens": 18214 + "node_count": 48, + "item_count": 48, + "soft_pair_count": 30, + "serialized_chars": 36827, + "estimated_input_tokens": 9207, + "estimated_eval_tokens": 18670 }, "overflow_policy": "truncate_then_skip", "trimmed": false, "aggressively_trimmed": false, "split": true, "analyst_result": { - "summary": "This cluster is mostly heterogeneous and does not form one broad duplicate set. The only plausible duplicate pair is the Qwen3.5 save_pretrained regression: issue 45216 and issue 45357. The rest are distinct bug reports across unrelated models, tokenizer paths, multimodal pipelines, and performance/regression topics.", - "confidence": 0.89, - "canonical_issue_reason": "issue:45216 is the best canonical anchor for the only likely duplicate subcluster: it describes the broader Qwen3.5 save_pretrained regression, while issue:45357 reads like a narrower follow-up symptom about incorrect visual encoder keys.", + "summary": "These issues are largely unrelated; the only strong duplicate-like pair is the Qwen3.5 `save_pretrained` regression reports. Most other pairs share a broad subsystem or model family but describe different failure modes.", + "confidence": 0.88, + "canonical_issue_reason": "issue:45216 is the clearest representative: it describes a concrete Qwen3.5 `save_pretrained` regression with a version boundary, and it aligns closely with the later follow-up report in 45357.", "canonical_pr_reason": null, - "best_issue_reason": "issue:45216 is the clearest representative issue in this set because it is the broader, earlier regression report with a concrete version scope; however, the cluster is too mixed for it to be a strong global representative beyond that subcluster.", + "best_issue_reason": "issue:45216 is the best issue candidate because it is specific, reproducible, and captures a user-visible regression rather than a vague compatibility or performance complaint.", "best_pr_reason": null, "soft_edge_verdicts": [ - { - "left": "issue:43322", - "right": "issue:43540", - "accept": false, - "reason": "Both are multimodal model failures, but they hit different models and different code paths: Llava Next loading segfault vs Qwen3OmniMoe video-processing ValueError." - }, { "left": "issue:36032", "right": "issue:41720", "accept": false, - "reason": "Tokenizer load-time AttributeError and Qwen3 auto device-mapping CUDA assert are unrelated failures." + "reason": "Different models and failure modes: T5 tokenizer init conflict vs Qwen3 auto-device-map CUDA assert." }, { "left": "issue:44479", "right": "issue:45325", "accept": false, - "reason": "Both involve Qwen VL/video logic, but one is a broad video-input regression across several models while the other is a still-image rope-position scaling bug; not the same concrete issue." + "reason": "Both involve Qwen VL, but one is a broad video-input regression and the other is a still-image RoPE/index scaling bug." }, { "left": "issue:43089", "right": "issue:45987", "accept": false, - "reason": "Generation performance overhead from GPU syncs is unrelated to the StaticCache return-type contract bug." + "reason": "Generation performance overhead vs a cache length return-type bug; not the same code-path issue." }, { "left": "issue:43425", "right": "issue:44403", "accept": false, - "reason": "Torch version incompatibility and noisy load output are different problems." - }, - { - "left": "issue:43296", - "right": "issue:43531", - "accept": false, - "reason": "PaddleOCR-VL loading in vLLM and Qwen3-MoE sliding_window behavior are unrelated model-specific bugs." + "reason": "Torch-version incompatibility is unrelated to the loading-noise complaint." }, { "left": "issue:45216", "right": "issue:45357", "accept": true, - "reason": "Same Qwen3.5 save_pretrained regression; 45357 appears to be a narrower manifestation of the same incorrect checkpoint serialization problem." + "reason": "Both report the same Qwen3.5 `save_pretrained` checkpoint regression; 45357 reads like a later, narrower manifestation of the same bug." + }, + { + "left": "issue:43296", + "right": "issue:43531", + "accept": false, + "reason": "Different problems: model loading failure for PaddleOCR-VL vs Qwen3-MoE sliding-window behavior." }, { "left": "issue:40444", "right": "issue:45325", "accept": false, - "reason": "Both are Qwen2.5-VL-related, but one is a finetuning/data-loading failure with multiple images per prompt and the other is a rope-index scaling bug for still images." + "reason": "IterableDataset/multi-image finetuning failure is different from a Qwen2.5-VL still-image position-id scaling bug." }, { "left": "issue:43611", "right": "issue:44534", "accept": false, - "reason": "General v5 loading regression with base_model_prefix and non-persistent buffer junk are different failure modes." + "reason": "Base-model-prefix loading regression and non-persistent buffer junk are separate model-initialization issues." }, { "left": "issue:44991", "right": "issue:46315", "accept": false, - "reason": "Tokenizer load failure for est-roberta and slow convert_tokens_to_ids performance are not the same bug." + "reason": "Tokenizer loading failure and slow `convert_tokens_to_ids` are different tokenizer regressions." }, { "left": "issue:44220", "right": "issue:44683", "accept": false, - "reason": "Audio feature extraction and compiled flex_attention failures are unrelated subsystems." + "reason": "Audio feature extraction issue vs compiled flex_attention on newer torch; unrelated code paths." }, { "left": "issue:43883", "right": "issue:44991", "accept": false, - "reason": "Missing all_tied_weights_keys on MolmoForCausalLM is unrelated to a tokenizer load regression." + "reason": "Missing tied-weights attribute in Molmo is unrelated to tokenizer loading for EMBEDDIA/est-roberta." }, { "left": "issue:43716", "right": "issue:46899", "accept": false, - "reason": "Both mention multimodal dtype issues, but they affect different models and different processing stages; not enough to treat as one bug." + "reason": "Both are multimodal dtype-related reports, but they affect different models and different processing stages." }, { "left": "issue:44403", "right": "issue:44991", "accept": false, - "reason": "Load-time logging noise and tokenizer loading failure are distinct issues." + "reason": "Loading noise is not the same as tokenizer loading failure." }, { "left": "issue:43381", "right": "issue:45923", "accept": false, - "reason": "Gradient checkpointing blocked in eval mode is not the same as a missing supports_gradient_checkpointing flag on a trust_remote_code variant." + "reason": "Eval-mode checkpointing restriction is distinct from a missing `supports_gradient_checkpointing` flag." }, { "left": "issue:42796", "right": "issue:43746", "accept": false, - "reason": "Tokenizer BPE-marker handling and GraniteSpeech PEFT checkpoint loading are unrelated." + "reason": "Tokenizer BPE-marker handling and PEFT local checkpoint loading are unrelated." }, { "left": "issue:45812", "right": "issue:46491", "accept": false, - "reason": "Both are tokenizer regressions, but one is wrong token IDs for Granite models and the other is a leading-space round-trip issue for CodeLlama; different behavior and likely different code paths." + "reason": "Granite token ID regression and CodeLlama leading-space round-trip regression are different tokenizer behaviors." }, { - "left": "issue:43632", - "right": "issue:44756", + "left": "issue:43792", + "right": "issue:44589", "accept": false, - "reason": "A Transformers v5 initialization-flag regression and an mmap OOM workaround on Strix Halo are unrelated." + "reason": "Whisper runtime failure is not the same as the float8 storage lookup error." }, { - "left": "issue:43792", - "right": "issue:44589", + "left": "issue:43632", + "right": "issue:44756", "accept": false, - "reason": "Whisper model runtime failure and missing Float8 storage support are different errors in different areas." + "reason": "`_is_hf_initialized` flag regression and Strix Halo mmap OOM are unrelated." }, { "left": "issue:33666", "right": "issue:40444", "accept": false, - "reason": "Both are Qwen VL training issues, but one is multi-GPU training and the other is iterable-dataset/multi-image prompt handling; not the same bug." + "reason": "Both mention Qwen VL training, but one is generic multi-GPU training and the other is an IterableDataset multi-image finetuning failure." }, { "left": "issue:44291", "right": "issue:44589", "accept": false, - "reason": "Unexpected _is_hf_initialized argument during init_empty_weights and missing Float8 storage object are unrelated TypeErrors." + "reason": "`init_empty_weights` argument handling and float8 storage resolution are separate v5 regressions." }, { - "left": "issue:42796", - "right": "issue:46710", + "left": "issue:40990", + "right": "issue:44811", "accept": false, - "reason": "Tokenizer marker handling and model accuracy regression after a version upgrade are different issues." + "reason": "Perplexity regression on GPT-OSS is unrelated to Whisper batch decode special-token handling." }, { - "left": "issue:40990", - "right": "issue:44811", + "left": "issue:42796", + "right": "issue:46710", "accept": false, - "reason": "Perplexity spike on GPT-OSS is unrelated to Whisper processor.batch_decode ignoring skip_special_tokens." + "reason": "Tokenizer BPE-marker bug and DeepSeek accuracy regression are different symptoms and likely different root causes." }, { "left": "issue:43482", "right": "issue:43525", "accept": false, - "reason": "Qwen2.5-GGUF loading failure under Transformers v5 and Llama4Config missing pad_token_id are different model/config bugs." + "reason": "Qwen2.5-GGUF loading failure and missing `pad_token_id` on Llama4Config are unrelated model/config bugs." }, { "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 failures." + "reason": "Progressive generation with `inputs_embeds`/`past_key_values` is unrelated to Qwen2.5-VL multi-image iterable finetuning." }, { "left": "issue:44079", "right": "issue:45663", "accept": false, - "reason": "ModelOutput key assignment logic and Gemma4 FSDP2 shared_kv_states rebuilding are unrelated." + "reason": "ModelOutput key assignment and Gemma4 FSDP2 shared_kv_state rebuild are different internal issues." }, { "left": "issue:43531", "right": "issue:44291", "accept": false, - "reason": "Qwen3-MoE sliding_window behavior and init_empty_weights _is_hf_initialized argument are different bugs." + "reason": "Qwen3-MoE sliding_window behavior and `init_empty_weights` type error do not match." }, { "left": "issue:43421", "right": "issue:44568", "accept": false, - "reason": "Runtime post-processor updates for special tokens and mdeberta-v3 add_special_tokens BOS/EOS regression are related in broad theme but not the same concrete bug." + "reason": "Runtime post-processor updates for tokenizers are unrelated to missing BOS/EOS insertion in mdeberta-v3." }, { "left": "issue:41720", "right": "issue:45081", "accept": false, - "reason": "Qwen3 auto device-mapping cudaErrorAssert and Mistral regex patch crashing on backend_tokenizer are unrelated." + "reason": "CUDA assert during device-mapped Qwen3 load is unrelated to a Mistral tokenizer regex crash." + }, + { + "left": "issue:42175", + "right": "issue:45542", + "accept": false, + "reason": "Backend packaging omission with `.[torch]` is distinct from a tensorboard-without-TF backend error." + }, + { + "left": "issue:43482", + "right": "issue:45538", + "accept": false, + "reason": "Qwen2.5-GGUF loading failure and CLIPTokenizer max-length default are unrelated." } ] }, "evaluator_result": { "accept": true, - "feedback": "Grounded overall. The summary stays conservative about heterogeneity, and the only accepted soft pair (issue:45216 vs issue:45357) is plausibly the same Qwen3.5 save_pretrained regression. No obvious overstatement in the rejected pairs." + "feedback": "The summary is grounded in the packet and stays conservative. The duplicate-like Qwen3.5 save_pretrained pair is the only clearly related one, and the rest of the soft-edge rejections are supported by the titles and issue scopes." }, "evaluator_used": true, "retried": false, @@ -4031,6 +4063,7 @@ "issue:42175", "issue:42796", "issue:42915", + "issue:43262", "issue:43296", "issue:43317", "issue:43377", @@ -4043,6 +4076,7 @@ "issue:43644", "issue:43646", "issue:43701", + "issue:43746", "issue:43818", "issue:43856", "issue:43873", @@ -4063,8 +4097,6 @@ "issue:45381", "issue:45479", "issue:45491", - "issue:45538", - "issue:45542", "issue:45584", "issue:45701", "issue:45812", @@ -4075,11 +4107,9 @@ "issue:46909" ], "soft_pairs": [ - "issue:42175|issue:45542", - "issue:43482|issue:45538", + "issue:44462|issue:45812", "issue:43296|issue:46821", "issue:44079|issue:44534", - "issue:44462|issue:45812", "issue:42175|issue:43296", "issue:43317|issue:43856", "issue:44509|issue:46909", @@ -4090,8 +4120,8 @@ "issue:40444|issue:45381", "issue:44265|issue:45072", "issue:43818|issue:45357", - "issue:45325|issue:45834", "issue:43493|issue:43994", + "issue:45325|issue:45834", "issue:45325|issue:45381", "issue:44829|issue:45479", "issue:43454|issue:44466", @@ -4100,223 +4130,215 @@ "issue:43701|issue:45325", "issue:43873|issue:45701", "issue:42796|issue:45381", - "issue:43606|issue:44265", "issue:41553|issue:43582", + "issue:43606|issue:44265", "issue:40990|issue:45584", - "issue:44479|issue:45823" + "issue:44479|issue:45823", + "issue:43262|issue:43746" ], - "prepared_review_unit_hash": "0b43b10bf5e45a019809079dfe98c5fb2553be3bad4cb92458a3d3d29db4f5df", + "prepared_review_unit_hash": "280237ba53cb1893388dab536f98af9ae75691fa408265f31a692778ed453c66", "component_budget": { "node_count": 514, "item_count": 514, "soft_pair_count": 1891, - "serialized_chars": 980926, - "estimated_input_tokens": 245232, - "estimated_eval_tokens": 490720 + "serialized_chars": 980895, + "estimated_input_tokens": 245224, + "estimated_eval_tokens": 490704 }, "budget": { "node_count": 48, "item_count": 48, - "soft_pair_count": 29, - "serialized_chars": 36481, - "estimated_input_tokens": 9121, - "estimated_eval_tokens": 18498 + "soft_pair_count": 28, + "serialized_chars": 36290, + "estimated_input_tokens": 9073, + "estimated_eval_tokens": 18402 }, "overflow_policy": "truncate_then_skip", "trimmed": false, "aggressively_trimmed": false, "split": true, "analyst_result": { - "summary": "Mostly a heterogeneous set of unrelated issues, with a few plausible duplicate pairs around tokenizer regressions and Qwen2.5-VL position-id handling. No pull requests are present.", - "confidence": 0.58, - "canonical_issue_reason": "Issue 44462 is the best anchor: it describes the broader AutoTokenizer regression, and 45812 looks like the same failure mode manifested on Granite models.", + "summary": "This cluster is heterogeneous: it mixes unrelated tokenizer, multimodal, loading/offload, training, and documentation regressions. There is no clear duplicate set or single canonical bug across the items.", + "confidence": 0.94, + "canonical_issue_reason": null, "canonical_pr_reason": null, - "best_issue_reason": "44462 is the clearest representative issue because it is specific, user-facing, and broad enough to subsume the matching Granite token-ID report.", + "best_issue_reason": "No issue is a strong global representative because the cluster spans many different models and code paths rather than one underlying bug.", "best_pr_reason": null, "soft_edge_verdicts": [ { - "left": "issue:42175", - "right": "issue:45542", - "accept": false, - "reason": "Both concern TensorFlow/backend packaging, but one is about the torch extra omitting TensorFlow and the other is a tf-backend runtime error when only tensorboard is installed; different bugs." - }, - { - "left": "issue:43482", - "right": "issue:45538", + "left": "issue:44462", + "right": "issue:45812", "accept": false, - "reason": "Different failure classes: Qwen2.5-GGUF loading vs CLIPTokenizer max_length handling." + "reason": "Both involve AutoTokenizer/tokenization, but one is about ignoring tokenizer.json while the other is a Granite-specific token-id regression; different failure modes and fixes." }, { "left": "issue:43296", "right": "issue:46821", "accept": false, - "reason": "Unrelated model families and errors; one is PaddleOCR-VL loading in vLLM, the other is a TimesFM window_size AttributeError." + "reason": "Completely different models and bugs: PaddleOCR-VL load failure vs TimesFM window_size AttributeError." }, { "left": "issue:44079", "right": "issue:44534", "accept": false, - "reason": "ModelOutput key assignment and non-persistent buffer junk are different code paths and symptoms." - }, - { - "left": "issue:44462", - "right": "issue:45812", - "accept": true, - "reason": "Same AutoTokenizer regression: ignoring tokenizer.json in the repo plausibly causes the wrong Granite token IDs reported in 45812." + "reason": "Different core issues: ModelOutput key assignment vs non-persistent buffer serialization in v5." }, { "left": "issue:42175", "right": "issue:43296", "accept": false, - "reason": "No shared bug beyond both being backend/model loading issues." + "reason": "Packaging/backend dependency problem vs a model-loading/runtime failure; not the same underlying bug." }, { "left": "issue:43317", "right": "issue:43856", "accept": false, - "reason": "Different problems: dequantized model offload load failure vs MoE training memory usage." + "reason": "Offload/load failure with dequantized models vs Qwen3 MoE training memory usage; distinct code paths." }, { "left": "issue:44509", "right": "issue:46909", "accept": false, - "reason": "Both are docs-related, but they concern different documents and different failures." + "reason": "Both are documentation-related, but they concern different docs content and are not the same concrete issue." }, { "left": "issue:43611", "right": "issue:43646", "accept": false, - "reason": "Both are v5 breakages, but one is base_model_prefix loading and the other is custom model initialization; not enough to treat as the same bug." + "reason": "Both mention Transformers v5 regressions, but one is about base_model_prefix loading and the other about custom model initialization." }, { "left": "issue:43377", "right": "issue:45491", "accept": false, - "reason": "Both involve batching/padding, but the models and root causes differ." + "reason": "Both involve batching/padding behavior, but they are different models with different symptoms and likely different root causes." }, { "left": "issue:43644", "right": "issue:44079", "accept": false, - "reason": "Non-persistent buffer junk and ModelOutput key assignment are unrelated regressions." + "reason": "Non-persistent buffer serialization and ModelOutput key handling are unrelated internals." }, { "left": "issue:43818", "right": "issue:46402", "accept": false, - "reason": "Different multimodal model bugs with no clear shared code path." + "reason": "Different model families and failures: Video-LLaVA architecture issue vs Sam3Model text-encoder weight loading." }, { "left": "issue:40444", "right": "issue:45381", "accept": false, - "reason": "Both mention Qwen2.5-VL, but one is finetuning with multiple images per prompt and the other is video position-id handling; distinct symptoms." + "reason": "Same model family, but one is finetuning with multiple images per prompt and the other is video position-id regression; not clearly the same bug." }, { "left": "issue:44265", "right": "issue:45072", "accept": false, - "reason": "torch.export/torch_compilable_check and bf16 dtype mismatches are different issues." + "reason": "torch.export/torch_compilable_check failure is unrelated to bfloat16 dtype mismatches." }, { "left": "issue:43818", "right": "issue:45357", "accept": false, - "reason": "Video-LLaVA temporal-attention/image-tower bug is unrelated to Qwen3.5 visual encoder key serialization." + "reason": "Both are multimodal model issues, but video_tower temporal attention/weight sharing is unrelated to incorrect visual encoder keys on save_pretrained." }, { - "left": "issue:45325", - "right": "issue:45834", + "left": "issue:43493", + "right": "issue:43994", "accept": false, - "reason": "Both are multimodal indexing bugs, but they affect different models and different input pipelines." + "reason": "Both are SigLIP2-related, but one reports HF-vs-JAX implementation discrepancy while the other reports bad AutoModel/pipeline outputs; not enough evidence they are the same defect." }, { - "left": "issue:43493", - "right": "issue:43994", - "accept": true, - "reason": "Same SigLIP2 regression: HF-vs-JAX discrepancy in 43493 matches the bad AutoModel/pipeline outputs reported in 43994." + "left": "issue:45325", + "right": "issue:45834", + "accept": false, + "reason": "Different models and errors: Qwen2.5-VL rope-index scaling vs Kosmos2.5 long-OCR index error." }, { "left": "issue:45325", "right": "issue:45381", - "accept": true, - "reason": "Both describe the same Qwen2.5-VL v5.3.0 position-id regression in the video/image rope-indexing path." + "accept": false, + "reason": "Both are Qwen2.5-VL position-id bugs, but one is still-image temporal scaling and the other is video vision_position_ids; related area, but not clearly one duplicate." }, { "left": "issue:44829", "right": "issue:45479", "accept": false, - "reason": "Different classification-loss issues: flash_attention_3 training degeneration vs num_labels=1 zero-loss behavior." + "reason": "Different classification regressions across different model families; one is FA3 training degeneration, the other is zero-loss single-label classification behavior." }, { "left": "issue:43454", "right": "issue:44466", "accept": false, - "reason": "AyaVision lm_head tying and lm_head.weight serialization are different bugs." + "reason": "AyaVision lm_head tying bug is unrelated to Granite lm_head serialization inconsistency." }, { "left": "issue:44291", "right": "issue:44877", "accept": false, - "reason": "init_empty_weights/_is_hf_initialized TypeError and strict config loading are separate failure modes." + "reason": "Both concern loading/config handling in v5, but one is init_empty_weights argument mismatch and the other is strict config rejection for granite_speech." }, { "left": "issue:42915", "right": "issue:43482", "accept": false, - "reason": "Different model-specific loading/configuration problems." + "reason": "Qwen3 MoE FP8 config failure and Qwen2.5-GGUF loading with transformers v5 are separate model/format compatibility issues." }, { "left": "issue:43701", "right": "issue:45325", "accept": false, - "reason": "Checkpoint resume key mismatch is unrelated to Qwen2.5-VL position ids." + "reason": "resume_from_checkpoint key mismatch is a training/checkpointing bug, not the same as Qwen2.5-VL rope index scaling." }, { "left": "issue:43873", "right": "issue:45701", "accept": false, - "reason": "Quantized offloading and tokenization version changes are unrelated." + "reason": "Quantization/offloading behavior and tokenization changes are different subsystems and different failure modes." }, { "left": "issue:42796", "right": "issue:45381", "accept": false, - "reason": "Tokenizer BPE-marker handling and Qwen2.5-VL video position ids are different bugs." + "reason": "Tokenizer BPE-marker handling and Qwen2.5-VL video position IDs are unrelated." }, { - "left": "issue:43606", - "right": "issue:44265", + "left": "issue:41553", + "right": "issue:43582", "accept": false, - "reason": "CPU-offload device mismatch and torch.export export failures are unrelated." + "reason": "AutoTokenizer error messaging for Voxtral and an Apple Silicon TypeError in caching_allocator_warmup are unrelated." }, { - "left": "issue:41553", - "right": "issue:43582", + "left": "issue:43606", + "right": "issue:44265", "accept": false, - "reason": "AutoTokenizer Voxtral error messaging and AppleSilicon caching_allocator_warmup TypeError are unrelated." + "reason": "CPU-offload device mismatch and torch.export export failure via torch_compilable_check are distinct problems." }, { "left": "issue:40990", "right": "issue:45584", "accept": false, - "reason": "High perplexity on gpt-oss and Whisper empty-transcription failure are unrelated." + "reason": "High perplexity on GPT-OSS with WikiText and Whisper empty-transcription generation failure are unrelated model behaviors." }, { "left": "issue:44479", "right": "issue:45823", "accept": false, - "reason": "Qwen v5.3.0 video-input regression and Gemma4 device mismatch are different model bugs." + "reason": "Both are multimodal regressions, but one is video-input behavior for Qwen variants and the other is a Gemma4 PLE device-mismatch issue." + }, + { + "left": "issue:43262", + "right": "issue:43746", + "accept": false, + "reason": "Audio processor sampling-rate defaulting and GraniteSpeech PEFT checkpoint loading are different code paths." } ] }, - "evaluator_result": { - "accept": true, - "feedback": "The summary is grounded in the packet and stays appropriately cautious overall. The rejected soft-edge pairs are conservative, and the accepted pairs are at least plausible duplicates based on the issue titles, especially the Qwen2.5-VL position-id regression pair. No PR pairs are present, so there is no mergeability concern." - }, - "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 }, @@ -4345,7 +4367,6 @@ "issue:42915", "issue:42947", "issue:43257", - "issue:43262", "issue:43317", "issue:43322", "issue:43377", @@ -4354,14 +4375,15 @@ "issue:43572", "issue:43606", "issue:43720", - "issue:43746", "issue:43749", "issue:43819", "issue:43828", "issue:43873", + "issue:43881", "issue:44186", "issue:44265", "issue:44387", + "issue:44488", "issue:44534", "issue:44568", "issue:44610", @@ -4383,7 +4405,6 @@ "issue:46821" ], "soft_pairs": [ - "issue:43262|issue:43746", "issue:43828|issue:46821", "issue:35707|issue:41720", "issue:34634|issue:35707", @@ -4406,236 +4427,237 @@ "issue:43454|issue:43720", "issue:43749|issue:44534", "issue:43317|issue:43873", - "issue:43322|issue:46050", "issue:36331|issue:42617", + "issue:43322|issue:46050", "issue:41720|issue:43322", "issue:36331|issue:42915", - "issue:41720|issue:44977", "issue:44877|issue:44991", + "issue:41720|issue:44977", "issue:44387|issue:45704", + "issue:43881|issue:44488", "issue:40990|issue:45701", "issue:41720|issue:42617" ], - "prepared_review_unit_hash": "1c26a14e6ad8827ac6206ab91f20fb1302f03e8cafe726385fb3760c15643c2c", + "prepared_review_unit_hash": "c5abfe8aa4c104088ee1f73e831d6e77c2d634f82345db6ff76cd0b7216c9fe5", "component_budget": { "node_count": 514, "item_count": 514, "soft_pair_count": 1891, - "serialized_chars": 980926, - "estimated_input_tokens": 245232, - "estimated_eval_tokens": 490720 + "serialized_chars": 980895, + "estimated_input_tokens": 245224, + "estimated_eval_tokens": 490704 }, "budget": { "node_count": 48, "item_count": 48, "soft_pair_count": 32, - "serialized_chars": 37382, - "estimated_input_tokens": 9346, - "estimated_eval_tokens": 18948 + "serialized_chars": 37224, + "estimated_input_tokens": 9306, + "estimated_eval_tokens": 18868 }, "overflow_policy": "truncate_then_skip", "trimmed": false, "aggressively_trimmed": false, "split": true, "analyst_result": { - "summary": "The cluster is heterogeneous: the soft-similarity pairs mostly connect unrelated regressions across audio, tokenization, quantization/offload, vision, and training code paths. I would not merge any pair as duplicates.", - "confidence": 0.82, - "canonical_issue_reason": "No strong canonical issue: there is no single underlying bug shared by the set, and the issues span multiple independent subsystems.", + "summary": "This is a loose grab bag of unrelated Transformers bugs spanning tokenization, multimodal processors, device mapping/offloading, quantization, training APIs, and generation. None of the soft pairs look like the same underlying defect, so no duplicate merges are justified.", + "confidence": 0.89, + "canonical_issue_reason": "Issue 43873 is the best umbrella representative: it is open, discusses a broad offloading/quantization regression, and sits closest to the dominant runtime-loading theme in this cluster.", "canonical_pr_reason": null, - "best_issue_reason": "issue:43873 is the broadest representative of the mixed set because it concerns offloading + quantization and has multiple inbound references, but it is still not a duplicate of the others.", + "best_issue_reason": "Issue 43873 is the strongest standalone issue for triage because it is concrete, high-signal, and broadly representative of the cluster\u2019s most common failure mode (offloading/quantization/device-map regressions).", "best_pr_reason": null, "soft_edge_verdicts": [ - { - "left": "issue:43262", - "right": "issue:43746", - "accept": false, - "reason": "Audio chat-template sampling-rate default bug vs GraniteSpeech PEFT checkpoint loading; different model families and failure modes." - }, { "left": "issue:43828", "right": "issue:46821", "accept": false, - "reason": "Different runtime errors in unrelated models: autocast dtype mismatch for Phi-tiny-MoE vs TimesFM window_size AttributeError." + "reason": "Autocast dtype mismatch vs TimesFM AttributeError: different model families and failure modes." }, { "left": "issue:35707", "right": "issue:41720", "accept": false, - "reason": "Both are generation/runtime issues, but one is inputs_embeds/past_key_values behavior and the other is auto device mapping causing CUDA assert; not the same bug." + "reason": "Progressive generation cache bug vs Qwen3 CUDA assert; unrelated generation vs device-mapping crash." }, { "left": "issue:34634", "right": "issue:35707", "accept": false, - "reason": "BarkProcessor voice preset failure is unrelated to progressive generation with inputs_embeds/past_key_values." + "reason": "Bark voice_preset processor bug vs progressive generation inputs_embeds/past_key_values bug." }, { "left": "issue:39692", "right": "issue:40444", "accept": false, - "reason": "SigLIP2 doc example errors vs Qwen2.5-VL finetuning with iterable multimodal batches; different surface area and fix." + "reason": "SigLIP2 doc/model mismatch vs Qwen2.5-VL multi-image iterable dataset failure; different workflows." }, { "left": "issue:41720", "right": "issue:45823", "accept": false, - "reason": "Both involve device_map/dispatch, but Qwen3 auto mapping CUDA assert and Gemma4 PLE forward device mismatch are different code paths and likely different causes." + "reason": "Both mention device mismatch, but these are different models and distinct code paths." }, { "left": "issue:44186", "right": "issue:44265", "accept": false, - "reason": "Tokenizer crash on NER/padding vs torch.export failing on torch_compilable_check; unrelated tokenizer vs export functionality." + "reason": "Tokenizer padding/NER crash vs torch.export failure on torch_compilable_check; unrelated." }, { "left": "issue:43606", "right": "issue:43819", "accept": false, - "reason": "CPU offload device mismatch for Bark-small is unrelated to DAC.from_latents missing STE parity." + "reason": "CPU-offload device mismatch vs DAC STE mismatch; unrelated." }, { "left": "issue:43454", "right": "issue:43550", "accept": false, - "reason": "AyaVision lm_head tying bug and Bamba torch.compile+SDPA failure are different model-specific defects." + "reason": "Weight tying bug vs torch.compile/SDPA failure; unrelated." }, { "left": "issue:43257", "right": "issue:45923", "accept": false, - "reason": "Qwen3 MoE weight conversion under accelerate+deepspeed vs missing supports_gradient_checkpointing flag on Nemotron remote-code variant; not the same issue." + "reason": "MoE weight conversion under accelerate/deepspeed vs missing gradient-checkpointing flag; unrelated." }, { "left": "issue:42796", "right": "issue:46209", "accept": false, - "reason": "Tokenizer BPE-marker handling bug vs FP8 inference failure on B200; different components entirely." + "reason": "Tokenizer BPE-marker handling vs FP8 inference on B200; unrelated." }, { "left": "issue:45823", "right": "issue:46821", "accept": false, - "reason": "Gemma4 device placement mismatch vs TimesFM window_size AttributeError; unrelated failures." + "reason": "Gemma4 device mismatch vs TimesFM window_size AttributeError; unrelated." }, { "left": "issue:44568", "right": "issue:45920", "accept": false, - "reason": "mdeberta-v3-base add_special_tokens regression is a specific tokenizer bug, while 45920 is a broader AutoTokenizer v5 ID mapping regression affecting multiple models." + "reason": "Both are tokenizer regressions, but one is missing BOS/EOS insertion and the other is wrong token IDs across models; too broad." }, { "left": "issue:43377", "right": "issue:44610", "accept": false, - "reason": "MIMI batched-vs-single padding-mask inconsistency is unrelated to OmDet-Turbo processor/model input-size mismatch." + "reason": "MIMI batching/padding-mask bug vs OmDet processor size mismatch; unrelated." }, { "left": "issue:43454", "right": "issue:45072", "accept": false, - "reason": "AyaVision weight-tying bug and dtype mismatches in SwitchTransformers/TimmWrapperModel are distinct issues." + "reason": "lm_head weight tying bug vs bf16 dtype mismatch; unrelated." }, { "left": "issue:42617", "right": "issue:42915", "accept": false, - "reason": "3d_parallel.py execution problem vs Qwen3Moe failing with FineGrainedFP8Config; different scripts and failure mechanisms." + "reason": "3d_parallel launch failure vs Qwen3MoE FineGrainedFP8Config failure; unrelated." }, { "left": "issue:40990", "right": "issue:42947", "accept": false, - "reason": "High perplexity on gpt-oss-20b is a model/eval issue, not the same as LoRA gradient-checkpointing ineffectiveness." + "reason": "High perplexity regression vs LoRA gradient-checkpointing ineffectiveness; unrelated." }, { "left": "issue:46305", "right": "issue:46315", "accept": false, - "reason": "MoonshineStreaming off-by-one mask logic and slow convert_tokens_to_ids regression are unrelated performance/logic bugs." + "reason": "Streaming mask off-by-one vs slow convert_tokens_to_ids regression; unrelated." }, { "left": "issue:43572", "right": "issue:45759", "accept": false, - "reason": "StableLmConfig missing pad_token_idx vs AutoModelForCausalLM not unwrapping text_config for composite Qwen multimodal configs; different config-layer bugs." + "reason": "Missing pad_token_idx config vs text_config unwrapping bug; unrelated." }, { "left": "issue:43377", "right": "issue:44898", "accept": false, - "reason": "MIMI padding-mask bug and Perceiver non-default resolution failure are separate model behaviors." + "reason": "MIMI encoder padding-mask bug vs Perceiver resolution handling; unrelated." }, { "left": "issue:43454", "right": "issue:43720", "accept": false, - "reason": "AyaVision lm_head tying issue is unrelated to BitNet packed-weight unpacking during accelerate loading." + "reason": "Weight tying bug vs BitNet packed-weight unpacking bug; unrelated." }, { "left": "issue:43749", "right": "issue:44534", "accept": false, - "reason": "FSDP CPU RAM efficient loading failure and non-persistent buffers filled with junk are both v5 regressions, but different subsystems and fixes." + "reason": "FSDP CPU-RAM loading bug vs non-persistent buffer junk; unrelated." }, { "left": "issue:43317", "right": "issue:43873", "accept": false, - "reason": "Both touch offload/loading, but dequantized model loading with device_map=auto and generic quantization offloading problems are not the same concrete bug." - }, - { - "left": "issue:43322", - "right": "issue:46050", - "accept": false, - "reason": "Llava Next load-time segfault vs sinq quantized save/load corruption; different models and failure points." + "reason": "Both touch offloading/quantization, but one is a specific dequantized-model load failure and the other is a broader offloading complaint; not the same concrete bug." }, { "left": "issue:36331", "right": "issue:42617", "accept": false, - "reason": "Trainer compute_loss signature breakage is unrelated to 3d_parallel.py runtime failure." + "reason": "CustomTrainer API break vs 3d_parallel launch issue; unrelated." + }, + { + "left": "issue:43322", + "right": "issue:46050", + "accept": false, + "reason": "Llava Next load segfault vs sinq save/load regression; unrelated." }, { "left": "issue:41720", "right": "issue:43322", "accept": false, - "reason": "Qwen3 auto device mapping CUDA assert and Llava Next segmentation fault are both runtime crashes but not the same code path." + "reason": "Qwen3 device-map assert vs Llava Next segfault; unrelated." }, { "left": "issue:36331", "right": "issue:42915", "accept": false, - "reason": "Trainer API incompatibility vs Qwen3Moe FP8 loading failure; unrelated training vs inference bugs." + "reason": "Trainer API break vs Qwen3MoE FP8 config failure; unrelated." }, { - "left": "issue:41720", - "right": "issue:44977", + "left": "issue:44877", + "right": "issue:44991", "accept": false, - "reason": "Auto device mapping CUDA assert and Qwen3.5 flash-attention generation issue are distinct model/runtime paths." + "reason": "Strict config loading vs tokenizer loading regression; unrelated." }, { - "left": "issue:44877", - "right": "issue:44991", + "left": "issue:41720", + "right": "issue:44977", "accept": false, - "reason": "Strict config rejecting granite_speech and tokenizer loading failure for EMBEDDIA/est-roberta are different config/tokenizer regressions." + "reason": "Same model family, but device-map crash and flash-attention generation bug are different failures." }, { "left": "issue:44387", "right": "issue:45704", "accept": false, - "reason": "Int4 quantization OOM in Transformers 5.x is unrelated to T5 using apex.FusedRMSNorm memory leak." + "reason": "Int4 quantization memory growth vs apex RMSNorm memory leak; unrelated." + }, + { + "left": "issue:43881", + "right": "issue:44488", + "accept": false, + "reason": "GLM-4V load failure vs sleng-bert tokenizer load failure; unrelated." }, { "left": "issue:40990", "right": "issue:45701", "accept": false, - "reason": "High perplexity on one model vs version-dependent tokenization changes; too broad to be duplicates." + "reason": "Perplexity regression vs tokenizer-version behavior change; related theme but not the same bug." }, { "left": "issue:41720", "right": "issue:42617", "accept": false, - "reason": "Different failures: auto device mapping CUDA assert versus a script execution issue." + "reason": "Qwen3 auto-device-map crash vs 3d_parallel launch issue; unrelated." } ] }, @@ -4671,6 +4693,7 @@ "issue:43317", "issue:43404", "issue:43450", + "issue:43526", "issue:43531", "issue:43550", "issue:43572", @@ -4681,12 +4704,12 @@ "issue:43720", "issue:43873", "issue:43881", + "issue:43883", "issue:44008", "issue:44186", "issue:44423", "issue:44451", "issue:44466", - "issue:44488", "issue:44610", "issue:44734", "issue:44756", @@ -4708,7 +4731,6 @@ "issue:47020" ], "soft_pairs": [ - "issue:43881|issue:44488", "issue:43450|issue:45701", "issue:44466|issue:45357", "issue:43065|issue:46402", @@ -4718,221 +4740,222 @@ "issue:46612|issue:46858", "issue:45685|issue:46821", "issue:43295|issue:45701", - "issue:43577|issue:43708", "issue:43550|issue:43720", + "issue:43577|issue:43708", "issue:43646|issue:46665", "issue:36010|issue:42617", "issue:44451|issue:47020", - "issue:44186|issue:44610", "issue:43550|issue:43606", + "issue:44186|issue:44610", "issue:42915|issue:43317", "issue:43404|issue:45357", - "issue:44610|issue:44898", "issue:46821|issue:46869", + "issue:44610|issue:44898", "issue:28530|issue:33666", "issue:41720|issue:45161", "issue:44734|issue:46858", - "issue:44008|issue:45823", "issue:43881|issue:44991", + "issue:44008|issue:45823", + "issue:43531|issue:43572", "issue:44423|issue:45406", - "issue:43531|issue:43572" + "issue:43526|issue:43883" ], - "prepared_review_unit_hash": "b875e7e62fc3bd535132fc36a0823c11442ad2a0136984835a2e0fcf840d4f98", + "prepared_review_unit_hash": "361e15b453e83b57c677efadba9a8cd38a288436a97fef29bc08d11258b07893", "component_budget": { "node_count": 514, "item_count": 514, "soft_pair_count": 1891, - "serialized_chars": 980926, - "estimated_input_tokens": 245232, - "estimated_eval_tokens": 490720 + "serialized_chars": 980895, + "estimated_input_tokens": 245224, + "estimated_eval_tokens": 490704 }, "budget": { - "node_count": 47, - "item_count": 47, + "node_count": 48, + "item_count": 48, "soft_pair_count": 28, - "serialized_chars": 35568, - "estimated_input_tokens": 8892, - "estimated_eval_tokens": 18040 + "serialized_chars": 36191, + "estimated_input_tokens": 9048, + "estimated_eval_tokens": 18352 }, "overflow_policy": "truncate_then_skip", "trimmed": false, "aggressively_trimmed": false, "split": true, "analyst_result": { - "summary": "This cluster is mostly heterogeneous issue reports that only share superficial keywords (loading, tokenizer, device_map, cache, etc.). I don't see any pair that is safely the same underlying bug, so all soft edges should be rejected.", - "confidence": 0.93, - "canonical_issue_reason": null, + "summary": "The proposed soft links are mostly broad topical similarity rather than true duplicates; none look like the same underlying bug/change. The set is dominated by unrelated Transformers regressions, and the closest pairs still differ in concrete failure mode and code path.", + "confidence": 0.67, + "canonical_issue_reason": "issue:43873 is the best representative open issue here: it is broad, active, and centered on offloading/quantization/device-map behavior, which is the closest recurring theme among the overbroad soft matches.", "canonical_pr_reason": null, - "best_issue_reason": "Issue 43708 is the clearest, most actionable standalone bug report in the set (trainer resume/max_steps regression), but the overall cluster is too diverse to treat it as a true canonical duplicate.", + "best_issue_reason": "issue:43873 is the strongest global issue candidate because it is the most general and reusable triage target among the set; several other items are much narrower or clearly unrelated.", "best_pr_reason": null, "soft_edge_verdicts": [ - { - "left": "issue:43881", - "right": "issue:44488", - "accept": false, - "reason": "Both are load failures, but for different models and likely different tokenizer/config paths." - }, { "left": "issue:43450", "right": "issue:45701", "accept": false, - "reason": "Video batch shape bug vs tokenizer/version-dependent tokenization change; unrelated failure modes." + "reason": "Video processor batched-shape bug vs tokenizer/version-change report; different subsystem and failure mode." }, { "left": "issue:44466", "right": "issue:45357", "accept": false, - "reason": "Both concern serialization, but one is tied lm_head weights and the other is visual encoder keys." + "reason": "Both mention serialization-like behavior, but one is lm_head/tied weights and the other is visual encoder key saving; not the same bug." }, { "left": "issue:43065", "right": "issue:46402", "accept": false, - "reason": "Different Sam3 components: dummy Conv2d in the pixel decoder vs missing text encoder weights." + "reason": "Both are Sam3-related, but one is a dummy Conv2d in the pixel decoder while the other is missing text encoder weight loading." }, { "left": "issue:26413", "right": "issue:28530", "accept": false, - "reason": "Trainer checkpoint resume iterator bug vs early-stopping metric lookup; distinct code paths." + "reason": "Checkpoint resume/epoch iterator failure is unrelated to early stopping metric lookup." }, { "left": "issue:43873", "right": "issue:44756", "accept": false, - "reason": "Quantization/offload behavior vs Strix Halo mmap OOM; not the same bug." + "reason": "Both involve memory/offload concerns, but one is quantization offloading behavior and the other is mmap OOM avoidance on Strix Halo." }, { "left": "issue:44186", "right": "issue:45072", "accept": false, - "reason": "Tokenizer padding/truncation crash vs bf16 dtype mismatch in different model wrappers." + "reason": "Tokenizer padding/truncation crash is unrelated to bf16 dtype mismatch in model wrappers." }, { "left": "issue:46612", "right": "issue:46858", "accept": false, - "reason": "Both touch cache logic, but one is beam-cache reorder across legacy models and the other is prefill seq-length tensor arithmetic." + "reason": "Beam-search cache reorder omission and static-cache seq_length tensor arithmetic are different cache bugs." }, { "left": "issue:45685", "right": "issue:46821", "accept": false, - "reason": "MPS histogram op limitation vs TimesFM window_size AttributeError; unrelated." + "reason": "MPS histogram/int issue in MoE is unrelated to a TimesFM window_size attribute error." }, { "left": "issue:43295", "right": "issue:45701", "accept": false, - "reason": "Processor/tokenizer API regression vs tokenization changes across versions; different concrete issues." + "reason": "Regression around processor.tokenizer access is not the same as a tokenizer output changing across versions." }, { - "left": "issue:43577", - "right": "issue:43708", + "left": "issue:43550", + "right": "issue:43720", "accept": false, - "reason": "Blip2 dtype propagation bug vs Trainer resume max_steps miscalculation; unrelated." + "reason": "One is a torch.compile/SDPA failure, the other is packed-weight unpacking during accelerate loading." }, { - "left": "issue:43550", - "right": "issue:43720", + "left": "issue:43577", + "right": "issue:43708", "accept": false, - "reason": "torch.compile/SDPA failure vs accelerate loading of packed BitNet weights; different code paths." + "reason": "Model dtype loading problem vs resume_from_checkpoint max_steps miscalculation; different code paths." }, { "left": "issue:43646", "right": "issue:46665", "accept": false, - "reason": "General 5.0 custom-model init breakage vs from_pretrained behavior inside Python classes; too broad and not the same bug." + "reason": "Custom model init breakage and from_pretrained-inside-class inconsistency are related only at a very broad API level." }, { "left": "issue:36010", "right": "issue:42617", "accept": false, - "reason": "GenerationMixin import error vs failure running 3d_parallel.py; unrelated." + "reason": "ImportError for GenerationMixin is unrelated to not being able to run 3d_parallel.py." }, { "left": "issue:44451", "right": "issue:47020", "accept": false, - "reason": "Both are tokenizer-loading problems, but different tokenizers and different root causes." + "reason": "Both are tokenizer-loading issues, but one is a specific model load failure and the other is a Unigram constructor type mismatch." }, { - "left": "issue:44186", - "right": "issue:44610", + "left": "issue:43550", + "right": "issue:43606", "accept": false, - "reason": "LayoutLMv2 tokenizer crash vs OmDet-Turbo processor image-size mismatch; not the same bug." + "reason": "Both involve loading/dispatch behavior, but one is compile+SDPA and the other is CPU-offload device mismatch." }, { - "left": "issue:43550", - "right": "issue:43606", + "left": "issue:44186", + "right": "issue:44610", "accept": false, - "reason": "Both involve inference/runtime issues, but SDPA compile failure and CPU-offload device mismatch are different problems." + "reason": "Tokenizer crash on batched padding/truncation is not the same as a processor producing the wrong image size." }, { "left": "issue:42915", "right": "issue:43317", "accept": false, - "reason": "Quantization config failure vs dequantized model load with gpu+cpu offload; separate offload paths." + "reason": "Qwen3Moe FP8 config failure and dequantized model offload load failure are distinct quantization bugs." }, { "left": "issue:43404", "right": "issue:45357", "accept": false, - "reason": "Weight tying bug vs incorrect save_pretrained key serialization; distinct model behaviors." + "reason": "Untied lm_head weights in Mistral3 are unrelated to incorrect visual encoder keys in Qwen3.5 save_pretrained." }, { - "left": "issue:44610", - "right": "issue:44898", + "left": "issue:46821", + "right": "issue:46869", "accept": false, - "reason": "Both are vision preprocessing issues, but one is fixed-size processor output and the other is Perceiver resolution handling." + "reason": "TimesFM window_size attribute error and YOLO multi-GPU device-map failure are different models and code paths." }, { - "left": "issue:46821", - "right": "issue:46869", + "left": "issue:44610", + "right": "issue:44898", "accept": false, - "reason": "TimesFM config AttributeError vs YOLO multi-GPU device-map failure; unrelated." + "reason": "Both concern image sizes, but one is an OmDet-Turbo processor/model mismatch and the other is Perceiver interpolation not working at non-default resolution." }, { "left": "issue:28530", "right": "issue:33666", "accept": false, - "reason": "Early stopping metric resolution vs Qwen2-VL multi-GPU training; different subsystems." + "reason": "Early stopping metric handling does not match Qwen2-VL multi-GPU training." }, { "left": "issue:41720", "right": "issue:45161", "accept": false, - "reason": "Auto device-map cuda assert vs tensor-parallel MoE issue; not the same concrete bug." + "reason": "Auto device mapping cuda assert on Qwen3 and TP not working on GPT-OSS MoE are only loosely related by distributed inference." }, { "left": "issue:44734", "right": "issue:46858", "accept": false, - "reason": "Serve KV-continuation indexing bug vs static-cache seq-length device tensor bug; related area, different defect." + "reason": "Both mention cache/prefill, but one is a responses/KV-continuation indexing crash and the other is a seq_length device-tensor arithmetic bug." + }, + { + "left": "issue:43881", + "right": "issue:44991", + "accept": false, + "reason": "glm-4v-9b load failure and tokenizer load failure for est-roberta are unrelated." }, { "left": "issue:44008", "right": "issue:45823", "accept": false, - "reason": "Gemma3n forward variable collision vs Gemma4 PLE device mismatch; different models and root causes." + "reason": "Gemma3n variable-name collision and Gemma4 PLE device mismatch are different model bugs." }, { - "left": "issue:43881", - "right": "issue:44991", + "left": "issue:43531", + "right": "issue:43572", "accept": false, - "reason": "Two tokenizer/model loading failures for different models; no shared underlying bug." + "reason": "Qwen3-MoE sliding_window behavior and missing pad_token_idx in StableLmConfig are unrelated regressions." }, { "left": "issue:44423", "right": "issue:45406", "accept": false, - "reason": "Both are serve/multimodal crashes, but one is a string-device conversion issue and the other is missing _tokenizer on Gemma4Processor." + "reason": "Both are serve/processor-related crashes, but the failing attribute and model families are different." }, { - "left": "issue:43531", - "right": "issue:43572", + "left": "issue:43526", + "right": "issue:43883", "accept": false, - "reason": "Qwen3-MoE sliding_window bug vs StableLmConfig pad_token_idx regression; unrelated." + "reason": "BeitImageProcessorFast reduce_labels and Molmo all_tied_weights_keys are unrelated issues in different components." } ] }, @@ -4971,9 +4994,9 @@ "issue:43316", "issue:43322", "issue:43329", + "issue:43388", "issue:43425", "issue:43452", - "issue:43526", "issue:43531", "issue:43540", "issue:43550", @@ -4983,7 +5006,6 @@ "issue:43746", "issue:43782", "issue:43873", - "issue:43883", "issue:44186", "issue:44361", "issue:44464", @@ -4993,6 +5015,7 @@ "issue:44671", "issue:44743", "issue:44913", + "issue:45059", "issue:45632", "issue:45684", "issue:45701", @@ -5006,12 +5029,11 @@ "issue:47145" ], "soft_pairs": [ - "issue:43526|issue:43883", "issue:42915|issue:46291", "issue:36331|issue:41720", "issue:44464|issue:46858", - "issue:45632|issue:45684", "issue:44671|issue:44743", + "issue:45632|issue:45684", "issue:44186|issue:44361", "issue:43425|issue:46821", "issue:28530|issue:30064", @@ -5025,199 +5047,206 @@ "issue:43540|issue:43782", "issue:43531|issue:45759", "issue:44492|issue:44509", + "issue:43452|issue:47020", "issue:43550|issue:43653", "issue:44534|issue:46665", - "issue:43452|issue:47020", "issue:41720|issue:43746", "issue:43606|issue:43720", "issue:42490|issue:43329", "issue:34567|issue:35707", - "issue:43531|issue:46821" + "issue:43531|issue:46821", + "issue:43388|issue:45059" ], - "prepared_review_unit_hash": "577c80c033ac739c704a4ef11e5f1d2be96ebf91869fdea3d80cea6b312ddcbb", + "prepared_review_unit_hash": "e36020859446628ec0d6c1c58e9e32085ec14d30a09ad6cb171eae0c896f3424", "component_budget": { "node_count": 514, "item_count": 514, "soft_pair_count": 1891, - "serialized_chars": 980926, - "estimated_input_tokens": 245232, - "estimated_eval_tokens": 490720 + "serialized_chars": 980895, + "estimated_input_tokens": 245224, + "estimated_eval_tokens": 490704 }, "budget": { "node_count": 48, "item_count": 48, "soft_pair_count": 27, - "serialized_chars": 35839, - "estimated_input_tokens": 8960, - "estimated_eval_tokens": 18176 + "serialized_chars": 35871, + "estimated_input_tokens": 8968, + "estimated_eval_tokens": 18192 }, "overflow_policy": "truncate_then_skip", "trimmed": false, "aggressively_trimmed": false, "split": true, "analyst_result": { - "summary": "This cluster is a broad false-positive mix of unrelated Transformers issues. The soft pairs share only superficial subsystem or wording overlap, not the same underlying bug or change. No PRs are present.", - "confidence": 0.98, - "canonical_issue_reason": "No single canonical issue stands out; the cluster is too heterogeneous, so collapsing it to one duplicate would be arbitrary.", + "summary": "This cluster is a loose collection of unrelated Transformers issues with a few superficial topical similarities (tokenizers, config loading, multimodal models, generation, offloading). None of the soft pairs look like the same underlying bug/change, so no duplicate merges are justified.", + "confidence": 0.88, + "canonical_issue_reason": "issue:43531 is the best anchor among a heterogeneous set because it is one of the more central nodes in the similarity graph and is a concrete, code-path-specific bug report; however, the cluster as a whole is not a strong duplicate set.", "canonical_pr_reason": null, - "best_issue_reason": "No issue is a strong global representative for the same reason; the closest thematic anchor is a generation/cache bug, but it still does not cover the rest of the cluster.", + "best_issue_reason": "issue:43531 is the least-bad representative for the small Qwen/config-related subcluster, but it should not be treated as a global duplicate umbrella for the rest of the items.", "best_pr_reason": null, "soft_edge_verdicts": [ - { - "left": "issue:43526", - "right": "issue:43883", - "accept": false, - "reason": "Different components: BEiT image processor label reduction vs Molmo loading/tied-weights attribute error." - }, { "left": "issue:42915", "right": "issue:46291", "accept": false, - "reason": "Different bugs: Qwen3Moe FP8 config failure vs kernels import/version requirement on pipelines." + "reason": "Different failures: Qwen3Moe FP8/load behavior vs pipelines import requiring kernels version/revision. Same broad release window, not the same bug." }, { "left": "issue:36331", "right": "issue:41720", "accept": false, - "reason": "Trainer loss-signature change and Qwen3 device-mapping CUDA assert are unrelated." + "reason": "Trainer custom compute_loss signature mismatch vs Qwen3 auto device-mapping CUDA assert. Different layers and root causes." }, { "left": "issue:44464", "right": "issue:46858", "accept": false, - "reason": "Both involve generation/compile/cache, but one is inconsistent chunked generation and the other is a device-tensor seq-length arithmetic bug; not the same fix." + "reason": "Both involve generation/cache, but one is inconsistent chunked generation under compiled forward and the other is a device-tensor seq_length arithmetic bug in static cache prefill. Not the same code path." }, { - "left": "issue:45632", - "right": "issue:45684", + "left": "issue:44671", + "right": "issue:44743", "accept": false, - "reason": "Cache path collision and read-only permission propagation are distinct save/load problems." + "reason": "CamemBERT masked LM prediction regression and Qwen recurrent-state reset are unrelated model-specific bugs." }, { - "left": "issue:44671", - "right": "issue:44743", + "left": "issue:45632", + "right": "issue:45684", "accept": false, - "reason": "CamemBERT masked-LM prediction regression vs Qwen3 recurrent-state reset under cache." + "reason": "One is cache-path collision for trust_remote_code; the other is read-only permissions propagation in save_pretrained/register_for_auto_class. Different filesystem issues." }, { "left": "issue:44186", "right": "issue:44361", "accept": false, - "reason": "Both are tokenizer crashes, but they affect different tokenizers with different task paths and failure modes." + "reason": "Both are tokenizer bugs, but one crashes on NER/batched padding and the other raises an AttributeError on tasks. Distinct failures and code paths." }, { "left": "issue:43425", "right": "issue:46821", "accept": false, - "reason": "Torch 2.10 compatibility and TimesFM window_size AttributeError are unrelated." + "reason": "Torch 2.10 compatibility issue vs TimesFM window_size AttributeError. Different APIs and failure modes." }, { "left": "issue:28530", "right": "issue:30064", "accept": false, - "reason": "Early-stopping metric selection and void segmentation-map processing are different subsystems." + "reason": "Early stopping metric selection vs image processor handling of void segmentation maps. No shared underlying bug." }, { "left": "issue:43316", "right": "issue:44913", "accept": false, - "reason": "Gemma3TextConfig API mismatch is unrelated to GPTNeoX rotary_pct persistence on reload." + "reason": "Gemma3TextConfig API discrepancy vs GPTNeoXConfig rotary_pct reload defaulting. Separate config serialization/deserialization bugs." }, { "left": "issue:45701", "right": "issue:47145", "accept": false, - "reason": "Tokenization-version drift and FP8 weight_scale loading corruption are different classes of bugs." + "reason": "Tokenizer behavior changes across versions vs FP8 compressed-tensors checkpoints missing weight_scale application. Different subsystems." }, { "left": "issue:42915", "right": "issue:43322", "accept": false, - "reason": "Qwen3Moe fine-grained FP8 config and Llava Next segmentation fault do not point to the same code path." + "reason": "Qwen3Moe FineGrainedFP8Config failure vs Llava Next segmentation fault. Different models and failure mechanisms." }, { "left": "issue:33357", "right": "issue:35707", "accept": false, - "reason": "MacOS bus error with CLIP loading is unrelated to progressive generation with inputs_embeds/past_key_values." + "reason": "MacOS CLIP bus error vs progressive generation with inputs_embeds/past_key_values. Unrelated runtime crash vs generation semantics." }, { "left": "issue:29159", "right": "issue:36032", "accept": false, - "reason": "Both concern tokenizers, but one is slow/fast inconsistency and the other is a T5 add_special_tokens constructor conflict." + "reason": "Slow/fast tokenizer inconsistency vs T5Tokenizer add_special_tokens name conflict. Both tokenizer-related, but not the same defect." }, { "left": "issue:30064", "right": "issue:34634", "accept": false, - "reason": "Image processor void-segmentation handling and BarkProcessor voice_preset are unrelated." + "reason": "Void segmentation-map processing vs BarkProcessor voice_preset handling. Different processor code paths and symptoms." }, { "left": "issue:43873", "right": "issue:45704", "accept": false, - "reason": "Quantization/offloading behavior and an apex FusedRMSNorm memory leak are different problems." + "reason": "Quantization offloading issue vs T5 apex FusedRMSNorm memory leak. Different components and root causes." }, { "left": "issue:43540", "right": "issue:43782", "accept": false, - "reason": "Qwen3OmniMoe video-input processing and Qwen3VL weight_only loading are separate code paths." + "reason": "Qwen3OmniMoe video input ValueError vs Qwen3VL weight_only load error. Same family, but different failing paths." }, { "left": "issue:43531", "right": "issue:45759", "accept": false, - "reason": "Both are Qwen3 family issues, but sliding_window behavior and text_config unwrapping are not the same bug." + "reason": "Both touch Qwen3 config handling, but one is a sliding_window bug and the other is from_config not unwrapping text_config for multimodal configs. Not the same bug." }, { "left": "issue:44492", "right": "issue:44509", "accept": false, - "reason": "A cache-strategies typo and stale docs for removed pipelines are not duplicates." + "reason": "Both are docs-related, but they cover different removed/deprecated pipeline tasks. Not duplicates." + }, + { + "left": "issue:43452", + "right": "issue:47020", + "accept": false, + "reason": "gguf_file loading issue vs XLMRobertaTokenizer passing a dict to Unigram. Different loading stacks and errors." }, { "left": "issue:43550", "right": "issue:43653", "accept": false, - "reason": "torch.compile+SDPA failure and BigBirdTokenizer special-token registration are unrelated." + "reason": "Bamba torch.compile/SDPA failure vs BigBirdTokenizer special-token registration bug. Unrelated model/runtime vs tokenizer issue." }, { "left": "issue:44534", "right": "issue:46665", "accept": false, - "reason": "Non-persistent buffer corruption in v5 and from_pretrained-in-class behavior are different regressions." - }, - { - "left": "issue:43452", - "right": "issue:47020", - "accept": false, - "reason": "gguf_file loading breakage and XLMRobertaTokenizer Unigram argument typing are unrelated tokenizer/model-loading bugs." + "reason": "Non-persistent buffers filled with junk in v5 vs inconsistent from_pretrained behavior inside Python classes. Different mechanisms." }, { "left": "issue:41720", "right": "issue:43746", "accept": false, - "reason": "Qwen3 auto device-mapping CUDA assert and GraniteSpeech PEFT local-checkpoint loading are distinct issues." + "reason": "Qwen3 auto device mapping CUDA assert vs GraniteSpeech PEFT local-checkpoint loading. Different model families and failure modes." }, { "left": "issue:43606", "right": "issue:43720", "accept": false, - "reason": "Both involve loading/offloading, but device mismatch on CPU offload and packed-weight unpacking are not the same underlying bug." + "reason": "CPU offload device mismatch vs BitNet packed weights not unpacked during accelerate loading. Related loading context, but distinct bugs." }, { "left": "issue:42490", "right": "issue:43329", "accept": false, - "reason": "Trailing-slash import warning and undefined video-token-count variables are unrelated." + "reason": "Trailing-slash local path import warning vs undefined video-branch helper names in multimodal token counting. Different concerns." }, { "left": "issue:34567", "right": "issue:35707", "accept": false, - "reason": "Trainer token-count tracking and progressive generation cache handling are separate areas." + "reason": "TrainerState token counter not updating vs generation with inputs_embeds and past_key_values. No shared underlying bug." + }, + { + "left": "issue:43531", + "right": "issue:46821", + "accept": false, + "reason": "Both mention window-related settings, but one is Qwen3-MoE sliding_window behavior and the other is TimesFM window_size AttributeError. Different models and APIs." + }, + { + "left": "issue:43388", + "right": "issue:45059", + "accept": false, + "reason": "gather_for_metrics batch truncation bug vs SAM3 text+bounding-box odd behavior. Different evaluation and multimodal paths." } ] }, @@ -5246,7 +5275,6 @@ "issue:41720", "issue:42886", "issue:43381", - "issue:43388", "issue:43421", "issue:43526", "issue:43531", @@ -5263,6 +5291,8 @@ "issue:44077", "issue:44368", "issue:44451", + "issue:44514", + "issue:44545", "issue:44589", "issue:44610", "issue:44625", @@ -5273,7 +5303,6 @@ "issue:44987", "issue:44991", "issue:45042", - "issue:45059", "issue:45072", "issue:45357", "issue:45406", @@ -5291,14 +5320,13 @@ "issue:46858" ], "soft_pairs": [ - "issue:43388|issue:45059", - "issue:34634|issue:41720", "issue:45406|issue:45464", + "issue:34634|issue:41720", "issue:43606|issue:43653", "issue:43531|issue:44589", "issue:45759|issue:45859", - "issue:43540|issue:46858", "issue:44451|issue:44991", + "issue:43540|issue:46858", "issue:44077|issue:44625", "issue:45357|issue:46402", "issue:43644|issue:46665", @@ -5306,229 +5334,230 @@ "issue:43526|issue:43881", "issue:44062|issue:44451", "issue:44898|issue:45072", - "issue:44625|issue:45479", "issue:43381|issue:43844", + "issue:44625|issue:45479", "issue:39692|issue:44610", "issue:41720|issue:45704", - "issue:43746|issue:46050", "issue:44368|issue:45357", + "issue:43746|issue:46050", "issue:43381|issue:45704", - "issue:46655|issue:46693", "issue:44918|issue:44936", + "issue:46655|issue:46693", "issue:43421|issue:43927", - "issue:45704|issue:45923", "issue:45759|issue:45923", + "issue:45704|issue:45923", "issue:43746|issue:43873", - "issue:42886|issue:44843" + "issue:42886|issue:44843", + "issue:44514|issue:44545" ], - "prepared_review_unit_hash": "ed1a3cec6e954b341ce4b479e4c283d1b4c4b192a280d7119258237f2503e2f2", + "prepared_review_unit_hash": "fd654b47f520ad7c350704b7033611316dc6b87a012cfc57967f54a4e7eb8c5b", "component_budget": { "node_count": 514, "item_count": 514, "soft_pair_count": 1891, - "serialized_chars": 980926, - "estimated_input_tokens": 245232, - "estimated_eval_tokens": 490720 + "serialized_chars": 980895, + "estimated_input_tokens": 245224, + "estimated_eval_tokens": 490704 }, "budget": { "node_count": 48, "item_count": 48, "soft_pair_count": 29, - "serialized_chars": 36647, - "estimated_input_tokens": 9162, - "estimated_eval_tokens": 18580 + "serialized_chars": 36606, + "estimated_input_tokens": 9152, + "estimated_eval_tokens": 18560 }, "overflow_policy": "truncate_then_skip", "trimmed": false, "aggressively_trimmed": false, "split": true, "analyst_result": { - "summary": "This cluster is mostly heterogeneous and does not look like a single duplicate set. The only strong duplicate-like match is the offline tokenizer loading regression pair (42886/44843); the rest are distinct model-, processor-, training-, or config-specific bugs.", - "confidence": 0.91, - "canonical_issue_reason": "issue:44843 is the clearest root-cause report in the cluster for the one genuine duplicate-like theme: it explains the offline tokenizer failure and points to the unconditional model_info() call in _patch_mistral_regex.", + "summary": "Mostly false-positive similarity edges across unrelated bugs. Two pairs look like true duplicates: the HF_HUB_OFFLINE tokenizer/cache regression (42886\u219444843) and the exact Qwen2_5_VLProcessor batched padding=False crash (44514\u219444545). Everything else is only broad subsystem/model overlap, not the same underlying issue.", + "confidence": 0.79, + "canonical_issue_reason": "No single canonical issue cleanly represents the whole set because the soft edges split into at least two separate duplicate pairs plus many unrelated false positives.", "canonical_pr_reason": null, - "best_issue_reason": "issue:44843 is the best issue anchor because it is explicit about the failing code path, generalizes the offline-mode tokenizer regression, and cleanly subsumes issue:42886.", + "best_issue_reason": "No global best issue across the full set; the items are heterogeneous, so any single representative would be misleading.", "best_pr_reason": null, "soft_edge_verdicts": [ { - "left": "issue:43388", - "right": "issue:45059", + "left": "issue:45406", + "right": "issue:45464", "accept": false, - "reason": "Different bugs: one is gather_for_metrics dropping label elements in the last batch; the other is a SAM3 text+box prompting issue." + "reason": "Different bugs: Gemma4Processor missing _tokenizer vs Qwen3.5 chat/streaming failure." }, { "left": "issue:34634", "right": "issue:41720", "accept": false, - "reason": "Unrelated subsystems and failures: BarkProcessor voice_preset vs Qwen3 auto device mapping cudaErrorAssert." - }, - { - "left": "issue:45406", - "right": "issue:45464", - "accept": false, - "reason": "Different models and code paths: Gemma4Processor missing _tokenizer vs Qwen3.5 streaming inference failure." + "reason": "Bark voice_preset behavior and Qwen3 auto device mapping cudaErrorAssert are unrelated." }, { "left": "issue:43606", "right": "issue:43653", "accept": false, - "reason": "One is Bark CPU-offload device mismatch; the other is a BigBirdTokenizer special-token registration/empty decode issue." + "reason": "CPU offload device mismatch in Bark-small vs BigBirdTokenizer special-token decode issue." }, { "left": "issue:43531", "right": "issue:44589", "accept": false, - "reason": "Different root causes: Qwen3-MoE sliding_window behavior vs a Float8 storage lookup/type error." + "reason": "Qwen3-MoE sliding_window bug and Float8 storage lookup error are different code paths." }, { "left": "issue:45759", "right": "issue:45859", "accept": false, - "reason": "Both are Qwen3-related, but the bugs differ: composite config unwrapping vs CPU-offload incompatibility in rotary embeddings." + "reason": "Qwen multimodal config unwrapping vs Qwen3.5 rotary embedding CPU-offload incompatibility." }, { - "left": "issue:43540", - "right": "issue:46858", + "left": "issue:44451", + "right": "issue:44991", "accept": false, - "reason": "Different problems: Qwen3OmniMoe video-input validation vs static-cache slice arithmetic on non-CUDA backends." + "reason": "Both are load failures, but for different models and different root causes." }, { - "left": "issue:44451", - "right": "issue:44991", + "left": "issue:43540", + "right": "issue:46858", "accept": false, - "reason": "Both are model-loading failures, but the affected models and likely causes differ; not the same underlying bug." + "reason": "Qwen3OmniMoe video ValueError vs static-cache seq_length tensor arithmetic on non-CUDA backends." }, { "left": "issue:44077", "right": "issue:44625", "accept": false, - "reason": "PatchTST post_init validation and Qwen3.5 num_labels propagation are unrelated issues." + "reason": "PatchTSTMixer post_init policy and Qwen3.5 num_labels propagation are unrelated." }, { "left": "issue:45357", "right": "issue:46402", "accept": false, - "reason": "Both involve multimodal checkpoint loading/saving, but they target different models and distinct state-dict loading problems." + "reason": "Qwen3.5 visual encoder save_pretrained keys vs Sam3Model text-encoder loading failure." }, { "left": "issue:43644", "right": "issue:46665", "accept": false, - "reason": "Non-persistent buffer junk in Transformers 5.0 is not the same as inconsistent from_pretrained behavior inside Python classes." + "reason": "Non-persistent buffer junk and from_pretrained-in-Python-class inconsistency are distinct issues." }, { "left": "issue:44987", "right": "issue:45042", "accept": false, - "reason": "Different bugs: loading physical-intelligence/fast vs PIL image processors incorrectly requiring torchvision." + "reason": "Loading physical-intelligence/fast and PIL processors requiring torchvision are different bugs." }, { "left": "issue:43526", "right": "issue:43881", "accept": false, - "reason": "BeitImageProcessorFast reduce_labels truncation is unrelated to glm-4v-9b loading failure." + "reason": "BeitImageProcessorFast reduce_labels bug vs glm-4v-9b loading failure." }, { "left": "issue:44062", "right": "issue:44451", "accept": false, - "reason": "The AddedToken special-argument error is a tokenizer construction bug, but not the same as ScandiBERT loading failure." + "reason": "AddedToken special-kwarg TypeError is unrelated to ScandiBERT loading failure." }, { "left": "issue:44898", "right": "issue:45072", "accept": false, - "reason": "Perceiver resolution handling and bfloat16 dtype mismatches are separate failure modes." + "reason": "Perceiver interpolation failure and SwitchTransformers/TimmWrapper dtype mismatch are not the same bug." }, { - "left": "issue:44625", - "right": "issue:45479", + "left": "issue:43381", + "right": "issue:43844", "accept": false, - "reason": "One is config propagation of num_labels; the other is a degenerate loss bug when num_labels=1. Related theme, but not the same bug." + "reason": "Eval-mode gradient checkpointing restriction vs ZeRO-3 gradient blow-up are different problems." }, { - "left": "issue:43381", - "right": "issue:43844", + "left": "issue:44625", + "right": "issue:45479", "accept": false, - "reason": "Gradient checkpointing in eval mode and ZeRO-3 gradient growth are different training/runtime problems." + "reason": "Qwen3.5 num_labels propagation and single-label zero-loss behavior are separate classification bugs." }, { "left": "issue:39692", "right": "issue:44610", "accept": false, - "reason": "Documentation example/model-processor mismatch and quantization failure is not the same as OmDet-Turbo image-size mismatch." + "reason": "Documentation model/processor mismatch is different from OmDet-Turbo processor/model size mismatch." }, { "left": "issue:41720", "right": "issue:45704", "accept": false, - "reason": "Qwen3 CUDA assert on A800 and T5 apex RMSNorm memory leak are unrelated." + "reason": "Qwen3 CUDA assert and T5 apex RMSNorm memory leak are unrelated." }, { - "left": "issue:43746", - "right": "issue:46050", + "left": "issue:44368", + "right": "issue:45357", "accept": false, - "reason": "Both involve save/load after model modification, but PEFT adapter local checkpoints and sinq quantization persistence are distinct bugs." + "reason": "LoRA tie_word_embeddings warning is not the same as incorrect visual encoder keys on save." }, { - "left": "issue:44368", - "right": "issue:45357", + "left": "issue:43746", + "right": "issue:46050", "accept": false, - "reason": "A warning about tie_word_embeddings in Qwen3.5 LoRA fine-tuning is unrelated to incorrect visual encoder keys on save_pretrained." + "reason": "PEFT local-checkpoint loading and sinq save/load breakage are different checkpointing bugs." }, { "left": "issue:43381", "right": "issue:45704", "accept": false, - "reason": "No shared failure mechanism: checkpointing/eval-mode restriction vs apex RMSNorm memory leak." + "reason": "Gradient checkpointing eval-mode issue vs T5 apex memory leak are unrelated." }, { - "left": "issue:46655", - "right": "issue:46693", + "left": "issue:44918", + "right": "issue:44936", "accept": false, - "reason": "SmolVLM video padding bug and FA2 long-context performance regression are different issues." + "reason": "TRL SFT embedding unpacking and trainer.evaluate-after-train failure are different trainer bugs." }, { - "left": "issue:44918", - "right": "issue:44936", + "left": "issue:46655", + "right": "issue:46693", "accept": false, - "reason": "Both mention trainer flows, but one is an embedding unpacking failure in SFT training and the other is evaluate() failing after train()." + "reason": "Video padding calculation bug vs FA2 long-context performance regression." }, { "left": "issue:43421", "right": "issue:43927", "accept": false, - "reason": "Runtime post-processor updates for special tokens and config persistence of custom token IDs are related but not the same bug." + "reason": "Runtime special-token post-processor updates and config token-ID persistence are separate tokenizer issues." }, { - "left": "issue:45704", + "left": "issue:45759", "right": "issue:45923", "accept": false, - "reason": "T5 RMSNorm memory leak and missing supports_gradient_checkpointing on a trust_remote_code variant are unrelated." + "reason": "Config unwrapping for composite Qwen models vs missing supports_gradient_checkpointing flag." }, { - "left": "issue:45759", + "left": "issue:45704", "right": "issue:45923", "accept": false, - "reason": "Composite Qwen config unwrapping and missing gradient-checkpointing support do not share a concrete code-path bug." + "reason": "Apex RMSNorm memory leak and missing gradient-checkpointing flag are unrelated." }, { "left": "issue:43746", "right": "issue:43873", "accept": false, - "reason": "PEFT checkpoint loading and quantization/offloading behavior are different code paths and failure modes." + "reason": "PEFT local checkpoint loading vs quantization/offloading behavior with quantization." }, { "left": "issue:42886", "right": "issue:44843", "accept": true, - "reason": "These describe the same offline tokenizer regression: loading from cache fails when HF_HUB_OFFLINE=1, and 44843 identifies the root cause as an unconditional model_info() call in _patch_mistral_regex." + "reason": "Same HF_HUB_OFFLINE regression: tokenizer cache load fails because AutoTokenizer unconditionally hits model_info() in _patch_mistral_regex." + }, + { + "left": "issue:44514", + "right": "issue:44545", + "accept": true, + "reason": "Verbatim duplicate of the same Qwen2_5_VLProcessor.apply_chat_template crash with padding=False." } ] }, "evaluator_result": { "accept": true, - "feedback": "The summary is grounded in the packet and the conservative pattern holds: the cluster is mostly heterogeneous, and the 42886/44843 pair is the only clearly supported duplicate-like match from the titles. The other soft-edge rejections are cautious and consistent with the packet evidence." + "feedback": "Grounded and conservative overall. The summary correctly treats the set as heterogeneous and only upgrades the two clearly duplicated pairs supported by the titles: 42886/44843 and 44514/44545. The remaining soft-edge rejections stay appropriately cautious and do not overstate evidence." }, "evaluator_used": true, "retried": false, @@ -5553,6 +5582,7 @@ "issue:36010", "issue:36331", "issue:39692", + "issue:41720", "issue:43295", "issue:43421", "issue:43493", @@ -5575,11 +5605,11 @@ "issue:44361", "issue:44464", "issue:44488", - "issue:44514", - "issue:44545", "issue:44625", + "issue:44661", "issue:44671", "issue:44857", + "issue:44991", "issue:44995", "issue:45081", "issue:45356", @@ -5598,14 +5628,13 @@ "issue:47050" ], "soft_pairs": [ - "issue:44514|issue:44545", "issue:30064|issue:36010", "issue:39692|issue:43493", "issue:43421|issue:43653", "issue:43994|issue:44671", "issue:44464|issue:47050", - "issue:43782|issue:45381", "issue:46032|issue:46463", + "issue:43782|issue:45381", "issue:44857|issue:47050", "issue:45081|issue:45356", "issue:43421|issue:44186", @@ -5616,237 +5645,242 @@ "issue:43720|issue:43819", "issue:43295|issue:45834", "issue:43653|issue:43720", - "issue:46829|issue:46832", "issue:43550|issue:43927", - "issue:43550|issue:44361", "issue:43596|issue:44220", - "issue:44625|issue:47050", - "issue:43540|issue:46294", + "issue:46829|issue:46832", + "issue:43550|issue:44361", "issue:43531|issue:43643", + "issue:44625|issue:47050", "issue:44062|issue:44488", + "issue:43540|issue:46294", "issue:45381|issue:46693", "issue:45987|issue:46829", "issue:43611|issue:46665", - "issue:30064|issue:36331" + "issue:30064|issue:36331", + "issue:41720|issue:44991", + "issue:44220|issue:44661" ], - "prepared_review_unit_hash": "b167f766cff7736623ee5a38c1568672a1959a2b3091b2c77cdf74472a773af2", + "prepared_review_unit_hash": "68fcff3788f1eede3b1a9c37eef128abf245bf7566349074708238ef6cb0c0c0", "component_budget": { "node_count": 514, "item_count": 514, "soft_pair_count": 1891, - "serialized_chars": 980926, - "estimated_input_tokens": 245232, - "estimated_eval_tokens": 490720 + "serialized_chars": 980895, + "estimated_input_tokens": 245224, + "estimated_eval_tokens": 490704 }, "budget": { - "node_count": 47, - "item_count": 47, - "soft_pair_count": 30, - "serialized_chars": 36421, - "estimated_input_tokens": 9106, - "estimated_eval_tokens": 18468 + "node_count": 48, + "item_count": 48, + "soft_pair_count": 31, + "serialized_chars": 37263, + "estimated_input_tokens": 9316, + "estimated_eval_tokens": 18888 }, "overflow_policy": "truncate_then_skip", "trimmed": false, "aggressively_trimmed": false, "split": true, "analyst_result": { - "summary": "This cluster is mostly a set of unrelated HF issues that only superficially overlap by model family or error type. The only clear duplicate is the two Qwen2_5_VLProcessor apply_chat_template crashes with padding=False.", - "confidence": 0.95, - "canonical_issue_reason": "issue:44514 is the clearest canonical issue: it describes the exact batched-input padding=False crash, and issue:44545 is an obvious duplicate with nearly identical title/body.", + "summary": "These items are heterogeneous issues across unrelated code paths. The suggested soft edges mostly share broad keywords or subsystem names, but not the same underlying bug or fix.", + "confidence": 0.96, + "canonical_issue_reason": "No canonical issue: the set spans unrelated tokenizer, vision, training, quantization, cache, and loading regressions, so there is no single representative duplicate.", "canonical_pr_reason": null, - "best_issue_reason": "issue:44514 is the best representative because it is the earlier report of the exact failure mode and matches the duplicate pair most precisely.", + "best_issue_reason": "No safe global best issue. Even the closest-looking pairs are only superficially related and would not be merged as the same bug.", "best_pr_reason": null, "soft_edge_verdicts": [ - { - "left": "issue:44514", - "right": "issue:44545", - "accept": true, - "reason": "Same underlying bug: identical Qwen2_5_VLProcessor.apply_chat_template crash on batched input when padding=False." - }, { "left": "issue:30064", "right": "issue:36010", "accept": false, - "reason": "Different subsystems and failures: image processor segmentation maps vs GenerationMixin import error." + "reason": "ImageProcessor segmentation-map handling and a GenerationMixin import error are unrelated subsystems and failure modes." }, { "left": "issue:39692", "right": "issue:43493", "accept": false, - "reason": "Both concern SigLIP2, but one is a docs/example mismatch and quantization failure; the other is an implementation discrepancy vs JAX." + "reason": "Both mention SigLIP2, but one is a docs/example mismatch and quantization failure while the other is an implementation discrepancy vs JAX." }, { "left": "issue:43421", "right": "issue:43653", "accept": false, - "reason": "Different tokenizer bugs: runtime special-token updates vs BigBirdTokenizer mask token decode/registration." + "reason": "Special-token post-processor updates at runtime and BigBirdTokenizer mask-token registration causing empty decode are different tokenizer bugs." }, { "left": "issue:43994", "right": "issue:44671", "accept": false, - "reason": "Both are model-quality regressions, but they affect different models and different failure modes." + "reason": "Different model families and symptoms: SigLIP2 nonsensical outputs vs CamemBERT masked-LM prediction regression." }, { "left": "issue:44464", "right": "issue:47050", "accept": false, - "reason": "Unrelated issues: compiled chunked generation inconsistency vs ONNX fp16 dtype mismatch in attention." + "reason": "Compiled forward generation inconsistency and fp16 ONNX attention dtype mismatch are different execution paths and root causes." }, { - "left": "issue:43782", - "right": "issue:45381", + "left": "issue:46032", + "right": "issue:46463", "accept": false, - "reason": "Different Qwen3 VL issues: weight_only load error vs wrong video vision_position_ids." + "reason": "Mamba2Mixer cache misuse with seq_len > 1 is unrelated to T5 decoder input validation error messaging." }, { - "left": "issue:46032", - "right": "issue:46463", + "left": "issue:43782", + "right": "issue:45381", "accept": false, - "reason": "Different code paths: Mamba2Mixer cache correctness vs T5 decoder-input validation error message." + "reason": "Both are Qwen vision-language issues, but one is a weight_only loading failure and the other is a video vision_position_ids regression; not the same bug." }, { "left": "issue:44857", "right": "issue:47050", "accept": false, - "reason": "Different models and bugs: LwDetr float16 AMP crash vs CanineModel ONNX export dtype mismatch." + "reason": "AMP/CUDA crash in LwDetrImageLoss is unrelated to Canine ONNX export dtype mismatch in attention." }, { "left": "issue:45081", "right": "issue:45356", "accept": false, - "reason": "Both involve tokenizer regex handling, but one is a backend_tokenizer AttributeError and the other is a Kimi-K2.5 regression with codec handling/warning behavior." + "reason": "Tokenizer backend attribute error in Mistral regex patching and Kimi tokenizer codec-handling regression are separate tokenizer issues." }, { "left": "issue:43421", "right": "issue:44186", "accept": false, - "reason": "Both are tokenizer-related, but the first is about post-processor updates and the second about LayoutLMv2 NER padding/truncation crashes." + "reason": "Runtime special-token post-processing and LayoutLMv2Tokenizer crashes on NER padding/truncation are different code paths and symptoms." }, { "left": "issue:45356", "right": "issue:45907", "accept": false, - "reason": "Tokenizer regression vs repo file listing API forwarding; no shared underlying bug." + "reason": "A tokenizer regression is unrelated to list_repo_files not forwarding revision in CI." }, { "left": "issue:44995", "right": "issue:46858", "accept": false, - "reason": "Both mention cache/indexing, but one is GlmMoeDsa stale indexer cache and the other is StaticCache sequence length/device-tensor behavior." + "reason": "Both involve cache length, but one is a stale indexer cache on second forward and the other is a device-tensor return breaking static-cache arithmetic under compile." }, { "left": "issue:39692", "right": "issue:43873", "accept": false, - "reason": "SigLIP2 docs/example issue vs generic offloading-with-quantization problem; not the same bug." + "reason": "SigLIP2 docs/quantization issues are not the same as generic offloading not working with quantization." }, { "left": "issue:39692", "right": "issue:47050", "accept": false, - "reason": "Different model families and failure modes; no shared concrete code path." + "reason": "A SigLIP2 documentation/quantization report is not the same as a Canine fp16 ONNX attention dtype mismatch." }, { "left": "issue:43720", "right": "issue:43819", "accept": false, - "reason": "Different architectures and bugs: BitNet packed weights loading vs DAC latent reconstruction/STE mismatch." + "reason": "BitNet packed-weight unpacking during accelerate loading and DAC from_latents missing STE are different model internals and bugs." }, { "left": "issue:43295", "right": "issue:45834", "accept": false, - "reason": "Both are regression-style issues, but one is processor/tokenizer API breakage and the other is a long OCR input index error." + "reason": "Processor/tokenizer regression in v4.57.5 and Kosmos2.5 long OCR index error are unrelated multimodal bugs." }, { "left": "issue:43653", "right": "issue:43720", "accept": false, - "reason": "Different tokenizer CI bug vs BitNet loading bug; only broad loading/tokenization overlap." - }, - { - "left": "issue:46829", - "right": "issue:46832", - "accept": false, - "reason": "Both are output-shape/type issues in vision models, but they affect different models and code paths." + "reason": "BigBirdTokenizer special-token registration and BitNet packed-weight loading are different domains and code paths." }, { "left": "issue:43550", "right": "issue:43927", "accept": false, - "reason": "Different failure modes: Bamba torch.compile/SDPA crash vs DiaConfig losing custom token IDs on save/load." - }, - { - "left": "issue:43550", - "right": "issue:44361", - "accept": false, - "reason": "No shared bug: torch.compile/SDPA failure vs MLukeTokenizer AttributeError." + "reason": "Bamba torch.compile SDPA failure and DiaConfig custom token IDs lost after save/load are unrelated failures." }, { "left": "issue:43596", "right": "issue:44220", "accept": false, - "reason": "Different bugs and model areas: ZeRO-3/BertModel init failure vs audio feature extraction helper issue." + "reason": "Deepspeed ZeRO-3/BertModel index error and _torch_extract_fbank_features() are different subsystems." }, { - "left": "issue:44625", - "right": "issue:47050", + "left": "issue:46829", + "right": "issue:46832", "accept": false, - "reason": "Qwen3.5 config propagation bug vs CanineModel fp16 ONNX export dtype mismatch." + "reason": "VideoPrism output field misassignment and RT-DETR forward TypeError are distinct model bugs." }, { - "left": "issue:43540", - "right": "issue:46294", + "left": "issue:43550", + "right": "issue:44361", "accept": false, - "reason": "Both involve video input in Qwen-family models, but the symptoms and implicated APIs differ; not clearly the same bug." + "reason": "Bamba compile/SDPA and MLukeTokenizer AttributeError are unrelated issues." }, { "left": "issue:43531", "right": "issue:43643", "accept": false, - "reason": "Different issues: sliding_window behavior in Qwen3-MoE vs trust_remote_code missing fields in AutoConfig.from_pretrained." + "reason": "Qwen3 sliding_window handling and AutoConfig trust_remote_code missing fields are different configuration/loading problems." }, { - "left": "issue:44062", - "right": "issue:44488", + "left": "issue:44625", + "right": "issue:47050", "accept": false, - "reason": "Different tokenizer problems: AddedToken special-arg conflict vs a model load failure for cjvt/sleng-bert." + "reason": "Qwen3.5 num_labels propagation and Canine ONNX export dtype mismatch are unrelated bugs." }, { - "left": "issue:45381", - "right": "issue:46693", + "left": "issue:44062", + "right": "issue:44488", "accept": false, - "reason": "Qwen2.5-VL video position IDs bug vs FA2 long-context performance regression; different concerns." + "reason": "AddedToken special-argument conflict and failure to load a specific tokenizer are different tokenizer-loading problems." + }, + { + "left": "issue:43540", + "right": "issue:46294", + "accept": false, + "reason": "Both concern Qwen video inputs, but one is a processing ValueError in Qwen3OmniMoe and the other is chat-template feature/token mismatch in Qwen3.5-9B." + }, + { + "left": "issue:45381", + "right": "issue:46693", + "accept": false, + "reason": "Qwen2.5-VL video position IDs and FA2 long-context performance regression are unrelated issues." }, { "left": "issue:45987", "right": "issue:46829", "accept": false, - "reason": "StaticCache return type bug vs VideoPrism hidden_states/output packaging bug." + "reason": "StaticCache sequence-length return type and VideoPrism hidden_states output are not the same bug." }, { "left": "issue:43611", "right": "issue:46665", "accept": false, - "reason": "Both are loading-related, but base_model_prefix loading regression is distinct from from_pretrained behavior inside Python classes." + "reason": "Base_model_prefix loading breakage and inconsistent from_pretrained usage inside Python classes are different loading semantics issues." }, { "left": "issue:30064", "right": "issue:36331", "accept": false, - "reason": "Image segmentation-map processing bug vs CustomTrainer compute_loss signature mismatch." + "reason": "Void segmentation-map processing and CustomTrainer.compute_loss signature mismatch are unrelated." + }, + { + "left": "issue:41720", + "right": "issue:44991", + "accept": false, + "reason": "Qwen3 auto device-mapping CUDA assert and tokenizer loading for EMBEDDIA/est-roberta are unrelated failures." + }, + { + "left": "issue:44220", + "right": "issue:44661", + "accept": false, + "reason": "Audio feature extraction and add-new-model-like failing inside TOKENIZER_MAPPING_NAMES are different code paths." } ] }, - "evaluator_result": { - "accept": true, - "feedback": "The cluster summary is well grounded in the packet. The only accepted edge, issue:44514 \u2194 issue:44545, is strongly supported by the near-identical titles and matching failure mode, with 44514 correctly chosen as the earlier report. The remaining soft edges are phrased conservatively and do not overclaim duplication." - }, - "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 }, @@ -5866,12 +5900,12 @@ "issue:34634", "issue:36010", "issue:39692", - "issue:41720", "issue:43322", "issue:43335", "issue:43421", "issue:43540", "issue:43550", + "issue:43577", "issue:43606", "issue:43618", "issue:43643", @@ -5880,15 +5914,15 @@ "issue:43697", "issue:43749", "issue:43756", + "issue:43818", "issue:43819", "issue:44008", "issue:44186", - "issue:44220", "issue:44442", "issue:44448", + "issue:44479", "issue:44488", "issue:44561", - "issue:44661", "issue:44671", "issue:44811", "issue:44821", @@ -5912,18 +5946,16 @@ "issue:47020" ], "soft_pairs": [ - "issue:41720|issue:44991", - "issue:44220|issue:44661", "issue:44186|issue:44898", "issue:43653|issue:43819", "issue:44488|issue:47020", - "issue:45987|issue:46932", "issue:43550|issue:43819", "issue:43335|issue:43756", "issue:43618|issue:43697", "issue:43643|issue:44877", - "issue:44008|issue:46858", + "issue:45987|issue:46932", "issue:44821|issue:45084", + "issue:44008|issue:46858", "issue:34634|issue:36010", "issue:44843|issue:45081", "issue:39692|issue:45834", @@ -5934,216 +5966,221 @@ "issue:45183|issue:45479", "issue:44991|issue:47020", "issue:43322|issue:45685", - "issue:43644|issue:43749", "issue:44991|issue:46291", - "issue:43606|issue:45072", + "issue:43644|issue:43749", "issue:44561|issue:45020", + "issue:43606|issue:45072", "issue:43421|issue:44811", - "issue:43550|issue:45072" + "issue:43550|issue:45072", + "issue:43577|issue:43818", + "issue:44479|issue:45834" ], - "prepared_review_unit_hash": "3b48b67e3fb4104e8b52d1ad8d3411f5f73f195f63457368e7da8a385585f138", + "prepared_review_unit_hash": "6647d02ca62f5ce78c29802d8ee42cd6429c239b9658e3324c5a98c1a2ddaa72", "component_budget": { "node_count": 514, "item_count": 514, "soft_pair_count": 1891, - "serialized_chars": 980926, - "estimated_input_tokens": 245232, - "estimated_eval_tokens": 490720 + "serialized_chars": 980895, + "estimated_input_tokens": 245224, + "estimated_eval_tokens": 490704 }, "budget": { "node_count": 47, "item_count": 47, "soft_pair_count": 28, - "serialized_chars": 35695, - "estimated_input_tokens": 8924, - "estimated_eval_tokens": 18104 + "serialized_chars": 35859, + "estimated_input_tokens": 8965, + "estimated_eval_tokens": 18186 }, "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 bug reports. The soft pairs share broad keywords (tokenizer, config, remote_code, v5 regressions, loading) but not the same concrete failure mode or code path, so none should be merged as duplicates. No PRs are present.", - "confidence": 0.93, - "canonical_issue_reason": "If a single issue must represent the cluster, #45020 is the broadest umbrella because it captures the general theme of v5/remote_code breakage; however, the items are not truly one duplicate cluster.", + "summary": "Mostly heterogeneous Transformers bug reports spanning tokenizer/config loading, remote_code regressions, multimodal/video failures, and dtype/cache/offload issues. Only one soft pair looks like the same underlying bug; the rest are surface-level similarities across different code paths.", + "confidence": 0.71, + "canonical_issue_reason": "Issue 45020 is the broadest umbrella report here, covering recent remote_code regressions and serving as the closest cluster-level representative, though the set is otherwise mixed.", "canonical_pr_reason": null, - "best_issue_reason": "#45020 is the most broadly representative issue in the set and has the clearest high-level summary of a recurring regression class, even though most other items are not duplicates of it.", + "best_issue_reason": "Issue 45020 has the widest apparent blast radius and is the most representative high-level issue among these reports.", "best_pr_reason": null, "soft_edge_verdicts": [ - { - "left": "issue:41720", - "right": "issue:44991", - "accept": false, - "reason": "Different models and failures: Qwen3 device-mapping CUDA assert vs tokenizer loading for est-roberta." - }, - { - "left": "issue:44220", - "right": "issue:44661", - "accept": false, - "reason": "Unrelated code paths: feature extraction helper bug vs tokenizer mapping-name generation failure." - }, { "left": "issue:44186", "right": "issue:44898", "accept": false, - "reason": "Different subsystems and symptoms: LayoutLMv2 tokenization/NER crash vs Perceiver image-size interpolation failure." + "reason": "Both are model/input handling bugs, but LayoutLMv2 tokenizer NER/padding crashes and Perceiver non-default-resolution failures are unrelated code paths." }, { "left": "issue:43653", "right": "issue:43819", "accept": false, - "reason": "Tokenizer special-token registration issue vs DAC latent reconstruction mismatch; not the same bug." + "reason": "BigBirdTokenizer special-token registration and DAC from_latents/STE mismatch are different subsystems and failures." }, { "left": "issue:44488", "right": "issue:47020", "accept": false, - "reason": "Both are tokenizer-load issues, but for different models and different root causes; not mergeable as one fix." - }, - { - "left": "issue:45987", - "right": "issue:46932", - "accept": false, - "reason": "Cache sequence-length return-type bug is unrelated to Gemma4 quantized embedder dtype casting." + "reason": "Both involve tokenizer loading, but one is a model-specific load failure and the other is an XLMRobertaTokenizer init bug with a different root cause." }, { "left": "issue:43550", "right": "issue:43819", "accept": false, - "reason": "Different areas entirely: Bamba torch.compile/SDPA failure vs DAC.from_latents mismatch." + "reason": "torch.compile/SDPA failure in Bamba is unrelated to DAC latent conversion mismatch." }, { "left": "issue:43335", "right": "issue:43756", "accept": false, - "reason": "Separate model/config bugs: sparse-layer creation vs RoPE layer dropping." + "reason": "SwitchTransformers sparse-layer config and Smollm3 RoPE-layer dropout are separate architecture bugs." }, { "left": "issue:43618", "right": "issue:43697", "accept": false, - "reason": "CLIP output field regression vs RTDetrV2 output difference; no shared code-path problem." + "reason": "Missing CLIPOutput attentions and RTDetrV2 output drift in v5 are distinct model regressions." }, { "left": "issue:43643", "right": "issue:44877", "accept": false, - "reason": "Both concern config loading, but one is trust_remote_code field loss and the other is strict config rejection for granite_speech." + "reason": "Both touch config/remote-code loading, but one is missing fields from trust_remote_code and the other is strict config blocking granite_speech; different failure modes." }, { - "left": "issue:44008", - "right": "issue:46858", + "left": "issue:45987", + "right": "issue:46932", "accept": false, - "reason": "Gemma3n variable-name collision is unrelated to cache.get_seq_length() returning a device tensor." + "reason": "StaticCache seq-length type/shape bug and Gemma4 quantized-embedder dtype casting are unrelated." }, { "left": "issue:44821", "right": "issue:45084", "accept": false, - "reason": "AutoImageProcessor-from-URL loading and a template compilation TypeError are unrelated." + "reason": "AutoImageProcessor URL loading and template-node compile errors do not share the same underlying bug." + }, + { + "left": "issue:44008", + "right": "issue:46858", + "accept": false, + "reason": "Gemma3n variable collision in forward() and cache.get_seq_length() returning a device tensor are different bugs." }, { "left": "issue:34634", "right": "issue:36010", "accept": false, - "reason": "BarkProcessor voice_preset handling is unrelated to the GenerationMixin import error." + "reason": "BarkProcessor voice_preset behavior and GenerationMixin import failure are unrelated." }, { "left": "issue:44843", "right": "issue:45081", "accept": false, - "reason": "Both hit _patch_mistral_regex, but one is offline hub access and the other is a backend_tokenizer attribute crash; different fixes." + "reason": "Both involve _patch_mistral_regex, but one is an offline model_info call and the other is a backend_tokenizer attribute crash; distinct defects." }, { "left": "issue:39692", "right": "issue:45834", "accept": false, - "reason": "SigLIP2 docs/example problems and Kosmos2.5 long-input index errors are unrelated." + "reason": "SigLIP2 doc-example errors and Kosmos2.5 long-OCR index errors are different models and code paths." }, { "left": "issue:43540", "right": "issue:44008", "accept": false, - "reason": "Different multimodal models and different failures: Qwen3OmniMoe video processing vs Gemma3n audio mask attribute collision." + "reason": "Qwen3OmniMoe video-input ValueError and Gemma3n audio_mel_mask collision are unrelated multimodal bugs." }, { "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": "LayoutLMv2 tokenizer crashes and AutoTokenizer failing to load FastSpeech2ConformerTokenizer are different tokenizer regressions." }, { "left": "issue:44448", "right": "issue:44671", "accept": false, - "reason": "Both are v5 regressions, but the model behavior and underlying code paths are distinct." + "reason": "Pegasus v4/v5 output changes and CamemBERT masked-LM prediction drift are separate model regressions." }, { "left": "issue:45584", "right": "issue:45834", "accept": false, - "reason": "Whisper empty-transcription generation bug is unrelated to Kosmos2.5 OCR length/index handling." + "reason": "Whisper empty-transcription handling and Kosmos2.5 OCR length errors are unrelated." }, { "left": "issue:45183", "right": "issue:45479", "accept": false, - "reason": "Validation error messaging bug vs degenerate zero-loss classification behavior; different underlying problems." + "reason": "Input validation XOR messaging and single-label zero-loss classification are different layers of behavior." }, { "left": "issue:44991", "right": "issue:47020", - "accept": false, - "reason": "Different tokenizer implementations and failure modes; not the same loading bug." + "accept": true, + "reason": "These look like the same tokenizer-loading regression: the est-roberta load failure is a concrete manifestation of the XLMRobertaTokenizer Unigram-vocab bug." }, { "left": "issue:43322", "right": "issue:45685", "accept": false, - "reason": "Llava Next loading segfault and MPS histogram dtype error are unrelated." - }, - { - "left": "issue:43644", - "right": "issue:43749", - "accept": false, - "reason": "Both involve loading/runtime stability, but the non-persistent buffer junk issue is not the same as FSDP CPU RAM efficient loading." + "reason": "Llava Next load-time segfault and moe MPS histogram failure are unrelated." }, { "left": "issue:44991", "right": "issue:46291", "accept": false, - "reason": "One is a tokenizer load failure for a model repo; the other is a pipelines import break caused by kernels dependency metadata." + "reason": "Tokenizer loading failure for est-roberta and kernels package version/revision import failure are different issues." }, { - "left": "issue:43606", - "right": "issue:45072", + "left": "issue:43644", + "right": "issue:43749", "accept": false, - "reason": "Different model families and different device/dtype problems; not one concrete bug." + "reason": "Non-persistent buffer junk in v5 and broken FSDP CPU RAM efficient loading are separate loading regressions." }, { "left": "issue:44561", "right": "issue:45020", "accept": false, - "reason": "Both mention remote_code regressions, but the specific breakage and code path differ; too broad to treat as duplicates." + "reason": "Both mention remote_code regressions, but 44561 is specifically the is_torch_fx_available removal while 45020 is a broader umbrella report; not clearly the same concrete bug." + }, + { + "left": "issue:43606", + "right": "issue:45072", + "accept": false, + "reason": "CPU offload device mismatch for bark-small and bfloat16 dtype mismatch in SwitchTransformers/TimmWrapperModel are different failure modes." }, { "left": "issue:43421", "right": "issue:44811", "accept": false, - "reason": "Special-token runtime updates and Whisper batch_decode skip_special_tokens behavior are separate tokenizer/processor issues." + "reason": "Runtime post-processor updates for special tokens and Whisper batch_decode skip_special_tokens handling are unrelated tokenizer/processor behaviors." }, { "left": "issue:43550", "right": "issue:45072", "accept": false, - "reason": "Bamba SDPA torch.compile failure and bfloat16 dtype mismatches are different bugs despite both being model/runtime issues." + "reason": "Bamba torch.compile SDPA failure and bfloat16 dtype mismatches in other models are not the same bug." + }, + { + "left": "issue:43577", + "right": "issue:43818", + "accept": false, + "reason": "BLIP2 dtype propagation and Video-LLaVA temporal-attention/weight-sharing issues are unrelated." + }, + { + "left": "issue:44479", + "right": "issue:45834", + "accept": false, + "reason": "Qwen video-input regression and Kosmos2.5 long-OCR index error are different multimodal paths." } ] }, - "evaluator_result": null, - "evaluator_used": false, + "evaluator_result": { + "accept": true, + "feedback": "The summary is grounded in the packet: issue 45020 is the clearest umbrella report, and the cluster is indeed mostly heterogeneous across tokenizer/config loading, remote_code, multimodal, and dtype/cache/offload bugs. The soft-edge calls are conservative overall, and the single accepted pair (44991 \u2194 47020) is plausibly the same tokenizer-loading regression." + }, + "evaluator_used": true, "retried": false, - "accepted_nontrivial_soft_edge": false, + "accepted_nontrivial_soft_edge": true, "error_kind": null, "error_message": null }, @@ -6171,11 +6208,11 @@ "issue:43577", "issue:43582", "issue:43645", + "issue:43653", "issue:43716", "issue:43720", "issue:43749", "issue:43782", - "issue:43818", "issue:43819", "issue:43825", "issue:43827", @@ -6183,11 +6220,11 @@ "issue:43994", "issue:44008", "issue:44291", - "issue:44479", "issue:44610", "issue:44671", "issue:44805", "issue:44857", + "issue:44912", "issue:44936", "issue:44977", "issue:45003", @@ -6196,7 +6233,6 @@ "issue:45325", "issue:45381", "issue:45701", - "issue:45834", "issue:46032", "issue:46175", "issue:46285", @@ -6210,8 +6246,6 @@ "issue:46932" ], "soft_pairs": [ - "issue:43577|issue:43818", - "issue:44479|issue:45834", "issue:43582|issue:46512", "issue:43782|issue:44977", "issue:44805|issue:44936", @@ -6229,205 +6263,200 @@ "issue:43825|issue:43827", "issue:34634|issue:36331", "issue:42175|issue:45081", - "issue:43577|issue:43582", "issue:44671|issue:45701", + "issue:43577|issue:43582", "issue:45003|issue:46285", "issue:46175|issue:46932", - "issue:43262|issue:43819", - "issue:46697|issue:46728", "issue:43782|issue:45325", + "issue:46697|issue:46728", + "issue:43262|issue:43819", + "issue:43653|issue:44912", "issue:43582|issue:44857" ], - "prepared_review_unit_hash": "fcc3f3c1e3d8a4e6d74bc8bab5a9129fa30a5c1bf9ea6de7149f1a13b7ae24ac", + "prepared_review_unit_hash": "14891676e0305ff7e3528f63512b3e1d480cf3b1ffc1588329350ab5fc10b015", "component_budget": { "node_count": 514, "item_count": 514, "soft_pair_count": 1891, - "serialized_chars": 980926, - "estimated_input_tokens": 245232, - "estimated_eval_tokens": 490720 + "serialized_chars": 980895, + "estimated_input_tokens": 245224, + "estimated_eval_tokens": 490704 }, "budget": { - "node_count": 48, - "item_count": 48, - "soft_pair_count": 27, - "serialized_chars": 36362, - "estimated_input_tokens": 9091, - "estimated_eval_tokens": 18438 + "node_count": 47, + "item_count": 47, + "soft_pair_count": 26, + "serialized_chars": 35469, + "estimated_input_tokens": 8868, + "estimated_eval_tokens": 17992 }, "overflow_policy": "truncate_then_skip", "trimmed": false, "aggressively_trimmed": false, "split": true, "analyst_result": { - "summary": "The cluster is highly heterogeneous: most soft links are topical similarities, not true duplicates. I would not merge any of the proposed pairs.", - "confidence": 0.97, - "canonical_issue_reason": "No true canonical duplicate stands out because the cluster mixes unrelated bugs; if forced to anchor the set, #43749 is the strongest representative by discussion/inbound refs, but it does not represent the rest.", + "summary": "This set is mostly heterogeneous: tokenizer, multimodal, training, loading, and backend issues are all mixed together. I don\u2019t see any soft pair that clearly describes the same underlying bug/change, so the cluster looks like false-positive similarity rather than true duplicates. #28218 is only the closest local representative for the tokenizer-related subset.", + "confidence": 0.72, + "canonical_issue_reason": "#28218 is the clearest, most concrete symptom report in the set and the oldest issue with some inbound references; if a single issue must represent the only loosely related tokenizer subset, this is the best fit.", "canonical_pr_reason": null, - "best_issue_reason": "#43749 (\"FSDP_CPU_RAM_EFFICIENT_LOADING broken\") is the best single issue to anchor on because it has the most activity and a concrete bug report, even though it is not a duplicate hub for this cluster.", + "best_issue_reason": "No issue is a strong cluster-wide canonical because the items span unrelated subsystems; #28218 is merely the best representative for the narrow tokenizer-space symptom, not the whole cluster.", "best_pr_reason": null, "soft_edge_verdicts": [ - { - "left": "issue:43577", - "right": "issue:43818", - "accept": false, - "reason": "Both are multimodal-model bugs, but one is a Blip2 dtype-loading issue and the other is a Video-LLaVA architecture/weight-sharing issue; different code paths." - }, - { - "left": "issue:44479", - "right": "issue:45834", - "accept": false, - "reason": "Both involve video/multimodal input regressions, but they affect different models and symptoms; not the same underlying bug." - }, { "left": "issue:43582", "right": "issue:46512", "accept": false, - "reason": "Apple Silicon allocator TypeError and stale `config.dtype` on `from_config()` are unrelated initialization bugs." + "reason": "Different dtype/state bugs: Apple Silicon warmup TypeError vs from_config leaving config.dtype stale." }, { "left": "issue:43782", "right": "issue:44977", "accept": false, - "reason": "One is a `from_pretrained(..., weight_only=True)` loading failure, the other is a flash-attention generation problem; different failure modes." + "reason": "Unrelated failures: Qwen3VL from_pretrained weight_only error vs Qwen3.5 flash-attention generation issues." }, { "left": "issue:44805", "right": "issue:44936", "accept": false, - "reason": "Both are training/evaluation errors, but one is a tensor-mask shape mismatch and the other is an `evaluate()` after `train()` regression." + "reason": "Mask/index shape mismatch bug is unrelated to trainer.evaluate() failing after trainer.train()." }, { "left": "issue:43825", "right": "issue:43994", "accept": false, - "reason": "The first is an error-message/docs issue about removed translation pipelines; the second is a model inference quality bug." + "reason": "One is a pipeline v5 error-message issue; the other is a model output correctness issue for SigLIP2." }, { "left": "issue:43388", "right": "issue:46032", "accept": false, - "reason": "Metrics gathering dropping tuple labels and Mamba2 cache corruption are separate training/runtime bugs." + "reason": "Different code paths: gather_for_metrics dropping tuple labels vs Mamba2 use_cache correctness." }, { "left": "issue:46294", "right": "issue:46463", "accept": false, - "reason": "Qwen3.5-VL video-template mismatch and T5 decoder-input validation messaging are unrelated issues." + "reason": "Both mention input validation, but Qwen3.5 video/chat-template mismatch is not the same bug as T5 decoder-input error messaging." }, { "left": "issue:43716", "right": "issue:44008", "accept": false, - "reason": "Both are multimodal dtype/forward bugs, but they affect different models and different internal components." + "reason": "Both are multimodal dtype/forward-path bugs, but they affect different models and different concrete failures." }, { "left": "issue:43931", "right": "issue:46846", "accept": false, - "reason": "Both mention shape/oom problems, but one is a Qwen3-VL weight-shape loading error and the other is a tensor-parallel plan omission." + "reason": "Qwen3-VL weight-shape mismatch and Qwen3.5 tensor-parallel plan omissions are different loading/parallelism problems." }, { "left": "issue:43540", "right": "issue:45381", "accept": false, - "reason": "Both are Qwen video-input regressions, but they concern different model families and different position/token handling paths." + "reason": "Video-input processing failure in Qwen3OmniMoe is not the same as Qwen2.5-VL rope/position-id scaling." }, { "left": "issue:28218", "right": "issue:29159", "accept": false, - "reason": "These are both tokenizer-related, but one is about an extra space after an added token and the other is a broader slow/fast tokenizer inconsistency; not enough evidence they are the same bug." + "reason": "Tokenizer-space behavior and slow/fast tokenizer inconsistency may be related, but the evidence is too broad to conclude the same underlying bug." }, { "left": "issue:44610", "right": "issue:45072", "accept": false, - "reason": "Image-size mismatch in OmDet-Turbo and dtype mismatch in SwitchTransformers/TimmWrapperModel are unrelated." + "reason": "Different models and failures: processor size mismatch vs bfloat16 dtype mismatch in inference wrappers." }, { "left": "issue:43749", "right": "issue:44291", "accept": false, - "reason": "Both are model-loading bugs, but one is broken FSDP CPU RAM-efficient loading and the other is an unexpected internal argument with `init_empty_weights()`." + "reason": "FSDP CPU RAM loading breakage is a different path from init_empty_weights rejecting an unexpected argument." }, { "left": "issue:43720", "right": "issue:45072", "accept": false, - "reason": "BitNet packed-weight loading and bfloat16 dtype mismatches are different problems in different code paths." + "reason": "BitNet packed-weight loading and dtype mismatch inference issues are unrelated." }, { "left": "issue:43645", "right": "issue:46665", "accept": false, - "reason": "Jupyter custom-model initialization breakage and `from_pretrained` behavior inside Python classes are related only broadly, not the same defect." + "reason": "Both involve custom model initialization, but the notebook regression and from_pretrained-in-classes inconsistency are not the same concrete bug." }, { "left": "issue:43825", "right": "issue:43827", "accept": false, - "reason": "One is a runtime error-message regression, the other is stale documentation after pipeline removal; different artifacts and fixes." + "reason": "Related to the same v5 pipeline removals, but one is an error message problem and the other is stale docs; not the same artifact change." }, { "left": "issue:34634", "right": "issue:36331", "accept": false, - "reason": "BarkProcessor voice preset and CustomTrainer loss signature are completely unrelated." + "reason": "BarkProcessor voice_preset and CustomTrainer.compute_loss are unrelated bugs in different subsystems." }, { "left": "issue:42175", "right": "issue:45081", "accept": false, - "reason": "Packaging/backends installation and a tokenizer regex patch crash are separate issues." + "reason": "Backend dependency packaging and Mistral regex patch crashes are different problems." }, { - "left": "issue:43577", - "right": "issue:43582", + "left": "issue:44671", + "right": "issue:45701", "accept": false, - "reason": "BLIP2 dtype loading and Apple Silicon allocator warmup are unrelated." + "reason": "CamemBERT masked-LM prediction regression and generic tokenization version changes are not the same issue." }, { - "left": "issue:44671", - "right": "issue:45701", + "left": "issue:43577", + "right": "issue:43582", "accept": false, - "reason": "CamemBERT masked-LM regression and a generic tokenization-version change are too broad to treat as the same bug." + "reason": "BLIP2 dtype mismatch and Apple Silicon allocator warmup TypeError are unrelated." }, { "left": "issue:45003", "right": "issue:46285", "accept": false, - "reason": "Unsafe `sys.modules` access and CLIP structure inconsistency are different modeling-utils/API-shape issues." + "reason": "Unsafe sys.modules access is not the same bug as CLIP text-model structural inconsistency." }, { "left": "issue:46175", "right": "issue:46932", "accept": false, - "reason": "Gemma4 one-hot position-embedding memory blowup and quantized embedder dtype casting are distinct bugs." + "reason": "Gemma4 one-hot memory blowup and Gemma4 quantized embedders dtype casting are different issues." }, { - "left": "issue:43262", - "right": "issue:43819", + "left": "issue:43782", + "right": "issue:45325", "accept": false, - "reason": "Audio chat-template sample-rate defaulting and DAC latent reconstruction mismatch are unrelated audio bugs." + "reason": "Qwen3VL weight_only loading error and Qwen2.5-VL still-image rope scaling are different model/version bugs." }, { "left": "issue:46697", "right": "issue:46728", "accept": false, - "reason": "Over-padding in Swin2SR and Aria patch-count miscalculation are both image-shape issues, but not the same code path." + "reason": "Swin2SR over-padding and Aria patch-count miscalculation are both image-processing bugs, but not the same concrete failure." }, { - "left": "issue:43782", - "right": "issue:45325", + "left": "issue:43262", + "right": "issue:43819", + "accept": false, + "reason": "Audio apply_chat_template sampling-rate defaulting and DAC from_latents/STE mismatch are unrelated." + }, + { + "left": "issue:43653", + "right": "issue:44912", "accept": false, - "reason": "Qwen3VL loading failure and Qwen2.5-VL rope-index scaling bug are different failures." + "reason": "BigBirdTokenizer special-token registration and MXFP4 quantization fallback are different bugs." }, { "left": "issue:43582", "right": "issue:44857", "accept": false, - "reason": "Apple Silicon TypeError in allocator warmup and CUDA float16 AMP crash in LwDetrImageLoss are unrelated." + "reason": "Apple Silicon caching warmup TypeError is unrelated to LwDetrImageLoss float16 AMP/CUDA crashes." } ] }, @@ -6467,7 +6496,7 @@ "issue:43638", "issue:43643", "issue:43645", - "issue:43653", + "issue:43696", "issue:43697", "issue:43742", "issue:43746", @@ -6501,7 +6530,6 @@ "issue:47050" ], "soft_pairs": [ - "issue:43653|issue:44912", "issue:44589|issue:45042", "issue:43742|issue:45020", "issue:45372|issue:46326", @@ -6512,220 +6540,221 @@ "issue:43906|issue:46209", "issue:43531|issue:45042", "issue:43526|issue:43697", + "issue:43638|issue:44936", "issue:43452|issue:44991", "issue:42175|issue:46291", - "issue:43638|issue:44936", - "issue:43819|issue:43927", "issue:45081|issue:45406", + "issue:43819|issue:43927", "issue:44060|issue:44368", "issue:43329|issue:45600", "issue:45084|issue:45538", "issue:43526|issue:44554", "issue:42490|issue:43257", "issue:43746|issue:44811", - "issue:45859|issue:47050", - "issue:43854|issue:44977", "issue:43643|issue:45759", + "issue:43854|issue:44977", + "issue:45859|issue:47050", "issue:38175|issue:44610", + "issue:39692|issue:44912", "issue:43262|issue:44811", - "issue:39692|issue:44912" + "issue:43257|issue:43696" ], - "prepared_review_unit_hash": "ae4373ef7525d063a91d055ce0264fbce385aad16d55866ecab03f04fa143d81", + "prepared_review_unit_hash": "fd8eb8f2161879218a0ed25ca9e42fa5c0e1d183c3cd360a5050b8918261d47e", "component_budget": { "node_count": 514, "item_count": 514, "soft_pair_count": 1891, - "serialized_chars": 980926, - "estimated_input_tokens": 245232, - "estimated_eval_tokens": 490720 + "serialized_chars": 980895, + "estimated_input_tokens": 245224, + "estimated_eval_tokens": 490704 }, "budget": { "node_count": 48, "item_count": 48, "soft_pair_count": 28, - "serialized_chars": 36553, - "estimated_input_tokens": 9139, - "estimated_eval_tokens": 18534 + "serialized_chars": 36518, + "estimated_input_tokens": 9130, + "estimated_eval_tokens": 18516 }, "overflow_policy": "truncate_then_skip", "trimmed": false, "aggressively_trimmed": false, "split": true, "analyst_result": { - "summary": "The cluster is highly heterogeneous; the soft links are mostly broad topical similarities rather than true duplicates. No convincing duplicate issue/PR canonical emerges from these items.", - "confidence": 0.21, - "canonical_issue_reason": "If forced to pick a representative, issue 45020 is the broadest umbrella report among the loading/remote-code regression items.", + "summary": "The cluster is mostly a grab bag of unrelated Transformers bugs across loading, tokenizers, processors, training, quantization, and model-specific regressions. I did not find any soft pair that clearly shares the same underlying bug or fix path.", + "confidence": 0.91, + "canonical_issue_reason": "issue:45020 is the broadest regression-style report in the set and the closest thing to an umbrella issue, though the cluster is not a true duplicate group.", "canonical_pr_reason": null, - "best_issue_reason": "Issue 45020 is the least specific and most cluster-like representative, but the set is still too diverse to treat as one duplicate family.", + "best_issue_reason": "issue:45020 is the most general issue here for broad load-time regressions, but it is still too vague to serve as a strong canonical for the whole cluster.", "best_pr_reason": null, "soft_edge_verdicts": [ - { - "left": "issue:43653", - "right": "issue:44912", - "accept": false, - "reason": "Different bugs: BigBirdTokenizer special-token decode regression vs MXFP4 quantization fallback/load behavior." - }, { "left": "issue:44589", "right": "issue:45042", "accept": false, - "reason": "Unrelated failures: missing Float8 storage type vs PIL image processors incorrectly requiring torchvision." + "reason": "Different bugs: FP8 storage lookup vs PIL/torchvision backend requirement." }, { "left": "issue:43742", "right": "issue:45020", "accept": false, - "reason": "Specific MobileLLM loading key error vs a broad remote_code regression; same area, not the same bug." + "reason": "Both are loading-related, but one is a specific MobileLLM key error and the other is a broad remote_code regression." }, { "left": "issue:45372", "right": "issue:46326", "accept": false, - "reason": "Both involve mistral_common, but one is an import failure and the other is a chat-template/continue_final_message bug." + "reason": "Both mention mistral_common, but one is an import failure and the other is a tokenizer continue_final_message behavior bug." }, { "left": "issue:43550", "right": "issue:44912", "accept": false, - "reason": "torch.compile + SDPA on Bamba is unrelated to the MXFP4 quantization fallback issue." + "reason": "Unrelated: SDPA torch.compile failure for Bamba vs MXFP4 quantization fallback for git-oss." }, { "left": "issue:36010", "right": "issue:43645", "accept": false, - "reason": "GenerationMixin import error and Jupyter custom-model initialization breakage are different code paths." + "reason": "Different failure modes: GenerationMixin import path vs custom model init in Jupyter notebooks." }, { "left": "issue:36010", "right": "issue:36331", "accept": false, - "reason": "ImportError for GenerationMixin is unrelated to the custom trainer kwarg incompatibility." + "reason": "ImportError for GenerationMixin is unrelated to the CustomTrainer compute_loss signature change." }, { "left": "issue:43257", "right": "issue:43329", "accept": false, - "reason": "Qwen3 MoE weight-conversion/loading issue is not the same as the multimodal token-counting video-branch bug." + "reason": "Qwen3 MoE weight conversion/loading issue vs undefined video-token helper variables in multimodal token counting." }, { "left": "issue:43906", "right": "issue:46209", "accept": false, - "reason": "An isolated reproduction of an older issue is not the same as the FP8/B200 inference failure." + "reason": "Isolated reproduction of an older issue is not the same bug as FP8 inference failure on B200." }, { "left": "issue:43531", "right": "issue:45042", "accept": false, - "reason": "Sliding-window handling in Qwen3-MoE is unrelated to the torchvision dependency regression in PIL processors." + "reason": "Different subsystems: Qwen3-MoE sliding window vs PIL image processor torchvision dependency." }, { "left": "issue:43526", "right": "issue:43697", "accept": false, - "reason": "BEiT reduce_labels behavior and RTDetrV2 output divergence are different processor/model bugs." + "reason": "BEiT label reduction bug is unrelated to RTDetrV2 output differences." + }, + { + "left": "issue:43638", + "right": "issue:44936", + "accept": false, + "reason": "Different training problems: ZeRO-3/non-pretrained Bert IndexError vs trainer.evaluate() after train()." }, { "left": "issue:43452", "right": "issue:44991", "accept": false, - "reason": "gguf_file handling for Auto* loaders is not the same as a specific tokenizer-loading regression." + "reason": "Both involve model/tokenizer loading, but gguf_file handling and est-roberta tokenizer loading are distinct issues." }, { "left": "issue:42175", "right": "issue:46291", "accept": false, - "reason": "Backend extras missing TensorFlow and kernels import requirements are unrelated dependency issues." + "reason": "Packaging/backends issue vs kernels package version/revision import failure." }, { - "left": "issue:43638", - "right": "issue:44936", + "left": "issue:45081", + "right": "issue:45406", "accept": false, - "reason": "Zero3 index error during Bert training is not the same as trainer.evaluate() failing after train()." + "reason": "Mistral regex patch crash and Gemma4Processor missing _tokenizer are different code paths and symptoms." }, { "left": "issue:43819", "right": "issue:43927", "accept": false, - "reason": "DAC latent reconstruction mismatch and DiaConfig save/load token-ID loss are unrelated." - }, - { - "left": "issue:45081", - "right": "issue:45406", - "accept": false, - "reason": "Both raise AttributeError, but one is Mistral regex patching and the other is Gemma4Processor tokenizer initialization." + "reason": "DAC latent reconstruction mismatch is unrelated to DiaConfig losing custom token IDs." }, { "left": "issue:44060", "right": "issue:44368", "accept": false, - "reason": "Both mention Qwen3 tied-weight warnings, but the underlying problems and APIs differ." + "reason": "Both mention tied-weights warnings in Qwen3, but one is a wrong weight-tying warning and the other is a repeated fine-tuning warning; not the same concrete bug." }, { "left": "issue:43329", "right": "issue:45600", "accept": false, - "reason": "Undefined video token helper variables and a removed auto-mapping reference are not the same bug." + "reason": "Multimodal video token counting bug vs auto_mappings.py CI breakage from removed configs." }, { "left": "issue:45084", "right": "issue:45538", "accept": false, - "reason": "Template compilation error and CLIPTokenizer max-length constant are unrelated." + "reason": "Template compilation error is unrelated to CLIPTokenizer model_max_length behavior." }, { "left": "issue:43526", "right": "issue:44554", "accept": false, - "reason": "Image label reduction bug vs MPS attention correctness issue; no shared code path." + "reason": "BEiT fast processor label reduction and MPS attention correctness are different bugs." }, { "left": "issue:42490", "right": "issue:43257", "accept": false, - "reason": "Trailing-slash local-path warning is unrelated to Qwen3 MoE loading/conversion." + "reason": "Local-path import warning with trailing slash is unrelated to Qwen3 MoE weight conversion under accelerate+deepspeed." }, { "left": "issue:43746", "right": "issue:44811", "accept": false, - "reason": "PEFT local-checkpoint loading and Whisper batch_decode skip_special_tokens are different subsystems and failures." + "reason": "PEFT checkpoint loading for GraniteSpeech is unrelated to Whisper batch_decode skipping special tokens." }, { - "left": "issue:45859", - "right": "issue:47050", + "left": "issue:43643", + "right": "issue:45759", "accept": false, - "reason": "CPU offload incompatibility in Qwen3_5MoeTextRotaryEmbedding is unrelated to Canine ONNX dtype mismatch." + "reason": "Both involve config handling, but trust_remote_code missing fields and from_config text_config unwrapping are different failures." }, { "left": "issue:43854", "right": "issue:44977", "accept": false, - "reason": "GLM-4.7-Flash loading/unit-test problems and Qwen3.5 flash-attention generation bugs are distinct." + "reason": "Model loading in unit tests for GLM-4.7-Flash is unrelated to Qwen3.5 flash-attention generation issues." }, { - "left": "issue:43643", - "right": "issue:45759", + "left": "issue:45859", + "right": "issue:47050", "accept": false, - "reason": "Both are config-loading regressions, but one is missing fields from trust_remote_code and the other is text_config unwrapping in from_config." + "reason": "CPU offload incompatibility in Qwen3_5MoeTextRotaryEmbedding is unrelated to Canine fp16 ONNX dtype mismatch." }, { "left": "issue:38175", "right": "issue:44610", "accept": false, - "reason": "SigLIP2 zero-probability output and OmDet processor size mismatch are different model/processor issues." + "reason": "SigLIP2 zero-probabilities and OmDet-Turbo processor/model size mismatch are separate vision bugs." + }, + { + "left": "issue:39692", + "right": "issue:44912", + "accept": false, + "reason": "SigLIP2 docs/model-processor mismatch and git-oss MXFP4 quantization fallback are different issues." }, { "left": "issue:43262", "right": "issue:44811", "accept": false, - "reason": "Audio apply_chat_template sampling-rate defaulting and Whisper batch_decode skip_special_tokens are unrelated." + "reason": "Audio chat-template sample-rate defaulting and Whisper batch_decode skip_special_tokens are unrelated processor behaviors." }, { - "left": "issue:39692", - "right": "issue:44912", + "left": "issue:43257", + "right": "issue:43696", "accept": false, - "reason": "SigLIP2 documentation/example errors are not the same as MXFP4 quantization fallback behavior." + "reason": "Qwen3 MoE loading/conversion bug is unrelated to GPT-oss CUDA OOM." } ] }, @@ -6768,7 +6797,6 @@ "issue:43540", "issue:43577", "issue:43650", - "issue:43696", "issue:43697", "issue:43742", "issue:43749", @@ -6783,6 +6811,7 @@ "issue:44617", "issue:44756", "issue:44821", + "issue:44857", "issue:44912", "issue:44987", "issue:45042", @@ -6795,233 +6824,234 @@ "issue:46421", "issue:46424", "issue:46701", + "issue:46832", "issue:47145" ], "soft_pairs": [ - "issue:43257|issue:43696", "issue:43329|issue:43450", "issue:43329|issue:44479", - "issue:43257|issue:43931", "issue:42915|issue:43742", - "issue:44387|issue:44617", + "issue:43257|issue:43931", "issue:44912|issue:47145", "issue:44291|issue:45042", "issue:34567|issue:40990", "issue:42915|issue:44821", - "issue:29159|issue:34567", + "issue:44387|issue:44617", "issue:45459|issue:45803", - "issue:46421|issue:46424", + "issue:29159|issue:34567", "issue:43650|issue:43906", + "issue:46421|issue:46424", "issue:43525|issue:46291", "issue:28282|issue:29870", "issue:43749|issue:44756", "issue:44987|issue:45538", - "issue:28282|issue:34689", "issue:43425|issue:46291", + "issue:28282|issue:34689", "issue:43540|issue:46701", "issue:43454|issue:43577", "issue:44568|issue:45356", "issue:33357|issue:34567", "issue:39379|issue:44554", "issue:42175|issue:43824", - "issue:43697|issue:45357" + "issue:43697|issue:45357", + "issue:44857|issue:46832" ], - "prepared_review_unit_hash": "4aacf4b0b9dfbc3f574485766b05d0418559a4059b07b757198d1416a8bd4256", + "prepared_review_unit_hash": "fc3552733b4f21e70dae1accb347fe8de82f9d3dcd65d83824f7f991060d3539", "component_budget": { "node_count": 514, "item_count": 514, "soft_pair_count": 1891, - "serialized_chars": 980926, - "estimated_input_tokens": 245232, - "estimated_eval_tokens": 490720 + "serialized_chars": 980895, + "estimated_input_tokens": 245224, + "estimated_eval_tokens": 490704 }, "budget": { - "node_count": 47, - "item_count": 47, + "node_count": 48, + "item_count": 48, "soft_pair_count": 27, - "serialized_chars": 35241, - "estimated_input_tokens": 8811, - "estimated_eval_tokens": 17878 + "serialized_chars": 35833, + "estimated_input_tokens": 8959, + "estimated_eval_tokens": 18174 }, "overflow_policy": "truncate_then_skip", "trimmed": false, "aggressively_trimmed": false, "split": true, "analyst_result": { - "summary": "The set is mostly heterogeneous, with no single cluster-wide duplicate. The only plausible duplicate pair is the video-input regression in the Qwen/VL multimodal path.", - "confidence": 0.87, - "canonical_issue_reason": "issue:44479 is the broadest and most user-facing report in the only coherent subgroup: it describes the video-input regression across multiple Qwen VL variants, while issue:43329 points to the likely lower-level helper bug behind it.", + "summary": "This cluster is mostly a collection of unrelated issue reports across loaders, tokenizers, multimodal/video, quantization, cache, and training utilities. Only one soft edge looks like a likely true duplicate around the Qwen video-input regression path.", + "confidence": 0.69, + "canonical_issue_reason": "issue:44479 is the best canonical issue here because it is the broader, user-facing regression report spanning multiple Qwen VL variants and a specific release window, and it likely subsumes the more specific video-token bug described in issue:43329.", "canonical_pr_reason": null, - "best_issue_reason": "issue:44479 is the best representative issue overall because it captures the shared video-regression symptom across several models and versions, rather than a narrower internal failure mode.", + "best_issue_reason": "issue:44479 is the strongest representative issue overall: it clearly states the affected models, regression scope, and concrete failure mode, making it the most useful anchor if the video regression cluster is deduplicated.", "best_pr_reason": null, "soft_edge_verdicts": [ - { - "left": "issue:43257", - "right": "issue:43696", - "accept": false, - "reason": "Different failures: Qwen3 MoE weight conversion/loading vs GPT-oss OOM." - }, { "left": "issue:43329", "right": "issue:43450", "accept": false, - "reason": "Both involve video/multimodal processing, but one is an undefined-variable bug in token counting and the other is a batched-shape bug in video processors; not the same concrete bug." + "reason": "Both involve video inputs, but one is a tokenizer/token-counting bug and the other is a batched processor shape issue; different code paths and symptoms." }, { "left": "issue:43329", "right": "issue:44479", "accept": true, - "reason": "Strong overlap on the video-input multimodal regression path; 43329 names the likely helper defect, and 44479 reports the broader v5.3.0 regression across the same Qwen VL family." - }, - { - "left": "issue:43257", - "right": "issue:43931", - "accept": false, - "reason": "Different Qwen-related loading problems: MoE weight conversion vs a shape-mismatch load error." + "reason": "Both point to the same Qwen VLM video-input regression; 43329 identifies the likely root cause in multimodal token counting, while 44479 reports the broader user-facing breakage." }, { "left": "issue:42915", "right": "issue:43742", "accept": false, - "reason": "Distinct model-loading failures with different symptoms and likely different code paths." + "reason": "Different models and failure modes: FP8 config handling vs a key error during MobileLLM loading." }, { - "left": "issue:44387", - "right": "issue:44617", + "left": "issue:43257", + "right": "issue:43931", "accept": false, - "reason": "Both mention OOM, but one is int4 quantization memory growth and the other is a video model running out of memory; not the same bug." + "reason": "Both are Qwen loading problems, but one is accelerate/deepspeed MOE conversion and the other is a VL weight-shape mismatch; not the same bug." }, { "left": "issue:44912", "right": "issue:47145", "accept": false, - "reason": "Both are quantization/loading bugs, but one falls back to bf16 for MXFP4 while the other ignores FP8 weight scales; different failure modes." + "reason": "Both are quantization/loading correctness issues, but they involve different formats and code paths (MXFP4 fallback vs FP8 weight_scale application)." }, { "left": "issue:44291", "right": "issue:45042", "accept": false, - "reason": "Unrelated issues: init_empty_weights signature/type error vs image processors incorrectly requiring torchvision." + "reason": "Unrelated subsystems: init_empty_weights loading failure vs an image processor torchvision dependency regression." }, { "left": "issue:34567", "right": "issue:40990", "accept": false, - "reason": "Trainer token-count accounting vs a perplexity regression on gpt-oss; unrelated." + "reason": "TrainerState token counting is unrelated to the GPT-OSS perplexity regression." }, { "left": "issue:42915", "right": "issue:44821", "accept": false, - "reason": "Different areas: Qwen3 MoE FP8 loading vs AutoImageProcessor URL loading." + "reason": "No concrete overlap: Qwen3Moe FP8 config vs AutoImageProcessor URL loading." }, { - "left": "issue:29159", - "right": "issue:34567", + "left": "issue:44387", + "right": "issue:44617", "accept": false, - "reason": "Tokenizer inconsistency vs TrainerState token-count update bug; different subsystems." + "reason": "Both mention memory/quantization-like stress, but one is int4 reserved-memory growth and the other is a SAM video OOM report; different bugs." }, { "left": "issue:45459", "right": "issue:45803", "accept": false, - "reason": "Both are error-handling issues, but they affect different code paths and are not the same underlying bug." + "reason": "Both involve exception handling, but they are separate tokenizer/data conversion code paths with different triggers and fixes." }, { - "left": "issue:46421", - "right": "issue:46424", + "left": "issue:29159", + "right": "issue:34567", "accept": false, - "reason": "Both are cache-related, but one is lazy-init recompilation and the other is max_cache_length handling; not duplicates." + "reason": "Tokenizer consistency and TrainerState token accounting are unrelated." }, { "left": "issue:43650", "right": "issue:43906", "accept": false, - "reason": "Placeholder/data request vs isolated reproduction of a separate issue." + "reason": "A vague request and an isolated reproduction are not the same bug." + }, + { + "left": "issue:46421", + "right": "issue:46424", + "accept": false, + "reason": "Both are cache-related, but one is lazy init recompilation and the other is max_cache_length behavior for quantized cache only." }, { "left": "issue:43525", "right": "issue:46291", "accept": false, - "reason": "Llama4 config attribute error vs pipeline import failure due to kernels version/revision requirements." + "reason": "Config attribute missing vs pipelines import failure due to kernels dependency metadata; unrelated." }, { "left": "issue:28282", "right": "issue:29870", "accept": false, - "reason": "ImportError for missing PyTorch vs Bart generate() UnboundLocalError; unrelated." + "reason": "PyTorch missing ImportError vs Bart generate UnboundLocalError are unrelated." }, { "left": "issue:43749", "right": "issue:44756", "accept": false, - "reason": "Different memory/loading problems with no shared concrete code path." + "reason": "Both concern memory/loading, but one is FSDP CPU RAM-efficient loading and the other is mmap on Strix Halo; different fixes." }, { "left": "issue:44987", "right": "issue:45538", "accept": false, - "reason": "Model-loading failure for a specific model vs CLIPTokenizer max length bug." + "reason": "Loading a specific model under transformers>=5.1.0 is unrelated to CLIPTokenizer's max_length constant." }, { - "left": "issue:28282", - "right": "issue:34689", + "left": "issue:43425", + "right": "issue:46291", "accept": false, - "reason": "Generic missing-torch import error vs Llama 3.2 Vision model-loading regression." + "reason": "Torch version incompatibility is not the same as kernels version/revision requirements for pipelines import." }, { - "left": "issue:43425", - "right": "issue:46291", + "left": "issue:28282", + "right": "issue:34689", "accept": false, - "reason": "Torch 2.10 compatibility issue vs kernels import failure; different dependencies and symptoms." + "reason": "Generic PyTorch missing-library error vs Llama 3.2 Vision loading regression are not the same issue." }, { "left": "issue:43540", "right": "issue:46701", "accept": false, - "reason": "Different multimodal pipeline failures involving different models and failure modes." + "reason": "Different multimodal model families and different failure modes (video processing ValueError vs image-text pipeline output handling)." }, { "left": "issue:43454", "right": "issue:43577", "accept": false, - "reason": "Weight tying bug in AyaVision vs dtype propagation bug in BLIP-2 loading." + "reason": "Weight tying/generation corruption in AyaVision is unrelated to Blip2 dtype propagation." }, { "left": "issue:44568", "right": "issue:45356", "accept": false, - "reason": "Both tokenizer regressions, but one is missing BOS/EOS insertion and the other is codec/warning handling for a different tokenizer." + "reason": "Both are tokenizer regressions, but they affect different models and behaviors; not the same concrete bug." }, { "left": "issue:33357", "right": "issue:34567", "accept": false, - "reason": "MacOS CLIP bus error vs TrainerState token counter bug; unrelated." + "reason": "MacOS CLIP bus error is unrelated to TrainerState token tracking." }, { "left": "issue:39379", "right": "issue:44554", "accept": false, - "reason": "Different correctness bugs in different attention/multimodal paths." + "reason": "mllama multimodal mask handling and MPS attention correctness are different bugs." }, { "left": "issue:42175", "right": "issue:43824", "accept": false, - "reason": "Backend packaging issue vs missing model class import; not the same bug." + "reason": "Backend installation metadata issue vs missing exported model class; unrelated." }, { "left": "issue:43697", "right": "issue:45357", "accept": false, - "reason": "RTDetrV2 output regression vs Qwen3.5 save_pretrained key regression; unrelated." + "reason": "RT-DETR output differences and Qwen3.5 save_pretrained key mapping are distinct issues." + }, + { + "left": "issue:44857", + "right": "issue:46832", + "accept": false, + "reason": "Both touch RT-DETR-ish code, but one is an AMP/CUDA loss crash and the other is a forward shape/type error when feature levels exceed backbone outputs." } ] }, "evaluator_result": { "accept": true, - "feedback": "The summary is grounded in the packet, and the verdicts stay conservative overall. The only accepted soft edge, 43329 vs 44479, is plausibly the same video-path regression bug; the rest are reasonably rejected as distinct failures." + "feedback": "The summary is grounded in the packet and stays conservative. The only accepted soft edge, issue:43329 \u2194 issue:44479, is supported by the issue descriptions and the broader vs. specific video-regression framing is appropriately cautious. The remaining pairs are rejected with plausible distinctions rather than overclaiming duplication." }, "evaluator_used": true, "retried": false, @@ -7069,12 +7099,12 @@ "issue:44792", "issue:44821", "issue:44849", - "issue:44857", "issue:44898", "issue:44936", "issue:44964", "issue:44987", "issue:44991", + "issue:45042", "issue:45081", "issue:45127", "issue:45137", @@ -7087,20 +7117,18 @@ "issue:46512", "issue:46693", "issue:46821", - "issue:46832", "issue:46869", "issue:47020" ], "soft_pairs": [ - "issue:44857|issue:46832", "issue:45137|issue:45405", "issue:43404|issue:45127", "issue:45081|issue:47020", "issue:28282|issue:29127", "issue:43742|issue:44291", "issue:44936|issue:46821", - "issue:44479|issue:46693", "issue:42915|issue:45538", + "issue:44479|issue:46693", "issue:45538|issue:46869", "issue:44062|issue:44991", "issue:24643|issue:29159", @@ -7112,224 +7140,225 @@ "issue:44849|issue:45701", "issue:42175|issue:44488", "issue:43540|issue:44560", - "issue:43540|issue:44479", "issue:44987|issue:45907", + "issue:43540|issue:44479", "issue:44291|issue:44821", "issue:44479|issue:45381", - "issue:38175|issue:44898", "issue:34567|issue:36331", + "issue:38175|issue:44898", "issue:44792|issue:44964", + "issue:44062|issue:45081", "issue:44568|issue:45488", - "issue:44062|issue:45081" + "issue:45042|issue:45081" ], - "prepared_review_unit_hash": "ecb93465c250a16be6c703a23bf35cc8cf7f35c9a0d16f19fa607add3b846429", + "prepared_review_unit_hash": "e8dd45ad15ae06cf75063db789668692ea6cb54800e8a910804d1c9801637394", "component_budget": { "node_count": 514, "item_count": 514, "soft_pair_count": 1891, - "serialized_chars": 980926, - "estimated_input_tokens": 245232, - "estimated_eval_tokens": 490720 + "serialized_chars": 980895, + "estimated_input_tokens": 245224, + "estimated_eval_tokens": 490704 }, "budget": { - "node_count": 48, - "item_count": 48, + "node_count": 47, + "item_count": 47, "soft_pair_count": 29, - "serialized_chars": 36412, - "estimated_input_tokens": 9103, - "estimated_eval_tokens": 18462 + "serialized_chars": 35832, + "estimated_input_tokens": 8958, + "estimated_eval_tokens": 18172 }, "overflow_policy": "truncate_then_skip", "trimmed": false, "aggressively_trimmed": false, "split": true, "analyst_result": { - "summary": "This cluster is mostly heterogeneous: it mixes tokenizer regressions, model-loading errors, trainer issues, and video-input bugs. Only one soft edge looks like the same underlying bug (the Qwen2.5-VL video-input regression).", - "confidence": 0.74, - "canonical_issue_reason": "No single canonical issue fits the whole cluster; the items span multiple unrelated bugs, with only a small Qwen video-input subgroup looking duplicate-like.", + "summary": "This cluster is mostly a loose semantic group of unrelated Transformers bug reports. Only the Qwen2.5-VL / Qwen3-VL video-input regression pair looks like the same underlying issue; the rest are distinct bugs in different models/code paths.", + "confidence": 0.84, + "canonical_issue_reason": "issue:44479 is the broadest and clearest report of the shared video-input regression, covering multiple Qwen multimodal variants; issue:45381 appears to be a narrower symptom/report of the same problem.", "canonical_pr_reason": null, - "best_issue_reason": "issue:45488 is the best representative of the tokenizer-regression subgroup because it is broad, active, and has multiple inbound references, but it is not a true canonical for the whole cluster.", + "best_issue_reason": "issue:44479 is the best representative issue because it captures the regression at the right abstraction level and scope, making it the most useful duplicate target.", "best_pr_reason": null, "soft_edge_verdicts": [ - { - "left": "issue:44857", - "right": "issue:46832", - "accept": false, - "reason": "Both are vision/detection failures, but they hit different models and code paths: LwDetrImageLoss AMP vs RTDetrModel feature-level handling." - }, { "left": "issue:45137", "right": "issue:45405", "accept": false, - "reason": "DeepSpeed ZeRO3 deque error and a MIN_PEFT_VERSION packaging/versioning issue are unrelated." + "reason": "DeepSpeed ZeRO3 deque/index failure and a PEFT version bump are unrelated." }, { "left": "issue:43404", "right": "issue:45127", "accept": false, - "reason": "Both mention tied embeddings, but one is a Mistral3 weight-tying bug and the other is LoRA merge collapse with extended vocab; different failure modes." + "reason": "Both mention tied weights, but one is a Mistral lm_head tying bug while the other is LoRA merge collapse with extended vocab; different code paths." }, { "left": "issue:45081", "right": "issue:47020", "accept": false, - "reason": "Different tokenizer classes and different bugs: Mistral regex patch backend access vs XLMRoberta/Unigram init mismatch." + "reason": "Different tokenizer classes and different failures: Mistral regex patch crash vs XLMRoberta Unigram init error." }, { "left": "issue:28282", "right": "issue:29127", "accept": false, - "reason": "Missing PyTorch import and a LayoutLMv3 box-validation error are unrelated." + "reason": "PyTorch missing import error is unrelated to LayoutLMV3's error-message clarity issue." }, { "left": "issue:43742", "right": "issue:44291", "accept": false, - "reason": "Both are loading-time TypeErrors, but one is a missing key for MobileLLM and the other is an init_empty_weights argument mismatch." + "reason": "MobileLLM loading key error and init_empty_weights unexpected argument are separate loading bugs." }, { "left": "issue:44936", "right": "issue:46821", "accept": false, - "reason": "Trainer evaluate-after-train failure and TimesFM window_size AttributeError are different subsystems and symptoms." + "reason": "Trainer evaluate-after-train failure is unrelated to TimesFM window_size AttributeError." }, { - "left": "issue:44479", - "right": "issue:46693", + "left": "issue:42915", + "right": "issue:45538", "accept": false, - "reason": "One is a video-input regression, the other is a long-context FA2 performance regression; not the same bug." + "reason": "FineGrainedFP8Config failure and CLIPTokenizer max-length behavior are unrelated." }, { - "left": "issue:42915", - "right": "issue:45538", + "left": "issue:44479", + "right": "issue:46693", "accept": false, - "reason": "Qwen3Moe FP8 failure and CLIPTokenizer model_max_length behavior are unrelated." + "reason": "Qwen video-input regression and FA2 long-context performance regression are different problems." }, { "left": "issue:45538", "right": "issue:46869", "accept": false, - "reason": "Tokenizer max-length behavior and YOLO multi-GPU device-map failure are unrelated." + "reason": "Tokenizer max-length handling is unrelated to YOLO multi-GPU device-map failure." }, { "left": "issue:44062", "right": "issue:44991", "accept": false, - "reason": "AddedToken special-argument duplication and est-roberta tokenizer loading are different tokenizer issues." + "reason": "AddedToken special-argument collision and est-roberta tokenizer loading failure are not the same bug." }, { "left": "issue:24643", "right": "issue:29159", "accept": false, - "reason": "DeepSpeed training weight shape error and slow/fast tokenizer inconsistency are unrelated." + "reason": "DeepSpeed training weight-shape error is unrelated to tokenizer slow/fast inconsistency." }, { "left": "issue:43452", "right": "issue:43582", "accept": false, - "reason": "GGUF from_pretrained loading breakage and Apple Silicon allocator warmup TypeError are unrelated." + "reason": "gguf_file loading and Apple Silicon caching_allocator_warmup TypeError are unrelated." }, { "left": "issue:27957", "right": "issue:28282", "accept": false, - "reason": "Flash Attention 2 on XLMRoberta and missing PyTorch are not the same problem." + "reason": "Flash Attention 2 with XLMRoberta is unrelated to the PyTorch-not-found import error." }, { "left": "issue:42915", "right": "issue:43749", "accept": false, - "reason": "Qwen3Moe FP8 failure and FSDP CPU RAM efficient loading are different training/loading paths." + "reason": "Qwen3Moe FP8 config failure and FSDP CPU RAM efficient loading are different training/loading bugs." }, { "left": "issue:43577", "right": "issue:46512", "accept": false, - "reason": "Both involve dtype, but one is a loaded-model dtype/qformer mismatch and the other is stale config dtype after from_config." + "reason": "Both involve dtype, but one is BLIP2 load-time dtype propagation and the other is stale config dtype from from_config; not the same concrete bug." }, { "left": "issue:41720", "right": "issue:43540", "accept": false, - "reason": "Auto device-map CUDA assert for Qwen3 and Qwen3OmniMoe video-input ValueError are different issues." + "reason": "Auto device-mapping cudaErrorAssert and Qwen3OmniMoe video ValueError are different failures." }, { "left": "issue:44849", "right": "issue:45701", "accept": false, - "reason": "Qwen3.5 output_hidden_states bug and generic tokenization changes are not the same underlying bug." + "reason": "Qwen3.5 hidden-states bug and version-dependent tokenization changes are unrelated." }, { "left": "issue:42175", "right": "issue:44488", "accept": false, - "reason": "Missing TensorFlow backend on pip install and failure to load a specific model are unrelated." + "reason": "Packaging/backend dependency issue and loading a specific model are unrelated." }, { "left": "issue:43540", "right": "issue:44560", "accept": false, - "reason": "Both are video-input problems, but they affect different Qwen model families and report different failure modes." + "reason": "Both are video-input issues in Qwen models, but they affect different models and error paths; not enough evidence of the same bug." }, { - "left": "issue:43540", - "right": "issue:44479", + "left": "issue:44987", + "right": "issue:45907", "accept": false, - "reason": "Same broad area, but these are separate model-specific video-input failures with different symptoms." + "reason": "Model loading failure and list_repo_files revision forwarding are unrelated." }, { - "left": "issue:44987", - "right": "issue:45907", + "left": "issue:43540", + "right": "issue:44479", "accept": false, - "reason": "Model-loading failure and list_repo_files revision forwarding are unrelated." + "reason": "Different Qwen multimodal models and different failures; too broad to treat as the same bug." }, { "left": "issue:44291", "right": "issue:44821", "accept": false, - "reason": "init_empty_weights argument mismatch and AutoImageProcessor-from-URL failure are unrelated." + "reason": "init_empty_weights constructor compatibility and AutoImageProcessor URL loading are separate issues." }, { "left": "issue:44479", "right": "issue:45381", "accept": true, - "reason": "These look like the same Qwen2.5-VL video-input regression: 45381 is a narrower report of the broader 44479 video-position/video-input bug." + "reason": "Both report the same Qwen2.5-VL/Qwen3-VL video-input regression, with 45381 describing the wrong vision_position_ids symptom of the broader 44479 issue." }, { - "left": "issue:38175", - "right": "issue:44898", + "left": "issue:34567", + "right": "issue:36331", "accept": false, - "reason": "Unexpected zero probabilities in SigLIP2 and Perceiver non-default-resolution failure are unrelated vision-model bugs." + "reason": "TrainerState token-count tracking and compute_loss signature incompatibility are unrelated trainer bugs." }, { - "left": "issue:34567", - "right": "issue:36331", + "left": "issue:38175", + "right": "issue:44898", "accept": false, - "reason": "TrainerState token-count tracking and CustomTrainer.compute_loss signature breakage are different trainer issues." + "reason": "Unexpected zero probabilities and Perceiver interpolation failure are different model-specific inference bugs." }, { "left": "issue:44792", "right": "issue:44964", "accept": false, - "reason": "Both involve multimodal models, but one is a failing test for Janus image generation and the other is a model-loading error for Phi-4 multimodal." + "reason": "Janus image-generation test failure and Phi-4 multimodal loading failure are unrelated." + }, + { + "left": "issue:44062", + "right": "issue:45081", + "accept": false, + "reason": "AddedToken special handling and Mistral regex patch crash are different tokenizer regressions." }, { "left": "issue:44568", "right": "issue:45488", "accept": false, - "reason": "Both are tokenizer regressions, but one is special-token addition for mDeBERTa and the other is LlamaTokenizer overriding tokenizer.json pre-tokenization." + "reason": "MDeBERTa special-token insertion and LlamaTokenizer ByteLevel/Metaspace override are distinct tokenizer regressions." }, { - "left": "issue:44062", + "left": "issue:45042", "right": "issue:45081", "accept": false, - "reason": "AddedToken keyword duplication and Mistral regex patch AttributeError are different tokenizer bugs." + "reason": "PIL/torchvision dependency behavior is unrelated to Mistral tokenizer regex patching." } ] }, "evaluator_result": { "accept": true, - "feedback": "The analyst summary is grounded in the packet and stays conservative. The cluster is indeed mostly heterogeneous, and the one accepted soft edge (issue:44479 vs issue:45381) is a plausible duplicate-like Qwen2.5-VL video-input regression. The rejected soft edges are also framed cautiously and do not overstate evidence." + "feedback": "The summary is well grounded in the packet and stays conservative. The canonical issue choice is supported by the titles: 44479 is the broader v5.3.0 video-input regression report, while 45381 is a narrower qwen2.5-vl symptom. The other soft-pair rejections are appropriately cautious given only title-level evidence." }, "evaluator_used": true, "retried": false, @@ -7374,11 +7403,10 @@ "issue:44821", "issue:44877", "issue:44913", - "issue:45042", "issue:45072", - "issue:45081", "issue:45216", "issue:45325", + "issue:45357", "issue:45405", "issue:45479", "issue:45588", @@ -7387,6 +7415,7 @@ "issue:45672", "issue:45674", "issue:45823", + "issue:45920", "issue:46121", "issue:46399", "issue:46463", @@ -7400,231 +7429,231 @@ "issue:46899" ], "soft_pairs": [ - "issue:45042|issue:45081", "issue:43257|issue:43756", - "issue:43582|issue:46821", "issue:43742|issue:44589", + "issue:43582|issue:46821", "issue:43531|issue:44821", "issue:43756|issue:46399", "issue:43756|issue:45325", "issue:45405|issue:45600", "issue:44737|issue:46463", "issue:43716|issue:46701", - "issue:43577|issue:45072", - "issue:45823|issue:46463", "issue:43906|issue:46693", + "issue:45823|issue:46463", + "issue:43577|issue:45072", "issue:44568|issue:46491", "issue:24643|issue:29870", "issue:45672|issue:45674", "issue:44625|issue:44877", - "issue:43540|issue:46693", "issue:43452|issue:44488", + "issue:43540|issue:46693", "issue:43906|issue:46468", "issue:28282|issue:32090", "issue:43697|issue:46829", "issue:44913|issue:46121", "issue:45663|issue:46846", - "issue:44737|issue:45588", "issue:43526|issue:45479", + "issue:44737|issue:45588", "issue:46463|issue:46899", - "issue:43697|issue:45216" + "issue:43697|issue:45216", + "issue:45357|issue:45920" ], - "prepared_review_unit_hash": "69bd00f67a1b3fec1585c8bca74216819a8791c4d2645bafaba9d66a4fb92c29", + "prepared_review_unit_hash": "c2621cb83b2c6171bd50bb6ab21de5b6642411dcad80b72e6cb3c649bdeb5b3f", "component_budget": { "node_count": 514, "item_count": 514, "soft_pair_count": 1891, - "serialized_chars": 980926, - "estimated_input_tokens": 245232, - "estimated_eval_tokens": 490720 + "serialized_chars": 980895, + "estimated_input_tokens": 245224, + "estimated_eval_tokens": 490704 }, "budget": { "node_count": 48, "item_count": 48, "soft_pair_count": 28, - "serialized_chars": 36615, - "estimated_input_tokens": 9154, - "estimated_eval_tokens": 18564 + "serialized_chars": 36581, + "estimated_input_tokens": 9146, + "estimated_eval_tokens": 18548 }, "overflow_policy": "truncate_then_skip", "trimmed": false, "aggressively_trimmed": false, "split": true, "analyst_result": { - "summary": "This cluster is highly heterogeneous: it mixes unrelated issues across loading, generation, quantization, tokenizer, config, and device/dtype bugs. None of the suggested soft pairs look like true duplicates or the same underlying fix target.", - "confidence": 0.93, - "canonical_issue_reason": "No single issue is a safe canonical target because the cluster does not center on one bug; the items span many distinct code paths and model families.", + "summary": "This is a heterogeneous issue cluster with superficial lexical overlap (mostly around model loading, dtype, rope, and config terms), but none of the soft pairs look like true duplicates of the same underlying bug.", + "confidence": 0.9, + "canonical_issue_reason": null, "canonical_pr_reason": null, - "best_issue_reason": "No issue is a strong global representative for deduping this cluster; any choice would be arbitrary because the topics are not the same bug.", + "best_issue_reason": "issue:43756 is the most connected node in the soft-similarity graph, but that connectivity is spurious; the cluster is too mixed for a strong canonical issue.", "best_pr_reason": null, "soft_edge_verdicts": [ - { - "left": "issue:45042", - "right": "issue:45081", - "accept": false, - "reason": "Both involve loading behavior, but one is about PIL image processors wrongly requiring torchvision and the other is a Mistral tokenizer regex patch crash; different code paths." - }, { "left": "issue:43257", "right": "issue:43756", "accept": false, - "reason": "Qwen3 MoE weight conversion with accelerate/deepspeed is unrelated to Smollm3 RoPE-layer dropping." + "reason": "Qwen3 MOE weight-conversion/loading bug vs. Smollm3 RoPE-layer mismatch; different models and failure modes." }, { - "left": "issue:43582", - "right": "issue:46821", + "left": "issue:43742", + "right": "issue:44589", "accept": false, - "reason": "Apple Silicon warmup TypeError and TimesFM window_size AttributeError are separate runtime issues in different components." + "reason": "MobileLLM key-loading error vs. Float8 storage lookup TypeError; unrelated load paths." }, { - "left": "issue:43742", - "right": "issue:44589", + "left": "issue:43582", + "right": "issue:46821", "accept": false, - "reason": "MobileLLM key lookup failure and Float8 storage lookup failure are different loading/type errors." + "reason": "Apple Silicon warmup TypeError vs. TimesFM window_size AttributeError; different APIs and devices." }, { "left": "issue:43531", "right": "issue:44821", "accept": false, - "reason": "Qwen3-MoE sliding_window behavior and AutoImageProcessor URL loading are unrelated." + "reason": "Qwen3-MoE sliding-window behavior vs. AutoImageProcessor URL loading; no shared code-path bug." }, { "left": "issue:43756", "right": "issue:46399", "accept": false, - "reason": "Smollm3 RoPE-layer count bug and Qwen3.5 kernel registration crash are different model-specific issues." + "reason": "Smollm3 RoPE-layer dropping vs. Qwen3.5 apply_rotary_pos_emb registration error; both rope-related but not the same bug." }, { "left": "issue:43756", "right": "issue:45325", "accept": false, - "reason": "Smollm3 RoPE-layer omission and Qwen2.5-VL temporal position scaling are distinct rotary-position bugs in different models." + "reason": "RoPE layer-count mismatch vs. Qwen2.5-VL temporal position scaling; distinct model-specific rotary issues." }, { "left": "issue:45405", "right": "issue:45600", "accept": false, - "reason": "PEFT minimum version bump and stale auto_mappings config references are unrelated maintenance/config issues." + "reason": "PEFT version-pin problem vs. stale auto_mappings reference breakage; unrelated maintenance issues." }, { "left": "issue:44737", "right": "issue:46463", "accept": false, - "reason": "XLNet CPU placement bug and T5 decoder-input validation error are different code paths and symptoms." + "reason": "XLNet CPU tensor allocation bug vs. T5 decoder input validation message; different subsystems entirely." }, { "left": "issue:43716", "right": "issue:46701", "accept": false, - "reason": "Mistral-3 image preprocessor dtype mismatch and DiffusionGemma pipeline failure are unrelated multimodal issues." + "reason": "Mistral-3 image dtype mismatch vs. DiffusionGemma pipeline output bug; different multimodal paths." }, { - "left": "issue:43577", - "right": "issue:45072", + "left": "issue:43906", + "right": "issue:46693", "accept": false, - "reason": "BLIP2 dtype persistence and SwitchTransformers/TimmWrapperModel bfloat16 mismatch are separate dtype-loading problems." + "reason": "A reproduction of an older issue vs. a FA2 long-context performance regression; not the same concrete defect." }, { "left": "issue:45823", "right": "issue:46463", "accept": false, - "reason": "Gemma4 device-map forward mismatch and T5 decoder validation are unrelated model execution bugs." + "reason": "Gemma4 PLE device-map mismatch vs. T5 decoder validation error; unrelated model families." }, { - "left": "issue:43906", - "right": "issue:46693", + "left": "issue:43577", + "right": "issue:45072", "accept": false, - "reason": "Isolated reproduction of another issue and FA2 long-context performance regression are not the same underlying bug." + "reason": "BLIP2 dtype propagation issue vs. SwitchTransformers/TimmWrapper bfloat16 mismatch; similar theme, different code paths." }, { "left": "issue:44568", "right": "issue:46491", "accept": false, - "reason": "Tokenizer special-token regression for mdeberta-v3-base and CodeLlama leading-space round-trip regression are different tokenizer behaviors." + "reason": "Tokenizer special-token regression vs. CodeLlama space round-trip regression; both tokenizer bugs but not the same behavior." }, { "left": "issue:24643", "right": "issue:29870", "accept": false, - "reason": "DeepSpeed 2-D weight error and Bart decoder_attention_mask assignment error are unrelated training/generation failures." + "reason": "DeepSpeed training weight-shape RuntimeError vs. Bart generate unbound variable; distinct training/generation failures." }, { "left": "issue:45672", "right": "issue:45674", "accept": false, - "reason": "Both are quantization-related, but one is a Gemma4 uint8 torch.finfo crash and the other is a BitsAndBytes lm_head exclusion bug; different fixes." + "reason": "Gemma4 4-bit audio/vision dtype cast bug vs. BitsAndBytes skip-module exclusion bug; different quantization failures." }, { "left": "issue:44625", "right": "issue:44877", "accept": false, - "reason": "Qwen3.5 num_labels propagation and strict config blocking granite_speech loading are different config issues." + "reason": "Qwen3.5 num_labels propagation vs. strict config blocking granite_speech; different configuration validations." }, { - "left": "issue:43540", - "right": "issue:46693", + "left": "issue:43452", + "right": "issue:44488", "accept": false, - "reason": "Qwen3OmniMoe video-processing ValueError and FA2 long-context performance regression are different problems." + "reason": "gguf_file loading regression vs. sleng-bert model loading failure; overlapping symptom only, not same root cause." }, { - "left": "issue:43452", - "right": "issue:44488", + "left": "issue:43540", + "right": "issue:46693", "accept": false, - "reason": "gguf_file breaking tokenizer/model loading and inability to load a specific checkpoint are separate model-loading issues." + "reason": "Qwen3OmniMoe video-input ValueError vs. FA2 long-context generation slowdown; unrelated bugs." }, { "left": "issue:43906", "right": "issue:46468", "accept": false, - "reason": "A reproduction of an existing issue is not the same as the official Nemotron example being broken." + "reason": "Reproduction of issue 38071 vs. broken Nemotron example; not enough evidence of same defect." }, { "left": "issue:28282", "right": "issue:32090", "accept": false, - "reason": "PyTorch-missing ImportError and Trainer _gpu_broadcast_one NoneType error are unrelated environment/runtime failures." + "reason": "Missing PyTorch import error vs. NoneType broadcast TypeError in Trainer; different layers of the stack." }, { "left": "issue:43697", "right": "issue:46829", "accept": false, - "reason": "RTDetrV2 output differences in v5 and VideoPrism hidden_states return bug are different model-output issues." + "reason": "RTDetrV2 output divergence vs. VideoPrism output field mismatch; different models and outputs." }, { "left": "issue:44913", "right": "issue:46121", "accept": false, - "reason": "GPTNeoX rotary_pct reload regression and rope-validation list TypeError are separate config serialization problems." + "reason": "GPTNeoX config reload issue vs. rope-parameter JSON list TypeError; unrelated config serialization bugs." }, { "left": "issue:45663", "right": "issue:46846", "accept": false, - "reason": "Gemma4 FSDP2 shared_kv_states bug and Qwen3.5 TP plan missing linear_attn entries are distinct training/sharding issues." + "reason": "Gemma4 FSDP2 shared_kv_states KeyError vs. Qwen3.5 tensor-parallel plan omission; different failure modes." }, { - "left": "issue:44737", - "right": "issue:45588", + "left": "issue:43526", + "right": "issue:45479", "accept": false, - "reason": "XLNet CPU arange bug and flash_attention s_aux=None crash are different code paths." + "reason": "BEiT image processor label reduction bug vs. sequence-classification zero-loss bug; not the same code path." }, { - "left": "issue:43526", - "right": "issue:45479", + "left": "issue:44737", + "right": "issue:45588", "accept": false, - "reason": "BeitImageProcessorFast label reduction bug and single-label classification zero-loss bug are unrelated processing vs loss-calculation issues." + "reason": "XLNet CPU allocation issue vs. flash-attention None-handling crash; different components." }, { "left": "issue:46463", "right": "issue:46899", "accept": false, - "reason": "T5 decoder-input validation and Gemma4 4-bit audio/vision dtype-cast bug are unrelated." + "reason": "T5 decoder validation message bug vs. Gemma4 multimodal 4-bit dtype cast; unrelated models and failures." }, { "left": "issue:43697", "right": "issue:45216", "accept": false, - "reason": "RTDetrV2 output mismatch and Qwen3.5 save_pretrained regression are different model behaviors and fixes." + "reason": "RTDetrV2 output mismatch vs. Qwen3.5 save_pretrained regression; no shared underlying bug." + }, + { + "left": "issue:45357", + "right": "issue:45920", + "accept": false, + "reason": "Qwen3.5 incorrect visual-encoder checkpoint saving vs. wrong token IDs in AutoTokenizer; separate serialization/tokenization issues." } ] }, @@ -7663,6 +7692,7 @@ "issue:43720", "issue:43742", "issue:43749", + "issue:43756", "issue:43761", "issue:43819", "issue:43994", @@ -7679,16 +7709,15 @@ "issue:45005", "issue:45200", "issue:45356", - "issue:45357", "issue:45372", "issue:45375", "issue:45542", "issue:45701", "issue:45704", - "issue:45920", "issue:46050", "issue:46170", "issue:46178", + "issue:46402", "issue:46463", "issue:46701", "issue:46829", @@ -7697,218 +7726,224 @@ "issue:47000" ], "soft_pairs": [ - "issue:45357|issue:45920", - "issue:44466|issue:45005", - "issue:43720|issue:44898", "issue:41762|issue:44936", + "issue:43720|issue:44898", + "issue:44466|issue:45005", "issue:45200|issue:46463", - "issue:44493|issue:46050", "issue:43643|issue:45375", + "issue:44493|issue:46050", + "issue:44361|issue:44442", "issue:43688|issue:43749", "issue:38175|issue:43295", - "issue:44361|issue:44442", "issue:43653|issue:44361", "issue:43452|issue:44843", - "issue:43819|issue:44811", - "issue:43994|issue:44898", "issue:43531|issue:43742", - "issue:44857|issue:47000", + "issue:43994|issue:44898", + "issue:43819|issue:44811", "issue:45372|issue:46170", "issue:43761|issue:46829", - "issue:29159|issue:33357", + "issue:44857|issue:47000", "issue:43525|issue:46701", + "issue:29159|issue:33357", "issue:45542|issue:46178", "issue:46832|issue:46932", - "issue:43688|issue:45704", "issue:43295|issue:45356", - "issue:43688|issue:45701", + "issue:43688|issue:45704", "issue:43653|issue:44568", - "issue:43646|issue:43749" + "issue:43688|issue:45701", + "issue:43646|issue:43749", + "issue:43756|issue:46402" ], - "prepared_review_unit_hash": "2803bfed095f975482fda7d201d5938e7939991de1c06af12ea65eec8522f5bd", + "prepared_review_unit_hash": "0c11754a2a6b92c4e089c9267e8696e84b37b35b3604bf9b773e9024af8e251a", "component_budget": { "node_count": 514, "item_count": 514, "soft_pair_count": 1891, - "serialized_chars": 980926, - "estimated_input_tokens": 245232, - "estimated_eval_tokens": 490720 + "serialized_chars": 980895, + "estimated_input_tokens": 245224, + "estimated_eval_tokens": 490704 }, "budget": { "node_count": 47, "item_count": 47, "soft_pair_count": 27, - "serialized_chars": 35560, - "estimated_input_tokens": 8890, - "estimated_eval_tokens": 18036 + "serialized_chars": 35499, + "estimated_input_tokens": 8875, + "estimated_eval_tokens": 18006 }, "overflow_policy": "truncate_then_skip", "trimmed": false, "aggressively_trimmed": false, "split": true, "analyst_result": { - "summary": "These items are mostly distinct bug reports spanning tokenizers, vision processors, model loading, and backend/import edge cases. I found no soft pair that is clearly the same underlying issue with enough confidence to merge; the closest cases are related by theme but still appear to hit different code paths or symptoms. The broadest, most representative issue is the v5 AutoTokenizer token-ID regression affecting many model families.", - "confidence": 0.78, - "canonical_issue_reason": "issue:45920 is the most canonical because it describes a broad, concrete regression with multiple affected model families and a clear shared failure mode (wrong token IDs in AutoTokenizer).", + "summary": "The soft pairs are mostly false positives: they share broad subsystems (tokenizers, loading, vision models, losses) but not the same concrete bug or code path. Overall this looks like a heterogeneous set of unrelated issues rather than a duplicate cluster.", + "confidence": 0.96, + "canonical_issue_reason": "No single issue cleanly represents a true duplicate set here; the items are heterogeneous. If one must be chosen, #43646 is the broadest umbrella-style regression report about v5 breaking custom model initialization, though it does not subsume the others.", "canonical_pr_reason": null, - "best_issue_reason": "issue:45920 is the best single issue to anchor this set: it is high-signal, regression-focused, and broader than the more model-specific reports.", + "best_issue_reason": "#43646 is the most general and representative issue among the set, but only as a broad regression summary rather than a duplicate canonical.", "best_pr_reason": null, "soft_edge_verdicts": [ { - "left": "issue:45357", - "right": "issue:45920", - "accept": false, - "reason": "Both are tokenizer/model-regression flavored, but one is about incorrect visual encoder keys on save_pretrained and the other about wrong token IDs from AutoTokenizer; different symptoms and likely different fixes." - }, - { - "left": "issue:44466", - "right": "issue:45005", + "left": "issue:41762", + "right": "issue:44936", "accept": false, - "reason": "Both mention tied weights in v5, but one is about device-dependent lm_head.weight serialization while the other is broader translation-model tied-weight behavior; related area, not clearly the same bug." + "reason": "Different failures: Gemma3 load-time ZeRO-3 index error vs trainer.evaluate() after training. No shared code path." }, { "left": "issue:43720", "right": "issue:44898", "accept": false, - "reason": "Different subsystems and failures: BitNet packed-weight loading vs Perceiver positional interpolation at non-default resolutions." + "reason": "Unrelated models and bugs: BitNet packed-weight loading vs Perceiver non-default resolution handling." }, { - "left": "issue:41762", - "right": "issue:44936", + "left": "issue:44466", + "right": "issue:45005", "accept": false, - "reason": "Gemma3 ZeRO-3 loading IndexError and trainer.evaluate-after-train failure are unrelated runtime paths." + "reason": "Both mention tied weights, but one is device-dependent lm_head serialization and the other is translation-model tied-weight issues; too broad to treat as the same bug." }, { "left": "issue:45200", "right": "issue:46463", "accept": false, - "reason": "Gemma 4 mm_token_type_ids handling is unrelated to T5 decoder input validation error messaging." + "reason": "Different model families and symptoms: Gemma 4 mm_token_type_ids defaulting vs T5 decoder-input validation message." + }, + { + "left": "issue:43643", + "right": "issue:45375", + "accept": false, + "reason": "Both are config-field related, but one is a generic AutoConfig trust_remote_code issue and the other is a specific strict-drop of a Gemma/Qwen vision field." }, { "left": "issue:44493", "right": "issue:46050", "accept": false, - "reason": "Position-id unexpected-key loading issue and sinq quantization save/load breakage are different model-state serialization problems." + "reason": "State-dict/key mismatch vs quantized save/load breakage; similar theme but not the same underlying defect." }, { - "left": "issue:43643", - "right": "issue:45375", + "left": "issue:44361", + "right": "issue:44442", "accept": false, - "reason": "Both concern missing config fields, but one is a generic AutoConfig trust_remote_code return issue and the other is a specific Qwen vision config field dropped by strict parsing." + "reason": "Both are tokenizer-loading failures, but they affect different tokenizers and different failure modes." }, { "left": "issue:43688", "right": "issue:43749", "accept": false, - "reason": "Auxiliary-loss normalization in MoE models is unrelated to FSDP CPU RAM efficient loading." + "reason": "Aux-loss normalization bug vs FSDP CPU RAM efficient loading regression; unrelated subsystems." }, { "left": "issue:38175", "right": "issue:43295", "accept": false, - "reason": "SigLIP2 zero-probability output and processor.tokenizer regression are both vision-related but not the same code-path bug." - }, - { - "left": "issue:44361", - "right": "issue:44442", - "accept": false, - "reason": "Different tokenizers and failures: MLuke task-time AttributeError vs AutoTokenizer failing to load FastSpeech2ConformerTokenizer." + "reason": "Different problems: SigLIP2 zero probabilities vs processor.tokenizer/custom image-tokenizer regression." }, { "left": "issue:43653", "right": "issue:44361", "accept": false, - "reason": "BigBird special-token registration bug and MLuke task AttributeError are not the same underlying issue." + "reason": "BigBird special-token registration bug and MLukeTokenizer AttributeError are separate tokenizer issues." }, { "left": "issue:43452", "right": "issue:44843", "accept": false, - "reason": "gguf_file loading breakage and unconditional model_info() call in offline mode are separate AutoTokenizer/from_pretrained problems." + "reason": "Both involve AutoTokenizer.from_pretrained, but the causes differ: gguf_file loading vs an unconditional model_info() call breaking offline mode." }, { - "left": "issue:43819", - "right": "issue:44811", + "left": "issue:43531", + "right": "issue:43742", "accept": false, - "reason": "DAC.from_latents forward mismatch and Whisper batch_decode skip_special_tokens handling affect different components." + "reason": "Qwen3-MoE sliding_window behavior and MobileLLM key error are unrelated." }, { "left": "issue:43994", "right": "issue:44898", "accept": false, - "reason": "Both are vision-model inference issues, but SigLIP2 nonsensical outputs and Perceiver non-default-resolution failure are different bugs." - }, - { - "left": "issue:43531", - "right": "issue:43742", - "accept": false, - "reason": "Qwen3-MoE sliding_window behavior and MobileLLM-125M load KeyError are unrelated." + "reason": "Both are vision-related inference issues, but with different models and symptoms; not the same bug." }, { - "left": "issue:44857", - "right": "issue:47000", + "left": "issue:43819", + "right": "issue:44811", "accept": false, - "reason": "Both are detection-loss crashes, but one is float16/CUDA AMP in LwDetrImageLoss and the other is infeasible cost matrix in RTDetrLoss; not the same defect." + "reason": "DAC forward/from_latents mismatch is unrelated to Whisper batch_decode skip_special_tokens handling." }, { "left": "issue:45372", "right": "issue:46170", "accept": false, - "reason": "Gemma 4 processor loading import failure from mistral_common and Windows ROCm torch.distributed dependency errors are different platform/dependency issues." + "reason": "Different import/dependency failures on different platforms and libraries." }, { "left": "issue:43761", "right": "issue:46829", "accept": false, - "reason": "CLIPVisionModel missing hidden_states with a flag set and VideoPrism returning last_hidden_state as hidden_states are different output-contract bugs." + "reason": "Both concern hidden_states outputs, but they are different models and different return-path bugs." }, { - "left": "issue:29159", - "right": "issue:33357", + "left": "issue:44857", + "right": "issue:47000", "accept": false, - "reason": "Tokenizer inconsistency and a MacOS bus error in a CLIP model are unrelated." + "reason": "Different detection-loss crashes with different error conditions; not one concrete defect." }, { "left": "issue:43525", "right": "issue:46701", "accept": false, - "reason": "Missing pad_token_id on Llama4Config and DiffusionGemma generation/pipeline failure are different model-initialization problems." + "reason": "Llama4Config missing pad_token_id is unrelated to DiffusionGemma pipeline failure." + }, + { + "left": "issue:29159", + "right": "issue:33357", + "accept": false, + "reason": "Tokenizer consistency issue vs MacOS bus error in a CLIP community model; unrelated." }, { "left": "issue:45542", "right": "issue:46178", "accept": false, - "reason": "Both are backend-selection/import issues, but tensorboard-without-TensorFlow and tensorflow_text backend mapping are separate dependency bugs." + "reason": "Both concern optional TensorFlow-related backends, but they are distinct backend-resolution bugs with different offending packages." }, { "left": "issue:46832", "right": "issue:46932", "accept": false, - "reason": "RT-DETR feature-level mismatch and Gemma4 quantized embedder dtype casting are unrelated." + "reason": "RT-DETR feature-level mismatch and Gemma4 dtype casting for quantized embedders are unrelated." + }, + { + "left": "issue:43295", + "right": "issue:45356", + "accept": false, + "reason": "Both are regressions around tokenizers/processors, but the concrete breakage and affected models differ." }, { "left": "issue:43688", "right": "issue:45704", "accept": false, - "reason": "MoE auxiliary-loss normalization is unrelated to T5 silently using apex.FusedRMSNorm." + "reason": "Aux-loss normalization and T5 apex.FusedRMSNorm memory leak are unrelated." }, { - "left": "issue:43295", - "right": "issue:45356", + "left": "issue:43653", + "right": "issue:44568", "accept": false, - "reason": "Processor/tokenizer regression in custom model code and Kimi-K2.5 codec handling/regression are different tokenizer issues." + "reason": "Both are special-token issues, but they affect different tokenizers and different token types." }, { "left": "issue:43688", "right": "issue:45701", "accept": false, - "reason": "Auxiliary-loss normalization and version-dependent tokenization behavior are unrelated." + "reason": "Aux-loss normalization bug is not the same as a broad tokenization-version-change report." }, { - "left": "issue:43653", - "right": "issue:44568", + "left": "issue:43646", + "right": "issue:43749", + "accept": false, + "reason": "Generic v5 custom-model initialization regression vs FSDP loading breakage; different code paths." + }, + { + "left": "issue:43756", + "right": "issue:46402", "accept": false, - "reason": "BigBird mask-token registration and MDeBERTa BOS/EOS add_special_tokens regression are different tokenizer special-token behaviors." + "reason": "Smollm3 RoPE-layer handling and Sam3Model text encoder weight loading are unrelated model-specific bugs." } ] }, @@ -7950,7 +7985,6 @@ "issue:43653", "issue:43688", "issue:43720", - "issue:43756", "issue:43761", "issue:43782", "issue:43927", @@ -7960,6 +7994,7 @@ "issue:44484", "issue:44514", "issue:44554", + "issue:44610", "issue:44625", "issue:44704", "issue:44734", @@ -7973,7 +8008,6 @@ "issue:45491", "issue:46002", "issue:46050", - "issue:46402", "issue:46491", "issue:46612", "issue:46693", @@ -7982,245 +8016,245 @@ "issue:47013" ], "soft_pairs": [ - "issue:43756|issue:46402", "issue:44464|issue:44743", - "issue:43452|issue:43577", "issue:43653|issue:43927", "issue:44821|issue:46821", "issue:42898|issue:46491", + "issue:43452|issue:43577", "issue:44843|issue:44991", "issue:44479|issue:44625", "issue:43761|issue:47013", "issue:33666|issue:36010", - "issue:32090|issue:33666", "issue:44484|issue:44704", + "issue:32090|issue:33666", "issue:44466|issue:46050", - "issue:41720|issue:44464", "issue:43688|issue:45440", "issue:44977|issue:46693", - "issue:44464|issue:44977", + "issue:41720|issue:44464", "issue:44734|issue:46612", + "issue:44464|issue:44977", "issue:43504|issue:43550", "issue:33666|issue:36331", - "issue:43262|issue:43295", - "issue:43278|issue:43688", "issue:43504|issue:43720", + "issue:43278|issue:43688", + "issue:43262|issue:43295", "issue:43782|issue:46693", "issue:45491|issue:46002", "issue:43720|issue:43927", "issue:43577|issue:44554", "issue:43611|issue:46846", "issue:43606|issue:43927", + "issue:43504|issue:44610", "issue:44514|issue:45290" ], - "prepared_review_unit_hash": "3daafd649e64ec2dde90a93dd826a1cb46721d235689d19b2da23d81ac48b15c", + "prepared_review_unit_hash": "889eaa35b3e971e30337e7b1557019daf59d33f754a221e674f459fe30a34ab9", "component_budget": { "node_count": 514, "item_count": 514, "soft_pair_count": 1891, - "serialized_chars": 980926, - "estimated_input_tokens": 245232, - "estimated_eval_tokens": 490720 + "serialized_chars": 980895, + "estimated_input_tokens": 245224, + "estimated_eval_tokens": 490704 }, "budget": { - "node_count": 48, - "item_count": 48, + "node_count": 47, + "item_count": 47, "soft_pair_count": 30, - "serialized_chars": 37020, - "estimated_input_tokens": 9255, - "estimated_eval_tokens": 18766 + "serialized_chars": 36445, + "estimated_input_tokens": 9112, + "estimated_eval_tokens": 18480 }, "overflow_policy": "truncate_then_skip", "trimmed": false, "aggressively_trimmed": false, "split": true, "analyst_result": { - "summary": "This is a loose semantic cluster, not a true duplicate set: the issues span many unrelated Transformers regressions across loading, tokenization, processors, training, generation, and model-specific paths. All soft edges look superficial and should be rejected.", - "confidence": 0.12, - "canonical_issue_reason": "No single canonical issue: the set mixes unrelated bugs across tokenizers, processors, generation, training, and model-loading paths.", + "summary": "This is a loose themed collection of Transformers issues, but the proposed soft links mostly connect different bugs in different models or code paths (loading, generation, tokenization, dtype, device placement, etc.). I would reject all soft edges as duplicates.", + "confidence": 0.94, + "canonical_issue_reason": "No canonical issue: the set fragments into many unrelated bugs, so no single issue cleanly represents the cluster.", "canonical_pr_reason": null, - "best_issue_reason": "No issue is globally suitable as a cluster anchor; this should be split by subsystem/model family rather than merged around one report.", + "best_issue_reason": "If forced to pick a loose anchor, issue 44514 is the closest to a repeated apply_chat_template crash pattern, but it still only covers a small slice of the set and is not a true duplicate hub.", "best_pr_reason": null, "soft_edge_verdicts": [ - { - "left": "issue:43756", - "right": "issue:46402", - "accept": false, - "reason": "Different models and failures: Smollm3 RoPE-layer behavior vs Sam3 text-encoder weights not loading." - }, { "left": "issue:44464", "right": "issue:44743", "accept": false, - "reason": "Both are generation bugs, but one is compiled-forward chunked decoding and the other is recurrent-state reset in Qwen3.5 cache handling." - }, - { - "left": "issue:43452", - "right": "issue:43577", - "accept": false, - "reason": "Unrelated paths: GGUF loading breaks tokenizer/model imports, while BLIP2 has a dtype propagation issue." + "reason": "Both involve generation/cache behavior, but one is a compiled-forward inconsistency and the other is a recurrent-state unpacking bug in a different model path." }, { "left": "issue:43653", "right": "issue:43927", "accept": false, - "reason": "Both touch tokenization/config, but BigBird special-token decode and DiaConfig token IDs are different bugs." + "reason": "BigBird tokenizer special-token decoding and DiaConfig token-ID persistence are different failure modes and different subsystems." }, { "left": "issue:44821", "right": "issue:46821", "accept": false, - "reason": "Different subsystems: AutoImageProcessor URL loading vs TimesFM window_size attribute access." + "reason": "AutoImageProcessor URL loading and TimesFM window_size AttributeError are unrelated loader/model issues." }, { "left": "issue:42898", "right": "issue:46491", "accept": false, - "reason": "Both concern tokenizer behavior, but one is a v5 cleanup default change and the other is CodeLlama leading-space round-trip regression." + "reason": "Tokenization-space cleanup regression and CodeLlama leading-space round-trip regression are similar only at a very high level, not the same bug." + }, + { + "left": "issue:43452", + "right": "issue:43577", + "accept": false, + "reason": "GGUF loading breakage and BLIP2 dtype propagation are distinct loading vs dtype issues." }, { "left": "issue:44843", "right": "issue:44991", "accept": false, - "reason": "Offline hub lookup in _patch_mistral_regex is unrelated to EMBEDDIA tokenizer loading failure." + "reason": "Offline model_info() behavior in tokenizer patching and a specific tokenizer load failure for EMBEDDIA/est-roberta are not the same root cause." }, { "left": "issue:44479", "right": "issue:44625", "accept": false, - "reason": "Qwen3.5 video-input regression and num_labels propagation are separate code paths." + "reason": "Qwen video-input regression and Qwen3.5 num_labels propagation affect different code paths and symptoms." }, { "left": "issue:43761", "right": "issue:47013", "accept": false, - "reason": "Different model families and bugs: CLIPVision hidden_states vs xLSTM train-mode return shape." + "reason": "CLIPVision hidden_states handling and xLSTM train-mode state unpacking are unrelated model bugs." }, { "left": "issue:33666", "right": "issue:36010", "accept": false, - "reason": "Multi-GPU Qwen2-VL training is unrelated to the GenerationMixin import error." - }, - { - "left": "issue:32090", - "right": "issue:33666", - "accept": false, - "reason": "Trainer GPU broadcast NoneType crash is a different training failure than Qwen2-VL multi-GPU behavior." + "reason": "Qwen2-VL multi-GPU training and GenerationMixin import failure are completely different issues." }, { "left": "issue:44484", "right": "issue:44704", "accept": false, - "reason": "A save_pretrained shard-size question is unrelated to AutoProcessor kwargs not being forwarded." + "reason": "save_pretrained shard-size behavior and AutoProcessor cached_file kwarg forwarding are unrelated API problems." }, { - "left": "issue:44466", - "right": "issue:46050", + "left": "issue:32090", + "right": "issue:33666", "accept": false, - "reason": "Different persistence bugs: lm_head serialization/tied weights vs sinq quantization save/load breakage." + "reason": "Trainer gpu_broadcast NoneType error and Qwen2-VL multi-GPU training are not the same defect." }, { - "left": "issue:41720", - "right": "issue:44464", + "left": "issue:44466", + "right": "issue:46050", "accept": false, - "reason": "Qwen3 auto device-mapping CUDA assert is unrelated to compiled-forward chunked generation inconsistency." + "reason": "Serialization inconsistency of lm_head.weight and sinq quantized save/load breakage are different persistence bugs." }, { "left": "issue:43688", "right": "issue:45440", "accept": false, - "reason": "Aux-loss normalization in MoE models is a different problem from DeepSeekV3 implementation divergence." + "reason": "Auxiliary-loss normalization and DeepSeekV3MoE divergence are separate model-behavior issues." }, { "left": "issue:44977", "right": "issue:46693", "accept": false, - "reason": "Both involve generation regressions, but one is Qwen3.5 flash-attention correctness and the other is a FA2 long-context performance regression." + "reason": "Qwen3.5 flash-attention generation failure and a long-context FA2 performance regression are different symptoms and likely different causes." }, { - "left": "issue:44464", - "right": "issue:44977", + "left": "issue:41720", + "right": "issue:44464", "accept": false, - "reason": "Compiled-forward chunked generation and Qwen3.5 flash-attention failures do not share the same underlying bug." + "reason": "Auto device mapping cuda assert on Qwen3 and chunked-generation compiled-forward inconsistency are unrelated generation/runtime bugs." }, { "left": "issue:44734", "right": "issue:46612", "accept": false, - "reason": "Both are cache/generation issues, but KV-cache continuation indexing and beam-search cache reorder skipping are different mechanisms." + "reason": "Serving KV-cache continuation indexing and beam-search cache reorder being skipped are both cache-related, but they are distinct code paths and bugs." + }, + { + "left": "issue:44464", + "right": "issue:44977", + "accept": false, + "reason": "Compiled-forward chunked generation inconsistency and Qwen3.5 flash-attention output failure are not the same underlying problem." }, { "left": "issue:43504", "right": "issue:43550", "accept": false, - "reason": "Beit legacy-field loading and Bamba torch.compile SDPA failures are unrelated." + "reason": "Legacy field loading for Beit and Bamba torch.compile+SDPA failure are unrelated model-loading/runtime bugs." }, { "left": "issue:33666", "right": "issue:36331", "accept": false, - "reason": "Multi-GPU training and CustomTrainer compute_loss signature mismatch are separate trainer issues." + "reason": "Qwen2-VL multi-GPU training and CustomTrainer.compute_loss signature mismatch are different Trainer-side issues." }, { - "left": "issue:43262", - "right": "issue:43295", + "left": "issue:43504", + "right": "issue:43720", "accept": false, - "reason": "Processor chat-template sampling-rate defaults and processor.tokenizer regression are different API breakages." + "reason": "Beit legacy-field load failure and BitNet packed-weight unpacking during accelerate loading are separate loading problems." }, { "left": "issue:43278", "right": "issue:43688", "accept": false, - "reason": "Embedding dtype drift in eval is unrelated to MoE auxiliary-loss normalization." + "reason": "Embedding dtype drift in eval and auxiliary-loss normalization are different model numerics issues." }, { - "left": "issue:43504", - "right": "issue:43720", + "left": "issue:43262", + "right": "issue:43295", "accept": false, - "reason": "Model preset loading and BitNet packed-weight unpacking are distinct load-path bugs." + "reason": "Audio chat-template sampling-rate default and processor.tokenizer/image-handling regression are unrelated processor bugs." }, { "left": "issue:43782", "right": "issue:46693", "accept": false, - "reason": "Qwen3VL weight_only loading error is unrelated to FA2 long-context performance regression." + "reason": "Qwen3VL weight_only loading error and FA2 long-context performance regression do not share a concrete code-path failure." }, { "left": "issue:45491", "right": "issue:46002", "accept": false, - "reason": "Both mention NaNs, but Gemma3 sliding-window attention and classification-head checkout/init issues are different root causes." + "reason": "Gemma3 mixed-length NaN embeddings and classification-head NaNs on checkout/init are similar only by symptom, not root cause." }, { "left": "issue:43720", "right": "issue:43927", "accept": false, - "reason": "BitNet packed-weight handling has nothing to do with DiaConfig custom token IDs." + "reason": "BitNet packed-weight loading and DiaConfig token-ID persistence are unrelated." }, { "left": "issue:43577", "right": "issue:44554", "accept": false, - "reason": "BLIP2 dtype loading and the MPS attention correctness bug are unrelated." + "reason": "BLIP2 dtype mismatch and an MPS attention correctness issue are different bugs in different execution paths." }, { "left": "issue:43611", "right": "issue:46846", "accept": false, - "reason": "base_model_prefix loading regression is not the same as Qwen3.5 linear_attn TP-plan OOM/shape errors." + "reason": "base_model_prefix loading regression and missing linear_attn TP-plan entries are separate model-compatibility issues." }, { "left": "issue:43606", "right": "issue:43927", "accept": false, - "reason": "CPU offload device mismatch for bark-small is unrelated to DiaConfig token-ID loss." + "reason": "CPU offload device mismatch for bark-small and DiaConfig token-ID loss are unrelated." + }, + { + "left": "issue:43504", + "right": "issue:44610", + "accept": false, + "reason": "Beit legacy-field loading and OmDet processor/model-size mismatch are not the same underlying bug." }, { "left": "issue:44514", "right": "issue:45290", "accept": false, - "reason": "Both are apply_chat_template crashes, but batched padding=False and assistant tool-call/no-content are different edge cases." + "reason": "Both hit apply_chat_template, but one is batched padding=False and the other is assistant tool-call/no-content handling; these are distinct failures." } ] }, @@ -8253,7 +8287,6 @@ "issue:43295", "issue:43452", "issue:43493", - "issue:43504", "issue:43525", "issue:43572", "issue:43575", @@ -8274,7 +8307,6 @@ "issue:44561", "issue:44568", "issue:44589", - "issue:44610", "issue:44625", "issue:44792", "issue:44843", @@ -8291,10 +8323,11 @@ "issue:46209", "issue:46365", "issue:46489", - "issue:46710" + "issue:46710", + "issue:46821", + "issue:46832" ], "soft_pairs": [ - "issue:43504|issue:44610", "issue:43525|issue:43572", "issue:43643|issue:44561", "issue:43720|issue:44361", @@ -8307,226 +8340,233 @@ "issue:43452|issue:44451", "issue:45823|issue:46050", "issue:39692|issue:43295", - "issue:43994|issue:46192", "issue:42796|issue:44977", + "issue:43994|issue:46192", "issue:43493|issue:43844", "issue:43278|issue:43844", - "issue:43646|issue:44466", "issue:43278|issue:43701", + "issue:43646|issue:44466", "issue:46209|issue:46710", "issue:44568|issue:46489", "issue:33357|issue:33666", "issue:43927|issue:44361", - "issue:43575|issue:45405", - "issue:44792|issue:45405", - "issue:44568|issue:45812", "issue:44849|issue:46209", - "issue:44488|issue:44843", + "issue:44568|issue:45812", + "issue:44792|issue:45405", + "issue:43575|issue:45405", "issue:43572|issue:44291", - "issue:44589|issue:45520" + "issue:44488|issue:44843", + "issue:44589|issue:45520", + "issue:46821|issue:46832" ], - "prepared_review_unit_hash": "c08c45d9e3ee5f36eda7e934b1c9c5223e224d13d58cce6a151545676e16d634", + "prepared_review_unit_hash": "e9051c3688997d75a744d1b8a4f7f258a9cd5dc8afbfab9ef3fa98f6a715768f", "component_budget": { "node_count": 514, "item_count": 514, "soft_pair_count": 1891, - "serialized_chars": 980926, - "estimated_input_tokens": 245232, - "estimated_eval_tokens": 490720 + "serialized_chars": 980895, + "estimated_input_tokens": 245224, + "estimated_eval_tokens": 490704 }, "budget": { "node_count": 48, "item_count": 48, "soft_pair_count": 30, - "serialized_chars": 36759, - "estimated_input_tokens": 9190, - "estimated_eval_tokens": 18636 + "serialized_chars": 36734, + "estimated_input_tokens": 9184, + "estimated_eval_tokens": 18624 }, "overflow_policy": "truncate_then_skip", "trimmed": false, "aggressively_trimmed": false, "split": true, "analyst_result": { - "summary": "This cluster is a loose mix of unrelated Transformers regressions, mostly around tokenizer/config/load-path breakages in v5. None of the soft pairs look like true duplicates, so I reject all of them; the best representative issue is the broad Granite tokenizer regression.", - "confidence": 0.31, - "canonical_issue_reason": "issue:45812 is the broadest, most representative user-facing regression here: a silent v4\u2192v5 tokenizer break that affects all Granite models and matches the cluster\u2019s overall tokenizer/config regression theme.", + "summary": "The set is mostly a collection of unrelated standalone issues rather than one duplicate cluster. I would not merge any of the soft pairs; no PRs are present.", + "confidence": 0.97, + "canonical_issue_reason": null, "canonical_pr_reason": null, - "best_issue_reason": "Same reason: it is the clearest, widest-scope issue in the set and the most plausible anchor for the surrounding tokenizer-related reports.", + "best_issue_reason": "issue:45812 is the strongest standalone bug report here: it is concrete, broad in impact, and describes a clear v4\u2192v5 tokenizer regression affecting all Granite models.", "best_pr_reason": null, "soft_edge_verdicts": [ - { - "left": "issue:43504", - "right": "issue:44610", - "accept": false, - "reason": "Different models and failure modes: BEiT legacy-field loading bug vs OmDet processor/image-size mismatch." - }, { "left": "issue:43525", "right": "issue:43572", "accept": false, - "reason": "Both are config-attribute regressions, but they affect different fields and model families; no concrete shared bug path." + "reason": "Both mention missing config token fields after v5, but they affect different models and different fields (`pad_token_id` vs `pad_token_idx`), so this does not look like the same bug." }, { "left": "issue:43643", "right": "issue:44561", "accept": false, - "reason": "Both mention trust_remote_code, but one is missing fields in AutoConfig and the other is a removed utility breaking imports." + "reason": "Both involve `trust_remote_code`, but one is about `AutoConfig` returning missing fields while the other is about removal of `is_torch_fx_available`; these are different regressions and code paths." }, { "left": "issue:43720", "right": "issue:44361", "accept": false, - "reason": "Unrelated subsystems: BitNet accelerate-loading bug vs MLuke tokenizer task AttributeError." + "reason": "Completely different subsystems: BitNet loading vs MLukeTokenizer task AttributeError." }, { "left": "issue:44964", "right": "issue:45405", "accept": false, - "reason": "Model load failure vs dependency version bump issue; not the same bug." + "reason": "A model-loading bug versus a dependency-version bump issue; no shared underlying defect." }, { "left": "issue:44625", "right": "issue:45584", "accept": false, - "reason": "Different models and paths: Qwen3.5 num_labels propagation vs Whisper empty-transcription generation." + "reason": "Different models and failures: `num_labels` propagation vs Whisper empty-transcription generation." }, { "left": "issue:43493", "right": "issue:43825", "accept": false, - "reason": "SigLIP2 implementation mismatch and a pipeline error message regression are unrelated." + "reason": "Both are around transformers behavior, but one is SigLIP2 correctness versus a pipeline error message regression; not the same underlying bug." }, { "left": "issue:44977", "right": "issue:46209", "accept": false, - "reason": "Different hardware/runtime issues: flash-attention generation bug vs FP8/B200 inference failure." + "reason": "Unrelated failures: Qwen3.5 flash-attention generation vs FP8 inference on B200." }, { "left": "issue:44554", "right": "issue:46365", "accept": false, - "reason": "MPS attention correctness and Trainer import dependency conflict are unrelated." + "reason": "Different areas entirely: MPS attention correctness vs import-time Trainer/datasets issue." }, { "left": "issue:28218", "right": "issue:33357", "accept": false, - "reason": "Tokenizer spacing bug vs MacOS CLIP bus error; no shared code-path." + "reason": "Tokenizer spacing bug versus a MacOS bus error in CLIP loading; no common code-path bug." }, { "left": "issue:43452", "right": "issue:44451", "accept": false, - "reason": "Both are load failures, but for different causes and models; not a duplicate." + "reason": "Both are model/tokenizer loading complaints, but one is about `gguf_file` handling and the other is a specific BERT model load failure; not the same defect." }, { "left": "issue:45823", "right": "issue:46050", "accept": false, - "reason": "Device-mapped Gemma4 forward mismatch is unrelated to sinq save/load quantization breakage." + "reason": "Device mismatch in Gemma4 forward pass versus broken quantized model save/load; unrelated." }, { "left": "issue:39692", "right": "issue:43295", "accept": false, - "reason": "SigLIP2 docs/example errors and processor/tokenizer regression are different problems." + "reason": "Documentation example errors versus a regression in custom model/processor code handling." }, { - "left": "issue:43994", - "right": "issue:46192", + "left": "issue:42796", + "right": "issue:44977", "accept": false, - "reason": "AutoModel/pipeline nonsense outputs vs return_tensors being ignored are separate API issues." + "reason": "Tokenizer BPE-marker handling versus flash-attention generation; different bugs." }, { - "left": "issue:42796", - "right": "issue:44977", + "left": "issue:43994", + "right": "issue:46192", "accept": false, - "reason": "Tokenizer BPE-marker handling vs Qwen3.5 flash-attention generation bug are unrelated." + "reason": "SigLIP2 inference outputs versus `return_tensors` being ignored; unrelated API bugs." }, { "left": "issue:43493", "right": "issue:43844", "accept": false, - "reason": "SigLIP2 implementation mismatch vs ZeRO-3 gradient inflation are unrelated training/model issues." + "reason": "Model implementation discrepancy versus a DeepSpeed gradient increase issue; different failures." }, { "left": "issue:43278", "right": "issue:43844", "accept": false, - "reason": "BF16\u2192FP32 dtype drift during evaluate is unrelated to the gradient increase report." + "reason": "Embedding dtype change during evaluate is unrelated to the ZeRO-3 gradient regression." }, { - "left": "issue:43646", - "right": "issue:44466", + "left": "issue:43278", + "right": "issue:43701", "accept": false, - "reason": "Custom model init breakage and tied-weight serialization are different regression classes." + "reason": "Different problems: dtype mismatch versus checkpoint resume key mismatch." }, { - "left": "issue:43278", - "right": "issue:43701", + "left": "issue:43646", + "right": "issue:44466", "accept": false, - "reason": "Embedding dtype change and checkpoint key mismatch are unrelated." + "reason": "Custom model initialization breakage versus inconsistent serialization of tied weights; not the same bug." }, { "left": "issue:46209", "right": "issue:46710", "accept": false, - "reason": "Both are performance/quality regressions, but they describe different failure modes and code paths." + "reason": "FP8 inference failure on B200 and an accuracy regression after upgrading Transformers are different issues." }, { "left": "issue:44568", "right": "issue:46489", "accept": false, - "reason": "Different tokenizer bugs on different models; no evidence of the same underlying defect." + "reason": "Both concern tokenizer behavior in v5+, but one is special-token insertion for mDeBERTa and the other is incorrect output for DeepSeek-Coder; too different to treat as one bug." }, { "left": "issue:33357", "right": "issue:33666", "accept": false, - "reason": "MacOS bus error on CLIP and Qwen2-VL multi-GPU training are unrelated." + "reason": "A runtime bus error on MacOS versus a multi-GPU training question; unrelated." }, { "left": "issue:43927", "right": "issue:44361", "accept": false, - "reason": "DiaConfig save/load token-id loss and MLukeTokenizer AttributeError are different bugs." + "reason": "DiaConfig save/load token ID loss is unrelated to MLukeTokenizer task AttributeError." }, { - "left": "issue:43575", - "right": "issue:45405", + "left": "issue:44849", + "right": "issue:46209", "accept": false, - "reason": "TP OOM on a large Qwen model is unrelated to a PEFT version bump." + "reason": "Qwen3.5 hidden-state behavior versus FP8 inference failure; different code paths and symptoms." }, { - "left": "issue:44792", - "right": "issue:45405", + "left": "issue:44568", + "right": "issue:45812", "accept": false, - "reason": "Janus image-generation test failure is unrelated to dependency pinning." + "reason": "Both are tokenizer regressions, but one is about missing BOS/EOS insertion and the other is wrong token IDs for Granite models; they are not the same underlying bug." }, { - "left": "issue:44849", - "right": "issue:46209", + "left": "issue:44792", + "right": "issue:45405", "accept": false, - "reason": "Qwen3.5 hidden-states bug and FP8/B200 inference issue are unrelated." + "reason": "Janus image-generation test failure is unrelated to the PEFT version bump." }, { - "left": "issue:44488", - "right": "issue:44843", + "left": "issue:43575", + "right": "issue:45405", "accept": false, - "reason": "Repo-specific load failure and offline-mode model_info call are different tokenizer-loading issues." + "reason": "OOM when loading a large model with tensor parallelism is unrelated to a dependency version bump." }, { "left": "issue:43572", "right": "issue:44291", "accept": false, - "reason": "Missing pad_token_idx in StableLmConfig and unexpected _is_hf_initialized arg are unrelated compatibility regressions." + "reason": "Missing config field after update versus `init_empty_weights` rejecting an unexpected argument; unrelated regressions." + }, + { + "left": "issue:44488", + "right": "issue:44843", + "accept": false, + "reason": "A specific model load failure is not the same as an offline-mode failure caused by an unconditional `model_info()` call." }, { "left": "issue:44589", "right": "issue:45520", "accept": false, - "reason": "Float8 storage lookup failure and flash_attn import key error are unrelated backend issues." + "reason": "Different import/storage problems: missing Float8 storage type versus a Python 3.13 `flash_attn` import-utils KeyError." + }, + { + "left": "issue:46821", + "right": "issue:46832", + "accept": false, + "reason": "TimesFM `window_size` AttributeError and RT-DETR feature-level mismatch are unrelated model bugs." } ] }, @@ -8594,237 +8634,229 @@ "issue:46701", "issue:46710", "issue:46752", - "issue:46821", "issue:46832", "issue:46846", "issue:47145" ], "soft_pairs": [ - "issue:46821|issue:46832", "issue:43334|issue:45759", "issue:43824|issue:44589", "issue:44849|issue:45357", "issue:43504|issue:43606", - "issue:45245|issue:46710", "issue:43334|issue:43572", + "issue:45245|issue:46710", "issue:43257|issue:43299", "issue:44493|issue:46285", "issue:43504|issue:43653", "issue:43606|issue:44112", "issue:45072|issue:46832", - "issue:46399|issue:46846", "issue:44849|issue:47145", + "issue:46399|issue:46846", "issue:44464|issue:45584", + "issue:43844|issue:44964", "issue:43873|issue:46050", "issue:43994|issue:44493", - "issue:43844|issue:44964", "issue:29159|issue:32090", "issue:44869|issue:46752", - "issue:45200|issue:45229", - "issue:45588|issue:46050", "issue:43937|issue:46701", - "issue:45229|issue:45823", + "issue:45200|issue:45229", "issue:42490|issue:43696", + "issue:45229|issue:45823", "issue:45084|issue:45520", + "issue:45588|issue:46050", "issue:29942|issue:30333", "issue:28218|issue:34567" ], - "prepared_review_unit_hash": "76a4878114b240cdad5c72b5bedcd6decd22031a38c8f123e4a598dcf7620609", + "prepared_review_unit_hash": "ac34198e6274b1e865f2ec5fd1c15ceccce3043f58e80b3ea5c4993d8bd66e50", "component_budget": { "node_count": 514, "item_count": 514, "soft_pair_count": 1891, - "serialized_chars": 980926, - "estimated_input_tokens": 245232, - "estimated_eval_tokens": 490720 + "serialized_chars": 980895, + "estimated_input_tokens": 245224, + "estimated_eval_tokens": 490704 }, "budget": { - "node_count": 48, - "item_count": 48, - "soft_pair_count": 28, - "serialized_chars": 36148, - "estimated_input_tokens": 9037, - "estimated_eval_tokens": 18330 + "node_count": 47, + "item_count": 47, + "soft_pair_count": 27, + "serialized_chars": 35288, + "estimated_input_tokens": 8822, + "estimated_eval_tokens": 17900 }, "overflow_policy": "truncate_then_skip", "trimmed": false, "aggressively_trimmed": false, "split": true, "analyst_result": { - "summary": "These items are mostly unrelated model/tokenizer/runtime bugs across different subsystems. None of the soft-similarity pairs look like the same underlying defect, so I would not merge any of them.", - "confidence": 0.78, - "canonical_issue_reason": "issue:43873 is the broadest, highest-signal open issue here, with strong discussion and inbound references around a user-facing runtime problem.", + "summary": "This cluster is mostly a loose semantic collection of unrelated bug reports. The soft pairs share broad themes like Qwen/Gemma/model-loading or tokenizer issues, but not the same concrete failure mode, so none should be merged as duplicates.", + "confidence": 0.62, + "canonical_issue_reason": "issue:43334 is a concrete, reproducible loading regression with a specific stack trace and model family; among the set it is the clearest actionable bug report.", "canonical_pr_reason": null, - "best_issue_reason": "issue:43873 is the most representative issue in the set: active, broad impact, and clearly about a concrete runtime failure mode.", + "best_issue_reason": "issue:43334 has the cleanest failure signal and most specific diagnosis, making it the strongest representative issue in this set.", "best_pr_reason": null, "soft_edge_verdicts": [ - { - "left": "issue:46821", - "right": "issue:46832", - "accept": false, - "reason": "Different models and distinct failures: TimesFM window_size attribute error vs RT-DETR feature-level shape/type mismatch." - }, { "left": "issue:43334", "right": "issue:45759", "accept": false, - "reason": "Both involve Qwen multimodal configs, but one is a missing pad_token_id attribute and the other is AutoModelForCausalLM config unwrapping; different code paths." + "reason": "Both involve Qwen multimodal loading, but one is a pad_token_id attribute error and the other is AutoModelForCausalLM config unwrapping; different code paths and fixes." }, { "left": "issue:43824", "right": "issue:44589", "accept": false, - "reason": "Unrelated bugs: a missing Qwen2.5-VL import symbol versus a Float8 storage lookup/type error." + "reason": "Importing a missing class and a Float8 storage lookup failure are unrelated bugs." }, { "left": "issue:44849", "right": "issue:45357", "accept": false, - "reason": "Different Qwen3.5 regressions with different surfaces: output_hidden_states behavior vs save_pretrained visual-encoder keys." + "reason": "Both are Qwen3.5 regressions, but one affects output_hidden_states and the other save_pretrained visual encoder keys; not the same issue." }, { "left": "issue:43504", "right": "issue:43606", "accept": false, - "reason": "Completely different subsystems: BEiT pretrained loading with a legacy field vs CPU offload device mismatch for bark-small." + "reason": "Different models and failure modes: legacy field loading vs CPU offload device mismatch." }, { - "left": "issue:45245", - "right": "issue:46710", + "left": "issue:43334", + "right": "issue:43572", "accept": false, - "reason": "A category-cardinality runtime error is not the same as an accuracy regression after a version upgrade." + "reason": "Both mention missing token id fields after a version update, but they affect different configs and model families." }, { - "left": "issue:43334", - "right": "issue:43572", + "left": "issue:45245", + "right": "issue:46710", "accept": false, - "reason": "Similar missing-config-field theme, but different models and different missing attributes (pad_token_id vs pad_token_idx)." + "reason": "A category-count runtime limit and an accuracy regression in a different model are unrelated." }, { "left": "issue:43257", "right": "issue:43299", "accept": false, - "reason": "Both are Qwen MoE-related, but one is a conversion/loading issue with accelerate+deepspeed and the other is a separate Qwen3VL loading regression." + "reason": "Both concern Qwen MOE models, but one is weight conversion with accelerate/deepspeed and the other is a loading breakage in dev; different concrete problems." }, { "left": "issue:44493", "right": "issue:46285", "accept": false, - "reason": "Different problems: unexpected position-id keys in many models vs CLIP text model structural inconsistency." + "reason": "Unexpected position-id keys and CLIP text-model structure changes are unrelated bugs." }, { "left": "issue:43504", "right": "issue:43653", "accept": false, - "reason": "Legacy-field model loading bug and BigBirdTokenizer special-token registration bug are unrelated." + "reason": "Legacy field loading for BEiT is unrelated to BigBirdTokenizer special-token registration." }, { "left": "issue:43606", "right": "issue:44112", "accept": false, - "reason": "Offload device mismatch for a model is not the same as a stale CI test issue in GraniteSpeech." + "reason": "A real inference device mismatch and a stale CI test failure are not the same issue." }, { "left": "issue:45072", "right": "issue:46832", "accept": false, - "reason": "Both are inference/runtime failures, but they affect different models and different failure modes (dtype mismatch vs feature-level mismatch)." + "reason": "Different model subsystems and failures: dtype mismatch vs RTDetr feature-level indexing." }, { - "left": "issue:46399", - "right": "issue:46846", + "left": "issue:44849", + "right": "issue:47145", "accept": false, - "reason": "Both mention Qwen3.5, but one is a kernel registration crash and the other is a tensor-parallel plan/OOM issue." + "reason": "Qwen3.5 hidden-state handling and FP8 weight_scale application are unrelated." }, { - "left": "issue:44849", - "right": "issue:47145", + "left": "issue:46399", + "right": "issue:46846", "accept": false, - "reason": "Different bugs: Qwen3.5 hidden-state output behavior vs FP8 compressed-tensor weight-scale loading corruption." + "reason": "Both are Qwen3.5-related, but one is kernel registration and the other is tensor-parallel planning; not mergeable as one fix." }, { "left": "issue:44464", "right": "issue:45584", "accept": false, - "reason": "Chunked generation with compiled forward and Whisper empty-transcription failure are unrelated code paths." + "reason": "Compiled-forward chunked generation mismatch and Whisper empty-transcription decoding are different code paths." + }, + { + "left": "issue:43844", + "right": "issue:44964", + "accept": false, + "reason": "ZeRO-3 gradient behavior and Phi-4 multimodal loading are unrelated." }, { "left": "issue:43873", "right": "issue:46050", "accept": false, - "reason": "Both touch quantization/offloading, but one is runtime offloading behavior and the other is save/load corruption for sinq quantization." + "reason": "Quantization/offloading broadly overlap, but one is runtime offload behavior and the other is save/load corruption after sinq quantization." }, { "left": "issue:43994", "right": "issue:44493", "accept": false, - "reason": "Different model families and distinct issues: nonsensical SigLIP pipeline output vs position-id key warnings." - }, - { - "left": "issue:43844", - "right": "issue:44964", - "accept": false, - "reason": "Separate failures: ZeRO-3 gradient growth vs model import/loading for Phi-4 multimodal." + "reason": "SigLIP2 nonsensical outputs and unexpected position-id keys are different bugs." }, { "left": "issue:29159", "right": "issue:32090", "accept": false, - "reason": "Tokenizer consistency bug and Trainer/_gpu_broadcast_one TypeError are unrelated." + "reason": "Tokenizer fast/slow inconsistency and NoneType GPU broadcast are unrelated." }, { "left": "issue:44869", "right": "issue:46752", "accept": false, - "reason": "Whisper word-timestamp decode crash and apply_chat_template empty-list IndexError are different input validation bugs." + "reason": "Both are edge-case crashes, but in different APIs and failure mechanisms." }, { - "left": "issue:45200", - "right": "issue:45229", + "left": "issue:43937", + "right": "issue:46701", "accept": false, - "reason": "One is a missing default token-type tensor for fine-tuning; the other is a multi-GPU OOM on Gemma4 inference." + "reason": "GLM generation config validation and DiffusionGemma pipeline failure are unrelated." }, { - "left": "issue:45588", - "right": "issue:46050", + "left": "issue:45200", + "right": "issue:45229", "accept": false, - "reason": "Flash-attention crash on s_aux=None and sinq save/load breakage are different defects." + "reason": "Same model family, but text-only fine-tuning defaults and multi-GPU OOM are different issues." }, { - "left": "issue:43937", - "right": "issue:46701", + "left": "issue:42490", + "right": "issue:43696", "accept": false, - "reason": "GenerationConfig validation for GLM-5 is not the same as an image-text-to-text pipeline failure for DiffusionGemmaGenerationOutput." + "reason": "A path-format import warning is unrelated to CUDA OOM." }, { "left": "issue:45229", "right": "issue:45823", "accept": false, - "reason": "Both are Gemma4-related, but one is OOM during multi-GPU inference and the other is a device-map forward mismatch." + "reason": "Both are Gemma4 issues, but one is OOM and the other is a device-map forward mismatch; not the same bug." }, { - "left": "issue:42490", - "right": "issue:43696", + "left": "issue:45084", + "right": "issue:45520", "accept": false, - "reason": "Local-path import warning and GPT-oss-20b CUDA OOM are unrelated." + "reason": "Compile-time non-template-node error and a Python 3.13 import-utils KeyError are unrelated." }, { - "left": "issue:45084", - "right": "issue:45520", + "left": "issue:45588", + "right": "issue:46050", "accept": false, - "reason": "Compilation/template-node TypeError and Python 3.13 flash_attn import KeyError are separate issues." + "reason": "Flash-attention crash on sink-less models and sinq checkpoint save/load corruption are different code paths." }, { "left": "issue:29942", "right": "issue:30333", "accept": false, - "reason": "Flash Attention 2 CI failures and MLFlow job-finalization/reporting are unrelated." + "reason": "Flash Attention test failures and MLflow reporting after job failure are unrelated." }, { "left": "issue:28218", "right": "issue:34567", "accept": false, - "reason": "Tokenizer spacing after added token and TrainerState token-count tracking are different problems." + "reason": "Tokenizer spacing behavior and TrainerState token-count tracking are different concerns." } ] }, @@ -8911,38 +8943,38 @@ "issue:45406|issue:46121", "issue:45588|issue:45759", "issue:43540|issue:45200", - "issue:43295|issue:43873", "issue:43572|issue:46821", + "issue:43295|issue:43873", "issue:43295|issue:44610", "issue:44186|issue:46192", "issue:45823|issue:46192", "issue:43550|issue:43844", - "issue:43450|issue:43575", "issue:43572|issue:44987", + "issue:43450|issue:43575", "issue:29870|issue:29942", "issue:45520|issue:46178", "issue:37428|issue:43824", "issue:45200|issue:45823", "issue:42490|issue:43493", "issue:39692|issue:46192", - "issue:43295|issue:43493", "issue:44521|issue:45290", + "issue:43295|issue:43493", "issue:42915|issue:43824" ], - "prepared_review_unit_hash": "04453221808d00cf5f11479a3a95d6e7587620a76b09b05b9aca6b626fc80d3e", + "prepared_review_unit_hash": "478f9f7fc72c594908b058e90e7d3c8c61f02d85e1edb62887cc8a4587f3d1bf", "component_budget": { "node_count": 514, "item_count": 514, "soft_pair_count": 1891, - "serialized_chars": 980926, - "estimated_input_tokens": 245232, - "estimated_eval_tokens": 490720 + "serialized_chars": 980895, + "estimated_input_tokens": 245224, + "estimated_eval_tokens": 490704 }, "budget": { "node_count": 47, "item_count": 47, "soft_pair_count": 31, - "serialized_chars": 36584, + "serialized_chars": 36583, "estimated_input_tokens": 9146, "estimated_eval_tokens": 18548 }, @@ -8951,198 +8983,192 @@ "aggressively_trimmed": false, "split": true, "analyst_result": { - "summary": "This set is not a true duplicate cluster: the items span unrelated bugs across JIT, generation, flash-attention, multimodal processors, quantization, device mapping, and import/load regressions. None of the proposed soft links look like the same underlying issue.", - "confidence": 0.91, - "canonical_issue_reason": null, - "canonical_pr_reason": null, - "best_issue_reason": "issue:43295 is the most broadly framed regression report and has enough generality to serve as a weak representative of the cluster, but it is still not a real canonical for the other items.", - "best_pr_reason": null, + "summary": "This cluster is a loose similarity set across many unrelated Transformers issues (multimodal processors, flash-attention, serving, quantization, configs, and model-specific regressions). None of the soft pairs look like true duplicates of the same underlying bug/change.", + "confidence": 0.93, + "canonical_issue_reason": "Issue 43295 is the broadest representative anchor: a recent regression in processor/tokenizer handling that affects custom multimodal code. The rest of the cluster is too heterogeneous for a single true duplicate canonical.", + "canonical_pr_reason": "No pull requests are present in this cluster.", + "best_issue_reason": "Issue 43295 is the best global issue anchor because it describes a concrete regression with broad multimodal impact and is the most generally representative of the set.", + "best_pr_reason": "No pull requests are present in this cluster.", "soft_edge_verdicts": [ { "left": "issue:43065", "right": "issue:46369", "accept": false, - "reason": "Both mention Sam3, but one is about a dummy Conv2d implementation while the other is a memory overflow when calling sam3-litetext; different failure modes and likely different fixes." + "reason": "Same model family (Sam3), but different bugs: one is a dummy Conv2d implementation detail, the other is a memory overflow in the litetext model." }, { "left": "issue:15354", "right": "issue:38175", "accept": false, - "reason": "GeneratorExp/JIT export failure is unrelated to SigLIP2 zero probabilities; different models, code paths, and symptoms." + "reason": "Both involve vision models, but one is a torch.jit.script export limitation and the other is an inference/probability issue in SigLIP2." }, { "left": "issue:44554", "right": "issue:45446", "accept": false, - "reason": "MPS attention correctness with mismatched head dims is unrelated to an incorrect PyTorch version gate for AuxRequest import in flex_attention." + "reason": "Different code paths: MPS attention correctness vs a PyTorch version-check/import bug in flex_attention." }, { "left": "issue:43493", "right": "issue:43701", "accept": false, - "reason": "SigLIP2 HF-vs-JAX discrepancy is unrelated to resume_from_checkpoint key mismatch; both are broad regressions but not the same bug." + "reason": "Unrelated symptoms: SigLIP2 implementation discrepancy vs resume-from-checkpoint key mismatch." }, { "left": "issue:41720", "right": "issue:46710", "accept": false, - "reason": "Qwen3 auto device map cudaErrorAssert and DeepSeek accuracy regression are different models and different classes of failure." + "reason": "Different models and failures: Qwen3 auto device-mapping CUDA assert vs an accuracy regression in DeepSeek-R1-Distill-Llama-8B." }, { "left": "issue:44977", "right": "issue:46701", "accept": false, - "reason": "Qwen3.5 flash-attention generation issues and DiffusionGemma pipeline failure are separate model-specific problems." + "reason": "Both are generation-related, but one is a flash-attention bug in Qwen3.5 and the other is an image-text-to-text pipeline failure in DiffusionGemma." }, { "left": "issue:44423", "right": "issue:44734", "accept": false, - "reason": "Both affect serve, but one crashes on multimodal continuous batching with a string .to() call while the other is a KV-cache continuation tensor indexing bug; not the same code-path problem." + "reason": "Both hit serve, but the failures are distinct: multimodal input conversion on strings vs KV-cache continuation tensor indexing." }, { "left": "issue:37428", "right": "issue:43957", "accept": false, - "reason": "ImportError for a missing flash-attention helper is unrelated to meta-device loading regressions, despite both involving loading/integration paths." + "reason": "Import-time flash-attention symbol error vs model loading breakage with torch.device(\"meta\"); different root causes." }, { "left": "issue:43295", "right": "issue:43450", "accept": false, - "reason": "Processor.tokenizer regression and incorrect video batch output shapes are different multimodal processing bugs." + "reason": "Both are multimodal processor issues, but one is a processor/tokenizer regression and the other is a batched video-shape bug." }, { "left": "issue:44977", "right": "issue:46710", "accept": false, - "reason": "Flash-attention generation failure in Qwen3.5 and accuracy regression in DeepSeek are not the same underlying issue." + "reason": "Different bugs: flash-attention generation failure in Qwen3.5 vs a model accuracy regression after version upgrade." }, { "left": "issue:45072", "right": "issue:45198", "accept": false, - "reason": "dtype mismatches in CI for SwitchTransformers/TimmWrapperModel are unrelated to Wav2Vec2 save_pretrained/tokenization failures." + "reason": "One is a bfloat16 dtype mismatch in inference, the other is save_pretrained/tokenization failure; not the same bug." }, { "left": "issue:45406", "right": "issue:46121", "accept": false, - "reason": "Gemma4Processor missing _tokenizer in serve is unrelated to rope-parameter dict conversion with JSON-loaded lists." + "reason": "Gemma4Processor _tokenizer attribute crash and rope-parameter JSON list typing are unrelated." }, { "left": "issue:45588", "right": "issue:45759", "accept": false, - "reason": "flash_attention.py crashing on s_aux=None is a sink-less flash-attention integration bug, not an AutoModelForCausalLM.from_config text_config unwrapping issue." + "reason": "Different paths: flash-attention sink-less model crash vs AutoModelForCausalLM.from_config not unwrapping text_config." }, { "left": "issue:43540", "right": "issue:45200", "accept": false, - "reason": "Qwen3OmniMoe video processing ValueError and Gemma 4 mm_token_type_ids defaulting are different multimodal-input handling issues." + "reason": "Both are multimodal-model issues, but one is a video input processing ValueError and the other is a Gemma4 fine-tuning token-ID defaulting problem." }, { - "left": "issue:43295", - "right": "issue:43873", + "left": "issue:43572", + "right": "issue:46821", "accept": false, - "reason": "Processor/tokenizer regression and quantization offloading behavior are distinct problems even though both involve multimodal/model-loading infrastructure." + "reason": "Different config/attribute failures in unrelated model families (StableLm vs TimesFM)." }, { - "left": "issue:43572", - "right": "issue:46821", + "left": "issue:43295", + "right": "issue:43873", "accept": false, - "reason": "Missing pad_token_idx in StableLmConfig is unrelated to TimesFM window_size AttributeError; different models and APIs." + "reason": "Processor/tokenizer regression vs offloading/quantization behavior; they do not share the same code path." }, { "left": "issue:43295", "right": "issue:44610", "accept": false, - "reason": "Custom model code regression around processor.tokenizer is unrelated to OmDet-Turbo processor/model image-size mismatch." + "reason": "Both involve processors, but one is a tokenizer access regression and the other is a wrong image size output in OmDet-Turbo." }, { "left": "issue:44186", "right": "issue:46192", "accept": false, - "reason": "LayoutLMv2Tokenizer crashes on NER/batching and return_tensors being ignored with text_kwargs are different tokenizer behaviors." + "reason": "LayoutLMv2 tokenizer crashes on NER/padding are not the same as return_tensors being ignored when text_kwargs is passed." }, { "left": "issue:45823", "right": "issue:46192", "accept": false, - "reason": "Gemma4 device_map='auto' forward mismatch is unrelated to return_tensors being ignored when text_kwargs is passed." + "reason": "Gemma4 device mismatch during forward is unrelated to return_tensors being ignored in multimodal preprocessing." }, { "left": "issue:43550", "right": "issue:43844", "accept": false, - "reason": "Bamba torch.compile SDPA failure and gradient growth under HfDeepSpeedConfig + ZeRO-3 are separate training/runtime issues." + "reason": "torch.compile+SDPA failure and ZeRO-3 gradient growth are distinct training/runtime issues." }, { - "left": "issue:43450", - "right": "issue:43575", + "left": "issue:43572", + "right": "issue:44987", "accept": false, - "reason": "Video processor batched-shape bug and Qwen2-57B-A14B-Instruct tp OOM are unrelated." + "reason": "StableLm missing config field and loading physical-intelligence/fast on transformers>=5.1.0 are unrelated regression types." }, { - "left": "issue:43572", - "right": "issue:44987", + "left": "issue:43450", + "right": "issue:43575", "accept": false, - "reason": "StableLmConfig pad_token_idx regression and loading physical-intelligence/fast on transformers>=5.1 are different compatibility issues." + "reason": "Video processor shape bug vs tensor-parallel loading OOM; different subsystems and symptoms." }, { "left": "issue:29870", "right": "issue:29942", "accept": false, - "reason": "BART decoder_attention_mask UnboundLocalError and failing Flash Attention 2 tests are different bugs in different code paths." + "reason": "Bart generate unbound-local error and Flash Attention 2 test failures are unrelated." }, { "left": "issue:45520", "right": "issue:46178", "accept": false, - "reason": "flash_attn import KeyError on Python 3.13 is unrelated to an unsupported tensorflow_text backend mapping error." + "reason": "Both are import/backend related, but one is a flash_attn KeyError in import_utils and the other is a missing BACKENDS_MAPPING entry for tensorflow_text." }, { "left": "issue:37428", "right": "issue:43824", "accept": false, - "reason": "Missing _flash_supports_window_size import and Qwen2.5-VL import failure are both import errors but for different symbols and modules." + "reason": "Flash-attention utility import failure and missing Qwen2_5_VL export are unrelated import errors." }, { "left": "issue:45200", "right": "issue:45823", "accept": false, - "reason": "Gemma 4 mm_token_type_ids defaulting and PLE device mismatch under auto device mapping are separate Gemma4 issues." - }, - { - "left": "issue:42490", - "right": "issue:43493", - "accept": false, - "reason": "Trailing-slash import warning for local models is unrelated to SigLIP2 implementation discrepancy." + "reason": "Gemma4 mm_token_type_ids defaulting for text-only fine-tuning is a different bug from Gemma4 PLE device mismatch under device_map=auto." }, { "left": "issue:39692", "right": "issue:46192", "accept": false, - "reason": "SigLIP2 documentation example errors are unrelated to return_tensors being ignored with text_kwargs." + "reason": "SigLIP2 documentation example errors are unrelated to return_tensors being silently ignored." }, { - "left": "issue:43295", - "right": "issue:43493", + "left": "issue:44521", + "right": "issue:45290", "accept": false, - "reason": "processor.tokenizer regression and SigLIP2 HF-vs-JAX discrepancy are not the same bug." + "reason": "Both touch apply_chat_template, but one is assistant-mask generation for multimodal inputs while the other is a crash on tool-call assistant messages with no content." }, { - "left": "issue:44521", - "right": "issue:45290", + "left": "issue:43295", + "right": "issue:43493", "accept": false, - "reason": "All-zero assistant masks for multimodal inputs and crashes on tool-call assistant messages are different apply_chat_template failures." + "reason": "Processor/tokenizer regression and SigLIP2 HF-vs-JAX discrepancy are not the same underlying bug." }, { "left": "issue:42915", "right": "issue:43824", "accept": false, - "reason": "Qwen3Moe FineGrainedFP8Config failure and Qwen2.5-VL import failure are unrelated model/config problems." + "reason": "Qwen3Moe FineGrainedFP8Config failure and Qwen2_5_VL import error are unrelated." } ] }, @@ -9217,48 +9243,48 @@ ], "soft_pairs": [ "issue:45381|issue:46209", - "issue:44466|issue:45127", "issue:28218|issue:35141", + "issue:44466|issue:45127", "issue:44737|issue:46894", "issue:43278|issue:43856", - "issue:43452|issue:46512", "issue:43653|issue:44112", - "issue:43262|issue:43450", + "issue:43452|issue:46512", "issue:43824|issue:43975", + "issue:43262|issue:43450", "issue:43950|issue:46285", "issue:43334|issue:43525", "issue:44442|issue:45198", "issue:43526|issue:46192", "issue:45200|issue:46899", - "issue:45200|issue:46932", "issue:44898|issue:45412", + "issue:45200|issue:46932", "issue:44792|issue:45563", "issue:43819|issue:44112", "issue:45200|issue:45325", "issue:43335|issue:43550", "issue:41720|issue:46209", "issue:44561|issue:46234", - "issue:44610|issue:44857", "issue:43827|issue:43901", + "issue:44610|issue:44857", "issue:43819|issue:45083", - "issue:44857|issue:45593", "issue:43388|issue:43526", + "issue:44857|issue:45593", "issue:43606|issue:45083" ], - "prepared_review_unit_hash": "00128302475710cc9603784edc577b7db198dbcb83034ed1d20a87240f4028e3", + "prepared_review_unit_hash": "f14e2af978cb6715a846377477108e0cd41050b7e2135a3f08792b47f1944a27", "component_budget": { "node_count": 514, "item_count": 514, "soft_pair_count": 1891, - "serialized_chars": 980926, - "estimated_input_tokens": 245232, - "estimated_eval_tokens": 490720 + "serialized_chars": 980895, + "estimated_input_tokens": 245224, + "estimated_eval_tokens": 490704 }, "budget": { "node_count": 48, "item_count": 48, "soft_pair_count": 28, - "serialized_chars": 36571, + "serialized_chars": 36570, "estimated_input_tokens": 9143, "estimated_eval_tokens": 18542 }, @@ -9267,186 +9293,180 @@ "aggressively_trimmed": false, "split": true, "analyst_result": { - "summary": "The cluster is mostly a loose mix of unrelated issues. I found one likely duplicate pair about Gemma4 quantized multimodal embedders casting to the wrong dtype; the rest are distinct bugs or docs issues and should stay separate.", - "confidence": 0.95, - "canonical_issue_reason": "issue:46899 is the clearest and most complete report of the shared quantized vision/audio embedder dtype-cast bug; issue:46932 reads like a narrower follow-up/fix note for the same problem.", + "summary": "This cluster is mostly heterogeneous and does not appear to contain broader duplicates. I found one strong duplicate-like pair around Gemma4 quantized multimodal embedders; the rest are separate bugs/issues. No PRs are present.", + "confidence": 0.89, + "canonical_issue_reason": "issue:46899 is the clearest representative of the only true duplicate-like problem in the set: Gemma4 quantized audio/vision embedders failing due to unsafe dtype casting. issue:46932 reads like a closely related follow-up/fix note rather than a separate underlying bug.", "canonical_pr_reason": null, - "best_issue_reason": "issue:46899 best represents the duplicated bug because it describes the end-user failure, covers both audio and vision paths, and has the fuller repro context.", + "best_issue_reason": "issue:46899 is the best global issue anchor because it cleanly captures the shared Gemma4 quantized-embedder dtype-cast defect; issue:46932 is effectively the nearest duplicate candidate.", "best_pr_reason": null, "soft_edge_verdicts": [ { "left": "issue:45381", "right": "issue:46209", "accept": false, - "reason": "Different Qwen2.5-VL bugs: bad vision_position_ids vs FP8 inference failure on B200." + "reason": "Different failures: Qwen2.5-VL vision position IDs vs an FP8 inference failure on B200." }, { - "left": "issue:44466", - "right": "issue:45127", + "left": "issue:28218", + "right": "issue:35141", "accept": false, - "reason": "Both involve tied-weights/LM head behavior, but the concrete failures differ (serialization inconsistency vs LoRA merge collapse)." + "reason": "Tokenizer spacing after added tokens is unrelated to embedding reinitialization after resize." }, { - "left": "issue:28218", - "right": "issue:35141", + "left": "issue:44466", + "right": "issue:45127", "accept": false, - "reason": "Tokenizer spacing bug and embedding reinit bug are unrelated code paths." + "reason": "Both involve tied-weight behavior, but one is serialization/device-dependent while the other is LoRA merge collapse with extended vocab; not the same bug." }, { "left": "issue:44737", "right": "issue:46894", "accept": false, - "reason": "XLNet CPU arange issue is unrelated to CohereASR double-shift training loss." + "reason": "XLNet CPU tensor creation and CohereASR double-shift loss are unrelated code paths." }, { "left": "issue:43278", "right": "issue:43856", "accept": false, - "reason": "Different problems: dtype drift in evaluate vs Qwen3 MoE memory usage." - }, - { - "left": "issue:43452", - "right": "issue:46512", - "accept": false, - "reason": "from_pretrained gguf loading bug is unrelated to from_config(dtype=...) stale config dtype." + "reason": "Different problems: eval-time dtype drift vs Qwen3 MoE memory usage." }, { "left": "issue:43653", "right": "issue:44112", "accept": false, - "reason": "Tokenizer special-token registration bug and GraniteSpeech CI stale test are different." + "reason": "Tokenizer special-token registration bug is unrelated to a GraniteSpeech CI test failure." }, { - "left": "issue:43262", - "right": "issue:43450", + "left": "issue:43452", + "right": "issue:46512", "accept": false, - "reason": "Audio chat-template sampling-rate bug is unrelated to batched video processor shape handling." + "reason": "GGUF loading failures and stale config dtype after from_config are different issues." }, { "left": "issue:43824", "right": "issue:43975", "accept": false, - "reason": "Import error for Qwen2.5-VL is unrelated to DeepSeek-Coder detokenization." + "reason": "Import error for a Qwen2.5-VL class is unrelated to DeepSeek detokenization in v5." + }, + { + "left": "issue:43262", + "right": "issue:43450", + "accept": false, + "reason": "Audio chat-template sampling-rate default and batched video shape handling are different processor bugs." }, { "left": "issue:43950", "right": "issue:46285", "accept": false, - "reason": "Non-persistent buffer corruption and CLIPTextModel structure inconsistency are different regressions." + "reason": "Non-persistent buffer corruption is unrelated to CLIP class structure inconsistencies." }, { "left": "issue:43334", "right": "issue:43525", "accept": false, - "reason": "Same symptom class (missing pad_token_id), but different model configs and not clearly the same underlying bug." + "reason": "Same symptom class (missing pad_token_id), but different model configs and likely different fixes." }, { "left": "issue:44442", "right": "issue:45198", "accept": false, - "reason": "AutoTokenizer load failure and Wav2Vec2 save/tokenization failure are distinct issues." + "reason": "AutoTokenizer loading failure and Wav2Vec2 save/tokenization failure are separate tokenizer issues." }, { "left": "issue:43526", "right": "issue:46192", "accept": false, - "reason": "Label reduction bug and return_tensors being ignored are unrelated processor behaviors." + "reason": "Label reduction in BeitImageProcessorFast is unrelated to return_tensors being ignored with text_kwargs." }, { "left": "issue:45200", "right": "issue:46899", "accept": false, - "reason": "Gemma4 mm_token_type_ids defaulting issue is separate from quantized embedder dtype casting." + "reason": "Gemma4 token-type-id defaults and Gemma4 quantized embedder dtype casting are different bugs." }, { - "left": "issue:45200", - "right": "issue:46932", + "left": "issue:44898", + "right": "issue:45412", "accept": false, - "reason": "Gemma4 mm_token_type_ids defaulting issue is separate from quantized embedder dtype casting." + "reason": "Perceiver interpolation failure and RT-DETR memory leak are unrelated." }, { - "left": "issue:44898", - "right": "issue:45412", + "left": "issue:45200", + "right": "issue:46932", "accept": false, - "reason": "Perceiver interpolation failure and RT-DETR memory release bug are unrelated." + "reason": "Different Gemma4 problems: default mm_token_type_ids vs quantized vision/audio dtype casting." }, { "left": "issue:44792", "right": "issue:45563", "accept": false, - "reason": "Janus image generation test failure and paged-generate warning issue are different." + "reason": "Janus image-generation test failure is unrelated to a stale generate() warning." }, { "left": "issue:43819", "right": "issue:44112", "accept": false, - "reason": "DAC forward/from_latents mismatch is unrelated to GraniteSpeech CI test flakiness." + "reason": "DAC latent mismatch and GraniteSpeech CI device override are unrelated." }, { "left": "issue:45200", "right": "issue:45325", "accept": false, - "reason": "Different multimodal bugs: Gemma4 token type ids vs Qwen2.5-VL rope index scaling." + "reason": "Token-type-id defaults and Qwen2.5-VL rope-index scaling are different model bugs." }, { "left": "issue:43335", "right": "issue:43550", "accept": false, - "reason": "SwitchTransformers sparse-layer config bug and Bamba SDPA compile bug are unrelated." + "reason": "Sparse-layer construction in SwitchTransformers is unrelated to Bamba torch.compile SDPA failure." }, { "left": "issue:41720", "right": "issue:46209", "accept": false, - "reason": "Different hardware/model failures: A800 cuda assert vs B200 FP8 inference." + "reason": "Different hardware/model-specific inference failures; not the same underlying bug." }, { "left": "issue:44561", "right": "issue:46234", "accept": false, - "reason": "Removed utility API and missing Falcon-Mamba utility flag are unrelated." - }, - { - "left": "issue:44610", - "right": "issue:44857", - "accept": false, - "reason": "Processor output size mismatch and AMP float16 crash are different bugs." + "reason": "Removal of a utility function breaking trust_remote_code is unrelated to a missing import_utils symbol." }, { "left": "issue:43827", "right": "issue:43901", "accept": false, - "reason": "Both are docs regressions, but they concern different pipeline APIs and are not the same change." + "reason": "Both are docs regressions, but they reference different API changes and different pages." }, { - "left": "issue:43819", - "right": "issue:45083", + "left": "issue:44610", + "right": "issue:44857", "accept": false, - "reason": "DAC latent mismatch and qwen3_omni_moe length helper behavior are unrelated." + "reason": "Processor input-size mismatch and AMP CUDA crash are separate issues." }, { - "left": "issue:44857", - "right": "issue:45593", + "left": "issue:43819", + "right": "issue:45083", "accept": false, - "reason": "AMP/CUDA crash and D-FINE auxiliary-loss behavior are unrelated." + "reason": "DAC forward/from_latents mismatch and qwen3_omni_moe helper behavior are unrelated." }, { "left": "issue:43388", "right": "issue:43526", "accept": false, - "reason": "Mask2Former gather_for_metrics label dropping and BeitImageProcessorFast reduce_labels bug differ." + "reason": "Both concern labels, but one is gather_for_metrics truncation and the other is BeitImageProcessorFast label reduction." }, { - "left": "issue:43606", - "right": "issue:45083", + "left": "issue:44857", + "right": "issue:45593", "accept": false, - "reason": "CPU offload device mismatch and qwen3_omni_moe feature-length helper behavior are unrelated." + "reason": "Float16 AMP crash and missing auxiliary losses are different D-FINE issues." }, { - "left": "issue:46899", - "right": "issue:46932", - "accept": true, - "reason": "Both describe the same Gemma4 quantized multimodal embedder dtype-casting bug; 46932 is a narrower fix-oriented duplicate of 46899." + "left": "issue:43606", + "right": "issue:45083", + "accept": false, + "reason": "CPU-offload device mismatch and qwen3_omni_moe output-length helper behavior are unrelated." } ] }, @@ -9520,9 +9540,9 @@ "issue:46858" ], "soft_pairs": [ + "issue:37428|issue:43425", "issue:46512|issue:46821", "issue:44206|issue:45584", - "issue:37428|issue:43425", "issue:43335|issue:43653", "issue:45759|issue:46285", "issue:38175|issue:45412", @@ -9537,26 +9557,26 @@ "issue:37428|issue:44351", "issue:44448|issue:44743", "issue:43335|issue:45083", - "issue:43335|issue:44112", "issue:45290|issue:46326", + "issue:43335|issue:44112", "issue:43335|issue:45593", - "issue:44514|issue:44521", "issue:43278|issue:43825", + "issue:44514|issue:44521", "issue:34689|issue:35141", - "issue:45588|issue:45859", "issue:42490|issue:43854", + "issue:45588|issue:45859", "issue:29127|issue:32090", - "issue:43819|issue:44493", - "issue:39692|issue:43526" + "issue:39692|issue:43526", + "issue:43819|issue:44493" ], - "prepared_review_unit_hash": "8cf2376b93879e21b8c8e9521f69366ec8f3e6ef2889aa32a801164b3567e796", + "prepared_review_unit_hash": "65794a7ec90cb34f2f42a15796433cd7bc6aec92415937a39f6904842d34dea0", "component_budget": { "node_count": 514, "item_count": 514, "soft_pair_count": 1891, - "serialized_chars": 980926, - "estimated_input_tokens": 245232, - "estimated_eval_tokens": 490720 + "serialized_chars": 980895, + "estimated_input_tokens": 245224, + "estimated_eval_tokens": 490704 }, "budget": { "node_count": 48, @@ -9571,174 +9591,180 @@ "aggressively_trimmed": false, "split": true, "analyst_result": { - "summary": "No soft pair looks like a true duplicate: the cluster is broad and the proposed matches usually share only a subsystem name or a vague regression theme, not the same concrete bug.", - "confidence": 0.98, - "canonical_issue_reason": "No single issue is a good canonical representative here; the items span unrelated bugs across loading, caching, chat templating, CI, and model-specific edge cases.", + "summary": "This cluster is heterogeneous: the soft pairs mostly share broad Transformers themes but not the same underlying bug. I would not merge any of the suggested pairs as duplicates.", + "confidence": 0.93, + "canonical_issue_reason": null, "canonical_pr_reason": null, - "best_issue_reason": "No single issue is especially suitable as the global best issue because there is no coherent underlying duplicate cluster to represent.", + "best_issue_reason": "issue:44521 is the strongest standalone representative: it has active discussion, inbound references, and a clear reproducible failure in a common API (`apply_chat_template`) affecting multimodal inputs.", "best_pr_reason": null, "soft_edge_verdicts": [ + { + "left": "issue:37428", + "right": "issue:43425", + "accept": false, + "reason": "Both are import/version problems, but one is a missing internal flash-attention symbol and the other is Torch 2.10 incompatibility; different root causes and fixes." + }, { "left": "issue:46512", "right": "issue:46821", "accept": false, - "reason": "Different bugs: one is dtype/config state propagation in from_config, the other is a TimesFM window_size AttributeError." + "reason": "`from_config(dtype=...)` leaving `model.config.dtype` stale is unrelated to TimesFM `window_size` raising `AttributeError`; different APIs and failure modes." }, { "left": "issue:44206", "right": "issue:45584", "accept": false, - "reason": "Different models and failure modes: LasrFeatureExtractor center-arg crash vs Whisper empty-transcription behavior after align_special_tokens." - }, - { - "left": "issue:37428", - "right": "issue:43425", - "accept": false, - "reason": "One is a missing internal import symbol; the other is a broad Torch 2.10 compatibility problem." + "reason": "`LasrFeatureExtractor` rejecting an unsupported `center` arg is unrelated to Whisper empty-transcription behavior after `align_special_tokens`." }, { "left": "issue:43335", "right": "issue:43653", "accept": false, - "reason": "Different code paths: SwitchTransformers sparse-layer construction vs BigBirdTokenizer special-token registration/decoding." + "reason": "SwitchTransformers sparse-layer creation and BigBirdTokenizer special-token registration are separate bugs in different components." }, { "left": "issue:45759", "right": "issue:46285", "accept": false, - "reason": "Both involve model config structure, but they are different architectures and different bugs (Qwen multimodal text_config unwrapping vs CLIP text model structure)." + "reason": "Composite Qwen multimodal config unwrapping is not the same bug as CLIP text-model structure divergence." }, { "left": "issue:38175", "right": "issue:45412", "accept": false, - "reason": "Unrelated issues: SigLIP2 zero probabilities vs RT-DETR memory not being released." + "reason": "Unexpected zero probabilities in SigLIP2 inference is unrelated to RT-DETR memory not being released on deletion." }, { "left": "issue:46424", "right": "issue:46858", "accept": false, - "reason": "Different cache bugs: max_cache_length handling vs device-tensor seq_length breaking slice arithmetic under compile." + "reason": "Ignoring `max_cache_length` for quantized cache and returning a device tensor from `get_seq_length()` are different cache bugs." }, { "left": "issue:29942", "right": "issue:32090", "accept": false, - "reason": "Different areas and symptoms: Flash Attention 2 test failures vs Trainer/_gpu_broadcast_one NoneType handling." + "reason": "Failing Flash Attention 2 tests and a Trainer `_gpu_broadcast_one` `NoneType` error are unrelated code paths." }, { "left": "issue:28282", "right": "issue:33357", "accept": false, - "reason": "Different environment/runtime problems: missing PyTorch import vs MacOS bus error loading a community CLIP model." + "reason": "Missing PyTorch import for `AutoModel` is unrelated to the MacOS bus error loading a community CLIP model." }, { "left": "issue:43611", "right": "issue:44466", "accept": false, - "reason": "One concerns model loading with base_model_prefix, the other inconsistent lm_head.weight serialization; not the same bug." + "reason": "Both touch v5 model loading/serialization, but `base_model_prefix` loading failure and `lm_head.weight` serialization inconsistency are distinct bugs." }, { "left": "issue:43335", "right": "issue:43504", "accept": false, - "reason": "Different failures in different model code paths: sparse layer creation vs legacy-field preset loading for BEiT segmentation." + "reason": "SwitchTransformers sparse-layer creation with `num_sparse_encoder_layers=0` is unrelated to BEiT legacy-field load failure." }, { "left": "issue:44855", "right": "issue:45083", "accept": false, - "reason": "Python 3.13 import/parsing error vs qwen3_omni_moe helper behavior; unrelated defects." + "reason": "Python 3.13 `torch.jit` parsing/indentation failure is unrelated to `_get_feat_extract_output_lengths` behavior in qwen3_omni_moe." }, { "left": "issue:43550", "right": "issue:44112", "accept": false, - "reason": "A torch.compile/SDPA model bug and a stale CI device-override test are not the same underlying issue." + "reason": "`torch.compile`/SDPA failure in Bamba-9B-v2 is unrelated to the stale device-override CI test in GraniteSpeech." }, { "left": "issue:45520", "right": "issue:46170", "accept": false, - "reason": "Different platform/dependency regressions: flash_attn import lookup on Python 3.13 vs Windows ROCm torch.distributed import behavior." + "reason": "Python 3.13 `flash_attn` `KeyError` is a different issue from Windows ROCm import/type errors involving `torch.distributed`." }, { "left": "issue:37428", "right": "issue:44351", "accept": false, - "reason": "Both are import errors, but they refer to different missing symbols and different subsystems." + "reason": "Missing `_flash_supports_window_size` and missing `HybridCache` are different import/export regressions." }, { "left": "issue:44448", "right": "issue:44743", "accept": false, - "reason": "Different behavior regressions: v4/v5 output mismatch for Pegasus vs recurrent state reset in qwen3_5 modular cache handling." + "reason": "Pegasus output differences across versions do not match the recurrent-state reset bug in modular_qwen3_5 cache handling." }, { "left": "issue:43335", "right": "issue:45083", "accept": false, - "reason": "The titles point to different model families and different bugs; only a broad Transformer theme is shared." + "reason": "SwitchTransformers sparse-layer creation is unrelated to the qwen3_omni_moe helper-length bug." }, { - "left": "issue:43335", - "right": "issue:44112", + "left": "issue:45290", + "right": "issue:46326", "accept": false, - "reason": "Sparse-layer config bug vs CI test failure in GraniteSpeech; no shared concrete code-path." + "reason": "Both involve chat templating, but one crashes on assistant tool-call messages with no content while the other is a `continue_final_message` bug in a mistral_common tokenizer path." }, { - "left": "issue:45290", - "right": "issue:46326", + "left": "issue:43335", + "right": "issue:44112", "accept": false, - "reason": "Both are chat-template related, but one is a tool-call/no-content crash and the other is continue_final_message on mistral_common tokenizers." + "reason": "SwitchTransformers config bug and GraniteSpeech CI device-override test failure are unrelated." }, { - "left": "issue:44514", - "right": "issue:44521", + "left": "issue:43335", + "right": "issue:45593", "accept": false, - "reason": "Same broad API, but different concrete failures: batched padding=False crash vs all-zero assistant_masks for multimodal inputs." + "reason": "Sparse-layer construction in SwitchTransformers is unrelated to D-FINE losing auxiliary losses when denoising is off." }, { "left": "issue:43278", "right": "issue:43825", "accept": false, - "reason": "Different bug classes: dtype mismatch during train/eval vs incorrect pipeline translation messaging." + "reason": "A dtype mismatch in `from_config` is unrelated to the pipeline translation-task warning regression." }, { - "left": "issue:34689", - "right": "issue:35141", + "left": "issue:44514", + "right": "issue:44521", "accept": false, - "reason": "Model-loading regression for a specific model vs token-embedding reinitialization after resize; not the same defect." + "reason": "Both are `apply_chat_template` issues, but one is a batched `padding=False` crash and the other is all-zero assistant masks for multimodal inputs; different bugs." }, { - "left": "issue:45588", - "right": "issue:45859", + "left": "issue:34689", + "right": "issue:35141", "accept": false, - "reason": "Flash-attention s_aux=None crash and CPU-offload rotary embedding incompatibility are separate issues." + "reason": "Llama 3.2 Vision model-loading regression and output-embedding reinitialization after resizing token embeddings are different problems." }, { "left": "issue:42490", "right": "issue:43854", "accept": false, - "reason": "Trailing-slash local-path import warning vs failing to load a specific GLM-4.7 model in tests." + "reason": "Trailing-slash local-path import warning is unrelated to the GLM-4.7-Flash unit-test loading failure." }, { - "left": "issue:29127", - "right": "issue:32090", + "left": "issue:45588", + "right": "issue:45859", "accept": false, - "reason": "Different failures: LayoutLMv3 error-message clarity vs Trainer GPU broadcast NoneType type error." + "reason": "`s_aux=None` crash in flash_attention integration and CPU-offload incompatibility in Qwen3_5MoeTextRotaryEmbedding are different code paths." }, { - "left": "issue:43819", - "right": "issue:44493", + "left": "issue:29127", + "right": "issue:32090", "accept": false, - "reason": "DAC forward/latent mismatch and unexpected position-id keys are unrelated model bugs." + "reason": "LayoutLMv3 error-message clarity and Trainer GPU broadcast `NoneType` failure are unrelated." }, { "left": "issue:39692", "right": "issue:43526", "accept": false, - "reason": "Documentation example issues for SigLIP2 vs a BEiT fast image-processor label-reduction bug." + "reason": "SigLIP2 documentation-example problems are not the same as the BeitImageProcessorFast `reduce_labels` bug." + }, + { + "left": "issue:43819", + "right": "issue:44493", + "accept": false, + "reason": "DAC forward/from_latents mismatch is unrelated to the unexpected position-id key issue in models loading." } ] }, @@ -9811,47 +9837,47 @@ "issue:47000" ], "soft_pairs": [ - "issue:46431|issue:46566", "issue:43606|issue:44857", + "issue:46431|issue:46566", "issue:43720|issue:44112", - "issue:39692|issue:44493", "issue:28218|issue:32090", "issue:45310|issue:45478", + "issue:39692|issue:44493", "issue:44448|issue:44779", - "issue:44206|issue:45823", "issue:43701|issue:45083", - "issue:44617|issue:46697", + "issue:44206|issue:45823", "issue:26413|issue:31501", + "issue:44617|issue:46697", "issue:45127|issue:45245", "issue:45459|issue:46234", "issue:32090|issue:33357", - "issue:44857|issue:45072", "issue:42490|issue:43575", + "issue:44857|issue:45072", "issue:45593|issue:47000", "issue:43550|issue:45083", + "issue:44464|issue:46424", "issue:27957|issue:34689", "issue:38175|issue:44493", - "issue:44464|issue:46424", "issue:29942|issue:35141", "issue:29870|issue:32090", "issue:45290|issue:45381", "issue:43278|issue:43329", "issue:10000|issue:15354" ], - "prepared_review_unit_hash": "c7f0b2f989d6cdf3beaa5ae2fb8bd6f57a6db2bedc47d59f0d4bab174dd4ba8b", + "prepared_review_unit_hash": "417a639bf6ccc9b47783fa8223db35ffa7a2c2dab5d6561f069e59e4136ec211", "component_budget": { "node_count": 514, "item_count": 514, "soft_pair_count": 1891, - "serialized_chars": 980926, - "estimated_input_tokens": 245232, - "estimated_eval_tokens": 490720 + "serialized_chars": 980895, + "estimated_input_tokens": 245224, + "estimated_eval_tokens": 490704 }, "budget": { "node_count": 47, "item_count": 47, "soft_pair_count": 26, - "serialized_chars": 34959, + "serialized_chars": 34960, "estimated_input_tokens": 8740, "estimated_eval_tokens": 17736 }, @@ -9860,174 +9886,174 @@ "aggressively_trimmed": false, "split": true, "analyst_result": { - "summary": "This cluster is mostly heterogeneous issue reports. The only clear duplicate pair is the repeated Qwen3.5 MoE from_pretrained error (issues 45310 and 45478); the rest are distinct bugs or regressions in different code paths.", - "confidence": 0.88, - "canonical_issue_reason": "Issue 45310 is the best canonical issue because it has an almost exact duplicate in 45478 with the same title and underlying Qwen3.5 MoE from_pretrained failure, making it the clearest representative bug report in this set.", + "summary": "This cluster is mostly a grab bag of unrelated issue reports. The only clear duplicate pair is the two identical Qwen3.5 MoE from_pretrained errors; everything else looks like distinct bugs or regressions.", + "confidence": 0.89, + "canonical_issue_reason": "issue:45310 is the earliest report of the only clear duplicate pair and has the stronger cluster anchor role for the identical Qwen3.5 MoE from_pretrained error.", "canonical_pr_reason": null, - "best_issue_reason": "Issue 45310 is the strongest issue to keep as the representative bug report: it is specific, well-scoped, and has a near-identical duplicate confirming the same underlying regression.", + "best_issue_reason": "issue:45310 is the best issue representative because it is the clearest duplicate candidate and the earliest, most canonical report among the items.", "best_pr_reason": null, "soft_edge_verdicts": [ - { - "left": "issue:46431", - "right": "issue:46566", - "accept": false, - "reason": "Different failures: missing chat template vs. model weights staying on meta with auto device map." - }, { "left": "issue:43606", "right": "issue:44857", "accept": false, - "reason": "Unrelated bugs affecting different models and paths; one is CPU offload device mismatch, the other AMP loss crash." + "reason": "Different failures and code paths: CPU offload device mismatch in bark-small vs float16 AMP crash in LwDetrImageLoss." }, { - "left": "issue:43720", - "right": "issue:44112", + "left": "issue:46431", + "right": "issue:46566", "accept": false, - "reason": "Different code paths: accelerate loading of packed BitNet weights vs. GraniteSpeech CI device override test." + "reason": "One is missing chat template support; the other is a device_map/meta-device loading issue in DiffusionGemma." }, { - "left": "issue:39692", - "right": "issue:44493", + "left": "issue:43720", + "right": "issue:44112", "accept": false, - "reason": "Different symptoms and subsystems: SigLIP2 doc example errors vs. a broad position-id key regression." + "reason": "Both are CI/loading-related, but the root causes differ: unpacking packed weights vs a stale device override in GraniteSpeech." }, { "left": "issue:28218", "right": "issue:32090", "accept": false, - "reason": "Tokenizer spacing bug and Trainer broadcast NoneType error are unrelated." + "reason": "Tokenizer spacing after added tokens and Trainer _gpu_broadcast_one NoneType errors are unrelated bugs." }, { "left": "issue:45310", "right": "issue:45478", "accept": true, - "reason": "Near-identical reports of the same Qwen3.5 MoE from_pretrained regression with matching title and failure mode." + "reason": "Same title and same concrete failure: transformers>=5.4.0 breaking Qwen3.5 MoE from_pretrained." }, { - "left": "issue:44448", - "right": "issue:44779", + "left": "issue:39692", + "right": "issue:44493", "accept": false, - "reason": "Both mention output differences, but one is Pegasus v4/v5 behavior and the other is Deepseek tokenizer correctness; not the same bug." + "reason": "SigLIP2 docs/example errors do not match the broad unexpected position-id key regressions." }, { - "left": "issue:44206", - "right": "issue:45823", + "left": "issue:44448", + "right": "issue:44779", "accept": false, - "reason": "Different models and errors: unsupported center arg in audio feature extraction vs. Gemma4 device mismatch." + "reason": "Pegasus v4/v5 output differences and Deepseek tokenizer regressions are different subsystems and symptoms." }, { "left": "issue:43701", "right": "issue:45083", "accept": false, - "reason": "Resume-from-checkpoint key mismatch is unrelated to qwen3_omni_moe feature-extractor length behavior." + "reason": "resume_from_checkpoint key mismatch is a checkpoint/resume bug; _get_feat_extract_output_lengths in qwen3_omni_moe is unrelated." }, { - "left": "issue:44617", - "right": "issue:46697", + "left": "issue:44206", + "right": "issue:45823", "accept": false, - "reason": "CUDA OOM in Sam3Video is not the same underlying issue as Swin2SR image over-padding." + "reason": "Unsupported center arg in LasrFeatureExtractor and Gemma4 PLE device mismatch are distinct issues." }, { "left": "issue:26413", "right": "issue:31501", "accept": false, - "reason": "Resume-from-checkpoint epoch iterator failure and FineWeb training startup failure are distinct training issues." + "reason": "Both affect training start/resume, but the epoch_iterator empty error and FineWeb training startup failure are different underlying problems." + }, + { + "left": "issue:44617", + "right": "issue:46697", + "accept": false, + "reason": "CUDA OOM in Sam3Video and over-padding in Swin2SR image processing are unrelated." }, { "left": "issue:45127", "right": "issue:45245", "accept": false, - "reason": "LoRA merge / tied embeddings collapse is unrelated to the categorical limit runtime error." + "reason": "LoRA merge/model collapse and category-count RuntimeError are entirely different bugs." }, { "left": "issue:45459", "right": "issue:46234", "accept": false, - "reason": "Tokenizer protobuf error handling and a missing import_utils helper are different bugs." + "reason": "Tokenizer protobuf error masking and a missing import_utils symbol are unrelated maintenance bugs." }, { "left": "issue:32090", "right": "issue:33357", "accept": false, - "reason": "Trainer _gpu_broadcast_one NoneType error is not the same as a MacOS bus error in CLIP loading." + "reason": "Trainer NoneType broadcast failure and MacOS CLIP bus error are different failure modes." }, { - "left": "issue:44857", - "right": "issue:45072", + "left": "issue:42490", + "right": "issue:43575", "accept": false, - "reason": "Both involve dtype issues, but they affect different models and concrete failures; not mergeable as one fix." + "reason": "Local-path import warning and tensor-parallel OOM on Qwen2-57B are unrelated." }, { - "left": "issue:42490", - "right": "issue:43575", + "left": "issue:44857", + "right": "issue:45072", "accept": false, - "reason": "Local-path trailing-slash import warning is unrelated to Qwen2-57B tensor-parallel OOM." + "reason": "Both mention dtype-ish failures, but one is an AMP crash in a loss and the other is a bfloat16 inference mismatch; not the same code-path issue." }, { "left": "issue:45593", "right": "issue:47000", "accept": false, - "reason": "Different detection-loss bugs with different triggers and failure modes." + "reason": "Different detection-loss bugs: denoising-off auxiliary loss omission vs RTDetr cost matrix infeasibility." }, { "left": "issue:43550", "right": "issue:45083", "accept": false, - "reason": "Torch.compile/SDPA failure in Bamba-9B-v2 is unrelated to qwen3_omni_moe output-length behavior." + "reason": "torch.compile/SDPA failure in Bamba is unrelated to qwen3_omni_moe feature-length behavior." + }, + { + "left": "issue:44464", + "right": "issue:46424", + "accept": false, + "reason": "Chunked generation inconsistency under compiled forward is a different bug from cache max length only applying to Quantized Cache." }, { "left": "issue:27957", "right": "issue:34689", "accept": false, - "reason": "Flash Attention 2 support for XLMRoberta and Llama 3.2 model loading regression are different problems." + "reason": "XLMRoberta Flash Attention 2 support and Llama 3.2 Vision model loading regression are distinct issues." }, { "left": "issue:38175", "right": "issue:44493", "accept": false, - "reason": "SigLIP2 zero probabilities and unexpected position-id keys are different model bugs." - }, - { - "left": "issue:44464", - "right": "issue:46424", - "accept": false, - "reason": "Compiled forward chunked generation inconsistency is not the same as max_cache_length only applying to quantized cache." + "reason": "siglip2 zero probabilities and unexpected position-id keys are not the same failure." }, { "left": "issue:29942", "right": "issue:35141", "accept": false, - "reason": "Failing Flash Attention 2 tests and output embedding reinitialization on resize are unrelated." + "reason": "Failing FA2 tests and embedding reinitialization after resize are different bugs." }, { "left": "issue:29870", "right": "issue:32090", "accept": false, - "reason": "Bart generate decoder_attention_mask bug is unrelated to Trainer broadcast NoneType error." + "reason": "Bart generate decoder_attention_mask bug and Trainer broadcast NoneType bug are unrelated." }, { "left": "issue:45290", "right": "issue:45381", "accept": false, - "reason": "Chat template crash with tool calls is a different path from Qwen2.5-VL video position-id handling." + "reason": "Chat template/tool-call crash and qwen2.5-vl video position-id issue are both multimodal-related but clearly different paths." }, { "left": "issue:43278", "right": "issue:43329", "accept": false, - "reason": "Embedding dtype drift in evaluation and an undefined-variable video token-count bug are unrelated." + "reason": "Embedding dtype drift in evaluate and an undefined video-branch multimodal token helper are unrelated." }, { "left": "issue:10000", "right": "issue:15354", "accept": false, - "reason": "German DistilBertModel issue and torch.jit.script GeneratorExp export failure are unrelated legacy bugs." + "reason": "German DistilBertModel issue and GeneratorExp/JIT export failure are unrelated model-specific problems." } ] }, "evaluator_result": { "accept": true, - "feedback": "The summary is grounded in the packet: 45310/45478 is the only clearly duplicated pair by title/body, and the remaining soft-pairs are conservatively rejected as different issues. I don\u2019t see evidence of overstatement, and there are no PR-pair mergeability concerns here." + "feedback": "Grounded and conservative overall. The only accepted soft edge, issue:45310 \u2194 issue:45478, is well supported by the identical title and clear duplicate signal, and choosing issue:45310 as the canonical/earliest representative is reasonable. The remaining pairs are rejected cautiously with distinctions that track the packet titles, and the summary does not overclaim broader duplication beyond the one clear pair." }, "evaluator_used": true, "retried": false, @@ -10085,7 +10111,6 @@ "issue:45290", "issue:45440", "issue:45491", - "issue:45563", "issue:45588", "issue:45698", "issue:45803", @@ -10100,23 +10125,23 @@ "soft_pairs": [ "issue:37428|issue:42175", "issue:45290|issue:46655", - "issue:29870|issue:33290", "issue:44448|issue:44977", + "issue:29870|issue:33290", "issue:44360|issue:44561", "issue:43278|issue:43381", "issue:44855|issue:46399", "issue:40444|issue:43526", "issue:45920|issue:46192", "issue:32090|issue:33290", - "issue:44977|issue:45229", "issue:43927|issue:45698", + "issue:44977|issue:45229", "issue:46421|issue:46858", "issue:42490|issue:43873", - "issue:43873|issue:45229", "issue:25251|issue:32090", + "issue:43873|issue:45229", + "issue:43976|issue:45440", "issue:44448|issue:44466", "issue:43381|issue:43701", - "issue:43976|issue:45440", "issue:44871|issue:44977", "issue:29127|issue:34567", "issue:33357|issue:39692", @@ -10127,223 +10152,216 @@ "issue:41720|issue:45229", "issue:44186|issue:45198", "issue:44361|issue:45198", - "issue:45198|issue:45803", - "issue:44977|issue:45563" + "issue:45198|issue:45803" ], - "prepared_review_unit_hash": "f4256ea590fe9cd4074c0b787e726b057c940cba2e93bd3736bc909f7d068983", + "prepared_review_unit_hash": "b827f2565d0d9de0c34183ff609a60fa9d06a476455c0a4caaf220e3a991941d", "component_budget": { "node_count": 514, "item_count": 514, "soft_pair_count": 1891, - "serialized_chars": 980926, - "estimated_input_tokens": 245232, - "estimated_eval_tokens": 490720 + "serialized_chars": 980895, + "estimated_input_tokens": 245224, + "estimated_eval_tokens": 490704 }, "budget": { - "node_count": 48, - "item_count": 48, - "soft_pair_count": 31, - "serialized_chars": 37321, - "estimated_input_tokens": 9331, - "estimated_eval_tokens": 18918 + "node_count": 47, + "item_count": 47, + "soft_pair_count": 30, + "serialized_chars": 36452, + "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 not a duplicate set: it mixes unrelated issues across tokenizers, generation, training, caching, vision, and packaging. No soft edge looks like the same underlying bug/change.", - "confidence": 0.99, - "canonical_issue_reason": "No canonical issue stands out because the items are heterogeneous and do not share a single root cause.", + "summary": "This cluster is largely heterogeneous: the suggested soft pairs mostly share only broad subsystem labels (tokenization, cache, training, serialization) but not the same concrete bug. I would not merge any of the proposed pairs as duplicates.", + "confidence": 0.94, + "canonical_issue_reason": null, "canonical_pr_reason": null, - "best_issue_reason": "No single issue is a good global representative; at best, issue 45920 is the broadest regression report, but it still does not match most of the cluster.", + "best_issue_reason": null, "best_pr_reason": null, "soft_edge_verdicts": [ { "left": "issue:37428", "right": "issue:42175", "accept": false, - "reason": "ImportError for flash attention utils vs missing TensorFlow backend package." + "reason": "Different problems in unrelated areas: flash-attention import error vs missing TensorFlow backend in a torch install." }, { "left": "issue:45290", "right": "issue:46655", "accept": false, - "reason": "Chat template tool-call crash vs video processor padding bug." + "reason": "Both are processing-related, but one is a chat template crash and the other is a video padding bug; different code paths." }, { - "left": "issue:29870", - "right": "issue:33290", + "left": "issue:44448", + "right": "issue:44977", "accept": false, - "reason": "BART generate() mask bug vs DeepSpeed Adafactor OOM." + "reason": "Generation mismatch for Pegasus v4/v5 is unrelated to Qwen3.5 flash-attention failures." }, { - "left": "issue:44448", - "right": "issue:44977", + "left": "issue:29870", + "right": "issue:33290", "accept": false, - "reason": "Pegasus output regression vs Qwen3.5 flash-attn generation crash." + "reason": "Bart generation UnboundLocalError and DeepSpeed Adafactor OOM are separate training/inference failures." }, { "left": "issue:44360", "right": "issue:44561", "accept": false, - "reason": "DSA indexer/ReLU discussion vs trust_remote_code FX helper removal." + "reason": "A DSA ReLU discussion and a v5 trust_remote_code regression are not the same bug." }, { "left": "issue:43278", "right": "issue:43381", "accept": false, - "reason": "Eval dtype drift vs gradient checkpointing unsupported in eval." + "reason": "dtype drift between train/eval is different from checkpointing being disallowed in eval mode." }, { "left": "issue:44855", "right": "issue:46399", "accept": false, - "reason": "Python 3.13 IndentationError vs missing decorator for kernel registration." + "reason": "Both mention decorators, but they fail in different models and for different reasons." }, { "left": "issue:40444", "right": "issue:43526", "accept": false, - "reason": "IterableDataset multi-image finetuning failure vs BEiT label reduction bug." + "reason": "Multi-image VL dataset finetuning failure is unrelated to BeitImageProcessorFast label reduction." }, { "left": "issue:45920", "right": "issue:46192", "accept": false, - "reason": "Tokenizer regression across many models vs ignored return_tensors kwarg." + "reason": "Wrong token IDs in AutoTokenizer is a different bug from return_tensors being ignored with text_kwargs." }, { "left": "issue:32090", "right": "issue:33290", "accept": false, - "reason": "Trainer broadcast NoneType error vs optimizer-related OOM." + "reason": "Trainer GPU broadcast NoneType error is not the same as Adafactor/DeepSpeed OOM." }, { - "left": "issue:44977", - "right": "issue:45229", + "left": "issue:43927", + "right": "issue:45698", "accept": false, - "reason": "Flash-attention generation crash vs multi-GPU CUDA OOM." + "reason": "Both involve save/load behavior, but one loses custom token IDs while the other loads the wrong custom module." }, { - "left": "issue:43927", - "right": "issue:45698", + "left": "issue:44977", + "right": "issue:45229", "accept": false, - "reason": "Config token IDs lost vs wrong custom module loaded after save/load." + "reason": "Flash-attention generation bug and multi-GPU CUDA OOM are different failure modes." }, { "left": "issue:46421", "right": "issue:46858", "accept": false, - "reason": "Spurious recompilation vs static-cache slice arithmetic failure." + "reason": "Cache lazy-init recompilation and device-tensor seq-length arithmetic breakage are distinct cache/compile issues." }, { "left": "issue:42490", "right": "issue:43873", "accept": false, - "reason": "Local-path import warning vs quantization offloading failure." + "reason": "Local-path import warning is unrelated to quantization offloading not working." + }, + { + "left": "issue:25251", + "right": "issue:32090", + "accept": false, + "reason": "Pipeline top_k nesting behavior is unrelated to Trainer GPU broadcast errors." }, { "left": "issue:43873", "right": "issue:45229", "accept": false, - "reason": "Offloading/quantization issue vs Gemma4 CUDA OOM." + "reason": "Quantization offloading failure is not the same as Gemma4 multi-GPU OOM." }, { - "left": "issue:25251", - "right": "issue:32090", + "left": "issue:43976", + "right": "issue:45440", "accept": false, - "reason": "Pipeline top_k shape change vs Trainer GPU broadcast error." + "reason": "Python version support regression and DeepSeekV3MoE implementation divergence are unrelated." }, { "left": "issue:44448", "right": "issue:44466", "accept": false, - "reason": "Generation output mismatch vs tied-weight serialization inconsistency." + "reason": "Output differences for Pegasus and lm_head serialization inconsistency are separate bugs." }, { "left": "issue:43381", "right": "issue:43701", "accept": false, - "reason": "Eval checkpointing restriction vs checkpoint resume key mismatch." - }, - { - "left": "issue:43976", - "right": "issue:45440", - "accept": false, - "reason": "Python version compatibility break vs DeepSeekV3MoE divergence." + "reason": "Eval-mode gradient checkpointing and resume_from_checkpoint key mismatch are different problems." }, { "left": "issue:44871", "right": "issue:44977", "accept": false, - "reason": "Gemma eos_token_id config mismatch vs Qwen3.5 flash-attn crash." + "reason": "Gemma eos_token_id config mismatch is unrelated to Qwen3.5 flash-attention generation failures." }, { "left": "issue:29127", "right": "issue:34567", "accept": false, - "reason": "LayoutLMv3 error clarity issue vs token counter not updating." + "reason": "LayoutLMv3 box-validation messaging and TrainerState token counting are unrelated." }, { "left": "issue:33357", "right": "issue:39692", "accept": false, - "reason": "MacOS bus error vs documentation example errors." + "reason": "MacOS bus error on CLIP is not the same as SigLIP2 documentation/example errors." }, { "left": "issue:43278", "right": "issue:43575", "accept": false, - "reason": "Eval dtype change vs tensor-parallel load OOM." + "reason": "Embedding dtype drift and Qwen2-57B tensor-parallel OOM are unrelated." }, { "left": "issue:44561", "right": "issue:44779", "accept": false, - "reason": "Removed FX helper breakage vs tokenizer regression in v5." + "reason": "Removal of is_torch_fx_available breaking trust_remote_code is a different regression from DeepSeek tokenizer output changes." }, { "left": "issue:45588", "right": "issue:45823", "accept": false, - "reason": "Flash-attention None handling vs device-map PLE mismatch." + "reason": "Flash-attention sink-less model crash and Gemma4 device-map mismatch are different forward-path bugs." }, { "left": "issue:44568", "right": "issue:45491", "accept": false, - "reason": "Special-token insertion bug vs Gemma3 NaN embeddings." + "reason": "Tokenizer special-token insertion bug is unrelated to Gemma3 NaN embeddings from sliding-window batching." }, { "left": "issue:41720", "right": "issue:45229", "accept": false, - "reason": "Device-mapping assert vs multi-GPU OOM." + "reason": "Qwen3 auto device mapping assert and Gemma4 multi-GPU OOM are different device-placement failures." }, { "left": "issue:44186", "right": "issue:45198", "accept": false, - "reason": "LayoutLMv2 tokenizer crash vs Wav2Vec2 save/tokenization failure." + "reason": "LayoutLMv2Tokenizer NER/padding crash and Wav2Vec2 save/tokenization failure are separate tokenizer issues." }, { "left": "issue:44361", "right": "issue:45198", "accept": false, - "reason": "MLuke tokenizer AttributeError vs Wav2Vec2 save/tokenization failure." + "reason": "MLukeTokenizer AttributeError is unrelated to Wav2Vec2 save_pretrained/tokenization failure." }, { "left": "issue:45198", "right": "issue:45803", "accept": false, - "reason": "Wav2Vec2 serialization/tokenization bug vs Fuyu exception handling bug." - }, - { - "left": "issue:44977", - "right": "issue:45563", - "accept": false, - "reason": "Generation crash vs stale warning text." + "reason": "Wav2Vec2 tokenizer/save bug and Fuyu's bare except swallowing errors are different defects." } ] }, @@ -10396,6 +10414,7 @@ "issue:44462", "issue:44479", "issue:44743", + "issue:44977", "issue:45072", "issue:45198", "issue:45245", @@ -10417,6 +10436,7 @@ ], "soft_pairs": [ "issue:43611|issue:46121", + "issue:44977|issue:45563", "issue:29870|issue:34689", "issue:29942|issue:30064", "issue:43653|issue:44743", @@ -10425,243 +10445,249 @@ "issue:43756|issue:45440", "issue:43653|issue:44442", "issue:43065|issue:43994", - "issue:43065|issue:45412", "issue:44265|issue:45198", + "issue:43065|issue:45412", "issue:43575|issue:43701", "issue:29127|issue:33357", - "issue:43412|issue:45588", - "issue:45372|issue:45588", "issue:45491|issue:46693", + "issue:45372|issue:45588", + "issue:43412|issue:45588", "issue:42490|issue:43381", - "issue:45245|issue:45563", "issue:29127|issue:35141", + "issue:45245|issue:45563", "issue:45459|issue:45712", "issue:45440|issue:45593", "issue:43994|issue:45412", "issue:45372|issue:45406", "issue:45406|issue:46701", - "issue:43575|issue:44387", "issue:43421|issue:45491", - "issue:45412|issue:46369", + "issue:43575|issue:44387", "issue:43066|issue:44462", + "issue:45412|issue:46369", "issue:45372|issue:45459", "issue:44442|issue:45072", "issue:43526|issue:45834" ], - "prepared_review_unit_hash": "9391ddd2879d5067b0737b0a1d886ba69cb782e2a0c0e5dad5d7aa4869ac2c4f", + "prepared_review_unit_hash": "2c2b63fa9e8f73b6f11e9ac9cf16cdd2960d118ea5c4fdd17572efcb7966dd17", "component_budget": { "node_count": 514, "item_count": 514, "soft_pair_count": 1891, - "serialized_chars": 980926, - "estimated_input_tokens": 245232, - "estimated_eval_tokens": 490720 + "serialized_chars": 980895, + "estimated_input_tokens": 245224, + "estimated_eval_tokens": 490704 }, "budget": { - "node_count": 47, - "item_count": 47, - "soft_pair_count": 31, - "serialized_chars": 36514, - "estimated_input_tokens": 9129, - "estimated_eval_tokens": 18514 + "node_count": 48, + "item_count": 48, + "soft_pair_count": 32, + "serialized_chars": 37361, + "estimated_input_tokens": 9341, + "estimated_eval_tokens": 18938 }, "overflow_policy": "truncate_then_skip", "trimmed": false, "aggressively_trimmed": false, "split": true, "analyst_result": { - "summary": "These items are a collection of unrelated Transformers bugs across different models and subsystems; none of the soft pairs look like the same underlying issue, so all should be rejected as duplicates.", - "confidence": 0.96, - "canonical_issue_reason": "issue:44462 is the clearest broadly relevant bug report in the set: it has concrete reproduction details, impacts AutoTokenizer loading, and is phrased as a precise user-facing regression.", + "summary": "These issues are mostly unrelated topical matches rather than true duplicates. I would not merge any soft pair. If a representative issue is needed, issue 46693 is the strongest standalone regression report by activity and references.", + "confidence": 0.79, + "canonical_issue_reason": "No clear duplicate cluster exists; issue 46693 is the most active, concrete regression report and is the least arbitrary anchor among the set.", "canonical_pr_reason": null, - "best_issue_reason": "issue:44462 is the best issue candidate because it is specific, actionable, and well-scoped to a real loading-path problem with likely wider impact than the more model-specific reports.", + "best_issue_reason": "Issue 46693 has the richest discussion and inbound references, and it describes a user-visible regression with a well-defined failure mode.", "best_pr_reason": null, "soft_edge_verdicts": [ { "left": "issue:43611", "right": "issue:46121", "accept": false, - "reason": "Both are loading/config-related, but one is a base_model_prefix regression and the other is a rope-validation TypeError; different code paths and failures." + "reason": "Different bugs: model loading via base_model_prefix vs rope-validation TypeError from JSON-loaded lists." + }, + { + "left": "issue:44977", + "right": "issue:45563", + "accept": false, + "reason": "Different failure modes: flash-attention generation breakage vs a stale warning in paged generate()." }, { "left": "issue:29870", "right": "issue:34689", "accept": false, - "reason": "BART generate() decoder mask initialization and Llama 3.2 vision model loading are unrelated bugs." + "reason": "Unrelated subsystems: BART generate() decoder mask bug vs Llama 3.2 Vision model loading regression." }, { "left": "issue:29942", "right": "issue:30064", "accept": false, - "reason": "Flash Attention 2 test failures and void segmentation map processing are different subsystems with different failure modes." + "reason": "One is Flash Attention 2 test failures; the other is image processor handling of void segmentation maps." }, { "left": "issue:43653", "right": "issue:44743", "accept": false, - "reason": "Tokenizer special-token registration and Qwen3.5 recurrent-state caching are not the same bug." + "reason": "Tokenizer special-token decode bug vs recurrent-state reset in modular_qwen3_5 cache logic." }, { "left": "issue:44206", "right": "issue:44479", "accept": false, - "reason": "Unsupported center argument in a feature extractor is unrelated to the Qwen VL video-input regression." + "reason": "Feature-extractor arg crash vs video-input regression across Qwen VL variants." }, { "left": "issue:42915", "right": "issue:42994", "accept": false, - "reason": "Qwen3Moe FP8 config failure and quantized model saving are separate quantization bugs." + "reason": "Both involve quantization broadly, but one is a FineGrainedFP8 config failure and the other is quantized model saving." }, { "left": "issue:43756", "right": "issue:45440", "accept": false, - "reason": "Smollm3 RoPE-layer dropping and DeepseekV3MoE divergence are different model implementation issues." + "reason": "Different models and defects: missing RoPE layers in Smollm3 vs divergence in DeepseekV3MoE implementation." }, { "left": "issue:43653", "right": "issue:44442", "accept": false, - "reason": "BigBird tokenizer special-token registration and FastSpeech2ConformerTokenizer loading are distinct tokenizer bugs." + "reason": "Tokenizer decode-output bug vs AutoTokenizer failing to load a specific tokenizer class." }, { "left": "issue:43065", "right": "issue:43994", "accept": false, - "reason": "A dummy Conv2d placeholder in Sam3PixelDecoder is not the same as SigLIP2 producing nonsensical outputs." + "reason": "Different vision-model issues: dummy Conv2d in Sam3PixelDecoder vs nonsensical SigLIP2 inference." }, { - "left": "issue:43065", - "right": "issue:45412", + "left": "issue:44265", + "right": "issue:45198", "accept": false, - "reason": "Sam3PixelDecoder dummy layer and RT-DETR memory release are unrelated issues." + "reason": "torch.export failure from torch_compilable_check is unrelated to Wav2Vec2 save_pretrained/tokenization failures." }, { - "left": "issue:44265", - "right": "issue:45198", + "left": "issue:43065", + "right": "issue:45412", "accept": false, - "reason": "torch.export failures with torch_compilable_check are unrelated to Wav2Vec2 save_pretrained/tokenization failures." + "reason": "Sam3PixelDecoder dummy layer issue is unrelated to RT-DETR memory not being released on deletion." }, { "left": "issue:43575", "right": "issue:43701", "accept": false, - "reason": "Tensor-parallel OOM on Qwen2-57B and resume_from_checkpoint key mismatch are different problems." + "reason": "OOM on tensor-parallel model loading is unrelated to resume_from_checkpoint key mismatch." }, { "left": "issue:29127", "right": "issue:33357", "accept": false, - "reason": "LayoutLMv3 box info error messaging and a CLIP MacOS bus error are not the same underlying bug." + "reason": "LayoutLMv3 box-information error and MacOS CLIP bus error are different models, symptoms, and code paths." }, { - "left": "issue:43412", - "right": "issue:45588", + "left": "issue:45491", + "right": "issue:46693", "accept": false, - "reason": "Gemma3n export guard/erfinv support and flash_attention s_aux=None crashes occur in different code paths." + "reason": "Both mention generation context, but one is a Gemma3 NaN-embedding bug and the other is a FA2 long-context performance regression." }, { "left": "issue:45372", "right": "issue:45588", "accept": false, - "reason": "Gemma 4 processor loading import error and flash_attention sink handling are unrelated." + "reason": "Gemma 4 processor import failure is unrelated to flash_attention.py crashing on s_aux=None." }, { - "left": "issue:45491", - "right": "issue:46693", + "left": "issue:43412", + "right": "issue:45588", "accept": false, - "reason": "Gemma3 NaN embeddings from all-padding windows and FA2 long-context performance regression are different symptoms with different root causes." + "reason": "Executorch export guard/erfinv support issue is a different code path from the flash_attention.py AttributeError." }, { "left": "issue:42490", "right": "issue:43381", "accept": false, - "reason": "Path-with-slash import warning and gradient-checkpointing eval-mode behavior are not the same issue." + "reason": "Local-path import warning is unrelated to gradient checkpointing being disallowed in eval mode." }, { - "left": "issue:45245", - "right": "issue:45563", + "left": "issue:29127", + "right": "issue:35141", "accept": false, - "reason": "Category-cardinality overflow and a stale generate() warning are different generate/save behaviors." + "reason": "LayoutLMv3 error messaging and embedding reinitialization after resize_embeddings are separate bugs." }, { - "left": "issue:29127", - "right": "issue:35141", + "left": "issue:45245", + "right": "issue:45563", "accept": false, - "reason": "LayoutLMv3 error clarity and token-embedding reinitialization after resize are unrelated." + "reason": "Category-count RuntimeError is unrelated to the paged generate stale warning." }, { "left": "issue:45459", "right": "issue:45712", "accept": false, - "reason": "Protobuf decode error masking tokenizer failures and leftover dummy classes in dummy_pt_objects are separate cleanup/exception-handling issues." + "reason": "Tokenizer protobuf error handling is unrelated to leftover dummy PT classes leaking into dir(transformers)." }, { "left": "issue:45440", "right": "issue:45593", "accept": false, - "reason": "DeepseekV3MoE divergence and D-FINE auxiliary-loss behavior are different model bugs." + "reason": "DeepseekV3MoE divergence is unrelated to D-FINE auxiliary-loss behavior." }, { "left": "issue:43994", "right": "issue:45412", "accept": false, - "reason": "SigLIP2 bad outputs and RT-DETR memory leak do not share the same underlying cause." + "reason": "SigLIP2 inference quality issue and RT-DETR memory leak are different subsystems and symptoms." }, { "left": "issue:45372", "right": "issue:45406", "accept": false, - "reason": "ReasoningEffort import failure and Gemma4Processor missing _tokenizer are distinct processor-loading regressions." + "reason": "Gemma 4 processor loading import error is unrelated to transformers serve lacking Gemma4Processor._tokenizer." }, { "left": "issue:45406", "right": "issue:46701", "accept": false, - "reason": "Gemma4Processor attribute error and DiffusionGemma image-text pipeline failure are different bugs." - }, - { - "left": "issue:43575", - "right": "issue:44387", - "accept": false, - "reason": "TP OOM on Qwen2-57B and int4 reserved-memory regression are both memory-related but not the same code-path problem." + "reason": "Missing _tokenizer in serve is unrelated to DiffusionGemmaGenerationOutput pipeline breakage." }, { "left": "issue:43421", "right": "issue:45491", "accept": false, - "reason": "Runtime post-processor updates for special tokens and Gemma3 NaN embeddings are unrelated." + "reason": "Runtime post-processor update behavior is unrelated to Gemma3 NaN embeddings on mixed-length batches." }, { - "left": "issue:45412", - "right": "issue:46369", + "left": "issue:43575", + "right": "issue:44387", "accept": false, - "reason": "RT-DETR garbage-collection memory leak and sam3-litetext memory overflow are different memory issues." + "reason": "Model-loading OOM for Qwen2-57B is not the same as increased CUDA reserved memory under int4 quantization." }, { "left": "issue:43066", "right": "issue:44462", "accept": false, - "reason": "Wrong tokenizer decoder type in v5 and AutoTokenizer ignoring tokenizer.json are different tokenizer-loading problems." + "reason": "Wrong tokenizer decoder type in v5 and ignoring tokenizer.json are related to tokenizers, but they are distinct loading bugs." + }, + { + "left": "issue:45412", + "right": "issue:46369", + "accept": false, + "reason": "RT-DETR deletion/GC memory leak is different from SAM3 litetext memory overflow during model use." }, { "left": "issue:45372", "right": "issue:45459", "accept": false, - "reason": "Gemma 4 processor import failure and protobuf-missing tokenizer decode masking are unrelated exception paths." + "reason": "Gemma 4 processor import failure is unrelated to the protobuf decode-error masking bug." }, { "left": "issue:44442", "right": "issue:45072", "accept": false, - "reason": "FastSpeech2ConformerTokenizer loading and bfloat16 dtype mismatches in SwitchTransformers/TimmWrapperModel are separate bugs." + "reason": "FastSpeech2ConformerTokenizer loading failure is unrelated to bfloat16 dtype mismatches in SwitchTransformers/TimmWrapperModel." }, { "left": "issue:43526", "right": "issue:45834", "accept": false, - "reason": "BeitImageProcessorFast reduce_labels truncation and Kosmos2.5 long-OCR index errors are not the same underlying issue." + "reason": "BeitImageProcessorFast reduce_labels bug is unrelated to Kosmos2.5 OCR index errors." } ] }, @@ -10737,14 +10763,14 @@ "soft_pairs": [ "issue:27957|issue:33290", "issue:27957|issue:29127", - "issue:43526|issue:46693", "issue:43824|issue:44279", + "issue:43526|issue:46693", "issue:33290|issue:34689", "issue:33290|issue:35141", "issue:44857|issue:45588", "issue:43994|issue:46369", - "issue:42915|issue:43931", "issue:43526|issue:44186", + "issue:42915|issue:43931", "issue:43334|issue:43531", "issue:45250|issue:46640", "issue:43381|issue:43873", @@ -10752,208 +10778,214 @@ "issue:43646|issue:44077", "issue:46121|issue:46399", "issue:44514|issue:45381", - "issue:44977|issue:45245", "issue:43761|issue:45859", + "issue:44977|issue:45245", "issue:44743|issue:44811", - "issue:43262|issue:44206", "issue:44442|issue:44898", + "issue:43262|issue:44206", "issue:42490|issue:44743", + "issue:42175|issue:46234", "issue:43720|issue:44442", "issue:27957|issue:29870", - "issue:42175|issue:46234", "issue:43316|issue:43525", "issue:45003|issue:45440" ], - "prepared_review_unit_hash": "3ad74388bdf2ae424ec8c2cc4e58eee4b89be867808ea2805686ccfaac05bbf5", + "prepared_review_unit_hash": "6ca5e913b3bae805e06fbc2af5cbb9b7513da3986fc6625509ce951710e60d1b", "component_budget": { "node_count": 514, "item_count": 514, "soft_pair_count": 1891, - "serialized_chars": 980926, - "estimated_input_tokens": 245232, - "estimated_eval_tokens": 490720 + "serialized_chars": 980895, + "estimated_input_tokens": 245224, + "estimated_eval_tokens": 490704 }, "budget": { "node_count": 48, "item_count": 48, "soft_pair_count": 28, - "serialized_chars": 35955, - "estimated_input_tokens": 8989, - "estimated_eval_tokens": 18234 + "serialized_chars": 35948, + "estimated_input_tokens": 8987, + "estimated_eval_tokens": 18230 }, "overflow_policy": "truncate_then_skip", "trimmed": false, "aggressively_trimmed": false, "split": true, "analyst_result": { - "summary": "This cluster is heterogeneous: most pairs are only superficially similar by subsystem or model family, and none look like safe duplicates. I\u2019d use the open Flash Attention 2 report as the best anchor, but it does not unify the rest of the set.", - "confidence": 0.91, - "canonical_issue_reason": "issue:46640 is the most representative concrete issue in the only loosely related flash-attention subgroup; it is more substantive than the vague FA2 placeholder issue:45250.", + "summary": "Mostly a heterogeneous set of unrelated Transformers issues. The only recurring theme is Flash Attention 2, but the paired items describe different failures and code paths, so no soft pair should be merged.", + "confidence": 0.86, + "canonical_issue_reason": "Issue 46640 is the clearest current anchor for the Flash Attention 2 subgroup: it is explicitly about FA2, open, and broad enough to collect related reports.", "canonical_pr_reason": null, - "best_issue_reason": "issue:46640 has the clearest actionable bug report and the broadest relevance within the cluster\u2019s small flash-attention subset, while still being specific enough to serve as the anchor issue.", + "best_issue_reason": "Issue 46640 is the best representative issue in this set because it is the most directly on-topic and general within the only visible cluster theme (FA2).", "best_pr_reason": null, "soft_edge_verdicts": [ { "left": "issue:27957", "right": "issue:33290", "accept": false, - "reason": "Unrelated failures: XLM-RoBERTa Flash Attention 2 support vs AdaFactor/DeepSpeed OOM." + "reason": "XLMRoberta/Flash Attention 2 and deepspeed Adafactor OOM are different bugs in different paths." }, { "left": "issue:27957", "right": "issue:29127", "accept": false, - "reason": "Different models and symptoms: FA2 support on XLM-RoBERTa vs LayoutLMv3 error messaging." + "reason": "FA2 model support vs LayoutLMv3 error messaging; unrelated failure modes." }, { - "left": "issue:43526", - "right": "issue:46693", + "left": "issue:43824", + "right": "issue:44279", "accept": false, - "reason": "Different bugs: BeitImageProcessorFast label reduction vs FA2 long-context performance regression." + "reason": "Importing a missing Qwen2.5-VL class is not the same as a generic dependency issue." }, { - "left": "issue:43824", - "right": "issue:44279", + "left": "issue:43526", + "right": "issue:46693", "accept": false, - "reason": "Different problem scopes: missing Qwen2.5-VL import vs a generic dependency issue." + "reason": "BeitImageProcessorFast label reduction bug and FA2 long-context performance regression are unrelated." }, { "left": "issue:33290", "right": "issue:34689", "accept": false, - "reason": "OOM during training with AdaFactor/DeepSpeed is not the same as Llama 3.2 Vision model loading breakage." + "reason": "Adafactor/deepspeed OOM and Llama 3.2 Vision load break are distinct problems." }, { "left": "issue:33290", "right": "issue:35141", "accept": false, - "reason": "Training OOM and embedding reinitialization are separate code paths and failures." + "reason": "Optimizer memory usage and embedding reinitialization during post_init are unrelated code paths." }, { "left": "issue:44857", "right": "issue:45588", "accept": false, - "reason": "Different failures: LwDetrImageLoss AMP/CUDA crash vs flash_attention sink-less model AttributeError." + "reason": "AMP/CUDA loss crash and flash_attention.py s_aux=None crash do not describe the same bug." }, { "left": "issue:43994", "right": "issue:46369", "accept": false, - "reason": "Different models and symptoms: SigLIP2 bad outputs vs SAM3-LiteText memory overflow." + "reason": "Siglip2 bad outputs and sam3-litetext memory overflow are different issues." }, { - "left": "issue:42915", - "right": "issue:43931", + "left": "issue:43526", + "right": "issue:44186", "accept": false, - "reason": "Qwen3MoE FineGrainedFP8Config failure is not the same as Qwen3-VL weight-shape mismatch." + "reason": "ImageProcessorFast label reduction and LayoutLMv2Tokenizer NER/padding crashes are unrelated." }, { - "left": "issue:43526", - "right": "issue:44186", + "left": "issue:42915", + "right": "issue:43931", "accept": false, - "reason": "Separate image-processing/tokenizer bugs with no shared concrete code path." + "reason": "Qwen3Moe FineGrainedFP8Config failure and Qwen3-VL weight-shape mismatch are separate model-loading issues." }, { "left": "issue:43334", "right": "issue:43531", "accept": false, - "reason": "Qwen3-VL missing pad_token_id and Qwen3-MoE sliding_window are different config/runtime issues." + "reason": "Missing pad_token_id on Qwen3-VL text config is unrelated to sliding_window behavior in Qwen3-MoE." }, { "left": "issue:45250", "right": "issue:46640", "accept": false, - "reason": "Both mention Flash Attention 2, but the titles are too vague to show the same underlying bug or mergeable fix." + "reason": "Both mention Flash Attention 2, but 45250 is a generic topic/title while 46640 is a specific crash report; not enough evidence of the same bug." }, { "left": "issue:43381", "right": "issue:43873", "accept": false, - "reason": "Gradient checkpointing eval-mode behavior is unrelated to quantized offloading." + "reason": "Gradient checkpointing in eval mode and quantization offloading misbehavior are different problems." }, { "left": "issue:45588", "right": "issue:46821", "accept": false, - "reason": "flash_attention.py sink handling is unrelated to TimesFM window_size attribute errors." + "reason": "flash_attention.py AttributeError on sink-less models and TimesFM window_size AttributeError are unrelated." }, { "left": "issue:43646", "right": "issue:44077", "accept": false, - "reason": "Custom model initialization breakage and patchtsmixer post_init policy are different issues, not one concrete fix." + "reason": "Custom model initialization break in Transformers 5.0 and patchtsmixer post_init policy are not the same change." }, { "left": "issue:46121", "right": "issue:46399", "accept": false, - "reason": "Both touch rotary/rope-related areas, but one is a TypeError in validation and the other is a registration/decorator crash." + "reason": "JSON-loaded list handling in rope validation and Qwen3.5 rotary embedding registration are different code paths." }, { "left": "issue:44514", "right": "issue:45381", "accept": false, - "reason": "Different Qwen2.5-VL bugs: batched apply_chat_template crash vs incorrect vision_position_ids for video input." + "reason": "Batched chat-template padding crash and wrong video vision_position_ids are distinct Qwen2.5-VL issues." }, { - "left": "issue:44977", - "right": "issue:45245", + "left": "issue:43761", + "right": "issue:45859", "accept": false, - "reason": "Flash-attention generation failure and category-cardinality runtime error are unrelated." + "reason": "CLIPVisionModel hidden_states regression and Qwen3_5MoeTextRotaryEmbedding CPU offload incompatibility are unrelated." }, { - "left": "issue:43761", - "right": "issue:45859", + "left": "issue:44977", + "right": "issue:45245", "accept": false, - "reason": "CLIPVisionModel hidden_states regression is unrelated to Qwen3_5MoE text rotary offload compatibility." + "reason": "Qwen3.5 flash-attention generation bug and category-count overflow are unrelated." }, { "left": "issue:44743", "right": "issue:44811", "accept": false, - "reason": "Recurrent-state cache reset in Qwen3.5 modular code is unrelated to Whisper batch_decode skip_special_tokens handling." + "reason": "Qwen3.5 recurrent-state reset and Whisper batch_decode skip_special_tokens are different components and symptoms." + }, + { + "left": "issue:44442", + "right": "issue:44898", + "accept": false, + "reason": "FastSpeech2ConformerTokenizer loading and Perceiver non-default resolution failure are unrelated." }, { "left": "issue:43262", "right": "issue:44206", "accept": false, - "reason": "Audio chat-template sampling-rate defaulting and unsupported center arg in LasrFeatureExtractor are different bugs." + "reason": "Audio chat-template sample-rate default and LasrFeatureExtractor center-arg crash are different audio issues." }, { - "left": "issue:44442", - "right": "issue:44898", + "left": "issue:42490", + "right": "issue:44743", + "accept": false, + "reason": "Local path trailing-slash import warning and Qwen3.5 recurrent-state reset are unrelated." + }, + { + "left": "issue:42175", + "right": "issue:46234", "accept": false, - "reason": "Tokenizer loading failure and Perceiver interpolation failure are separate model/processor issues." + "reason": "Missing TensorFlow backend in extra install and missing import_utils helper are not the same dependency bug." }, { "left": "issue:43720", "right": "issue:44442", "accept": false, - "reason": "BitNet packed-weight loading and FastSpeech2Conformer tokenizer loading are unrelated." + "reason": "BitNet packed-weight loading and tokenizer load failure are different failures." }, { "left": "issue:27957", "right": "issue:29870", "accept": false, - "reason": "Different failures: XLM-RoBERTa FA2 support vs BART generate() decoder_attention_mask unbound variable." - }, - { - "left": "issue:42175", - "right": "issue:46234", - "accept": false, - "reason": "Packaging backend omission and missing import_utils helper are unrelated dependency issues." + "reason": "FA2 support for XLMRoberta and Bart generate decoder_attention_mask UnboundLocalError are unrelated." }, { "left": "issue:43316", "right": "issue:43525", "accept": false, - "reason": "Both are config-API discrepancies, but they concern different models and different missing attributes." + "reason": "Gemma3TextConfig API discrepancy and Llama4Config missing pad_token_id are separate config issues." }, { "left": "issue:45003", "right": "issue:45440", "accept": false, - "reason": "Unsafe sys.modules access in modeling_utils is unrelated to DeepSeekV3MoE implementation divergence." + "reason": "Unsafe sys.modules access in modeling_utils and DeepseekV3MoE divergence are unrelated bugs." } ] }, @@ -11029,54 +11061,54 @@ "issue:44077|issue:44479", "issue:43493|issue:43827", "issue:43873|issue:44871", - "issue:45325|issue:46697", "issue:44062|issue:45907", - "issue:43643|issue:44913", + "issue:45325|issue:46697", "issue:44855|issue:45137", + "issue:43643|issue:44913", "issue:43316|issue:45759", "issue:43749|issue:43931", "issue:44987|issue:46178", "issue:44190|issue:44496", "issue:43931|issue:45216", - "issue:42617|issue:43575", "issue:43701|issue:46506", "issue:44442|issue:45803", + "issue:42617|issue:43575", "issue:43232|issue:43388", "issue:43550|issue:43701", "issue:43575|issue:43742", "issue:43761|issue:44079", - "issue:27957|issue:32090", "issue:43761|issue:44188", + "issue:27957|issue:32090", "issue:43606|issue:43701", "issue:25251|issue:28282", "issue:46829|issue:46932", "issue:42907|issue:45698", "issue:16998|issue:22355" ], - "prepared_review_unit_hash": "b34f724d98d9b7f3e133fafc17468d7089dd2feea4c4a2b0686b9043689ca109", + "prepared_review_unit_hash": "c455cdbc834ebb07e8bfa03a525abea5fe9f81de7207058c4eb40a9709de9efd", "component_budget": { "node_count": 514, "item_count": 514, "soft_pair_count": 1891, - "serialized_chars": 980926, - "estimated_input_tokens": 245232, - "estimated_eval_tokens": 490720 + "serialized_chars": 980895, + "estimated_input_tokens": 245224, + "estimated_eval_tokens": 490704 }, "budget": { "node_count": 47, "item_count": 47, "soft_pair_count": 26, - "serialized_chars": 34972, - "estimated_input_tokens": 8743, - "estimated_eval_tokens": 17742 + "serialized_chars": 34974, + "estimated_input_tokens": 8744, + "estimated_eval_tokens": 17744 }, "overflow_policy": "truncate_then_skip", "trimmed": false, "aggressively_trimmed": false, "split": true, "analyst_result": { - "summary": "This cluster is a heterogeneous set of unrelated transformer issues. None of the soft pairs look like the same underlying bug or change, so no merges are warranted.", - "confidence": 0.97, + "summary": "The cluster is mostly a set of unrelated Transformers issues that only share superficial semantic similarity; I do not see a true duplicate group here.", + "confidence": 0.95, "canonical_issue_reason": null, "canonical_pr_reason": null, "best_issue_reason": null, @@ -11086,157 +11118,157 @@ "left": "issue:44077", "right": "issue:44479", "accept": false, - "reason": "Different models and failures: an API removal for patchtsmixer vs a video-input regression for Qwen VL variants." + "reason": "Different bugs: one is about `post_init` API policy for patchtsmixer, the other is a v5 video-input regression across Qwen-VL models." }, { "left": "issue:43493", "right": "issue:43827", "accept": false, - "reason": "SigLIP2 implementation mismatch is unrelated to docs still referencing `pipeline()` after v5 removals." + "reason": "SigLIP2 implementation mismatch vs documentation still referencing removed `pipeline()` calls; unrelated topics and code paths." }, { "left": "issue:43873", "right": "issue:44871", "accept": false, - "reason": "Quantization/offloading behavior and Gemma-3 EOS token config inconsistency are separate issues." - }, - { - "left": "issue:45325", - "right": "issue:46697", - "accept": false, - "reason": "Both concern preprocessing math, but one is rope/temporal position scaling and the other is image over-padding." + "reason": "Quantization/offloading behavior versus Gemma-3 eos token config inconsistency; not the same underlying defect." }, { "left": "issue:44062", "right": "issue:45907", "accept": false, - "reason": "Tokenizer construction error is unrelated to `list_repo_files` not forwarding `revision`." + "reason": "Tokenizer `AddedToken` argument collision and `list_repo_files` not forwarding `revision` are unrelated API bugs." }, { - "left": "issue:43643", - "right": "issue:44913", + "left": "issue:45325", + "right": "issue:46697", "accept": false, - "reason": "`AutoConfig.from_pretrained` missing fields with remote code is not the same as GPTNeoX rotary_pct not persisting on reload." + "reason": "Qwen2.5-VL rope/position-id scaling bug and Swin2SR over-padding bug affect different models and functions." }, { "left": "issue:44855", "right": "issue:45137", "accept": false, - "reason": "Python 3.13 TorchScript parsing error is unrelated to the DeepSpeed ZeRO3 deque underflow." + "reason": "Python 3.13 `torch.jit.script` parsing issue in DebertaV2 import is unrelated to DeepSpeed ZeRO3 deque handling." + }, + { + "left": "issue:43643", + "right": "issue:44913", + "accept": false, + "reason": "`trust_remote_code` config field loss and GPTNeoX `rotary_pct` not persisting on reload are different config-loading problems." }, { "left": "issue:43316", "right": "issue:45759", "accept": false, - "reason": "Both involve config APIs, but one is a Gemma3TextConfig discrepancy and the other is `from_config` not unwrapping `text_config` for Qwen multimodal configs." + "reason": "Both involve config APIs, but one is a Gemma3TextConfig discrepancy and the other is unwrapping `text_config` for composite Qwen multimodal configs; not the same bug." }, { "left": "issue:43749", "right": "issue:43931", "accept": false, - "reason": "Broken FSDP CPU RAM-efficient loading is a different path from Qwen3-VL shape mismatch on model load." + "reason": "FSDP RAM-efficient loading regression and Qwen3-VL weight-shape mismatch are distinct model-loading failures." }, { "left": "issue:44987", "right": "issue:46178", "accept": false, - "reason": "Model-loading regression for a specific repo is unrelated to a backend mapping error for `tensorflow_text`." + "reason": "Loading `physical-intelligence/fast` failing in newer Transformers is unrelated to an unsupported backend entry for `tensorflow_text`." }, { "left": "issue:44190", "right": "issue:44496", "accept": false, - "reason": "Local dataset loading in a training script is unrelated to unrecognized model config loading." + "reason": "Local dataset loading in a training script and unrecognized model config are separate issues." }, { "left": "issue:43931", "right": "issue:45216", "accept": false, - "reason": "Loading shape mismatch and incorrect checkpoint saving are different Qwen3-VL/Qwen3.5 problems." - }, - { - "left": "issue:42617", - "right": "issue:43575", - "accept": false, - "reason": "`3d_parallel.py` runtime failure is not the same concrete issue as TP loading OOM for Qwen2-57B." + "reason": "One is a load-time shape mismatch for Qwen3-VL, the other is a save-pretrained checkpoint regression for Qwen3.5; different directions and code paths." }, { "left": "issue:43701", "right": "issue:46506", "accept": false, - "reason": "Checkpoint key mismatch and CPU/CUDA output divergence are unrelated." + "reason": "Resume-from-checkpoint key mismatch and CPU/CUDA divergence in PPDocLayoutV3 are unrelated." }, { "left": "issue:44442", "right": "issue:45803", "accept": false, - "reason": "Tokenizer loading failure is unrelated to Fuyu tensor conversion error handling." + "reason": "FastSpeech2Conformer tokenizer loading failure and a bare `except` in Fuyu tensor conversion are not the same bug." + }, + { + "left": "issue:42617", + "right": "issue:43575", + "accept": false, + "reason": "`3d_parallel.py` run failure and Qwen2-57B-A14B-Instruct tensor-parallel OOM are different runtime problems." }, { "left": "issue:43232", "right": "issue:43388", "accept": false, - "reason": "Generation kwargs sync bug and metric-gathering label truncation are separate code paths." + "reason": "Generation kwargs sync issue and `gather_for_metrics` dropping tuple labels affect different training/generation code paths." }, { "left": "issue:43550", "right": "issue:43701", "accept": false, - "reason": "torch.compile/SDPA failure in Bamba is unrelated to resume-from-checkpoint key mismatch." + "reason": "Bamba torch.compile/SDPA failure is unrelated to resume-from-checkpoint key mismatch." }, { "left": "issue:43575", "right": "issue:43742", "accept": false, - "reason": "Qwen2 TP OOM and MobileLLM key error are distinct model-loading failures." + "reason": "Large-model OOM on load and MobileLLM key error are different model-loading failures." }, { "left": "issue:43761", "right": "issue:44079", "accept": false, - "reason": "A model-specific hidden_states regression is not the same as a generic ModelOutput key assignment bug." + "reason": "CLIPVisionModel not returning hidden states and ModelOutput key assignment when a key was previously `None` are related in theme but not clearly the same underlying bug." }, { - "left": "issue:27957", - "right": "issue:32090", + "left": "issue:43761", + "right": "issue:44188", "accept": false, - "reason": "Flash Attention support for XLM-Roberta is unrelated to the Trainer `_gpu_broadcast_one` TypeError." + "reason": "Hidden-states regression in CLIPVisionModel is unrelated to attention-kernel divergence under `torch.compile`." }, { - "left": "issue:43761", - "right": "issue:44188", + "left": "issue:27957", + "right": "issue:32090", "accept": false, - "reason": "Hidden_states output regression and attention-kernel divergence under torch.compile are different bugs." + "reason": "Flash Attention 2 support for XLMRoberta and a Trainer `_gpu_broadcast_one` NoneType error are unrelated." }, { "left": "issue:43606", "right": "issue:43701", "accept": false, - "reason": "CPU offload device mismatch is unrelated to resume checkpoint key mismatch." + "reason": "CPU-offload device mismatch for bark-small and resume-from-checkpoint key mismatch do not share the same cause." }, { "left": "issue:25251", "right": "issue:28282", "accept": false, - "reason": "Pipeline `top_k` output nesting bug is unrelated to missing PyTorch dependency detection." + "reason": "Pipeline `top_k` output nesting and missing PyTorch for AutoModel are unrelated." }, { "left": "issue:46829", "right": "issue:46932", "accept": false, - "reason": "VideoPrism output field mix-up is unrelated to Gemma4 quantized embedder dtype casting." + "reason": "VideoPrism hidden_states misassignment and Gemma4 quantized embedder dtype casting are distinct bugs." }, { "left": "issue:42907", "right": "issue:45698", "accept": false, - "reason": "Saving dequantized models and loading the wrong custom module after save/load are different serialization bugs." + "reason": "Saving dequantized models and loading the wrong custom module after save_pretrained are both serialization-related, but not the same concrete defect." }, { "left": "issue:16998", "right": "issue:22355", "accept": false, - "reason": "A model_max_length question for DeBERTa-V3 is unrelated to missing `transformers.onnx` import." + "reason": "A question about `model_max_length` and a missing `transformers.onnx` module are unrelated." } ] }, @@ -11260,7 +11292,6 @@ "model": "gpt-5.4-mini?service_tier=flex", "cluster_id": "cluster-10000-514", "nodes": [ - "issue:42175", "issue:43065", "issue:43295", "issue:43316", @@ -11307,6 +11338,7 @@ "issue:46421", "issue:46424", "issue:46640", + "issue:46821", "issue:46894" ], "soft_pairs": [ @@ -11314,18 +11346,18 @@ "issue:43295|issue:44871", "issue:44871|issue:45584", "issue:46234|issue:46291", - "issue:44077|issue:44534", "issue:43316|issue:43716", - "issue:42175|issue:43316", + "issue:44077|issue:44534", + "issue:43316|issue:46821", "issue:44464|issue:46421", - "issue:43646|issue:44913", "issue:43452|issue:44496", + "issue:43646|issue:44913", "issue:44871|issue:46050", "issue:45803|issue:46365", "issue:44162|issue:44792", "issue:43404|issue:43441", - "issue:43065|issue:44617", "issue:44242|issue:46894", + "issue:43065|issue:44617", "issue:43540|issue:44514", "issue:44610|issue:46369", "issue:45632|issue:46424", @@ -11339,202 +11371,202 @@ "issue:43352|issue:46640", "issue:45127|issue:45200" ], - "prepared_review_unit_hash": "59168b96a37a9ec0a328105bc1e5fa8cdf9d4eeccaeee491e8de5511a2d468c6", + "prepared_review_unit_hash": "15a013321fbc76ce26218c8cd8142b2d979182412b9abe118c048ba4c12fabd4", "component_budget": { "node_count": 514, "item_count": 514, "soft_pair_count": 1891, - "serialized_chars": 980926, - "estimated_input_tokens": 245232, - "estimated_eval_tokens": 490720 + "serialized_chars": 980895, + "estimated_input_tokens": 245224, + "estimated_eval_tokens": 490704 }, "budget": { "node_count": 48, "item_count": 48, "soft_pair_count": 28, - "serialized_chars": 35938, - "estimated_input_tokens": 8985, - "estimated_eval_tokens": 18226 + "serialized_chars": 35927, + "estimated_input_tokens": 8982, + "estimated_eval_tokens": 18220 }, "overflow_policy": "truncate_then_skip", "trimmed": false, "aggressively_trimmed": false, "split": true, "analyst_result": { - "summary": "This set is mostly a grab bag of unrelated model- and subsystem-specific bugs. A few pairs share broad themes (tokenizers, Flash Attention, cache/prefill, loading), but the concrete failure modes are different, so I do not see any true duplicate clusters here.", - "confidence": 0.29, - "canonical_issue_reason": "issue:43646 is the broadest umbrella regression in the list ('Transformers 5.0.0 breaks custom model initialization'), but the other issues are largely separate, model-specific bugs rather than duplicates of it.", + "summary": "This is not a true duplicate cluster: the items span many unrelated regressions, model-specific bugs, and import/load-path issues. None of the soft pairs look like the same underlying bug or change.", + "confidence": 0.95, + "canonical_issue_reason": "No single canonical duplicate exists; if one issue must anchor the cluster, #43646 is the broadest v5/custom-initialization regression.", "canonical_pr_reason": null, - "best_issue_reason": "issue:43646 is the most general/high-level issue and the closest thing to a cluster anchor, though it still does not actually subsume the rest of the set.", + "best_issue_reason": "#43646 is the best representative because it is the broadest umbrella for the few adjacent custom-model/config regressions, though the cluster is still heterogeneous.", "best_pr_reason": null, "soft_edge_verdicts": [ { "left": "issue:43295", "right": "issue:43825", "accept": false, - "reason": "Both are Transformers v4.57/v5 regressions, but one is about `processor.tokenizer`/image handling and the other is about an incorrect pipeline error message for translation tasks; different code paths." + "reason": "Different bugs: processor.tokenizer/image-input regression vs pipeline translation-task error messaging." }, { "left": "issue:43295", "right": "issue:44871", "accept": false, - "reason": "`processor.tokenizer` access regression and Gemma-3 eos token config mismatch are different bugs affecting different APIs/models." + "reason": "Custom processor/tokenizer regression and Gemma eos_token_id mismatch are unrelated model/config issues." }, { "left": "issue:44871", "right": "issue:45584", "accept": false, - "reason": "Both mention tokenizer/template behavior, but Gemma-3 eos_token_id inconsistency and Whisper empty transcription after `align_special_tokens` are distinct model-specific failures." + "reason": "Gemma eos_token_id inconsistency vs Whisper empty-transcription crash after align_special_tokens; different code paths." }, { "left": "issue:46234", "right": "issue:46291", "accept": false, - "reason": "Missing `is_falcon_mamba_ssm_available` in `import_utils` and a kernels version/revision import failure in `transformers.pipelines` are different dependency/import issues." + "reason": "Missing import helper vs kernels dependency import failure; both are import issues but not the same defect." }, { - "left": "issue:44077", - "right": "issue:44534", + "left": "issue:43316", + "right": "issue:43716", "accept": false, - "reason": "`patchtsmixer` post_init policy and non-persistent buffer junk in v5 are unrelated changes with different symptoms and fixes." + "reason": "Gemma3TextConfig API mismatch vs Mistral-3 dtype mismatch; different models and failure modes." }, { - "left": "issue:43316", - "right": "issue:43716", + "left": "issue:44077", + "right": "issue:44534", "accept": false, - "reason": "Gemma3TextConfig API inconsistency and Mistral-3 image-preprocessor dtype mismatch are different model-specific bugs." + "reason": "Optional post_init restriction and junk non-persistent buffers are unrelated transformer-v5 behaviors." }, { - "left": "issue:42175", - "right": "issue:43316", + "left": "issue:43316", + "right": "issue:46821", "accept": false, - "reason": "A packaging/backend issue with `pip install '.[torch]'` missing TensorFlow is unrelated to the Gemma3TextConfig API discrepancy." + "reason": "Different model families and bugs: config API discrepancy vs TimesFM window_size AttributeError." }, { "left": "issue:44464", "right": "issue:46421", "accept": false, - "reason": "Both touch chunked generation/cache behavior, but one reports inconsistent outputs under compiled forward while the other is about lazy cache init causing recompilation; not the same underlying bug." + "reason": "Chunked generation inconsistency with compiled forward is not the same as lazy cache initialization recompilation." }, { - "left": "issue:43646", - "right": "issue:44913", + "left": "issue:43452", + "right": "issue:44496", "accept": false, - "reason": "Generic custom model initialization breakage is different from GPTNeoXConfig `rotary_pct` not persisting on reload." + "reason": "Both touch model loading, but one is gguf_file/from_pretrained breakage and the other is unrecognized model config.json; not the same root cause." }, { - "left": "issue:43452", - "right": "issue:44496", + "left": "issue:43646", + "right": "issue:44913", "accept": false, - "reason": "`gguf_file` breaking `from_pretrained` and an unrecognized model missing `model_type` are both loading problems, but with different root causes and failure surfaces." + "reason": "Broad custom model initialization regression vs GPTNeoX rotary_pct reload persistence bug; different issues." }, { "left": "issue:44871", "right": "issue:46050", "accept": false, - "reason": "Gemma-3 eos token mismatch and quantized model save/load breakage are unrelated." + "reason": "Gemma eos_token_id mismatch and sinq quantized save/load corruption are unrelated." }, { "left": "issue:45803", "right": "issue:46365", "accept": false, - "reason": "Bare `except` in Fuyu tensor conversion and Trainer import failure due to datasets are completely different issues." + "reason": "Bare except in tensor conversion vs Trainer import failure due to datasets; unrelated problems." }, { "left": "issue:44162", "right": "issue:44792", "accept": false, - "reason": "ESM2 workflow breakage and a Janus image-generation test failure do not describe the same bug." + "reason": "ESM2 breakage and janus image-generation test failure are different models and symptoms." }, { "left": "issue:43404", "right": "issue:43441", "accept": false, - "reason": "Mistral3 weight tying and Ministral-3 FlashAttention failure are both Mistral-family issues, but they are different concrete code-path problems." + "reason": "lm_head weight tying bug vs FlashAttention-2 unsupported error; not the same code-path problem." }, { - "left": "issue:43065", - "right": "issue:44617", + "left": "issue:44242", + "right": "issue:46894", "accept": false, - "reason": "Sam3PixelDecoder dummy Conv2d and Sam3Video CUDA OOM are different Sam3 problems with different causes." + "reason": "Load-balancing loss omission and ASR double-shift bug are distinct training issues." }, { - "left": "issue:44242", - "right": "issue:46894", + "left": "issue:43065", + "right": "issue:44617", "accept": false, - "reason": "Missing load-balancing loss when `output_router_logits=False` and CohereASR double-shift training loss are separate loss-calculation bugs." + "reason": "Dummy Conv2d in Sam3PixelDecoder vs Sam3Video CUDA OOM; unrelated Sam3 problems." }, { "left": "issue:43540", "right": "issue:44514", "accept": false, - "reason": "Both are Qwen multimodal input bugs, but Qwen3OmniMoe video processing and Qwen2_5_VL batched chat-template padding behavior are different failure modes." + "reason": "Qwen3OmniMoe video-processing ValueError and Qwen2_5_VLProcessor batching crash are different processors and failures." }, { "left": "issue:44610", "right": "issue:46369", "accept": false, - "reason": "OmDet-Turbo processor image size mismatch and Sam3-LiteText memory overflow are unrelated model/processor problems." + "reason": "OmDet-Turbo input-size mismatch vs sam3-litetext memory overflow; not the same bug." }, { "left": "issue:45632", "right": "issue:46424", "accept": false, - "reason": "A cache-path collision for local models and `max_cache_length` only applying to quantized cache are different cache/config issues." + "reason": "Remote-code cache-path collision vs max_cache_length only honored for Quantized Cache; separate cache bugs." }, { "left": "issue:43572", "right": "issue:46178", "accept": false, - "reason": "StableLmConfig missing `pad_token_idx` and an invalid `tensorflow_text` backend mapping are unrelated bugs." + "reason": "Missing StableLmConfig pad_token_idx and tensorflow_text backend mapping error are unrelated." }, { "left": "issue:45542", "right": "issue:46234", "accept": false, - "reason": "Undefined tf backend when only tensorboard is installed and a missing import-utils helper for Falcon Mamba are different backend/import issues." + "reason": "TensorBoard-without-TF backend error and missing Falcon Mamba availability helper are different dependency issues." }, { "left": "issue:45803", "right": "issue:46170", "accept": false, - "reason": "Swallowing exceptions in Fuyu tensor conversion is unrelated to the Windows ROCm `torch.distributed` dependency problem." + "reason": "Swallowed KeyboardInterrupt in Fuyu conversion vs Windows ROCm torch.distributed import failure; unrelated." }, { "left": "issue:44792", "right": "issue:45600", "accept": false, - "reason": "A Janus generation test failure and stale auto-mappings for removed Sam3LiteText configs are different CI failures." + "reason": "Janus test failure vs stale auto_mappings reference to removed Sam3LiteText configs; different code paths." }, { "left": "issue:44792", "right": "issue:44977", "accept": false, - "reason": "Janus image generation failure and Qwen3.5 FlashAttention generation issues are model-specific and not duplicates." + "reason": "Janus generation test failure and Qwen3.5 flash-attention behavior are unrelated model-specific bugs." }, { "left": "issue:43504", "right": "issue:45698", "accept": false, - "reason": "BEiT legacy field loading and wrong custom-module loading after `save_pretrained`/`from_pretrained` both affect loading, but they are separate mechanisms." + "reason": "Legacy preset field in BEiT segmentation and wrong custom module after save_pretrained are both load-related but not the same defect." }, { "left": "issue:43650", "right": "issue:46202", "accept": false, - "reason": "These issues are entirely unrelated; one is effectively non-informative, the other is a Gemma pipeline roundtrip bug." + "reason": "Placeholder issue text and Gemma pipeline roundtrip failure are clearly unrelated." }, { "left": "issue:43352", "right": "issue:46640", "accept": false, - "reason": "Both mention Flash Attention 2, but one says NemotronHForCausalLM does not support it yet while the other is an unspecified flash-attention issue; not enough to treat as the same bug." + "reason": "Specific unsupported FlashAttention-2 error for NemotronH vs a generic open Flash Attention 2 issue; not the same concrete bug." }, { "left": "issue:45127", "right": "issue:45200", "accept": false, - "reason": "LoRA merge collapse with tied embeddings and Gemma 4 `mm_token_type_ids` defaulting are different model-behavior issues." + "reason": "LoRA merge collapse with tied embeddings and Gemma 4 mm_token_type_ids defaulting are different model behaviors." } ] }, @@ -11562,7 +11594,6 @@ "issue:27957", "issue:29870", "issue:35141", - "issue:41720", "issue:41950", "issue:42490", "issue:43335", @@ -11570,6 +11601,7 @@ "issue:43475", "issue:43525", "issue:43531", + "issue:43550", "issue:43644", "issue:43673", "issue:43688", @@ -11610,222 +11642,222 @@ "soft_pairs": [ "issue:43525|issue:43531", "issue:44534|issue:44913", - "issue:35141|issue:42490", - "issue:43673|issue:44743", "issue:41950|issue:43475", + "issue:43673|issue:44743", + "issue:35141|issue:42490", "issue:43388|issue:43688", "issue:44190|issue:44964", "issue:44855|issue:45588", "issue:29870|issue:35141", - "issue:45072|issue:46932", - "issue:45632|issue:46617", "issue:44869|issue:46832", + "issue:45632|issue:46617", + "issue:45072|issue:46932", + "issue:46082|issue:46697", "issue:45245|issue:46858", "issue:44734|issue:45406", - "issue:46082|issue:46697", "issue:43827|issue:43994", - "issue:44617|issue:45584", "issue:43644|issue:44913", + "issue:44617|issue:45584", "issue:25251|issue:27957", - "issue:43931|issue:44060", "issue:43335|issue:44855", + "issue:43931|issue:44060", "issue:44242|issue:46897", "issue:44077|issue:44877", "issue:43673|issue:46421", "issue:45381|issue:46369", "issue:45563|issue:45584", - "issue:41720|issue:45245" + "issue:43550|issue:44190" ], - "prepared_review_unit_hash": "f3a598c1bee416b04bf42e76e1291002caa0bf15bdb0eff1b186bef7ee862ade", + "prepared_review_unit_hash": "3fbf127c1021af26959296009c24d2565a289eabc2b529bbc65224a561840fd5", "component_budget": { "node_count": 514, "item_count": 514, "soft_pair_count": 1891, - "serialized_chars": 980926, - "estimated_input_tokens": 245232, - "estimated_eval_tokens": 490720 + "serialized_chars": 980895, + "estimated_input_tokens": 245224, + "estimated_eval_tokens": 490704 }, "budget": { "node_count": 48, "item_count": 48, "soft_pair_count": 27, - "serialized_chars": 36031, - "estimated_input_tokens": 9008, - "estimated_eval_tokens": 18272 + "serialized_chars": 36027, + "estimated_input_tokens": 9007, + "estimated_eval_tokens": 18270 }, "overflow_policy": "truncate_then_skip", "trimmed": false, "aggressively_trimmed": false, "split": true, "analyst_result": { - "summary": "This is a heterogeneous semantic cluster, not a duplicate set: the items span unrelated model, cache, config, pipeline, and documentation bugs. None of the soft pairs look like the same underlying issue, so all should stay separate.", - "confidence": 0.94, - "canonical_issue_reason": "No single canonical issue stands out; the cluster is too heterogeneous, and the soft pairs are only broad subsystem similarities rather than duplicates.", + "summary": "This set is heterogeneous: the items touch unrelated models, pipelines, cache paths, config behaviors, and decoding bugs. None of the soft pairs look like the same underlying issue/change; most only share a broad subsystem keyword.", + "confidence": 0.96, + "canonical_issue_reason": "No canonical issue emerges: the items cover many unrelated bugs across different models, pipelines, and subsystems.", "canonical_pr_reason": null, - "best_issue_reason": "issue:43644 is the broadest representative of the only cluster-wide regression theme here (non-persistent buffer state in v5) and is more central than the model-specific reports.", + "best_issue_reason": "No single issue is a better global representative because there is no coherent duplicate cluster here.", "best_pr_reason": null, "soft_edge_verdicts": [ { "left": "issue:43525", "right": "issue:43531", "accept": false, - "reason": "Different problems: missing pad_token_id vs sliding_window handling in Qwen3-MoE." + "reason": "Different models and failures: Llama4 pad_token_id attribute error vs Qwen3-MoE sliding_window behavior." }, { "left": "issue:44534", "right": "issue:44913", "accept": false, - "reason": "Both are config/runtime regressions, but one is buffer initialization junk and the other is rotary_pct not persisting on reload." + "reason": "Both are config/runtime regressions in Transformers v5, but they concern unrelated mechanics: non-persistent buffers vs GPTNeoX rotary_pct reload." }, { - "left": "issue:35141", - "right": "issue:42490", + "left": "issue:41950", + "right": "issue:43475", "accept": false, - "reason": "Unrelated: embedding reinit after resize vs import warning from trailing slash paths." + "reason": "A pipeline processor-selection bug is not the same as a Sam3VisionEncoderOutput missing-attribute error." }, { "left": "issue:43673", "right": "issue:44743", "accept": false, - "reason": "Both mention cache, but one is missing GenerationMixin cache and the other is recurrent state reset in modular_qwen3_5." + "reason": "Both involve cache handling, but one is missing GenerationMixin cache in chunked prefill while the other resets recurrent states in Qwen3 modular code." }, { - "left": "issue:41950", - "right": "issue:43475", + "left": "issue:35141", + "right": "issue:42490", "accept": false, - "reason": "Different subsystems: video-classification processor lookup vs missing fpn_position_embeddings on Sam3VisionEncoderOutput." + "reason": "Completely different areas: tied embeddings/post_init reinitialization vs local-path import warning." }, { "left": "issue:43388", "right": "issue:43688", "accept": false, - "reason": "Different bugs: gather_for_metrics drops tuple labels, while the other is aux-loss normalization." + "reason": "Different training/eval bugs: gather_for_metrics tuple truncation vs auxiliary loss normalization in MoE models." }, { "left": "issue:44190", "right": "issue:44964", "accept": false, - "reason": "Local dataset loading script failure vs model weight-shape mismatch on Phi-4 multimodal." + "reason": "Unrelated failures: a local dataset-loading script issue vs a multimodal model loading problem." }, { "left": "issue:44855", "right": "issue:45588", "accept": false, - "reason": "Python 3.13 JIT parsing/indentation issue vs flash_attention crash on s_aux=None." + "reason": "Python 3.13 decorator parsing/import error is unrelated to a flash_attention crash on s_aux=None." }, { "left": "issue:29870", "right": "issue:35141", "accept": false, - "reason": "Different code paths: Bart generate decoder mask bug vs output embedding reinit after resizing." + "reason": "Bart generate() decoder_attention_mask initialization bug is unrelated to embedding reinitialization in post_init." }, { - "left": "issue:45072", - "right": "issue:46932", + "left": "issue:44869", + "right": "issue:46832", "accept": false, - "reason": "Both are dtype-related, but one is CI inference mismatches and the other is Gemma4 quantized embedders casting." + "reason": "Whisper trailing-replacement-character decoding crash and RT-DETR feature-level mismatch are different code paths." }, { "left": "issue:45632", "right": "issue:46617", "accept": false, - "reason": "Both involve trust_remote_code cache loading, but one is leaf-path collision and the other is FileNotFoundError from cache." + "reason": "Both concern trust_remote_code/cache, but one is a leaf-directory cache collision and the other is a FileNotFoundError from cache loading." }, { - "left": "issue:44869", - "right": "issue:46832", + "left": "issue:45072", + "right": "issue:46932", + "accept": false, + "reason": "Different model families and causes: bfloat16 dtype mismatches in inference vs quantized Gemma4 vision/audio embedder casting." + }, + { + "left": "issue:46082", + "right": "issue:46697", "accept": false, - "reason": "Whisper decode crash vs RT-DETR forward TypeError; no shared underlying bug." + "reason": "LlamaConfig head_dim validation and Swin2SR over-padding are unrelated." }, { "left": "issue:45245", "right": "issue:46858", "accept": false, - "reason": "Category-count limit error vs cache.get_seq_length() device-tensor arithmetic; unrelated." + "reason": "Category-count overflow and cache.get_seq_length() returning a device tensor are unrelated bugs." }, { "left": "issue:44734", "right": "issue:45406", "accept": false, - "reason": "Serve KV-cache indexing crash vs Gemma4Processor missing _tokenizer." - }, - { - "left": "issue:46082", - "right": "issue:46697", - "accept": false, - "reason": "LlamaConfig validation vs Swin2SR over-padding; different layers and symptoms." + "reason": "KV-cache continuation indexing crash and Gemma4Processor missing _tokenizer are different serve failures." }, { "left": "issue:43827", "right": "issue:43994", "accept": false, - "reason": "Docs still referencing pipeline() vs bad outputs from AutoModel/pipeline for siglip2; not the same bug." + "reason": "Docs still referencing removed pipeline APIs is not the same as a model producing bad results through AutoModel/pipeline usage." }, { - "left": "issue:44617", - "right": "issue:45584", + "left": "issue:43644", + "right": "issue:44913", "accept": false, - "reason": "CUDA OOM in Sam3Video vs Whisper empty-transcription crash after align_special_tokens." + "reason": "Non-persistent buffer initialization junk and GPTNeoXConfig rotary_pct reload regression are distinct issues." }, { - "left": "issue:43644", - "right": "issue:44913", + "left": "issue:44617", + "right": "issue:45584", "accept": false, - "reason": "Related to loading/config state, but not the same issue: non-persistent buffer junk vs rotary_pct resetting on reload." + "reason": "CUDA OOM in Sam3Video is unrelated to Whisper empty-transcription generation failure." }, { "left": "issue:25251", "right": "issue:27957", "accept": false, - "reason": "Pipeline top_k output shape bug vs XLMRoberta Flash Attention 2 support issue." + "reason": "A pipeline output-shape change and XLMRoberta Flash Attention 2 support are unrelated." }, { - "left": "issue:43931", - "right": "issue:44060", + "left": "issue:43335", + "right": "issue:44855", "accept": false, - "reason": "Model weight shape mismatch vs incorrect tied-weights warning in Qwen3-Next." + "reason": "SwitchTransformers sparse-layer construction and DebertaV2 import parsing are unrelated." }, { - "left": "issue:43335", - "right": "issue:44855", + "left": "issue:43931", + "right": "issue:44060", "accept": false, - "reason": "Sparse-layer creation bug vs Python 3.13 JIT indentation error." + "reason": "Both mention Qwen models, but weight-shape loading errors and tied-weights warning misbinding are different bugs." }, { "left": "issue:44242", "right": "issue:46897", "accept": false, - "reason": "Load-balancing loss omission vs Florence2 training-loss double-shift; both are loss bugs but distinct." + "reason": "MoE load-balancing loss gating and Florence2 training-loss double-shift are not the same bug." }, { "left": "issue:44077", "right": "issue:44877", "accept": false, - "reason": "patchtsmixer post_init policy vs strict-config loading failure for granite_speech." + "reason": "A model-specific post_init policy issue is unrelated to strict config blocking granite_speech loading." }, { "left": "issue:43673", "right": "issue:46421", "accept": false, - "reason": "Both involve chunked prefill/cache, but one is missing cache in v5 and the other is lazy cache init causing recompilation." + "reason": "Both involve chunked prefill/cache, but one is missing cache state and the other is spurious recompilation from lazy cache init." }, { "left": "issue:45381", "right": "issue:46369", "accept": false, - "reason": "Qwen2.5-VL video position ids bug vs Sam3-litetext memory overflow." + "reason": "Qwen2.5-VL video position IDs and Sam3-litetext memory overflow are unrelated." }, { "left": "issue:45563", "right": "issue:45584", "accept": false, - "reason": "Paged generate warning regression vs Whisper empty transcription crash." + "reason": "Paged-generate stale warning and Whisper empty-transcription failure are different generation bugs." }, { - "left": "issue:41720", - "right": "issue:45245", + "left": "issue:43550", + "right": "issue:44190", "accept": false, - "reason": "Qwen3 auto device mapping cuda assert vs category-count limit in another model path; unrelated." + "reason": "Bamba torch.compile+SDPA failure is unrelated to the local dataset-loading script issue." } ] }, @@ -11852,14 +11884,13 @@ "issue:25251", "issue:29870", "issue:39692", + "issue:41720", "issue:43065", "issue:43295", "issue:43335", "issue:43526", "issue:43540", - "issue:43550", "issue:43643", - "issue:43673", "issue:43854", "issue:44112", "issue:44190", @@ -11879,6 +11910,7 @@ "issue:44964", "issue:45003", "issue:45127", + "issue:45245", "issue:45276", "issue:45290", "issue:45357", @@ -11895,234 +11927,226 @@ "issue:46655", "issue:46697", "issue:46752", - "issue:46858", "issue:47013" ], "soft_pairs": [ - "issue:43550|issue:44190", + "issue:41720|issue:45245", "issue:44933|issue:45362", "issue:43065|issue:44610", "issue:43335|issue:44190", "issue:45588|issue:47013", "issue:44479|issue:44734", - "issue:44617|issue:46369", - "issue:45381|issue:46655", "issue:43854|issue:44792", + "issue:45381|issue:46655", + "issue:44617|issue:46369", "issue:39692|issue:44190", "issue:45127|issue:45357", "issue:43295|issue:45276", "issue:44464|issue:45563", "issue:44779|issue:46489", "issue:45276|issue:45584", - "issue:43540|issue:44734", "issue:45290|issue:46752", + "issue:43540|issue:44734", "issue:44112|issue:45600", - "issue:44521|issue:45381", "issue:25251|issue:29870", + "issue:44521|issue:45381", "issue:44462|issue:45907", "issue:43643|issue:44496", "issue:43526|issue:46697", "issue:44938|issue:45542", "issue:43854|issue:44964", - "issue:44246|issue:45003", - "issue:43673|issue:46858" + "issue:44246|issue:45003" ], - "prepared_review_unit_hash": "b8ba91dccb2e2ad672923e62fda5f19bf5f4be1a815ee03278876011cd9f87b2", + "prepared_review_unit_hash": "b9e35787215526ca67c2fc01e77a0e3e09fd9482a1acec614a47efe659504f83", "component_budget": { "node_count": 514, "item_count": 514, "soft_pair_count": 1891, - "serialized_chars": 980926, - "estimated_input_tokens": 245232, - "estimated_eval_tokens": 490720 + "serialized_chars": 980895, + "estimated_input_tokens": 245224, + "estimated_eval_tokens": 490704 }, "budget": { - "node_count": 48, - "item_count": 48, - "soft_pair_count": 27, - "serialized_chars": 35942, - "estimated_input_tokens": 8986, - "estimated_eval_tokens": 18228 + "node_count": 47, + "item_count": 47, + "soft_pair_count": 26, + "serialized_chars": 35009, + "estimated_input_tokens": 8753, + "estimated_eval_tokens": 17762 }, "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 bug reports. The only soft pair that looks like the same underlying issue is the DeepSeek tokenizer regression pair (44779/46489); the rest differ by model, subsystem, or failure mode.", - "confidence": 0.84, - "canonical_issue_reason": "issue:46489 is the most specific of the only true duplicate pair: it clearly describes the DeepSeek-Coder v1 tokenizer regression in v5+ and ties it to a gap in a prior fix, making it the best canonical issue.", + "summary": "Most pairs are clearly unrelated despite some superficial overlap in subsystems. The only strong duplicate-like pair is the DeepSeek tokenizer regression reports, which describe the same v5+ tokenizer breakage.", + "confidence": 0.89, + "canonical_issue_reason": "issue:46489 is the best canonical anchor because it is the most specific DeepSeek tokenizer regression report, explicitly ties the problem to v5+, and notes it fills a gap in PR #44801\u2019s fix.", "canonical_pr_reason": null, - "best_issue_reason": "issue:46489 is the strongest representative issue here because it pinpoints the concrete tokenizer regression and its relation to the earlier fix, which makes it the clearest anchor for triage.", + "best_issue_reason": "issue:46489 is the strongest issue to represent the cluster\u2019s one real duplicate pair: concrete failure, clear affected scope, and direct linkage to the prior fix.", "best_pr_reason": null, "soft_edge_verdicts": [ { - "left": "issue:43550", - "right": "issue:44190", + "left": "issue:41720", + "right": "issue:45245", "accept": false, - "reason": "Different models and failures: torch.compile/SDPA vs local dataset loading." + "reason": "Different failures and subsystems: Qwen3 device-mapping CUDA assert vs category-count overflow." }, { "left": "issue:44933", "right": "issue:45362", "accept": false, - "reason": "Unrelated subsystems: missing image_utils import vs Qwen chat crash." + "reason": "Unrelated: missing import in image_utils vs Qwen3.5 chat crash." }, { "left": "issue:43065", "right": "issue:44610", "accept": false, - "reason": "Both are SAM/vision issues, but one is a dummy Conv2d and the other is a processor/model image-size mismatch." + "reason": "Both involve vision models, but one is a dummy Conv2d in Sam3PixelDecoder and the other is an OmDet-Turbo processor/model image-size mismatch." }, { "left": "issue:43335", "right": "issue:44190", "accept": false, - "reason": "Completely unrelated: SwitchTransformers config bug vs dataset loading script issue." + "reason": "Different code paths: SwitchTransformers config layer construction vs local dataset loading in an example script." }, { "left": "issue:45588", "right": "issue:47013", "accept": false, - "reason": "Different code paths and failure modes: flash attention sink handling vs xLSTM unpacking in train mode." + "reason": "Different bugs: flash-attention sinkless-model s_aux handling vs xLSTM train-mode unpacking." }, { "left": "issue:44479", "right": "issue:44734", "accept": false, - "reason": "Different regressions: video input handling vs serve KV-cache continuation indexing." + "reason": "Both are generation bugs, but one is a video-input regression in Qwen VL variants and the other is a serve KV-cache indexing crash." }, { - "left": "issue:44617", - "right": "issue:46369", + "left": "issue:43854", + "right": "issue:44792", "accept": false, - "reason": "Both mention memory pressure in SAM3, but one is video OOM and the other is sam3-litetext memory overflow; not clearly the same bug." + "reason": "Different models and failures: GLM-4.7 load/test issue vs Janus image-generation test failure." }, { "left": "issue:45381", "right": "issue:46655", "accept": false, - "reason": "Both are video-processing bugs, but they affect different models and likely different preprocessing logic." + "reason": "Both are video-processing issues, but the reported symptoms and models differ (vision_position_ids vs padding calculation)." }, { - "left": "issue:43854", - "right": "issue:44792", + "left": "issue:44617", + "right": "issue:46369", "accept": false, - "reason": "Different model-loading/test failures for different models." + "reason": "Both mention memory pressure in Sam3 variants, but one is Sam3Video CUDA OOM and the other is Sam3-LiteText memory overflow; not clearly the same bug." }, { "left": "issue:39692", "right": "issue:44190", "accept": false, - "reason": "Documentation example errors vs local dataset loading; unrelated." + "reason": "Doc-example model/processor mismatch and quantization problems are unrelated to local dataset loading in an example script." }, { "left": "issue:45127", "right": "issue:45357", "accept": false, - "reason": "Different mechanisms: LoRA merge/vocab collapse vs incorrect saved visual encoder keys." + "reason": "Different issues: LoRA merge with extended vocab causing collapse vs incorrect visual encoder keys saved by Qwen3.5." }, { "left": "issue:43295", "right": "issue:45276", "accept": false, - "reason": "Different bugs: processor/tokenizer regression vs resize_token_embeddings not updating outputs." + "reason": "Processor/tokenizer API regression is not the same as gemma4 resize_token_embeddings not propagating to embeddings." }, { - "left": "issue:44464", + "left": "issue:44462", "right": "issue:45563", "accept": false, - "reason": "Different generation issues: compiled forward inconsistency vs a stale warning in paged generate()." + "reason": "Tokenizer-file precedence bug is unrelated to a stale generate warning about num_return_sequences." }, { "left": "issue:44779", "right": "issue:46489", "accept": true, - "reason": "Same underlying DeepSeek tokenizer regression in v5+; 46489 is a more specific instance tied to the earlier fix gap." + "reason": "Same underlying bug: DeepSeek tokenizer output is wrong on v5/v5+ and works on v4; 46489 is a more specific report of the same regression with an explicit gap in PR #44801\u2019s fix." }, { "left": "issue:45276", "right": "issue:45584", "accept": false, - "reason": "Unrelated: embedding resize bug vs Whisper empty-transcription generation failure." + "reason": "Resize-token-embedding propagation bug is unrelated to Whisper empty-transcription handling after align_special_tokens." }, { - "left": "issue:43540", - "right": "issue:44734", + "left": "issue:45290", + "right": "issue:46752", "accept": false, - "reason": "Different subsystems: multimodal video input parsing vs response KV-cache continuation indexing." + "reason": "Both hit apply_chat_template edge cases, but one is assistant tool-call/no-content handling and the other is an empty-list crash; too different to merge." }, { - "left": "issue:45290", - "right": "issue:46752", + "left": "issue:43540", + "right": "issue:44734", "accept": false, - "reason": "Both are apply_chat_template crashes, but one is tool-call/no-content handling and the other is empty-list input; different edge cases." + "reason": "Different failures: Qwen3OmniMoe video input ValueError vs serve KV-cache continuation tensor indexing." }, { "left": "issue:44112", "right": "issue:45600", "accept": false, - "reason": "CI failures in different areas: GraniteSpeech device override test vs stale auto_mappings entries." + "reason": "Both are CI-related, but the stale GraniteSpeech device-override test and removed Sam3LiteText auto-mapping configs are different problems." }, { - "left": "issue:44521", - "right": "issue:45381", + "left": "issue:25251", + "right": "issue:29870", "accept": false, - "reason": "Both concern multimodal chat/video, but assistant_masks and vision_position_ids are different code paths and bugs." + "reason": "Unrelated: pipeline top_k nesting change vs Bart generate decoder_attention_mask assignment error." }, { - "left": "issue:25251", - "right": "issue:29870", + "left": "issue:44521", + "right": "issue:45381", "accept": false, - "reason": "Unrelated bugs: pipeline top_k nesting vs Bart generate decoder_attention_mask initialization." + "reason": "Both involve multimodal chat/video inputs, but one is assistant mask generation and the other is video position IDs." }, { "left": "issue:44462", "right": "issue:45907", "accept": false, - "reason": "Tokenizer loading behavior vs repo file listing revision forwarding; not the same issue." + "reason": "Tokenizer resolution bug is unrelated to list_repo_files not forwarding revision in a cookbook CI path." }, { "left": "issue:43643", "right": "issue:44496", "accept": false, - "reason": "Different config-loading problems: missing fields with trust_remote_code vs unrecognized model_type." + "reason": "Both are config-loading failures, but missing fields from trust_remote_code and missing model_type are not the same bug." }, { "left": "issue:43526", "right": "issue:46697", "accept": false, - "reason": "Both are image-processor bugs, but one is label reduction and the other is over-padding." + "reason": "Different image-processor bugs: reduce_labels truncation vs over-padding when dimensions are already divisible." }, { "left": "issue:44938", "right": "issue:45542", "accept": false, - "reason": "Python 3.14 import failure is unrelated to tensorboard-without-TF backend behavior." + "reason": "Unrelated import/load failures: Python 3.14 compatibility vs tensorboard-only undefined tf backend." }, { "left": "issue:43854", "right": "issue:44964", "accept": false, - "reason": "Different model loading failures for different models." + "reason": "Both are model-loading issues, but they affect different models and failure modes." }, { "left": "issue:44246", "right": "issue:45003", "accept": false, - "reason": "Import slowness vs unsafe sys.modules access; related only loosely, not the same bug." - }, - { - "left": "issue:43673", - "right": "issue:46858", - "accept": false, - "reason": "Both involve generation cache, but 43673 is missing cache in chunked_prefill while 46858 is a static-cache slice arithmetic bug on non-CUDA backends." + "reason": "Import slowness and unsafe sys.modules access are not the same concrete bug." } ] }, "evaluator_result": { "accept": true, - "feedback": "Grounded overall. The single accepted soft pair (44779/46489) is the only clearly supported duplicate from the titles, and the rest of the false verdicts are conservative rather than overreaching." + "feedback": "Grounded overall. The only clearly duplicate-like pair in the packet is issue:44779 \u2194 issue:46489, and the other soft-edge rejections stay conservative by distinguishing different models, code paths, or failure modes. The canonical anchor choice is also well supported by the packet." }, "evaluator_used": true, "retried": false, @@ -12154,7 +12178,6 @@ "issue:42796", "issue:42907", "issue:43295", - "issue:43381", "issue:43404", "issue:43441", "issue:43526", @@ -12180,7 +12203,6 @@ "issue:45216", "issue:45335", "issue:45446", - "issue:45593", "issue:45600", "issue:45632", "issue:45759", @@ -12190,9 +12212,11 @@ "issue:46234", "issue:46424", "issue:46506", + "issue:46858", "issue:46932" ], "soft_pairs": [ + "issue:43673|issue:46858", "issue:44737|issue:45803", "issue:27957|issue:35141", "issue:44964|issue:45600", @@ -12210,221 +12234,220 @@ "issue:41950|issue:44016", "issue:43673|issue:46424", "issue:43295|issue:45335", - "issue:43582|issue:46932", "issue:45083|issue:46506", "issue:44496|issue:45042", "issue:43976|issue:44987", + "issue:43582|issue:46932", "issue:43526|issue:43901", - "issue:43901|issue:44496", "issue:43673|issue:44336", - "issue:44496|issue:45759", + "issue:43901|issue:44496", "issue:44683|issue:45446", + "issue:44496|issue:45759", "issue:44016|issue:44220", - "issue:43381|issue:45593", "issue:42907|issue:43441" ], - "prepared_review_unit_hash": "d8b4417316e5408c01d313f048bc5d464f2788fb7f65e173c9927a6d4e34071d", + "prepared_review_unit_hash": "04924de299411dc032dc9040dc1cccec6dad6766b6d8acf1958880b34870142a", "component_budget": { "node_count": 514, "item_count": 514, "soft_pair_count": 1891, - "serialized_chars": 980926, - "estimated_input_tokens": 245232, - "estimated_eval_tokens": 490720 + "serialized_chars": 980895, + "estimated_input_tokens": 245224, + "estimated_eval_tokens": 490704 }, "budget": { - "node_count": 48, - "item_count": 48, + "node_count": 47, + "item_count": 47, "soft_pair_count": 29, - "serialized_chars": 36526, - "estimated_input_tokens": 9132, - "estimated_eval_tokens": 18520 + "serialized_chars": 36050, + "estimated_input_tokens": 9013, + "estimated_eval_tokens": 18282 }, "overflow_policy": "truncate_then_skip", "trimmed": false, "aggressively_trimmed": false, "split": true, "analyst_result": { - "summary": "These look like a loose collection of unrelated Transformers issues rather than a duplicate cluster. None of the soft pairs appear to be the same underlying bug or change.", - "confidence": 0.86, - "canonical_issue_reason": "issue:44521 is the most active and referenced item in the set, and it is a broad user-facing multimodal regression, so it is the best single issue anchor here.", + "summary": "These issues span many unrelated Transformers bugs and regressions; none of the proposed soft pairs looks like the same underlying defect, so all candidate edges should be rejected.", + "confidence": 0.94, + "canonical_issue_reason": null, "canonical_pr_reason": null, - "best_issue_reason": "issue:44521 has the strongest signal for reuse as a representative issue: it is open, has inbound references, and describes a concrete end-user bug.", + "best_issue_reason": "If a representative issue is needed, #44521 is the strongest cluster anchor: it is open, actively discussed, has inbound references, and concerns a broadly impactful multimodal API bug.", "best_pr_reason": null, "soft_edge_verdicts": [ + { + "left": "issue:43673", + "right": "issue:46858", + "accept": false, + "reason": "Both involve cache behavior, but one is about a missing cache in chunked prefill and the other is about a device-tensor seq_length breaking arithmetic in static cache; different bugs." + }, { "left": "issue:44737", "right": "issue:45803", "accept": false, - "reason": "XLNet CPU device-placement bug vs Fuyu exception-swallowing bug; different code paths and fixes." + "reason": "XLNet CPU placement bug and Fuyu exception-swallowing are unrelated code paths and failure modes." }, { "left": "issue:27957", "right": "issue:35141", "accept": false, - "reason": "FlashAttention 2 support for XLM-Roberta is unrelated to embedding reinitialization after resize." + "reason": "Flash Attention 2 support for XLM-Roberta is unrelated to token embedding reinitialization when tie_word_embedding is false." }, { "left": "issue:44964", "right": "issue:45600", "accept": false, - "reason": "Phi-4 multimodal load failure vs stale auto_mappings references/CI breakage; not the same bug." + "reason": "Phi-4 multimodal loading failure and stale auto_mappings references are different issues in different loader paths." }, { "left": "issue:44112", "right": "issue:44964", "accept": false, - "reason": "GraniteSpeech CI stale device override test is unrelated to loading Phi-4 multimodal models." + "reason": "A GraniteSpeech CI test failure is not the same as a Phi-4 multimodal model loading regression." }, { "left": "issue:42175", "right": "issue:44279", "accept": false, - "reason": "Missing TensorFlow backend after torch-only install is not the same as a generic dependency issue." + "reason": "Package extras/backend dependency complaints are too broad here; these reports do not describe the same concrete dependency bug." }, { "left": "issue:45632", "right": "issue:46234", "accept": false, - "reason": "Local trust_remote_code cache collision and missing import_utils symbol are different problems." + "reason": "A local trust_remote_code cache-path collision is unrelated to a missing import_utils availability helper." }, { "left": "issue:42796", "right": "issue:44521", "accept": false, - "reason": "Tokenizer BPE-marker handling is unrelated to all-zero assistant masks in multimodal chat templates." + "reason": "Tokenizer BPE-marker handling and multimodal assistant-mask generation are different tokenizer behaviors." }, { "left": "issue:32090", "right": "issue:35141", "accept": false, - "reason": "Trainer GPU broadcast NoneType error is unrelated to post-init embedding reinitialization." + "reason": "Trainer GPU broadcast NoneType error and embedding resize/post_init reinitialization are unrelated." }, { "left": "issue:36246", "right": "issue:41628", "accept": false, - "reason": "Missing Qwen2_5_VLImageProcessor import and missing AutoImageProcessor import are different symbols/code paths." + "reason": "Both are import errors, but for different symbols and different package surfaces; they are not the same missing-export bug." }, { "left": "issue:36246", "right": "issue:38617", "accept": false, - "reason": "Both are import errors, but they involve different missing names and different modules." + "reason": "Different missing imports from different modules; no shared underlying fix." }, { "left": "issue:43404", "right": "issue:45216", "accept": false, - "reason": "Mistral3 lm_head tying bug and Qwen3.5 save_pretrained regression are distinct model/config issues." + "reason": "Mistral3 lm_head tying is unrelated to the Qwen3.5 save_pretrained checkpoint regression." }, { "left": "issue:43531", "right": "issue:45600", "accept": false, - "reason": "Qwen3-MoE sliding_window behavior is unrelated to stale auto_mappings config references." + "reason": "Qwen3-MoE sliding-window behavior and stale auto_mappings config references are separate bugs." }, { "left": "issue:44295", "right": "issue:46506", "accept": false, - "reason": "position_ids buffer access error vs CPU/CUDA output divergence are different failures." + "reason": "Reading a registered buffer incorrectly and CPU/CUDA output divergence are not the same defect." }, { "left": "issue:46082", "right": "issue:46121", "accept": false, - "reason": "LlamaConfig head_dim validation and rope-parameter list handling are separate validation bugs." + "reason": "Llama head_dim validation and rope-parameter JSON-list handling are distinct config-validation problems." }, { "left": "issue:41950", "right": "issue:44016", "accept": false, - "reason": "Pipeline image-processor lookup bug is unrelated to a notebook syntax error." + "reason": "Video-classification processor lookup and a notebook syntax error are unrelated." }, { "left": "issue:43673", "right": "issue:46424", "accept": false, - "reason": "GenerationMixin cache missing and cache max_length semantics are related to caching but not the same concrete bug." + "reason": "Generation cache missing in chunked prefill and max_cache_length being honored only for quantized cache are different cache semantics." }, { "left": "issue:43295", "right": "issue:45335", "accept": false, - "reason": "Processor.tokenizer regression with images is unrelated to t5gemma resize_token_embeddings not affecting decoder.embed_tokens." - }, - { - "left": "issue:43582", - "right": "issue:46932", - "accept": false, - "reason": "Apple Silicon warmup TypeError and Gemma4 dtype-casting fix are unrelated." + "reason": "Processor.tokenizer regression and resize_token_embeddings not updating decoder.embed_tokens are separate model/processor bugs." }, { "left": "issue:45083", "right": "issue:46506", "accept": false, - "reason": "Helper output-length computation bug is unrelated to CPU/CUDA result differences." + "reason": "A helper returning unexpected lengths in qwen3_omni_moe is unrelated to PPDocLayoutV3 CPU/CUDA divergence." }, { "left": "issue:44496", "right": "issue:45042", "accept": false, - "reason": "Unrecognized model config error is not the same as PIL image processors incorrectly requiring torchvision." + "reason": "Unrecognized-model config loading and PIL backend requiring torchvision are different failure modes." }, { "left": "issue:43976", "right": "issue:44987", "accept": false, - "reason": "Python version support regression and model-load failure are different issues." + "reason": "Python version compatibility and loading a specific model checkpoint are unrelated." }, { - "left": "issue:43526", - "right": "issue:43901", + "left": "issue:43582", + "right": "issue:46932", "accept": false, - "reason": "BeitImageProcessorFast label reduction bug is unrelated to a docs mismatch about return_all_scores." + "reason": "Apple Silicon warmup TypeError and Gemma4 quantized-embedder dtype casting are separate issues." }, { - "left": "issue:43901", - "right": "issue:44496", + "left": "issue:43526", + "right": "issue:43901", "accept": false, - "reason": "Documentation wording issue vs model config import failure; different domains entirely." + "reason": "BeitImageProcessorFast label reduction and TextClassificationPipeline docs mismatch are not the same bug." }, { "left": "issue:43673", "right": "issue:44336", "accept": false, - "reason": "Cache-missing regression and ANSI escape handling in loading_report are unrelated." + "reason": "Generation cache behavior and ANSI codes emitted by loading_report are unrelated." }, { - "left": "issue:44496", - "right": "issue:45759", + "left": "issue:43901", + "right": "issue:44496", "accept": false, - "reason": "Unrecognized model config error is unrelated to AutoModelForCausalLM.from_config not unwrapping text_config." + "reason": "A docs/behavior mismatch for return_all_scores is unrelated to an unrecognized-model config error." }, { "left": "issue:44683", "right": "issue:45446", "accept": false, - "reason": "Both touch flex_attention, but one is a torch>=2.9 compiled-failure and the other is an AuxRequest version-check/import bug." + "reason": "Both touch flex_attention, but one is a compile failure on torch>=2.9 and the other is an incorrect version check for AuxRequest import; different concrete problems." }, { - "left": "issue:44016", - "right": "issue:44220", + "left": "issue:44496", + "right": "issue:45759", "accept": false, - "reason": "Notebook syntax error and _torch_extract_fbank_features bug are unrelated." + "reason": "Unrecognized model config and from_config failing to unwrap text_config are different model-loading issues." }, { - "left": "issue:43381", - "right": "issue:45593", + "left": "issue:44016", + "right": "issue:44220", "accept": false, - "reason": "Gradient checkpointing eval-mode restriction and missing auxiliary losses when denoising is off are different behaviors." + "reason": "A notebook syntax error and a feature-extraction helper issue are unrelated." }, { "left": "issue:42907", "right": "issue:43441", "accept": false, - "reason": "Dequantized save failure and FlashAttention failure are separate Ministral issues with different fixes." + "reason": "Saving dequantized Ministral/Devstral models and FlashAttention failures in Ministral are separate problems." } ] }, @@ -12479,11 +12502,11 @@ "issue:45276", "issue:45278", "issue:45335", - "issue:45357", "issue:45375", "issue:45405", "issue:45584", "issue:45588", + "issue:45593", "issue:45812", "issue:46082", "issue:46123", @@ -12498,19 +12521,20 @@ "issue:47085" ], "soft_pairs": [ + "issue:43381|issue:45593", "issue:43650|issue:44220", "issue:44016|issue:44062", - "issue:43976|issue:44279", "issue:44792|issue:46424", + "issue:43976|issue:44279", "issue:43441|issue:45588", - "issue:44521|issue:46693", "issue:43824|issue:44933", + "issue:44521|issue:46693", "issue:44279|issue:46515", "issue:43479|issue:45375", "issue:45335|issue:45584", "issue:43299|issue:45216", - "issue:44327|issue:46752", "issue:46123|issue:46369", + "issue:44327|issue:46752", "issue:46614|issue:47085", "issue:44279|issue:45405", "issue:43295|issue:44779", @@ -12522,207 +12546,206 @@ "issue:41950|issue:45020", "issue:43976|issue:45405", "issue:44779|issue:45812", - "issue:43381|issue:47000", "issue:46082|issue:46399", - "issue:44466|issue:45276", - "issue:43299|issue:45357" + "issue:43381|issue:47000", + "issue:44466|issue:45276" ], - "prepared_review_unit_hash": "43625d2df4f56ed34f09dd13bb28def182507b66e5bfc2cb8c82029b9dab6229", + "prepared_review_unit_hash": "a4f346ee05595b957b3282a43119b39f96547ba812b68d12c137a1afc22ee581", "component_budget": { "node_count": 514, "item_count": 514, "soft_pair_count": 1891, - "serialized_chars": 980926, - "estimated_input_tokens": 245232, - "estimated_eval_tokens": 490720 + "serialized_chars": 980895, + "estimated_input_tokens": 245224, + "estimated_eval_tokens": 490704 }, "budget": { "node_count": 48, "item_count": 48, "soft_pair_count": 28, - "serialized_chars": 36069, - "estimated_input_tokens": 9018, - "estimated_eval_tokens": 18292 + "serialized_chars": 36041, + "estimated_input_tokens": 9011, + "estimated_eval_tokens": 18278 }, "overflow_policy": "truncate_then_skip", "trimmed": false, "aggressively_trimmed": false, "split": true, "analyst_result": { - "summary": "The set is not a true duplicate cluster: the pairs span unrelated regressions, docs bugs, tokenizer issues, and import errors. None of the soft edges look like the same underlying bug/change, so all should be rejected.", - "confidence": 0.84, - "canonical_issue_reason": "issue:44521 is the strongest representative issue here: it is open, has multiple inbound references, and describes a concrete, reproducible multimodal API bug with broad user impact.", + "summary": "Mostly unrelated issue reports across different Transformer subsystems. The cluster is dominated by v5 regressions, but the soft pairs are generally distinct enough that none should be merged as duplicates.", + "confidence": 0.43, + "canonical_issue_reason": "Issue 45278 is the broadest umbrella-style report in the set, covering widespread breakage after a version upgrade; it is more suitable as the cluster anchor than the narrower model- or tokenizer-specific bugs.", "canonical_pr_reason": null, - "best_issue_reason": "issue:46693 is the best issue for triage priority overall because it is open, has substantial discussion, and reports a high-impact performance regression.", + "best_issue_reason": "Issue 45278 best represents the overall cluster because it summarizes the common theme of post-upgrade regressions, while the others are mostly specific manifestations in single models, tokenizers, or pipelines.", "best_pr_reason": null, "soft_edge_verdicts": [ + { + "left": "issue:43381", + "right": "issue:45593", + "accept": false, + "reason": "Different bugs: eval-mode gradient checkpointing behavior vs D-FINE auxiliary-loss handling." + }, { "left": "issue:43650", "right": "issue:44220", "accept": false, - "reason": "Completely different problems: a vague data request versus a specific fbank feature extraction bug." + "reason": "No clear overlap; one is a vague data request, the other is an fbank feature extraction issue." }, { "left": "issue:44016", "right": "issue:44062", "accept": false, - "reason": "Notebook syntax error and AddedToken keyword collision are unrelated failure modes." + "reason": "Unrelated failures: notebook syntax error vs AddedToken keyword argument conflict." }, { - "left": "issue:43976", - "right": "issue:44279", + "left": "issue:44792", + "right": "issue:46424", "accept": false, - "reason": "Different topics: Python-version support breakage versus a generic dependency issue." + "reason": "Different code paths: janus test failure vs cache max length handling." }, { - "left": "issue:44792", - "right": "issue:46424", + "left": "issue:43976", + "right": "issue:44279", "accept": false, - "reason": "Janus test failure is unrelated to cache max length behavior in quantized cache." + "reason": "One is a Python version compatibility bug; the other is a generic dependency issue." }, { "left": "issue:43441", "right": "issue:45588", "accept": false, - "reason": "Both mention flash attention, but they affect different models and different code paths; not the same bug." + "reason": "Both mention flash attention, but they are different model-specific crashes with different root causes." }, { - "left": "issue:44521", - "right": "issue:46693", + "left": "issue:43824", + "right": "issue:44933", "accept": false, - "reason": "Multimodal chat-template masks and FA2 long-context performance regression are different issues." + "reason": "Different import problems for different modules; not the same missing symbol." }, { - "left": "issue:43824", - "right": "issue:44933", + "left": "issue:44521", + "right": "issue:46693", "accept": false, - "reason": "Different import failures involving different modules and symbols." + "reason": "Multimodal chat-template masking bug is unrelated to the FA2 long-context performance regression." }, { "left": "issue:44279", "right": "issue:46515", "accept": false, - "reason": "Generic dependency issue and torch 2.4.1 incompatibility are not the same underlying defect." + "reason": "Both are dependency-related, but they concern different external packages and failures." }, { "left": "issue:43479", "right": "issue:45375", "accept": false, - "reason": "Different config bugs: default vision/audio config initialization versus a missing vision-index field being dropped by strict parsing." + "reason": "Different config classes and different missing/defaulted fields." }, { "left": "issue:45335", "right": "issue:45584", "accept": false, - "reason": "Token embedding resize propagation and Whisper empty-transcription generation failure are unrelated." + "reason": "Resize-token-embeddings propagation bug is unrelated to Whisper empty-transcription generation failure." }, { "left": "issue:43299", "right": "issue:45216", "accept": false, - "reason": "Both are regressions, but one breaks Qwen3VL model loading and the other breaks Qwen3.5 checkpoint saving." + "reason": "Both are Qwen-related regressions, but one is loading failure and the other is checkpoint saving behavior." }, { - "left": "issue:44327", - "right": "issue:46752", + "left": "issue:46123", + "right": "issue:46369", "accept": false, - "reason": "QA decode_spans top-k crash and empty apply_chat_template crash are different code paths and failure conditions." + "reason": "Pipeline batch-tracking bug and SAM memory overflow are unrelated." }, { - "left": "issue:46123", - "right": "issue:46369", + "left": "issue:44327", + "right": "issue:46752", "accept": false, - "reason": "Partial-batch result dropping and SAM3 LiteText memory overflow are not the same bug." + "reason": "Different crashes in different APIs: QA span decoding vs empty chat-template input." }, { "left": "issue:46614", "right": "issue:47085", "accept": false, - "reason": "Phonemizer language state bug and tokenizer thread-safety regression are unrelated." + "reason": "Tokenizer backend language-state bug is unrelated to tokenizer thread-safety regression." }, { "left": "issue:44279", "right": "issue:45405", "accept": false, - "reason": "Dependency trouble and an unreleased PEFT version bump are different problems." + "reason": "Generic dependency issue vs a specific PEFT minimum-version bump; not the same bug." }, { "left": "issue:43295", "right": "issue:44779", "accept": false, - "reason": "Processor/tokenizer regression with images and Deepseek tokenizer output corruption are distinct tokenizer issues, not one fix." + "reason": "Both are tokenizer regressions, but they affect different APIs and models with different failure modes." }, { "left": "issue:39692", "right": "issue:43901", "accept": false, - "reason": "SigLIP2 docs quantization/model mismatch and TextClassificationPipeline return_all_scores docs mismatch are unrelated documentation problems." + "reason": "Both are documentation-related, but they describe different examples and different behaviors." }, { "left": "issue:42796", "right": "issue:43650", "accept": false, - "reason": "Tokenizer BPE-marker handling and a vague 'ADD THE DATA' request are not duplicates." + "reason": "Tokenizer BPE-marker bug is unrelated to the data request." }, { "left": "issue:44075", "right": "issue:44368", "accept": false, - "reason": "SGD optimizer args not used is unrelated to a tie_word_embeddings warning during Qwen3.5 fine-tuning." + "reason": "SGD optimizer args not being used is unrelated to the tie_word_embeddings warning." }, { "left": "issue:43577", "right": "issue:45216", "accept": false, - "reason": "Blip2 dtype propagation and Qwen3.5 save_pretrained corruption are separate model-loading/saving bugs." + "reason": "Blip2 dtype propagation bug is unrelated to Qwen3.5 checkpoint serialization." }, { "left": "issue:44351", "right": "issue:45278", "accept": false, - "reason": "Specific HybridCache import failure versus broad import errors after upgrading to 5.5.0 are not the same issue." + "reason": "45278 is a broad import-regression report, but 44351 is a single specific missing import and not clearly the same underlying issue." }, { "left": "issue:41950", "right": "issue:45020", "accept": false, - "reason": "Video-classification processor lookup and remote_code breakage are different failures." + "reason": "Video-classification processor lookup bug is different from remote_code model breakage." }, { "left": "issue:43976", "right": "issue:45405", "accept": false, - "reason": "Python support regression and an unreleased dependency version bump are unrelated." + "reason": "Python support regression vs unreleased PEFT dependency bump; unrelated." }, { "left": "issue:44779", "right": "issue:45812", "accept": false, - "reason": "Deepseek tokenizer regression and Granite AutoTokenizer ID regression are different model-specific tokenization bugs." - }, - { - "left": "issue:43381", - "right": "issue:47000", - "accept": false, - "reason": "Gradient checkpointing in eval mode and RTDetrLoss infeasible cost matrix are unrelated training/runtime bugs." + "reason": "Different tokenizer regressions affecting different model families and different symptoms." }, { "left": "issue:46082", "right": "issue:46399", "accept": false, - "reason": "LlamaConfig head_dim validation and Qwen3.5 rotary-embedding registration are different code paths." + "reason": "LlamaConfig validation issue is unrelated to the Qwen3.5 rotary-pos-emb registration crash." }, { - "left": "issue:44466", - "right": "issue:45276", + "left": "issue:43381", + "right": "issue:47000", "accept": false, - "reason": "Tied-weight serialization inconsistency and resize_token_embeddings not updating Gemma4 layers are not the same bug." + "reason": "Gradient checkpointing eval-mode issue is unrelated to RTDetr loss infeasibility." }, { - "left": "issue:43299", - "right": "issue:45357", + "left": "issue:44466", + "right": "issue:45276", "accept": false, - "reason": "Model loading breakage and incorrect visual encoder keys on save_pretrained are separate Qwen3.5 regressions." + "reason": "Both touch tied/embedding weights, but one is serialization and the other is resize propagation; not the same bug." } ] }, @@ -12777,6 +12800,7 @@ "issue:45081", "issue:45103", "issue:45200", + "issue:45357", "issue:45372", "issue:45397", "issue:45542", @@ -12795,6 +12819,7 @@ "issue:47145" ], "soft_pairs": [ + "issue:43299|issue:45357", "issue:46082|issue:46512", "issue:43653|issue:45593", "issue:41950|issue:45823", @@ -12802,17 +12827,17 @@ "issue:44242|issue:45200", "issue:43644|issue:47145", "issue:43299|issue:43854", - "issue:43931|issue:47145", "issue:44462|issue:45081", - "issue:45588|issue:47000", + "issue:43931|issue:47145", "issue:44683|issue:45542", + "issue:45588|issue:47000", "issue:44016|issue:44291", "issue:43479|issue:43525", "issue:45081|issue:45907", "issue:33290|issue:45656", "issue:42994|issue:43317", - "issue:43441|issue:45834", "issue:44336|issue:46851", + "issue:43441|issue:45834", "issue:44623|issue:44749", "issue:43873|issue:47145", "issue:45103|issue:45372", @@ -12822,190 +12847,196 @@ "issue:42175|issue:43976", "issue:43901|issue:44246" ], - "prepared_review_unit_hash": "fbcb211b56aa1866ce2fae4bab45fbb1fb763c39d30f15298662b1d108b02653", + "prepared_review_unit_hash": "b21a9762f9e0570fc3f3fd298ce5365ba5de7f3db77c19addc56c65e9dbd5557", "component_budget": { "node_count": 514, "item_count": 514, "soft_pair_count": 1891, - "serialized_chars": 980926, - "estimated_input_tokens": 245232, - "estimated_eval_tokens": 490720 + "serialized_chars": 980895, + "estimated_input_tokens": 245224, + "estimated_eval_tokens": 490704 }, "budget": { - "node_count": 47, - "item_count": 47, - "soft_pair_count": 26, - "serialized_chars": 34849, - "estimated_input_tokens": 8713, - "estimated_eval_tokens": 17682 + "node_count": 48, + "item_count": 48, + "soft_pair_count": 27, + "serialized_chars": 35740, + "estimated_input_tokens": 8935, + "estimated_eval_tokens": 18126 }, "overflow_policy": "truncate_then_skip", "trimmed": false, "aggressively_trimmed": false, "split": true, "analyst_result": { - "summary": "The items are largely unrelated Transformers issues; none of the soft pairs look like true duplicates or mergeable same-bug cases. The only broadly notable open issue is the quantization/offloading report, but this set does not form a coherent duplicate cluster.", - "confidence": 0.97, - "canonical_issue_reason": "Issue 43873 is the most active and broadly framed open bug (quantization/offloading with inbound references), but it is still not a true canonical duplicate for the rest of this mixed set.", + "summary": "This is a loose cluster of mostly unrelated issue reports; the nearest shared theme is loading/offload/quantization regressions. I would use issue #43873 as the canonical issue because it is the broadest representative of that theme, but none of the soft pairs look like true duplicates or the same underlying fix.", + "confidence": 0.87, + "canonical_issue_reason": "#43873 is the broadest, most central issue in the set: it covers quantization/offloading behavior in a way that best matches the cluster\u2019s recurring loading/offload reports.", "canonical_pr_reason": null, - "best_issue_reason": "Issue 43873 is the best anchor only because it is open, high-signal, and broadly scoped; otherwise the cluster is too heterogeneous to pick a stronger representative.", + "best_issue_reason": "#43873 is the best umbrella issue for this cluster because it is broadly scoped and closest to the shared loading/offload/quantization thread seen in the other nearby reports.", "best_pr_reason": null, "soft_edge_verdicts": [ + { + "left": "issue:43299", + "right": "issue:45357", + "accept": false, + "reason": "Both involve Qwen/VL models, but one is a loading failure and the other is a save_pretrained regression with different symptoms and code paths." + }, { "left": "issue:46082", "right": "issue:46512", "accept": false, - "reason": "Different bugs: LlamaConfig validation vs dtype persistence in from_config." + "reason": "Different config bugs: explicit head_dim validation versus stale dtype metadata after from_config(dtype=...)." }, { "left": "issue:43653", "right": "issue:45593", "accept": false, - "reason": "Tokenizer special-token registration vs D-FINE auxiliary-loss behavior; unrelated code paths." + "reason": "Tokenizer special-token decode bug and D-FINE auxiliary-loss behavior are unrelated." }, { "left": "issue:41950", "right": "issue:45823", "accept": false, - "reason": "Pipeline processor lookup bug vs Gemma4 device-mismatch forward bug." + "reason": "Different subsystems and failures: video pipeline processor lookup versus Gemma4 device_map forward mismatch." }, { "left": "issue:43650", "right": "issue:46209", "accept": false, - "reason": "Placeholder issue text vs FP8 inference failure; no shared underlying bug." + "reason": "No clear shared bug; one is a vague data request, the other is an FP8 inference failure on B200." }, { "left": "issue:44242", "right": "issue:45200", "accept": false, - "reason": "MoE load-balancing loss omission vs multimodal token-id defaulting; different failures." + "reason": "One is MoE load-balancing loss gating, the other is Gemma 4 multimodal token defaults; not the same defect." }, { "left": "issue:43644", "right": "issue:47145", "accept": false, - "reason": "Both involve model loading/weights, but one is buffer initialization and the other is FP8 weight_scale application." + "reason": "Both mention loading/corruption, but one is non-persistent buffers filled with junk while the other is FP8 weight_scale not applied on frozen checkpoints." }, { "left": "issue:43299", "right": "issue:43854", "accept": false, - "reason": "Different models and different loading failures; no shared concrete fix." - }, - { - "left": "issue:43931", - "right": "issue:47145", - "accept": false, - "reason": "Shape mismatch on one model vs silent FP8 scale corruption on another." + "reason": "Different model families and failures: Qwen3VL MoE loading versus GLM-4.7-Flash unit-test loading." }, { "left": "issue:44462", "right": "issue:45081", "accept": false, - "reason": "Tokenizer.json selection bug vs mistral regex patch crash; separate loading paths." + "reason": "Tokenizer loading issues, but one ignores tokenizer.json and the other crashes patching Mistral regex due to backend_tokenizer access." }, { - "left": "issue:45588", - "right": "issue:47000", + "left": "issue:43931", + "right": "issue:47145", "accept": false, - "reason": "FlashAttention sink-less crash vs RTDetr loss infeasibility; unrelated subsystems." + "reason": "A weight-shape mismatch for Qwen3-VL is unrelated to FP8 compressed-tensors weight_scale handling." }, { "left": "issue:44683", "right": "issue:45542", "accept": false, - "reason": "Torch flex_attention compile failure vs missing TensorFlow backend dependency." + "reason": "Completely different problems: compiled flex_attention on torch>=2.9 versus a missing TensorFlow backend when only tensorboard is installed." + }, + { + "left": "issue:45588", + "right": "issue:47000", + "accept": false, + "reason": "Different failure modes in different code paths: flash_attention s_aux=None crash versus RTDetrLoss infeasible cost matrix." }, { "left": "issue:44016", "right": "issue:44291", "accept": false, - "reason": "Notebook syntax error vs init_empty_weights argument mismatch; different layers." + "reason": "A notebook syntax error is unrelated to the init_empty_weights TypeError regression." }, { "left": "issue:43479", "right": "issue:43525", "accept": false, - "reason": "Different config objects and fields; not the same bug." + "reason": "Both are config-related, but one concerns multimodal default sub-configs and the other a missing pad_token_id attribute." }, { "left": "issue:45081", "right": "issue:45907", "accept": false, - "reason": "Tokenizer regex patch crash vs repository revision forwarding; unrelated APIs." + "reason": "Tokenizer patching crash and list_repo_files revision forwarding are unrelated APIs." }, { "left": "issue:33290", "right": "issue:45656", "accept": false, - "reason": "Deepspeed OOM from Adafactor vs double optimizer.step calls; distinct issues." + "reason": "Both mention deepspeed, but one is Adafactor OOM and the other is a double optimizer-step bug; not the same underlying issue." }, { "left": "issue:42994", "right": "issue:43317", "accept": false, - "reason": "Quantized saving failure vs dequantized model offload loading failure." + "reason": "Both are quantization/offload adjacent, but one is save_pretrained failure and the other is load failure with gpu/cpu offload." }, { - "left": "issue:43441", - "right": "issue:45834", + "left": "issue:44336", + "right": "issue:46851", "accept": false, - "reason": "Model-specific FlashAttention bug vs OCR index error; no shared code path." + "reason": "ANSI escape handling in loading_report is unrelated to Qwen3.5 variable-length position_ids support." }, { - "left": "issue:44336", - "right": "issue:46851", + "left": "issue:43441", + "right": "issue:45834", "accept": false, - "reason": "ANSI logging formatting issue vs variable-length position_ids bug." + "reason": "Different model-specific bugs: FlashAttention failure for Ministral-3 versus an OCR index error in Kosmos2.5." }, { "left": "issue:44623", "right": "issue:44749", "accept": false, - "reason": "Processor save_pretrained missing files vs unrelated performance regression report." + "reason": "processor.save_pretrained missing files and a slowdown report after upgrading Transformers are not the same bug." }, { "left": "issue:43873", "right": "issue:47145", "accept": false, - "reason": "Both mention quantization/offload, but one is broad offloading behavior and the other is a specific missing weight_scale application." + "reason": "Both are in the broad offload/quantization area, but one is a general offloading behavior issue and the other is a specific FP8 checkpoint scaling bug." }, { "left": "issue:45103", "right": "issue:45372", "accept": false, - "reason": "Auto-docstring AttributeError vs Gemma 4 processor import error; separate components." + "reason": "Auto-docstring annotation handling and Gemma 4 processor import failure are unrelated." }, { "left": "issue:45859", "right": "issue:46851", "accept": false, - "reason": "CPU-offload rotary embedding bug vs variable-length position_ids bug; different fixes." + "reason": "Both are Qwen3.5 issues, but one is CPU offload incompatibility and the other is variable-length position_ids support; different code paths." }, { "left": "issue:44261", "right": "issue:44263", "accept": false, - "reason": "LayerNorm epsilon precision bug vs torch.split return-value handling; unrelated logic." + "reason": "Different model components and defects: MLA RMS norm epsilon precision issue versus torch.split return values in a MoE indexer." }, { "left": "issue:44623", "right": "issue:45397", "accept": false, - "reason": "Processor save_pretrained issue vs gemma-4 zero3 from_pretrained issue." + "reason": "Missing processor save files and Gemma-4 zero3 from_pretrained are separate loading/saving problems." }, { "left": "issue:42175", "right": "issue:43976", "accept": false, - "reason": "Missing TensorFlow backend on install vs Python version compatibility; different root causes." + "reason": "Package dependency installation and Python-version compatibility are different release/build issues." }, { "left": "issue:43901", "right": "issue:44246", "accept": false, - "reason": "Docs mismatch vs import slowness; not the same bug." + "reason": "A docs mismatch about return_all_scores is unrelated to intermittent import latency." } ] }, @@ -13081,47 +13112,47 @@ "soft_pairs": [ "issue:45600|issue:45712", "issue:44671|issue:44779", - "issue:43673|issue:43761", "issue:43525|issue:43881", - "issue:43575|issue:47145", + "issue:43673|issue:43761", "issue:44016|issue:45823", + "issue:43575|issue:47145", "issue:44186|issue:47000", "issue:44188|issue:44336", - "issue:44987|issue:45542", "issue:43824|issue:43976", + "issue:44987|issue:45542", "issue:43525|issue:43643", - "issue:42796|issue:43441", "issue:43723|issue:44162", + "issue:42796|issue:43441", "issue:44336|issue:44373", "issue:44589|issue:45362", "issue:45325|issue:46851", "issue:44246|issue:44496", "issue:46291|issue:46399", - "issue:44871|issue:47145", "issue:44297|issue:45397", + "issue:44871|issue:47145", "issue:45083|issue:46851", - "issue:46655|issue:46726", "issue:41628|issue:43976", + "issue:46655|issue:46726", "issue:43784|issue:45542", "issue:46399|issue:46851", "issue:43526|issue:45003", "issue:45561|issue:45563", "issue:43540|issue:46752" ], - "prepared_review_unit_hash": "ec1a071b0328010f570cd46191739d030c2b3f391e2f0ce76bc7c5dd8c06d843", + "prepared_review_unit_hash": "6e46cd65ae32f3db0c7e69495de98c5921090c42673b1e9bf60f5803f6ee5007", "component_budget": { "node_count": 514, "item_count": 514, "soft_pair_count": 1891, - "serialized_chars": 980926, - "estimated_input_tokens": 245232, - "estimated_eval_tokens": 490720 + "serialized_chars": 980895, + "estimated_input_tokens": 245224, + "estimated_eval_tokens": 490704 }, "budget": { "node_count": 48, "item_count": 48, "soft_pair_count": 28, - "serialized_chars": 36040, + "serialized_chars": 36038, "estimated_input_tokens": 9010, "estimated_eval_tokens": 18276 }, @@ -13130,180 +13161,180 @@ "aggressively_trimmed": false, "split": true, "analyst_result": { - "summary": "This cluster is heterogeneous: most items are unrelated bug reports spanning different models, APIs, and failure modes. The soft pairs share only broad themes like imports, tokenizers, or loading, but not the same underlying defect, so none should be merged as duplicates.", - "confidence": 0.95, - "canonical_issue_reason": "Issue 45561 is the most actionable anchor in this set: it is open, has the strongest ongoing discussion/inbound attention, and describes a concrete, reproducible failure in patched testing utilities.", + "summary": "This cluster is heterogeneous: the soft pairs mostly share broad keywords (tokenizer, model loading, position_ids, kernels, processors) but describe different concrete failures, models, or code paths. I do not see a true duplicate family dominating the cluster, so there is no strong canonical issue candidate here. All soft edges are false positives.", + "confidence": 0.9, + "canonical_issue_reason": null, "canonical_pr_reason": null, - "best_issue_reason": "Issue 45561 is the best representative issue because it is specific, currently active, and has clear downstream impact; the rest are mostly isolated model- or API-specific regressions.", + "best_issue_reason": null, "best_pr_reason": null, "soft_edge_verdicts": [ { "left": "issue:45600", "right": "issue:45712", "accept": false, - "reason": "Both affect repo hygiene, but one is an auto-mapping CI break and the other is leftover dummy PT classes; different files and different bugs." + "reason": "Both are repo-maintenance/regression issues, but one is stale auto-mappings for Sam3LiteText configs and the other is leftover dummy classes leaking into exports; different files and fixes." }, { "left": "issue:44671", "right": "issue:44779", "accept": false, - "reason": "Both are tokenizer regressions, but for different models and different failure modes: CamemBERT MLM predictions vs Deepseek BPE handling." - }, - { - "left": "issue:43673", - "right": "issue:43761", - "accept": false, - "reason": "One is a GenerationMixin cache/chunked_prefill bug; the other is a CLIPVisionModel hidden_states regression. Same broad version, different code paths." + "reason": "Both concern incorrect model behavior in v5, but CamemBERT MLM predictions and Deepseek tokenizer outputs are different model families and code paths." }, { "left": "issue:43525", "right": "issue:43881", "accept": false, - "reason": "These are separate loading failures for different models: Llama4Config pad_token_id missing vs glm-4v-9b loading failure." + "reason": "Separate loading failures: missing pad_token_id on Llama4Config vs a GLM-4V-9B load failure." }, { - "left": "issue:43575", - "right": "issue:47145", + "left": "issue:43673", + "right": "issue:43761", "accept": false, - "reason": "OOM during tensor-parallel load is unrelated to silent FP8 checkpoint corruption from missing weight_scale application." + "reason": "Chunked prefill cache regression in GenerationMixin is unrelated to CLIPVisionModel not returning hidden_states." }, { "left": "issue:44016", "right": "issue:45823", "accept": false, - "reason": "A notebook syntax error and a Gemma4 device-mismatch forward bug are not the same underlying issue." + "reason": "Notebook syntax error and Gemma4 device-mismatch forward bug are unrelated." + }, + { + "left": "issue:43575", + "right": "issue:47145", + "accept": false, + "reason": "Qwen2 OOM on tensor parallel loading and FP8 checkpoint weight_scale loading are different problems." }, { "left": "issue:44186", "right": "issue:47000", "accept": false, - "reason": "LayoutLMv2 tokenizer crashes and RTDetrLoss infeasible cost-matrix errors are distinct model-specific failures." + "reason": "LayoutLMv2 tokenizer NER/padding crash is unrelated to RTDetrLoss infeasible cost matrix." }, { "left": "issue:44188", "right": "issue:44336", "accept": false, - "reason": "One is a torch.compile attention-kernel branching bug; the other is ANSI escape codes in loading_report output." + "reason": "Torch.compile attention branching divergence and ANSI code leakage in loading_report are unrelated." }, { - "left": "issue:44987", - "right": "issue:45542", + "left": "issue:43824", + "right": "issue:43976", "accept": false, - "reason": "These are different loading problems: one is a model load regression for physical-intelligence/fast, the other is a backend-detection error from tensorboard without TensorFlow." + "reason": "Import error for a missing Qwen2.5-VL class is unrelated to the Python 3.9/3.10 compatibility issue." }, { - "left": "issue:43824", - "right": "issue:43976", + "left": "issue:44987", + "right": "issue:45542", "accept": false, - "reason": "ImportError for a missing Qwen2.5-VL class and a Python-version compatibility problem are unrelated." + "reason": "One is a model-loading failure for physical-intelligence/fast; the other is a TensorBoard/TensorFlow backend import issue." }, { "left": "issue:43525", "right": "issue:43643", "accept": false, - "reason": "Missing pad_token_id in Llama4Config and missing fields from AutoConfig with trust_remote_code are distinct config-handling bugs." + "reason": "Both involve config fields, but they are different bugs: missing pad_token_id vs trust_remote_code returning incomplete AutoConfig fields." }, { - "left": "issue:42796", - "right": "issue:43441", + "left": "issue:43723", + "right": "issue:44162", "accept": false, - "reason": "Both mention Ministral-3, but one is tokenizer/BPE-marker handling and the other is a FlashAttention failure in v5 RC." + "reason": "Generic tokenizer loading regression and ESM2-specific breakage are not the same underlying bug." }, { - "left": "issue:43723", - "right": "issue:44162", + "left": "issue:42796", + "right": "issue:43441", "accept": false, - "reason": "Tokenizer loading in v5 and ESM2 breakage are different components and different reported symptoms." + "reason": "Ministral tokenizer BPE-marker handling and Ministral flash-attention failure are different components." }, { "left": "issue:44336", "right": "issue:44373", "accept": false, - "reason": "ANSI-code emission in loading_report is unrelated to a wrong docstring for position_ids." + "reason": "ANSI code emission in loading_report is unrelated to a wrong docstring for position_ids." }, { "left": "issue:44589", "right": "issue:45362", "accept": false, - "reason": "A missing Float8 storage type and a Qwen3.5 chat crash are different runtime failures." + "reason": "Float8 storage lookup failure and Qwen3.5 chat crash are unrelated." }, { "left": "issue:45325", "right": "issue:46851", "accept": false, - "reason": "Both involve position_ids, but one is Qwen2.5-VL still-image rope scaling and the other is Qwen3.5 variable-length input handling." + "reason": "Both mention position_ids, but one is Qwen2.5-VL temporal scaling for still images and the other is Qwen3.5 variable-length input handling." }, { "left": "issue:44246", "right": "issue:44496", "accept": false, - "reason": "Import taking too long and unrecognized model/config loading are unrelated issues." + "reason": "Slow import and unrecognized model/config load failure are unrelated." }, { "left": "issue:46291", "right": "issue:46399", "accept": false, - "reason": "Kernel-version/revision import requirements and a missing decorator on apply_rotary_pos_emb are different failures in different layers." + "reason": "Both mention kernels, but one fails during transformers.pipelines import due to kernels package metadata and the other is a Qwen3.5 kernel-registration decorator issue." }, { - "left": "issue:44871", - "right": "issue:47145", + "left": "issue:44297", + "right": "issue:45397", "accept": false, - "reason": "Gemma-3 EOS config inconsistency and FP8 weight_scale omission are separate checkpoint/config bugs." + "reason": "Tokenizer save_pretrained metadata mismatch is unrelated to Gemma-4 zero3 from_pretrained." }, { - "left": "issue:44297", - "right": "issue:45397", + "left": "issue:44871", + "right": "issue:47145", "accept": false, - "reason": "Tokenizer_config class mismatch on save and gemma-4 zero3 from_pretrained are not the same bug." + "reason": "Gemma eos_token_id inconsistency and FP8 weight_scale loading are different model/config bugs." }, { "left": "issue:45083", "right": "issue:46851", "accept": false, - "reason": "Unexpected helper output lengths in qwen3_omni_moe is unrelated to Qwen3.5 variable-length position_ids behavior." + "reason": "Unexpected feature-extract output lengths in qwen3_omni_moe is not the same as Qwen3.5 variable-length position_ids handling." }, { - "left": "issue:46655", - "right": "issue:46726", + "left": "issue:41628", + "right": "issue:43976", "accept": false, - "reason": "SmolVLM video padding miscalculation and AutoProcessor fallback when vision backend is missing are different processor issues." + "reason": "AutoImageProcessor import failure and Python version compatibility regression are unrelated." }, { - "left": "issue:41628", - "right": "issue:43976", + "left": "issue:46655", + "right": "issue:46726", "accept": false, - "reason": "An AutoImageProcessor import error and a Python 3.9/3.10 compatibility report are unrelated." + "reason": "Both involve processors, but one is a video padding-calculation bug and the other is an AutoProcessor error when a vision backend is missing." }, { "left": "issue:43784", "right": "issue:45542", "accept": false, - "reason": "Sentence-transformers import failing with NameError nn and undefined tf backend error are separate dependency/import issues." + "reason": "sentence-transformers import NameError and TensorBoard/TensorFlow backend error are unrelated." }, { "left": "issue:46399", "right": "issue:46851", "accept": false, - "reason": "A missing decorator on kernel registration is unrelated to variable-length input position_ids handling." + "reason": "Qwen3.5 kernel registration crash and Qwen3.5 variable-length position_ids are different code paths." }, { "left": "issue:43526", "right": "issue:45003", "accept": false, - "reason": "BeitImageProcessorFast reduce_labels behavior and unsafe sys.modules access in modeling_utils are different code paths." + "reason": "BEiT label reduction bug and unsafe sys.modules access are unrelated." }, { "left": "issue:45561", "right": "issue:45563", "accept": false, - "reason": "One is a pytest-xdist race on captured_info.txt; the other is a stale warning in paged generate()." + "reason": "pytest-xdist race on captured_info.txt and a stale num_return_sequences warning are unrelated." }, { "left": "issue:43540", "right": "issue:46752", "accept": false, - "reason": "Qwen3OmniMoe video-input ValueError and empty-list apply_chat_template IndexError are distinct crashes." + "reason": "Qwen3OmniMoe video-input ValueError and empty-list apply_chat_template IndexError are unrelated." } ] }, @@ -13339,6 +13370,7 @@ "issue:43540", "issue:43550", "issue:43901", + "issue:44162", "issue:44220", "issue:44242", "issue:44246", @@ -13356,11 +13388,11 @@ "issue:45072", "issue:45103", "issue:45137", + "issue:45200", "issue:45292", "issue:45362", "issue:45375", "issue:45397", - "issue:45561", "issue:45588", "issue:45593", "issue:45759", @@ -13378,108 +13410,108 @@ "soft_pairs": [ "issue:45103|issue:46399", "issue:45072|issue:46829", - "issue:44246|issue:44871", "issue:44295|issue:45397", + "issue:44246|issue:44871", "issue:43531|issue:45362", - "issue:46614|issue:46665", "issue:45588|issue:46851", + "issue:46614|issue:46665", "issue:42491|issue:43299", "issue:44336|issue:44485", "issue:45003|issue:46821", "issue:44242|issue:45593", "issue:43550|issue:46506", - "issue:44779|issue:45292", "issue:42886|issue:46132", + "issue:44779|issue:45292", "issue:36246|issue:42757", "issue:36246|issue:44351", "issue:43901|issue:45003", "issue:43257|issue:45137", "issue:44496|issue:45003", - "issue:43540|issue:46655", "issue:41950|issue:44220", - "issue:43262|issue:44246", + "issue:43540|issue:46655", "issue:44451|issue:46752", - "issue:44479|issue:46655", + "issue:43262|issue:44246", "issue:45375|issue:45759", + "issue:44479|issue:46655", "issue:44295|issue:44749", - "issue:45003|issue:45561" + "issue:44162|issue:45200" ], - "prepared_review_unit_hash": "77b80df9ccd611f9ce6ea50f788e9e0ffaf4d2c44e382a0fd4cf8946a2726d99", + "prepared_review_unit_hash": "cd8dddf845478ea7a94744ac0a1dfd3e00510399fe0f877429d5c7cb565ef7fc", "component_budget": { "node_count": 514, "item_count": 514, "soft_pair_count": 1891, - "serialized_chars": 980926, - "estimated_input_tokens": 245232, - "estimated_eval_tokens": 490720 + "serialized_chars": 980895, + "estimated_input_tokens": 245224, + "estimated_eval_tokens": 490704 }, "budget": { - "node_count": 47, - "item_count": 47, + "node_count": 48, + "item_count": 48, "soft_pair_count": 27, - "serialized_chars": 35294, - "estimated_input_tokens": 8824, - "estimated_eval_tokens": 17904 + "serialized_chars": 35864, + "estimated_input_tokens": 8966, + "estimated_eval_tokens": 18188 }, "overflow_policy": "truncate_then_skip", "trimmed": false, "aggressively_trimmed": false, "split": true, "analyst_result": { - "summary": "This cluster is mostly heterogeneous: import/export errors, Qwen-family model regressions, video/audio processor bugs, and unrelated infra/docs issues. None of the proposed soft pairs look like the same underlying bug or change, so no merges are justified.", - "confidence": 0.96, - "canonical_issue_reason": "No single canonical issue fits this cluster; the items are too diverse to safely consolidate under one bug report.", + "summary": "These items are mostly distinct regressions or documentation/config issues that only share broad subsystem/model-family overlap. No soft pair looks like the same concrete bug/change, so I rejected all duplicate candidates.", + "confidence": 0.77, + "canonical_issue_reason": "issue:44479 is the broadest concrete regression in the set: it describes a multi-model v5.3.0 video-input breakage with a clear common failure mode, making it the best single anchor if one issue must represent this cluster.", "canonical_pr_reason": null, - "best_issue_reason": "No global best issue stands out as a true duplicate hub or superset of the others.", + "best_issue_reason": "issue:44479 is the most representative issue because it spans multiple related Qwen video models and names a specific regression rather than a narrow, model-specific symptom.", "best_pr_reason": null, "soft_edge_verdicts": [ { "left": "issue:45103", "right": "issue:46399", "accept": false, - "reason": "Different failures: auto_docstring annotation handling vs Qwen3.5 kernel registration/decorator error." + "reason": "Both mention loading/registration errors, but one is an auto-docstring crash from future annotations and the other is a Qwen3.5 kernel-registration problem; different code paths and bugs." }, { "left": "issue:45072", "right": "issue:46829", "accept": false, - "reason": "Unrelated code paths: bfloat16 dtype mismatch in inference vs a VideoPrism output field mix-up." + "reason": "These are both output/hidden-state mismatches, but they affect different models and different inference paths; not the same underlying defect." }, { - "left": "issue:44246", - "right": "issue:44871", + "left": "issue:44295", + "right": "issue:45397", "accept": false, - "reason": "Import-time slowness is unrelated to the Gemma eos_token_id config inconsistency." + "reason": "A buffer registration access error is unrelated to the Gemma-4 ZeRO3 from_pretrained failure; only generic training/loading context overlaps." }, { - "left": "issue:44295", - "right": "issue:45397", + "left": "issue:44246", + "right": "issue:44871", "accept": false, - "reason": "Buffer access error and ZeRO3 from_pretrained failure are distinct issues." + "reason": "One is an import-performance complaint and the other is an eos_token_id config inconsistency; no shared concrete bug." }, { "left": "issue:43531", "right": "issue:45362", "accept": false, - "reason": "Both involve Qwen models, but one is sliding_window behavior and the other is chat loading failure; not the same bug." + "reason": "Both are Qwen-related, but one concerns sliding_window handling in MoE models while the other is a chat crash for Qwen3.5; different failure modes." }, { - "left": "issue:46614", - "right": "issue:46665", + "left": "issue:45588", + "right": "issue:46851", "accept": false, - "reason": "Stale phonemizer language state is unrelated to from_pretrained behavior inside Python classes." + "reason": "Flash-attention sink handling and variable-length position_ids are separate code paths and symptoms." }, { - "left": "issue:45588", - "right": "issue:46851", + "left": "issue:46614", + "right": "issue:46665", "accept": false, - "reason": "Flash-attention s_aux handling and variable-length position_ids are different failure modes." + "reason": "Stale phonemizer language state is unrelated to from_pretrained behavior inside Python classes." }, { "left": "issue:42491", "right": "issue:43299", "accept": false, - "reason": "Both are Qwen MoE-related, but they target different loading/conversion problems and versions." + "reason": "Both involve Qwen MoE loading, but one is LoRA compatibility across hf4.x/hf5.x while the other is a Qwen3VL MoE load regression; not the same bug." }, { "left": "issue:44336", @@ -13491,109 +13523,109 @@ "left": "issue:45003", "right": "issue:46821", "accept": false, - "reason": "sys.modules access safety and TimesFM window_size AttributeError are unrelated." + "reason": "Unsafe sys.modules access in modeling_utils and a TimesFM window_size AttributeError are different issues with no shared code path." }, { "left": "issue:44242", "right": "issue:45593", "accept": false, - "reason": "Both mention missing auxiliary losses, but they are model-specific and caused by different conditions." + "reason": "Both concern losses, but one is router/load-balancing loss gating and the other is D-FINE auxiliary loss suppression; distinct models and logic." }, { "left": "issue:43550", "right": "issue:46506", "accept": false, - "reason": "torch.compile/SDPA failure and CPU-vs-CUDA output divergence are unrelated." + "reason": "torch.compile+SDPA on Bamba is unrelated to CPU/CUDA output differences in PPDocLayoutV3." }, { - "left": "issue:44779", - "right": "issue:45292", + "left": "issue:42886", + "right": "issue:46132", "accept": false, - "reason": "Tokenizer correctness regression is not the same as resize_token_embeddings not updating output embeddings." + "reason": "Offline cache loading and AttentionInterface.register behavior changes are unrelated." }, { - "left": "issue:42886", - "right": "issue:46132", + "left": "issue:44779", + "right": "issue:45292", "accept": false, - "reason": "Offline cache loading and AttentionInterface.register behavior do not share a concrete code-path bug." + "reason": "Deepseek tokenizer regression and resize_token_embeddings not updating output embeddings are different functional bugs." }, { "left": "issue:36246", "right": "issue:42757", "accept": false, - "reason": "Both are import errors, but they involve different missing symbols from different packages." + "reason": "These are both import errors, but they involve different missing symbols from different packages and unrelated fixes." }, { "left": "issue:36246", "right": "issue:44351", "accept": false, - "reason": "Both are missing imports from transformers, but the symbols and underlying fixes differ." + "reason": "Both are import/export issues, but Qwen2_5_VLImageProcessor and HybridCache are different symbols with different causes." }, { "left": "issue:43901", "right": "issue:45003", "accept": false, - "reason": "Docs mismatch for return_all_scores is unrelated to unsafe sys.modules access." + "reason": "Docs mentioning return_all_scores versus unsafe sys.modules access are not the same issue." }, { "left": "issue:43257", "right": "issue:45137", "accept": false, - "reason": "Qwen3 MoE weight conversion and ZeRO3 deque IndexError are different DeepSpeed-related failures." + "reason": "DeepSpeed ZeRO3 deque pop error and Qwen3 MoE weight conversion under accelerate+deepspeed are both training-stack issues but not the same concrete bug." }, { "left": "issue:44496", "right": "issue:45003", "accept": false, - "reason": "Model config recognition error is unrelated to sys.modules access in modeling_utils." - }, - { - "left": "issue:43540", - "right": "issue:46655", - "accept": false, - "reason": "Different video-processing bugs affecting different models and code paths." + "reason": "Unrecognized model config and unsafe sys.modules access are unrelated loading problems." }, { "left": "issue:41950", "right": "issue:44220", "accept": false, - "reason": "Video-classification processor lookup and fbank feature extraction are unrelated." + "reason": "Video-classification pipeline image-processor lookup and _torch_extract_fbank_features are separate media pipeline bugs." }, { - "left": "issue:43262", - "right": "issue:44246", + "left": "issue:43540", + "right": "issue:46655", "accept": false, - "reason": "Audio chat template sampling-rate behavior is unrelated to import slowness." + "reason": "Both touch video processing, but one is a Qwen3OmniMoe value error and the other is a SmolVLM padding-calculation bug; different code paths." }, { "left": "issue:44451", "right": "issue:46752", "accept": false, - "reason": "Model loading failure and empty-chat-template IndexError are different issues." + "reason": "Model loading failure for ScandiBERT is unrelated to apply_chat_template on an empty list." }, { - "left": "issue:44479", - "right": "issue:46655", + "left": "issue:43262", + "right": "issue:44246", "accept": false, - "reason": "Both touch video inputs, but they are model-specific regressions with different symptoms." + "reason": "Audio processor chat-template sampling rate defaults and slow imports are unrelated." }, { "left": "issue:45375", "right": "issue:45759", "accept": false, - "reason": "Config field dropping and text_config unwrapping are distinct multimodal config issues." + "reason": "Both involve Qwen 3.5/3.6 multimodal config handling, but one is a strict-config field drop and the other is from_config failing to unwrap text_config; different underlying bugs." + }, + { + "left": "issue:44479", + "right": "issue:46655", + "accept": false, + "reason": "Both are video-input regressions, but they affect different models and likely different preprocessing logic; too broad to treat as one duplicate." }, { "left": "issue:44295", "right": "issue:44749", "accept": false, - "reason": "Buffer registration error and major slowdown after upgrade are unrelated." + "reason": "A buffer/position_ids access issue is unrelated to the Chinese-language performance regression after upgrading transformers." }, { - "left": "issue:45003", - "right": "issue:45561", + "left": "issue:44162", + "right": "issue:45200", "accept": false, - "reason": "Unsafe sys.modules access has nothing to do with the pytest-xdist captured_info race." + "reason": "ESM2 breakage and Gemma 4 mm_token_type_ids defaults are different model-specific problems." } ] }, @@ -13627,7 +13659,6 @@ "issue:43408", "issue:43475", "issue:43526", - "issue:43540", "issue:43596", "issue:43653", "issue:44038", @@ -13643,8 +13674,8 @@ "issue:44877", "issue:44977", "issue:44987", + "issue:45003", "issue:45137", - "issue:45200", "issue:45250", "issue:45292", "issue:45335", @@ -13652,6 +13683,7 @@ "issue:45397", "issue:45412", "issue:45459", + "issue:45561", "issue:45600", "issue:45656", "issue:45859", @@ -13667,18 +13699,18 @@ "issue:47000" ], "soft_pairs": [ - "issue:44162|issue:45200", + "issue:45003|issue:45561", "issue:45292|issue:47000", - "issue:45859|issue:46399", "issue:45137|issue:45600", + "issue:45859|issue:46399", "issue:43295|issue:45292", "issue:36246|issue:43596", "issue:43475|issue:46369", "issue:44466|issue:45335", "issue:44623|issue:46726", + "issue:44877|issue:45375", "issue:44242|issue:44857", "issue:44038|issue:46506", - "issue:44877|issue:45375", "issue:43526|issue:46897", "issue:44987|issue:45459", "issue:29870|issue:30333", @@ -13692,199 +13724,192 @@ "issue:45397|issue:45656", "issue:45412|issue:46123", "issue:43065|issue:43408", - "issue:44877|issue:46170", - "issue:43475|issue:43540" + "issue:44877|issue:46170" ], - "prepared_review_unit_hash": "f485e580c194d8ce210754cdb3c2c9df5a29a47e869b05f488036b19449fba19", + "prepared_review_unit_hash": "a9f1c0428a4e1e6bc71c92dd075158cc02a20a895f46f90efcb9e8285b7f7051", "component_budget": { "node_count": 514, "item_count": 514, "soft_pair_count": 1891, - "serialized_chars": 980926, - "estimated_input_tokens": 245232, - "estimated_eval_tokens": 490720 + "serialized_chars": 980895, + "estimated_input_tokens": 245224, + "estimated_eval_tokens": 490704 }, "budget": { "node_count": 48, "item_count": 48, - "soft_pair_count": 27, - "serialized_chars": 35932, - "estimated_input_tokens": 8983, - "estimated_eval_tokens": 18222 + "soft_pair_count": 26, + "serialized_chars": 35612, + "estimated_input_tokens": 8903, + "estimated_eval_tokens": 18062 }, "overflow_policy": "truncate_then_skip", "trimmed": false, "aggressively_trimmed": false, "split": true, "analyst_result": { - "summary": "Most items are unrelated issue reports spanning different models, subsystems, and failure modes. I found one plausible duplicate pair around `resize_token_embeddings` not propagating updated embeddings; the rest should stay separate.", - "confidence": 0.81, - "canonical_issue_reason": "Issue 45292 is the broad, model-agnostic report for the `resize_token_embeddings` propagation bug; 45335 looks like a model-specific manifestation of the same underlying problem.", + "summary": "This cluster is highly heterogeneous: it spans unrelated bugs in generation, deepspeed, processors/tokenizers, SAM3, Qwen, DETR, config handling, and CI. None of the soft edges look like the same underlying defect or a single mergeable PR.", + "confidence": 0.98, + "canonical_issue_reason": "No clear canonical issue: the items do not share one underlying bug, so there is no meaningful duplicate hub for the cluster.", "canonical_pr_reason": null, - "best_issue_reason": "If consolidating the only credible duplicate pair, 45292 is the best representative because it states the general embedding-resize failure rather than a single architecture-specific symptom.", + "best_issue_reason": "No strong representative issue. The closest generic-looking item is 45292, but it is still unrelated to most others and should not be treated as the cluster center.", "best_pr_reason": null, "soft_edge_verdicts": [ { - "left": "issue:44162", - "right": "issue:45200", + "left": "issue:45003", + "right": "issue:45561", "accept": false, - "reason": "Different models and different failures; both are training-related but not the same bug." + "reason": "Different failures: unsafe sys.modules access in modeling_utils vs a pytest-xdist race on captured_info.txt in testing utilities." }, { "left": "issue:45292", "right": "issue:47000", "accept": false, - "reason": "`resize_token_embeddings` mismatch is unrelated to RT-DETR loss infeasibility." + "reason": "Unrelated bugs: token embedding resize not updating output embeddings vs RT-DETR loss crashing with infeasible cost matrix." }, { - "left": "issue:45859", - "right": "issue:46399", + "left": "issue:45137", + "right": "issue:45600", "accept": false, - "reason": "Both touch Qwen3.5 rotary code, but one is CPU offload compatibility and the other is a kernel registration/decorator crash." + "reason": "Different code paths and symptoms: DeepSpeed ZeRO3 deque underflow vs stale auto_mappings references to removed Sam3LiteText configs." }, { - "left": "issue:45137", - "right": "issue:45600", + "left": "issue:45859", + "right": "issue:46399", "accept": false, - "reason": "DeepSpeed ZeRO3 deque error and stale auto_mappings config references are unrelated." + "reason": "Both involve Qwen3.5-family models, but one is CPU offload incompatibility in rotary embedding and the other is a missing registration decorator crash with use_kernels=True." }, { "left": "issue:43295", "right": "issue:45292", "accept": false, - "reason": "Processor/tokenizer regression is unrelated to embedding resizing." + "reason": "Different regressions: processor/tokenizer API breakage vs resize_token_embeddings not propagating to output embeddings." }, { "left": "issue:36246", "right": "issue:43596", "accept": false, - "reason": "Importing a missing Qwen2.5-VL image processor is unrelated to the BertModel ZeRO3 indexing error." + "reason": "Completely different problems: missing Qwen2_5_VLImageProcessor import vs DeepSpeed ZeRO3/BertModel IndexError." }, { "left": "issue:43475", "right": "issue:46369", "accept": false, - "reason": "Same SAM3 family, but one is a missing vision-encoder attribute and the other is memory overflow." + "reason": "Same broad SAM3 area, but distinct bugs: missing fpn_position_embeddings attribute vs memory overflow when calling sam3-litetext." }, { "left": "issue:44466", "right": "issue:45335", "accept": false, - "reason": "Serialization of tied weights is a different bug from resize_token_embeddings not updating decoder embeddings." + "reason": "Related to tied embeddings/resize behavior, but not the same bug: device-dependent lm_head serialization vs t5gemma decoder.embed_tokens not updating." }, { "left": "issue:44623", "right": "issue:46726", "accept": false, - "reason": "Saving missing processor files and backend-dependent AutoProcessor loading are different problems." + "reason": "Different processor issues: save_pretrained missing files vs AutoProcessor complaining about missing vision backend." + }, + { + "left": "issue:44877", + "right": "issue:45375", + "accept": false, + "reason": "Both mention strict config, but the concrete defects differ: granite_speech loading blocked vs Qwen3_5MoeVisionConfig field being silently dropped." }, { "left": "issue:44242", "right": "issue:44857", "accept": false, - "reason": "Router load-balancing loss logic is unrelated to the RT-DETR AMP crash." + "reason": "Unrelated modeling bugs: MoE load-balancing loss omission vs LwDetrImageLoss crashing under float16 AMP on CUDA." }, { "left": "issue:44038", "right": "issue:46506", "accept": false, - "reason": "Qwen3-VL-Moe loading issue and PPDocLayoutV3 CPU/CUDA divergence are unrelated." - }, - { - "left": "issue:44877", - "right": "issue:45375", - "accept": false, - "reason": "Both involve strict config handling, but they affect different model configs and different missing fields." + "reason": "Different models and failures: Qwen3-VL-Moe / transformers 5.0 breakage vs CPU/CUDA output mismatch in PPDocLayoutV3." }, { "left": "issue:43526", "right": "issue:46897", "accept": false, - "reason": "BEiT label reduction bug is unrelated to the Florence2 double-shift training-loss bug." + "reason": "Different label-processing bug vs ASR training-loss double-shift bug; same general ML area but not the same defect." }, { "left": "issue:44987", "right": "issue:45459", "accept": false, - "reason": "Model loading failure and protobuf-related tokenizer error masking are unrelated." + "reason": "Loading regression for physical-intelligence/fast vs tokenizer error masking when protobuf is missing; no shared code-path bug." }, { "left": "issue:29870", "right": "issue:30333", "accept": false, - "reason": "Bart generate-time UnboundLocalError and MLflow reporting after training failure are unrelated." + "reason": "Generation-time Bart unbound local error vs MLflow reporting not updated after training failure; unrelated functionality." }, { "left": "issue:43352", "right": "issue:45250", "accept": false, - "reason": "Generic Flash Attention 2.0 topic is not the same concrete bug as NemotronHForCausalLM lacking support." + "reason": "One is a model-specific unsupported Flash Attention 2 guard, the other is a generic Flash Attention 2 topic; not the same concrete bug." }, { "left": "issue:43526", "right": "issue:46894", "accept": false, - "reason": "Label reduction and CohereASR double-shift are different training/data-processing bugs." + "reason": "Different issues despite similar 'double' wording: BEiT label reduction bug vs CohereASR training-loss double-shift bug." }, { "left": "issue:44479", "right": "issue:46170", "accept": false, - "reason": "Video-input regression in Qwen models is unrelated to the Windows/ROCm torch.distributed dependency issue." + "reason": "Unrelated: video-input regression for Qwen vision-language models vs Windows/ROCm import/type dependency problems." }, { "left": "issue:42199", "right": "issue:46851", "accept": false, - "reason": "DETR cardinality handling and Qwen3.5 variable-length position_ids are unrelated." + "reason": "Different model families and failure modes: DETR cardinality/background-class error vs Qwen3.5 variable-length position_ids behavior." }, { "left": "issue:44060", "right": "issue:44077", "accept": false, - "reason": "Tied-weights warning and patchtsmixer post_init validation are unrelated." + "reason": "Different issues in the Qwen/patchtsmixer ecosystem: tied-weights warning vs optional post_init policy change." }, { "left": "issue:44368", "right": "issue:44977", "accept": false, - "reason": "Tie-word-embeddings warning during LoRA fine-tuning is unrelated to the flash-attention generation bug." + "reason": "One is a tied-embeddings warning in LoRA fine-tuning, the other is a flash-attention generation failure for Qwen3.5." }, { "left": "issue:43653", "right": "issue:44162", "accept": false, - "reason": "BigBird tokenizer special-token registration and ESM2 breakage are unrelated." + "reason": "Different tokenizer/model bugs: BigBirdTokenizer special-token decode problem vs ESM2 being broken." }, { "left": "issue:45397", "right": "issue:45656", "accept": false, - "reason": "ZeRO3 from_pretrained failure and optimizer step being called twice are different DeepSpeed issues." + "reason": "Different DeepSpeed problems: gemma-4 from_pretrained failure vs optimizer.step being called twice." }, { "left": "issue:45412", "right": "issue:46123", "accept": false, - "reason": "RT-DETR GC/memory retention and MaskGenerationPipeline final-batch handling are unrelated." + "reason": "Memory/resource management in RT-DETR vs a MaskGenerationPipeline partial-batch bookkeeping bug; not the same defect." }, { "left": "issue:43065", "right": "issue:43408", "accept": false, - "reason": "SAM3 pixel-decoder dummy layer and SAM3 model-type mismatch are different config/model bugs." + "reason": "Different SAM3 issues: dummy Conv2d in Sam3PixelDecoder vs wrong model type warning when loading sam3_video into sam3_tracker." }, { "left": "issue:44877", "right": "issue:46170", "accept": false, - "reason": "Strict config loading and Windows ROCm import/type errors are unrelated." - }, - { - "left": "issue:43475", - "right": "issue:43540", - "accept": false, - "reason": "Missing SAM3 vision-encoder attribute is unrelated to Qwen3OmniMoe video-input validation." + "reason": "Strict config loading issue vs Windows/ROCm torch.distributed dependency failure; unrelated code paths." } ] }, @@ -13909,9 +13934,9 @@ "cluster_id": "cluster-10000-514", "nodes": [ "issue:42886", - "issue:42994", "issue:43065", - "issue:43482", + "issue:43475", + "issue:43540", "issue:43646", "issue:43701", "issue:43723", @@ -13958,21 +13983,22 @@ "issue:47145" ], "soft_pairs": [ + "issue:43475|issue:43540", "issue:46170|issue:46832", "issue:42886|issue:45229", "issue:43723|issue:46710", "issue:43931|issue:44368", "issue:44188|issue:46202", - "issue:44995|issue:45200", "issue:44368|issue:45563", + "issue:44995|issue:45200", "issue:46132|issue:47145", "issue:44483|issue:46202", "issue:43906|issue:46202", "issue:44188|issue:46421", - "issue:44995|issue:45588", "issue:45593|issue:47068", - "issue:45685|issue:46431", + "issue:44995|issue:45588", "issue:45987|issue:46082", + "issue:45685|issue:46431", "issue:44361|issue:44749", "issue:45103|issue:46858", "issue:44297|issue:44623", @@ -13985,217 +14011,216 @@ "issue:46752|issue:47020", "issue:44871|issue:46132", "issue:44964|issue:45137", - "issue:45375|issue:45478", - "issue:42994|issue:43482" + "issue:45375|issue:45478" ], - "prepared_review_unit_hash": "7aaddb3661c1ce613724f1ac20b829dc71a038e089761a5e8980a79cb9f02f19", + "prepared_review_unit_hash": "6cd7e63e15c22fd5aef32fab4a7934487e4176fbf51590d5e25e89b176676ec5", "component_budget": { "node_count": 514, "item_count": 514, "soft_pair_count": 1891, - "serialized_chars": 980926, - "estimated_input_tokens": 245232, - "estimated_eval_tokens": 490720 + "serialized_chars": 980895, + "estimated_input_tokens": 245224, + "estimated_eval_tokens": 490704 }, "budget": { "node_count": 48, "item_count": 48, "soft_pair_count": 29, - "serialized_chars": 36543, - "estimated_input_tokens": 9136, - "estimated_eval_tokens": 18528 + "serialized_chars": 36618, + "estimated_input_tokens": 9155, + "estimated_eval_tokens": 18566 }, "overflow_policy": "truncate_then_skip", "trimmed": false, "aggressively_trimmed": false, "split": true, "analyst_result": { - "summary": "Most pairs are unrelated symptom-level issues. The only strong duplicate signal is the Qwen3.5 MoE config/loading regression, where a missing config field appears to drive the from_pretrained failure.", - "confidence": 0.78, - "canonical_issue_reason": "issue:45375 is the most specific root-cause report: it identifies the missing `deepstack_visual_indexes` field and explains why strict config handling drops it.", + "summary": "This cluster is mostly heterogeneous: most pairs are different bugs across unrelated models, pipelines, or subsystems. The only clear duplicate-looking pair is the Qwen3.5 MoE config/load failure; everything else should stay separate.", + "confidence": 0.79, + "canonical_issue_reason": "issue:45375 is the best canonical issue because it identifies the likely root cause directly: Qwen3_5MoeVisionConfig is missing deepstack_visual_indexes and @strict drops it, which explains the broader from_pretrained failure.", "canonical_pr_reason": null, - "best_issue_reason": "issue:45375 is the best representative issue because it states the concrete config defect, which is more actionable than the broader downstream loading-error report.", + "best_issue_reason": "issue:45375 is the strongest representative issue in the set: it is specific, root-cause oriented, and directly describes the underlying config/schema regression.", "best_pr_reason": null, "soft_edge_verdicts": [ + { + "left": "issue:43475", + "right": "issue:43540", + "accept": false, + "reason": "Different models and failures: SAM 3 video attribute error vs Qwen3OmniMoe video input validation." + }, { "left": "issue:46170", "right": "issue:46832", "accept": false, - "reason": "Different subsystems and failures: Windows/ROCm import dependency issue vs RT-DETR feature-level mismatch." + "reason": "Unrelated subsystems: Windows/ROCm import dependency error vs RT-DETR forward shape/type error." }, { "left": "issue:42886", "right": "issue:45229", "accept": false, - "reason": "Tokenizer offline cache loading and Gemma4 multi-GPU OOM are unrelated." + "reason": "Both are transformer bugs, but one is offline tokenizer cache loading and the other is Gemma multi-GPU OOM." }, { "left": "issue:43723", "right": "issue:46710", "accept": false, - "reason": "Tokenizer loading in v5 vs accuracy regression in a specific model after upgrade; not the same bug." + "reason": "Tokenizer loading regression vs model accuracy regression after upgrade; not the same bug." }, { "left": "issue:43931", "right": "issue:44368", "accept": false, - "reason": "Weight-shape mismatch during model load is unrelated to a config warning about `tie_word_embeddings`." + "reason": "Weight-shape mismatch on Qwen3-VL load is unrelated to a warning about tie_word_embeddings in fine-tuning." }, { "left": "issue:44188", "right": "issue:46202", "accept": false, - "reason": "Torch.compile attention-kernel divergence and Gemma pipeline roundtrip tokenization are different problems." + "reason": "Different code paths: torch.compile attention-kernel divergence vs Gemma pipeline roundtrip/tokenization issue." }, { - "left": "issue:44995", - "right": "issue:45200", + "left": "issue:44368", + "right": "issue:45563", "accept": false, - "reason": "Stale cache crash on second forward pass is not the same as missing `mm_token_type_ids` defaults." + "reason": "A config warning about tied embeddings is unrelated to a stale generate() warning on num_return_sequences." }, { - "left": "issue:44368", - "right": "issue:45563", + "left": "issue:44995", + "right": "issue:45200", "accept": false, - "reason": "Both are warnings, but they concern different fields and behaviors; not the same underlying issue." + "reason": "Stale MoE indexer cache on second forward is not the same as Gemma 4 missing mm_token_type_ids defaults." }, { "left": "issue:46132", "right": "issue:47145", "accept": false, - "reason": "Registered-function behavior change is unrelated to FP8 checkpoint weight_scale loading corruption." + "reason": "AttentionInterface registration behavior change is unrelated to FP8 checkpoint loading missing weight_scale." }, { "left": "issue:44483", "right": "issue:46202", "accept": false, - "reason": "API request rejection and Gemma pipeline roundtrip failure are different code paths." + "reason": "Chat/completions API rejection and Gemma pipeline roundtrip failure are different bugs." }, { "left": "issue:43906", "right": "issue:46202", "accept": false, - "reason": "An isolated reproduction of an earlier issue is not the same as the Gemma pipeline token roundtrip bug." + "reason": "An isolated reproduction of another issue is not the same as a Gemma tokenizer roundtrip failure." }, { "left": "issue:44188", "right": "issue:46421", "accept": false, - "reason": "Both touch compilation/caching, but one is attention-kernel divergence and the other is spurious recompilation from lazy cache init." - }, - { - "left": "issue:44995", - "right": "issue:45588", - "accept": false, - "reason": "Second-forward stale indexer cache is unrelated to flash_attention crashing on `s_aux=None`." + "reason": "torch.compile attention divergence and cache lazy-init recompilation are distinct problems." }, { "left": "issue:45593", "right": "issue:47068", "accept": false, - "reason": "Auxiliary-loss behavior in D-FINE is not the same as a host-sync concern in DeepSpeed loss computation." + "reason": "D-FINE auxiliary-loss logic is unrelated to a deepspeed loss host-sync concern." }, { - "left": "issue:45685", - "right": "issue:46431", + "left": "issue:44995", + "right": "issue:45588", "accept": false, - "reason": "MPS histogram kernel limitation and missing chat template are unrelated failures." + "reason": "Stale MoE cache on second forward is unrelated to flash_attention crashing on s_aux=None." }, { "left": "issue:45987", "right": "issue:46082", "accept": false, - "reason": "Cache length return-type bug is unrelated to LlamaConfig head_dim validation." + "reason": "Cache return-type contract bug and LlamaConfig head_dim validation are different issues." + }, + { + "left": "issue:45685", + "right": "issue:46431", + "accept": false, + "reason": "MPS histogram failure in MoE has no overlap with missing chat-template support in a processor." }, { "left": "issue:44361", "right": "issue:44749", "accept": false, - "reason": "Tokenizer AttributeError and a performance regression after upgrade are different issues." + "reason": "Tokenizer AttributeError on tasks is unrelated to a performance regression after upgrade." }, { "left": "issue:45103", "right": "issue:46858", "accept": false, - "reason": "auto_docstring annotation handling and static-cache seq-length device-tensor behavior are unrelated." + "reason": "auto_docstring crashes with future annotations; this is unrelated to static cache seq-length behavior under torch.compile." }, { "left": "issue:44297", "right": "issue:44623", "accept": false, - "reason": "Both are save_pretrained issues, but one is wrong tokenizer_class metadata and the other is missing processor files; distinct bugs." + "reason": "Tokenizer save_pretrained metadata mismatch and processor save_pretrained missing files are related only superficially." }, { "left": "issue:43065", "right": "issue:46123", "accept": false, - "reason": "Dummy Conv2d in SAM decoder is unrelated to mask generation pipeline batch handling." + "reason": "SAM3 pixel decoder dummy Conv2d is unrelated to MaskGenerationPipeline dropping the final partial batch." }, { "left": "issue:43994", "right": "issue:46726", "accept": false, - "reason": "Nonsensical model outputs and missing vision backend handling are not the same defect." + "reason": "Siglip2 nonsensical outputs and missing-vision-backend processor detection are different failure modes." }, { "left": "issue:44749", "right": "issue:45656", "accept": false, - "reason": "Data-filtering slowdown after upgrade is unrelated to deepspeed causing double optimizer steps." + "reason": "Performance slowdown after upgrade is not the same as deepspeed calling optimizer.step() twice." }, { "left": "issue:43701", "right": "issue:44749", "accept": false, - "reason": "Checkpoint resume key mismatch is unrelated to a performance regression in filtering." + "reason": "Checkpoint key mismatch and post-upgrade slowdown do not share a concrete code-path bug." }, { "left": "issue:43646", "right": "issue:44162", "accept": false, - "reason": "Both mention broad v5 breakage, but they concern different models and different failure modes." + "reason": "Custom model initialization breakage and ESM2 breakage are both upgrade regressions but not the same defect." }, { "left": "issue:44623", "right": "issue:45656", "accept": false, - "reason": "Processor save_pretrained file omissions are unrelated to double optimizer stepping under deepspeed." + "reason": "Processor serialization missing files is unrelated to deepspeed double optimizer stepping." }, { "left": "issue:46752", "right": "issue:47020", "accept": false, - "reason": "Empty chat-template input crash and XLMRobertaTokenizer constructor type error are different tokenizer bugs." + "reason": "Empty chat-template input IndexError and XLMRobertaTokenizer Unigram construction error are unrelated." }, { "left": "issue:44871", "right": "issue:46132", "accept": false, - "reason": "Gemma eos_token_id mismatch is unrelated to AttentionInterface.register behavior changes." + "reason": "Gemma eos_token_id inconsistency does not match AttentionInterface.register behavior change." }, { "left": "issue:44964", "right": "issue:45137", "accept": false, - "reason": "Phi-4 multimodal loading error and DeepSpeed ZeRO3 deque pop crash are unrelated." + "reason": "Phi-4 multimodal load failure and DeepSpeed ZeRO3 deque underflow are unrelated." }, { "left": "issue:45375", "right": "issue:45478", "accept": true, - "reason": "Same Qwen3.5 MoE loading regression: one report identifies the missing config field/root cause, the other reports the resulting from_pretrained failure." - }, - { - "left": "issue:42994", - "right": "issue:43482", - "accept": false, - "reason": "Quantized model saving failure and Qwen2.5-GGUF loading failure are separate serialization/loading problems." + "reason": "Likely same Qwen3.5 MoE loading regression: the config issue explains the broader from_pretrained error." } ] }, "evaluator_result": { "accept": true, - "feedback": "The analysis is well grounded in the packet. The 45375/45478 pairing is the only clearly supported duplicate signal, and the reasons stay conservative and specific to the missing `deepstack_visual_indexes` config field versus the resulting `from_pretrained` error. The other soft-edge rejections are generally cautious and do not overclaim linkage from superficial similarity. Summary language is also appropriately restrained." + "feedback": "The summary is grounded in the packet and stays conservative. The canonical pair 45375/45478 is well supported by the titles and looks mergeable as the same Qwen3.5 MoE config/loading fix. The other soft-edge rejections are cautious and consistent with the visible evidence; no overstatement stands out." }, "evaluator_used": true, "retried": false, @@ -14225,6 +14250,7 @@ "issue:42994", "issue:43232", "issue:43475", + "issue:43482", "issue:43525", "issue:43643", "issue:43673", @@ -14238,22 +14264,22 @@ "issue:44351", "issue:44373", "issue:44623", + "issue:44661", "issue:44857", "issue:44869", "issue:44933", "issue:44945", - "issue:45081", "issue:45092", - "issue:45103", "issue:45137", "issue:45230", "issue:45310", "issue:45356", "issue:45375", "issue:45397", + "issue:45406", "issue:45561", - "issue:45593", "issue:45753", + "issue:46202", "issue:46326", "issue:46399", "issue:46463", @@ -14265,255 +14291,259 @@ "issue:47059" ], "soft_pairs": [ - "issue:39692|issue:46726", + "issue:42994|issue:43482", "issue:43475|issue:46821", "issue:43475|issue:43643", + "issue:45230|issue:45397", "issue:44188|issue:46851", + "issue:39692|issue:46726", "issue:43232|issue:43673", - "issue:45230|issue:45397", - "issue:43986|issue:46726", "issue:28282|issue:30990", - "issue:46877|issue:47000", - "issue:45310|issue:45375", "issue:44373|issue:46851", + "issue:45310|issue:45375", + "issue:43986|issue:46726", "issue:44038|issue:44623", - "issue:46463|issue:46877", + "issue:46877|issue:47000", "issue:41628|issue:44933", + "issue:46463|issue:46877", "issue:42994|issue:43525", "issue:44869|issue:45356", "issue:43723|issue:46326", "issue:42371|issue:45092", - "issue:43873|issue:44945", "issue:43784|issue:46399", + "issue:43873|issue:44945", "issue:45137|issue:47000", "issue:41628|issue:44351", "issue:36010|issue:41628", - "issue:44945|issue:45753", "issue:42994|issue:43824", - "issue:44857|issue:46877", + "issue:44945|issue:45753", "issue:45561|issue:47059", "issue:44038|issue:45310", - "issue:45081|issue:45103", - "issue:45593|issue:46877" + "issue:44857|issue:46877", + "issue:44661|issue:47059", + "issue:45406|issue:46202" ], - "prepared_review_unit_hash": "f5d19afbea4c8a4117706ff265b9526f758ad81527fed03b863ac8c5cba2b9a3", + "prepared_review_unit_hash": "08fb64bae1a853ebd56a7044612aa6d2f2592ffa2d4247288ebfce24e54727ae", "component_budget": { "node_count": 514, "item_count": 514, "soft_pair_count": 1891, - "serialized_chars": 980926, - "estimated_input_tokens": 245232, - "estimated_eval_tokens": 490720 + "serialized_chars": 980895, + "estimated_input_tokens": 245224, + "estimated_eval_tokens": 490704 }, "budget": { - "node_count": 47, - "item_count": 47, - "soft_pair_count": 30, - "serialized_chars": 35937, - "estimated_input_tokens": 8985, - "estimated_eval_tokens": 18226 + "node_count": 48, + "item_count": 48, + "soft_pair_count": 31, + "serialized_chars": 36646, + "estimated_input_tokens": 9162, + "estimated_eval_tokens": 18580 }, "overflow_policy": "truncate_then_skip", "trimmed": false, "aggressively_trimmed": false, "split": true, "analyst_result": { - "summary": "Mostly a heterogeneous mix of unrelated Transformers bugs; only one clear duplicate-like subcluster emerged around AutoProcessor failing when a required vision backend is missing. The remaining soft matches are mostly keyword/model-family similarity rather than the same concrete bug.", - "confidence": 0.68, - "canonical_issue_reason": null, + "summary": "This cluster is a loose mix of transformer import, generation, tokenizer, and model-specific regression reports. The soft links are mostly superficial wording overlap rather than the same underlying bug, so I would not merge any of them as duplicates.", + "confidence": 0.84, + "canonical_issue_reason": "issue:41628 is the broadest and most central import-regression report in the set, and it best represents the small import-error subset of this loose cluster.", "canonical_pr_reason": null, - "best_issue_reason": "Issue 46726 is the best representative of the only credible duplicate subcluster: AutoProcessor/vision-backend missing handling. It describes the concrete failure clearly and is closest to 43986.", + "best_issue_reason": "issue:41628 is the most generally framed user-facing failure here and the closest thing to a cluster anchor, though the overall set is still heterogeneous.", "best_pr_reason": null, "soft_edge_verdicts": [ { - "left": "issue:39692", - "right": "issue:46726", + "left": "issue:42994", + "right": "issue:43482", "accept": false, - "reason": "Different bugs: a SigLIP2 docs/example mismatch and quantization failure versus AutoProcessor backend-missing error handling." + "reason": "Save failure for quantized models and GGUF loading failure are different code paths and different operations." }, { "left": "issue:43475", "right": "issue:46821", "accept": false, - "reason": "Both are AttributeError reports, but they affect different models and different missing fields/code paths." + "reason": "Different models and different missing attributes; SAM3 video output bug is unrelated to TimesFM window_size." }, { "left": "issue:43475", "right": "issue:43643", "accept": false, - "reason": "SAM3 output attribute loss is unrelated to AutoConfig returning incomplete fields under trust_remote_code." + "reason": "One is a model output attribute bug, the other is a generic remote-code config field loss issue." }, { - "left": "issue:44188", - "right": "issue:46851", + "left": "issue:45230", + "right": "issue:45397", "accept": false, - "reason": "Different runtime problems: torch.compile attention-kernel divergence versus Qwen3.5 position_ids handling." + "reason": "Generic bug report versus a specific gemma-4 ZeRO3 from_pretrained failure; not the same underlying issue." }, { - "left": "issue:43232", - "right": "issue:43673", + "left": "issue:44188", + "right": "issue:46851", "accept": false, - "reason": "Both involve generation internals, but one is sync_gpus/update_model_kwargs behavior and the other is chunked_prefill cache handling." + "reason": "torch.compile attention-kernel divergence is unrelated to Qwen3.5 variable-length position_ids handling." }, { - "left": "issue:45230", - "right": "issue:45397", + "left": "issue:39692", + "right": "issue:46726", "accept": false, - "reason": "Issue 45230 is too generic to establish the same bug, and the other is a specific Gemma-4 ZeRO3 loading failure." + "reason": "Documentation example errors are separate from runtime missing-vision-backend processor behavior." }, { - "left": "issue:43986", - "right": "issue:46726", - "accept": true, - "reason": "Same underlying failure mode: AutoProcessor crashes or misreports when the required vision backend/torchvision is missing." + "left": "issue:43232", + "right": "issue:43673", + "accept": false, + "reason": "Both mention generation, but they target different failures: sync_gpus kwargs handling vs chunked_prefill cache access." }, { "left": "issue:28282", "right": "issue:30990", "accept": false, - "reason": "An AutoModel PyTorch ImportError is unrelated to Sentence Transformers getting stuck while loading." + "reason": "Missing PyTorch dependency and sentence-transformers hanging on load are different problems." }, { - "left": "issue:46877", - "right": "issue:47000", + "left": "issue:44373", + "right": "issue:46851", "accept": false, - "reason": "Completely unrelated: auto_docstring import-time diagnostics versus RTDetrLoss infeasible cost matrix." + "reason": "A wrong docstring is not the same as a runtime bug in variable-length position_ids handling." }, { "left": "issue:45310", "right": "issue:45375", "accept": false, - "reason": "Same model family, but one is a generic from_pretrained error and the other is a specific config-field omission; not enough evidence of the same defect." + "reason": "Related model family, but one is a from_pretrained error and the other is a silent config-field drop." }, { - "left": "issue:44373", - "right": "issue:46851", + "left": "issue:43986", + "right": "issue:46726", "accept": false, - "reason": "A docstring issue about position_ids is not the same as a runtime variable-length input failure." + "reason": "Both mention vision backends, but they describe different loaders and different missing-backend failures." }, { "left": "issue:44038", "right": "issue:44623", "accept": false, - "reason": "Different areas: Qwen3-VL-Moe behavior versus processor.save_pretrained missing files." + "reason": "A Qwen3-VL-Moe bug and processor.save_pretrained missing files are unrelated failures." }, { - "left": "issue:46463", - "right": "issue:46877", + "left": "issue:46877", + "right": "issue:47000", "accept": false, - "reason": "Incorrect T5 validation messaging is unrelated to auto_docstring printing lint diagnostics." + "reason": "Import-time doc-lint output and an RTDetr loss crash are entirely unrelated." }, { "left": "issue:41628", "right": "issue:44933", "accept": false, - "reason": "Both are import-related, but they concern different missing symbols/modules and different failure causes." + "reason": "Both are import errors, but they are for different symbols and different modules." + }, + { + "left": "issue:46463", + "right": "issue:46877", + "accept": false, + "reason": "T5 decoder-input validation messaging has nothing to do with doc-lint output at import time." }, { "left": "issue:42994", "right": "issue:43525", "accept": false, - "reason": "Quantized model saving and a missing pad_token_id on Llama4Config are different bugs." + "reason": "Quantized saving failure and missing pad_token_id on Llama4Config are different bugs." }, { "left": "issue:44869", "right": "issue:45356", "accept": false, - "reason": "Both mention tokenizer/codec behavior, but they are different models and different failure modes." + "reason": "Whisper timestamp decoding crash and Gemma/Kimi tokenizer regression are separate tokenizer issues." }, { "left": "issue:43723", "right": "issue:46326", "accept": false, - "reason": "AutoTokenizer loading in v5 and continue_final_message for mistral_common tokenizers are separate issues." + "reason": "Tokenizer loading in v5 and mistral_common continue_final_message are different tokenizer paths." }, { "left": "issue:42371", "right": "issue:45092", "accept": false, - "reason": "TF32 settings guidance is unrelated to InternVL2 remote-code/meta-init incompatibility." + "reason": "TF32 API guidance and InternVL2 meta-initialization incompatibility are unrelated." }, { - "left": "issue:43873", - "right": "issue:44945", + "left": "issue:43784", + "right": "issue:46399", "accept": false, - "reason": "Quantization offloading and pipeline-parallelism output corruption are different code paths." + "reason": "Sentence-transformers import NameError and Qwen3.5 kernel registration failure are unrelated." }, { - "left": "issue:43784", - "right": "issue:46399", + "left": "issue:43873", + "right": "issue:44945", "accept": false, - "reason": "Sentence-Transformers import failure and Qwen3.5 kernel registration crash are unrelated." + "reason": "Quantization offloading behavior and pipeline-parallel output correctness are different issues." }, { "left": "issue:45137", "right": "issue:47000", "accept": false, - "reason": "DeepSpeed ZeRO3 deque underflow and RTDetrLoss infeasible cost matrix are unrelated." + "reason": "DeepSpeed ZeRO3 deque underflow and RTDetr loss infeasibility are unrelated crashes." }, { "left": "issue:41628", "right": "issue:44351", "accept": false, - "reason": "Both are import errors from transformers, but the missing symbols are different and imply different regressions." + "reason": "Different missing imports: AutoImageProcessor versus HybridCache." }, { "left": "issue:36010", "right": "issue:41628", "accept": false, - "reason": "PyTorch missing for AutoModel is unrelated to an AutoImageProcessor import failure." - }, - { - "left": "issue:44945", - "right": "issue:45753", - "accept": false, - "reason": "Pipeline-parallelism output errors and an image-specific infinite loop are different problems." + "reason": "GenerationMixin import failure and AutoImageProcessor import failure are distinct compatibility issues." }, { "left": "issue:42994", "right": "issue:43824", "accept": false, - "reason": "Quantized model saving and missing Qwen2.5-VL import are unrelated." + "reason": "Quantized saving failure and missing Qwen2_5_VLForConditionalGeneration import are unrelated." }, { - "left": "issue:44857", - "right": "issue:46877", + "left": "issue:44945", + "right": "issue:45753", "accept": false, - "reason": "A CUDA/AMP loss crash is unrelated to auto_docstring import-time stdout noise." + "reason": "Pipeline-parallel output errors and a Qwen3.5 infinite-loop case are not the same bug." }, { "left": "issue:45561", "right": "issue:47059", "accept": false, - "reason": "Both are test failures, but one is a pytest-xdist file race and the other is a huggingface_hub CLI API break." + "reason": "pytest-xdist file-race in testing utils and a huggingface_hub CLI attribute error are unrelated." }, { "left": "issue:44038", "right": "issue:45310", "accept": false, - "reason": "Same broad family, but the model variants and failure details differ too much to treat as one bug." + "reason": "Same model family, but different versions and different failure modes; not mergeable as one bug." }, { - "left": "issue:45081", - "right": "issue:45103", + "left": "issue:44857", + "right": "issue:46877", "accept": false, - "reason": "Tokenizer regex patching and auto_docstring future-annotations handling are unrelated." + "reason": "CUDA AMP loss crash and doc-lint stdout at import time are completely different problems." }, { - "left": "issue:45593", - "right": "issue:46877", + "left": "issue:44661", + "right": "issue:47059", + "accept": false, + "reason": "Model-generation mapping logic and CLI completion compatibility are unrelated." + }, + { + "left": "issue:45406", + "right": "issue:46202", "accept": false, - "reason": "D-FINE auxiliary-loss behavior is unrelated to auto_docstring diagnostics output." + "reason": "Gemma4 processor missing _tokenizer and Gemma pipeline roundtrip issues are different defects." } ] }, - "evaluator_result": { - "accept": true, - "feedback": "The summary is well grounded in the packet and stays conservative. The only accepted soft edge, issue:43986 \u2194 issue:46726, is supported by both titles describing AutoProcessor failures when a required vision backend is missing. The remaining pair rejections are appropriately cautious and do not overclaim duplicate status." - }, - "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 }, @@ -14543,8 +14573,6 @@ "issue:43479", "issue:43704", "issue:43723", - "issue:43784", - "issue:43824", "issue:43976", "issue:44016", "issue:44079", @@ -14553,16 +14581,17 @@ "issue:44496", "issue:44568", "issue:44623", - "issue:44661", "issue:44743", "issue:44779", "issue:44792", "issue:44869", "issue:44933", + "issue:45081", + "issue:45103", "issue:45310", "issue:45405", - "issue:45406", "issue:45561", + "issue:45593", "issue:45600", "issue:45712", "issue:45812", @@ -14580,12 +14609,12 @@ "issue:47085" ], "soft_pairs": [ - "issue:45406|issue:46202", - "issue:44661|issue:47059", + "issue:45593|issue:46877", "issue:44222|issue:45310", + "issue:45081|issue:45103", "issue:46202|issue:46701", - "issue:42673|issue:43704", "issue:43232|issue:44792", + "issue:42673|issue:43704", "issue:42831|issue:44222", "issue:42175|issue:44933", "issue:44496|issue:45561", @@ -14594,9 +14623,9 @@ "issue:44743|issue:47068", "issue:43976|issue:44933", "issue:30990|issue:35141", - "issue:42898|issue:44568", "issue:45561|issue:45600", "issue:43479|issue:46829", + "issue:42898|issue:44568", "issue:42175|issue:45712", "issue:44568|issue:47085", "issue:44623|issue:45310", @@ -14609,236 +14638,229 @@ "issue:43479|issue:44869", "issue:44016|issue:46231", "issue:39692|issue:46909", - "issue:43784|issue:43824", - "issue:42898|issue:44779", - "issue:43479|issue:44079" + "issue:43479|issue:44079", + "issue:42898|issue:44779" ], - "prepared_review_unit_hash": "c9d0a6216a15d2f59f552247640040b5212bdd70be084496719e4a6964224102", + "prepared_review_unit_hash": "fceeb80d80ee08c311bd8abfa74948b00f997a653319e2e83a9d46f7de96b2be", "component_budget": { "node_count": 514, "item_count": 514, "soft_pair_count": 1891, - "serialized_chars": 980926, - "estimated_input_tokens": 245232, - "estimated_eval_tokens": 490720 + "serialized_chars": 980895, + "estimated_input_tokens": 245224, + "estimated_eval_tokens": 490704 }, "budget": { - "node_count": 48, - "item_count": 48, - "soft_pair_count": 32, - "serialized_chars": 37341, - "estimated_input_tokens": 9336, - "estimated_eval_tokens": 18928 + "node_count": 47, + "item_count": 47, + "soft_pair_count": 31, + "serialized_chars": 36553, + "estimated_input_tokens": 9139, + "estimated_eval_tokens": 18534 }, "overflow_policy": "truncate_then_skip", "trimmed": false, "aggressively_trimmed": false, "split": true, "analyst_result": { - "summary": "This cluster is mostly heterogeneous, with one clear duplicate pair around the Qwen3ForCausalLM VRAM leak; the rest are separate bugs in different subsystems and should not be merged.", - "confidence": 0.91, - "canonical_issue_reason": "No single canonical issue fits the whole set: most items are unrelated, and only issue:42673 / issue:43704 form a true duplicate pair.", - "canonical_pr_reason": "No pull requests are present in the input.", - "best_issue_reason": "issue:42673 is the strongest anchor among the issues because it has a clear exact duplicate (issue:43704) and a concrete, reproducible bug report.", + "summary": "This set is mostly heterogeneous; the only clear duplicate pair is the Qwen3ForCausalLM VRAM leak report. No pull requests are present.", + "confidence": 0.95, + "canonical_issue_reason": "issue:42673 is the stronger canonical issue because it is the more detailed, earlier report of the same Qwen3ForCausalLM VRAM leak in multi-dataloader-thread usage; issue:43704 is a later near-identical repeat.", + "canonical_pr_reason": null, + "best_issue_reason": "issue:42673 is the best representative issue for the only confirmed duplicate subcluster and has the richest original description.", "best_pr_reason": null, "soft_edge_verdicts": [ { - "left": "issue:45406", - "right": "issue:46202", + "left": "issue:45593", + "right": "issue:46877", "accept": false, - "reason": "Both are Gemma-related, but one is a missing processor attribute and the other is a pipeline roundtrip/tokenization issue; different bugs." + "reason": "Different bugs: FP8/auxiliary-loss behavior vs auto_docstring import-time stderr logging." }, { - "left": "issue:44661", - "right": "issue:47059", + "left": "issue:44222", + "right": "issue:45310", "accept": false, - "reason": "Unrelated subsystems: model-mapping logic versus CLI completion breakage from huggingface_hub." + "reason": "Both mention MoE/FP8, but one is save_pretrained behavior and the other is from_pretrained loading failure; not the same bug." }, { - "left": "issue:44222", - "right": "issue:45310", + "left": "issue:45081", + "right": "issue:45103", "accept": false, - "reason": "Both mention MoE/FP8, but one is a save_pretrained issue and the other is a from_pretrained loading regression; not the same bug." + "reason": "Both are AttributeError crashes, but in unrelated code paths (Mistral regex patching vs auto_docstring kwargs processing)." }, { "left": "issue:46202", "right": "issue:46701", "accept": false, - "reason": "Different Gemma pipeline failures with different symptoms and code paths; not mergeable as one fix." - }, - { - "left": "issue:42673", - "right": "issue:43704", - "accept": true, - "reason": "Same title, same symptom, same underlying Qwen3ForCausalLM VRAM leak in multiple dataloader threads; clear duplicate." + "reason": "Both are Gemma-related, but one is tokenizer roundtrip on \u2581 and the other is an image-text pipeline output issue." }, { "left": "issue:43232", "right": "issue:44792", "accept": false, - "reason": "Generation kwargs handling versus a Janus model test failure; no shared root cause." + "reason": "Unrelated generation-state bug vs a janus test failure." + }, + { + "left": "issue:42673", + "right": "issue:43704", + "accept": true, + "reason": "Same issue: identical title and same Qwen3ForCausalLM VRAM leak when used across multiple dataloader threads." }, { "left": "issue:42831", "right": "issue:44222", "accept": false, - "reason": "Both involve FP8, but one reports accuracy degradation and the other save_pretrained behavior for MoE; different issues." + "reason": "Different FP8 problems: accuracy regression vs save_pretrained on MoE." }, { "left": "issue:42175", "right": "issue:44933", "accept": false, - "reason": "TensorFlow backend packaging and an image_utils import error are unrelated." + "reason": "Different subsystems: backend extras installation vs image_utils import regression." }, { "left": "issue:44496", "right": "issue:45561", "accept": false, - "reason": "Model auto-loading/config resolution versus a pytest-xdist race in testing utils; unrelated." + "reason": "Unrelated: model loading/config error vs pytest-xdist file race in testing utils." }, { "left": "issue:45405", "right": "issue:46877", "accept": false, - "reason": "PEFT version pinning and docstring lint output are unrelated maintenance issues." + "reason": "Unrelated packaging/version pin issue vs auto_docstring import-time diagnostics." }, { "left": "issue:30333", "right": "issue:33290", "accept": false, - "reason": "MLFlow reporting on failed training jobs is unrelated to a DeepSpeed Adafactor OOM." + "reason": "Different failure modes: MLflow reporting after training job failure vs Deepspeed/Adafactor OOM." }, { "left": "issue:44743", "right": "issue:47068", "accept": false, - "reason": "Recurrent-state reset in modular Qwen3.5 and a possible CUDA host sync in loss computation are different code paths and symptoms." + "reason": "Different bugs in different generation/training paths." }, { "left": "issue:43976", "right": "issue:44933", "accept": false, - "reason": "Python version support breakage versus an import error in image_utils; unrelated." + "reason": "Python version compatibility issue vs nonexistent image_utils import." }, { "left": "issue:30990", "right": "issue:35141", "accept": false, - "reason": "Sentence Transformers loading hang and token embedding reinitialization are unrelated." - }, - { - "left": "issue:42898", - "right": "issue:44568", - "accept": false, - "reason": "Both are tokenizer regressions in v5, but one is clean-up spacing behavior and the other is missing BOS/EOS insertion; different bugs." + "reason": "Unrelated: loading hang vs embedding resize/post_init reinitialization." }, { "left": "issue:45561", "right": "issue:45600", "accept": false, - "reason": "Pytest-xdist race on a generated file versus stale auto_mappings references; unrelated." + "reason": "Different CI/test problems: xdist file contention vs removed Sam3LiteText config references." }, { "left": "issue:43479", "right": "issue:46829", "accept": false, - "reason": "Multimodal config defaults and VideoPrism output field wiring are different defects." + "reason": "Different model-output/config initialization bug vs output field mismatch in VideoPrism classification." + }, + { + "left": "issue:42898", + "right": "issue:44568", + "accept": false, + "reason": "Both tokenizer regressions, but one is clean-up spacing behavior and the other is special-token addition for a specific tokenizer." }, { "left": "issue:42175", "right": "issue:45712", "accept": false, - "reason": "Packaging/backend dependency issue versus leftover dummy classes leaking into exports; not the same change." + "reason": "Different installation/backend extras issue vs dummy_pt_objects leakage into dir(transformers)." }, { "left": "issue:44568", "right": "issue:47085", "accept": false, - "reason": "Tokenizer special-token behavior and tokenizer thread-safety are distinct problems." + "reason": "Different tokenizer behaviors; no shared concrete code-path failure." }, { "left": "issue:44623", "right": "issue:45310", "accept": false, - "reason": "Processor save_pretrained file handling and Qwen3.5 MoE loading are unrelated." + "reason": "Different loading/saving bugs: processor.save_pretrained missing files vs Qwen3.5 MoE from_pretrained failure." }, { "left": "issue:42673", "right": "issue:46693", "accept": false, - "reason": "VRAM leak in dataloader threads and a long-context performance regression are not the same bug." + "reason": "Both mention performance, but one is a VRAM leak in dataloader-thread usage and the other is a long-context FA2 regression." }, { "left": "issue:44186", "right": "issue:47085", "accept": false, - "reason": "LayoutLMv2 tokenizer crashes on NER inputs versus fast tokenizer thread-safety regressions; different issues." + "reason": "Different tokenizer issues; no evidence they share the same underlying thread-safety bug." }, { "left": "issue:45600", "right": "issue:47059", "accept": false, - "reason": "CI break from removed configs versus CLI completion attribute error; unrelated." + "reason": "Different CI failures: removed config references vs huggingface_hub CLI incompatibility." }, { "left": "issue:46077", "right": "issue:46123", "accept": false, - "reason": "Gemma4 attention mask construction and MaskGenerationPipeline final-batch handling are different pipeline bugs." + "reason": "Different pipeline behaviors: bidirectional mask construction vs partial-batch result dropping." }, { "left": "issue:42898", "right": "issue:45812", "accept": false, - "reason": "Both are v5 tokenizer regressions, but one concerns cleanup-space behavior and the other wrong token IDs for Granite models; not the same root cause." + "reason": "Both are tokenizer regressions, but different concrete failures and models." }, { "left": "issue:43723", "right": "issue:44568", "accept": false, - "reason": "Tokenizer loading failure and special-token insertion regression are separate tokenizer issues." + "reason": "Different v5 tokenizer loading issue vs special-token/BOS-EOS behavior on one tokenizer." }, { "left": "issue:43479", "right": "issue:44869", "accept": false, - "reason": "Multimodal config defaults versus Whisper timestamp decoding crash are unrelated." + "reason": "Different bugs: config default initialization vs Whisper word-timestamp decoding crash." }, { "left": "issue:44016", "right": "issue:46231", "accept": false, - "reason": "Notebook syntax error and documentation parameter mismatch are different documentation problems." + "reason": "Notebook syntax error vs documentation parameter-name mismatch." }, { "left": "issue:39692", "right": "issue:46909", "accept": false, - "reason": "Both are broken docs examples, but they concern different models and different failure modes; not a duplicate pair." + "reason": "Both are broken docs examples, but for different models and different failure modes." }, { - "left": "issue:43784", - "right": "issue:43824", + "left": "issue:43479", + "right": "issue:44079", "accept": false, - "reason": "One is a NameError importing sentence-transformers, the other is an ImportError for a missing model class; unrelated." + "reason": "Different issues: multimodal config defaults vs ModelOutput key assignment semantics." }, { "left": "issue:42898", "right": "issue:44779", "accept": false, - "reason": "Different tokenizer regressions: cleanup-space behavior versus DeepSeek tokenization correctness." - }, - { - "left": "issue:43479", - "right": "issue:44079", - "accept": false, - "reason": "Multimodal config initialization and ModelOutput key assignment are unrelated." + "reason": "Both are tokenizer regressions, but different models and different incorrect outputs." } ] }, "evaluator_result": { "accept": true, - "feedback": "The summary is grounded in the packet: issue:42673 and issue:43704 are an exact duplicate pair with identical titles, and the rest of the cluster is reasonably characterized as heterogeneous. The soft-edge rejections are conservative and do not overclaim shared root cause. No pull requests are present, so the PR reasoning is also consistent." + "feedback": "The summary is grounded in the packet. The only confirmed duplicate pair is correctly identified as issue:42673 / issue:43704, with conservative reasoning and the earlier, more detailed report chosen as canonical. The soft-edge verdicts stay cautious and are supported by the titles and metadata. No pull requests are present, matching the packet." }, "evaluator_used": true, "retried": false, @@ -14877,6 +14899,8 @@ "issue:43479", "issue:43644", "issue:43723", + "issue:43784", + "issue:43824", "issue:43874", "issue:44060", "issue:44075", @@ -14897,7 +14921,6 @@ "issue:45125", "issue:45200", "issue:45245", - "issue:45310", "issue:45356", "issue:45405", "issue:45656", @@ -14908,14 +14931,15 @@ "issue:47085" ], "soft_pairs": [ + "issue:43784|issue:43824", "issue:44843|issue:47085", - "issue:27957|issue:30333", "issue:42673|issue:47085", + "issue:27957|issue:30333", "issue:30333|issue:34689", - "issue:42898|issue:43295", "issue:43874|issue:45092", - "issue:42222|issue:45200", + "issue:42898|issue:43295", "issue:44295|issue:45656", + "issue:42222|issue:45200", "issue:44734|issue:45103", "issue:42898|issue:45356", "issue:30333|issue:35141", @@ -14925,221 +14949,220 @@ "issue:43116|issue:44190", "issue:45405|issue:45712", "issue:44297|issue:45656", - "issue:44933|issue:45712", "issue:42831|issue:44623", + "issue:44933|issue:45712", "issue:44448|issue:46132", - "issue:44060|issue:45125", "issue:44222|issue:44623", + "issue:44060|issue:45125", "issue:42796|issue:43723", "issue:43479|issue:43644", "issue:45245|issue:46132", "issue:44373|issue:44485", "issue:44360|issue:46762", - "issue:41628|issue:42757", - "issue:44297|issue:45310" + "issue:41628|issue:42757" ], - "prepared_review_unit_hash": "5b96407e197a76650d9940108cd76a60f4900065ce3ed40c162c2f7dd8808c2c", + "prepared_review_unit_hash": "fb4aed9b77579f719f1630f1684b7f2767cd2550b045cfcf6d6e48c39fb97c85", "component_budget": { "node_count": 514, "item_count": 514, "soft_pair_count": 1891, - "serialized_chars": 980926, - "estimated_input_tokens": 245232, - "estimated_eval_tokens": 490720 + "serialized_chars": 980895, + "estimated_input_tokens": 245224, + "estimated_eval_tokens": 490704 }, "budget": { - "node_count": 47, - "item_count": 47, + "node_count": 48, + "item_count": 48, "soft_pair_count": 29, - "serialized_chars": 35704, - "estimated_input_tokens": 8926, - "estimated_eval_tokens": 18108 + "serialized_chars": 36332, + "estimated_input_tokens": 9083, + "estimated_eval_tokens": 18422 }, "overflow_policy": "truncate_then_skip", "trimmed": false, "aggressively_trimmed": false, "split": true, "analyst_result": { - "summary": "The soft pairs are mostly keyword-similar but do not describe the same underlying bug or change. All candidate duplicate edges should be rejected, and there is no clear canonical issue/PR for this heterogeneous set.", - "confidence": 0.96, - "canonical_issue_reason": "No canonical issue: the items span unrelated topics (tokenizers, multimodal models, optimizers, serving, docs), and the soft similarities do not indicate one shared root cause.", + "summary": "This is a heterogeneous grab bag of unrelated Transformers issues (imports, tokenizer/model regressions, training/runtime bugs, docs, and config problems). None of the soft-similarity pairs look like the same underlying bug or change, so all proposed duplicate links should be rejected.", + "confidence": 0.91, + "canonical_issue_reason": "issue:30333 is the least-arbitrary anchor only because it has the most soft links, but the cluster is too heterogeneous for a true canonical duplicate.", "canonical_pr_reason": null, - "best_issue_reason": "No best issue target: none of these issues is a plausible umbrella duplicate for the rest.", + "best_issue_reason": "issue:30333 is the best available representative by graph connectivity, though it is still not a genuine duplicate hub.", "best_pr_reason": null, "soft_edge_verdicts": [ { - "left": "issue:44843", - "right": "issue:47085", + "left": "issue:43784", + "right": "issue:43824", "accept": false, - "reason": "Offline HF hub lookup in `AutoTokenizer.from_pretrained` is unrelated to tokenizer thread-safety regressions." + "reason": "Both are import errors, but one is about sentence-transformers/nn and the other about a missing Qwen VL class; unrelated failures." }, { - "left": "issue:27957", - "right": "issue:30333", + "left": "issue:44843", + "right": "issue:47085", "accept": false, - "reason": "Flash Attention 2 model support and MLflow reporting on failed jobs are different subsystems and bugs." + "reason": "Offline tokenizer patching vs Rust tokenizer thread-safety are different code paths and bugs." }, { "left": "issue:42673", "right": "issue:47085", "accept": false, - "reason": "VRAM leakage in dataloader threads is not the same as Rust tokenizer thread-safety." + "reason": "VRAM leak with Qwen3-Next in dataloader threads is not the same as a tokenizer thread-safety regression." }, { - "left": "issue:30333", - "right": "issue:34689", + "left": "issue:27957", + "right": "issue:30333", "accept": false, - "reason": "MLflow reporting failure and Llama 3.2 vision model-loading regression have different causes and code paths." + "reason": "Flash Attention 2 on XLMRoberta and MLFlow reporting after failed training are unrelated." }, { - "left": "issue:42898", - "right": "issue:43295", + "left": "issue:30333", + "right": "issue:34689", "accept": false, - "reason": "Tokenization cleanup behavior changes and processor/tokenizer image-handling regression are distinct issues." + "reason": "Training-job MLFlow reporting and Llama 3.2 Vision model loading are different issues." }, { "left": "issue:43874", "right": "issue:45092", "accept": false, - "reason": "Missing image patch-count method and old remote-code/meta-init incompatibility are different multimodal bugs." + "reason": "Missing image-processor method in GLM46V is unrelated to InternVL2 meta-init incompatibility." }, { - "left": "issue:42222", - "right": "issue:45200", + "left": "issue:42898", + "right": "issue:43295", "accept": false, - "reason": "VitPose model breakage and Gemma 4 multimodal token type defaults are unrelated." + "reason": "Both are v5 regressions, but one is tokenizer cleanup behavior and the other is processor/tokenizer image handling." }, { "left": "issue:44295", "right": "issue:45656", "accept": false, - "reason": "Reading `position_ids` after buffer registration and double optimizer stepping under deepspeed are unrelated." + "reason": "Buffer registration access error and double optimizer step with deepspeed are unrelated." + }, + { + "left": "issue:42222", + "right": "issue:45200", + "accept": false, + "reason": "VitPose model breakage and Gemma 4 mm_token_type_ids defaults are different multimodal bugs." }, { "left": "issue:44734", "right": "issue:45103", "accept": false, - "reason": "KV-cache tensor indexing crash in serving and auto-docstring annotation crash are unrelated failures." + "reason": "Serve/KV-cache tensor indexing crash and auto_docstring future-annotations crash are unrelated." }, { "left": "issue:42898", "right": "issue:45356", "accept": false, - "reason": "`clean_up_tokenization_spaces` behavior and Kimi-K2.5 tokenizer codec/regex regression are not the same bug." + "reason": "Tokenizer cleanup behavior change and Kimi-K2.5 codec/regression warning are different tokenizer issues." }, { "left": "issue:30333", "right": "issue:35141", "accept": false, - "reason": "MLflow failure reporting and embedding reinitialization in `post_init` are unrelated." + "reason": "MLFlow reporting on failed jobs and embedding resize/post_init reinitialization are unrelated." }, { "left": "issue:22355", "right": "issue:28282", "accept": false, - "reason": "Missing `transformers.onnx` module and missing PyTorch dependency error are different import problems." + "reason": "Missing transformers.onnx module and missing PyTorch dependency are distinct import problems." }, { "left": "issue:44222", "right": "issue:44297", "accept": false, - "reason": "FP8 `save_pretrained` for MoE models and tokenizer class mismatch on save are separate save-path bugs." + "reason": "FP8 MoE save_pretrained behavior and tokenizer_class metadata mismatch are different save_pretrained issues." }, { "left": "issue:44075", "right": "issue:46231", "accept": false, - "reason": "Ignored SGD optimizer args is a runtime bug; the other is only a docstring parameter-name mismatch." + "reason": "SGD optimizer arguments being ignored is not the same as a documentation parameter-name typo." }, { "left": "issue:43116", "right": "issue:44190", "accept": false, - "reason": "Multi-label classification example returning empty results and local dataset loading failure are different scripts and issues." + "reason": "run_classification.py multi-label output bug and local dataset loading in image classification are unrelated example-script issues." }, { "left": "issue:45405", "right": "issue:45712", "accept": false, - "reason": "PEFT version bump and leftover dummy classes leaking into `dir(transformers)` are unrelated maintenance issues." + "reason": "PEFT version pinning and leftover dummy classes are completely different repository maintenance problems." }, { "left": "issue:44297", "right": "issue:45656", "accept": false, - "reason": "Tokenizer save metadata mismatch has nothing to do with deepspeed causing optimizer step calls." + "reason": "Tokenizer config save mismatch and double optimizer stepping are unrelated." }, { - "left": "issue:44933", - "right": "issue:45712", + "left": "issue:42831", + "right": "issue:44623", "accept": false, - "reason": "Missing import from `image_utils` and dummy class leakage are both import-surface issues but not the same root cause." + "reason": "FineGrainedFP8 accuracy issues and missing processor save files are different bugs." }, { - "left": "issue:42831", - "right": "issue:44623", + "left": "issue:44933", + "right": "issue:45712", "accept": false, - "reason": "FP8 accuracy regression and missing files from `processor.save_pretrained` are unrelated." + "reason": "Bad import from image_utils and stray dummy classes in dummy_pt_objects.py are unrelated cleanup/import issues." }, { "left": "issue:44448", "right": "issue:46132", "accept": false, - "reason": "Different outputs between v4/v5 for Pegasus and `AttentionInterface.register` changing behavior are separate model/API bugs." + "reason": "Different outputs across v4/v5 for Pegasus and AttentionInterface.register changing behavior are not the same fix." }, { - "left": "issue:44060", - "right": "issue:45125", + "left": "issue:44222", + "right": "issue:44623", "accept": false, - "reason": "Wrong tied-weights warning in Qwen3-Next and missing `_tp_plan` in Qwen3.5 MoE are unrelated." + "reason": "FP8 MoE save_pretrained bug and processor.save_pretrained missing files are different serialization problems." }, { - "left": "issue:44222", - "right": "issue:44623", + "left": "issue:44060", + "right": "issue:45125", "accept": false, - "reason": "FP8 MoE `save_pretrained` and processor file omissions are different persistence bugs." + "reason": "Wrong tied-weights warning in Qwen3-Next and missing _tp_plan in Qwen3_5Moe are unrelated Qwen model bugs." }, { "left": "issue:42796", "right": "issue:43723", "accept": false, - "reason": "BPE marker handling in a specific tokenizer and `AutoTokenizer.from_pretrained` loading failure in v5 are not the same issue." + "reason": "Tokenizer BPE-marker handling and AutoTokenizer.from_pretrained v5 loading are different tokenizer failures." }, { "left": "issue:43479", "right": "issue:43644", "accept": false, - "reason": "Incorrect default multimodal config initialization and junk-filled non-persistent buffers are unrelated." + "reason": "Default multimodal config initialization and non-persistent buffer junk filling are unrelated v5 issues." }, { "left": "issue:45245", "right": "issue:46132", "accept": false, - "reason": "Category-count runtime error and `AttentionInterface.register` behavior change are different code paths." + "reason": "Category cardinality limit and AttentionInterface.register behavior are not the same defect." }, { "left": "issue:44373", "right": "issue:44485", "accept": false, - "reason": "Wrong `position_ids` docstring and GLM-5 RoPE implementation discussion are not duplicates." + "reason": "Docstring typo for position_ids and GLM-5 RoPE implementation are unrelated." }, { "left": "issue:44360", "right": "issue:46762", "accept": false, - "reason": "DSA indexer missing ReLU and Minimax attention numerical differences are unrelated." + "reason": "DSA indexer lacking ReLU and Minimax attention numerical differences are different low-level math issues." }, { "left": "issue:41628", "right": "issue:42757", "accept": false, - "reason": "Missing `AutoImageProcessor` import and missing `is_offline_mode` import are distinct symbol-export regressions." - }, - { - "left": "issue:44297", - "right": "issue:45310", - "accept": false, - "reason": "Tokenizer config class mismatch and Qwen3.5 MoE from_pretrained failure are not the same concrete bug." + "reason": "Missing AutoImageProcessor import and missing is_offline_mode import are different API/export regressions." } ] }, @@ -15172,7 +15195,6 @@ "issue:42907", "issue:43122", "issue:43232", - "issue:43408", "issue:43475", "issue:43504", "issue:43723", @@ -15184,17 +15206,18 @@ "issue:44263", "issue:44279", "issue:44295", + "issue:44297", "issue:44360", "issue:44373", "issue:44484", "issue:44623", "issue:44829", + "issue:44964", "issue:44977", "issue:45071", "issue:45092", "issue:45310", "issue:45357", - "issue:45381", "issue:45834", "issue:46002", "issue:46050", @@ -15203,7 +15226,6 @@ "issue:46305", "issue:46328", "issue:46369", - "issue:46402", "issue:46424", "issue:46431", "issue:46614", @@ -15213,238 +15235,238 @@ "issue:47059" ], "soft_pairs": [ + "issue:44297|issue:45310", "issue:45071|issue:46082", "issue:43504|issue:47059", - "issue:44222|issue:44484", "issue:43723|issue:44977", + "issue:44222|issue:44484", "issue:43122|issue:46209", "issue:46614|issue:46932", "issue:44263|issue:45092", - "issue:46209|issue:46431", "issue:44295|issue:44623", + "issue:46209|issue:46431", "issue:43232|issue:46424", "issue:43232|issue:45834", "issue:44263|issue:44373", "issue:30990|issue:38175", "issue:44190|issue:46082", - "issue:46369|issue:46909", "issue:43475|issue:46305", "issue:44263|issue:44360", "issue:46305|issue:46655", + "issue:46369|issue:46909", "issue:42831|issue:45310", - "issue:42907|issue:46050", "issue:22355|issue:27957", + "issue:42907|issue:46050", "issue:44279|issue:46328", "issue:45092|issue:45357", "issue:42831|issue:44038", "issue:44829|issue:46002", - "issue:41628|issue:43824", "issue:44261|issue:45092", - "issue:45381|issue:46305", - "issue:43408|issue:46402" + "issue:41628|issue:43824", + "issue:44964|issue:47059" ], - "prepared_review_unit_hash": "1cf6e4eeedba9cca3798bc4f6e7c49036ad367442c603d82621b3700a4258e8d", + "prepared_review_unit_hash": "09b3f77fc2a6aaab6370321b916b1a50799ca25e1ba22d3d64d2ede9fe9d50c8", "component_budget": { "node_count": 514, "item_count": 514, "soft_pair_count": 1891, - "serialized_chars": 980926, - "estimated_input_tokens": 245232, - "estimated_eval_tokens": 490720 + "serialized_chars": 980895, + "estimated_input_tokens": 245224, + "estimated_eval_tokens": 490704 }, "budget": { - "node_count": 48, - "item_count": 48, + "node_count": 47, + "item_count": 47, "soft_pair_count": 29, - "serialized_chars": 36014, - "estimated_input_tokens": 9004, - "estimated_eval_tokens": 18264 + "serialized_chars": 35451, + "estimated_input_tokens": 8863, + "estimated_eval_tokens": 17982 }, "overflow_policy": "truncate_then_skip", "trimmed": false, "aggressively_trimmed": false, "split": true, "analyst_result": { - "summary": "The cluster is highly heterogeneous: the soft pairs span unrelated loading, config, quantization, generation, vision, tokenizer, and CLI problems. None of the candidate edges look like the same underlying bug/change, so I would not merge any of them or pick a single canonical item for the whole set.", - "confidence": 0.93, - "canonical_issue_reason": "There is no single underlying issue shared by this set; the items cover many unrelated failure modes across different subsystems.", + "summary": "The set is heterogeneous: most pairs share only a broad transformers-v5 / model-loading / serialization theme, but not the same concrete bug or fix path. I would not merge any of the soft-edge candidates.", + "confidence": 0.72, + "canonical_issue_reason": "If one issue must represent the cluster, issue 45092 is the best fit: it is a concrete, user-facing v5 compatibility regression with a clear failure mode (remote-code checkpoints and meta initialization) and stronger signal than the narrower or unrelated reports.", "canonical_pr_reason": null, - "best_issue_reason": "No issue is a good global representative because the cluster is not coherent and does not center on one bug.", + "best_issue_reason": "Issue 45092 is the strongest canonical candidate because it is specific, reproducible, and central to the recurring v5 loading-compatibility theme; the other issues are either narrower symptoms or clearly orthogonal.", "best_pr_reason": null, "soft_edge_verdicts": [ + { + "left": "issue:44297", + "right": "issue:45310", + "accept": false, + "reason": "Both involve model/tokenizer loading, but one is a tokenizer_class save-pretrained mismatch and the other is a Qwen3.5 MoE from_pretrained failure; different bug surfaces and likely different fixes." + }, { "left": "issue:45071", "right": "issue:46082", "accept": false, - "reason": "Both involve config handling, but one is a v5 type-checking regression and the other is a Llama head_dim validation edge case." + "reason": "These are different config-validation regressions: PretrainedConfig type checking versus LlamaConfig head_dim handling." }, { "left": "issue:43504", "right": "issue:47059", "accept": false, - "reason": "Legacy preset loading for Beit is unrelated to a CLI failure caused by huggingface_hub compatibility." + "reason": "BEiT pretrained loading with a legacy field and a CLI completion error from huggingface_hub are unrelated." }, { - "left": "issue:44222", - "right": "issue:44484", + "left": "issue:43723", + "right": "issue:44977", "accept": false, - "reason": "One is an FP8 save_pretrained bug; the other asks about the default max_shard_size value." + "reason": "Generic AutoTokenizer loading in v5 and Qwen3.5 flash-attention generation issues are different code paths." }, { - "left": "issue:43723", - "right": "issue:44977", + "left": "issue:44222", + "right": "issue:44484", "accept": false, - "reason": "Tokenizer import/load failure is unrelated to a flash-attention generation problem in Qwen3.5." + "reason": "FP8 save_pretrained for MoE and the default max_shard_size question are not the same bug." }, { "left": "issue:43122", "right": "issue:46209", "accept": false, - "reason": "Tokenization changes across versions are unrelated to FP8 inference failing on B200." + "reason": "Tokenizer behavior regression and FP8 inference failure on B200 are unrelated." }, { "left": "issue:46614", "right": "issue:46932", "accept": false, - "reason": "A stale phonemizer backend bug is unrelated to a Gemma4 dtype-casting fix." + "reason": "Stale phonemizer language state and Gemma4 dtype casting for quantized embedders are different subsystems." }, { "left": "issue:44263", "right": "issue:45092", "accept": false, - "reason": "DSA indexer split semantics are unrelated to remote-code checkpoint meta-init compatibility." + "reason": "torch.split return semantics in a DSA indexer and remote-code checkpoint meta-init incompatibility are not the same concrete bug." }, { - "left": "issue:46209", - "right": "issue:46431", + "left": "issue:44295", + "right": "issue:44623", "accept": false, - "reason": "FP8 inference on B200 is unrelated to processor chat-template handling." + "reason": "A position_ids buffer read error and missing files from processor.save_pretrained are different save/load problems." }, { - "left": "issue:44295", - "right": "issue:44623", + "left": "issue:46209", + "right": "issue:46431", "accept": false, - "reason": "Reading position_ids after buffer registration is unrelated to missing files in processor.save_pretrained." + "reason": "FP8 inference on B200 and apply_chat_template failing because a processor has no chat template are unrelated." }, { "left": "issue:43232", "right": "issue:46424", "accept": false, - "reason": "Generation kwargs sync_gpus behavior is unrelated to cache max_cache_length semantics." + "reason": "Generation kwargs after sync_gpus and max_cache_length honoring only quantized cache are separate generation/cache behaviors." }, { "left": "issue:43232", "right": "issue:45834", "accept": false, - "reason": "Generation kwargs handling is unrelated to Kosmos2.5 OCR index errors." + "reason": "Generation kwargs handling and a Kosmos2.5 OCR index error are unrelated." }, { "left": "issue:44263", "right": "issue:44373", "accept": false, - "reason": "A torch.split return-value issue in DSA indexing is unrelated to a docstring correction." + "reason": "A torch.split / DSA indexer issue and a wrong docstring for position_ids are not duplicates." }, { "left": "issue:30990", "right": "issue:38175", "accept": false, - "reason": "Sentence Transformers loading stalls and zero probabilities in SigLIP2 are different bugs." + "reason": "Sentence Transformers loading stuck and zero probabilities in siglip2 are different bugs." }, { "left": "issue:44190", "right": "issue:46082", "accept": false, - "reason": "Local dataset loading in an example script is unrelated to a LlamaConfig validation error." - }, - { - "left": "issue:46369", - "right": "issue:46909", - "accept": false, - "reason": "SAM3 LiteText memory overflow is unrelated to a broken documentation example." + "reason": "A local dataset loading failure and a LlamaConfig head_dim validation issue are unrelated." }, { "left": "issue:43475", "right": "issue:46305", "accept": false, - "reason": "SAM3 vision output attribute errors are unrelated to a MoonshineStreaming off-by-one mask bug." + "reason": "SAM3VisionEncoderOutput missing an attribute and a streaming mask off-by-one are different code paths." }, { "left": "issue:44263", "right": "issue:44360", "accept": false, - "reason": "Both mention the DSA indexer, but they describe different defects: split outputs vs missing ReLU." + "reason": "Both touch the DSA indexer, but one is about torch.split outputs and the other about a missing ReLU; too different to be the same bug." }, { "left": "issue:46305", "right": "issue:46655", "accept": false, - "reason": "Two padding/masking issues in different models, but they affect different processors and code paths." + "reason": "A sliding-window off-by-one in streaming and incorrect video padding calculation are distinct." }, { - "left": "issue:42831", - "right": "issue:45310", + "left": "issue:46369", + "right": "issue:46909", "accept": false, - "reason": "FP8 accuracy issues are unrelated to Qwen3.5 Moe from_pretrained errors." + "reason": "Memory overflow in sam3-litetext and a broken DIA documentation example are unrelated." }, { - "left": "issue:42907", - "right": "issue:46050", + "left": "issue:42831", + "right": "issue:45310", "accept": false, - "reason": "Saving dequantized Ministrals/Devstrals is unrelated to sinq quantization round-trip breakage." + "reason": "FineGrainedFP8 accuracy loss and Qwen3.5 Moe from_pretrained errors are not the same bug." }, { "left": "issue:22355", "right": "issue:27957", "accept": false, - "reason": "Missing transformers.onnx module import and XLMRoberta Flash Attention 2 are unrelated issues." + "reason": "Missing transformers.onnx module and XLMRoberta flash-attention behavior are unrelated." + }, + { + "left": "issue:42907", + "right": "issue:46050", + "accept": false, + "reason": "Dequantized model save failure and sinq save/load breakage are both serialization-related but not the same concrete code-path problem." }, { "left": "issue:44279", "right": "issue:46328", "accept": false, - "reason": "A dependency problem is unrelated to SequenceFeatureExtractor.pad inefficiency." + "reason": "A generic dependency issue and a SequenceFeatureExtractor.pad performance issue are unrelated." }, { "left": "issue:45092", "right": "issue:45357", "accept": false, - "reason": "Both touch Transformers v5 / Qwen-family compatibility, but one is checkpoint meta-init loading and the other is save_pretrained key mapping." + "reason": "Old InternVL2 meta-init incompatibility and Qwen3.5 incorrect visual encoder keys in save_pretrained are separate model-specific regressions." }, { - "left": "issue:42831", + "left": "issue:43824", "right": "issue:44038", "accept": false, - "reason": "Both are FP8-related, but one concerns accuracy and the other a specific Qwen3-VL-Moe bug." + "reason": "Missing import for Qwen2_5_VLForConditionalGeneration and a Qwen3-VL-Moe bug are different failures." }, { "left": "issue:44829", "right": "issue:46002", "accept": false, - "reason": "Flash Attention 3 training degeneration is unrelated to NaNs appearing in classification heads during checkout/init." - }, - { - "left": "issue:41628", - "right": "issue:43824", - "accept": false, - "reason": "AutoImageProcessor import failure and Qwen2.5-VL import failure are different missing-symbol issues." + "reason": "Degenerate training with flash_attention_3 and NaNs in classification heads upon checkout/init are different symptoms with different likely causes." }, { "left": "issue:44261", "right": "issue:45092", "accept": false, - "reason": "MLA rms_norm_eps precision error is unrelated to InternVL2 remote-code checkpoint meta initialization." + "reason": "Missing rms_norm_eps in MLA q_a_layernorm and remote-code checkpoint meta-init incompatibility are unrelated." }, { - "left": "issue:45381", - "right": "issue:46305", + "left": "issue:41628", + "right": "issue:43824", "accept": false, - "reason": "Qwen2.5-VL video vision_position_ids mismatch is unrelated to MoonshineStreaming masking." + "reason": "Both are import errors, but they concern different symbols and likely different export/registration causes." }, { - "left": "issue:43408", - "right": "issue:46402", + "left": "issue:44964", + "right": "issue:47059", "accept": false, - "reason": "A SAM3 tracker/model-type warning is unrelated to SAM3 model text encoder weights not being loaded." + "reason": "Phi-4 multimodal loading on latest transformers and a hub CLI completion attribute error are unrelated." } ] }, @@ -15513,27 +15535,29 @@ "issue:46172", "issue:46305", "issue:46369", + "issue:46402", "issue:46909", "issue:47059" ], "soft_pairs": [ - "issue:44964|issue:47059", - "issue:39379|issue:46172", + "issue:43408|issue:46402", + "issue:45381|issue:46305", "issue:44483|issue:45406", - "issue:43644|issue:46305", + "issue:39379|issue:46172", "issue:42831|issue:45753", + "issue:43644|issue:46305", "issue:22355|issue:28218", "issue:44327|issue:47059", "issue:39401|issue:44568", - "issue:42757|issue:45278", "issue:44188|issue:44373", + "issue:42757|issue:45278", "issue:43408|issue:46369", - "issue:42886|issue:43066", "issue:42886|issue:42994", + "issue:42886|issue:43066", "issue:27957|issue:30990", "issue:42796|issue:43122", - "issue:45381|issue:46909", "issue:43232|issue:45593", + "issue:45381|issue:46909", "issue:43540|issue:46909", "issue:43646|issue:46002", "issue:42907|issue:44945", @@ -15542,206 +15566,212 @@ "issue:43576|issue:43784", "issue:43976|issue:44945", "issue:42907|issue:44964", - "issue:41628|issue:45278", "issue:42371|issue:44263", + "issue:41628|issue:45278", "issue:43329|issue:47059" ], - "prepared_review_unit_hash": "6fd39962edecf0261058641c87fdc86b1b7d0268a5fe157acb96a93e50a507c0", + "prepared_review_unit_hash": "ba232d7ba95fd0c34d39ffa24f963dfa433f5919393bafab133241ee68c5ad90", "component_budget": { "node_count": 514, "item_count": 514, "soft_pair_count": 1891, - "serialized_chars": 980926, - "estimated_input_tokens": 245232, - "estimated_eval_tokens": 490720 + "serialized_chars": 980895, + "estimated_input_tokens": 245224, + "estimated_eval_tokens": 490704 }, "budget": { - "node_count": 47, - "item_count": 47, - "soft_pair_count": 28, - "serialized_chars": 35009, - "estimated_input_tokens": 8753, - "estimated_eval_tokens": 17762 + "node_count": 48, + "item_count": 48, + "soft_pair_count": 29, + "serialized_chars": 35873, + "estimated_input_tokens": 8969, + "estimated_eval_tokens": 18194 }, "overflow_policy": "truncate_then_skip", "trimmed": false, "aggressively_trimmed": false, "split": true, "analyst_result": { - "summary": "The cluster is a loose collection of mostly unrelated issues across tokenizer behavior, model loading, multimodal inference, import regressions, and CLI/API breakages. None of the soft pairs look like the same underlying bug or change, so I reject them all as duplicates.", - "confidence": 0.47, - "canonical_issue_reason": "Issue 45278 is the broadest representative item in this set: it describes widespread import breakage after a Transformers version upgrade, which is a recurring theme among several nearby issues in the cluster.", + "summary": "The cluster is a grab bag of unrelated bug reports. None of the soft-similarity pairs look like the same underlying bug or change, so all soft edges are rejected.", + "confidence": 0.91, + "canonical_issue_reason": "issue:45278 is the broadest, clearest regression report in the set: multiple import errors after upgrading from 4.57.0 to 5.5.0. It is more representative than the narrower, single-symbol import bugs.", "canonical_pr_reason": null, - "best_issue_reason": "Issue 45278 is the best single issue to anchor the cluster because it is high-scope, clearly user-facing, and closest to the repeated import/regression failures seen elsewhere in the set.", + "best_issue_reason": "issue:45278 is the best global issue candidate because it describes a concrete, user-facing regression with a clear version jump and broad impact, making it a stronger canonical representative than the other isolated reports.", "best_pr_reason": null, "soft_edge_verdicts": [ { - "left": "issue:44964", - "right": "issue:47059", + "left": "issue:43408", + "right": "issue:46402", "accept": false, - "reason": "Different failures: multimodal model loading vs a CLI completion attribute error from huggingface_hub." + "reason": "Different problems: a sam3_model/sam3_tracker type mismatch warning vs text encoder weights not loading in Sam3Model." }, { - "left": "issue:39379", - "right": "issue:46172", + "left": "issue:45381", + "right": "issue:46305", "accept": false, - "reason": "Different code paths and symptoms: mllama mask handling vs a DeepSeek tensor transpose bug." + "reason": "Both involve vision/video, but one is Qwen2.5-VL position IDs and the other is a MoonshineStreaming sliding-window off-by-one; different code paths and fixes." }, { "left": "issue:44483", "right": "issue:45406", "accept": false, - "reason": "One is an API request acceptance problem; the other is a serve-time processor attribute crash." + "reason": "API request handling for /v1/chat/completions is unrelated to Gemma4Processor missing _tokenizer." }, { - "left": "issue:43644", - "right": "issue:46305", + "left": "issue:39379", + "right": "issue:46172", "accept": false, - "reason": "Non-persistent buffer initialization corruption is unrelated to a streaming window off-by-one mask bug." + "reason": "mllama masked-out-mask handling and deepseek-v4 tensor transpose are unrelated implementation bugs." }, { "left": "issue:42831", "right": "issue:45753", "accept": false, - "reason": "Accuracy loss in FineGrainedFP8 is not the same as an infinite-loop bug on a specific image." + "reason": "FineGrainedFP8 accuracy regression and a Qwen3_5 image infinite loop are not the same bug." + }, + { + "left": "issue:43644", + "right": "issue:46305", + "accept": false, + "reason": "Non-persistent buffer junk in v5.0.0 is unrelated to a sliding_window_mask off-by-one in MoonshineStreaming." }, { "left": "issue:22355", "right": "issue:28218", "accept": false, - "reason": "Import failure for transformers.onnx is unrelated to tokenizer spacing after added tokens." + "reason": "Missing transformers.onnx module and tokenizer adds an extra space after added tokens are different issues." }, { "left": "issue:44327", "right": "issue:47059", "accept": false, - "reason": "QA decode_spans top-k crash and a CLI completion attribute error are unrelated." + "reason": "QA decode_spans top-k bounds error and a broken CLI completion attribute are unrelated failures." }, { "left": "issue:39401", "right": "issue:44568", "accept": false, - "reason": "Both are tokenizer issues, but they affect different models and different behaviors (offset mapping vs BOS/EOS insertion)." + "reason": "Qwen3 tokenizer offset_mapping is a tokenizer alignment bug; add_special_tokens not adding BOS/EOS is a different tokenizer behavior regression." }, { - "left": "issue:42757", - "right": "issue:45278", + "left": "issue:44188", + "right": "issue:44373", "accept": false, - "reason": "Both mention imports, but one is a specific missing huggingface_hub symbol while the other is a broad upgrade regression with many import errors." + "reason": "Diverging attention kernels under torch.compile and a wrong docstring are not the same code-path problem." }, { - "left": "issue:44188", - "right": "issue:44373", + "left": "issue:42757", + "right": "issue:45278", "accept": false, - "reason": "A torch.compile attention-kernel divergence bug is unrelated to a docstring issue." + "reason": "Both are import failures, but one is missing is_offline_mode from huggingface_hub while the other is many import errors after a version upgrade; too broad and not the same root cause." }, { "left": "issue:43408", "right": "issue:46369", "accept": false, - "reason": "These involve different SAM3 model variants and different failures: type warning vs memory overflow." + "reason": "A model-type warning and a memory overflow in sam3-litetext are different symptoms and likely different fixes." }, { "left": "issue:42886", - "right": "issue:43066", + "right": "issue:42994", "accept": false, - "reason": "Offline tokenizer cache loading and wrong tokenizer decoder type are different bugs." + "reason": "Offline tokenizer cache loading and quantized model saving are unrelated." }, { "left": "issue:42886", - "right": "issue:42994", + "right": "issue:43066", "accept": false, - "reason": "Tokenizer cache/offline loading is not the same as quantized model saving failure." + "reason": "Tokenizer cache/offline loading and wrong tokenizer decoder type in v5 are distinct tokenizer issues." }, { "left": "issue:27957", "right": "issue:30990", "accept": false, - "reason": "Flash Attention support for XLM-Roberta is unrelated to Sentence Transformers loading hangs." + "reason": "Flash Attention 2 support for XLMRoberta and Sentence Transformers getting stuck loading are unrelated." }, { "left": "issue:42796", "right": "issue:43122", "accept": false, - "reason": "Model-specific BPE marker handling is not the same as a broader version-to-version tokenization difference." + "reason": "Both concern tokenization, but one is Ministral BPE marker handling and the other is a version-to-version tokenization change; different models and likely different causes." }, { - "left": "issue:45381", - "right": "issue:46909", + "left": "issue:43232", + "right": "issue:45593", "accept": false, - "reason": "A video position_ids bug in Qwen2.5-VL is unrelated to a broken DIA documentation example." + "reason": "Generation kwargs after sync_gpus and D-FINE auxiliary-loss behavior are unrelated." }, { - "left": "issue:43232", - "right": "issue:45593", + "left": "issue:45381", + "right": "issue:46909", "accept": false, - "reason": "Generation kwargs handling after sync_gpus and D-FINE auxiliary-loss behavior are unrelated subsystems." + "reason": "Qwen2.5-VL video position IDs and a broken DIA documentation example are not the same bug." }, { "left": "issue:43540", "right": "issue:46909", "accept": false, - "reason": "A Qwen3OmniMoe video-input ValueError is unrelated to a documentation example issue." + "reason": "Qwen3OmniMoe video input ValueError and a documentation example issue are unrelated." }, { "left": "issue:43646", "right": "issue:46002", "accept": false, - "reason": "Custom model initialization breakage and NaNs in classification heads are different initialization-time problems." + "reason": "Custom model initialization breakage and NaNs in classification heads are different failures." }, { "left": "issue:42907", "right": "issue:44945", "accept": false, - "reason": "Quantized model save failures are unrelated to incorrect outputs under pipeline parallelism." + "reason": "Saving dequantized Ministrals and incorrect pipeline-parallel LLM output do not look like the same underlying bug." }, { "left": "issue:44295", "right": "issue:44297", "accept": false, - "reason": "Reading a buffer-marked position_ids tensor and tokenizer_class metadata mismatches are different serialization issues." + "reason": "Reading a buffer-backed position_ids tensor and tokenizer_config tokenizer_class mismatch are separate serialization issues." }, { "left": "issue:30990", "right": "issue:33357", "accept": false, - "reason": "Both are load-time failures, but for different models and different root causes (hang vs bus error)." + "reason": "Sentence Transformers loading hang and a MacOS bus error with a CLIP model are different runtime failures." }, { "left": "issue:43576", "right": "issue:43784", "accept": false, - "reason": "A broken transformers env CLI command is unrelated to a NameError when importing sentence-transformers." + "reason": "Broken transformers env CLI and sentence-transformers import NameError are unrelated." }, { "left": "issue:43976", "right": "issue:44945", "accept": false, - "reason": "Python version compatibility regressions are not the same as incorrect LLM output under pipeline parallelism." + "reason": "Python version support regression and pipeline-parallel output corruption are not the same issue." }, { "left": "issue:42907", "right": "issue:44964", "accept": false, - "reason": "Saving dequantized weights and loading a multimodal model are different operations and failure modes." + "reason": "Saving dequantized models and loading Phi-4 multimodal are different model workflow failures." }, { - "left": "issue:41628", - "right": "issue:45278", + "left": "issue:42371", + "right": "issue:44263", "accept": false, - "reason": "Missing AutoImageProcessor is a specific import symbol issue; 45278 is a broad collection of import errors after upgrade, not the same concrete bug." + "reason": "TF32 configuration guidance and a torch.split return-value issue are unrelated." }, { - "left": "issue:42371", - "right": "issue:44263", + "left": "issue:41628", + "right": "issue:45278", "accept": false, - "reason": "TF32 settings guidance and a torch.split return-value bug are unrelated." + "reason": "Cannot import AutoImageProcessor and many import errors after upgrade share a broad theme, but they are different concrete import breakages with different likely fixes." }, { "left": "issue:43329", "right": "issue:47059", "accept": false, - "reason": "A multimodal token counting video-branch bug is unrelated to a CLI completion attribute regression." + "reason": "Undefined variables in multimodal token counting and a CLI completion attribute error are unrelated." } ] }, @@ -15815,23 +15845,23 @@ "issue:46762" ], "soft_pairs": [ - "issue:44117|issue:44704", "issue:43756|issue:44060", + "issue:44117|issue:44704", "issue:42907|issue:43937", "issue:43784|issue:45600", "issue:39401|issue:44488", "issue:39401|issue:44521", "issue:44625|issue:44829", - "issue:44704|issue:45290", "issue:46172|issue:46762", "issue:44360|issue:44485", + "issue:44704|issue:45290", "issue:43122|issue:46326", "issue:43408|issue:43475", "issue:43122|issue:44521", "issue:44038|issue:44297", "issue:42907|issue:43452", - "issue:22355|issue:29159", "issue:44038|issue:44295", + "issue:22355|issue:29159", "issue:44038|issue:44222", "issue:45216|issue:45478", "issue:43824|issue:44351", @@ -15846,214 +15876,214 @@ "issue:39379|issue:43232", "issue:43493|issue:45753" ], - "prepared_review_unit_hash": "3bc7e524c0629d0c67929961e95838beb341425407e67c0498893b68566aaa97", + "prepared_review_unit_hash": "6580a57c72761926cfc74bd6ab70357e21e3a4b015a772379b90232eb1bfe04b", "component_budget": { "node_count": 514, "item_count": 514, "soft_pair_count": 1891, - "serialized_chars": 980926, - "estimated_input_tokens": 245232, - "estimated_eval_tokens": 490720 + "serialized_chars": 980895, + "estimated_input_tokens": 245224, + "estimated_eval_tokens": 490704 }, "budget": { "node_count": 48, "item_count": 48, "soft_pair_count": 30, - "serialized_chars": 36442, - "estimated_input_tokens": 9111, - "estimated_eval_tokens": 18478 + "serialized_chars": 36431, + "estimated_input_tokens": 9108, + "estimated_eval_tokens": 18472 }, "overflow_policy": "truncate_then_skip", "trimmed": false, "aggressively_trimmed": false, "split": true, "analyst_result": { - "summary": "The pairs are mostly same-subsystem but not the same underlying bug; I would reject all soft edges. For a canonical issue, the clearest and most self-contained bug report is the Deepseek tokenizer regression.", - "confidence": 0.91, - "canonical_issue_reason": "issue:44779 is a concrete, versioned regression report with a specific model and observable behavior, making it a strong anchor issue for triage.", + "summary": "The cluster is highly heterogeneous: most items are unrelated model-specific bugs, tokenizer regressions, import errors, or generation issues. None of the soft pairs look like the same underlying defect, so I would not merge any of them as duplicates.", + "confidence": 0.95, + "canonical_issue_reason": "No single canonical issue stands out; the items span unrelated bug families across tokenizers, loaders, multimodal handling, generation, and model-specific regressions.", "canonical_pr_reason": null, - "best_issue_reason": "issue:44779 is the most actionable issue here: it is narrowly scoped, reproducible, and clearly states the regression boundary (works in v4, breaks in v5).", + "best_issue_reason": "No issue is globally suitable as a canonical duplicate representative for this cluster because it is not one coherent bug family.", "best_pr_reason": null, "soft_edge_verdicts": [ { - "left": "issue:44117", - "right": "issue:44704", + "left": "issue:43756", + "right": "issue:44060", "accept": false, - "reason": "Different failure modes: tokenizer mapping fallback vs AutoProcessor not forwarding kwargs." + "reason": "Different problems: Smollm3 RoPE layer selection vs Qwen3-Next tied-weights warning." }, { - "left": "issue:43756", - "right": "issue:44060", + "left": "issue:44117", + "right": "issue:44704", "accept": false, - "reason": "Both touch model internals, but one is RoPE layer composition and the other is a tied-weights warning bug." + "reason": "Tokenizer mapping fallback and AutoProcessor kwargs forwarding are different code paths." }, { "left": "issue:42907", "right": "issue:43937", "accept": false, - "reason": "Distinct code paths: dequantized save_pretrained vs invalid GenerationConfig on GLM-5." + "reason": "Model save/dequantization bug vs invalid GenerationConfig validation." }, { "left": "issue:43784", "right": "issue:45600", "accept": false, - "reason": "Both are import/compatibility issues, but one is a missing nn import and the other is stale auto_mappings references." + "reason": "Missing nn import in sentence-transformers integration vs stale auto_mappings references." }, { "left": "issue:39401", "right": "issue:44488", "accept": false, - "reason": "Tokenizer offset_mapping bug vs model loading failure for a specific checkpoint." + "reason": "Qwen3 offset_mapping bug vs failure to load a specific model repository." }, { "left": "issue:39401", "right": "issue:44521", "accept": false, - "reason": "Tokenizer offset_mapping issue and multimodal assistant_masks issue are separate bugs." + "reason": "Tokenizer offset mapping issue and multimodal assistant_mask bug are different defects." }, { "left": "issue:44625", "right": "issue:44829", "accept": false, - "reason": "num_labels propagation and flash_attention_3 training degeneration are unrelated." - }, - { - "left": "issue:44704", - "right": "issue:45290", - "accept": false, - "reason": "Different API surfaces: cached_file kwargs forwarding vs apply_chat_template crash with tool calls." + "reason": "num_labels propagation mismatch vs flash_attention_3 training degeneration." }, { "left": "issue:46172", "right": "issue:46762", "accept": false, - "reason": "Both concern numerical behavior, but they refer to different models and different implementations." + "reason": "DeepSeek-V4 dtype/transpose bug is unrelated to Minimax attention numerical differences." }, { "left": "issue:44360", "right": "issue:44485", "accept": false, - "reason": "One is about a DSA indexer activation, the other about GLM-5 RoPE implementation." + "reason": "DSA indexer activation discussion is not the same as GLM-5 RoPE implementation." + }, + { + "left": "issue:44704", + "right": "issue:45290", + "accept": false, + "reason": "Processor file-resolution kwargs vs chat-template tool-call crash are different failures." }, { "left": "issue:43122", "right": "issue:46326", "accept": false, - "reason": "Both are tokenizer-related, but one is a v4-to-v5 regression and the other is continue_final_message for mistral_common tokenizers." + "reason": "Different tokenizer bugs in different subsystems; not the same underlying change." }, { "left": "issue:43408", "right": "issue:43475", "accept": false, - "reason": "Same model family, but a warning about model type mismatch is not the same as a missing output attribute error." - }, - { - "left": "issue:43122", - "right": "issue:44521", - "accept": false, - "reason": "Tokenization regression and multimodal chat-template assistant mask bug are different behaviors." + "reason": "Model-type warning on SAM3 tracker vs missing fpn_position_embeddings attribute." }, { "left": "issue:44038", "right": "issue:44297", "accept": false, - "reason": "Qwen3-VL-Moe import/breakage vs tokenizer save_pretrained metadata mismatch." + "reason": "Qwen3-VL-Moe loading bug vs tokenizer save_pretrained metadata mismatch." }, { "left": "issue:42907", "right": "issue:43452", "accept": false, - "reason": "Dequantized saving problem is unrelated to gguf_file breaking model/tokenizer loading." + "reason": "Saving a dequantized model and gguf_file load handling are distinct issues." }, { - "left": "issue:22355", - "right": "issue:29159", + "left": "issue:44038", + "right": "issue:44295", "accept": false, - "reason": "Missing transformers.onnx module import vs slow/fast tokenizer inconsistency are unrelated." + "reason": "Qwen3-VL-Moe regression vs position_ids buffer read error; different code paths." }, { - "left": "issue:44038", - "right": "issue:44295", + "left": "issue:22355", + "right": "issue:29159", "accept": false, - "reason": "Qwen3-VL-Moe issue vs reading position_ids after buffer registration are different bugs." + "reason": "Importing transformers.onnx and slow/fast tokenizer inconsistency are unrelated." }, { "left": "issue:44038", "right": "issue:44222", "accept": false, - "reason": "Different failures: model import/config issue vs FP8 save_pretrained for moe." + "reason": "Qwen3-VL-Moe load bug and FP8 save_pretrained bug do not match." }, { "left": "issue:45216", "right": "issue:45478", "accept": false, - "reason": "Both involve Qwen3.5, but one is save_pretrained regression and the other is from_pretrained loading error." + "reason": "Both involve Qwen3.5, but one is save_pretrained corruption and the other is from_pretrained failure; not the same fix." }, { "left": "issue:43824", "right": "issue:44351", "accept": false, - "reason": "Both are import errors, but for different symbols and different failure causes." + "reason": "Missing Qwen2_5_VLForConditionalGeneration export vs missing HybridCache export." }, { "left": "issue:43116", "right": "issue:45200", "accept": false, - "reason": "Example-script multi-label output bug vs Gemma 4 mm_token_type_ids defaulting issue." + "reason": "Example script multi-label output bug vs Gemma 4 mm_token_type_ids defaulting issue." }, { "left": "issue:39401", "right": "issue:46614", "accept": false, - "reason": "Different tokenizer implementations and different bugs: offset mapping vs stale phonemizer language state." + "reason": "Tokenizer offset mapping bug vs stale phonemizer language backend state." }, { "left": "issue:43122", "right": "issue:44779", "accept": false, - "reason": "Both are tokenizer regressions, but they affect different models and different behaviors." + "reason": "Different model tokenization regressions; not one concrete bug." }, { "left": "issue:42757", "right": "issue:44351", "accept": false, - "reason": "Missing huggingface_hub symbol and missing transformers symbol are separate import regressions." + "reason": "Different missing symbols from different packages/modules; too broad to deduplicate." }, { "left": "issue:36296", "right": "issue:43116", "accept": false, - "reason": "Tensor-parallel training bug is unrelated to the classification example returning empty results." + "reason": "Tensor parallel training bug is unrelated to the classification example returning empty results." }, { "left": "issue:43278", "right": "issue:44945", "accept": false, - "reason": "Dtype drift between train/eval and pipeline-parallel output corruption are different code paths." + "reason": "BF16->FP32 dtype drift in eval and pipeline-parallel output corruption are different defects." }, { "left": "issue:22355", "right": "issue:29127", "accept": false, - "reason": "Module import failure and LayoutLMv3 error-message clarity are unrelated." + "reason": "Import failure for transformers.onnx vs unclear LayoutLMv3 error messaging." }, { "left": "issue:42175", "right": "issue:44938", "accept": false, - "reason": "Backend packaging omission and Python 3.14 load failure are different environment issues." + "reason": "Missing tensorflow backend in extras vs Python 3.14 load failure are not the same bug." }, { "left": "issue:39379", "right": "issue:43232", "accept": false, - "reason": "mllama mask handling bug vs generation kwargs update after sync_gpus are unrelated." + "reason": "mllama mask handling and generation kwargs after sync_gpus are unrelated." }, { "left": "issue:43493", "right": "issue:45753", "accept": false, - "reason": "Different model-specific problems: SigLIP2 numerical mismatch vs Qwen3_5 image loop." + "reason": "SigLIP2 implementation discrepancy vs Qwen3_5 image-specific infinite loop." + }, + { + "left": "issue:44360", + "right": "issue:44485", + "accept": false, + "reason": "DSA indexer ReLU discussion does not match GLM-5 RoPE implementation." } ] }, @@ -16126,8 +16156,8 @@ "issue:47116" ], "soft_pairs": [ - "issue:46231|issue:46877", "issue:44188|issue:46328", + "issue:46231|issue:46877", "issue:43937|issue:44485", "issue:42898|issue:46614", "issue:38617|issue:45278", @@ -16140,217 +16170,217 @@ "issue:45357|issue:45478", "issue:44625|issue:46002", "issue:43116|issue:43493", - "issue:44610|issue:46909", "issue:45083|issue:45446", + "issue:44610|issue:46909", "issue:42371|issue:43874", "issue:44075|issue:44936", - "issue:46726|issue:46981", "issue:43618|issue:46829", + "issue:46726|issue:46981", "issue:46328|issue:47116", "issue:43992|issue:44704", "issue:43066|issue:46710", - "issue:43116|issue:43825", "issue:28348|issue:46640", - "issue:42673|issue:42796", + "issue:43116|issue:43825", "issue:43827|issue:45446", + "issue:42673|issue:42796", "issue:44995|issue:46697" ], - "prepared_review_unit_hash": "a4ae4b367ec558c0d153283c9e5b9ce2eeaa4fdd7e0181696d9e636e03316ddc", + "prepared_review_unit_hash": "14ef2d1104fc5d22140662a67451cc7d4e82f36f2d1d4cafdacaffed97f927aa", "component_budget": { "node_count": 514, "item_count": 514, "soft_pair_count": 1891, - "serialized_chars": 980926, - "estimated_input_tokens": 245232, - "estimated_eval_tokens": 490720 + "serialized_chars": 980895, + "estimated_input_tokens": 245224, + "estimated_eval_tokens": 490704 }, "budget": { "node_count": 47, "item_count": 47, "soft_pair_count": 28, - "serialized_chars": 35238, - "estimated_input_tokens": 8810, - "estimated_eval_tokens": 17876 + "serialized_chars": 35236, + "estimated_input_tokens": 8809, + "estimated_eval_tokens": 17874 }, "overflow_policy": "truncate_then_skip", "trimmed": false, "aggressively_trimmed": false, "split": true, "analyst_result": { - "summary": "The cluster is heterogeneous: most pairs are only loosely related by subsystem or model family, not the same underlying bug/change. I rejected all soft edges and found no safe canonical issue/PR representative.", - "confidence": 0.91, - "canonical_issue_reason": "No canonical issue: the items span unrelated regressions across tokenizers, model loading, vision/audio processors, docs, and runtime/version checks, so no single issue safely represents the cluster.", + "summary": "This cluster is very heterogeneous: most items are unrelated issues that only share broad Transformers/v5/model-family vocabulary. None of the soft pairs look like true duplicates of the same concrete bug or change, so all soft edges should be rejected.", + "confidence": 0.34, + "canonical_issue_reason": "Issue 45278 is the best umbrella representative because it captures the broad v5 regression/import-breakage theme that appears most often across the set, even though the cluster is not a true duplicate group.", "canonical_pr_reason": null, - "best_issue_reason": "No single issue is a reliable cluster representative; several candidates are superficially broad (for example, version/import regressions), but none clearly subsume the others.", + "best_issue_reason": "Issue 45278 is the most representative global issue candidate due to its broad cross-version breakage framing and overlap with several v5-related regression reports.", "best_pr_reason": null, "soft_edge_verdicts": [ { - "left": "issue:46231", - "right": "issue:46877", + "left": "issue:44188", + "right": "issue:46328", "accept": false, - "reason": "One is a docs parameter-name mismatch, the other is import-time doc-lint stdout noise; unrelated bugs." + "reason": "Different bugs: attention-kernel divergence under torch.compile vs a SequenceFeatureExtractor padding performance issue." }, { - "left": "issue:44188", - "right": "issue:46328", + "left": "issue:46231", + "right": "issue:46877", "accept": false, - "reason": "Attention-kernel branching under torch.compile is unrelated to a padding performance issue in SequenceFeatureExtractor." + "reason": "One is a docs parameter name mismatch; the other is import-time docstring lint output. No shared code-path bug." }, { "left": "issue:43937", "right": "issue:44485", "accept": false, - "reason": "Both mention GLM-5, but one is GenerationConfig validation and the other is RoPE implementation; different concrete failures." + "reason": "Both mention GLM-5, but one is invalid GenerationConfig handling and the other is a RoPE implementation concern." }, { "left": "issue:42898", "right": "issue:46614", "accept": false, - "reason": "Tokenizer cleanup behavior in v5 is unrelated to stale espeak backend state in Wav2Vec2PhonemeCTCTokenizer." + "reason": "Both are tokenizer-related regressions, but one is clean-up spacing behavior in v5 and the other is stale phonemizer language state." }, { "left": "issue:38617", "right": "issue:45278", "accept": false, - "reason": "45278 is a broad import-regression report, but not evidence of the same specific missing symbol/import path as 38617." + "reason": "45378 is a broad import-error umbrella; 38617 is a specific missing symbol import. Related context, but not the same concrete bug." }, { "left": "issue:43976", "right": "issue:45446", "accept": false, - "reason": "Different version-check bugs in different code paths: Python support vs PyTorch import gating." + "reason": "Different platform/version compatibility problems: Python version support vs a PyTorch version guard for an import." }, { "left": "issue:43611", "right": "issue:46002", "accept": false, - "reason": "Model-loading/base_model_prefix regression is unrelated to NaNs in classification heads during checkout/init." + "reason": "Model loading with base_model_prefix is unrelated to NaNs in classification heads." }, { "left": "issue:44038", "right": "issue:45478", "accept": false, - "reason": "Both involve Qwen3 Moe variants, but the model/version and failure mode differ; not the same underlying bug." + "reason": "Same broad Qwen3 MoE family, but different versions/models and different failure modes for from_pretrained." }, { "left": "issue:43066", "right": "issue:44779", "accept": false, - "reason": "Both are tokenizer regressions in v5, but for different tokenizers and symptoms; too broad to merge." + "reason": "Both are v5 tokenizer regressions, but they concern different tokenizers and distinct symptoms." }, { "left": "issue:42673", "right": "issue:45072", "accept": false, - "reason": "VRAM leakage in Qwen3ForCausalLM has nothing to do with bfloat16 dtype mismatches in other models." + "reason": "VRAM leakage in Qwen3ForCausalLM threads is unrelated to dtype mismatches in inference models." }, { "left": "issue:44355", "right": "issue:46231", "accept": false, - "reason": "Compiled Python file execution errors are unrelated to a documentation parameter-name mismatch." + "reason": "Compiled Python file errors have nothing to do with a documentation parameter typo." }, { "left": "issue:45357", "right": "issue:45478", "accept": false, - "reason": "Both are Qwen3.5-related, but one is save_pretrained key corruption and the other is from_pretrained loading failure; distinct code paths." + "reason": "Both involve Qwen3.5 Moe, but one is save_pretrained key corruption and the other is a from_pretrained load error." }, { "left": "issue:44625", "right": "issue:46002", "accept": false, - "reason": "num_labels propagation in Qwen3.5 is unrelated to NaNs in classification heads." + "reason": "num_labels propagation in Qwen3.5 is a config bug; NaNs in classification heads are a different training/init failure." }, { "left": "issue:43116", "right": "issue:43493", "accept": false, - "reason": "Example-script multi-label output bug is unrelated to a SigLIP2 implementation discrepancy." + "reason": "A classification example script returning empty results is unrelated to a SigLIP2 implementation discrepancy." }, { - "left": "issue:44610", - "right": "issue:46909", + "left": "issue:45083", + "right": "issue:45446", "accept": false, - "reason": "Processor padding/input-size bug is unrelated to a broken documentation example." + "reason": "Feature-length computation in qwen3_omni_moe is unrelated to a PyTorch version check in flex_attention." }, { - "left": "issue:45083", - "right": "issue:45446", + "left": "issue:44610", + "right": "issue:46909", "accept": false, - "reason": "A feature-extractor length-calculation issue in qwen3_omni_moe is unrelated to an AuxRequest import version check." + "reason": "An image processor sizing bug is unrelated to a broken documentation example." }, { "left": "issue:42371", "right": "issue:43874", "accept": false, - "reason": "TF32 behavior API guidance and a missing image-processor method are unrelated." + "reason": "TF32 API guidance and a missing multimodal-token helper method are different issues." }, { "left": "issue:44075", "right": "issue:44936", "accept": false, - "reason": "Unused SGD args and trainer.evaluate failing after train are different training/runtime issues." + "reason": "Optimizer argument handling and trainer train/evaluate sequencing are separate bugs." }, { - "left": "issue:46726", - "right": "issue:46981", + "left": "issue:43618", + "right": "issue:46829", "accept": false, - "reason": "Missing vision backend handling in AutoProcessor is unrelated to repeated HTTP 429 retries for cached files." + "reason": "Both are output-field assignment issues, but they affect different model classes and outputs, so they are not the same concrete bug." }, { - "left": "issue:43618", - "right": "issue:46829", + "left": "issue:46726", + "right": "issue:46981", "accept": false, - "reason": "Both are output-field regressions, but they affect different model output classes and different fields; not enough to treat as one bug." + "reason": "AutoProcessor backend-missing error handling is unrelated to retry behavior on HTTP 429 responses." }, { "left": "issue:46328", "right": "issue:47116", "accept": false, - "reason": "Padding inefficiency is unrelated to a streaming pipeline regression." + "reason": "A padding performance issue is unrelated to a streaming pipeline regression." }, { "left": "issue:43992", "right": "issue:44704", "accept": false, - "reason": "Missing embed_tokens.weight on encoder load is unrelated to AutoProcessor kwargs not being forwarded to cached_file." + "reason": "Missing embed_tokens weight loading is a model checkpoint issue; passing kwargs to cached_file is a separate AutoProcessor bug." }, { "left": "issue:43066", "right": "issue:46710", "accept": false, - "reason": "A tokenizer decoder-type regression is not the same as an accuracy regression in DeepSeek-R1 after upgrade." - }, - { - "left": "issue:43116", - "right": "issue:43825", - "accept": false, - "reason": "Example-script classification output and pipeline error-message wording are unrelated." + "reason": "Tokenizer decoder type regression and DeepSeek accuracy regression after upgrade are not the same bug." }, { "left": "issue:28348", "right": "issue:46640", "accept": false, - "reason": "Flash-attention support for GPT2LMHeadModel is not clearly the same bug as a generic flash-attention issue report." + "reason": "Flash Attention 2 support request for GPT2LMHeadModel is not the same as a generic flash-attention issue." }, { - "left": "issue:42673", - "right": "issue:42796", + "left": "issue:43116", + "right": "issue:43825", "accept": false, - "reason": "VRAM leakage in Qwen3ForCausalLM and a tokenizer BPE-marker bug are unrelated." + "reason": "Example-script multi-label output failure is unrelated to a pipeline error message regression." }, { "left": "issue:43827", "right": "issue:45446", "accept": false, - "reason": "Docs still referencing pipeline after removals is unrelated to a PyTorch version check bug." + "reason": "Docs still referencing pipeline() and a PyTorch version check bug are unrelated." + }, + { + "left": "issue:42673", + "right": "issue:42796", + "accept": false, + "reason": "VRAM leakage in multithreaded inference is unrelated to tokenizer BPE-marker handling." }, { "left": "issue:44995", "right": "issue:46697", "accept": false, - "reason": "A stale indexer cache crash on second forward pass is unrelated to an image over-padding preprocessing bug." + "reason": "A second-forward stale-cache crash is unrelated to image over-padding in Swin2SR." } ] }, @@ -16430,13 +16460,13 @@ "issue:43824|issue:46981", "issue:43064|issue:43673", "issue:43673|issue:46328", - "issue:42222|issue:43441", "issue:39379|issue:43688", + "issue:42222|issue:43441", "issue:42222|issue:47000", + "issue:44380|issue:46077", "issue:42222|issue:46752", - "issue:45405|issue:45446", "issue:43937|issue:45563", - "issue:44380|issue:46077", + "issue:45405|issue:45446", "issue:44117|issue:45588", "issue:44291|issue:45071", "issue:44466|issue:46002", @@ -16448,28 +16478,28 @@ "issue:44704|issue:47110", "issue:45071|issue:47059", "issue:45341|issue:46328", - "issue:44829|issue:45216", "issue:45071|issue:46512", + "issue:44829|issue:45216", "issue:45341|issue:45561", "issue:43408|issue:43742", "issue:45632|issue:46981", - "issue:42617|issue:45341", - "issue:44117|issue:45081" + "issue:44117|issue:45081", + "issue:42617|issue:45341" ], - "prepared_review_unit_hash": "96385758d1edf92929f495527dff7e2ea411b9e07248fab8d9707e5c9f72ff1a", + "prepared_review_unit_hash": "238445268bd15ceb72341a7a4c7aec402af5eb70c8d434d2810f219d2eb152b9", "component_budget": { "node_count": 514, "item_count": 514, "soft_pair_count": 1891, - "serialized_chars": 980926, - "estimated_input_tokens": 245232, - "estimated_eval_tokens": 490720 + "serialized_chars": 980895, + "estimated_input_tokens": 245224, + "estimated_eval_tokens": 490704 }, "budget": { "node_count": 48, "item_count": 48, "soft_pair_count": 31, - "serialized_chars": 37012, + "serialized_chars": 37011, "estimated_input_tokens": 9253, "estimated_eval_tokens": 18762 }, @@ -16478,198 +16508,198 @@ "aggressively_trimmed": false, "split": true, "analyst_result": { - "summary": "The soft pairs are mostly broad subsystem similarities rather than true duplicates. I don\u2019t see any pair that clearly describes the same underlying bug/change, so all soft edges should be rejected.", - "confidence": 0.83, - "canonical_issue_reason": "Issue 44779 is the clearest canonical candidate: it describes a concrete v5 tokenizer regression with a specific model, reproducible incorrect behavior, and a stable user-facing failure mode.", + "summary": "The cluster is loosely centered on Transformers issues, but the soft links mostly connect unrelated bugs. The broad tokenizer-loading regression is the best umbrella issue; no PRs are present.", + "confidence": 0.67, + "canonical_issue_reason": "Issue 33453 is the broadest and most generic: it describes a tokenizer-loading regression without narrowing to a specific model or code path, making it the best umbrella for nearby tokenizer-related reports.", "canonical_pr_reason": null, - "best_issue_reason": "Issue 44779 is the strongest standalone issue in this set because it is specific, reproducible, and tied to a single code-path regression; the others are generally either broader, less focused, or about different subsystems.", + "best_issue_reason": "Issue 33453 is the strongest representative for this cluster because it captures the general failure mode shared by several tokenizer/loading complaints, while being broad enough to serve as the canonical issue.", "best_pr_reason": null, "soft_edge_verdicts": [ { "left": "issue:45092", "right": "issue:46762", "accept": false, - "reason": "Different failures: meta-init incompatibility for remote-code checkpoints vs numerical differences in Minimax attention." + "reason": "Different failures: remote-code meta-initialization incompatibility vs Minimax MSA attention numerical differences." }, { "left": "issue:33453", "right": "issue:39401", "accept": false, - "reason": "Both are tokenizer-related, but one is a loading regression and the other is a Qwen3 offset_mapping bug; not the same defect." + "reason": "Both are tokenizer-related, but one is a broad loading regression and the other is a Qwen3 offset_mapping bug; different underlying defects." }, { "left": "issue:44263", "right": "issue:46762", "accept": false, - "reason": "Unrelated code paths: GLM MoE DSA indexing split behavior vs Minimax attention numerics." + "reason": "GlmMoeDsaIndexer torch.split behavior is unrelated to Minimax attention numerical drift." }, { "left": "issue:43824", "right": "issue:46981", "accept": false, - "reason": "ImportError for a missing class is unrelated to repeated HTTP 429 retry behavior in AutoProcessor." + "reason": "ImportError for a missing Qwen2.5-VL class is unrelated to AutoProcessor retry behavior on HTTP 429." }, { "left": "issue:43064", "right": "issue:43673", "accept": false, - "reason": "Optimizer-state corruption under FSDP/PEFT is unrelated to missing GenerationMixin cache during chunked prefill." + "reason": "FSDP2/PEFT optimizer-state corruption is a training/distributed bug, not the GenerationMixin cache issue." }, { "left": "issue:43673", "right": "issue:46328", "accept": false, - "reason": "Generation cache handling and SequenceFeatureExtractor padding performance are different problems." + "reason": "Generation cache missing and SequenceFeatureExtractor.pad performance are different code paths and symptoms." }, { - "left": "issue:42222", - "right": "issue:43441", + "left": "issue:39379", + "right": "issue:43688", "accept": false, - "reason": "Vitpose model breakage and Ministral FlashAttention failure affect different models and code paths." + "reason": "mllama mask handling and auxiliary-loss normalization are distinct model-specific bugs." }, { - "left": "issue:39379", - "right": "issue:43688", + "left": "issue:42222", + "right": "issue:43441", "accept": false, - "reason": "Incorrect mllama mask handling is unrelated to auxiliary-loss normalization in OLMoE/GPT Oss." + "reason": "vitpose model breakage and Ministral FlashAttention failure do not share the same root cause." }, { "left": "issue:42222", "right": "issue:47000", "accept": false, - "reason": "All-vitpose breakage is not the same as RTDetrLoss cost-matrix infeasibility." + "reason": "vitpose breakage is unrelated to RTDetrLoss cost-matrix infeasibility." }, { - "left": "issue:42222", - "right": "issue:46752", + "left": "issue:44380", + "right": "issue:46077", "accept": false, - "reason": "Model breakage in vitpose is unrelated to apply_chat_template failing on an empty list." + "reason": "Both involve attention behavior, but GPT2 attention scaling and Gemma4 bidirectional-mask construction are different bugs." }, { - "left": "issue:45405", - "right": "issue:45446", + "left": "issue:42222", + "right": "issue:46752", "accept": false, - "reason": "Dependency-version bump policy issue is unrelated to a PyTorch version check bug in flex_attention." + "reason": "vitpose breakage is unrelated to apply_chat_template crashing on an empty list." }, { "left": "issue:43937", "right": "issue:45563", "accept": false, - "reason": "Invalid GenerationConfig for GLM-5 is a different bug than a stale warning in paged generate()." + "reason": "GLM-5 GenerationConfig validation and paged-generate stale warning are separate issues." }, { - "left": "issue:44380", - "right": "issue:46077", + "left": "issue:45405", + "right": "issue:45446", "accept": false, - "reason": "Both involve attention settings, but GPT2 scaling being ignored is a different bug from Gemma4 building causal masks under bidirectional attention." + "reason": "A dependency version bump being too high is unrelated to the flex_attention PyTorch version check." }, { "left": "issue:44117", "right": "issue:45588", "accept": false, - "reason": "Tokenizer mapping returning None is unrelated to a flash_attention crash on s_aux=None." + "reason": "TOKENIZER_MAPPING_NAMES returning None and flash_attention crashing on s_aux=None are different failure sources." }, { "left": "issue:44291", "right": "issue:45071", "accept": false, - "reason": "Both are v5/type-related, but init_empty_weights argument handling is a different failure than PretrainedConfig type checking." + "reason": "Both are v5-related regressions, but one is about init_empty_weights arguments and the other about PretrainedConfig type checking." }, { "left": "issue:44466", "right": "issue:46002", "accept": false, - "reason": "Inconsistent lm_head serialization by device is unrelated to NaNs in classification heads after checkout/init." + "reason": "lm_head serialization/tied-weights behavior is unrelated to NaNs in classification heads." }, { "left": "issue:39379", "right": "issue:46897", "accept": false, - "reason": "These are different bugs: mllama mask semantics vs Florence2 double-shift training loss." + "reason": "mllama mask handling is unrelated to Florence2 training-loss double-shift behavior." }, { "left": "issue:44410", "right": "issue:45310", "accept": false, - "reason": "Missing Qwen3Next projections is an architectural/model-definition issue, not the same as Qwen3.5 MoE from_pretrained failure." + "reason": "Missing qwen3next projections and Qwen3.5 MoE from_pretrained errors are different model-specific problems." }, { "left": "issue:39401", "right": "issue:44779", "accept": false, - "reason": "Both are tokenizer regressions, but they affect different models and symptoms: offset_mapping vs general incorrect tokenization." + "reason": "Both are tokenizer regressions, but they affect different models and appear to come from different tokenizer logic." }, { "left": "issue:43450", "right": "issue:43986", "accept": false, - "reason": "Batched video processor shape output and missing torchvision crash are distinct processor-loading issues." + "reason": "Batched video processor shape issues are unrelated to the AutoProcessor/torchvision import crash." }, { "left": "issue:43937", "right": "issue:44977", "accept": false, - "reason": "GLM-5 GenerationConfig validation and Qwen3.5 flash-attention generation failure are different issues." + "reason": "Invalid GenerationConfig for GLM-5 is unrelated to Qwen3.5 flash-attention generation failures." }, { "left": "issue:44704", "right": "issue:47110", "accept": false, - "reason": "AutoProcessor kwargs forwarding is unrelated to special_tokens being the wrong type." + "reason": "Passing kwargs to cached_file and special_tokens being a list are unrelated API bugs." }, { "left": "issue:45071", "right": "issue:47059", "accept": false, - "reason": "PretrainedConfig type checking and cli completion compatibility with huggingface_hub are unrelated." + "reason": "PretrainedConfig type checking regression is unrelated to the hf hub CLI completion attribute breakage." }, { "left": "issue:45341", "right": "issue:46328", "accept": false, - "reason": "A generic testing_utils bug is not the same as a padding performance regression in SequenceFeatureExtractor." + "reason": "A testing_utils bug and a SequenceFeatureExtractor.pad performance issue are not the same defect." }, { - "left": "issue:44829", - "right": "issue:45216", + "left": "issue:45071", + "right": "issue:46512", "accept": false, - "reason": "Degenerate training with flash_attention_3 is unrelated to Qwen3.5 save_pretrained checkpoint serialization." + "reason": "PretrainedConfig type checking and stale config.dtype after from_config are separate configuration issues." }, { - "left": "issue:45071", - "right": "issue:46512", + "left": "issue:44829", + "right": "issue:45216", "accept": false, - "reason": "Stale config dtype after from_config is a different config-state bug than PretrainedConfig type checking failures." + "reason": "FlashAttention-3 training degeneration and Qwen3.5 checkpoint serialization are different bugs." }, { "left": "issue:45341", "right": "issue:45561", "accept": false, - "reason": "Both touch testing utils, but one is an unspecified bug and the other is a specific xdist race on captured_info.txt." + "reason": "The xdist captured_info.txt race is more specific than the unrelated general testing_utils bug." }, { "left": "issue:43408", "right": "issue:43742", "accept": false, - "reason": "sam3_video/sam3_tracker class mismatch is unrelated to a MobileLLM loading KeyError." + "reason": "sam3_tracker warning from sam3_video and MobileLLM-125M KeyError are unrelated loading issues." }, { "left": "issue:45632", "right": "issue:46981", "accept": false, - "reason": "Cache-path collision for local trust_remote_code models is unrelated to repeated 429 retries in AutoProcessor." + "reason": "Cache-path collisions for trust_remote_code and repeated 429 retries are unrelated loader/network problems." }, { - "left": "issue:42617", - "right": "issue:45341", + "left": "issue:44117", + "right": "issue:45081", "accept": false, - "reason": "3d_parallel.py execution failure is unrelated to the testing_utils bug." + "reason": "Both hit tokenizer loading, but one is a mapping-name assumption and the other is a Mistral regex patch crash." }, { - "left": "issue:44117", - "right": "issue:45081", + "left": "issue:42617", + "right": "issue:45341", "accept": false, - "reason": "TOKENIZER_MAPPING_NAMES returning None is a different bug from Mistral regex patching crashing on backend_tokenizer access." + "reason": "3d_parallel.py execution failure is unrelated to the testing_utils bug." } ] }, @@ -16699,13 +16729,11 @@ "issue:41628", "issue:42222", "issue:42371", - "issue:43012", "issue:43066", "issue:43097", "issue:43122", "issue:43208", "issue:43352", - "issue:43577", "issue:43673", "issue:43874", "issue:43931", @@ -16715,6 +16743,7 @@ "issue:44220", "issue:44261", "issue:44263", + "issue:44360", "issue:44410", "issue:44483", "issue:44484", @@ -16761,8 +16790,8 @@ "issue:43066|issue:43122", "issue:43208|issue:46894", "issue:43097|issue:44493", - "issue:44077|issue:46981", "issue:46752|issue:47116", + "issue:44077|issue:46981", "issue:43673|issue:45341", "issue:43931|issue:44410", "issue:41628|issue:45071", @@ -16771,220 +16800,211 @@ "issue:44484|issue:45478", "issue:27957|issue:43352", "issue:44263|issue:46172", - "issue:43012|issue:43577" + "issue:44360|issue:46172" ], - "prepared_review_unit_hash": "f6f4d11131fd8001848f68c82168156f8d39703c3442a0119640d22995a4ef39", + "prepared_review_unit_hash": "9cd87b2350cc1f9a7b7ecb270d96e24f174b409a237e847c84f380d0b29b6875", "component_budget": { "node_count": 514, "item_count": 514, "soft_pair_count": 1891, - "serialized_chars": 980926, - "estimated_input_tokens": 245232, - "estimated_eval_tokens": 490720 + "serialized_chars": 980895, + "estimated_input_tokens": 245224, + "estimated_eval_tokens": 490704 }, "budget": { - "node_count": 48, - "item_count": 48, + "node_count": 47, + "item_count": 47, "soft_pair_count": 29, - "serialized_chars": 35776, - "estimated_input_tokens": 8944, - "estimated_eval_tokens": 18144 + "serialized_chars": 35140, + "estimated_input_tokens": 8785, + "estimated_eval_tokens": 17826 }, "overflow_policy": "truncate_then_skip", "trimmed": false, "aggressively_trimmed": false, "split": true, "analyst_result": { - "summary": "The only accepted soft edge in the provided list is the Qwen3.5 streaming chat/completions regression (44483\u219445464). Separately, the packet also contains issues 45310 and 45478 with identical titles, so there is another obvious duplicate-like pair outside the soft-edge set; the packet is not limited to just one duplicate-like pair.", - "confidence": 0.79, - "canonical_issue_reason": "issue:45464 is the strongest canonical for the Qwen3.5 streaming chat/completions failure because it is the most detailed, reproducible report in that duplicate pair. Separately, issue:45310 and issue:45478 appear to be a second duplicate-like pair because they share an identical title.", + "summary": "These issues are largely heterogeneous regressions across imports, tokenizers, model loading, training, and API behavior. None of the proposed soft pairs look like the same concrete bug/change, so I would reject them all. If a representative issue must be named, #45278 is the broadest umbrella report, but it is not a true canonical duplicate target for this set.", + "confidence": 0.88, + "canonical_issue_reason": "The set is too heterogeneous to have one true canonical issue; the proposed pairs do not share a single concrete failure mode.", "canonical_pr_reason": null, - "best_issue_reason": "issue:45464 is the best issue candidate for the streaming chat/completions regression: it is specific, actionable, and clearly matches the same failure mode as issue:44483. The packet also contains a separate duplicate-like pair, issue:45310 and issue:45478, with identical titles.", + "best_issue_reason": "#45278 is the broadest umbrella report here ('many import errors after update from 4.57.0 to 5.5.0'), so it is the best representative issue if one must be picked.", "best_pr_reason": null, "soft_edge_verdicts": [ { "left": "issue:38617", "right": "issue:41628", "accept": false, - "reason": "Both are import errors, but they involve different missing symbols and different code paths." + "reason": "Both are import errors, but they are about different missing symbols and likely different code paths." }, { "left": "issue:43122", "right": "issue:46298", "accept": false, - "reason": "Tokenizer regression vs noisy logging; not the same underlying bug." + "reason": "Tokenizer regression vs noisy logging; unrelated bugs." }, { "left": "issue:42222", "right": "issue:46178", "accept": false, - "reason": "Vitpose model breakage and a backend-mapping error are unrelated failures." + "reason": "Model-specific breakage vs backend mapping/availability error; not the same fix." }, { "left": "issue:44877", "right": "issue:46231", "accept": false, - "reason": "Config-loading breakage and a documentation parameter mismatch are different issues." + "reason": "Config-loading strictness and a doc parameter typo are unrelated." }, { "left": "issue:39401", "right": "issue:46209", "accept": false, - "reason": "Tokenizer offset-mapping bug and FP8 inference failure are different code paths." + "reason": "Tokenizer offset mapping bug vs FP8/B200 inference failure; different layers." }, { "left": "issue:42371", "right": "issue:44261", "accept": false, - "reason": "TF32 settings/configuration issue vs MLA layernorm precision bug; not the same defect." + "reason": "TF32 settings/API issue vs MLA layernorm epsilon mismatch; different problems." }, { "left": "issue:43937", "right": "issue:44496", "accept": false, - "reason": "GenerationConfig validation error and unrecognized-model config error are different problems." - }, - { - "left": "issue:44484", - "right": "issue:45310", - "accept": false, - "reason": "Save-pretrained shard-size question vs Qwen3.5 MoE from_pretrained loading error are unrelated." + "reason": "GLM-5 generation config validation vs model-type recognition on load; different failure modes." }, { "left": "issue:44261", "right": "issue:46762", "accept": false, - "reason": "Both mention numeric behavior, but they affect different models and implementations." + "reason": "Both mention numerical differences, but they affect different models and different attention paths." }, { "left": "issue:44933", "right": "issue:45278", "accept": false, - "reason": "One missing import vs a broad import-error regression after upgrade are not the same bug." + "reason": "A single missing import is not the same bug as a broad multi-import regression." }, { "left": "issue:42371", "right": "issue:46617", "accept": false, - "reason": "TF32 behavior control and trust_remote_code cache FileNotFoundError are unrelated." + "reason": "TF32 behavior control and trust_remote_code cache loading are unrelated." }, { "left": "issue:44410", "right": "issue:45478", "accept": false, - "reason": "Qwen3Next missing projections and Qwen3.5 MoE loading failure are different model bugs." + "reason": "Different Qwen-family loading bugs in different modules; not one concrete code-path problem." }, { "left": "issue:44075", "right": "issue:45656", "accept": false, - "reason": "SGD args ignored and deepspeed double optimizer steps are separate optimizer issues." + "reason": "Unused SGD args and duplicate optimizer steps are different optimizer bugs." }, { "left": "issue:38617", "right": "issue:44220", "accept": false, - "reason": "Different import/attribute failures in different subsystems." + "reason": "ImportError vs audio feature extraction failure; unrelated." }, { "left": "issue:44263", "right": "issue:44485", "accept": false, - "reason": "torch.split return handling and GLM-5 RoPE implementation are not the same defect." + "reason": "Torch split return handling vs RoPE implementation; different GLM issues." }, { "left": "issue:43066", "right": "issue:43122", "accept": false, - "reason": "Both are tokenizer regressions, but the concrete symptoms and likely fixes differ." + "reason": "Both are tokenizer-related regressions, but one is decoder-type mismatch and the other is changed tokenization output; not clearly the same bug." }, { "left": "issue:43208", "right": "issue:46894", "accept": false, - "reason": "xLSTM training bugs and CohereASR double-shift are unrelated model-specific issues." + "reason": "Different training blockers in different model families; no shared concrete fix." }, { "left": "issue:43097", "right": "issue:44493", "accept": false, - "reason": "Embedding tying removal and unexpected position-id key errors are different regressions." + "reason": "Removed embedding-tie API vs unexpected position-id keys; unrelated regressions." }, { - "left": "issue:44077", - "right": "issue:46981", + "left": "issue:46752", + "right": "issue:47116", "accept": false, - "reason": "PatchTSMixer init policy and HTTP 429 retry behavior are unrelated." + "reason": "Empty chat-template crash vs streaming pipeline regression are different failures." }, { - "left": "issue:46752", - "right": "issue:47116", + "left": "issue:44077", + "right": "issue:46981", "accept": false, - "reason": "Empty chat-template crash and streaming pipeline regression are different problems." + "reason": "PatchTSMixer init policy vs HTTP 429 retry behavior; not the same underlying issue." }, { "left": "issue:43673", "right": "issue:45341", "accept": false, - "reason": "Generation cache regression and a testing_utils bug are unrelated." + "reason": "Missing generation cache in v5 chunked prefill vs a testing_utils bug; unrelated." }, { "left": "issue:43931", "right": "issue:44410", "accept": false, - "reason": "Qwen3-VL weight-shape mismatch and Qwen3Next missing projections are different loading issues." + "reason": "Qwen3-VL weight-shape mismatch and missing Qwen3Next projections are separate model-loading defects." }, { "left": "issue:41628", "right": "issue:45071", "accept": false, - "reason": "General import error and PreTrainedConfig type-checking breakage are not the same concrete bug." + "reason": "Missing AutoImageProcessor import and PreTrainedConfig type checking are different compatibility issues." }, { "left": "issue:44483", "right": "issue:45464", - "accept": true, - "reason": "Same model and same streaming chat/completions failure; 45464 is the richer duplicate report." + "accept": false, + "reason": "General chat/completions rejection vs streaming-specific Qwen3.5 failure; not one concrete bug." }, { "left": "issue:43874", "right": "issue:44261", "accept": false, - "reason": "Missing image-patch helper in GLM46V is unrelated to MLA q_a_layernorm precision handling." + "reason": "Different multimodal/token-count vs layernorm-epsilon bugs." }, { "left": "issue:44484", "right": "issue:45478", "accept": false, - "reason": "Different Qwen3.5-related issues: save_pretrained shard-size question vs from_pretrained loading failure." + "reason": "Save-pretrained shard size question vs Qwen3.5 load failure; unrelated." }, { "left": "issue:27957", "right": "issue:43352", "accept": false, - "reason": "Both mention Flash Attention 2, but they target different models and support checks." + "reason": "Both mention Flash Attention 2 support, but they concern different models and likely different implementations." }, { "left": "issue:44263", "right": "issue:46172", "accept": false, - "reason": "torch.split return-value handling and a DeepSeek tensor transpose issue are unrelated." + "reason": "Different GLM/DeepSeek code paths and different tensor-shape bugs." }, { - "left": "issue:43012", - "right": "issue:43577", + "left": "issue:44360", + "right": "issue:46172", "accept": false, - "reason": "General precision-warning behavior vs BLIP2 dtype staying float32 are similar only at a high level, not the same bug." + "reason": "Missing ReLU in the DSA indexer is unrelated to the DeepSeek transpose issue." } ] }, - "evaluator_result": { - "accept": false, - "feedback": "The soft-edge verdicts are mostly conservative, but the summary overstates uniqueness: the packet also contains issue:45310 and issue:45478 with identical titles ('[BUG] transformers>=5.4.0, Qwen3.5 Moe from_pretrained error'), which looks duplicate-like. Because of that, saying the Qwen3.5 streaming chat/completions pair is 'the only clear duplicate-like pair' is not fully grounded in the packet." - }, - "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 }, @@ -17012,6 +17032,7 @@ "issue:43441", "issue:43452", "issue:43531", + "issue:43577", "issue:43673", "issue:43873", "issue:43874", @@ -17024,7 +17045,6 @@ "issue:44263", "issue:44315", "issue:44351", - "issue:44360", "issue:44410", "issue:44484", "issue:44485", @@ -17051,8 +17071,8 @@ "issue:46981" ], "soft_pairs": [ - "issue:44360|issue:46172", "issue:43441|issue:44829", + "issue:43012|issue:43577", "issue:43931|issue:44038", "issue:43097|issue:43452", "issue:43986|issue:45341", @@ -17067,14 +17087,14 @@ "issue:43673|issue:46123", "issue:44410|issue:45375", "issue:45440|issue:46172", - "issue:38617|issue:42757", "issue:44038|issue:44410", + "issue:38617|issue:42757", "issue:43986|issue:45372", "issue:43097|issue:46050", "issue:43064|issue:44315", "issue:43208|issue:46897", - "issue:46655|issue:46664", "issue:43986|issue:44484", + "issue:46655|issue:46664", "issue:46298|issue:46752", "issue:38617|issue:44351", "issue:42890|issue:46298", @@ -17083,220 +17103,220 @@ "issue:43012|issue:46821", "issue:44117|issue:46701" ], - "prepared_review_unit_hash": "d5f7f5252c276887258d1abd9b53e81237239313360dd0821a6f68282a2b8f3c", + "prepared_review_unit_hash": "d5f2931e60060c4d371253c04616d3b11b8503b4b2182a026f6a0c71b12f0e9e", "component_budget": { "node_count": 514, "item_count": 514, "soft_pair_count": 1891, - "serialized_chars": 980926, - "estimated_input_tokens": 245232, - "estimated_eval_tokens": 490720 + "serialized_chars": 980895, + "estimated_input_tokens": 245224, + "estimated_eval_tokens": 490704 }, "budget": { "node_count": 48, "item_count": 48, "soft_pair_count": 31, - "serialized_chars": 36848, - "estimated_input_tokens": 9212, - "estimated_eval_tokens": 18680 + "serialized_chars": 36924, + "estimated_input_tokens": 9231, + "estimated_eval_tokens": 18718 }, "overflow_policy": "truncate_then_skip", "trimmed": false, "aggressively_trimmed": false, "split": true, "analyst_result": { - "summary": "This is a noisy mixed cluster of unrelated Transformers import, model, training, and processor bugs. None of the soft edges look like true duplicates or mergeable as the same underlying fix.", - "confidence": 0.96, - "canonical_issue_reason": "issue:43097 is the broadest representative of the cluster: a clear v5 breaking-change regression with multiple inbound refs and a general API-removal theme. It is the best cluster anchor, but it should not be treated as a duplicate of the other reports.", + "summary": "This cluster is noisy: the issues span many unrelated regressions, import errors, and model-specific bugs. I don\u2019t see a single underlying duplicate bug/change tying them together, so no soft edges should be merged.", + "confidence": 0.92, + "canonical_issue_reason": "No single canonical duplicate: the items are mostly distinct failures across different models, processors, and API surfaces rather than one shared root cause.", "canonical_pr_reason": null, - "best_issue_reason": "issue:43097 is the strongest single issue to keep as the cluster representative because it is a concrete, widely relevant regression report tied to the v5 transition. The other issues are mostly distinct model-specific or subsystem-specific bugs.", + "best_issue_reason": "Issue 43097 is the best representative if one must be chosen: it\u2019s a clear, user-facing breaking-change report with broader API impact and some discussion/reference activity.", "best_pr_reason": null, "soft_edge_verdicts": [ { - "left": "issue:44360", - "right": "issue:46172", + "left": "issue:43441", + "right": "issue:44829", "accept": false, - "reason": "Different bugs: one is about the DSA indexer lacking ReLU, the other about a deepseek-v4 tensor transpose/computation issue." + "reason": "Both mention FlashAttention, but one is a Ministral-3 failure and the other is degenerate training with flash_attention_3; different code paths and symptoms." }, { - "left": "issue:43441", - "right": "issue:44829", + "left": "issue:43012", + "right": "issue:43577", "accept": false, - "reason": "Both are model/training regressions, but the concrete failures differ: Ministral flash-attention crash vs. degenerate sequence-classification training." + "reason": "Both are dtype/precision complaints, but one is a PyTorch compile warning and the other is BLIP2 loading not casting qformer; not the same bug." }, { "left": "issue:43931", "right": "issue:44038", "accept": false, - "reason": "Same family only superficially; one is a Qwen3-VL-30B weight-shape mismatch, the other is a Qwen3-VL-Moe issue with a different failure mode." + "reason": "Different Qwen3-VL/MoE loading problems with different failure modes; too broad to be the same underlying issue." }, { "left": "issue:43097", "right": "issue:43452", "accept": false, - "reason": "One is a removed API in v5, the other is a gguf_file from_pretrained loading breakage; different code paths and fixes." + "reason": "One is a removed API/compatibility complaint, the other is a gguf_file preload regression; unrelated breakages." }, { "left": "issue:43986", "right": "issue:45341", "accept": false, - "reason": "Unrelated subsystems: video AutoProcessor/torchvision crash vs. a testing_utils bug." + "reason": "A processor import/dependency crash versus a testing_utils bug; no shared underlying code-path." }, { "left": "issue:44261", "right": "issue:44485", "accept": false, - "reason": "Different model internals: MLA q_a_layernorm precision/config issue vs. GLM-5 RoPE implementation bug." + "reason": "Different model internals: missing rms_norm_eps precision issue versus GLM-5 RoPE implementation concerns." }, { "left": "issue:45003", "right": "issue:45341", "accept": false, - "reason": "Different areas and symptoms: unsafe sys.modules access vs. a testing_utils issue." + "reason": "Unsafe sys.modules access in modeling_utils is unrelated to the testing_utils issue." }, { "left": "issue:42886", "right": "issue:46981", "accept": false, - "reason": "Offline cache loading failure and repeated HTTP 429 retries are distinct networking behaviors with different root causes." + "reason": "Both involve hub/network behavior, but offline cache loading and repeated 429 retries are different failure modes." }, { "left": "issue:45753", "right": "issue:46369", "accept": false, - "reason": "Different model-specific problems: an infinite loop on a specific image vs. memory overflow in sam3-litetext." + "reason": "Different model families and symptoms: infinite loop on a specific image versus memory overflow." }, { "left": "issue:43531", "right": "issue:45341", "accept": false, - "reason": "Qwen3-MoE sliding_window behavior is unrelated to a testing_utils bug." + "reason": "Sliding_window behavior in Qwen3-MoE has no relation to the testing utility bug." }, { "left": "issue:44117", "right": "issue:44987", "accept": false, - "reason": "TOKENIZER_MAPPING_NAMES returning None is a different failure than loading physical-intelligence/fast in >=5.1.0." + "reason": "One is a tokenizer mapping assumption bug, the other is a model-loading failure for a specific repo; not the same root cause." }, { "left": "issue:43874", "right": "issue:44263", "accept": false, - "reason": "Different implementations: missing multimodal image-patch method vs. torch.split return handling in the DSA indexer." + "reason": "Both touch GLM-related code, but image-patch counting and torch.split return values are separate bugs." }, { "left": "issue:43097", "right": "issue:45588", "accept": false, - "reason": "API removal in v5 and a flash_attention integration crash on s_aux=None are unrelated." + "reason": "Removed API visibility and flash_attention.py crashing on sink-less models are unrelated." }, { "left": "issue:43673", "right": "issue:46123", "accept": false, - "reason": "Cache handling during chunked_prefill and a final-partial-batch flag bug in MaskGenerationPipeline are different code paths." + "reason": "Cache handling in GenerationMixin and final-batch is_last behavior in MaskGenerationPipeline are different code paths." }, { "left": "issue:44410", "right": "issue:45375", "accept": false, - "reason": "Different Qwen3-family issues: missing projections in a layer vs. a missing config field silently dropped by strict validation." + "reason": "Both are Qwen3-family issues, but missing projections and a missing config field are different problems." }, { "left": "issue:45440", "right": "issue:46172", "accept": false, - "reason": "Different models and bugs: DeepseekV3MoE divergence vs. deepseek-v4 comb transpose behavior." - }, - { - "left": "issue:38617", - "right": "issue:42757", - "accept": false, - "reason": "Both are import errors, but they involve different missing symbols from different modules and different fixes." + "reason": "Different DeepSeek variants and different bugs; no shared concrete fix." }, { "left": "issue:44038", "right": "issue:44410", "accept": false, - "reason": "Both mention Qwen3, but one is a Qwen3-VL-Moe bug and the other is a qwen3next layer-definition issue." + "reason": "Qwen3-VL-Moe failure versus qwen3next missing projections; same broad family only, not the same bug." + }, + { + "left": "issue:38617", + "right": "issue:42757", + "accept": false, + "reason": "Both are import errors, but they concern different symbols/modules and likely different compatibility issues." }, { "left": "issue:43986", "right": "issue:45372", "accept": false, - "reason": "Different crash surfaces: missing torchvision for video processor loading vs. a mistral_common import error affecting Gemma 4 processor loading." + "reason": "Missing torchvision dependency and mistral_common ReasoningEffort import failure are distinct processor-loading issues." }, { "left": "issue:43097", "right": "issue:46050", "accept": false, - "reason": "A removed v5 API and a sinq quantization save/load regression are not the same underlying bug." + "reason": "API-removal complaint versus quantized model save/load corruption; unrelated." }, { "left": "issue:43064", "right": "issue:44315", "accept": false, - "reason": "Different training/integration paths: FSDP2+PEFT optimizer state corruption vs. Liger Kernel not being applied with model_init." + "reason": "Different training-stack problems: FSDP2/PEFT optimizer state corruption versus Liger Kernel not applied with model_init." }, { "left": "issue:43208", "right": "issue:46897", "accept": false, - "reason": "Separate model-specific reports: xLSTM training blockers vs. Florence2 double-shift loss bug." + "reason": "xLSTM training blockers and Florence2 double-shift loss bug are different model-specific issues." }, { - "left": "issue:46655", - "right": "issue:46664", + "left": "issue:43986", + "right": "issue:44484", "accept": false, - "reason": "Same subsystem only, but distinct bugs: incorrect padding calculation vs. inconsistent num_frames semantics." + "reason": "Processor crash without torchvision is unrelated to the max_shard_size question." }, { - "left": "issue:43986", - "right": "issue:44484", + "left": "issue:46655", + "right": "issue:46664", "accept": false, - "reason": "Video processor import failure and a save_pretrained shard-size question are unrelated." + "reason": "Both are SmolVLM video-processor issues, but one is padding calculation and the other is num_frames semantics; not the same concrete bug." }, { "left": "issue:46298", "right": "issue:46752", "accept": false, - "reason": "Noisy logging and apply_chat_template crashing on an empty list are unrelated issues." + "reason": "Noisy logging and apply_chat_template crashing on empty input are unrelated." }, { "left": "issue:38617", "right": "issue:44351", "accept": false, - "reason": "Both are import errors from transformers, but they concern different missing exports and different call sites." + "reason": "Two different missing-import symbols from different modules; same symptom class, not the same issue." }, { "left": "issue:42890", "right": "issue:46298", "accept": false, - "reason": "A test flake from missing set_seed and a noisy logging message are unrelated." + "reason": "A flaky test due to missing set_seed versus a logging message issue; unrelated." }, { "left": "issue:44062", "right": "issue:44117", "accept": false, - "reason": "Tokenizer AddedToken argument duplication and TOKENIZER_MAPPING_NAMES returning None are different tokenizer internals." + "reason": "Tokenizer AddedToken argument duplication and TOKENIZER_MAPPING_NAMES returning None are different tokenizer-loading bugs." }, { "left": "issue:43097", "right": "issue:43873", "accept": false, - "reason": "A removed API in v5 and a quantization/offloading bug are unrelated." + "reason": "Breaking API removal and quantization offloading behavior are not the same bug." }, { "left": "issue:43012", "right": "issue:46821", "accept": false, - "reason": "PyTorch precision warning for bf16 compilation is unrelated to a TimesFM window_size AttributeError." + "reason": "Precision warning during compile and TimesFM window_size AttributeError are unrelated." }, { "left": "issue:44117", "right": "issue:46701", "accept": false, - "reason": "Tokenizer mapping None and a DiffusionGemmaGenerationOutput pipeline failure are different bugs in different paths." + "reason": "Tokenizer mapping assumptions and DiffusionGemma pipeline failure do not share a concrete fix path." } ] }, @@ -17389,8 +17409,8 @@ "issue:43208|issue:43550", "issue:43352|issue:43976", "issue:43352|issue:44380", - "issue:43825|issue:44655", "issue:43066|issue:43906", + "issue:43825|issue:44655", "issue:44734|issue:47116", "issue:43097|issue:46298", "issue:45584|issue:46365", @@ -17399,237 +17419,237 @@ "issue:44077|issue:46365", "issue:42890|issue:47059", "issue:44222|issue:45478", - "issue:42757|issue:43824", "issue:44380|issue:45479", + "issue:42757|issue:43824", "issue:36683|issue:43582", "issue:43064|issue:44380" ], - "prepared_review_unit_hash": "95c97b8a174d2a322e1cc3fe25314fec0d185409f64ac6a98b5f742e9987af89", + "prepared_review_unit_hash": "ae360f47890d4d0efe43570df9688e951bcbf4500247495116fc4532d6cacfa9", "component_budget": { "node_count": 514, "item_count": 514, "soft_pair_count": 1891, - "serialized_chars": 980926, - "estimated_input_tokens": 245232, - "estimated_eval_tokens": 490720 + "serialized_chars": 980895, + "estimated_input_tokens": 245224, + "estimated_eval_tokens": 490704 }, "budget": { "node_count": 48, "item_count": 48, "soft_pair_count": 33, - "serialized_chars": 37473, - "estimated_input_tokens": 9369, - "estimated_eval_tokens": 18994 + "serialized_chars": 37467, + "estimated_input_tokens": 9367, + "estimated_eval_tokens": 18990 }, "overflow_policy": "truncate_then_skip", "trimmed": false, "aggressively_trimmed": false, "split": true, "analyst_result": { - "summary": "This cluster is mostly a grab-bag of unrelated Transformers issues; the soft links look like superficial keyword similarity rather than shared root causes. None of the candidate pairs are strong enough to merge, and there are no PRs here.", - "confidence": 0.44, - "canonical_issue_reason": "issue:47116 is the broadest, most central open report in the set, covering a v5 regression class rather than a narrow model-specific bug.", + "summary": "The set is a collection of mostly unrelated Transformers issues across imports, model configs, training, docs, and backend behavior. None of the soft pairs look like true duplicates or the same concrete change, and there are no PRs in the cluster.", + "confidence": 0.97, + "canonical_issue_reason": "No canonical issue: the items span many unrelated bugs and feature regressions, so there is no single underlying issue to dedupe against.", "canonical_pr_reason": null, - "best_issue_reason": "issue:47116 is the best umbrella issue because it is recent, open, and framed as a general regression rather than a one-off symptom.", + "best_issue_reason": "Issue 47116 is the broadest/recentest regression report and has inbound references, so it is the closest thing to a representative issue, but it is still not a true canonical for this mixed cluster.", "best_pr_reason": null, "soft_edge_verdicts": [ { "left": "issue:46172", "right": "issue:46710", "accept": false, - "reason": "Both mention DeepSeek, but one is a dtype/transpose bug and the other is an accuracy regression; different failure modes and likely different fixes." + "reason": "Different bugs: a DeepSeek-v4 tensor expression issue vs an accuracy regression after upgrading Transformers." }, { "left": "issue:43502", "right": "issue:43992", "accept": false, - "reason": "Unrelated areas: offline/local-files-only network behavior vs missing encoder weights during model loading." + "reason": "Different code paths: offline-file fetching vs missing UMT5 encoder weights on load." }, { "left": "issue:44704", "right": "issue:45103", "accept": false, - "reason": "AutoProcessor cached_file kwargs forwarding and auto_docstring annotation parsing are different subsystems and bugs." + "reason": "Both involve loading/documentation plumbing, but one is cached_file kwargs in AutoProcessor while the other is an auto_docstring annotation crash." }, { "left": "issue:44936", "right": "issue:46365", "accept": false, - "reason": "Training/evaluate lifecycle failure is unrelated to the import-path issue caused by datasets shadowing Trainer import." + "reason": "Trainer.evaluate failure after training is unrelated to importing Trainer being broken by installed datasets." }, { "left": "issue:44451", "right": "issue:47116", "accept": false, - "reason": "Model-loading failure for ScandiBERT is not the same as a streaming pipeline regression in v5." + "reason": "Model load failure for a specific checkpoint vs a streaming pipeline regression; not the same bug." }, { "left": "issue:44380", "right": "issue:46421", "accept": false, - "reason": "GPT-2 attention scaling under SDPA/FlashAttention is a model/backend bug, not the cache-lazy-init recompilation issue." + "reason": "GPT2 attention scaling ignored in SDPA/FlashAttention is unrelated to cache lazy initialization recompilation." }, { "left": "issue:44938", "right": "issue:46178", "accept": false, - "reason": "Python 3.14 load failure and tensorflow_text backend validation are unrelated compatibility problems." + "reason": "Python 3.14 import/load failure is unrelated to a backend mapping ValueError for tensorflow_text." }, { "left": "issue:43986", "right": "issue:44016", "accept": false, - "reason": "AutoProcessor/video-model torchvision dependency issue is unrelated to a notebook syntax error." + "reason": "AutoProcessor/torchvision crash is unrelated to a notebook syntax error." }, { "left": "issue:43064", "right": "issue:47116", "accept": false, - "reason": "Distributed optimizer-state corruption under FSDP2/PEFT is not the same as a streaming pipeline regression." + "reason": "Trainer/FSDP2/PEFT optimizer-state corruption is unrelated to streaming pipeline regressions." }, { "left": "issue:42907", "right": "issue:45753", "accept": false, - "reason": "Saving dequantized Ministral/Devstral models and a Qwen3.5 image infinite loop are different model-path failures." + "reason": "Dequantized save failure for Ministral/Devstral is unrelated to a Qwen3.5 infinite-loop inference bug." }, { "left": "issue:42907", "right": "issue:44655", "accept": false, - "reason": "Save-pretrained failure for dequantized models is not the same as inability to save Pipeline objects." + "reason": "Saving dequantized models is not the same as saving Pipeline objects with save_pretrained." }, { "left": "issue:43012", "right": "issue:43550", "accept": false, - "reason": "Both involve compilation/backends, but one is a bfloat16 warning and the other is a Bamba SDPA compile failure; not the same bug." + "reason": "bfloat16 compile warnings and Bamba SDPA compile failure are different backend problems." }, { "left": "issue:45712", "right": "issue:46298", "accept": false, - "reason": "Left is repo hygiene/dummy class leakage; right is a noisy logging message." + "reason": "Left is dummy PT object cleanup; right is a noisy logging message." }, { "left": "issue:43012", "right": "issue:44368", "accept": false, - "reason": "A PyTorch precision warning is not the same as a tie_word_embeddings warning from LoRA fine-tuning." + "reason": "bfloat16 compilation warning is unrelated to tie_word_embeddings warning spam during Qwen3.5 LoRA fine-tuning." }, { "left": "issue:43992", "right": "issue:44484", "accept": false, - "reason": "Missing embed tokens on load and a max_shard_size question are unrelated save/load concerns." + "reason": "Missing encoder weights on load is unrelated to save_pretrained shard size defaults." }, { "left": "issue:43827", "right": "issue:47116", "accept": false, - "reason": "Documentation still referencing pipeline() is not the same underlying problem as the runtime streaming pipeline regression." + "reason": "Docs still referencing pipeline() is separate from a runtime streaming pipeline regression." }, { "left": "issue:43208", "right": "issue:43550", "accept": false, - "reason": "xLSTM training blockers and Bamba SDPA compile failure are different model-specific issues." + "reason": "xLSTM training bugs are unrelated to Bamba torch.compile+SDPA failure." }, { "left": "issue:43352", "right": "issue:43976", "accept": false, - "reason": "Nemotron flash-attn support and Python version compatibility are unrelated." + "reason": "Flash Attention support for NemotronH is unrelated to a Python version compatibility issue." }, { "left": "issue:43352", "right": "issue:44380", "accept": false, - "reason": "Flash Attention support gating for Nemotron and GPT-2 attention scaling under SDPA/FlashAttention are distinct code paths." + "reason": "Nemotron FlashAttention support gating is unrelated to GPT2 attention-scaling being ignored." }, { - "left": "issue:43825", - "right": "issue:44655", + "left": "issue:43066", + "right": "issue:43906", "accept": false, - "reason": "Pipeline translation error messaging and saving Pipeline objects are different pipeline-related problems, not one bug." + "reason": "Wrong tokenizer decoder type in v5 is unrelated to an isolated reproduction of another issue." }, { - "left": "issue:43066", - "right": "issue:43906", + "left": "issue:43825", + "right": "issue:44655", "accept": false, - "reason": "Tokenizer decoder type mismatch and an isolated reproduction of another issue are not evidenced as the same underlying defect." + "reason": "pipeline() translation messaging is unrelated to inability to save Pipeline objects." }, { "left": "issue:44734", "right": "issue:47116", "accept": false, - "reason": "Server KV-cache continuation indexing failure is not the same as a generic streaming pipeline regression." + "reason": "KV-cache continuation tensor indexing in serve is a different runtime path from streaming pipeline regression." }, { "left": "issue:43097", "right": "issue:46298", "accept": false, - "reason": "Removed API warning/removal notice and a noisy logging message are unrelated." + "reason": "Removed tie_embeddings_and_encoder_decoder API is unrelated to a noisy logging message." }, { "left": "issue:45584", "right": "issue:46365", "accept": false, - "reason": "Whisper empty transcription failure has no relation to the Trainer import conflict." + "reason": "Whisper empty-transcription failure is unrelated to importing Trainer being broken by datasets." }, { "left": "issue:44380", "right": "issue:46132", "accept": false, - "reason": "Attention scaling under backend selection and AttentionInterface.register behavior change are different attention bugs." + "reason": "GPT2 SDPA/FlashAttention scaling bug is unrelated to AttentionInterface.register behavior changes." }, { "left": "issue:44162", "right": "issue:46172", "accept": false, - "reason": "ESM2 breakage and a DeepSeek tensor operation issue are unrelated model failures." + "reason": "ESM2 breakage is unrelated to a DeepSeek-v4 tensor operation bug." }, { "left": "issue:44077", "right": "issue:46365", "accept": false, - "reason": "patchtsmixer post_init validation is unrelated to the Trainer import problem." + "reason": "patchtsmixer post_init validation is unrelated to the Trainer import conflict." }, { "left": "issue:42890", "right": "issue:47059", "accept": false, - "reason": "Seed-related flaky tests and a huggingface_hub CLI API break are different classes of failures." + "reason": "Missing set_seed in a SAM HQ test is unrelated to a CLI breakage with huggingface_hub 1.22.0." }, { "left": "issue:44222", "right": "issue:45478", "accept": false, - "reason": "Both mention MoE/FP8/Qwen3.5, but one is save_pretrained in FP8 and the other is from_pretrained loading; insufficient evidence of one root cause." + "reason": "FP8 save_pretrained for MoE and Qwen3.5 MoE from_pretrained are related to different save/load paths and not the same bug." }, { - "left": "issue:42757", - "right": "issue:43824", + "left": "issue:44380", + "right": "issue:45479", "accept": false, - "reason": "huggingface_hub import break and a missing Transformers class import are separate import errors from different packages." + "reason": "Attention scaling ignored in backends is unrelated to degenerate zero-loss classification behavior." }, { - "left": "issue:44380", - "right": "issue:45479", + "left": "issue:42757", + "right": "issue:43824", "accept": false, - "reason": "Attention backend scaling and a degenerate zero-loss classification bug are unrelated model behaviors." + "reason": "Import failure for is_offline_mode from huggingface_hub is unrelated to missing Qwen2_5_VL exports from transformers." }, { "left": "issue:36683", "right": "issue:43582", "accept": false, - "reason": "Gemma3Config missing vocab_size and Apple Silicon allocator warmup TypeError are unrelated." + "reason": "Gemma3Config missing vocab_size is unrelated to an Apple Silicon caching_allocator_warmup TypeError." }, { "left": "issue:43064", "right": "issue:44380", "accept": false, - "reason": "Distributed optimizer-state corruption and GPT-2 attention scaling are unrelated code paths." + "reason": "FSDP2/PEFT optimizer-state corruption is unrelated to GPT2 attention scaling being ignored in SDPA/FlashAttention." } ] }, @@ -17710,8 +17730,8 @@ "issue:43756|issue:44485", "issue:43618|issue:43761", "issue:45276|issue:46077", - "issue:43208|issue:43526", "issue:36010|issue:42757", + "issue:43208|issue:43526", "issue:44038|issue:44230", "issue:42994|issue:44655", "issue:43299|issue:43867", @@ -17719,8 +17739,8 @@ "issue:44625|issue:46365", "issue:43352|issue:43937", "issue:44625|issue:46077", - "issue:44355|issue:46877", "issue:43611|issue:44655", + "issue:44355|issue:46877", "issue:36683|issue:43475", "issue:44655|issue:46050", "issue:43976|issue:44938", @@ -17734,20 +17754,20 @@ "issue:29870|issue:36296", "issue:43064|issue:44492" ], - "prepared_review_unit_hash": "74b1fd1c1a371ae732a688116631c1154e66912715f49c3ae1611d6203a73e1f", + "prepared_review_unit_hash": "69313572dbed1b267fe50b41222360a5b46106694280795a4b69b44427d43ff9", "component_budget": { "node_count": 514, "item_count": 514, "soft_pair_count": 1891, - "serialized_chars": 980926, - "estimated_input_tokens": 245232, - "estimated_eval_tokens": 490720 + "serialized_chars": 980895, + "estimated_input_tokens": 245224, + "estimated_eval_tokens": 490704 }, "budget": { "node_count": 47, "item_count": 47, "soft_pair_count": 31, - "serialized_chars": 36110, + "serialized_chars": 36109, "estimated_input_tokens": 9028, "estimated_eval_tokens": 18312 }, @@ -17756,198 +17776,198 @@ "aggressively_trimmed": false, "split": true, "analyst_result": { - "summary": "These are mostly false-positive similarities: the items span unrelated import errors, model/config regressions, backend issues, and feature-specific bugs, with no pair clearly showing the same root cause.", - "confidence": 0.96, - "canonical_issue_reason": "No clear canonical issue; the cluster is dominated by unrelated false-positive similarities across different components and failure modes.", + "summary": "No strong duplicate cluster emerges here: the items span unrelated import, config, saving, training, and model-specific regressions. The soft edges look like broad thematic similarity only, so I would not merge any of the candidate pairs.", + "confidence": 0.95, + "canonical_issue_reason": null, "canonical_pr_reason": null, - "best_issue_reason": "No single issue is a strong representative because the items do not share one underlying bug or change.", + "best_issue_reason": null, "best_pr_reason": null, "soft_edge_verdicts": [ { "left": "issue:42175", "right": "issue:46365", "accept": false, - "reason": "Different problems: one is missing TensorFlow in the torch extra, the other is an import failure for Trainer likely tied to datasets interaction." + "reason": "Both mention environment/backend setup, but one is about missing TensorFlow in a torch install and the other about Trainer import breaking with datasets." }, { "left": "issue:43012", "right": "issue:44683", "accept": false, - "reason": "Unrelated failures: bfloat16 compile warning versus compiled flex_attention failing on torch >= 2.9." + "reason": "Both involve compiled/precision-related warnings or failures, but they are different code paths and different failure modes." }, { "left": "issue:44483", "right": "issue:44514", "accept": false, - "reason": "Different request paths: chat/completions API rejection versus batched chat template crash with padding=False." + "reason": "Both concern request/chat handling, but one is an API accept/reject issue and the other is a batched chat template crash." }, { "left": "issue:38617", "right": "issue:43502", "accept": false, - "reason": "Both are import/network-related only at a high level, but one is a missing symbol import and the other is unwanted hub requests despite local_files_only=True." + "reason": "Both are import/offline-related at a high level, but one is an ImportError and the other is network calls despite local_files_only=True." }, { "left": "issue:43618", "right": "issue:44380", "accept": false, - "reason": "Different model behaviors: CLIP attentions missing versus GPT2 attention scaling ignored in SDPA/FlashAttention." + "reason": "Both are model backend regressions, but one is missing CLIP attentions and the other is GPT2 attention scaling being ignored." }, { "left": "issue:43756", "right": "issue:44485", "accept": false, - "reason": "Different model-specific issues: Smollm3 RoPE layer handling versus GLM-5 RoPE implementation discussion." + "reason": "Both mention RoPE/model behavior, but they target different models and different implementation bugs." }, { "left": "issue:43618", "right": "issue:43761", "accept": false, - "reason": "Both involve CLIP vision outputs, but one is attentions not assigned and the other is hidden_states missing; not the same bug." + "reason": "Both are CLIP regressions, but one is missing attentions and the other is hidden_states not returned; separate concrete bugs." }, { "left": "issue:45276", "right": "issue:46077", "accept": false, - "reason": "Different Gemma4 bugs: token embedding resize propagation versus bidirectional attention mask construction." + "reason": "Both are Gemma4 issues, but one is about token embedding resize propagation and the other about bidirectional attention masks." }, { - "left": "issue:43208", - "right": "issue:43526", + "left": "issue:36010", + "right": "issue:42757", "accept": false, - "reason": "xLSTM training bugs are unrelated to BeitImageProcessorFast reduce_labels returning only one label." + "reason": "Both are import errors, but they concern different symbols from different dependencies and unrelated breakages." }, { - "left": "issue:36010", - "right": "issue:42757", + "left": "issue:43208", + "right": "issue:43526", "accept": false, - "reason": "Both are import errors, but for different missing symbols from different packages and release breakages." + "reason": "These are unrelated: xLSTM training bugs versus a BeitImageProcessorFast label-reduction bug." }, { "left": "issue:44038", "right": "issue:44230", "accept": false, - "reason": "Related Qwen3-VL-Moe family, but one is a generic loading bug and the other is fp8 inference support; not the same change." + "reason": "Both are Qwen3-VL/MoE/fp8-related, but one is a general loading bug and the other is fp8 inference support." }, { "left": "issue:42994", "right": "issue:44655", "accept": false, - "reason": "Different save/load problems: quantized model saving versus saving Pipeline objects with save_pretrained." + "reason": "Both involve save_pretrained, but one is quantized model saving and the other is Pipeline serialization." }, { "left": "issue:43299", "right": "issue:43867", "accept": false, - "reason": "Different loading regressions: Qwen3VL MoE model loading versus model loading failure when state_dict is sorted." + "reason": "Both are model loading failures, but the roots differ: Qwen3VL MoE loading versus sorted state_dict handling." }, { "left": "issue:44483", "right": "issue:45290", "accept": false, - "reason": "Different chat/completions vs tokenizer chat template bugs; not the same code path." + "reason": "Both affect chat/completions-style input handling, but one is API request validation and the other is tool-call chat templating." }, { "left": "issue:44625", "right": "issue:46365", "accept": false, - "reason": "Qwen3.5 config propagation bug is unrelated to a Trainer import problem involving datasets." + "reason": "These are unrelated configuration propagation and import problems." }, { "left": "issue:43352", "right": "issue:43937", "accept": false, - "reason": "Nemotron FlashAttention support error is unrelated to GLM-5 generation config validation." + "reason": "Different model-specific issues: Flash Attention support for Nemotron versus GenerationConfig validation for GLM-5." }, { "left": "issue:44625", "right": "issue:46077", "accept": false, - "reason": "Different config propagation issues in different models; one is num_labels, the other is attention masking." + "reason": "Same model family, but different bugs: num_labels propagation versus attention-mask construction." }, { - "left": "issue:44355", - "right": "issue:46877", + "left": "issue:43611", + "right": "issue:44655", "accept": false, - "reason": "Compiled Python file execution errors are unrelated to auto_docstring emitting lint diagnostics at import time." + "reason": "Both touch model/pipeline APIs, but one is a v5 loading regression and the other is pipeline save_pretrained serialization." }, { - "left": "issue:43611", - "right": "issue:44655", + "left": "issue:44355", + "right": "issue:46877", "accept": false, - "reason": "Transformers 5.0 base_model_prefix loading regression is unrelated to Pipeline save_pretrained serialization." + "reason": "These are unrelated: compiled Python file errors versus import-time docstring lint output." }, { "left": "issue:36683", "right": "issue:43475", "accept": false, - "reason": "Different model families and symptoms: Gemma3Config missing vocab_size versus SAM 3 Video missing fpn_position_embeddings." + "reason": "Different models and failures: Gemma3Config missing vocab_size versus SAM 3 Video missing fpn_position_embeddings." }, { "left": "issue:44655", "right": "issue:46050", "accept": false, - "reason": "Both involve saving, but Pipeline save_pretrained and sinq quantized save/load failures are distinct code paths." + "reason": "Both are save/load issues, but one is Pipeline serialization and the other is sinq quantized model round-trip breakage." }, { "left": "issue:43976", "right": "issue:44938", "accept": false, - "reason": "Python version compatibility issue for 5.1.0 is not the same as Transformers failing to load in Python 3.14." + "reason": "Both are version/platform compatibility reports, but the Python version ranges and failure details do not match." }, { "left": "issue:44938", "right": "issue:47059", "accept": false, - "reason": "Both are environment/version breakages, but one is Python 3.14 loading and the other is a CLI incompatibility with huggingface_hub 1.22.0." + "reason": "Both are environment compatibility problems, but one is Python 3.14 loading and the other is a huggingface_hub CLI regression." }, { "left": "issue:36683", "right": "issue:44077", "accept": false, - "reason": "Gemma3Config missing vocab_size and patchtsmixer post_init validation are unrelated configuration bugs." + "reason": "Both are model/config regressions, but one is a missing config attribute and the other is an invalid optional post_init rule." }, { "left": "issue:43867", "right": "issue:45216", "accept": false, - "reason": "Model loading failure from sorted state_dict is unrelated to Qwen3.5 save_pretrained regression." + "reason": "Both mention loading/saving checkpoints, but the specific bugs are different and model-specific." }, { "left": "issue:36683", "right": "issue:47110", "accept": false, - "reason": "Gemma3Config vocab_size missing is unrelated to special_tokens being a list instead of a dict-like object." + "reason": "The titles indicate different failures: missing vocab_size versus special_tokens being a list instead of a mapping." }, { "left": "issue:46077", "right": "issue:46082", "accept": false, - "reason": "Gemma4 attention-mask construction and LlamaConfig head_dim validation are different model/config issues." + "reason": "Both are Gemma4 config issues, but one is attention-mask semantics and the other is a head_dim validation constraint." }, { "left": "issue:43618", "right": "issue:45216", "accept": false, - "reason": "CLIP attentions regression is unrelated to Qwen3.5 checkpoint saving correctness." + "reason": "Different regressions in different model paths: CLIP attentions versus Qwen3.5 checkpoint saving." }, { "left": "issue:36683", "right": "issue:43525", "accept": false, - "reason": "Gemma3Config missing vocab_size and Llama4Config missing pad_token_id are separate configuration attributes on different models." + "reason": "Both are missing-config-attribute bugs, but they affect different models and different attributes." }, { "left": "issue:29870", "right": "issue:36296", "accept": false, - "reason": "Bart generate() decoder_attention_mask bug is unrelated to a tensor parallel training bug." + "reason": "Both are training/generation bugs, but one is a BART generate() unbound variable and the other is a tensor-parallel training issue." }, { "left": "issue:43064", "right": "issue:44492", "accept": false, - "reason": "FSDP2/PEFT optimizer-state mismatch is unrelated to a typo in cache strategies." + "reason": "These are unrelated training-state versus typo/documentation issues." } ] }, @@ -18025,8 +18045,8 @@ "issue:43992|issue:45310", "issue:44075|issue:46365", "issue:44908|issue:46851", - "issue:29870|issue:33453", "issue:44393|issue:47000", + "issue:29870|issue:33453", "issue:43867|issue:46202", "issue:44355|issue:44855", "issue:43010|issue:45850", @@ -18052,220 +18072,220 @@ "issue:43010|issue:43065", "issue:43874|issue:46617" ], - "prepared_review_unit_hash": "bf6f3bae02d52265be5afa40531896629b37d5b869b13dd13b5cbb52f2d218b1", + "prepared_review_unit_hash": "e1736205b64d12d596a06a35b47bf06cf28a27f16e28d0750e2a18ed72c08683", "component_budget": { "node_count": 514, "item_count": 514, "soft_pair_count": 1891, - "serialized_chars": 980926, - "estimated_input_tokens": 245232, - "estimated_eval_tokens": 490720 + "serialized_chars": 980895, + "estimated_input_tokens": 245224, + "estimated_eval_tokens": 490704 }, "budget": { "node_count": 48, "item_count": 48, "soft_pair_count": 30, - "serialized_chars": 36303, - "estimated_input_tokens": 9076, - "estimated_eval_tokens": 18408 + "serialized_chars": 36307, + "estimated_input_tokens": 9077, + "estimated_eval_tokens": 18410 }, "overflow_policy": "truncate_then_skip", "trimmed": false, "aggressively_trimmed": false, "split": true, "analyst_result": { - "summary": "The cluster is mostly noisy and mixes unrelated transformer bugs; one clear subcluster is import-time regression/slow import complaints. I\u2019d use the broad import-regression report as the canonical issue, and there are no PRs here.", - "confidence": 0.46, - "canonical_issue_reason": "Issue 45278 is the broadest umbrella report in the set: it describes widespread import failures after a version upgrade, which best matches the cluster\u2019s import/regression subset and can absorb narrower import-related reports better than model-specific bugs.", + "summary": "Most pairs are unrelated and should stay separate. The only plausible duplicate is the import-time slowdown pair: one issue reports that `import transformers` is slow, and the other points to `import *` usage as the likely cause/fix.", + "confidence": 0.77, + "canonical_issue_reason": "issue:44246 is the clearest user-facing report of the import slowdown and is broad enough to serve as the anchor for the related performance regression discussion; issue:41669 is the more specific root-cause/fix-oriented counterpart.", "canonical_pr_reason": null, - "best_issue_reason": "Issue 45278 is the most representative user-facing report with the widest scope among the listed issues; it is concrete enough to anchor the cluster while still covering the general import regression theme.", + "best_issue_reason": "issue:44246 is the best overall issue candidate among the set because it captures the symptom most directly and has the strongest discussion/activity around the import-performance problem.", "best_pr_reason": null, "soft_edge_verdicts": [ { "left": "issue:43010", "right": "issue:45278", "accept": false, - "reason": "Different problems: cache/layer no_grad decoration vs general import failures after upgrade." + "reason": "Different bugs: cache/layer `update()` no-grad semantics vs broad import errors after upgrade." }, { "left": "issue:43992", "right": "issue:45310", "accept": false, - "reason": "Different model families and failures: missing UMT5 encoder weights vs Qwen3.5 MoE from_pretrained error." + "reason": "Different models and failure modes; one is UMT5 encoder weights, the other is Qwen3.5 MoE loading." }, { "left": "issue:44075", "right": "issue:46365", "accept": false, - "reason": "Optimizer argument handling and Trainer import failure are unrelated." + "reason": "Optimizer argument handling is unrelated to `Trainer` import failure." }, { "left": "issue:44908", "right": "issue:46851", "accept": false, - "reason": "Scheduler kwargs bug and Qwen3.5 variable-length position_ids bug are unrelated." + "reason": "Scheduler kwargs bug is unrelated to Qwen3.5 variable-length position-id handling." }, { - "left": "issue:29870", - "right": "issue:33453", + "left": "issue:44393", + "right": "issue:47000", "accept": false, - "reason": "BART generate decoder mask bug is not the same as tokenizer loading regression." + "reason": "Both are vision/detection-adjacent, but they are different concrete failures and code paths." }, { - "left": "issue:44393", - "right": "issue:47000", + "left": "issue:29870", + "right": "issue:33453", "accept": false, - "reason": "Both are vision/detection-related, but they are different failures in different code paths." + "reason": "Bart generate-time decoder mask bug is unrelated to tokenizer loading regression." }, { "left": "issue:43867", "right": "issue:46202", "accept": false, - "reason": "State-dict loading order issue and Gemma pipeline roundtrip issue are different bugs." + "reason": "Different loading/roundtrip problems; no clear shared underlying bug." }, { "left": "issue:44355", "right": "issue:44855", "accept": false, - "reason": "Both involve imports/compilation, but the root causes differ: compiled Python files vs a JIT parsing indentation error." + "reason": "Both involve import/compile-time failures, but the causes are different (`compiled` files vs `torch.jit.script` parsing on Python 3.13)." }, { "left": "issue:43010", "right": "issue:45850", "accept": false, - "reason": "General 'too many models break' complaint does not match cache update/no_grad decoration." + "reason": "No common code path; one is cache update semantics, the other is a broad complaints issue." }, { "left": "issue:36296", "right": "issue:43575", "accept": false, - "reason": "Tensor-parallel training bug and Qwen2 load-time OOM are not the same underlying issue." + "reason": "Tensor-parallel training bug vs OOM when loading a specific large model; not the same defect." }, { "left": "issue:43867", "right": "issue:45375", "accept": false, - "reason": "Missing state-dict key and missing config field are different model-loading regressions." + "reason": "Unrelated model-loading issue vs missing Qwen3.5 vision config field." }, { "left": "issue:41669", "right": "issue:44246", "accept": true, - "reason": "Both are about slow transformers import; one reports the symptom, the other identifies a concrete import-time cause." + "reason": "Same import-performance regression: `import transformers` is slow, and `import *` usage is presented as the underlying cause/fix for that slowdown." }, { "left": "issue:42890", "right": "issue:45561", "accept": false, - "reason": "Seed-related flaky integration test and xdist captured_info race are different test failures." + "reason": "Test flakiness from missing seeds is unrelated to xdist race on `captured_info.txt`." }, { "left": "issue:43618", "right": "issue:44077", "accept": false, - "reason": "CLIPOutput attentions assignment and patchtsmixer post_init constraints are unrelated model issues." + "reason": "Different model/API issues: CLIP attentions assignment vs patchtsmixer `post_init` policy." }, { "left": "issue:43010", "right": "issue:43976", "accept": false, - "reason": "No_grad decoration for caches is unrelated to Python-version load failures." + "reason": "No shared bug; cache/layer updates vs Python-version compatibility." }, { "left": "issue:43867", "right": "issue:45357", "accept": false, - "reason": "Different Qwen3.5 regressions: sorted state_dict loading vs incorrect visual encoder save keys." + "reason": "Generic model load error vs Qwen3.5 save-pretrained visual-key regression." }, { "left": "issue:43992", "right": "issue:45478", "accept": false, - "reason": "Same general area (Qwen3.5 loading), but the reported failures are different and not clearly the same bug." + "reason": "Different issues despite both being Qwen-related; `UMT5Encoder` weight loading is not the same as Qwen3.5 MoE `from_pretrained` failure." }, { "left": "issue:28282", "right": "issue:33453", "accept": false, - "reason": "PyTorch missing dependency error is unrelated to tokenizer loading regression." + "reason": "PyTorch-missing ImportError is unrelated to tokenizer-loading regression." }, { "left": "issue:43618", "right": "issue:46285", "accept": false, - "reason": "CLIP attentions assignment and CLIP text-model structure mismatch are different regressions." + "reason": "Different CLIP structure/attention issues; not the same underlying bug." }, { "left": "issue:44360", "right": "issue:45440", "accept": false, - "reason": "A DSA indexer/ReLU discussion is unrelated to DeepSeekV3MoE implementation divergence." + "reason": "A discussion about the DSA indexer is unrelated to DeepSeekV3MoE implementation divergence." }, { "left": "issue:42890", "right": "issue:44964", "accept": false, - "reason": "SAM HQ test seeding flake is unrelated to Phi-4 multimodal model loading." + "reason": "Unrelated: SamHQ test seed flakiness vs Phi-4 multimodal loading failure." }, { "left": "issue:42890", "right": "issue:44112", "accept": false, - "reason": "Different CI/test issues with different symptoms and scope." + "reason": "Test-seeding flakiness is unrelated to the GraniteSpeech device-override CI failure." }, { "left": "issue:43502", "right": "issue:44062", "accept": false, - "reason": "Local-files-only API requests and AddedToken keyword collision are unrelated tokenizer/loading bugs." + "reason": "Local-files-only network request bug is unrelated to `AddedToken` argument duplication." }, { "left": "issue:42898", "right": "issue:46664", "accept": false, - "reason": "Tokenization cleanup behavior change is unrelated to SmolVLM video processor frame semantics." + "reason": "Tokenization cleanup behavior change is unrelated to SmolVLM video `num_frames` semantics." }, { "left": "issue:44734", "right": "issue:46617", "accept": false, - "reason": "KV-cache continuation crash and cache FileNotFoundError are different runtime failures." + "reason": "KV-cache continuation indexing crash is unrelated to trust_remote_code cache FileNotFoundError." }, { "left": "issue:44222", "right": "issue:45125", "accept": false, - "reason": "FP8 save_pretrained issue and missing tensor-parallel plan are different MoE/tensor-parallel problems." + "reason": "Different Qwen3.5/MoE topics: FP8 save_pretrained vs missing tensor-parallel plan." }, { "left": "issue:43723", "right": "issue:44556", "accept": false, - "reason": "Both involve loading regressions after v5, but one is tokenizer loading and the other is checkpoint reload; too broad to treat as the same bug." + "reason": "Both are loading-related, but one is tokenizer-specific and the other is checkpoint reload after version upgrade; not the same concrete bug." }, { "left": "issue:43502", "right": "issue:44485", "accept": false, - "reason": "local_files_only bug is unrelated to GLM-5 RoPE implementation discussion." + "reason": "local_files_only API bug is unrelated to GLM-5 RoPE implementation discussion." }, { "left": "issue:43010", "right": "issue:43065", "accept": false, - "reason": "Cache/layer no_grad issue and dummy Conv2d in Sam3PixelDecoder are unrelated." + "reason": "Cache/layer update semantics are unrelated to Sam3PixelDecoder's dummy Conv2d." }, { "left": "issue:43874", "right": "issue:46617", "accept": false, - "reason": "Missing image-patch API and cache FileNotFoundError are different loading/runtime issues." + "reason": "ImageProcessorFast multimodal token counting bug is unrelated to cache FileNotFoundError." } ] }, "evaluator_result": { "accept": true, - "feedback": "The summary is grounded in the packet and stays conservative. The chosen canonical issue, 45278, is a reasonable umbrella for the import-regression/import-error subset, and the one accepted soft pair (41669 \u2194 44246) is plausibly the same import-slowness theme. No PR pairings are present, and the rejected pairs are generally conservative rather than overstated." + "feedback": "Grounded overall. The only accepted soft edge is the import-performance pair (41669/44246), which is well supported by the titles and activity; the other verdicts are conservatively rejected and do not overstate shared bugs." }, "evaluator_used": true, "retried": false, @@ -18336,13 +18356,13 @@ "issue:47110" ], "soft_pairs": [ - "issue:45850|issue:47110", "issue:43824|issue:46617", + "issue:45850|issue:47110", "issue:43576|issue:46132", "issue:44387|issue:44556", - "issue:42548|issue:44351", "issue:45125|issue:45478", "issue:30333|issue:33453", + "issue:42548|issue:44351", "issue:44355|issue:44749", "issue:30990|issue:31795", "issue:45125|issue:45375", @@ -18361,8 +18381,8 @@ "issue:43576|issue:47059", "issue:31795|issue:38617", "issue:43576|issue:43824", - "issue:44315|issue:44360", "issue:44961|issue:45468", + "issue:44315|issue:44360", "issue:44485|issue:44961", "issue:44351|issue:46617", "issue:43352|issue:44961", @@ -18372,271 +18392,271 @@ "issue:41628|issue:42548", "issue:45375|issue:47110", "issue:44222|issue:44230", - "issue:43867|issue:47110", - "issue:44315|issue:46209" + "issue:44315|issue:46209", + "issue:43867|issue:47110" ], - "prepared_review_unit_hash": "d289179a48b73d824aaea99b26317d6d7a73c5f83a245440a022a07d2c0188e1", + "prepared_review_unit_hash": "6df9452d9cf54538036f3907f51009574a8abfa7c4c6a26fa0ad599cb272f926", "component_budget": { "node_count": 514, "item_count": 514, "soft_pair_count": 1891, - "serialized_chars": 980926, - "estimated_input_tokens": 245232, - "estimated_eval_tokens": 490720 + "serialized_chars": 980895, + "estimated_input_tokens": 245224, + "estimated_eval_tokens": 490704 }, "budget": { "node_count": 48, "item_count": 48, "soft_pair_count": 38, - "serialized_chars": 38511, - "estimated_input_tokens": 9628, - "estimated_eval_tokens": 19512 + "serialized_chars": 38508, + "estimated_input_tokens": 9627, + "estimated_eval_tokens": 19510 }, "overflow_policy": "truncate_then_skip", "trimmed": false, "aggressively_trimmed": false, "split": true, "analyst_result": { - "summary": "This cluster is mostly heterogeneous transformer issues. The only clear duplicate-looking pair is the RoFormer bug around kwargs/attention-interface forwarding; the rest are separate bugs, regressions, or docs/import issues.", - "confidence": 0.91, - "canonical_issue_reason": "issue:46144 is the most concrete and code-path-specific RoFormer report, and it is open; it best represents the shared RoFormer forwarding/attention-interface problem with issue:46143.", + "summary": "Most candidate pairs are only loosely related by subsystem or version and do not look like the same underlying bug. The only plausible duplicate pair is the RoFormer attention/kwargs issue.", + "confidence": 0.84, + "canonical_issue_reason": "issue:46144 is the most concrete and code-path-specific RoFormer report; it names the attention-interface gap that likely explains the kwargs passthrough symptom in issue:46143.", "canonical_pr_reason": null, - "best_issue_reason": "issue:46144 is the best issue to anchor this cluster because it is the clearest, most actionable statement of the shared RoFormer bug and is still open.", + "best_issue_reason": "issue:46144 is the best representative issue here because it is specific, actionable, and points to the likely root cause rather than a downstream symptom.", "best_pr_reason": null, "soft_edge_verdicts": [ { - "left": "issue:45850", - "right": "issue:47110", + "left": "issue:43824", + "right": "issue:46617", "accept": false, - "reason": "Both mention broad breakage, but one is a generic v5 regression and the other is a specific special_tokens type error; different failure modes." + "reason": "Different failures: missing exported Qwen2_5_VL symbol vs trust_remote_code cache FileNotFoundError." }, { - "left": "issue:43824", - "right": "issue:46617", + "left": "issue:45850", + "right": "issue:47110", "accept": false, - "reason": "Different problems: missing exported Qwen2_5_VL class vs cache loading FileNotFoundError for trust_remote_code." + "reason": "One is a broad complaint about v5 regressions; the other is a specific special_tokens shape error." }, { "left": "issue:43576", "right": "issue:46132", "accept": false, - "reason": "One is the transformers env CLI, the other is AttentionInterface.register behavior; different surfaces and root causes." + "reason": "CLI/env command breakage is unrelated to AttentionInterface.register behavior changes." }, { "left": "issue:44387", "right": "issue:44556", "accept": false, - "reason": "Both are version regressions, but one is int4 memory/OOM and the other is checkpoint reload incompatibility; not the same bug." - }, - { - "left": "issue:42548", - "right": "issue:44351", - "accept": false, - "reason": "Both are import errors, but for different symbols and likely unrelated packaging/export paths." + "reason": "Int4 quantization OOM is a different problem from checkpoint reload incompatibility across versions." }, { "left": "issue:45125", "right": "issue:45478", "accept": false, - "reason": "Both involve Qwen3.5 MoE, but one is missing tensor-parallel plan and the other is a from_pretrained regression; not clearly the same issue." + "reason": "Both involve Qwen3.5 MoE, but one is tensor-parallel planning and the other is a from_pretrained loading error; not enough evidence of the same bug." }, { "left": "issue:30333", "right": "issue:33453", "accept": false, - "reason": "Unrelated: MLflow reporting on failed training vs tokenizer loading regression." + "reason": "MLFlow reporting on failed jobs is unrelated to tokenizer loading regression." + }, + { + "left": "issue:42548", + "right": "issue:44351", + "accept": false, + "reason": "Both are import errors, but for different symbols and likely different causes." }, { "left": "issue:44355", "right": "issue:44749", "accept": false, - "reason": "Completely different symptoms: compiled Python file errors vs slower filtering after upgrade." + "reason": "Compiled-file errors are unrelated to the data-filtering slowdown after upgrade." }, { "left": "issue:30990", "right": "issue:31795", "accept": false, - "reason": "Sentence-Transformers loading hang vs documentation confusion about forward args; no shared code bug." + "reason": "Loading hang and forward() documentation confusion do not describe the same bug." }, { "left": "issue:45125", "right": "issue:45375", "accept": false, - "reason": "Different Qwen3.5 MoE config/model issues: tensor-parallel metadata vs missing vision config field." + "reason": "Tensor-parallel plan support and a missing config field are distinct issues." }, { "left": "issue:43502", "right": "issue:44222", "accept": false, - "reason": "Local-files-only network leak vs FP8 moe save_pretrained issue; unrelated." + "reason": "Local-files-only network requests are unrelated to FP8 save_pretrained behavior." }, { "left": "issue:43502", "right": "issue:44961", "accept": false, - "reason": "Local-files-only API requests vs an unrelated issue titled 'racoon'." + "reason": "Local-files-only bug and an unrelated issue titled racoon are clearly different." }, { "left": "issue:43827", "right": "issue:44961", "accept": false, - "reason": "Docs mentioning pipeline() after v5 removals vs unrelated 'racoon' issue." + "reason": "Pipeline-removal docs cleanup is unrelated to racoon." }, { "left": "issue:46664", "right": "issue:46726", "accept": false, - "reason": "Video processor num_frames semantics vs AutoProcessor backend-missing image processor error; different layers." + "reason": "Video num_frames semantics and missing-vision-backend image processor handling are different bugs." }, { "left": "issue:43010", "right": "issue:43232", "accept": false, - "reason": "Both touch generation/caching, but one asks for no_grad on cache updates and the other is a sync_gpus generation bug." + "reason": "Both are generation-related, but one is about no_grad decoration of cache updates and the other about sync_gpus/model kwargs update ordering." }, { "left": "issue:30333", "right": "issue:36296", "accept": false, - "reason": "Unrelated training/reporting issue vs tensor parallel training bug." + "reason": "MLFlow reporting is unrelated to tensor-parallel training failures." }, { "left": "issue:43299", "right": "issue:44556", "accept": false, - "reason": "Both are loading regressions after version changes, but for different models and failure points." + "reason": "Qwen3VL MoE loading regression and general checkpoint reload regression are too broad and not clearly the same bug." }, { "left": "issue:29870", "right": "issue:31795", "accept": false, - "reason": "BART generate() UnboundLocalError vs forward-doc confusion; no shared underlying bug." + "reason": "Bart generate() UnboundLocalError is unrelated to forward() documentation wording." }, { "left": "issue:33453", "right": "issue:34689", "accept": false, - "reason": "Tokenizer loading regression vs Llama 3.2 Vision model loading breakage; different artifacts and code paths." + "reason": "Tokenizer loading regression and Llama 3.2 vision model loading breakage are different concrete failures." }, { "left": "issue:44355", "right": "issue:45071", "accept": false, - "reason": "Compiled Python file errors vs PretrainedConfig type-checking breakage; unrelated." + "reason": "Compiled-file execution errors are unrelated to PretrainedConfig type checking." }, { "left": "issue:42757", "right": "issue:43576", "accept": false, - "reason": "huggingface_hub import failure vs transformers env CLI breakage; different packages/features." + "reason": "huggingface_hub import failure and transformers env command breakage are different layers and symptoms." }, { "left": "issue:43576", "right": "issue:45071", "accept": false, - "reason": "CLI command breakage vs type-checking regression; unrelated." + "reason": "CLI command breakage is unrelated to type-checking behavior changes." }, { "left": "issue:43576", "right": "issue:47059", "accept": false, - "reason": "Both are CLI-related, but one is env command breakage and the other is a hub 1.22.0 completion API mismatch; not the same bug." + "reason": "Both mention CLI, but one is the env command in v5 and the other is a hub version-specific completion error." }, { "left": "issue:31795", "right": "issue:38617", "accept": false, - "reason": "Documentation confusion in a model forward method vs an import error for layer_type_validation." + "reason": "Documentation confusion and missing layer_type_validation import are unrelated." }, { "left": "issue:43576", "right": "issue:43824", "accept": false, - "reason": "Env CLI bug vs model import error; different domains." + "reason": "CLI env issues and missing model-class import are not the same bug." }, { - "left": "issue:44315", - "right": "issue:44360", + "left": "issue:44961", + "right": "issue:45468", "accept": false, - "reason": "Liger Kernel not applied with model_init vs DSA indexer missing ReLU; unrelated." + "reason": "Unrelated issue titles; no shared underlying bug is evident." }, { - "left": "issue:44961", - "right": "issue:45468", + "left": "issue:44315", + "right": "issue:44360", "accept": false, - "reason": "Unrelated issue titled 'racoon' vs Gemma-4 audio positional encoding bug." + "reason": "Liger Kernel application and DSA indexer ReLU are completely different areas." }, { "left": "issue:44485", "right": "issue:44961", "accept": false, - "reason": "GLM-5 RoPE implementation discussion vs unrelated 'racoon' issue." + "reason": "GLM-5 RoPE implementation discussion is unrelated to racoon." }, { "left": "issue:44351", "right": "issue:46617", "accept": false, - "reason": "Importing HybridCache vs FileNotFoundError when loading trust_remote_code from cache; different failure modes." + "reason": "HybridCache import failure and cached trust_remote_code FileNotFoundError are different import/load problems." }, { "left": "issue:43352", "right": "issue:44961", "accept": false, - "reason": "Flash Attention 2 support gap for Nemotron vs unrelated 'racoon' issue." + "reason": "Flash Attention 2 support error and racoon are unrelated." }, { "left": "issue:36296", "right": "issue:43901", "accept": false, - "reason": "Tensor parallel training bug vs docs mentioning return_all_scores; no shared bug." + "reason": "Tensor-parallel training bug and TextClassificationPipeline docs issue are not the same bug." }, { "left": "issue:43475", "right": "issue:46617", "accept": false, - "reason": "SAM 3 Video missing attribute vs trust_remote_code cache file missing; unrelated." + "reason": "SAM 3 Video attribute error and trust_remote_code cache FileNotFoundError are different failures." }, { "left": "issue:46143", "right": "issue:46144", "accept": true, - "reason": "Both report the same RoFormer implementation problem: kwargs/attention-interface forwarding is not respected in model methods and attention code." + "reason": "Both describe the same RoFormer attention-path problem: kwargs not propagating because the attention implementation does not use the attention interface." }, { "left": "issue:41628", "right": "issue:42548", "accept": false, - "reason": "Different import errors for different public symbols; not the same missing-export bug." + "reason": "AutoImageProcessor and PreTrainedModel import failures are different missing exports." }, { "left": "issue:45375", "right": "issue:47110", "accept": false, - "reason": "Qwen3_5MoeVisionConfig strict-field omission vs special_tokens list type error; unrelated." + "reason": "Missing deepstack_visual_indexes config field and special_tokens list/dict error are unrelated." }, { "left": "issue:44222", "right": "issue:44230", "accept": false, - "reason": "Both mention FP8/MoE, but one is save_pretrained behavior and the other is inference support; not mergeable as one fix." + "reason": "FP8 save_pretrained support and FP8 inference support are related broadly, but not the same concrete bug." }, { - "left": "issue:43867", - "right": "issue:47110", + "left": "issue:44315", + "right": "issue:46209", "accept": false, - "reason": "State-dict sorting load error vs special_tokens type error; different loading paths." + "reason": "model_init Liger Kernel application and FP8 inference on B200 are unrelated." }, { - "left": "issue:44315", - "right": "issue:46209", + "left": "issue:43867", + "right": "issue:47110", "accept": false, - "reason": "Liger Kernel/model_init issue vs FP8 inference on B200; unrelated." + "reason": "State-dict sorting load error and special_tokens key access error are different bugs." } ] }, "evaluator_result": { "accept": true, - "feedback": "The summary is grounded in the packet, and the one accepted soft-edge pair is reasonably conservative based on the RoFormer titles. The other rejected pairs are also appropriately cautious." + "feedback": "The summary is well grounded in the packet, and the only accepted duplicate pair (issue:46143 / issue:46144) is supported by the titles and the stated RoFormer attention-interface linkage. The remaining soft-edge rejections are appropriately conservative and do not overstate evidence." }, "evaluator_used": true, "retried": false, @@ -18685,7 +18705,6 @@ "issue:44556", "issue:44623", "issue:44704", - "issue:44829", "issue:44861", "issue:44908", "issue:45003", @@ -18722,8 +18741,8 @@ "issue:43576|issue:44556", "issue:43525|issue:45850", "issue:44556|issue:45092", - "issue:43572|issue:45850", "issue:44230|issue:46095", + "issue:43572|issue:45850", "issue:44492|issue:45850", "issue:45850|issue:47059", "issue:42371|issue:44492", @@ -18733,8 +18752,8 @@ "issue:42890|issue:44908", "issue:43576|issue:44908", "issue:43352|issue:44393", - "issue:43976|issue:44393", "issue:44295|issue:45230", + "issue:43976|issue:44393", "issue:44297|issue:45230", "issue:44908|issue:45071", "issue:44492|issue:46328", @@ -18745,241 +18764,270 @@ "issue:44077|issue:44861", "issue:31795|issue:33453", "issue:43576|issue:44273", - "issue:44360|issue:45468", - "issue:44315|issue:44829" + "issue:44360|issue:45468" ], - "prepared_review_unit_hash": "1fb00566925ccaec44695c9f1d15f43b644d6e1b72c978ac75dd7208b94a6058", + "prepared_review_unit_hash": "ca1cec2966b21f7439247281d99da6c246a1dbfe81fef01a51f402466fbeceb1", "component_budget": { "node_count": 514, "item_count": 514, "soft_pair_count": 1891, - "serialized_chars": 980926, - "estimated_input_tokens": 245232, - "estimated_eval_tokens": 490720 + "serialized_chars": 980895, + "estimated_input_tokens": 245224, + "estimated_eval_tokens": 490704 }, "budget": { - "node_count": 48, - "item_count": 48, - "soft_pair_count": 40, - "serialized_chars": 39113, - "estimated_input_tokens": 9779, - "estimated_eval_tokens": 19814 + "node_count": 47, + "item_count": 47, + "soft_pair_count": 39, + "serialized_chars": 38165, + "estimated_input_tokens": 9542, + "estimated_eval_tokens": 19340 }, "overflow_policy": "truncate_then_skip", "trimmed": false, "aggressively_trimmed": false, "split": true, "analyst_result": { - "summary": "These items are mostly unrelated issue reports. Several are superficially similar around v5 regressions, import/lazy-loading, or model-specific failures, but they do not share a concrete enough code path to treat as duplicates. I would not merge any of the soft edges.", - "confidence": 0.88, - "canonical_issue_reason": "issue:45850 is the broadest umbrella among the v5-regression reports, but it is too vague and high-level to safely absorb the others as duplicates.", + "summary": "These issues are mostly separate regressions or feature bugs across different models/subsystems; the soft links are topic-similar but not true duplicates. The broad v5 regression meta-report is the closest representative, but there is no strong single canonical bug here.", + "confidence": 0.84, + "canonical_issue_reason": "issue:45850 is the broadest umbrella report about many Transformers v5 breakages, so it best represents the cluster overall.", "canonical_pr_reason": null, - "best_issue_reason": "issue:45850 is the best global representative only because it captures the broadest symptom class ('v5 breaks too many models'), but it is still too generic to be a strong canonical target.", + "best_issue_reason": "issue:45850 is the most general and central issue in this set; the rest are narrower, distinct failures.", "best_pr_reason": null, "soft_edge_verdicts": [ { "left": "issue:43525", "right": "issue:47110", "accept": false, - "reason": "Different failures: missing pad_token_id on Llama4Config vs special_tokens being a list without keys()." + "reason": "Different failures: missing pad_token_id on Llama4Config vs a special_tokens structure error." }, { "left": "issue:45362", "right": "issue:45850", "accept": false, - "reason": "One is a specific Qwen3.5 chat crash; the other is a broad v5 breakage complaint with no shared code path." + "reason": "One is a Qwen3.5 chat crash; the other is a broad v5 complaint, not the same bug." }, { "left": "issue:43673", "right": "issue:44492", "accept": false, - "reason": "Different subsystems: chunked_prefill cache handling vs a typo in cache strategy docs/code." + "reason": "Chunked-prefill cache regression and a cache-strategy typo are unrelated." }, { "left": "issue:44485", "right": "issue:46664", "accept": false, - "reason": "GLM-5 RoPE implementation discussion is unrelated to SmolVLM video num_frames semantics." + "reason": "GLM-5 RoPE implementation discussion vs SmolVLM video num_frames semantics." }, { "left": "issue:44273", "right": "issue:45542", "accept": false, - "reason": "Lazy loading issues and a TensorBoard-without-TF backend error are not the same bug." + "reason": "Lazy loading malfunction and a TensorFlow/tensorboard backend error are different issues." }, { "left": "issue:44623", "right": "issue:45230", "accept": false, - "reason": "Processor save_pretrained missing files is not the same as a generic bug report." + "reason": "A specific processor.save_pretrained missing-files bug vs a generic placeholder report." }, { "left": "issue:44038", "right": "issue:45125", "accept": false, - "reason": "Different model-specific problems: Qwen3-VL-Moe bug vs missing tensor-parallel plan support." + "reason": "Qwen3-VL-Moe bug report vs missing _tp_plan for tensor parallelism." }, { "left": "issue:44273", "right": "issue:45003", "accept": false, - "reason": "Both touch import behavior, but lazy loading failure and unsafe sys.modules access are distinct code paths." + "reason": "Lazy loading failure and unsafe sys.modules access are distinct code paths." }, { "left": "issue:44360", "right": "issue:46664", "accept": false, - "reason": "DSA indexer ReLU discussion is unrelated to a video processor num_frames mismatch." + "reason": "DSA indexer activation issue vs SmolVLM video frame handling." }, { "left": "issue:41669", "right": "issue:43673", "accept": false, - "reason": "Import-star performance regression and missing GenerationMixin cache are different bugs." + "reason": "Import-* performance regression and missing GenerationMixin cache are not the same problem." }, { "left": "issue:44393", "right": "issue:45071", "accept": false, - "reason": "Qwen3-VL 2D bbox output bug is unrelated to PretrainedConfig type checking." + "reason": "Qwen3-VL bounding-box output bug vs PretrainedConfig type-checking regression." }, { "left": "issue:44393", "right": "issue:45753", "accept": false, - "reason": "Same model family, but hallucination/error in bbox output and an infinite-loop image case are not clearly the same defect." + "reason": "Bounding-box hallucination/error and an infinite-loop-on-image bug are different failures." }, { "left": "issue:43576", "right": "issue:44556", "accept": false, - "reason": "Transformers env command breakage and checkpoint reload incompatibility after upgrade are different failures." + "reason": "Broken `transformers env` command and checkpoint reload incompatibility are separate v5 regressions." }, { "left": "issue:43525", "right": "issue:45850", "accept": false, - "reason": "A specific Llama4Config attribute error is not the same as a generic statement that v5 breaks many models." + "reason": "Specific Llama4 config attribute bug vs a broad multi-model v5 breakage report." }, { "left": "issue:44556", "right": "issue:45092", "accept": false, - "reason": "Both are upgrade incompatibilities, but they affect different models and likely different initialization paths." + "reason": "Checkpoint reload after upgrade and remote-code meta-init incompatibility are different load-path issues." }, { - "left": "issue:43572", - "right": "issue:45850", + "left": "issue:44230", + "right": "issue:46095", "accept": false, - "reason": "Missing pad_token_idx in StableLmConfig is a specific config bug, not the broad v5 regression umbrella." + "reason": "FP8 inference support for Qwen3-VL/Qwen3.5 vs DeepSeekV4 weight conversion tooling." }, { - "left": "issue:44230", - "right": "issue:46095", + "left": "issue:43572", + "right": "issue:45850", "accept": false, - "reason": "fp8 support for Qwen models and a deepseekv4 weight-conversion question are unrelated feature requests." + "reason": "StableLmConfig pad_token_idx regression is a specific config bug, not the general v5 umbrella report." }, { "left": "issue:44492", "right": "issue:45850", "accept": false, - "reason": "A typo in cache strategies is too specific to be the same as a broad v5 breakage report." + "reason": "A cache-strategy typo is not the same as the broad v5 breakage meta-issue." }, { "left": "issue:45850", "right": "issue:47059", "accept": false, - "reason": "General v5 breakage and a huggingface_hub CLI compatibility failure are different issues." + "reason": "Broad v5 breakage report vs a CLI incompatibility with huggingface_hub 1.22.0." }, { "left": "issue:42371", "right": "issue:44492", "accept": false, - "reason": "TF32 API guidance and cache strategy typo are unrelated." + "reason": "TF32 settings deprecation and cache-strategy typo are unrelated." }, { "left": "issue:41669", "right": "issue:46218", "accept": false, - "reason": "Import-star slowdown and failure to import Qwen3_5ForCausalLM are both import-related, but not the same concrete bug." + "reason": "Import-* slowdown and failure to import Qwen3_5ForCausalLM are different symptoms." }, { "left": "issue:43408", "right": "issue:44315", "accept": false, - "reason": "sam3_video/sam3_tracker warning and Liger Kernel not applied with model_init are unrelated subsystems." + "reason": "sam3_tracker model-type warning and Liger Kernel not applied with model_init are unrelated." }, { "left": "issue:44246", "right": "issue:44273", "accept": false, - "reason": "Both mention import/lazy-loading behavior, but one is intermittent import slowness and the other is a lazy-loading correctness issue." + "reason": "General import slowdown and lazy loading failure are both performance/loading themes but not the same bug." }, { "left": "issue:42890", "right": "issue:44908", "accept": false, - "reason": "Missing set_seed calls in tests and inverse_sqrt scheduler kwargs handling are unrelated." + "reason": "Flaky SamHQ tests from missing seeds vs scheduler kwargs being ignored." }, { "left": "issue:43576", "right": "issue:44908", "accept": false, - "reason": "transformers env CLI breakage is unrelated to scheduler keyword propagation." + "reason": "Broken env command and inverse_sqrt kwargs handling are different regressions." + }, + { + "left": "issue:43352", + "right": "issue:44393", + "accept": false, + "reason": "Flash Attention 2 support error for Nemotron vs Qwen3-VL bounding-box output bug." + }, + { + "left": "issue:44295", + "right": "issue:45230", + "accept": false, + "reason": "position_ids buffer access error vs a generic placeholder issue." + }, + { + "left": "issue:43976", + "right": "issue:44393", + "accept": false, + "reason": "Python version compatibility break vs Qwen3-VL bbox hallucination." + }, + { + "left": "issue:44297", + "right": "issue:45230", + "accept": false, + "reason": "tokenizer_class mismatch on save_pretrained vs a generic bug report." + }, + { + "left": "issue:44908", + "right": "issue:45071", + "accept": false, + "reason": "Scheduler kwargs bug and PretrainedConfig type checking are unrelated." + }, + { + "left": "issue:44492", + "right": "issue:46328", + "accept": false, + "reason": "Cache-strategy typo and SequenceFeatureExtractor.pad performance issue are unrelated." }, { "left": "issue:41669", "right": "issue:44351", "accept": false, - "reason": "Wildcard imports causing slow import and missing HybridCache export are different API/export problems." + "reason": "Import-* slowdown and missing HybridCache import are different import problems." }, { "left": "issue:44704", "right": "issue:44908", "accept": false, - "reason": "AutoProcessor cached_file kwargs forwarding and scheduler kwargs handling are unrelated." + "reason": "AutoProcessor cached_file kwargs propagation and scheduler kwargs handling are unrelated." }, { "left": "issue:41669", "right": "issue:45341", "accept": false, - "reason": "Import-time slowdown and a testing_utils bug do not share the same underlying change." + "reason": "Import-* slowdown and a testing_utils bug are not the same issue." }, { "left": "issue:43976", "right": "issue:46218", "accept": false, - "reason": "Python version compatibility breakage and a model import failure are not the same issue." + "reason": "Python 3.9/3.10 compatibility break and missing Qwen3_5ForCausalLM import are different." }, { "left": "issue:44077", "right": "issue:44861", "accept": false, - "reason": "Optional post_init policy and _get_tied_weight_keys crashing on list input are distinct code-path bugs." + "reason": "patchtsmixer post_init policy and _get_tied_weight_keys AttributeError are unrelated." }, { "left": "issue:31795", "right": "issue:33453", "accept": false, - "reason": "Documentation confusion about forward args is unrelated to tokenizer loading regression." + "reason": "Documentation confusion in model.forward vs tokenizer loading regression are different problems." }, { "left": "issue:43576", "right": "issue:44273", "accept": false, - "reason": "CLI env command breakage and lazy loading malfunction are both v5-era issues, but not the same bug." + "reason": "Broken env command and lazy loading malfunction are both v5-related but not the same code-path bug." }, { "left": "issue:44360", "right": "issue:45468", "accept": false, - "reason": "DSA indexer discussion is unrelated to Gemma4AudioRelPositionalEncoding." - }, - { - "left": "issue:44315", - "right": "issue:44829", - "accept": false, - "reason": "Liger Kernel application with model_init and flash_attention_3 training degeneration are different acceleration/training problems." + "reason": "DSA indexer activation issue and Gemma-4 audio positional encoding bug are unrelated." } ] }, @@ -19022,16 +19070,17 @@ "issue:43576", "issue:43618", "issue:43704", - "issue:43717", "issue:44230", "issue:44261", "issue:44263", "issue:44273", + "issue:44315", "issue:44351", "issue:44485", "issue:44492", "issue:44623", "issue:44655", + "issue:44829", "issue:44861", "issue:44908", "issue:44938", @@ -19044,7 +19093,6 @@ "issue:45850", "issue:46018", "issue:46095", - "issue:46144", "issue:46172", "issue:46218", "issue:46655", @@ -19054,12 +19102,13 @@ ], "soft_pairs": [ "issue:43519|issue:46655", + "issue:44315|issue:44829", "issue:44273|issue:44908", "issue:45230|issue:45310", "issue:43519|issue:44485", "issue:45468|issue:46894", - "issue:44655|issue:44861", "issue:43576|issue:46218", + "issue:44655|issue:44861", "issue:43519|issue:44908", "issue:41669|issue:43704", "issue:31795|issue:34689", @@ -19094,295 +19143,294 @@ "issue:30333|issue:31515", "issue:31515|issue:31795", "issue:44230|issue:45310", - "issue:43298|issue:44230", - "issue:43717|issue:46144" + "issue:43298|issue:44230" ], - "prepared_review_unit_hash": "60b72ff4910b5a0192435d667ed106d96bce315ca33fd918b0f8ff8717ddd7e9", + "prepared_review_unit_hash": "1fbe5c4aa7ad398858be3714edb9dc3f4843ec27bd86de6fba39536afc5ab5b1", "component_budget": { "node_count": 514, "item_count": 514, "soft_pair_count": 1891, - "serialized_chars": 980926, - "estimated_input_tokens": 245232, - "estimated_eval_tokens": 490720 + "serialized_chars": 980895, + "estimated_input_tokens": 245224, + "estimated_eval_tokens": 490704 }, "budget": { "node_count": 48, "item_count": 48, "soft_pair_count": 43, - "serialized_chars": 39689, - "estimated_input_tokens": 9923, - "estimated_eval_tokens": 20102 + "serialized_chars": 39821, + "estimated_input_tokens": 9956, + "estimated_eval_tokens": 20168 }, "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 issue reports with only superficial keyword overlap (loading, tokenizer, model-specific bugs, Flash Attention, v5 regressions). I did not find any pair that clearly describes the same underlying bug/change, so no soft duplicates are accepted.", - "confidence": 0.72, - "canonical_issue_reason": "issue:43519 is the strongest canonical issue candidate because it is concrete, well-scoped, and actionable, with a clear failing code path in a specific processor/model rather than a vague or generic report.", + "summary": "This set is mostly heterogeneous Transformers issues, and the soft-similarity links look like broad topical overlap rather than true duplicates. I would not merge any of the candidate pairs; the only reasonable canonical choice is a concrete, well-scoped bug report rather than one of the vague or generic issues.", + "confidence": 0.61, + "canonical_issue_reason": "issue:43519 is the most concrete and actionable report in the set, and it also sits at the center of several soft-similarity candidates, so it is the best representative issue if one must be chosen.", "canonical_pr_reason": null, - "best_issue_reason": "issue:43519 is also the best global issue choice: it has a precise title and a specific repro target, making it the cleanest representative issue in an otherwise noisy set.", + "best_issue_reason": "issue:43519 has a specific failure mode, clear subsystem ownership, and enough detail to serve as the most useful anchor issue in this otherwise mixed group.", "best_pr_reason": null, "soft_edge_verdicts": [ { "left": "issue:43519", "right": "issue:46655", "accept": false, - "reason": "Different multimodal processor bugs: timestamp calculation vs video padding." + "reason": "Both are processor bugs, but one is timestamp math in Qwen3VL and the other is padding in SmolVLM video processing; different code paths and failure modes." + }, + { + "left": "issue:44315", + "right": "issue:44829", + "accept": false, + "reason": "Liger Kernel application during model_init is unrelated to flash_attention_3 degenerating training; different mechanisms and fixes." }, { "left": "issue:44273", "right": "issue:44908", "accept": false, - "reason": "Unrelated topics: lazy loading vs scheduler kwargs handling." + "reason": "Lazy loading behavior and inverse_sqrt scheduler kwargs are unrelated subsystems." }, { "left": "issue:45230", "right": "issue:45310", "accept": false, - "reason": "One is a generic bug report, the other a specific Qwen3.5 MoE from_pretrained failure." + "reason": "Generic bug report versus a specific Qwen3.5 MoE from_pretrained regression; no shared underlying defect." }, { "left": "issue:43519", "right": "issue:44485", "accept": false, - "reason": "Both are model-specific correctness bugs, but they affect different code paths: Qwen3VL timestamps vs GLM-5 RoPE." + "reason": "Both mention model internals, but one is timestamp calculation and the other is RoPE implementation for GLM-5; not the same bug." }, { "left": "issue:45468", "right": "issue:46894", "accept": false, - "reason": "Both are audio-related, but they describe different models and different bugs." + "reason": "Gemma-4 audio positional encoding and CohereASR training-loss shift bug are different models and different failures." }, { - "left": "issue:44655", - "right": "issue:44861", + "left": "issue:43576", + "right": "issue:46218", "accept": false, - "reason": "Serialization/save_pretrained failure is not the same as _tied_weight_keys attribute handling." + "reason": "Broken CLI env command and inability to import Qwen3_5ForCausalLM are unrelated." }, { - "left": "issue:43576", - "right": "issue:46218", + "left": "issue:44655", + "right": "issue:44861", "accept": false, - "reason": "Both are import/bootstrap problems, but they involve different commands and different missing symbols/modules." + "reason": "Saving Pipeline objects and _get_tied_weight_keys AttributeError are different save/load code paths." }, { "left": "issue:43519", "right": "issue:44908", "accept": false, - "reason": "Timestamp calculation bug and scheduler kwargs bug are unrelated." + "reason": "Timestamp calculation in a processor versus scheduler kwarg handling are unrelated." }, { "left": "issue:41669", "right": "issue:43704", "accept": false, - "reason": "Import-* cleanup/perf issue is unrelated to a VRAM leak in Qwen3." + "reason": "Import-star cleanup/performance issue is unrelated to VRAM leakage in Qwen3ForCausalLM." }, { "left": "issue:31795", "right": "issue:34689", "accept": false, - "reason": "Documentation confusion about forward args is unrelated to a model loading regression." + "reason": "Documentation confusion around forward arguments is not the same as a Llama 3.2 90B Vision model-loading regression." }, { "left": "issue:44485", "right": "issue:45468", "accept": false, - "reason": "Different model families and different numerical/architectural bugs." + "reason": "GLM-5 RoPE implementation and Gemma-4 audio positional encoding are distinct model-specific issues." }, { "left": "issue:43704", "right": "issue:44485", "accept": false, - "reason": "VRAM leak in Qwen3 is unrelated to GLM-5 RoPE implementation." + "reason": "VRAM leak in Qwen3ForCausalLM is not the same problem as GLM-5 RoPE behavior." }, { "left": "issue:46018", "right": "issue:46172", "accept": false, - "reason": "Both mention DeepSeek-V4, but one is a gating logic bug and the other is a tensor conversion/transpose issue." + "reason": "DeepSeek-V4 gating logic and tensor transpose behavior are unrelated." }, { "left": "issue:43502", "right": "issue:43519", "accept": false, - "reason": "Local-files-only network behavior is unrelated to Qwen3VL timestamp math." + "reason": "local_files_only network leakage is a loading/download issue, not a processor timestamp bug." }, { "left": "issue:42371", "right": "issue:43704", "accept": false, - "reason": "TF32 configuration guidance is unrelated to a VRAM leak bug." + "reason": "TF32 settings guidance and VRAM leakage are unrelated." }, { "left": "issue:45850", "right": "issue:46218", "accept": false, - "reason": "A broad v5 breakage complaint is not the same as a specific module import failure." + "reason": "Broad complaint about version 5 breaking models does not match a specific Qwen3.5 import failure." }, { "left": "issue:43519", "right": "issue:44623", "accept": false, - "reason": "Processor timestamp math and processor save_pretrained file omissions are different bugs." + "reason": "Processor timestamp math and missing save_pretrained files are different bugs in different APIs." }, { "left": "issue:44263", "right": "issue:45468", "accept": false, - "reason": "Different models and different code paths: GlmMoeDsaIndexer split values vs Gemma audio positional encoding." + "reason": "torch.split return values in GlmMoeDsaIndexer and Gemma-4 audio positional encoding are unrelated." }, { "left": "issue:44261", "right": "issue:45468", "accept": false, - "reason": "RMS norm epsilon precision issue is unrelated to Gemma audio positional encoding." + "reason": "Missing rms_norm_eps in MLA q_a_layernorm and Gemma-4 audio positional encoding are not the same issue." }, { "left": "issue:43298", "right": "issue:46095", "accept": false, - "reason": "Continuous batching audio support and DeepSeek-V4 weight conversion are unrelated concerns." + "reason": "Continuous batching audio support and DeepSeek-V4 weight conversion are different feature requests." }, { "left": "issue:43519", "right": "issue:43704", "accept": false, - "reason": "Different subsystems: processor timestamp bug vs multi-thread VRAM leak." + "reason": "Processor timestamp calculation and VRAM leakage are different bug classes." }, { "left": "issue:43352", "right": "issue:46218", "accept": false, - "reason": "Flash Attention 2 support in one model is unrelated to a module import error in another model." + "reason": "NemotronHForCausalLM flash-attn support and Qwen3_5 import failure are separate model-specific issues." }, { "left": "issue:44351", "right": "issue:46218", "accept": false, - "reason": "Both are import errors, but they are for different symbols and likely different breakages." + "reason": "HybridCache import failure and Qwen3_5 module import failure are both import-related, but not the same missing symbol or code path." }, { "left": "issue:44861", "right": "issue:47110", "accept": false, - "reason": "Tied-weight key handling and special_tokens list-key access are unrelated errors." + "reason": "Tied-weight key handling and special_tokens list/dict mismatch are different attribute/type errors." }, { "left": "issue:42913", "right": "issue:43066", "accept": false, - "reason": "Both are tokenizer v5 regressions, but one is a vague behavior change and the other is a specific decoder-type bug." + "reason": "Both concern tokenizer v5 behavior, but one is a broad regression report and the other is a specific decoder-type bug; not clearly the same underlying defect." }, { "left": "issue:44861", "right": "issue:47059", "accept": false, - "reason": "Tied-weight metadata handling is unrelated to the CLI completion attribute error." + "reason": "AttributeError in tied weights logic and CLI completion breakage from huggingface_hub are unrelated." }, { "left": "issue:28348", "right": "issue:45250", "accept": false, - "reason": "Both mention Flash Attention 2.0, but one is a GPT2 support request and the other is an unspecific standalone issue." + "reason": "GPT2 flash attention 2.0 support request and generic flash attention 2.0 issue are too broad to be the same concrete bug." }, { "left": "issue:44908", "right": "issue:44998", "accept": false, - "reason": "Scheduler kwargs bug and the unrelated 'Unemployment' issue do not match." + "reason": "Scheduler kwarg handling and an unrelated low-information issue title do not indicate the same bug." }, { "left": "issue:44492", "right": "issue:44998", "accept": false, - "reason": "Typo in cache strategies is unrelated to the unrelated placeholder issue." + "reason": "Cache strategy typo and the unrelated generic issue are not duplicates." }, { "left": "issue:44998", "right": "issue:45850", "accept": false, - "reason": "A generic '5.0 breaks models' complaint is not the same as an unrelated placeholder issue." + "reason": "Generic 'Unemployment' issue and broad '5.0 breaks too many models' complaint are not the same underlying problem." }, { "left": "issue:43618", "right": "issue:44998", "accept": false, - "reason": "CLIPOutput attentions regression is unrelated to the placeholder issue." + "reason": "CLIPOutput attentions regression is unrelated to the generic issue." }, { "left": "issue:44938", "right": "issue:44998", "accept": false, - "reason": "Python 3.14 load failure is unrelated to the placeholder issue." + "reason": "Python 3.14 load failure is unrelated to the generic issue." }, { "left": "issue:44861", "right": "issue:45290", "accept": false, - "reason": "Tied-weight key crash and chat template tool-call crash are different failure modes." + "reason": "Tied weight key crash and chat template tool-call crash are different code paths." }, { "left": "issue:31515", "right": "issue:38617", "accept": false, - "reason": "Checkpoint loading slowness is not the same as an import error for layer_type_validation." + "reason": "Slow from_pretrained checkpoint loading and missing layer_type_validation import are unrelated." }, { "left": "issue:44230", "right": "issue:44623", "accept": false, - "reason": "FP8 inference support request is unrelated to processor save_pretrained file omissions." + "reason": "FP8 support request for Qwen3-VL/Qwen3.5 MoE is not the same as processor.save_pretrained missing files." }, { "left": "issue:28348", "right": "issue:43352", "accept": false, - "reason": "Both are Flash Attention 2.0 feature gaps, but they target different models and are not the same change." + "reason": "Both mention flash attention 2.0, but they target different models and different support gaps." }, { "left": "issue:29870", "right": "issue:31515", "accept": false, - "reason": "BART generate mask initialization bug and from_pretrained slowness are unrelated." + "reason": "BART generate() decoder_attention_mask bug and slow checkpoint loading are unrelated." }, { "left": "issue:31515", "right": "issue:36296", "accept": false, - "reason": "Checkpoint loading performance and tensor-parallel training correctness are unrelated." + "reason": "Checkpoint load slowness and tensor parallel training bug are different subsystems." }, { "left": "issue:30333", "right": "issue:31515", "accept": false, - "reason": "MLflow reporting on failed jobs is unrelated to checkpoint loading speed." + "reason": "MLFlow reporting on failed training and checkpoint loading slowness are unrelated." }, { "left": "issue:31515", "right": "issue:31795", "accept": false, - "reason": "Loading performance issue and documentation confusion are unrelated." + "reason": "Load performance issue and forward-API documentation confusion are not the same bug." }, { "left": "issue:44230", "right": "issue:45310", "accept": false, - "reason": "FP8 inference support request and Qwen3.5 MoE from_pretrained failure are different issues." + "reason": "FP8 support request and Qwen3.5 MoE from_pretrained error are different reports." }, { "left": "issue:43298", "right": "issue:44230", "accept": false, - "reason": "Continuous batching audio support and FP8 support are different feature areas." - }, - { - "left": "issue:43717", - "right": "issue:46144", - "accept": false, - "reason": "Init-weights usage in Mamba is unrelated to RoFormer attention interface usage." + "reason": "Continuous batching support for audio models and FP8 support for Qwen3 VL/MoE are different feature requests." } ] }, @@ -19410,12 +19458,14 @@ "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", @@ -19428,100 +19478,106 @@ "issue:37934|issue:43668", "issue:42936|issue:43668" ], - "prepared_review_unit_hash": "96617343b4b632324a1949ecefeb3ab9ccefd32e7f588bcf87e285a6b7e29469", + "prepared_review_unit_hash": "7cffb2df2d35e5ca4cdaa220cb5526b42e91319e1799a96559fb9dbe8854eef8", "component_budget": { "node_count": 514, "item_count": 514, "soft_pair_count": 1891, - "serialized_chars": 980926, - "estimated_input_tokens": 245232, - "estimated_eval_tokens": 490720 + "serialized_chars": 980895, + "estimated_input_tokens": 245224, + "estimated_eval_tokens": 490704 }, "budget": { - "node_count": 8, - "item_count": 8, - "soft_pair_count": 11, - "serialized_chars": 7777, - "estimated_input_tokens": 1945, - "estimated_eval_tokens": 4146 + "node_count": 9, + "item_count": 9, + "soft_pair_count": 12, + "serialized_chars": 8608, + "estimated_input_tokens": 2152, + "estimated_eval_tokens": 4560 }, "overflow_policy": "truncate_then_skip", "trimmed": false, "aggressively_trimmed": false, "split": true, "analyst_result": { - "summary": "These issues are not duplicates; they span unrelated models and bug classes (normalization, loss plumbing, tokenizer warnings, quantization typo, DeepSeek gating, RoFormer attention). The cluster should be split rather than merged around a single canonical item.", - "confidence": 0.95, - "canonical_issue_reason": "No single canonical issue fits this cluster because the items describe different underlying bugs in different components, with no shared target or code path.", + "summary": "This cluster is heterogeneous: the issues cover unrelated bugs in different models/components (norms, tokenizers, loss kwargs, quantization typos, attention interfaces). None of the soft-similarity pairs look like true duplicates or the same underlying fix.", + "confidence": 0.97, + "canonical_issue_reason": "No clear canonical issue exists; the items are about distinct code paths and models rather than one shared bug.", "canonical_pr_reason": null, - "best_issue_reason": "No best issue for the whole cluster; the safest triage is to keep these as separate issues rather than selecting one representative.", + "best_issue_reason": "No single issue is a better global representative than the others because the cluster does not center on one problem.", "best_pr_reason": null, "soft_edge_verdicts": [ + { + "left": "issue:43717", + "right": "issue:46144", + "accept": false, + "reason": "Different model subsystems and failure modes: Mamba/Mamba-2 init_weights behavior vs RoFormer attention interface usage." + }, { "left": "issue:42936", "right": "issue:46018", "accept": false, - "reason": "Mask2former ignore_value handling and DeepSeek-V4 shared expert gating are unrelated bugs in different models/code paths." + "reason": "Mask2Former ignore_value handling is unrelated to DeepSeek-V4 shared-expert gating logic." }, { "left": "issue:43668", "right": "issue:44031", "accept": false, - "reason": "An incorrect type hint in ModernBERTConfig is unrelated to the tokenizer regex warning regression." + "reason": "A config type-hint bug in ModernBERT is unrelated to tokenizer regex warning behavior." }, { "left": "issue:42936", "right": "issue:43240", "accept": false, - "reason": "Mask2former ignore_value usage is a model-specific loss/config bug; passing kwargs to a loss function is a separate API-plumbing issue." + "reason": "One is about Mask2Former's ignore_value being unused; the other is about kwargs not reaching a loss function." }, { "left": "issue:43240", "right": "issue:46144", "accept": false, - "reason": "Missing kwargs in loss calculation and RoFormer attention not using the attention interface affect different subsystems and code paths." + "reason": "Loss-function kwargs plumbing and RoFormer attention interface usage are different code paths and not the same bug." }, { "left": "issue:42936", "right": "issue:44031", "accept": false, - "reason": "Mask2former ignore_value behavior has nothing to do with tokenizer regex warnings after 4.57.3." + "reason": "Mask2Former ignore_value handling does not match tokenizer regex warning regressions." }, { "left": "issue:42936", "right": "issue:46144", "accept": false, - "reason": "These are different model bugs: one concerns Mask2former ignore_value, the other RoFormer attention implementation." + "reason": "These concern unrelated areas: segmentation ignore_value vs RoFormer attention implementation." }, { "left": "issue:42936", "right": "issue:43813", "accept": false, - "reason": "Mask2former loss/config behavior and a quantization-operation typo are unrelated." + "reason": "Mask2Former loss/ignore_value handling is unrelated to a quantization attribute typo." }, { "left": "issue:43668", "right": "issue:43813", "accept": false, - "reason": "ModernBERT norm_eps type hint and the orig_conversion.quantization_operation typo are distinct, non-overlapping issues." + "reason": "ModernBERT norm_eps typing and orig_conversion.quantization_operation typo are separate issues." }, { "left": "issue:43240", "right": "issue:43668", "accept": false, - "reason": "Loss-function kwargs plumbing and a config type-hint mistake are different kinds of defects in separate areas." + "reason": "Loss kwargs propagation is unrelated to a ModernBERTConfig type-hint correction." }, { "left": "issue:37934", "right": "issue:43668", "accept": false, - "reason": "Llama4TextL2Norm possibly being RMSNorm is a model-implementation question, not the ModernBERT norm_eps type-hint issue." + "reason": "Llama4TextL2Norm implementation question is unrelated to ModernBERT norm_eps typing." }, { "left": "issue:42936", "right": "issue:43668", "accept": false, - "reason": "Mask2former ignore_value usage and ModernBERT config typing are unrelated and do not share a code path." + "reason": "Mask2Former ignore_value usage and ModernBERT config typing do not share the same underlying defect." } ] }, @@ -19554,52 +19610,52 @@ "issue:10519|issue:11005", "issue:11005|issue:11307" ], - "prepared_review_unit_hash": "f9bf23e21e81cb859796c7fdb4eda32de5db1c928691118c0603ae3b84785381", + "prepared_review_unit_hash": "c9c803e0465eee2115af7207519ccad5e8e469eaf27b2847e6a1dfb5c6ef924d", "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": 2873, + "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": 2873, + "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 about different topics: tokenizer truncation behavior, ReduceLROnPlateau-like scheduling functionality, and Wav2Vec2 word-level time offsets. They do not describe the same underlying bug or feature request, so none should be clustered as duplicates.", - "confidence": 0.98, - "canonical_issue_reason": "No canonical issue: the issues are unrelated and should remain separate.", + "summary": "These three issues are unrelated: tokenizer truncation behavior, learning-rate scheduler functionality, and Wav2Vec2 word boundary timestamps. No duplicate cluster is present.", + "confidence": 0.99, + "canonical_issue_reason": null, "canonical_pr_reason": null, - "best_issue_reason": "Issue 10519 is the most concrete, narrowly scoped feature request, but it is not a duplicate of the others.", + "best_issue_reason": null, "best_pr_reason": null, "soft_edge_verdicts": [ { "left": "issue:10519", "right": "issue:11307", "accept": false, - "reason": "Different domains and requests: truncation strategy vs. Wav2Vec2 word boundary offsets." + "reason": "Different domains and requests: truncation strategy vs. Wav2Vec2 word time offsets; no shared bug or code path." }, { "left": "issue:10519", "right": "issue:11005", "accept": false, - "reason": "Unrelated features: truncation from the beginning vs. a ReduceLROnPlateau-like scheduler." + "reason": "One is about tokenizer truncation options, the other about ReduceLROnPlateau-like scheduling behavior; unrelated functionality." }, { "left": "issue:11005", "right": "issue:11307", "accept": false, - "reason": "No shared underlying bug/change; one is learning-rate scheduling, the other is speech model time offsets." + "reason": "Learning-rate scheduler feature request and speech-model word offset timing are unrelated." } ] }, @@ -19631,12 +19687,12 @@ "issue:13244|issue:16618", "issue:15980|issue:16618" ], - "prepared_review_unit_hash": "6dcbbf1c75b46fa40685a43adc6a013348be127d51c7371972d165ec68605e0d", + "prepared_review_unit_hash": "75c108dea984b9477a33df43276a20d456a4de6a332c92836d6c9d9db894b0d8", "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 }, @@ -19644,7 +19700,7 @@ "node_count": 3, "item_count": 3, "soft_pair_count": 2, - "serialized_chars": 2477, + "serialized_chars": 2479, "estimated_input_tokens": 620, "estimated_eval_tokens": 1496 }, @@ -19653,24 +19709,24 @@ "aggressively_trimmed": false, "split": false, "analyst_result": { - "summary": "The cluster does not contain a clear duplicate set. Issue 13244 is unrelated. Issue 15980 is a narrow login/authentication error-message problem, while 16618 is a broader model-loading failure but not clearly the same bug, so both soft links should be rejected.", - "confidence": 0.78, - "canonical_issue_reason": "If one issue must represent the cluster, issue 16618 is the broadest and most central model-loading report; however, the evidence is weak and it is not a confirmed duplicate of the others.", + "summary": "These issues do not appear to be true duplicates. Issue 16618 is a broad model-loading failure report, while 15980 is specifically about an unhelpful error when downloading a private model without login; 13244 is about Tapas tokenization and is unrelated.", + "confidence": 0.86, + "canonical_issue_reason": "Issue 16618 is the broadest and most central report in the cluster, but it still appears to be a distinct model-loading problem rather than the same bug as the other issues.", "canonical_pr_reason": null, - "best_issue_reason": "Issue 16618 is the best issue to surface because it is the most general user-facing failure around model loading and has the strongest discussion signal among the three.", + "best_issue_reason": "16618 is the best candidate only because its title is the most general and likely to capture the core user-facing failure; however, it should not be treated as a duplicate of the others.", "best_pr_reason": null, "soft_edge_verdicts": [ { "left": "issue:13244", "right": "issue:16618", "accept": false, - "reason": "Tapas tokenization mismatch is a different problem than failing to load bert-base-uncased; no shared underlying bug is apparent." + "reason": "Tapas tokenization mismatches TensorFlow code is a separate preprocessing/tokenization issue, not a model-loading failure." }, { "left": "issue:15980", "right": "issue:16618", "accept": false, - "reason": "Private-model unauthenticated download errors are a specific auth/error-message issue, not the same as a generic inability to load a public model." + "reason": "Both mention model loading, but 15980 is specifically about a bad private-model auth/download error message, while 16618 is a different generic inability-to-load-model report." } ] }, @@ -19700,12 +19756,12 @@ "soft_pairs": [ "issue:26350|issue:33260" ], - "prepared_review_unit_hash": "a1d23823acba9cee165e0849f82f9cae4f00acc98082530a259e736b058c597f", + "prepared_review_unit_hash": "05b21069aa6759391557ab4327bc0537e58136d1223c57bcabed4e1482ddd03d", "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 }, @@ -19713,7 +19769,7 @@ "node_count": 2, "item_count": 2, "soft_pair_count": 1, - "serialized_chars": 1589, + "serialized_chars": 1590, "estimated_input_tokens": 398, "estimated_eval_tokens": 1052 }, @@ -19722,18 +19778,18 @@ "aggressively_trimmed": false, "split": false, "analyst_result": { - "summary": "These are not duplicates: one requests Flash Attention 2 support for more architectures, the other requests GGUF support for more architectures. They share a loose pattern (\u201ccommunity contribution\u201d support expansion) but target different features and code paths.", + "summary": "These are two separate open community feature requests: one for Flash Attention 2 support across more architectures, and one for GGUF support across more architectures. They share a similar framing but target different capabilities and code paths, so they are not duplicates.", "confidence": 0.97, "canonical_issue_reason": null, "canonical_pr_reason": null, - "best_issue_reason": null, + "best_issue_reason": "Neither issue is a good canonical representative for the other because they request different features; if forced to choose within this pair, neither should subsume the other.", "best_pr_reason": null, "soft_edge_verdicts": [ { "left": "issue:26350", "right": "issue:33260", "accept": false, - "reason": "Different underlying requests and implementation areas: Flash Attention 2 architecture support vs GGUF support. Similar framing is not enough to treat them as the same bug/change." + "reason": "Different feature requests (Flash Attention 2 vs GGUF support) with different implementation targets; similar wording alone is not enough to treat them as the same issue." } ] }, @@ -19798,214 +19854,214 @@ "issue:46222|issue:46931", "issue:46226|issue:46931" ], - "prepared_review_unit_hash": "9d351f1a5a3ba4e0578066adae0c03a05c25102b7b4bebb1c7c75c62d87c8801", + "prepared_review_unit_hash": "425f9c6e2531644cc11f4bdd9edbdd812c250eed2dc553101a54150fcf333a12", "component_budget": { "node_count": 9, "item_count": 9, "soft_pair_count": 29, - "serialized_chars": 15986, - "estimated_input_tokens": 3997, - "estimated_eval_tokens": 8250 + "serialized_chars": 15976, + "estimated_input_tokens": 3994, + "estimated_eval_tokens": 8244 }, "budget": { "node_count": 9, "item_count": 9, "soft_pair_count": 29, - "serialized_chars": 15986, - "estimated_input_tokens": 3997, - "estimated_eval_tokens": 8250 + "serialized_chars": 15976, + "estimated_input_tokens": 3994, + "estimated_eval_tokens": 8244 }, "overflow_policy": "truncate_then_skip", "trimmed": false, "aggressively_trimmed": false, "split": false, "analyst_result": { - "summary": "This cluster is mostly a false-positive mix of separate translation-request issues for different languages, plus two small true-duplicate subgroups: the generic placeholder translation issue trio and the Thai pair.", + "summary": "This cluster is mostly a set of separate docs-translation request issues for different languages. I only see duplicate-like evidence inside the exact same-title subclusters (the generic placeholder issues and the repeated Thai issue); the cross-language pairs should not be merged.", "confidence": 0.81, - "canonical_issue_reason": "43834 is the best representative of the true duplicate subgroup: it is the earliest of the near-identical placeholder translation issues, and 44365/44370 look like copies of the same templated request.", + "canonical_issue_reason": "Issue 43834 is the cleanest representative of the exact-duplicate generic-template subcluster: it has the same title/body pattern as 44365 and 44370 and appears to be the earliest copy.", "canonical_pr_reason": null, - "best_issue_reason": "43834 is the strongest issue candidate overall because it anchors the clearest duplicate subcluster and has the broadest exact/near-exact overlap with the other placeholder issues.", + "best_issue_reason": "43834 is the best issue to anchor the duplicate subcluster because it is one of the exact-title matches and is the most canonical/earliest-looking instance among them.", "best_pr_reason": null, "soft_edge_verdicts": [ { "left": "issue:27088", "right": "issue:44370", "accept": false, - "reason": "Different language request; similar translation-template wording is not enough to treat Turkish and the placeholder issue as the same underlying issue." + "reason": "Different language request (Turkish vs generic template); not the same underlying issue." }, { "left": "issue:27088", "right": "issue:44365", "accept": false, - "reason": "Turkish vs. a generic placeholder translation request are distinct issue targets, not the same bug/change." + "reason": "Different language request; similarity is only from the translation-template wording." }, { "left": "issue:27088", "right": "issue:43834", "accept": false, - "reason": "Different concrete translation request; only the template is shared." + "reason": "Specific Turkish request vs generic placeholder issue; no evidence they are the same report." }, { "left": "issue:27088", "right": "issue:38435", "accept": false, - "reason": "Turkish and Romanian translation requests are separate issues." + "reason": "Different language target (Turkish vs Romanian); distinct translation requests." }, { "left": "issue:27088", "right": "issue:28936", "accept": false, - "reason": "Different language targets (Turkish vs Spanish), so not duplicates." + "reason": "Different language target (Turkish vs Spanish); distinct requests despite similar template text." }, { "left": "issue:28936", "right": "issue:43834", "accept": false, - "reason": "Spanish translation request is distinct from the placeholder templated issue." + "reason": "Spanish request vs generic placeholder issue; same template, not the same language request." }, { "left": "issue:28936", "right": "issue:44365", "accept": false, - "reason": "Different language target; shared i18n wording is too broad to merge." + "reason": "Different language target; similarity comes from the shared docs-translation template." }, { "left": "issue:28936", "right": "issue:38435", "accept": false, - "reason": "Spanish vs Romanian translation requests are separate." + "reason": "Different language targets (Spanish vs Romanian); not duplicates." }, { "left": "issue:28936", "right": "issue:44370", "accept": false, - "reason": "Spanish translation issue and placeholder issue are not the same underlying request." + "reason": "Different language target; no evidence of one underlying issue." }, { "left": "issue:38435", "right": "issue:44370", "accept": false, - "reason": "Romanian request is separate from the generic placeholder translation issue." + "reason": "Romanian request vs generic template issue; not the same concrete translation request." }, { "left": "issue:38435", "right": "issue:44365", "accept": false, - "reason": "Different concrete translation target; not a duplicate." + "reason": "Romanian request vs generic template issue; similarity is superficial." }, { "left": "issue:38435", "right": "issue:43834", "accept": false, - "reason": "Romanian translation request is not the same as the placeholder templated request." + "reason": "Romanian request vs generic placeholder issue; different underlying request." }, { "left": "issue:38435", "right": "issue:46222", "accept": false, - "reason": "Different language/request context; shared template similarity is insufficient." + "reason": "Romanian vs Thai request; clearly different issues." }, { "left": "issue:38435", "right": "issue:46226", "accept": false, - "reason": "Different language/request context; do not merge." + "reason": "Romanian vs Thai request; clearly different issues." }, { "left": "issue:43834", "right": "issue:44370", "accept": true, - "reason": "Near-identical placeholder translation issue; looks like the same templated request duplicated." + "reason": "Exact same placeholder title/body pattern; looks like a duplicate copy of the same templated issue." }, { "left": "issue:43834", "right": "issue:44365", "accept": true, - "reason": "Same generic placeholder title/body pattern; plausibly one duplicated issue." + "reason": "Exact same placeholder title/body pattern; duplicate of the same templated issue." }, { "left": "issue:43834", "right": "issue:46222", "accept": false, - "reason": "The titles/content do not indicate the same concrete request; this looks like a different issue despite template similarity." + "reason": "Different title/language text; no strong evidence they are the same issue." }, { "left": "issue:43834", "right": "issue:46226", "accept": false, - "reason": "Same as above: too little evidence of the same underlying issue." + "reason": "Different title/language text; not the same underlying request." }, { "left": "issue:43834", "right": "issue:46931", "accept": false, - "reason": "Different title/content; not clearly the same request." + "reason": "Different language/text and no shared concrete request." }, { "left": "issue:44365", "right": "issue:44370", "accept": true, - "reason": "These are near-identical copies of the same placeholder translation issue." + "reason": "Exact same placeholder title/body pattern; likely duplicate issue reports." }, { "left": "issue:44365", "right": "issue:46222", "accept": false, - "reason": "Only template similarity; not enough to conclude the same issue." + "reason": "Generic placeholder issue vs Thai-language request; not the same request." }, { "left": "issue:44365", "right": "issue:46226", "accept": false, - "reason": "Not clearly the same concrete issue; reject." + "reason": "Generic placeholder issue vs Thai-language request; not the same request." }, { "left": "issue:44365", "right": "issue:46931", "accept": false, - "reason": "Different issue content and title." + "reason": "Generic placeholder issue vs different Thai request; not duplicates." }, { "left": "issue:44370", "right": "issue:46222", "accept": false, - "reason": "Shared i18n template language is too broad; the issues do not clearly describe the same request." + "reason": "Different language/text; only the broad template is shared." }, { "left": "issue:44370", "right": "issue:46226", "accept": false, - "reason": "Not the same concrete issue; reject." + "reason": "Different language/text; no evidence of same underlying issue." }, { "left": "issue:44370", "right": "issue:46931", "accept": false, - "reason": "Different issue content; no strong duplicate evidence." + "reason": "Different language/text; not the same request." }, { "left": "issue:46222", "right": "issue:46226", "accept": true, - "reason": "Exact duplicate subgroup: same Thai title and highly similar issue content indicate the same underlying issue." + "reason": "Identical Thai title/body pattern and near-identical metadata; strong duplicate signal." }, { "left": "issue:46222", "right": "issue:46931", "accept": false, - "reason": "Thai-titled issues are not clearly the same underlying request." + "reason": "Different Thai titles; not enough to treat as the same issue." }, { "left": "issue:46226", "right": "issue:46931", "accept": false, - "reason": "Different title/content; insufficient evidence of a true duplicate." + "reason": "Different Thai titles; similarity is not sufficient for duplication." } ] }, "evaluator_result": { "accept": true, - "feedback": "The cluster summary is grounded in the packet: the three placeholder issues have identical titles and near-identical bodies, the Thai pair has identical titles and exact similarity, and the Turkish/Spanish/Romanian items are distinct by title. The soft-edge verdicts are conservative overall, with rejections based on clear language/title differences and accepts limited to the strongest duplicate evidence." + "feedback": "The assessment is grounded in the packet and stays conservative overall. The exact-title generic-template subcluster (43834/44365/44370) and the identical Thai pair (46222/46226) are the only places with strong duplicate evidence, while the cross-language pairs are appropriately rejected. The only mild caution is that some reasons could have been phrased a bit more cautiously, but nothing here rises to the level of overstating the evidence." }, "evaluator_used": true, "retried": false, @@ -20098,256 +20154,268 @@ "issue:42503|issue:46653", "issue:45027|issue:46653" ], - "prepared_review_unit_hash": "e4eb40d2a080bb871845218a18d9148f6a23bb20ec50b3147f6cee0d42a4eb46", + "prepared_review_unit_hash": "be9be1871bfb9ba7ba01a55ff65a8fa742c68de0d448b2ece0f22adf02505cce", "component_budget": { "node_count": 21, "item_count": 21, "soft_pair_count": 48, - "serialized_chars": 28611, - "estimated_input_tokens": 7153, - "estimated_eval_tokens": 14562 + "serialized_chars": 28619, + "estimated_input_tokens": 7155, + "estimated_eval_tokens": 14566 }, "budget": { "node_count": 21, "item_count": 21, "soft_pair_count": 48, - "serialized_chars": 28611, - "estimated_input_tokens": 7153, - "estimated_eval_tokens": 14562 + "serialized_chars": 28619, + "estimated_input_tokens": 7155, + "estimated_eval_tokens": 14566 }, "overflow_policy": "truncate_then_skip", "trimmed": false, "aggressively_trimmed": false, "split": false, "analyst_result": { - "summary": "This cluster is a loose semantic group of model-request issues, not a duplicate set: the items span unrelated architectures, tasks, and modalities (keypoint matching, BERT variants, detection, audio ASR/TTS, vision-language, etc.).", - "confidence": 0.95, - "canonical_issue_reason": "No single canonical issue: the cluster mixes distinct model requests that do not share the same underlying change.", + "summary": "This cluster is mostly a set of superficially similar model-request issues, not true duplicates. The soft edges connect different models, tasks, and modalities, so none should be merged as the same underlying change.", + "confidence": 0.96, + "canonical_issue_reason": "No single canonical duplicate issue stands out; the items are distinct model requests that only share broad wording.", "canonical_pr_reason": null, - "best_issue_reason": "No single issue is a good global representative; the overlap is only broad topical similarity around \"add model support\".", + "best_issue_reason": "If a representative must be chosen, issue 36879 is the most generic model-add request, but it still does not represent the rest well enough to be a true canonical.", "best_pr_reason": null, "soft_edge_verdicts": [ { "left": "issue:36718", "right": "issue:47004", "accept": false, - "reason": "Similar keypoint terminology, but RoMa matcher and LoMa detector+matcher are different model requests." + "reason": "Both are keypoint-related, but RoMa matcher vs LoMa detector+matcher are different concrete requests and code paths." }, { "left": "issue:36718", "right": "issue:41115", "accept": false, - "reason": "Keypoint matcher request vs MiniCPM3 model support; unrelated code paths." + "reason": "RoMa keypoint matcher and MiniCPM3 architecture are unrelated model requests." }, { "left": "issue:36817", "right": "issue:41211", "accept": false, - "reason": "Different model families: EuroBert config addition vs DEIMv2 support." + "reason": "EuroBert config and DEIMv2 are different model families with no shared change." }, { "left": "issue:36817", "right": "issue:43671", "accept": false, - "reason": "Different tasks and models: EuroBert vs Qwen3-TTS." + "reason": "EuroBert config and Qwen3-TTS support are different modalities and fixes." }, { "left": "issue:36817", "right": "issue:36879", "accept": false, - "reason": "Both are model additions, but EuroBert and RF-DETR are unrelated changes." + "reason": "EuroBert config and RF-DETR model addition are unrelated requests." }, { "left": "issue:36817", "right": "issue:42954", "accept": false, - "reason": "EuroBert and XiaoMi MiMo are separate model families." + "reason": "EuroBert config and XiaoMi MiMo are distinct model-add requests." }, { "left": "issue:36879", "right": "issue:41211", "accept": false, - "reason": "Two distinct detection-model requests; not the same underlying change." + "reason": "RF-DETR and DEIMv2 are different detectors; not the same concrete addition." }, { "left": "issue:36879", "right": "issue:43909", "accept": false, - "reason": "RF-DETR detection support and LFM2.5 Audio support are unrelated." + "reason": "RF-DETR vision support and LFM2.5 Audio 1.5B are different modalities." }, { "left": "issue:36879", "right": "issue:44121", "accept": false, - "reason": "Different model families and implementation paths." + "reason": "RF-DETR model support and a weight-sparse transformer proposal are unrelated." }, { "left": "issue:36879", "right": "issue:43837", "accept": false, - "reason": "RF-DETR and Qwen3-ASR are different models/tasks." + "reason": "RF-DETR object detection and Qwen3-ASR audio transcription are different code paths." }, { "left": "issue:36879", "right": "issue:43671", "accept": false, - "reason": "RF-DETR detection support is unrelated to Qwen3-TTS." + "reason": "RF-DETR and Qwen3-TTS are unrelated model requests." }, { "left": "issue:37015", "right": "issue:43837", "accept": false, - "reason": "NeoBERT model addition vs Qwen3-ASR support; unrelated." + "reason": "NeoBERT and Qwen3-ASR do not target the same model or feature." }, { "left": "issue:37015", "right": "issue:44121", "accept": false, - "reason": "NeoBERT and OpenAI weight-sparse transformer are different model requests." + "reason": "NeoBERT support and the weight-sparse transformer request are different changes." }, { "left": "issue:37015", "right": "issue:43909", "accept": false, - "reason": "BERT model support vs audio model support; different changes." + "reason": "NeoBERT and LFM2.5 Audio 1.5B are unrelated model requests." }, { "left": "issue:36879", "right": "issue:37015", "accept": false, - "reason": "RF-DETR and NeoBERT are distinct model integrations." + "reason": "RF-DETR and NeoBERT are different models with separate implementation needs." }, { "left": "issue:37015", "right": "issue:42503", "accept": false, - "reason": "NeoBERT and ModernVBERT are different model families." + "reason": "NeoBERT and ModernVBERT are both BERT-like, but they are still distinct model families and not one duplicate request." }, { "left": "issue:41115", "right": "issue:43837", "accept": false, - "reason": "MiniCPM3 support and Qwen3-ASR support are unrelated." + "reason": "MiniCPM3 architecture and Qwen3-ASR support are unrelated." }, { "left": "issue:41115", "right": "issue:44121", "accept": false, - "reason": "MiniCPM3 vs weight-sparse transformer; different model work." + "reason": "MiniCPM3 architecture and weight-sparse transformer support are different concrete changes." }, { "left": "issue:41115", "right": "issue:42503", "accept": false, - "reason": "MiniCPM3 and ModernVBERT are separate model additions." + "reason": "MiniCPM3 architecture and ModernVBERT models are different model requests." }, { "left": "issue:41115", "right": "issue:45027", "accept": false, - "reason": "MiniCPM3 support and Voxtral TTS support are different tasks." + "reason": "MiniCPM3 architecture and Voxtral TTS support target different model paths." }, { "left": "issue:41211", "right": "issue:43671", "accept": false, - "reason": "DEIMv2 and Qwen3-TTS are unrelated model requests." + "reason": "DEIMv2 and Qwen3-TTS are unrelated model additions." }, { "left": "issue:42503", "right": "issue:43837", "accept": false, - "reason": "ModernVBERT models and Qwen3-ASR are different model families." + "reason": "ModernVBERT models and Qwen3-ASR support are different tasks and models." }, { "left": "issue:42503", "right": "issue:43671", "accept": false, - "reason": "ModernVBERT models and Qwen3-TTS are unrelated." + "reason": "ModernVBERT models and Qwen3-TTS support are unrelated." }, { "left": "issue:42503", "right": "issue:42954", "accept": false, - "reason": "ModernVBERT and XiaoMi MiMo are different model additions." + "reason": "ModernVBERT models and XiaoMi MiMo are different requests." }, { "left": "issue:42738", "right": "issue:45027", "accept": false, - "reason": "BERT-like RoPE support is not the same change as Voxtral TTS support." + "reason": "BERT-like models with RoPE and Voxtral TTS support are unrelated." }, { "left": "issue:42503", "right": "issue:42738", "accept": false, - "reason": "Both are transformer-related, but ModernVBERT models and BERT-like RoPE support are distinct requests." + "reason": "Both are BERT-related, but ModernVBERT models and BERT-like models with RoPE are not the same concrete change." }, { "left": "issue:42738", "right": "issue:42954", "accept": false, - "reason": "BERT-like RoPE models and XiaoMi MiMo are unrelated model support requests." + "reason": "BERT-like models with RoPE and XiaoMi MiMo are distinct requests." }, { "left": "issue:42738", "right": "issue:46653", "accept": false, - "reason": "Model-architecture request vs a non-model issue; no duplicate relation." + "reason": "A model architecture request and a person-named issue are clearly unrelated." }, { "left": "issue:42954", "right": "issue:44205", "accept": false, - "reason": "XiaoMi MiMo and SAM3-LiteText are different model families/modalities." + "reason": "XiaoMi MiMo and SAM3-LiteText are different models and modalities." }, { "left": "issue:42954", "right": "issue:43671", "accept": false, - "reason": "MiMo and Qwen3-TTS are separate model requests." + "reason": "MiMo and Qwen3-TTS are unrelated model requests." }, { "left": "issue:42954", "right": "issue:45177", "accept": false, - "reason": "MiMo and DeepSeek-OCR2 target different models." + "reason": "MiMo and DeepSeek-OCR2 are different model families." }, { "left": "issue:42954", "right": "issue:46653", "accept": false, - "reason": "Model request vs unrelated issue." + "reason": "MiMo support and a person-named issue have no shared change." }, { "left": "issue:43671", "right": "issue:43837", "accept": false, - "reason": "Qwen3-TTS and Qwen3-ASR are close in family, but they are different modalities and code paths." + "reason": "Qwen3-TTS and Qwen3-ASR are the same vendor family but different modalities and code paths, so they are not one duplicate fix." + }, + { + "left": "issue:43671", + "right": "issue:45027", + "accept": false, + "reason": "Qwen3-TTS and Voxtral TTS are different model families." }, { "left": "issue:43671", "right": "issue:44121", "accept": false, - "reason": "Qwen3-TTS support and weight-sparse transformer support are unrelated." + "reason": "Qwen3-TTS support and the weight-sparse transformer proposal are unrelated." }, { "left": "issue:43837", "right": "issue:43909", "accept": false, - "reason": "Both are audio-related requests, but Qwen3-ASR and LFM2.5 Audio are different model integrations." + "reason": "Qwen3-ASR and LFM2.5 Audio 1.5B are both audio-related, but they are different concrete models and implementations." }, { - "left": "issue:44121", + "left": "issue:43837", "right": "issue:45322", "accept": false, - "reason": "Weight-sparse transformer request vs EUPE vision encoder request; unrelated." + "reason": "Qwen3-ASR and EUPE are unrelated model requests." + }, + { + "left": "issue:43837", + "right": "issue:44121", + "accept": false, + "reason": "Qwen3-ASR support and the weight-sparse transformer request are not the same fix." }, { "left": "issue:43837", "right": "issue:45027", "accept": false, - "reason": "ASR and TTS support for different model families are not the same underlying change." + "reason": "Qwen3-ASR and Voxtral TTS are different tasks, not a shared underlying bug or change." }, { "left": "issue:37015", @@ -20359,25 +20427,25 @@ "left": "issue:43837", "right": "issue:45177", "accept": false, - "reason": "Qwen3-ASR and DeepSeek-OCR2 are different modalities and model families." + "reason": "ASR support and DeepSeek-OCR2 are different modalities and code paths." }, { "left": "issue:45322", "right": "issue:46133", "accept": false, - "reason": "EUPE and TIPSv2 are separate vision-language model requests." + "reason": "EUPE and TIPSv2 are both vision-language themed, but they are distinct model additions and not one mergeable change." }, { "left": "issue:42503", "right": "issue:46653", "accept": false, - "reason": "ModernVBERT models and an unrelated issue are not duplicates." + "reason": "ModernVBERT models and a person-named issue are unrelated." }, { "left": "issue:45027", "right": "issue:46653", "accept": false, - "reason": "Voxtral TTS support and an unrelated issue are not duplicates." + "reason": "Voxtral TTS support and a person-named issue are unrelated." } ] }, @@ -20407,7 +20475,7 @@ "soft_pairs": [ "issue:38740|issue:46443" ], - "prepared_review_unit_hash": "fedd33d170c1a59ab7a077d5482dd31b1af36249575d953d63bf931945e69da4", + "prepared_review_unit_hash": "1b784d7d1f6c281e5d561e3bc85c9f3ed43a1a4b684a297c9500f0d020b5aa8b", "component_budget": { "node_count": 2, "item_count": 2, @@ -20429,18 +20497,18 @@ "aggressively_trimmed": false, "split": false, "analyst_result": { - "summary": "The two issues are unrelated: one is a closed docs request to add Pruna as an optimization framework, and the other is an open Vision RoPE refactor. There is no shared code path, target, or substantive overlap indicating duplication.", - "confidence": 0.97, - "canonical_issue_reason": "issue:46443 is the more substantive technical item and remains open; issue:38740 is a separate closed documentation request and does not describe the same work.", + "summary": "The two issues are unrelated: one is a documentation request to add Pruna as an optimization framework, while the other is a Vision RoPE refactor. They do not describe the same underlying bug or change, so the soft duplicate edge should be rejected.", + "confidence": 0.98, + "canonical_issue_reason": "issue:46443 is the better representative issue because it is a substantive technical refactor with a clear code-change scope, whereas issue:38740 is a docs-only request.", "canonical_pr_reason": null, - "best_issue_reason": "issue:46443 is the better representative issue because it reflects the actual implementation refactor work, whereas issue:38740 is a docs-only request in a different area.", + "best_issue_reason": "issue:46443 is the strongest issue in the set for canonicalization since it concerns an actual implementation change rather than a documentation addition.", "best_pr_reason": null, "soft_edge_verdicts": [ { "left": "issue:38740", "right": "issue:46443", "accept": false, - "reason": "Reject: these are different topics entirely (docs request for an optimization framework vs. a Vision RoPE refactor) with no shared target, filenames, or code-path overlap." + "reason": "Reject: these are about different topics and code paths; a docs request for an optimization framework is not the same as a Vision RoPE refactor." } ] }, @@ -20470,12 +20538,12 @@ "soft_pairs": [ "issue:41084|issue:43037" ], - "prepared_review_unit_hash": "9c4b1d05ba749489449dd0dbe705a8dc29cfd8f02106ab0e6a72f0197b390dc4", + "prepared_review_unit_hash": "272f6d7d76045de2c1e4e4dffdbeb05e6eede6fe5ccaf2572d0d7310193f8bdb", "component_budget": { "node_count": 2, "item_count": 2, "soft_pair_count": 1, - "serialized_chars": 1479, + "serialized_chars": 1480, "estimated_input_tokens": 370, "estimated_eval_tokens": 996 }, @@ -20483,7 +20551,7 @@ "node_count": 2, "item_count": 2, "soft_pair_count": 1, - "serialized_chars": 1479, + "serialized_chars": 1480, "estimated_input_tokens": 370, "estimated_eval_tokens": 996 }, @@ -20492,18 +20560,18 @@ "aggressively_trimmed": false, "split": false, "analyst_result": { - "summary": "The two issues are unrelated: one is about block decoding behavior, the other is a DeepSeek v3.2 support request. This is not a duplicate cluster.", + "summary": "The two issues are not duplicates: one is about Set Block Decoding, while the other requests DeepSeek v3.2 support. They target different features and there is no shared implementation or explicit linkage.", "confidence": 0.98, "canonical_issue_reason": null, "canonical_pr_reason": null, - "best_issue_reason": "Neither issue is a clear duplicate of the other; if forced to pick a representative, issue 43037 is slightly more specific, but the cluster should not be merged.", + "best_issue_reason": null, "best_pr_reason": null, "soft_edge_verdicts": [ { "left": "issue:41084", "right": "issue:43037", "accept": false, - "reason": "Different underlying topics and no shared targets/files; a vague decoding issue is not the same as adding model-version support." + "reason": "Different underlying requests with no overlapping code path or explicit relation: decoding behavior vs model/version support." } ] }, @@ -20533,7 +20601,7 @@ "soft_pairs": [ "issue:41306|issue:41826" ], - "prepared_review_unit_hash": "edf802d892c21285ff6bba2fa77dbd6af2179aa4a7b6a8cc2330735555118e28", + "prepared_review_unit_hash": "08662c766402b314aa15b337c074af765dcb94face98ad64016c01bd1e56cb5c", "component_budget": { "node_count": 2, "item_count": 2, @@ -20555,18 +20623,18 @@ "aggressively_trimmed": false, "split": false, "analyst_result": { - "summary": "These are both Tiled-related feature requests, but they describe different changes: one asks for a `from_pretrained()` flag for TiledFusedLogitsLoss/FLCE, while the other requests integrating TiledMLP for lower memory use. They are not the same underlying bug or change.", - "confidence": 0.94, - "canonical_issue_reason": "Issue 41306 is the more concrete and narrowly scoped request; 41826 is a separate architecture/integration feature request.", + "summary": "These two issues are related only at a high level (memory-efficient tiled layers), but they describe different features and code paths: one is about exposing a TiledFusedLogitsLoss/FLCE flag in `from_pretrained()`, the other about adding TiledMLP for lower memory use. Not duplicates.", + "confidence": 0.95, + "canonical_issue_reason": "No clear canonical issue: the reports target different components and changes, so they should not be merged as one underlying bug or feature request.", "canonical_pr_reason": null, - "best_issue_reason": "41306 is the best representative issue because it is specific, actionable, and centered on a single API change.", + "best_issue_reason": "If one must be prioritized, #41306 is the more concrete and self-contained request because it specifies a precise `from_pretrained()` integration point and exact feature flag behavior.", "best_pr_reason": null, "soft_edge_verdicts": [ { "left": "issue:41306", "right": "issue:41826", "accept": false, - "reason": "Different requested changes and code paths: one concerns enabling FLCE via `from_pretrained()`, the other concerns adding TiledMLP for memory reduction. Similar theme, but not the same fix or mergeable duplicate." + "reason": "Different requested changes: FLCE/TiledFusedLogitsLoss flag support in `from_pretrained()` versus integrating TiledMLP. They share a broad memory-optimization theme but not the same underlying bug or implementation." } ] }, @@ -20596,7 +20664,7 @@ "soft_pairs": [ "issue:43598|issue:43600" ], - "prepared_review_unit_hash": "88e66d4d4cf5e287f8cfcf9b7bc0fe1bad76daeffd77d040ea8b8ceeb275d5e5", + "prepared_review_unit_hash": "ba450d8b0c9274afde801856d5273eeb15c7d80f42539319be1004283991c4d6", "component_budget": { "node_count": 2, "item_count": 2, @@ -20618,18 +20686,18 @@ "aggressively_trimmed": false, "split": false, "analyst_result": { - "summary": "These are related only at the broad Trainer/customization level, but they describe different problems: one is about `remove_unused_column` behavior, the other about FLOPS estimation for custom trainers. Not duplicates.", - "confidence": 0.95, - "canonical_issue_reason": "Issue 43598 is the clearer canonical issue because it is more specific and actionable around `remove_unused_column` and Trainer customizability.", + "summary": "These two closed issues are related only at a very broad Trainer-customization level, but they describe different underlying problems: column pruning customizability vs FLOPS estimation for custom trainers. They should not be treated as duplicates.", + "confidence": 0.93, + "canonical_issue_reason": "issue:43598 is the broader, more general Trainer customizability request and is the better representative of the cluster theme.", "canonical_pr_reason": null, - "best_issue_reason": "Issue 43598 is the better representative of the cluster only because it is more narrowly scoped; however, the two issues do not appear to describe the same bug.", + "best_issue_reason": "issue:43598 is the better standalone issue because it frames a clearer, wider Trainer behavior change; issue:43600 is a narrower performance-metrics concern.", "best_pr_reason": null, "soft_edge_verdicts": [ { "left": "issue:43598", "right": "issue:43600", "accept": false, - "reason": "Different underlying concerns: one changes how Trainer drops unused columns, the other adjusts FLOPS estimation for custom trainers. Shared Trainer context is too broad to treat as the same issue." + "reason": "Different code-path problems: one is about `remove_unused_column` behavior in Trainer, the other is about FLOPS estimation for custom trainers. They are related in subsystem but not the same bug/change." } ] }, @@ -20659,12 +20727,12 @@ "soft_pairs": [ "issue:43684|issue:44405" ], - "prepared_review_unit_hash": "04a866c3963ce37d1bc3dee4f8ae78bac35f2f790dadbff6f7b9b4157694be50", + "prepared_review_unit_hash": "e559a5ffa6d39d1c36f87267a112e745f34c0895ca4cbe79f65017f63a0928c7", "component_budget": { "node_count": 2, "item_count": 2, "soft_pair_count": 1, - "serialized_chars": 1611, + "serialized_chars": 1612, "estimated_input_tokens": 403, "estimated_eval_tokens": 1062 }, @@ -20672,7 +20740,7 @@ "node_count": 2, "item_count": 2, "soft_pair_count": 1, - "serialized_chars": 1611, + "serialized_chars": 1612, "estimated_input_tokens": 403, "estimated_eval_tokens": 1062 }, @@ -20681,18 +20749,18 @@ "aggressively_trimmed": false, "split": false, "analyst_result": { - "summary": "These are separate feature requests: one for Qwen3-Omni registration in AutoModel/AutoModelForConditionalGeneration, the other for Qwen3.5 sequence-classification support. They target different model families and different auto-model entry points, so they should not be deduplicated.", - "confidence": 0.95, - "canonical_issue_reason": "If one issue must represent the cluster, issue 44405 is the more specific and actionable request (a concrete AutoModelForSequenceClassification gap for Qwen3.5).", + "summary": "These two closed issues are both about expanding Qwen model support, but they target different model families and different AutoModel heads, so they do not look like duplicates.", + "confidence": 0.89, + "canonical_issue_reason": "issue #43684 is the broader, more general request and would be the better representative if this cluster were being summarized as a single issue topic.", "canonical_pr_reason": null, - "best_issue_reason": "Issue 44405 is the clearest standalone request because it names the exact missing task head and config, making it easier to implement and track.", + "best_issue_reason": "#43684 is the clearest canonical issue candidate because it describes a broader registration gap across two AutoModel classes, whereas #44405 is a narrower task-specific support request for a different model/config family.", "best_pr_reason": null, "soft_edge_verdicts": [ { "left": "issue:43684", "right": "issue:44405", "accept": false, - "reason": "Different underlying changes: Qwen3-Omni model registration for AutoModel/AutoModelForConditionalGeneration vs Qwen3.5 AutoModelForSequenceClassification support. Different model variants and different code paths." + "reason": "Different underlying changes: Qwen3-Omni AutoModel/AutoModelForConditionalGeneration registration vs Qwen3.5 sequence-classification support for Qwen3_5Config. Same broad theme, but not the same bug or change." } ] }, @@ -20729,77 +20797,80 @@ "issue:43845|issue:43859", "issue:43846|issue:43859" ], - "prepared_review_unit_hash": "447a03a4de289d4a1e6e579fe3f36b6d41353a525cf8cb4c9e934e8ad01521da", + "prepared_review_unit_hash": "5e885f5c0696f9e66b4e0269adfc40865e6a19833b484e4909ecc4dc333ea7ac", "component_budget": { "node_count": 4, "item_count": 4, "soft_pair_count": 6, - "serialized_chars": 4264, - "estimated_input_tokens": 1066, - "estimated_eval_tokens": 2388 + "serialized_chars": 4259, + "estimated_input_tokens": 1065, + "estimated_eval_tokens": 2386 }, "budget": { "node_count": 4, "item_count": 4, "soft_pair_count": 6, - "serialized_chars": 4264, - "estimated_input_tokens": 1066, - "estimated_eval_tokens": 2388 + "serialized_chars": 4259, + "estimated_input_tokens": 1065, + "estimated_eval_tokens": 2386 }, "overflow_policy": "truncate_then_skip", "trimmed": false, "aggressively_trimmed": false, "split": false, "analyst_result": { - "summary": "The cluster looks like one unrelated issue ('Fett') plus three very sparse issues with the same title ('huggingface '). There isn\u2019t enough substantive content to confidently merge any pair; the repeated titles are suggestive but not sufficient on their own.", + "summary": "Three issues share the exact same generic title \u201chuggingface \u201d and were created within about a day of each other, so they likely duplicate the same report; the \u201cFett\u201d issue looks unrelated and should stay separate.", "confidence": 0.71, - "canonical_issue_reason": "issue #43845 is the earliest of the repeated 'huggingface ' issues and is a reasonable representative if these turn out to be duplicates.", + "canonical_issue_reason": "Issue 43845 is the earliest of the near-identical \u201chuggingface \u201d reports and is a reasonable representative for that duplicate set.", "canonical_pr_reason": null, - "best_issue_reason": "issue #43845 is the best issue representative because it is the earliest matching title in the apparent mini-cluster and has at least as much signal as the others.", + "best_issue_reason": "Issue 43845 is the best issue candidate because it is an exact-title match with the other likely duplicates and is the earliest of the group.", "best_pr_reason": null, "soft_edge_verdicts": [ { "left": "issue:43835", "right": "issue:43845", "accept": false, - "reason": "Different titles and no concrete body/diff evidence tying them to the same underlying problem." + "reason": "Different titles with no shared substantive evidence; \u201cFett\u201d does not look like the same underlying issue as \u201chuggingface \u201d." }, { "left": "issue:43835", "right": "issue:43846", "accept": false, - "reason": "Different titles and no concrete body/diff evidence tying them to the same underlying problem." + "reason": "Different titles with no shared substantive evidence; this does not look like a duplicate pair." }, { "left": "issue:43835", "right": "issue:43859", "accept": false, - "reason": "Different titles and no concrete body/diff evidence tying them to the same underlying problem." + "reason": "Different titles with no shared substantive evidence; likely unrelated." }, { "left": "issue:43845", "right": "issue:43846", - "accept": false, - "reason": "Same short title, but there is no substantive description or explicit linkage proving the same bug." + "accept": true, + "reason": "Exact same title, very similar metadata, and close creation times strongly suggest the same underlying issue." }, { "left": "issue:43845", "right": "issue:43859", - "accept": false, - "reason": "Same short title, but the issues are too underspecified to confirm they describe the same underlying bug." + "accept": true, + "reason": "Exact same title and close timing; these look like duplicate reports of the same issue." }, { "left": "issue:43846", "right": "issue:43859", - "accept": false, - "reason": "Same short title, but the issues are too underspecified to confirm they describe the same underlying bug." + "accept": true, + "reason": "Exact same title and close timing; likely duplicate reports of the same issue." } ] }, - "evaluator_result": null, - "evaluator_used": false, + "evaluator_result": { + "accept": true, + "feedback": "The summary is grounded in the packet: issues 43845, 43846, and 43859 share the exact same title and are created close together, while 43835 has a different title and no other shared evidence. The representative selection of 43845 as the earliest of the same-title group is reasonable, and the duplicate verdicts are phrased conservatively as likely rather than certain." + }, + "evaluator_used": true, "retried": false, - "accepted_nontrivial_soft_edge": false, + "accepted_nontrivial_soft_edge": true, "error_kind": null, "error_message": null }, @@ -20837,106 +20908,106 @@ "issue:44573|issue:47036", "issue:44573|issue:44937" ], - "prepared_review_unit_hash": "16f6e753b07a5b907229b399434874de3dbf830fc85bffd5a0bbff41ae74eb82", + "prepared_review_unit_hash": "fc90d0a800bebf954041a1d5a2704d54281bda435c992099bf8dc5aa99dab37f", "component_budget": { "node_count": 6, "item_count": 6, "soft_pair_count": 12, - "serialized_chars": 7346, - "estimated_input_tokens": 1837, - "estimated_eval_tokens": 3930 + "serialized_chars": 7355, + "estimated_input_tokens": 1839, + "estimated_eval_tokens": 3934 }, "budget": { "node_count": 6, "item_count": 6, "soft_pair_count": 12, - "serialized_chars": 7346, - "estimated_input_tokens": 1837, - "estimated_eval_tokens": 3930 + "serialized_chars": 7355, + "estimated_input_tokens": 1839, + "estimated_eval_tokens": 3934 }, "overflow_policy": "truncate_then_skip", "trimmed": false, "aggressively_trimmed": false, "split": false, "analyst_result": { - "summary": "This cluster is dominated by spammy or generic issues with no shared concrete bug/change. The only mildly descriptive items are unrelated to each other, so there is no true duplicate set here.", - "confidence": 0.95, - "canonical_issue_reason": null, + "summary": "These issues do not look like a real duplicate set; they are mostly spammy or placeholder titles with no substantive body evidence. If a representative issue must be chosen, #47036 is the least spam-like, but there is no strong canonical duplicate here.", + "confidence": 0.94, + "canonical_issue_reason": "#47036 is the most specific and least spam-like item in the set, with at least one discussion activity point. Even so, it is not clearly a substantive canonical bug report.", "canonical_pr_reason": null, - "best_issue_reason": "issue:44418 is the least spam-like and has a real topical title, but it still does not appear related to the other items.", + "best_issue_reason": "#47036 is the best available representative only by elimination; the others are generic spam/placeholder titles with no useful details.", "best_pr_reason": null, "soft_edge_verdicts": [ { "left": "issue:44371", "right": "issue:44572", "accept": false, - "reason": "Both are just '' placeholders; no concrete underlying bug or change is described." + "reason": "Both are spam placeholders; no shared bug, code path, or concrete change to unify." }, { "left": "issue:44371", "right": "issue:44573", "accept": false, - "reason": "Both are spam placeholders with no actionable content to indicate the same issue." + "reason": "Both look like unrelated spam entries, not duplicates of a real issue." }, { "left": "issue:44371", "right": "issue:44418", "accept": false, - "reason": "A spam placeholder and a generic documentation suggestion are not the same underlying issue." + "reason": "One is spam and the other is a generic documentation suggestion; too little evidence of the same underlying issue." }, { "left": "issue:44371", "right": "issue:47036", "accept": false, - "reason": "A spam placeholder does not match the unrelated 'Community benchmark tip' issue." + "reason": "No shared concrete problem; only generic low-signal titles." }, { "left": "issue:44371", "right": "issue:44937", "accept": false, - "reason": "A spam placeholder and 'Check out \"Google LLC\"' are unrelated." + "reason": "Unrelated spam versus a specific quoted title; no evidence they describe the same issue." }, { "left": "issue:44418", "right": "issue:44572", "accept": false, - "reason": "Generic documentation enhancement and spam placeholder do not describe the same problem." + "reason": "Generic documentation suggestion and spam placeholder are not the same underlying bug/change." }, { "left": "issue:44418", "right": "issue:44573", "accept": false, - "reason": "Generic documentation suggestion vs. spam placeholder: no shared underlying issue." + "reason": "A docs suggestion and a spam issue do not match on a concrete problem statement." }, { "left": "issue:44572", "right": "issue:44573", "accept": false, - "reason": "Both are spam placeholders, but there is no evidence they refer to the same bug/change." + "reason": "Both are spammy placeholders, but there is no actual shared bug/change evidence to merge." }, { "left": "issue:44572", "right": "issue:47036", "accept": false, - "reason": "Spam placeholder and community benchmark tip are unrelated." + "reason": "No concrete overlap beyond generic similarity signals." }, { "left": "issue:44572", "right": "issue:44937", "accept": false, - "reason": "Spam placeholder and 'Check out \"Google LLC\"' are unrelated." + "reason": "The titles are unrelated and there is no issue-body evidence of the same underlying problem." }, { "left": "issue:44573", "right": "issue:47036", "accept": false, - "reason": "Spam placeholder and community benchmark tip are unrelated." + "reason": "No shared bug, code path, or explicit linkage; just low-signal similarity." }, { "left": "issue:44573", "right": "issue:44937", "accept": false, - "reason": "Spam placeholder and 'Check out \"Google LLC\"' are unrelated." + "reason": "Spam placeholder and quoted-title issue are not plausibly duplicates." } ] }, @@ -20966,46 +21037,46 @@ "soft_pairs": [ "issue:45145|issue:45151" ], - "prepared_review_unit_hash": "ac8a41dd98c89a155f78991cca4fef97c780cbfdb2a937615cc43f2955a01c6b", + "prepared_review_unit_hash": "2c661c03e57095f24a6ad276c1ee75acce601a1580da44b11e3962e2d942925a", "component_budget": { "node_count": 2, "item_count": 2, "soft_pair_count": 1, - "serialized_chars": 1646, - "estimated_input_tokens": 412, - "estimated_eval_tokens": 1080 + "serialized_chars": 1644, + "estimated_input_tokens": 411, + "estimated_eval_tokens": 1078 }, "budget": { "node_count": 2, "item_count": 2, "soft_pair_count": 1, - "serialized_chars": 1646, - "estimated_input_tokens": 412, - "estimated_eval_tokens": 1080 + "serialized_chars": 1644, + "estimated_input_tokens": 411, + "estimated_eval_tokens": 1078 }, "overflow_policy": "truncate_then_skip", "trimmed": false, "aggressively_trimmed": false, "split": false, "analyst_result": { - "summary": "The two issues have identical titles, highly similar bodies, and no distinguishing links/files, so they appear to be duplicates of the same energy-optimization proposal.", - "confidence": 0.98, - "canonical_issue_reason": "Issue #45145 is the better canonical candidate because it was created earlier and already has an inbound reference, while #45151 adds no unique identifying information.", + "summary": "The two closed issues have the same title and very similar bodies, with no differentiating targets or files, so they appear to be duplicate reports of the same underlying energy-optimization topic. Issue #45145 is the better canonical choice because it was opened earlier and has an inbound reference.", + "confidence": 0.99, + "canonical_issue_reason": "#45145 is earlier and already has an inbound reference, while #45151 adds no distinct scope or identifiers.", "canonical_pr_reason": null, - "best_issue_reason": "#45145 is the strongest single issue representative since it is the earliest instance of the duplicate pair and has slightly more reference activity.", + "best_issue_reason": "#45145 is the best issue to keep as the canonical duplicate because it is the earliest instance of the same report and has slightly more discussion linkage.", "best_pr_reason": null, "soft_edge_verdicts": [ { "left": "issue:45145", "right": "issue:45151", "accept": true, - "reason": "Same title and very similar body content with no conflicting details or unique scope; they look like duplicate reports of the same underlying issue." + "reason": "Same title and highly similar body with no distinct issue targets, filenames, or other differentiating evidence; these look like duplicate reports of the same issue." } ] }, "evaluator_result": { "accept": true, - "feedback": "The conclusion is grounded in the packet: the issues share an identical title, very similar body lengths, no files or explicit links, and a soft-similarity signal supports duplication. The canonical choice is also conservative and well-supported by earlier creation plus inbound reference. The wording is appropriately cautious overall, though the confidence is somewhat high for a soft-only match." + "feedback": "Grounded and conservative: both items share the exact title, have similar body sizes, no filenames or explicit targets, and only soft-similarity evidence is present. The canonical choice of #45145 is supported by earlier creation time and an inbound reference." }, "evaluator_used": true, "retried": false, @@ -21086,7 +21157,7 @@ "soft_pairs": [ "issue:45762|issue:45763" ], - "prepared_review_unit_hash": "10dd56511b12c8c968e005d9da0bb607ea30b817009b376d7d3902c72c6474bb", + "prepared_review_unit_hash": "b44917473bddedc1ab0ea264ae70723ba4d972aa137f4bcd526739f6ee702e00", "component_budget": { "node_count": 2, "item_count": 2, @@ -21108,8 +21179,8 @@ "aggressively_trimmed": false, "split": false, "analyst_result": { - "summary": "The two issues are unrelated closed items with different personal-name/date titles and no shared targets, files, or discussion, so they should not be clustered as duplicates.", - "confidence": 0.98, + "summary": "These two issues are not duplicates; they appear to be separate, unrelated personal-name/date entries with no shared bug report context or actionable technical content.", + "confidence": 0.96, "canonical_issue_reason": null, "canonical_pr_reason": null, "best_issue_reason": null, @@ -21119,7 +21190,7 @@ "left": "issue:45762", "right": "issue:45763", "accept": false, - "reason": "Only weak title similarity; no shared code, target, or discussion evidence, and they do not appear to describe the same underlying bug or request." + "reason": "Only superficial similarity in the titles; the names and dates differ and there is no shared issue context indicating the same underlying problem." } ] }, @@ -21149,7 +21220,7 @@ "soft_pairs": [ "issue:45764|issue:45765" ], - "prepared_review_unit_hash": "6853cc498a77a549f65c2927d2d67cc6c3de0ddd59d1c4f0012ac0c44344a5b3", + "prepared_review_unit_hash": "7eb00c8d765d382c58e73d7189567008d3bbeb11114ce44a4a1be6d582e62684", "component_budget": { "node_count": 2, "item_count": 2, @@ -21171,18 +21242,18 @@ "aggressively_trimmed": false, "split": false, "analyst_result": { - "summary": "The two closed issues are just unrelated personal-name/date titles with no body or linking evidence, so they do not look like duplicates of the same underlying bug or change.", - "confidence": 0.97, - "canonical_issue_reason": null, + "summary": "Two closed issues with unrelated personal-name/date titles and no body or code evidence; they do not look like duplicates of the same underlying bug.", + "confidence": 0.98, + "canonical_issue_reason": "Issue 45764 is the earliest-created item, so it is the only reasonable cluster representative if one must be chosen; however, it is not meaningfully better than 45765.", "canonical_pr_reason": null, - "best_issue_reason": "Neither issue is a good canonical representative for a duplicate cluster; if one must be chosen, issue 45764 is only marginally cleaner because the title is slightly more complete, but there is no substantive triage basis.", + "best_issue_reason": "Issue 45764 is marginally preferable as the cluster anchor only because it was created first and has the same low-information shape as 45765.", "best_pr_reason": null, "soft_edge_verdicts": [ { "left": "issue:45764", "right": "issue:45765", "accept": false, - "reason": "The titles refer to different people and different dates, with no shared bug report, code path, or change description; this is not the same underlying issue." + "reason": "They only share a superficial formatting pattern in the titles; there is no shared problem statement, code path, or other evidence that they describe the same issue." } ] }, @@ -21212,7 +21283,7 @@ "soft_pairs": [ "issue:45810|issue:46295" ], - "prepared_review_unit_hash": "d58777eb237bad2a02f476ad037e5901ad7638bf94665ea3b488727fb56addb6", + "prepared_review_unit_hash": "aba1664f0911fd1ff3a821cccfa763892c44ea5e4f818ba3fc81ef24195f12fc", "component_budget": { "node_count": 2, "item_count": 2, @@ -21234,8 +21305,8 @@ "aggressively_trimmed": false, "split": false, "analyst_result": { - "summary": "The two issues are unrelated and should not be clustered as duplicates: one asks for a new Qwen token-classification class, the other requests FP8 training support.", - "confidence": 0.98, + "summary": "The two issues are unrelated: one requests a new Qwen3_5 token-classification/value-model class, while the other asks for FP8 training support. They do not describe the same underlying bug or feature.", + "confidence": 0.99, "canonical_issue_reason": null, "canonical_pr_reason": null, "best_issue_reason": null, @@ -21245,7 +21316,7 @@ "left": "issue:45810", "right": "issue:46295", "accept": false, - "reason": "Different problem domains and code paths: model-head support for a specific architecture vs. training/runtime support for FP8. They do not describe the same underlying bug or change." + "reason": "Different requests in different areas: model-class addition for token classification vs FP8 training support. No shared code path or concrete implementation overlap." } ] }, @@ -21285,88 +21356,88 @@ "issue:45871|issue:45873", "issue:45872|issue:45873" ], - "prepared_review_unit_hash": "1fc5304a5052cf625466dce86ac61aef138ccb296d0846c170870543889c01cf", + "prepared_review_unit_hash": "16a18fb13f06f6a06b2f25033f00dfc7f5669d81a3c6d5aa012d90bc73016aa1", "component_budget": { "node_count": 5, "item_count": 5, "soft_pair_count": 8, - "serialized_chars": 5722, - "estimated_input_tokens": 1431, - "estimated_eval_tokens": 3118 + "serialized_chars": 5728, + "estimated_input_tokens": 1432, + "estimated_eval_tokens": 3120 }, "budget": { "node_count": 5, "item_count": 5, "soft_pair_count": 8, - "serialized_chars": 5722, - "estimated_input_tokens": 1431, - "estimated_eval_tokens": 3118 + "serialized_chars": 5728, + "estimated_input_tokens": 1432, + "estimated_eval_tokens": 3120 }, "overflow_policy": "truncate_then_skip", "trimmed": false, "aggressively_trimmed": false, "split": false, "analyst_result": { - "summary": "There are two exact duplicate issue pairs: 45870/45871 and 45872/45873. Issue 45869 is a separate dataset/sample post, and the two financial example topics differ enough to keep separate.", - "confidence": 0.94, - "canonical_issue_reason": "45870 is the cleanest representative of the exact duplicate 'Example: Financial time series prediction with DataForge' pair and is a better canonical label than the longer variants.", + "summary": "There are two clear duplicate issue pairs: 45870/45871 and 45872/45873. Issue 45869 is a separate dataset-promo topic and should not be merged with the DataForge example/service requests.", + "confidence": 0.96, + "canonical_issue_reason": "45870 is a clean representative of the exact-duplicate pair 45870/45871 and is the earliest item in that pair.", "canonical_pr_reason": null, - "best_issue_reason": "45870 is the strongest single issue to anchor the cluster because it is an exact duplicate target with a concise, neutral title; the others are either redundant copies or a different topic.", + "best_issue_reason": "45870 is the best single issue to keep as a representative, but note the cluster also contains a second, separate duplicate pair (45872/45873).", "best_pr_reason": null, "soft_edge_verdicts": [ { "left": "issue:45869", "right": "issue:45870", "accept": false, - "reason": "Financial dataset bundle promotion and a time-series example request are related to the same domain but not the same underlying issue/change." + "reason": "Different artifact types/topics: a dataset promotion versus an example request; not the same underlying issue." }, { "left": "issue:45869", "right": "issue:45871", "accept": false, - "reason": "The dataset listing and the example request are distinct; no evidence they describe one identical task or bug." + "reason": "Different artifact types/topics: a dataset promotion versus an example request; not the same underlying issue." }, { "left": "issue:45870", "right": "issue:45871", "accept": true, - "reason": "Exact same title and apparent same example request; these are duplicates." + "reason": "Exact title match and strong duplicate signals; same underlying issue." }, { "left": "issue:45870", "right": "issue:45872", "accept": false, - "reason": "Both mention financial DataForge examples, but one is a basic prediction example and the other adds a scraping service, so they are not clearly the same underlying request." + "reason": "Related financial/DataForge theme, but the titles describe different requests (example prediction vs scraping service); not the same issue." }, { "left": "issue:45870", "right": "issue:45873", "accept": false, - "reason": "Same as the 45870/45872 pair: related theme, but different concrete scope and not safe to merge as duplicates." + "reason": "Related financial/DataForge theme, but the titles describe different requests (example prediction vs scraping service); not the same issue." }, { "left": "issue:45871", "right": "issue:45872", "accept": false, - "reason": "The titles overlap in broad subject matter, but the requested example scopes differ enough to treat them as separate issues." + "reason": "Related theme only; these are different requests with different scopes, so they should not be deduplicated." }, { "left": "issue:45871", "right": "issue:45873", "accept": false, - "reason": "Not the same concrete issue; the scraping-service variant is broader and distinct from the plain prediction example." + "reason": "Related theme only; these are different requests with different scopes, so they should not be deduplicated." }, { "left": "issue:45872", "right": "issue:45873", "accept": true, - "reason": "Exact same title and scope; these are duplicates." + "reason": "Exact title match and strong duplicate signals; same underlying issue." } ] }, "evaluator_result": { "accept": true, - "feedback": "The conclusions are conservative and grounded in the packet. The duplicate calls for 45870/45871 and 45872/45873 are supported by identical titles and matching body lengths, and the rejections for the other pairs avoid overclaiming beyond the available evidence. The summary is appropriately cautious about 45869 and about the different scope of the remaining financial example issue." + "feedback": "The summary is grounded in the packet: 45870/45871 and 45872/45873 are clear duplicate pairs based on identical titles and matching body lengths, and 45869 is reasonably kept separate as a different dataset/sample topic. The soft-edge verdicts are conservative and do not overstate cross-pair duplication." }, "evaluator_used": true, "retried": false, @@ -21393,40 +21464,40 @@ "soft_pairs": [ "issue:46439|issue:46441" ], - "prepared_review_unit_hash": "8f3c9a6e3f283ee7d9175bd36bc49762daf8903f8255b55a3bf77e12a6e8fdaa", + "prepared_review_unit_hash": "6e8649b0f572a8725686f3a63a6d5c377b9c66c969ec0e6d94a117598b997c14", "component_budget": { "node_count": 2, "item_count": 2, "soft_pair_count": 1, - "serialized_chars": 1641, - "estimated_input_tokens": 411, - "estimated_eval_tokens": 1078 + "serialized_chars": 1640, + "estimated_input_tokens": 410, + "estimated_eval_tokens": 1076 }, "budget": { "node_count": 2, "item_count": 2, "soft_pair_count": 1, - "serialized_chars": 1641, - "estimated_input_tokens": 411, - "estimated_eval_tokens": 1078 + "serialized_chars": 1640, + "estimated_input_tokens": 410, + "estimated_eval_tokens": 1076 }, "overflow_policy": "truncate_then_skip", "trimmed": false, "aggressively_trimmed": false, "split": false, "analyst_result": { - "summary": "Both issues concern static cache behavior with linear-attention/Qwen3.5, but they describe different failure modes: one is a shape corruption RuntimeError during early initialization, the other is a KeyError for `linear_attention`. They look related in subsystem, not clearly the same underlying bug.", - "confidence": 0.82, - "canonical_issue_reason": "Issue 46439 is the better canonical issue because it describes the more general bug in `static` cache `early_initialization` affecting linear-attention layers, with a concrete mechanism and symptom.", + "summary": "Both issues concern `cache_implementation=\"static\"` and `linear_attention`, but they describe different failure modes: one is shape corruption during `early_initialization`, the other is a `KeyError` for `linear_attention` in Qwen3.5. They are related, but not clearly the same bug.", + "confidence": 0.76, + "canonical_issue_reason": "Issue 46439 is the better canonical issue because it describes a broader, more root-cause-like problem in static cache `early_initialization` affecting linear-attention layers, rather than a model-specific crash symptom.", "canonical_pr_reason": null, - "best_issue_reason": "Issue 46439 is the best representative of the cluster since it captures the broader root-cause description rather than a model-specific crash symptom.", + "best_issue_reason": "Issue 46439 is the best overall issue candidate because it is more general and appears to capture the underlying corruption mechanism, while 46441 reads more like a specific downstream manifestation or compatibility failure.", "best_pr_reason": null, "soft_edge_verdicts": [ { "left": "issue:46439", "right": "issue:46441", "accept": false, - "reason": "Related subsystem and likely overlapping fix area, but the reported failures are not the same concrete bug: one is wrong-shape corruption in `early_initialization`, the other is a `KeyError: 'linear_attention'`. Too little evidence they would be merged into one fix." + "reason": "Related subsystem, but not the same concrete bug: 46439 reports wrong-shape/runtime corruption in `early_initialization`, whereas 46441 reports a `KeyError: 'linear_attention'` for Qwen3.5. The failure modes and likely fixes differ." } ] }, @@ -21456,12 +21527,12 @@ "soft_pairs": [ "issue:46551|issue:46552" ], - "prepared_review_unit_hash": "e8c558a40f5a28fe12fd2fdf7a8ad3df38642a8efe19c7aff943ec525089307a", + "prepared_review_unit_hash": "e1fe5644e1faed07e4494b6e7a61fc0b5888d5d1d1d290c4ec380234cd842c8d", "component_budget": { "node_count": 2, "item_count": 2, "soft_pair_count": 1, - "serialized_chars": 1706, + "serialized_chars": 1707, "estimated_input_tokens": 427, "estimated_eval_tokens": 1110 }, @@ -21469,7 +21540,7 @@ "node_count": 2, "item_count": 2, "soft_pair_count": 1, - "serialized_chars": 1706, + "serialized_chars": 1707, "estimated_input_tokens": 427, "estimated_eval_tokens": 1110 }, @@ -21478,24 +21549,24 @@ "aggressively_trimmed": false, "split": false, "analyst_result": { - "summary": "The two issues are effectively duplicates: same title, same described tokenizer bug, and one is just a closed twin of the other. Keep the open issue as the canonical tracker.", + "summary": "The two issues are effectively identical reports of the same Wav2Vec2CTCTokenizer v5 special-token regression; they should be clustered together. The open issue is the better canonical tracker because it is newer and already has inbound references.", "confidence": 0.99, - "canonical_issue_reason": "Issue 46552 is the open instance and already has inbound references, so it is the better long-lived tracker; 46551 is closed and redundant.", + "canonical_issue_reason": "issue:46552 is the active/open report with inbound references, so it is the better canonical issue to keep as the cluster representative.", "canonical_pr_reason": null, - "best_issue_reason": "46552 is the best issue to surface because it remains open and appears to be the active discussion thread for the same bug.", + "best_issue_reason": "issue:46552 is the best single issue because it is open, current, and already attracting references, making it the most useful tracker for the bug.", "best_pr_reason": null, "soft_edge_verdicts": [ { "left": "issue:46551", "right": "issue:46552", "accept": true, - "reason": "Same title and same underlying Wav2Vec2CTCTokenizer v5 bug; the issues are duplicates rather than merely related." + "reason": "Same title and same underlying bug report; the issues are duplicate descriptions of the same tokenizer regression." } ] }, "evaluator_result": { "accept": true, - "feedback": "The decision is well grounded in the packet: the issues have identical titles, identical body length, and the open issue 46552 has inbound references while 46551 is closed. Treating 46551 as a duplicate and 46552 as the canonical tracker is conservative and supported." + "feedback": "The summary is grounded in the packet: the two issues have identical titles and body lengths, and the pair is marked with soft_similarity. The canonical choice of issue:46552 is also supported because it is open and has inbound references, while issue:46551 is closed. The soft-edge duplicate verdict is conservative and reasonable here." }, "evaluator_used": true, "retried": false, @@ -21522,7 +21593,7 @@ "soft_pairs": [ "issue:46657|issue:46692" ], - "prepared_review_unit_hash": "3b7774f6885f3f25976bdcebaccdfd7ae0d4d456b96e01ce17af6159656a1962", + "prepared_review_unit_hash": "00a3d2d76d9bed7b372649690cacfe2a97cc34e6ceb8a42a124f12a5df24d627", "component_budget": { "node_count": 2, "item_count": 2, @@ -21544,18 +21615,18 @@ "aggressively_trimmed": false, "split": false, "analyst_result": { - "summary": "The two issues both involve training-time problems in object detection, but they describe different failure modes: one is a tensor shape mismatch, the other is an out-of-memory condition after several epochs. They do not look like duplicates.", - "confidence": 0.94, - "canonical_issue_reason": "Neither issue is a clear duplicate of the other; if one had to be picked as the more general training bug report, #46657 is the cleaner canonical candidate because it describes a concrete error rather than a later-training resource failure.", + "summary": "The two issues are both training-related, but they describe different failure modes: a tensor shape mismatch in object detection training vs. an eventual out-of-memory condition in RT-DETR v2. They should not be treated as duplicates.", + "confidence": 0.93, + "canonical_issue_reason": "Issue #46657 is the cleaner representative if one must be chosen: it states a concrete, specific bug symptom (tensor size mismatch) and is already closed. #46692 is a different problem with a different failure mode.", "canonical_pr_reason": null, - "best_issue_reason": "#46657 is the better standalone issue because it names a specific reproducible failure mode (tensor size mismatch) and is more likely to be actionable for triage.", + "best_issue_reason": "#46657 is the best issue candidate because it is narrowly scoped and clearly describes a distinct code-path failure. #46692 appears unrelated enough that it should not be grouped under the same canonical issue.", "best_pr_reason": null, "soft_edge_verdicts": [ { "left": "issue:46657", "right": "issue:46692", "accept": false, - "reason": "Different underlying problems: a tensor dimension mismatch versus GPU memory exhaustion after epochs. They share a broad training context only, not the same bug." + "reason": "Do not merge: the symptoms and likely root causes differ. A tensor dimension mismatch during object detection training is not the same underlying bug as memory growth leading to OOM after epochs in RT-DETR v2." } ] }, @@ -21585,7 +21656,7 @@ "soft_pairs": [ "issue:46822|issue:46823" ], - "prepared_review_unit_hash": "f4a733d3e07670a4731d47df737745ec45428646f5ad34b5e78651858cf6eee0", + "prepared_review_unit_hash": "0bf2f8392e7576114b7f50c1a2997d46d6f4d8165d4f1612ca0a761aab22a7f6", "component_budget": { "node_count": 2, "item_count": 2, @@ -21607,18 +21678,18 @@ "aggressively_trimmed": false, "split": false, "analyst_result": { - "summary": "These two issues describe different bugs in different subsystems: ZeRO-3 model partitioning during load vs. 4-bit quantized loading with device_map/max_memory dispatch. They should not be clustered as duplicates.", + "summary": "These are two different bug reports, not duplicates: one is about ZeRO-3 zero.Init failing to partition a composite model submodule and causing OOM, the other is about 4-bit + auto device mapping incorrectly sending a fitting model to CPU and raising ValueError.", "confidence": 0.98, - "canonical_issue_reason": null, + "canonical_issue_reason": "No single canonical issue: the titles describe unrelated loading/dispatch bugs in different code paths.", "canonical_pr_reason": null, - "best_issue_reason": "No single canonical issue fits because the reports are unrelated; if one must be picked for prominence, issue 46823 has more inbound references and a clearer reproduction path.", + "best_issue_reason": "No clear best issue for the cluster because the items are not the same underlying problem.", "best_pr_reason": null, "soft_edge_verdicts": [ { "left": "issue:46822", "right": "issue:46823", "accept": false, - "reason": "Reject: the titles and described failure modes point to different code paths and different root causes, so they are not the same underlying bug." + "reason": "Only superficial similarity around model loading/memory placement; one is a ZeRO-3 partitioning/OOM problem, the other is a 4-bit device_map/max_memory dispatch error. Different root causes and different fixes." } ] }, @@ -21696,12 +21767,12 @@ "issue:46998|issue:47120", "issue:47091|issue:47120" ], - "prepared_review_unit_hash": "c9414c5ca6b278c708d5b48523a4c367af8ecde7870b4cd0f407b8536186b112", + "prepared_review_unit_hash": "6ea204b2d3853f0db4fdcda3031475bae1f8f419968a354a3b92735806966a94", "component_budget": { "node_count": 12, "item_count": 12, "soft_pair_count": 39, - "serialized_chars": 20780, + "serialized_chars": 20779, "estimated_input_tokens": 5195, "estimated_eval_tokens": 10646 }, @@ -21709,7 +21780,7 @@ "node_count": 12, "item_count": 12, "soft_pair_count": 39, - "serialized_chars": 20780, + "serialized_chars": 20779, "estimated_input_tokens": 5195, "estimated_eval_tokens": 10646 }, @@ -21718,246 +21789,252 @@ "aggressively_trimmed": false, "split": false, "analyst_result": { - "summary": "These look like a dated series of repetitive integration-failure triage issues for the same project, but each item is a separate daily report rather than a duplicate bug report. The open 2026-07-07 issue is the best current representative.", - "confidence": 0.97, - "canonical_issue_reason": "issue:47120 is the latest open triage report and already has inbound references, so it is the most useful cluster representative.", + "summary": "These are date-stamped Serde/Serge integration triage reports with highly similar template text, but they do not look like the same underlying bug report. They read more like separate daily tracking issues, so the soft duplicate edges should be rejected.", + "confidence": 0.93, + "canonical_issue_reason": "Issue 47120 is the best anchor if one must be chosen: it is the latest open triage thread and has the most recent inbound references, but the set is not actually duplicate-heavy.", "canonical_pr_reason": null, - "best_issue_reason": "Same reason: it is the current open daily triage issue and best matches the active tracking role for this cluster.", + "best_issue_reason": "Issue 47120 is the most useful single issue for follow-up because it is current, open, and appears to supersede the earlier daily triage entries.", "best_pr_reason": null, "soft_edge_verdicts": [ { "left": "issue:46906", "right": "issue:46940", "accept": false, - "reason": "Same triage template and project, but different dates indicate separate daily reports, not the same issue." + "reason": "Same triage template and adjacent dates, but this looks like a separate daily report rather than one bug duplicated." }, { "left": "issue:46906", "right": "issue:46948", "accept": false, - "reason": "Highly similar wording, but these are distinct day-stamped triage issues." + "reason": "Highly similar wording, but the date-stamped triage entries suggest distinct daily logs, not the same issue." }, { "left": "issue:46906", "right": "issue:46969", "accept": false, - "reason": "Different report dates; similarity comes from the recurring triage format, not a shared underlying bug." + "reason": "Shared template only; no evidence these describe the same concrete failure." }, { "left": "issue:46906", "right": "issue:46985", "accept": false, - "reason": "Separate daily integration triage entries, not one duplicate issue." + "reason": "These are separate date-specific triage issues, not a single duplicate bug." }, { "left": "issue:46906", "right": "issue:47037", "accept": false, - "reason": "Same series of daily triage reports, but each date corresponds to a different issue." + "reason": "Similarity comes from the repeated title format, but the issues are different daily reports." }, { "left": "issue:46940", "right": "issue:46948", "accept": false, - "reason": "Consecutive daily reports with templated content; not the same underlying report." + "reason": "Adjacent triage threads with the same template, but likely distinct day-by-day reports." }, { "left": "issue:46940", "right": "issue:46969", "accept": false, - "reason": "Different dates and likely different run data; do not collapse these into one issue." + "reason": "No concrete shared failure is shown beyond the recurring triage wording." }, { "left": "issue:46940", "right": "issue:46985", "accept": false, - "reason": "These are recurring triage tickets, not duplicates of one bug report." + "reason": "Similar title pattern only; not enough to treat as one duplicated issue." }, { "left": "issue:46940", "right": "issue:47049", "accept": false, - "reason": "Same subsystem and template, but distinct daily triage issues." + "reason": "Different dated triage entries, so they should not be merged as duplicates." }, { "left": "issue:46948", "right": "issue:46969", "accept": false, - "reason": "Similar triage text is expected here; the day-specific titles show separate reports." + "reason": "Same recurring triage format, but these appear to be separate daily log issues." }, { "left": "issue:46948", "right": "issue:46985", "accept": false, - "reason": "Recurring daily triage items, not a single issue repeated verbatim." + "reason": "The overlap is in template text, not a shared underlying bug report." }, { "left": "issue:46948", "right": "issue:46998", "accept": false, - "reason": "Different dates mean different triage reports, despite strong textual overlap." + "reason": "Date-stamped triage threads are not strong duplicate evidence by themselves." }, { "left": "issue:46969", "right": "issue:46985", "accept": false, - "reason": "Adjacent daily triage issues, but not the same underlying problem statement." + "reason": "They look like consecutive tracking reports, not one duplicate issue." }, { "left": "issue:46969", "right": "issue:47037", "accept": false, - "reason": "The similarity is from the shared template; the report date makes them distinct." + "reason": "Same subsystem and template, but no sign of the same concrete failure." }, { "left": "issue:46969", "right": "issue:47091", "accept": false, - "reason": "These are separate daily integration-failure triage reports." + "reason": "These are separate daily triage entries rather than duplicates of one bug." }, { "left": "issue:46985", "right": "issue:46998", "accept": false, - "reason": "Templated daily triage entries with different dates; not duplicates." + "reason": "Recurring title pattern only; no concrete duplicate bug evidence." }, { "left": "issue:46985", "right": "issue:47037", "accept": false, - "reason": "Same recurring triage series, but a different day/run." + "reason": "Looks like another daily triage note, not the same issue instance." }, { "left": "issue:46985", "right": "issue:47049", "accept": false, - "reason": "Different daily report, not the same issue instance." + "reason": "Date-based tracking issues should stay separate unless they share the same failure details." }, { - "left": "issue:46985", - "right": "issue:47091", - "accept": false, - "reason": "These appear to be day-by-day tracking issues rather than duplicate bug reports." - }, - { - "left": "issue:46985", - "right": "issue:47102", + "left": "issue:46998", + "right": "issue:47053", "accept": false, - "reason": "The date-stamped titles indicate separate daily triage tickets." + "reason": "The common wording is from the triage template, not a shared defect." }, { "left": "issue:46998", "right": "issue:47037", "accept": false, - "reason": "High textual overlap is due to the triage template, not a shared unique bug." + "reason": "Separate daily triage reports; not enough to merge as duplicates." }, { "left": "issue:46998", "right": "issue:47049", "accept": false, - "reason": "Distinct day-stamped integration triage issue." + "reason": "These appear to document different days of integration failures, not one bug." }, { - "left": "issue:46998", - "right": "issue:47053", + "left": "issue:46985", + "right": "issue:46998", "accept": false, - "reason": "Different daily report; should not be merged as duplicates." + "reason": "Duplicate-looking title alone is insufficient; these are distinct date-stamped triage issues." }, { "left": "issue:46998", "right": "issue:47091", "accept": false, - "reason": "Separate issue in the same recurring series." + "reason": "Same template, but no evidence of the same underlying problem." }, { - "left": "issue:46998", - "right": "issue:47120", + "left": "issue:47037", + "right": "issue:47049", "accept": false, - "reason": "Same recurring triage pattern, but each dated issue stands alone." + "reason": "Consecutive triage reports with similar wording are not duplicates by default." }, { "left": "issue:47037", - "right": "issue:47049", + "right": "issue:47091", "accept": false, - "reason": "Consecutive daily triage reports, not one underlying issue." + "reason": "These are separate daily triage threads, not one merged bug report." }, { "left": "issue:47037", "right": "issue:47053", "accept": false, - "reason": "Different date and likely different integration run; not a duplicate." + "reason": "Shared phrasing only; no concrete identical failure is shown." }, { - "left": "issue:47037", + "left": "issue:47049", + "right": "issue:47053", + "accept": false, + "reason": "Different dates and no specific shared bug details, so reject duplicate merge." + }, + { + "left": "issue:47049", "right": "issue:47091", "accept": false, - "reason": "Same template, different daily report." + "reason": "Looks like another daily tracking entry rather than the same issue." }, { - "left": "issue:47037", + "left": "issue:47053", + "right": "issue:47091", + "accept": false, + "reason": "Template similarity is expected here and does not indicate a duplicate." + }, + { + "left": "issue:47053", "right": "issue:47102", "accept": false, - "reason": "These are separate day-specific triage issues." + "reason": "Separate date-stamped triage issues; not the same underlying change or bug." }, { - "left": "issue:47049", - "right": "issue:47053", + "left": "issue:47091", + "right": "issue:47102", "accept": false, - "reason": "Different daily integration failure triage tickets." + "reason": "No shared concrete failure beyond the repeated triage title." }, { - "left": "issue:47049", + "left": "issue:46985", "right": "issue:47091", "accept": false, - "reason": "The overlap is from the recurring report format, not the same bug." + "reason": "Recurring triage format only; these should remain distinct issues." }, { - "left": "issue:47049", + "left": "issue:47037", "right": "issue:47102", "accept": false, - "reason": "Separate day-stamped triage issue." + "reason": "They read as independent daily reports, not duplicates." }, { - "left": "issue:47049", - "right": "issue:47120", + "left": "issue:46985", + "right": "issue:47102", "accept": false, - "reason": "Different dates mean different reports, even if the content is similar." + "reason": "Same subsystem and template, but no same-bug evidence." }, { - "left": "issue:47053", - "right": "issue:47091", + "left": "issue:47049", + "right": "issue:47102", "accept": false, - "reason": "Recurring triage template; not a duplicate issue." + "reason": "Different dated tracking issues; do not collapse them as duplicates." }, { - "left": "issue:47053", - "right": "issue:47102", + "left": "issue:47102", + "right": "issue:47120", "accept": false, - "reason": "Separate daily report, not the same underlying report." + "reason": "Adjacent daily triage threads, but still separate reports rather than one duplicate issue." }, { "left": "issue:47053", "right": "issue:47120", "accept": false, - "reason": "Different day and separate integration triage entry." + "reason": "The overlap is the shared triage template, not a shared concrete bug." }, { - "left": "issue:47091", - "right": "issue:47102", + "left": "issue:47049", + "right": "issue:47120", "accept": false, - "reason": "Distinct daily integration-failure triage issues." + "reason": "These look like sequential daily integration reports, not duplicate issues." }, { - "left": "issue:47091", + "left": "issue:46998", "right": "issue:47120", "accept": false, - "reason": "Same family of reports, but not the same issue." + "reason": "No evidence of the same underlying failure beyond the repeated title pattern." }, { - "left": "issue:47102", + "left": "issue:47091", "right": "issue:47120", "accept": false, - "reason": "Different dates and separate triage instances; keep both." + "reason": "Date-stamped triage entries should not be merged without identical bug details." } ] }, @@ -21987,12 +22064,12 @@ "soft_pairs": [ "issue:46983|issue:46984" ], - "prepared_review_unit_hash": "e928b9596ab5739ef97ef3f231fc3c6ca7e36f8ef7160f8df5923e5a98c5066b", + "prepared_review_unit_hash": "41f98382c30e7a786d511356165bd3ffea9d76d23a67bab147758d17f6031a11", "component_budget": { "node_count": 2, "item_count": 2, "soft_pair_count": 1, - "serialized_chars": 1490, + "serialized_chars": 1492, "estimated_input_tokens": 373, "estimated_eval_tokens": 1002 }, @@ -22000,7 +22077,7 @@ "node_count": 2, "item_count": 2, "soft_pair_count": 1, - "serialized_chars": 1490, + "serialized_chars": 1492, "estimated_input_tokens": 373, "estimated_eval_tokens": 1002 }, @@ -22009,27 +22086,27 @@ "aggressively_trimmed": false, "split": false, "analyst_result": { - "summary": "The two issues are highly similar RWKV7 support requests with identical titles, close creation times, and strong textual overlap, but there is no explicit linking evidence, so they should be treated as likely related/duplicate rather than confirmed identical. The open issue is the better live tracker.", - "confidence": 0.89, - "canonical_issue_reason": "Issue 46984 is open and is the better live tracker for this likely duplicate RWKV7 support request.", + "summary": "The two closed issues appear to be duplicates: same title, very similar body length, and high textual similarity. Issue #46983 is the better canonical representative because it was opened first.", + "confidence": 0.98, + "canonical_issue_reason": "#46983 and #46984 have the same title and closely matching descriptions; #46983 was created first, so it is the cleaner canonical issue.", "canonical_pr_reason": null, - "best_issue_reason": "Issue 46984 is the best representative because it is open and appears to cover the same RWKV7 model support request as 46983.", + "best_issue_reason": "#46983 is the best issue to keep as the canonical record because it is the earliest instance of the duplicate report and there is no evidence that #46984 adds distinct scope.", "best_pr_reason": null, "soft_edge_verdicts": [ { "left": "issue:46983", "right": "issue:46984", "accept": true, - "reason": "Identical title, near-identical timing, and very high textual similarity make these likely the same feature request, though the packet does not explicitly confirm it." + "reason": "Same issue title and highly similar content indicate the same underlying request; this is a clear duplicate pair." } ] }, "evaluator_result": { - "accept": false, - "feedback": "The packet supports treating these as highly similar reports with the same title and close creation times, and choosing the open issue as the better live tracker is reasonable. However, the analyst overstates the evidence by asserting they are definitively the same issue report and that the closed one is likely a duplicate/superseded report. There is no explicit link or stronger evidence in the packet, so the conclusion should stay conservative: likely related/duplicate, not confirmed identical." + "accept": true, + "feedback": "The conclusion is grounded in the packet: both issues have the same title, the soft-similarity score is high (Jaccard 0.837), and #46983 was created slightly earlier, which supports it as the canonical record. The duplicate verdict is conservative enough given the available evidence." }, "evaluator_used": true, - "retried": true, + "retried": false, "accepted_nontrivial_soft_edge": true, "error_kind": null, "error_message": null @@ -22053,12 +22130,12 @@ "soft_pairs": [ "issue:47007|issue:47015" ], - "prepared_review_unit_hash": "b972d39eab70cc3213ffe9eba75b260d565c3c231d44e6c900bc57064f14c450", + "prepared_review_unit_hash": "b1876cf596406f166329b82bd403b412872158a5fd60328c84575d8ee76b5622", "component_budget": { "node_count": 2, "item_count": 2, "soft_pair_count": 1, - "serialized_chars": 1626, + "serialized_chars": 1625, "estimated_input_tokens": 407, "estimated_eval_tokens": 1070 }, @@ -22066,7 +22143,7 @@ "node_count": 2, "item_count": 2, "soft_pair_count": 1, - "serialized_chars": 1626, + "serialized_chars": 1625, "estimated_input_tokens": 407, "estimated_eval_tokens": 1070 }, @@ -22075,18 +22152,18 @@ "aggressively_trimmed": false, "split": false, "analyst_result": { - "summary": "The two issues are topically related but do not look like the same underlying bug/change; they read more like separate article/proposal-style submissions about AI-driven vulnerability discovery.", - "confidence": 0.92, - "canonical_issue_reason": "No canonical duplicate: the titles and descriptions suggest two different writeups/proposals rather than the same underlying issue.", + "summary": "The two issues are topically related around AI-driven vulnerability discovery, but they do not look like the same underlying report. They read more like separate writeups/proposals with different emphasis and no explicit linkage.", + "confidence": 0.87, + "canonical_issue_reason": "No clear canonical duplicate: the titles point to different angles (lessons from LLM usage vs. a novel transformer-based cybersecurity system) and there is no shared evidence tying them to one issue.", "canonical_pr_reason": null, - "best_issue_reason": "Issue 47007 is the more specific and polished of the two, but only as a representative item; it is not a true duplicate target.", + "best_issue_reason": "If one issue must represent the cluster, 47007 is slightly more concrete and scoped, but it still is not a true duplicate of 47015.", "best_pr_reason": null, "soft_edge_verdicts": [ { "left": "issue:47007", "right": "issue:47015", "accept": false, - "reason": "They are only loosely similar at a thematic level. No shared code path, filenames, explicit targets, or concrete bug/change indicate the same underlying issue." + "reason": "Same broad theme, but not the same underlying bug or change; the titles suggest different content and intent, with no explicit issue links or shared artifacts." } ] }, @@ -22116,7 +22193,7 @@ "soft_pairs": [ "issue:47138|issue:47139" ], - "prepared_review_unit_hash": "01aaf023f332a5f24a6696a0c1b9aae4fe34dc99f9b9c24cf18e2a9d04c6b5bb", + "prepared_review_unit_hash": "71cb8b08b9036beee9481427d0067b51238c11f29ce42550103b65a64494125a", "component_budget": { "node_count": 2, "item_count": 2, @@ -22138,27 +22215,27 @@ "aggressively_trimmed": false, "split": false, "analyst_result": { - "summary": "These two issues are highly likely duplicates: they have the same title, identical body length, and were opened minutes apart, with issue 47139 already attracting inbound references. The packet does not prove the bodies are identical, so a cautious duplicate judgment is more appropriate than calling them exact duplicates.", - "confidence": 0.94, - "canonical_issue_reason": "issue:47139 is the better canonical candidate because it already has inbound references, suggesting others are pointing to it as the main report.", + "summary": "The two issues are effectively duplicates: same title, same apparent body length, and no distinguishing metadata beyond inbound references. Issue 47139 is the better canonical/best representative because it already has inbound references and likely serves as the more connected thread.", + "confidence": 0.99, + "canonical_issue_reason": "Issue 47139 is the better canonical issue because it has inbound references, making it the more established discussion thread for the same reported bug.", "canonical_pr_reason": null, - "best_issue_reason": "issue:47139 is the best representative issue since it appears slightly more established and likely serves as the duplicate target.", + "best_issue_reason": "Issue 47139 is the best issue candidate for the same reason: it is the more referenced duplicate and likely the preferred tracking thread.", "best_pr_reason": null, "soft_edge_verdicts": [ { "left": "issue:47138", "right": "issue:47139", "accept": true, - "reason": "Strong duplicate signal: identical title, matching body length, and near-simultaneous creation indicate the same underlying bug report, though the bodies are not explicitly confirmed identical." + "reason": "Same title and no visible differences in scope; they appear to report the same underlying bug, so this is a clear duplicate pair." } ] }, "evaluator_result": { - "accept": false, - "feedback": "The duplicate judgment is plausible, but the summary overstates the evidence by calling them \"exact duplicates.\" The packet only shows identical titles, identical body lengths, close timestamps, and one issue with inbound references; it does not confirm the bodies are the same. A more conservative phrasing like \"highly likely duplicates\" would be better." + "accept": true, + "feedback": "The assessment is grounded in the packet: the two issues have identical titles and body lengths, and issue 47139 has the stronger inbound-reference signal, so choosing it as the canonical/best representative is conservative and reasonable." }, "evaluator_used": true, - "retried": true, + "retried": false, "accepted_nontrivial_soft_edge": true, "error_kind": null, "error_message": null @@ -22174,7 +22251,7 @@ "source": "llm", "cache_hit": false, "model": "gpt-5.4-mini?service_tier=flex", - "cluster_id": "cluster-40546-46", + "cluster_id": "cluster-40546-42", "nodes": [ "pull_request:43451", "pull_request:44517", @@ -22192,7 +22269,6 @@ "pull_request:46958", "pull_request:46974", "pull_request:46995", - "pull_request:47025", "pull_request:47034", "pull_request:47035", "pull_request:47069", @@ -22210,14 +22286,13 @@ "pull_request:47114", "pull_request:47134", "pull_request:47135", - "pull_request:47136", - "pull_request:47137", "pull_request:47142", "pull_request:47143", - "pull_request:47148", - "pull_request:47149", "pull_request:47156", - "pull_request:47157" + "pull_request:47157", + "pull_request:47161", + "pull_request:47163", + "pull_request:47166" ], "soft_pairs": [ "pull_request:46696|pull_request:47134", @@ -22229,368 +22304,458 @@ "pull_request:47072|pull_request:47089", "pull_request:47077|pull_request:47157", "pull_request:47077|pull_request:47078", - "pull_request:47072|pull_request:47148", "pull_request:46266|pull_request:46957", - "pull_request:47114|pull_request:47148", "pull_request:47072|pull_request:47134", "pull_request:47114|pull_request:47135", "pull_request:44517|pull_request:47082", "pull_request:47034|pull_request:47035", + "pull_request:46995|pull_request:47163", "pull_request:47069|pull_request:47089", "pull_request:46196|pull_request:46995", - "pull_request:46266|pull_request:47148", + "pull_request:47135|pull_request:47163", "pull_request:47093|pull_request:47114", "pull_request:46707|pull_request:47108", "pull_request:46707|pull_request:47114", - "pull_request:47077|pull_request:47149", - "pull_request:46707|pull_request:47148", "pull_request:47089|pull_request:47113", + "pull_request:47093|pull_request:47161", "pull_request:46266|pull_request:46707", "pull_request:46995|pull_request:47143", "pull_request:47082|pull_request:47108", - "pull_request:47114|pull_request:47149", "pull_request:47108|pull_request:47114", "pull_request:47078|pull_request:47156", "pull_request:47077|pull_request:47114", - "pull_request:46912|pull_request:47025", "pull_request:47093|pull_request:47143", + "pull_request:47093|pull_request:47163", "pull_request:47134|pull_request:47135", + "pull_request:46995|pull_request:47161", + "pull_request:46707|pull_request:46957", "pull_request:46658|pull_request:47114", - "pull_request:47108|pull_request:47148", "pull_request:46707|pull_request:47082", "pull_request:46266|pull_request:47114", - "pull_request:47069|pull_request:47136", "pull_request:47114|pull_request:47156", - "pull_request:46912|pull_request:47078", "pull_request:46912|pull_request:47082", - "pull_request:47082|pull_request:47148", - "pull_request:47114|pull_request:47136", - "pull_request:47035|pull_request:47089", + "pull_request:46912|pull_request:47078", + "pull_request:47134|pull_request:47161", + "pull_request:43451|pull_request:47082", "pull_request:46895|pull_request:47134", + "pull_request:47113|pull_request:47161", "pull_request:47100|pull_request:47134", - "pull_request:46196|pull_request:47025", + "pull_request:47035|pull_request:47089", "pull_request:43451|pull_request:46995", "pull_request:47100|pull_request:47142", - "pull_request:46895|pull_request:47137", - "pull_request:46836|pull_request:46995", - "pull_request:46707|pull_request:47025", + "pull_request:47156|pull_request:47166", + "pull_request:47035|pull_request:47161", "pull_request:46266|pull_request:47034", - "pull_request:47100|pull_request:47113", + "pull_request:47034|pull_request:47108", "pull_request:47035|pull_request:47113", "pull_request:47035|pull_request:47069", - "pull_request:47034|pull_request:47108", + "pull_request:47100|pull_request:47113", "pull_request:46836|pull_request:47135", - "pull_request:46995|pull_request:47096", - "pull_request:47142|pull_request:47149", - "pull_request:47025|pull_request:47114", + "pull_request:46957|pull_request:47108", + "pull_request:47100|pull_request:47161", "pull_request:47069|pull_request:47113", - "pull_request:46912|pull_request:47136", + "pull_request:46196|pull_request:47161", "pull_request:47069|pull_request:47082", - "pull_request:46196|pull_request:47148", "pull_request:47089|pull_request:47134", - "pull_request:46707|pull_request:46957", - "pull_request:47100|pull_request:47149", + "pull_request:43451|pull_request:46266", "pull_request:43451|pull_request:47078", - "pull_request:47078|pull_request:47096" + "pull_request:47078|pull_request:47096", + "pull_request:47135|pull_request:47161", + "pull_request:43451|pull_request:44517", + "pull_request:47100|pull_request:47143", + "pull_request:46912|pull_request:47113", + "pull_request:46912|pull_request:47142", + "pull_request:44517|pull_request:47143", + "pull_request:44517|pull_request:47163" ], - "prepared_review_unit_hash": "5bbed105a4360f045f1ffa0d024b4299ae695c8096e7e141810ec02f0c7bd84a", + "prepared_review_unit_hash": "9adb1c334be1f69b501c576278b8e1d63e4954df1016345375d33531a285adb6", "component_budget": { - "node_count": 46, - "item_count": 46, - "soft_pair_count": 133, - "serialized_chars": 159768, - "estimated_input_tokens": 39942, - "estimated_eval_tokens": 80140 - }, - "budget": { "node_count": 42, "item_count": 42, + "soft_pair_count": 126, + "serialized_chars": 150672, + "estimated_input_tokens": 37668, + "estimated_eval_tokens": 75592 + }, + "budget": { + "node_count": 40, + "item_count": 40, "soft_pair_count": 72, - "serialized_chars": 118056, - "estimated_input_tokens": 29514, - "estimated_eval_tokens": 59284 + "serialized_chars": 115711, + "estimated_input_tokens": 28928, + "estimated_eval_tokens": 58112 }, "overflow_policy": "truncate_then_skip", "trimmed": false, "aggressively_trimmed": false, "split": true, "analyst_result": { - "summary": "This cluster is mostly a grab-bag of unrelated PRs. The closest-looking pairs are still different bug fixes or different docs/localization updates, so there is no safe duplicate canonical here.", - "confidence": 0.97, + "summary": "This cluster is mostly unrelated PRs spanning new model additions, docs, exporter/runtime tweaks, and small bug fixes. Only one soft edge looks like a true duplicate: the two CohereASR training-loss double-shift fixes.", + "confidence": 0.92, "canonical_issue_reason": null, - "canonical_pr_reason": "No single canonical PR: the cluster spans unrelated model additions, documentation edits, and distinct bug fixes; the nearest pairs are not the same concrete change.", + "canonical_pr_reason": "PR 47090 is the clearest representative of the only true duplicate pair: it fixes the same CohereASR loss double-shift bug as 46958, on the same files and issue target, with the most explicit rationale.", "best_issue_reason": null, - "best_pr_reason": "No global best PR stands out as a duplicate representative; the items do not share one underlying bug/change.", + "best_pr_reason": "PR 47090 is the best overall PR candidate because it is the most explicit and complete version of the CohereASR double-shift fix among the near-duplicates.", "soft_edge_verdicts": [ { "left": "pull_request:46696", "right": "pull_request:47134", "accept": false, - "reason": "Same FlashAttention file and issue target, but one fixes tracing behavior and the other fixes unpadding performance; different bugs." + "reason": "Same flash-attention area and issue target, but different concrete regressions: tracing/int sync vs unpadding max-seqlen performance." }, { "left": "pull_request:46903", "right": "pull_request:47090", "accept": false, - "reason": "Both touch CohereASR loss, but 46903 also spans PPFormulaNet and the fixes are not the same concrete change." + "reason": "Shared training-loss pattern, but different model families and fixes: PPFormulaNet plus CohereASR versus CohereASR only." }, { "left": "pull_request:46903", "right": "pull_request:46958", "accept": false, - "reason": "Related loss-shift symptom, but 46903 is a broader multi-model PR and 46958 is a separate CohereASR-only implementation." + "reason": "Both touch CohereASR code, but 46903 is a broader multi-model change and not the same concrete bug/fix as 46958." }, { "left": "pull_request:47156", "right": "pull_request:47157", "accept": false, - "reason": "Different Japanese doc pages and different translation changes; not the same underlying change." + "reason": "Both are docs PRs, but they translate different pages in different locales and are unrelated changes." }, { "left": "pull_request:46954", "right": "pull_request:46958", "accept": false, - "reason": "Unrelated GLM ASR Pytest fix vs CohereASR training-loss fix." + "reason": "Pytest fp32-keep behavior change versus CohereASR loss-shift fix; no shared underlying bug." }, { "left": "pull_request:46958", "right": "pull_request:46974", "accept": false, - "reason": "Different models and bugs: CohereASR loss shift vs MusicGen generation typo." + "reason": "Unrelated fixes in different models: CohereASR training loss versus a MusicGen generate typo." }, { "left": "pull_request:47072", "right": "pull_request:47089", "accept": false, - "reason": "Chat-template schema parsing vs AqlmConfig validation messages are unrelated." + "reason": "Chat-template JSON schema parsing fix and AqlmConfig error-message wording are unrelated." }, { "left": "pull_request:47077", "right": "pull_request:47157", "accept": false, - "reason": "Different Japanese documentation updates; only superficial localization similarity." + "reason": "Different Japanese docs updates in different pages; not the same underlying change." }, { "left": "pull_request:47077", "right": "pull_request:47078", "accept": false, - "reason": "Both are Japanese doc edits, but they translate different pages and are not one change." + "reason": "Both are Japanese docs PRs, but one is a translation effort and the other is wording cleanup." }, { - "left": "pull_request:47072", - "right": "pull_request:47148", + "left": "pull_request:46266", + "right": "pull_request:46957", "accept": false, - "reason": "Different subsystems: chat template schema parsing vs auto-mapping registration typing." + "reason": "DINO conversion mapping work and FSDP2 migration are unrelated subsystems and changes." }, { - "left": "pull_request:46266", - "right": "pull_request:46957", + "left": "pull_request:47072", + "right": "pull_request:47134", "accept": false, - "reason": "Different model families and fixes; only broad infra similarity." + "reason": "Chat schema enum handling and flash-attention performance work are unrelated." }, { "left": "pull_request:47114", - "right": "pull_request:47148", + "right": "pull_request:47089", "accept": false, - "reason": "Docs about conversion mapping vs auto-factory typing; related area, but not the same fix." + "reason": "Docs on custom-code conversion mapping and AqlmConfig error messages do not fix the same bug or feature." }, { "left": "pull_request:46196", "right": "pull_request:46995", "accept": false, - "reason": "Different attention backend changes on different code paths and hardware support." + "reason": "Both broaden attention/runtime support, but they target different backends and code paths." }, { - "left": "pull_request:46266", - "right": "pull_request:47148", + "left": "pull_request:47135", + "right": "pull_request:47163", "accept": false, - "reason": "DINO conversion mapping work vs lazy auto-mapping registration typing; not mergeable as one PR." + "reason": "Docs quote-rendering cleanup and XPU paged-attention tests are unrelated." }, { "left": "pull_request:47093", "right": "pull_request:47114", "accept": false, - "reason": "OpenPangu model addition vs conversion-mapping docs; different scope and code path." + "reason": "OpenPangu-V2 model addition versus documentation about custom-code conversion mapping; unrelated." }, { "left": "pull_request:46707", "right": "pull_request:47108", "accept": false, - "reason": "Native FSDP2 migration vs doc-builder pin comment cleanup; unrelated." - }, - { - "left": "pull_request:47114", - "right": "pull_request:47149", - "accept": false, - "reason": "Conversion-mapping docs vs MoE decode switching docs; same broad docs area only." + "reason": "Native FSDP2 migration and a doc-builder comment tweak are completely unrelated." }, { "left": "pull_request:46707", - "right": "pull_request:47148", + "right": "pull_request:47114", "accept": false, - "reason": "FSDP2 migration and auto-factory typing are unrelated changes." + "reason": "FSDP2 migration versus docs on conversion mapping lookup rules; not the same change." }, { "left": "pull_request:47089", "right": "pull_request:47113", "accept": false, - "reason": "Quantization config error-message cleanup vs audio feature/processor requirements; unrelated." + "reason": "AqlmConfig message wording and audio feature/proc requirements fixes are unrelated." }, { - "left": "pull_request:47035", - "right": "pull_request:47089", + "left": "pull_request:47093", + "right": "pull_request:47161", "accept": false, - "reason": "Torch compile backend default vs AqlmConfig validation messages; different bug classes." + "reason": "Model addition versus exporter min-version enforcement; no shared underlying change." }, { - "left": "pull_request:47034", - "right": "pull_request:47108", + "left": "pull_request:46266", + "right": "pull_request:46707", "accept": false, - "reason": "Neuron training experiment vs doc-builder pin comment; unrelated." + "reason": "DINO modernization and FSDP2 integration are different features with different targets." }, { - "left": "pull_request:46895", - "right": "pull_request:47134", + "left": "pull_request:46995", + "right": "pull_request:47143", "accept": false, - "reason": "CohereASR loss fix vs flash-attention performance fix; different code paths." + "reason": "Flash-linear-attention backend support and NeuCodec model support are unrelated." }, { - "left": "pull_request:47100", - "right": "pull_request:47134", + "left": "pull_request:47135", + "right": "pull_request:47161", + "accept": false, + "reason": "Docs quote rendering and exporter version gating are unrelated." + }, + { + "left": "pull_request:47035", + "right": "pull_request:47161", "accept": false, - "reason": "Kernel repo error handling vs flash-attention performance regression; unrelated." + "reason": "Neuron compile backend defaulting and exporter version requirements are unrelated." }, { "left": "pull_request:46196", - "right": "pull_request:47148", + "right": "pull_request:47161", "accept": false, - "reason": "SDPA dispatch support vs auto-factory str-key support; different functionality." + "reason": "Attention backend dispatch support and exporter version enforcement do not fix the same thing." }, { - "left": "pull_request:43451", - "right": "pull_request:47078", + "left": "pull_request:47034", + "right": "pull_request:47035", "accept": false, - "reason": "Molmo2 model addition vs Japanese DistilBERT docs; unrelated." + "reason": "The draft Neuron experiment script and torch-compile backend default are adjacent but not the same concrete change." }, { - "left": "pull_request:47078", - "right": "pull_request:47096", + "left": "pull_request:47069", + "right": "pull_request:47089", + "accept": false, + "reason": "Num_labels None typing/behavior and AqlmConfig error-message wording are unrelated." + }, + { + "left": "pull_request:47069", + "right": "pull_request:47082", "accept": false, - "reason": "Japanese doc translation vs exportable Kimi code export changes; unrelated." + "reason": "Num_labels property typing fix and RoMa model addition are unrelated." }, { "left": "pull_request:47077", "right": "pull_request:47114", "accept": false, - "reason": "Japanese DistilBERT docs vs conversion-mapping docs; not the same change." + "reason": "Japanese docs translation and RoMa model registration are unrelated." }, { - "left": "pull_request:46658", - "right": "pull_request:47114", + "left": "pull_request:47078", + "right": "pull_request:47156", "accept": false, - "reason": "Step 3.7 Flash model work vs conversion-mapping docs; unrelated." + "reason": "Japanese wording cleanup and Korean Optimum translation/typo fix are unrelated docs updates." }, { - "left": "pull_request:47069", - "right": "pull_request:47136", + "left": "pull_request:47077", + "right": "pull_request:47156", "accept": false, - "reason": "Config property typing change vs quicktour wording cleanup; unrelated." + "reason": "Different locales and different pages; not the same underlying documentation change." }, { - "left": "pull_request:46707", + "left": "pull_request:47089", + "right": "pull_request:47134", + "accept": false, + "reason": "AqlmConfig validation messages and flash-attention performance regression are unrelated." + }, + { + "left": "pull_request:46912", "right": "pull_request:47082", "accept": false, - "reason": "FSDP2 migration vs RoMa model addition; unrelated." + "reason": "RoPE test expansion and RoMa model addition are unrelated." }, { "left": "pull_request:46912", "right": "pull_request:47078", "accept": false, - "reason": "RoPE test/model changes vs Japanese docs; unrelated." + "reason": "RoPE tests and Japanese docs wording cleanup are unrelated." }, { - "left": "pull_request:46912", + "left": "pull_request:47134", + "right": "pull_request:47161", + "accept": false, + "reason": "Flash-attention unpadding performance fix and exporter version requirement changes are unrelated." + }, + { + "left": "pull_request:43451", "right": "pull_request:47082", "accept": false, - "reason": "RoPE test/model changes vs RoMa model addition; unrelated." + "reason": "Molmo2 model addition and RoMa keypoint matcher addition are unrelated model-introduction PRs." }, { - "left": "pull_request:47082", - "right": "pull_request:47148", + "left": "pull_request:46895", + "right": "pull_request:47134", "accept": false, - "reason": "RoMa model addition vs auto-factory string-key typing; unrelated." + "reason": "CohereASR loss-shift fix and flash-attention performance regression are unrelated bugs." }, { - "left": "pull_request:47069", + "left": "pull_request:47113", + "right": "pull_request:47161", + "accept": false, + "reason": "Audio requirements cleanup and exporter minimum-version enforcement are unrelated." + }, + { + "left": "pull_request:47093", + "right": "pull_request:47143", + "accept": false, + "reason": "OpenPangu-V2 model addition and NeuCodec support are different model-introduction PRs." + }, + { + "left": "pull_request:47093", + "right": "pull_request:47163", + "accept": false, + "reason": "OpenPangu-V2 model addition and XPU paged-attention tests are unrelated." + }, + { + "left": "pull_request:47134", + "right": "pull_request:47135", + "accept": false, + "reason": "Flash-attention performance fix and docs quote rendering cleanup are unrelated." + }, + { + "left": "pull_request:43451", + "right": "pull_request:46995", + "accept": false, + "reason": "Molmo2 model addition and flash-linear-attention backend support are unrelated." + }, + { + "left": "pull_request:47100", + "right": "pull_request:47143", + "accept": false, + "reason": "Hub-kernels API cleanup and NeuCodec support are unrelated." + }, + { + "left": "pull_request:46912", "right": "pull_request:47113", "accept": false, - "reason": "Configuration typing vs audio requirement decorators; unrelated." + "reason": "RoPE test coverage and audio requirements cleanup are unrelated." }, { "left": "pull_request:46912", - "right": "pull_request:47136", + "right": "pull_request:47142", "accept": false, - "reason": "RoPE changes vs quicktour wording cleanup; unrelated." + "reason": "RoPE tests and pipelines generator cleanup are unrelated." }, { - "left": "pull_request:47069", - "right": "pull_request:47082", + "left": "pull_request:47156", + "right": "pull_request:47166", "accept": false, - "reason": "Configuration property typing vs model addition; unrelated." + "reason": "Korean docs translation and quantization test updates are unrelated." }, { - "left": "pull_request:46196", + "left": "pull_request:46266", + "right": "pull_request:47034", + "accept": false, + "reason": "DINO conversion modernization and a Neuron experiment script are unrelated." + }, + { + "left": "pull_request:47034", "right": "pull_request:47108", "accept": false, - "reason": "SDPA dispatch support vs doc-builder pin comment cleanup; unrelated." + "reason": "Neuron experiment code and doc-builder comment tweak are unrelated." }, { - "left": "pull_request:47089", - "right": "pull_request:47134", + "left": "pull_request:47035", + "right": "pull_request:47113", "accept": false, - "reason": "AqlmConfig message wording vs flash-attention regression fix; unrelated." + "reason": "Torch-compile backend defaulting and audio requirements cleanup are unrelated." }, { - "left": "pull_request:46707", - "right": "pull_request:46957", + "left": "pull_request:47035", + "right": "pull_request:47069", "accept": false, - "reason": "FSDP2 migration vs finegrained FP8 decode optimization; unrelated." + "reason": "Neuron compile backend behavior and num_labels typing are unrelated." }, { - "left": "pull_request:47025", - "right": "pull_request:47149", + "left": "pull_request:47035", + "right": "pull_request:47082", "accept": false, - "reason": "Diffusion Gemma test cleanup vs MoE decode-switch docs; unrelated." + "reason": "Neuron compile backend defaulting and RoMa model addition are unrelated." }, { - "left": "pull_request:47142", - "right": "pull_request:47149", + "left": "pull_request:43451", + "right": "pull_request:46266", "accept": false, - "reason": "Pipeline generator cleanup vs MoE decode-switch docs; unrelated." + "reason": "Molmo2 model addition and DINO modern-standards work are unrelated." }, { - "left": "pull_request:47025", - "right": "pull_request:47114", + "left": "pull_request:47100", + "right": "pull_request:47142", + "accept": false, + "reason": "Hub-kernels error-path cleanup and pipeline generator handling are unrelated." + }, + { + "left": "pull_request:47069", + "right": "pull_request:47161", + "accept": false, + "reason": "Num_labels None behavior and exporter version requirements are unrelated." + }, + { + "left": "pull_request:47082", + "right": "pull_request:47108", + "accept": false, + "reason": "RoMa model addition and doc-builder comment tweak are unrelated." + }, + { + "left": "pull_request:44517", + "right": "pull_request:47143", + "accept": false, + "reason": "Qwen3-TTS model support and NeuCodec support are unrelated model additions." + }, + { + "left": "pull_request:44517", + "right": "pull_request:47082", + "accept": false, + "reason": "Qwen3-TTS model support and RoMa keypoint matcher addition are unrelated." + }, + { + "left": "pull_request:46957", + "right": "pull_request:47108", "accept": false, - "reason": "Diffusion Gemma tests vs conversion-mapping docs; unrelated." + "reason": "Finegrained FP8 generation changes and a doc-builder comment tweak are unrelated." }, { - "left": "pull_request:47069", - "right": "pull_request:47082", + "left": "pull_request:46995", + "right": "pull_request:47143", "accept": false, - "reason": "Config typing fix vs RoMa model addition; unrelated." + "reason": "Flash-linear-attention support and NeuCodec support are unrelated." }, { - "left": "pull_request:44517", - "right": "pull_request:47082", + "left": "pull_request:47093", + "right": "pull_request:47114", "accept": false, - "reason": "Qwen3-TTS model addition vs RoMa keypoint matcher; unrelated model additions." + "reason": "OpenPangu-V2 model support and docs about conversion lookup rules are unrelated." }, { - "left": "pull_request:47034", - "right": "pull_request:47035", - "accept": false, - "reason": "Neuron finetuning experiment vs torch-compile backend default; related platform, but different changes." + "left": "pull_request:47090", + "right": "pull_request:46958", + "accept": true, + "reason": "Same CohereASR training-loss double-shift bug, same files, same issue target, and the fixes are direct variants of the same change." } ] }, @@ -22612,7 +22777,7 @@ "source": "llm", "cache_hit": false, "model": "gpt-5.4-mini?service_tier=flex", - "cluster_id": "cluster-40546-46", + "cluster_id": "cluster-40546-42", "nodes": [ "pull_request:40546", "pull_request:43451", @@ -22621,470 +22786,402 @@ "pull_request:46658", "pull_request:46707", "pull_request:46836", - "pull_request:46880", "pull_request:46895", "pull_request:46912", "pull_request:46957", "pull_request:46958", "pull_request:47003", - "pull_request:47025", "pull_request:47034", "pull_request:47035", + "pull_request:47069", "pull_request:47077", "pull_request:47090", - "pull_request:47092", "pull_request:47096", - "pull_request:47100", - "pull_request:47113", - "pull_request:47134", - "pull_request:47135", - "pull_request:47136", - "pull_request:47137", "pull_request:47142", "pull_request:47143", - "pull_request:47148", - "pull_request:47149", - "pull_request:47156" + "pull_request:47156", + "pull_request:47161", + "pull_request:47163", + "pull_request:47166" ], "soft_pairs": [ - "pull_request:43451|pull_request:47136", - "pull_request:47096|pull_request:47136", - "pull_request:47135|pull_request:47137", - "pull_request:47100|pull_request:47143", - "pull_request:46912|pull_request:47113", - "pull_request:43451|pull_request:44517", - "pull_request:43451|pull_request:47135", - "pull_request:46912|pull_request:47142", - "pull_request:44517|pull_request:47143", "pull_request:44517|pull_request:46836", - "pull_request:44517|pull_request:47096", - "pull_request:47137|pull_request:47156", + "pull_request:46957|pull_request:47161", "pull_request:43451|pull_request:46957", "pull_request:46957|pull_request:47096", - "pull_request:47148|pull_request:47149", - "pull_request:46912|pull_request:47077", - "pull_request:47134|pull_request:47137", - "pull_request:46957|pull_request:47003", "pull_request:46658|pull_request:46895", + "pull_request:46912|pull_request:47077", "pull_request:43451|pull_request:47142", "pull_request:47096|pull_request:47142", - "pull_request:46880|pull_request:47092", + "pull_request:43451|pull_request:47069", + "pull_request:43451|pull_request:46895", + "pull_request:47003|pull_request:47142", "pull_request:46895|pull_request:47090", "pull_request:46958|pull_request:47090", "pull_request:46895|pull_request:46958", "pull_request:47077|pull_request:47156", - "pull_request:47025|pull_request:47092", + "pull_request:47163|pull_request:47166", "pull_request:46707|pull_request:47034", "pull_request:46658|pull_request:46836", "pull_request:43451|pull_request:46836", "pull_request:46658|pull_request:47143", "pull_request:46836|pull_request:47143", + "pull_request:46836|pull_request:47163", "pull_request:46196|pull_request:47143", - "pull_request:46836|pull_request:47092", - "pull_request:46836|pull_request:47096", + "pull_request:47143|pull_request:47163", + "pull_request:46196|pull_request:47163", + "pull_request:46836|pull_request:47161", + "pull_request:43451|pull_request:47077", + "pull_request:43451|pull_request:47161", "pull_request:43451|pull_request:46196", - "pull_request:46196|pull_request:47096", "pull_request:40546|pull_request:43451", + "pull_request:47143|pull_request:47161", + "pull_request:47161|pull_request:47163", "pull_request:43451|pull_request:46658", "pull_request:46658|pull_request:47096", "pull_request:43451|pull_request:47143", - "pull_request:43451|pull_request:47092", - "pull_request:47092|pull_request:47143", + "pull_request:43451|pull_request:47163", + "pull_request:46912|pull_request:47163", "pull_request:43451|pull_request:47096", "pull_request:43451|pull_request:47003", - "pull_request:43451|pull_request:47137", "pull_request:47003|pull_request:47096", - "pull_request:47096|pull_request:47137", - "pull_request:47096|pull_request:47143", "pull_request:43451|pull_request:47034", "pull_request:47034|pull_request:47096", - "pull_request:47092|pull_request:47096", + "pull_request:43451|pull_request:47166", + "pull_request:47096|pull_request:47166", "pull_request:43451|pull_request:47035", - "pull_request:47035|pull_request:47096", - "pull_request:43451|pull_request:46912", - "pull_request:47003|pull_request:47137", "pull_request:47003|pull_request:47034", + "pull_request:47003|pull_request:47166", + "pull_request:47096|pull_request:47143", "pull_request:47003|pull_request:47143", - "pull_request:47137|pull_request:47143", - "pull_request:46836|pull_request:47003", - "pull_request:46836|pull_request:47137" + "pull_request:47096|pull_request:47163", + "pull_request:47003|pull_request:47163", + "pull_request:47143|pull_request:47166", + "pull_request:46836|pull_request:47096", + "pull_request:46836|pull_request:47003" ], - "prepared_review_unit_hash": "d222be9d90fc16dc68cfd9a1e74f314fdc34016aaf27af0560b09f1e25860dc4", + "prepared_review_unit_hash": "7fc23bb3cc9747df1bc11de291a2355edbfaeeaf90e2dac86838377435f4fcf9", "component_budget": { - "node_count": 46, - "item_count": 46, - "soft_pair_count": 133, - "serialized_chars": 159768, - "estimated_input_tokens": 39942, - "estimated_eval_tokens": 80140 + "node_count": 42, + "item_count": 42, + "soft_pair_count": 126, + "serialized_chars": 150672, + "estimated_input_tokens": 37668, + "estimated_eval_tokens": 75592 }, "budget": { - "node_count": 31, - "item_count": 31, - "soft_pair_count": 61, - "serialized_chars": 94054, - "estimated_input_tokens": 23514, - "estimated_eval_tokens": 47284 + "node_count": 24, + "item_count": 24, + "soft_pair_count": 54, + "serialized_chars": 77931, + "estimated_input_tokens": 19483, + "estimated_eval_tokens": 39222 }, "overflow_policy": "truncate_then_skip", "trimmed": false, "aggressively_trimmed": false, "split": true, "analyst_result": { - "summary": "This cluster is mostly unrelated PRs spanning new model adds, docs, exporter tweaks, and test-only fixes. The only true duplicate set is the CohereASR training-loss double-shift fix, where three PRs target the same bug and code path.", + "summary": "This set is mostly unrelated PRs spanning new model additions, docs, exporters, tests, and infra. The only clear duplicate cluster is the CohereASR training-loss double-shift fixes; the rest share broad subsystems at most, not the same concrete change.", "confidence": 0.95, "canonical_issue_reason": null, - "canonical_pr_reason": "PR 47090 is the clearest representative of the real duplicate cluster: it directly targets the CohereASR double-shift loss bug, has the most explicit explanation, and links the shared issue.", + "canonical_pr_reason": "PR #47090 is the best canonical representative of the CohereASR duplicate cluster: it targets the same training-loss double-shift bug, explains the fix clearly, and uses the most complete loss-function call path.", "best_issue_reason": null, - "best_pr_reason": "PR 47090 is the best global representative because it is the most complete and clearly scoped fix among the only true duplicate group in this set.", + "best_pr_reason": "PR #47090 is the strongest overall representative among the duplicate candidates because it fixes the same concrete bug with the clearest rationale and preserves the intended loss-reduction behavior.", "soft_edge_verdicts": [ - { - "left": "pull_request:43451", - "right": "pull_request:47136", - "accept": false, - "reason": "Different scopes: new model docs/registration vs quicktour wording." - }, - { - "left": "pull_request:47096", - "right": "pull_request:47136", - "accept": false, - "reason": "Exportability changes vs inference docs clarification; unrelated." - }, - { - "left": "pull_request:47135", - "right": "pull_request:47137", - "accept": false, - "reason": "Different fixes: attention docs vs tokenizer type annotation/test." - }, - { - "left": "pull_request:47100", - "right": "pull_request:47143", - "accept": false, - "reason": "Kernel API fix vs adding a new model; unrelated code paths." - }, - { - "left": "pull_request:46912", - "right": "pull_request:47113", - "accept": false, - "reason": "RoPE behavior/tests vs audio feature/processor requirements; unrelated." - }, - { - "left": "pull_request:43451", - "right": "pull_request:44517", - "accept": false, - "reason": "Different model additions (VibeVoice vs Qwen3-TTS)." - }, - { - "left": "pull_request:43451", - "right": "pull_request:47135", - "accept": false, - "reason": "New model docs vs attention docs formatting; unrelated." - }, - { - "left": "pull_request:46912", - "right": "pull_request:47142", - "accept": false, - "reason": "RoPE tests vs pipeline generator handling; unrelated." - }, - { - "left": "pull_request:44517", - "right": "pull_request:47143", - "accept": false, - "reason": "Different model additions; not the same change." - }, { "left": "pull_request:44517", "right": "pull_request:46836", "accept": false, - "reason": "New model additions, but different models and code paths." + "reason": "Different model additions (VibeVoice vs Unlimited OCR); same docs/auto-registration patterns, but not the same change." }, { - "left": "pull_request:44517", - "right": "pull_request:47096", - "accept": false, - "reason": "Qwen3-TTS addition vs Kimi exportability work; unrelated." - }, - { - "left": "pull_request:47137", - "right": "pull_request:47156", + "left": "pull_request:46957", + "right": "pull_request:47161", "accept": false, - "reason": "Type annotation/tokenizer test vs translation/docs typo fix; unrelated." + "reason": "Both touch exporter-related code, but one is a finegrained FP8 kernel change and the other is exporter version gating; not the same bug or feature." }, { "left": "pull_request:43451", "right": "pull_request:46957", "accept": false, - "reason": "VibeVoice model add vs MoE decode-switching optimization; unrelated." + "reason": "Molmo2 model onboarding vs FP8/generation/export kernel work; no shared underlying change." }, { "left": "pull_request:46957", "right": "pull_request:47096", "accept": false, - "reason": "Different subsystems: generation optimization vs model exportability." + "reason": "OpenVINO exporter work and Kimi exportability are separate export paths with different code changes." }, { - "left": "pull_request:47148", - "right": "pull_request:47149", + "left": "pull_request:46658", + "right": "pull_request:46895", "accept": false, - "reason": "Auto factory key typing vs MoE docs; unrelated." + "reason": "Step-3.7 Flash model support and CohereASR loss fixing are unrelated." }, { "left": "pull_request:46912", "right": "pull_request:47077", "accept": false, - "reason": "Model performance/test changes vs translation docs; unrelated." + "reason": "RoPE test coverage work vs Japanese DistilBERT documentation translation; unrelated." }, { - "left": "pull_request:47134", - "right": "pull_request:47137", + "left": "pull_request:43451", + "right": "pull_request:47142", "accept": false, - "reason": "Flash-attention performance fix vs tokenizer method typing/test; unrelated." + "reason": "Molmo2 model addition vs pipeline generator cleanup; no duplicate change." }, { - "left": "pull_request:46957", - "right": "pull_request:47003", + "left": "pull_request:47096", + "right": "pull_request:47142", "accept": false, - "reason": "Different feature areas: decode switching vs OpenVINO exporter." + "reason": "Kimi export support vs pipeline generator handling; different code paths." }, { - "left": "pull_request:46658", - "right": "pull_request:46895", + "left": "pull_request:43451", + "right": "pull_request:47069", "accept": false, - "reason": "Step-3.7 Flash model bring-up vs CohereASR training-loss bug; unrelated." + "reason": "Model onboarding vs a small configuration property typing fix; unrelated." }, { "left": "pull_request:43451", - "right": "pull_request:47142", + "right": "pull_request:46895", "accept": false, - "reason": "New model docs vs pipeline generator cleanup; unrelated." + "reason": "Molmo2 addition vs CohereASR training-loss fix; unrelated." }, { - "left": "pull_request:47096", + "left": "pull_request:47003", "right": "pull_request:47142", "accept": false, - "reason": "Kimi export support vs pipeline generator cleanup; unrelated." - }, - { - "left": "pull_request:46880", - "right": "pull_request:47092", - "accept": false, - "reason": "Different test-only XPU expectations for different models." + "reason": "OpenVINO exporter addition and pipeline generator cleanup are distinct changes." }, { "left": "pull_request:46895", "right": "pull_request:47090", "accept": true, - "reason": "Same CohereASR double-shift loss bug, same files, same issue target." + "reason": "Same CohereASR double-shift bug in the same files; both fixes adjust the training loss path for the same concrete issue." }, { "left": "pull_request:46958", "right": "pull_request:47090", "accept": true, - "reason": "Same CohereASR training-loss bug and same code path; one is a variant of the same fix." + "reason": "Same CohereASR loss-shift bug and same code path; different implementation style, but same underlying fix." }, { "left": "pull_request:46895", "right": "pull_request:46958", "accept": true, - "reason": "Both fix the same CohereASR double-shift loss problem in the same files." + "reason": "Both are alternative fixes for the same CohereASR training-loss double-shift problem." }, { "left": "pull_request:47077", "right": "pull_request:47156", "accept": false, - "reason": "Japanese docs translation vs Optimum typo/i18n update; unrelated." + "reason": "Japanese DistilBERT docs translation vs a Furiosa typo/doc tweak in Optimum; unrelated." }, { - "left": "pull_request:47025", - "right": "pull_request:47092", + "left": "pull_request:47163", + "right": "pull_request:47166", "accept": false, - "reason": "Different test fixes for different models." + "reason": "Paged-attention XPU test additions vs quantization test extensions; both are tests, but not the same bug/change." }, { "left": "pull_request:46707", "right": "pull_request:47034", "accept": false, - "reason": "Native FSDP2 integration vs Neuron training experiment; unrelated." + "reason": "Native FSDP2 integration vs Neuron SFT experiment scripts; separate infrastructure work." }, { "left": "pull_request:46658", "right": "pull_request:46836", "accept": false, - "reason": "Different new model/code additions; not the same change." + "reason": "Step-3.7 Flash model support vs Unlimited OCR model support; different models and code paths." }, { "left": "pull_request:43451", "right": "pull_request:46836", "accept": false, - "reason": "Different model additions and registrations; unrelated." + "reason": "Molmo2 and Unlimited OCR are separate model additions, despite similar registration boilerplate." }, { "left": "pull_request:46658", "right": "pull_request:47143", "accept": false, - "reason": "Step-3.7 Flash model work vs NeuCodec support; unrelated." + "reason": "Step-3.7 Flash model support vs NeuCodec model support; unrelated model additions." }, { - "left": "pull_request:46196", + "left": "pull_request:46836", "right": "pull_request:47143", "accept": false, - "reason": "SDPA dispatch support vs new NeuCodec model addition; unrelated." + "reason": "Unlimited OCR and NeuCodec are distinct model onboarding PRs." }, { - "left": "pull_request:40546", - "right": "pull_request:43451", + "left": "pull_request:46836", + "right": "pull_request:47163", "accept": false, - "reason": "VibeVoice and Molmo2 are different model additions." + "reason": "Unlimited OCR model addition vs paged-attention test changes; unrelated." }, { - "left": "pull_request:43451", - "right": "pull_request:46658", + "left": "pull_request:46196", + "right": "pull_request:47143", "accept": false, - "reason": "Different model additions; no shared bug/change." + "reason": "SDPA dispatch support for wav2vec2-family models vs NeuCodec model addition; different changes." }, { - "left": "pull_request:46658", - "right": "pull_request:47096", + "left": "pull_request:47143", + "right": "pull_request:47163", "accept": false, - "reason": "Step-3.7 Flash model bring-up vs Kimi exportability; unrelated." + "reason": "NeuCodec model addition vs paged-attention XPU tests; unrelated." }, { "left": "pull_request:46196", - "right": "pull_request:47143", + "right": "pull_request:47163", "accept": false, - "reason": "Attention SDPA support vs NeuCodec model support; unrelated." + "reason": "Audio SDPA dispatch support vs paged-attention test additions; not the same underlying bug." }, { "left": "pull_request:46836", - "right": "pull_request:47092", + "right": "pull_request:47161", "accept": false, - "reason": "Unlimited OCR model add vs XPU test expectations for another model." + "reason": "Unlimited OCR model addition vs exporter minimum-version checks; unrelated." }, { - "left": "pull_request:46836", - "right": "pull_request:47096", + "left": "pull_request:43451", + "right": "pull_request:47077", "accept": false, - "reason": "Unlimited OCR model add vs exportability work for Kimi; unrelated." + "reason": "Molmo2 model onboarding vs Japanese translation docs; unrelated." }, { "left": "pull_request:43451", - "right": "pull_request:47143", + "right": "pull_request:47161", "accept": false, - "reason": "Different model additions (Molmo2 vs NeuCodec)." + "reason": "Molmo2 model onboarding vs exporter version requirements; unrelated." }, { - "left": "pull_request:47092", - "right": "pull_request:47143", + "left": "pull_request:43451", + "right": "pull_request:46196", "accept": false, - "reason": "Different test/model additions; unrelated." + "reason": "Molmo2 model support vs SDPA dispatch support for audio models; unrelated." }, { - "left": "pull_request:43451", - "right": "pull_request:47096", + "left": "pull_request:40546", + "right": "pull_request:43451", "accept": false, - "reason": "Molmo2 model add vs Kimi exportability; unrelated." + "reason": "VibeVoice and Molmo2 are distinct model additions with different files and implementations." }, { - "left": "pull_request:43451", - "right": "pull_request:47003", + "left": "pull_request:47143", + "right": "pull_request:47161", "accept": false, - "reason": "New model addition vs OpenVINO exporter; unrelated." + "reason": "NeuCodec onboarding and exporter version enforcement are unrelated." }, { - "left": "pull_request:43451", - "right": "pull_request:47137", + "left": "pull_request:47161", + "right": "pull_request:47163", "accept": false, - "reason": "New model docs vs tokenizer method typing/test; unrelated." + "reason": "Exporter dependency checks vs paged-attention test coverage; different concerns." }, { - "left": "pull_request:47003", - "right": "pull_request:47096", + "left": "pull_request:43451", + "right": "pull_request:46658", "accept": false, - "reason": "OpenVINO exporter addition vs Kimi exportability; different changes." + "reason": "Molmo2 and Step-3.7 Flash are separate model additions." }, { - "left": "pull_request:47096", - "right": "pull_request:47137", + "left": "pull_request:46658", + "right": "pull_request:47096", "accept": false, - "reason": "Exporter support vs tokenizer typing/test; unrelated." + "reason": "Step-3.7 Flash model support vs Kimi exportability changes; unrelated." }, { - "left": "pull_request:47096", + "left": "pull_request:43451", "right": "pull_request:47143", "accept": false, - "reason": "Kimi exportability vs NeuCodec model support; unrelated." + "reason": "Molmo2 and NeuCodec are different model onboarding PRs." }, { "left": "pull_request:43451", - "right": "pull_request:47034", + "right": "pull_request:47096", "accept": false, - "reason": "New model docs vs Neuron training experiment; unrelated." + "reason": "Molmo2 model addition vs exportable Kimi work; unrelated." }, { - "left": "pull_request:47034", - "right": "pull_request:47096", + "left": "pull_request:43451", + "right": "pull_request:47003", "accept": false, - "reason": "Neuron training script vs exportability work; unrelated." + "reason": "Molmo2 model support vs OpenVINO exporter support; different changes." }, { - "left": "pull_request:47092", + "left": "pull_request:47003", "right": "pull_request:47096", "accept": false, - "reason": "Different model-specific test expectations; unrelated." + "reason": "OpenVINO exporter introduction and Kimi export support are different exporter features." }, { "left": "pull_request:43451", - "right": "pull_request:47035", + "right": "pull_request:47034", "accept": false, - "reason": "New model add vs torch-compile backend default change; unrelated." + "reason": "Molmo2 model addition vs Neuron training experiment; unrelated." }, { - "left": "pull_request:47035", + "left": "pull_request:47034", "right": "pull_request:47096", "accept": false, - "reason": "Neuron compile-default tweak vs exportability work; unrelated." + "reason": "Neuron training experiment vs Kimi export support; different code paths." }, { - "left": "pull_request:43451", - "right": "pull_request:46912", + "left": "pull_request:47003", + "right": "pull_request:47034", "accept": false, - "reason": "New model add vs RoPE testing changes; unrelated." + "reason": "OpenVINO exporter support and Neuron experiment scripts are unrelated." }, { "left": "pull_request:47003", - "right": "pull_request:47137", + "right": "pull_request:47166", "accept": false, - "reason": "OpenVINO exporter vs tokenizer typing/test; unrelated." + "reason": "OpenVINO exporter work vs quantization test updates; not the same change." }, { - "left": "pull_request:47003", - "right": "pull_request:47034", + "left": "pull_request:47096", + "right": "pull_request:47143", "accept": false, - "reason": "OpenVINO exporter vs Neuron training experiment; unrelated." + "reason": "Kimi export support vs NeuCodec model onboarding; unrelated." }, { "left": "pull_request:47003", "right": "pull_request:47143", "accept": false, - "reason": "Exporter work vs new NeuCodec model support; unrelated." + "reason": "OpenVINO exporter addition vs NeuCodec model addition; unrelated." }, { - "left": "pull_request:47137", - "right": "pull_request:47143", + "left": "pull_request:47096", + "right": "pull_request:47163", + "accept": false, + "reason": "Kimi export support vs paged-attention XPU tests; different concerns." + }, + { + "left": "pull_request:47003", + "right": "pull_request:47163", + "accept": false, + "reason": "OpenVINO exporter support vs paged-attention tests; unrelated." + }, + { + "left": "pull_request:47143", + "right": "pull_request:47166", "accept": false, - "reason": "Tokenizer utility/test fix vs NeuCodec model addition; unrelated." + "reason": "NeuCodec model addition vs quantization test changes; unrelated." }, { "left": "pull_request:46836", - "right": "pull_request:47003", + "right": "pull_request:47096", "accept": false, - "reason": "Unlimited OCR model support vs OpenVINO exporter; unrelated." + "reason": "Unlimited OCR model onboarding and exportable Kimi are unrelated." }, { "left": "pull_request:46836", - "right": "pull_request:47137", + "right": "pull_request:47003", "accept": false, - "reason": "Unlimited OCR model support vs tokenizer utility/test fix; unrelated." + "reason": "Unlimited OCR model onboarding vs OpenVINO exporter introduction; unrelated." } ] }, @@ -23114,12 +23211,12 @@ "soft_pairs": [ "pull_request:40755|pull_request:41224" ], - "prepared_review_unit_hash": "b0f80f7168841ba89148fff1c21898b4caa4b5754389217c62f6ab4f97a26e9b", + "prepared_review_unit_hash": "a284edac83525c4dcd397d77f0edfa36defa5d3b9e21cf13944f9945ae49386f", "component_budget": { "node_count": 2, "item_count": 2, "soft_pair_count": 1, - "serialized_chars": 4676, + "serialized_chars": 4675, "estimated_input_tokens": 1169, "estimated_eval_tokens": 2594 }, @@ -23127,7 +23224,7 @@ "node_count": 2, "item_count": 2, "soft_pair_count": 1, - "serialized_chars": 4676, + "serialized_chars": 4675, "estimated_input_tokens": 1169, "estimated_eval_tokens": 2594 }, @@ -23136,18 +23233,18 @@ "aggressively_trimmed": false, "split": false, "analyst_result": { - "summary": "The two pull requests are unrelated: one adds TimesFM covariate forecasting support, the other adds DINOv3ViT image classification support. They do not target the same code path or underlying change, so this is not a duplicate cluster.", - "confidence": 0.98, + "summary": "These two PRs are unrelated feature additions in different model families, so they are not duplicates.", + "confidence": 0.99, "canonical_issue_reason": null, - "canonical_pr_reason": null, + "canonical_pr_reason": "No canonical PR: 40755 adds TimesFM covariate forecasting, while 41224 adds DINOv3 image-classification support; they touch different code paths and solve different problems.", "best_issue_reason": null, - "best_pr_reason": "Neither PR is a duplicate of the other; each addresses a different model and feature area, so there is no meaningful canonical PR in this cluster.", + "best_pr_reason": "Neither PR is a good global representative for a duplicate cluster because the changes are unrelated; if forced, 41224 is the more self-contained feature PR, but it does not subsume 40755.", "soft_edge_verdicts": [ { "left": "pull_request:40755", "right": "pull_request:41224", "accept": false, - "reason": "Different models, different tasks, and different files changed; no shared bug or mergeable single fix." + "reason": "Different underlying changes: TimesFM forecasting with covariates versus DINOv3 classification support. They modify different model implementations, docs, and tests, and could not plausibly be merged into one PR." } ] }, @@ -23266,12 +23363,12 @@ "pull_request:43823|pull_request:45213", "pull_request:43251|pull_request:43254" ], - "prepared_review_unit_hash": "64ed6e5c009319219a18822e19bd69d9dddde510a8394720db20c201d0515093", + "prepared_review_unit_hash": "21bd118a02a57abbc59ec784caa415e04b6b3b0ed81f6bb5474bcce55dfc2b55", "component_budget": { "node_count": 12, "item_count": 12, "soft_pair_count": 17, - "serialized_chars": 31599, + "serialized_chars": 31598, "estimated_input_tokens": 7900, "estimated_eval_tokens": 16056 }, @@ -23279,7 +23376,7 @@ "node_count": 12, "item_count": 12, "soft_pair_count": 17, - "serialized_chars": 31599, + "serialized_chars": 31598, "estimated_input_tokens": 7900, "estimated_eval_tokens": 16056 }, @@ -23288,126 +23385,120 @@ "aggressively_trimmed": false, "split": false, "analyst_result": { - "summary": "The cluster is mostly unrelated PRs; the only true duplicate/same-change pair is the two `fixed_cross_entropy` patches. Everything else targets different models, utilities, or test/doc work and should stay separate.", - "confidence": 0.78, + "summary": "This cluster is mostly a false-positive mix of unrelated PRs spanning tokenizer methods, processor/model registration, export tests, new model additions, refactors, and kernel/loss plumbing. The only clear duplicate pair is the two `fixed_cross_entropy` PRs, which both extend `loss_utils.py` to pass supported `cross_entropy` kwargs for the same issue.", + "confidence": 0.93, "canonical_issue_reason": null, - "canonical_pr_reason": "PR 43251 is the best canonical representative of the only real duplicate pair: it directly targets issue 43240 and makes the `cross_entropy` passthrough fix explicitly, while 43254 is just a narrower rewording of the same change.", + "canonical_pr_reason": "PR 43251 is the best canonical PR within the only true duplicate pair: it directly targets issue 43240, adds the missing `cross_entropy` kwargs in `fixed_cross_entropy`, and is the more complete/explicit fix compared with the narrower alternate implementation in 43254.", "best_issue_reason": null, - "best_pr_reason": "PR 43251 is the best global representative because it is the only clearly duplicate-worthy code change in the set and has the cleanest issue linkage to the underlying bug in `fixed_cross_entropy`.", + "best_pr_reason": "PR 43251 is the strongest representative of the cluster\u2019s only real duplicate subcluster because it has the explicit issue link, broader fix semantics, and clear alignment with the bug report; the other PRs are unrelated or too broad to serve as a canonical duplicate.", "soft_edge_verdicts": [ { "left": "pull_request:41855", "right": "pull_request:43424", "accept": false, - "reason": "Tokenizer method additions for Mistral are unrelated to an Executorch dynamic-shape export test." + "reason": "Different problems: Mistral tokenizer method additions vs ExecuTorch dynamic-shape export testing. No shared code-path bug." }, { "left": "pull_request:41855", "right": "pull_request:43823", "accept": false, - "reason": "Different model families and different changes; no shared bug or code path." + "reason": "Unrelated: tokenizer API gaps vs adding a new MobileLLM model package and weights. Not the same underlying change." }, { "left": "pull_request:42668", "right": "pull_request:43823", "accept": false, - "reason": "Processor auto-loading robustness is unrelated to adding MobileLLM support." + "reason": "Both touch processor/model auto-registration areas, but one is about robust `from_pretrained` handling while the other adds a new model. Different changes." }, { "left": "pull_request:43251", "right": "pull_request:43823", "accept": false, - "reason": "Cross-entropy kwargs plumbing is unrelated to a new model addition." + "reason": "`fixed_cross_entropy` kwargs handling is unrelated to adding MobileLLM model files and configs." }, { "left": "pull_request:43251", "right": "pull_request:44827", "accept": false, - "reason": "Different loss helper fix versus Mistral4 test/model changes." + "reason": "Loss-function kwargs support is unrelated to Mistral4 test/model fixes." }, { "left": "pull_request:43424", "right": "pull_request:44827", "accept": false, - "reason": "Executorch export tests and Mistral4 test fixes are different problems." + "reason": "ExecuTorch dynamic-shape export tests are unrelated to Mistral4 model/test fixes." }, { "left": "pull_request:43424", "right": "pull_request:43823", "accept": false, - "reason": "Dynamic export testing is unrelated to MobileLLM model scaffolding." + "reason": "Different subsystems and different bugs: exportability tests vs new model addition." }, { "left": "pull_request:43823", "right": "pull_request:45655", "accept": false, - "reason": "New model support is unrelated to config-resolution ordering in the base modeling utils." + "reason": "New MobileLLM model addition is unrelated to `cls.config` resolution order in `modeling_utils.py`." }, { "left": "pull_request:43823", "right": "pull_request:43995", "accept": false, - "reason": "Different subsystems and different concrete changes; not the same bug." - }, - { - "left": "pull_request:43823", - "right": "pull_request:45213", - "accept": false, - "reason": "A new model PR is unrelated to a meta skill/doc scaffold PR." + "reason": "New model addition vs Falcon output-collection refactor; not the same code-path issue." }, { "left": "pull_request:43995", "right": "pull_request:45213", "accept": false, - "reason": "Falcon output-collection refactor is unrelated to model-creation skill files." + "reason": "Falcon model refactoring is unrelated to the model-creation skill/doc scaffold PR." }, { "left": "pull_request:43995", "right": "pull_request:44827", "accept": false, - "reason": "Falcon refactor and Mistral4 test fixes are distinct model-specific changes." + "reason": "Falcon standardized output interface refactor is unrelated to Mistral4 test fixes." }, { "left": "pull_request:44827", "right": "pull_request:45170", "accept": false, - "reason": "Mistral4 test stabilization is unrelated to the `layrnorm`/`layernorm` rename fix." + "reason": "Mistral4 tests/model changes are unrelated to the `layrnorm` spelling fix across CLIP-like models." }, { "left": "pull_request:44827", "right": "pull_request:46048", "accept": false, - "reason": "Mistral4 test fixes and TDT loss kernel integration are unrelated changes." + "reason": "Mistral4 test fixes are unrelated to adding the TDT loss kernel." }, { "left": "pull_request:44827", "right": "pull_request:45655", "accept": false, - "reason": "Mistral4 test/model work is unrelated to class config resolution ordering." + "reason": "Mistral4 fixes and `cls.config` resolution order are different bugs in different areas." }, { "left": "pull_request:44827", "right": "pull_request:45213", "accept": false, - "reason": "A model-specific test fix is unrelated to the model-creation skill PR." + "reason": "Mistral4 test/model work is unrelated to the new model-creation skill PR." }, { "left": "pull_request:43823", "right": "pull_request:45213", "accept": false, - "reason": "New model code and an AI skill scaffold do not fix the same code-path problem." + "reason": "Adding MobileLLM is unrelated to a model-creation skill/documentation/tooling PR." }, { "left": "pull_request:43251", "right": "pull_request:43254", "accept": true, - "reason": "Both PRs fix the same `fixed_cross_entropy` pass-through bug for issue 43240 and could plausibly be merged into one patch." + "reason": "Both PRs fix the same concrete bug in `src/transformers/loss/loss_utils.py`: forwarding supported kwargs (`weight`, `label_smoothing`) to `nn.functional.cross_entropy` for issue 43240." } ] }, "evaluator_result": { "accept": true, - "feedback": "The summary is grounded in the packet: the only clearly same-bug/same-file pair is PR 43251 and PR 43254, both targeting `fixed_cross_entropy` and issue 43240. The other pairwise rejections stay conservative and are supported by the previews/filenames, so the cluster-level conclusion is reasonable." + "feedback": "The summary is grounded in the packet, and the only accepted duplicate pair is well supported by the shared issue target and identical loss-utils bugfix. The false soft-edge verdicts are conservative and the reasons do not overstate equivalence." }, "evaluator_used": true, "retried": false, @@ -23434,7 +23525,7 @@ "soft_pairs": [ "pull_request:42975|pull_request:42976" ], - "prepared_review_unit_hash": "76d05dbcbed5cb5214b1458bccec93dddd1162eef07ccf484ee4c0e1f13b3430", + "prepared_review_unit_hash": "1211cf2760ee585121d744d28bb3faa62d146c762f468f0eed2583c6b5376e35", "component_budget": { "node_count": 2, "item_count": 2, @@ -23456,18 +23547,18 @@ "aggressively_trimmed": false, "split": false, "analyst_result": { - "summary": "These are two different GitHub Actions maintenance PRs: one is a broad Node 24 compatibility sweep, the other is a narrower docker/build-push-action version bump. They overlap in workflow files but do not appear to fix the same concrete problem.", - "confidence": 0.9, + "summary": "Two open PRs both touch GitHub Actions workflow dependencies, but they are different maintenance changes: one is a broad Node 24 compatibility upgrade for many actions, while the other is a separate bump of docker/build-push-action to newer versions. They should not be treated as duplicates.", + "confidence": 0.91, "canonical_issue_reason": null, - "canonical_pr_reason": "PR 42975 is the better representative because it is the broader, more explicit maintenance change set, spanning many workflows and clearly targeting Node 24 compatibility; 42976 is a narrower action-version bump for docker workflows.", + "canonical_pr_reason": "PR 42975 is the better canonical representative because it has the clearest, broader stated purpose: upgrading workflow actions for Node 24 compatibility across many files.", "best_issue_reason": null, - "best_pr_reason": "PR 42975 is the strongest cluster representative: it has the wider scope and clearer purpose, while 42976 is more narrowly focused on docker build action upgrades.", + "best_pr_reason": "PR 42975 is the strongest standalone representative of the cluster since it addresses a concrete platform-compatibility goal rather than only refreshing a narrower action dependency.", "soft_edge_verdicts": [ { "left": "pull_request:42975", "right": "pull_request:42976", "accept": false, - "reason": "Shared workflow maintenance and some overlapping filenames are not enough: 42975 is about general GitHub Actions Node 24 compatibility, while 42976 upgrades docker/build-push-action in a different subset of workflows. They do not look like the same underlying change and would not plausibly be merged as one PR." + "reason": "Both are workflow dependency bumps, but they target different changes. 42975 upgrades several actions for Node 24 compatibility; 42976 specifically updates docker/build-push-action versions. They do not appear to fix the same underlying problem and would not naturally merge into one PR." } ] }, @@ -23502,58 +23593,58 @@ "pull_request:43448|pull_request:45490", "pull_request:45490|pull_request:46562" ], - "prepared_review_unit_hash": "b6246feef273d30409dd815d37ff631ba60000abb5b7b6fde2fe179e7b9e4aa6", + "prepared_review_unit_hash": "7ea4ada15fd5532a27102bd16f521e758e3bad537708b411ccbf37a9c1595651", "component_budget": { "node_count": 4, "item_count": 4, "soft_pair_count": 4, - "serialized_chars": 13047, - "estimated_input_tokens": 3262, - "estimated_eval_tokens": 6780 + "serialized_chars": 13044, + "estimated_input_tokens": 3261, + "estimated_eval_tokens": 6778 }, "budget": { "node_count": 4, "item_count": 4, "soft_pair_count": 4, - "serialized_chars": 13047, - "estimated_input_tokens": 3262, - "estimated_eval_tokens": 6780 + "serialized_chars": 13044, + "estimated_input_tokens": 3261, + "estimated_eval_tokens": 6778 }, "overflow_policy": "truncate_then_skip", "trimmed": false, "aggressively_trimmed": false, "split": false, "analyst_result": { - "summary": "These are four unrelated model-addition PRs (ViT NEPA, Molmo, CTSM, Step3p7). They share only generic Transformers scaffolding like docs, auto mappings, and tests, but they implement different models and code paths, so none are duplicates.", - "confidence": 0.97, + "summary": "These are four separate model-addition PRs that only overlap on shared framework registration/docs files; they do not fix the same bug or implement the same model change, so none should be clustered as duplicates.", + "confidence": 0.98, "canonical_issue_reason": null, - "canonical_pr_reason": null, + "canonical_pr_reason": "No single canonical PR for the cluster: each PR adds a different model (ViT NEPA, Molmo, CTSM, Step3p7) with only generic infrastructure overlap.", "best_issue_reason": null, - "best_pr_reason": null, + "best_pr_reason": "46562 is the strongest single representative if one must be picked: it is a substantial, self-contained model addition with broad integration across auto-mapping, processing, modeling, and tests.", "soft_edge_verdicts": [ { "left": "pull_request:42978", "right": "pull_request:43448", "accept": false, - "reason": "Different models and implementations (ViT NEPA vs Molmo); overlap is only generic model-registration/docs scaffolding." + "reason": "Different model additions (ViT NEPA vs Molmo); shared auto/docs files are expected for registration, not evidence of the same underlying change." }, { "left": "pull_request:42978", "right": "pull_request:45490", "accept": false, - "reason": "ViT NEPA and CTSM are distinct model families with different code paths; shared auto/docs files are superficial." + "reason": "Different architectures and domains (vision model vs time-series CTSM); overlap is only in generic Transformers plumbing." }, { "left": "pull_request:43448", "right": "pull_request:45490", "accept": false, - "reason": "Molmo and CTSM add different architectures and files; similarity is limited to common Transformers integration boilerplate." + "reason": "Molmo and CTSM are unrelated model implementations; they share only common repo-wide integration files." }, { "left": "pull_request:45490", "right": "pull_request:46562", "accept": false, - "reason": "CTSM and Step3p7 are unrelated model additions; shared auto-mapping/docs updates do not indicate the same change." + "reason": "CTSM and Step3p7 target different code paths and model families; common auto/mapping changes are not enough to treat them as duplicates." } ] }, @@ -23602,12 +23693,12 @@ "pull_request:46539|pull_request:46922", "pull_request:46886|pull_request:46922" ], - "prepared_review_unit_hash": "2e8ef4acd4422dbfcc3306da6c120c10fae8be0a91f5916b694b31fb35a0f2db", + "prepared_review_unit_hash": "2a7f30f19a63ee0c2cd04e5cca0fc50df2684d4a61e55ae4eb9ee88889c0b7ea", "component_budget": { "node_count": 7, "item_count": 7, "soft_pair_count": 15, - "serialized_chars": 19604, + "serialized_chars": 19601, "estimated_input_tokens": 4901, "estimated_eval_tokens": 10058 }, @@ -23615,7 +23706,7 @@ "node_count": 7, "item_count": 7, "soft_pair_count": 15, - "serialized_chars": 19604, + "serialized_chars": 19601, "estimated_input_tokens": 4901, "estimated_eval_tokens": 10058 }, @@ -23624,108 +23715,108 @@ "aggressively_trimmed": false, "split": false, "analyst_result": { - "summary": "This cluster is mostly unrelated PRs spanning test cleanups, fp8 dequant, generation config, CI infrastructure, and one YOLOS device-mismatch fix. The only clear duplicate pair is the two YOLOS PRs, which target the same issue and code path.", - "confidence": 0.96, + "summary": "Only the two YOLOS PRs are true duplicates; the rest are generic fix/test PRs touching different models or subsystems and should stay separate.", + "confidence": 0.94, "canonical_issue_reason": null, - "canonical_pr_reason": "pull_request:46922 is the best canonical PR: it directly fixes the YOLOS device-mismatch path, matches the shared issue target, and is the more complete of the two overlapping YOLOS fixes.", + "canonical_pr_reason": "PR 46922 is the most complete YOLOS fix: it addresses the device-mismatch path directly and also includes the no-split-module change, subsuming PR 46886.", "best_issue_reason": null, - "best_pr_reason": "pull_request:46922 is the strongest representative overall because it addresses a concrete bug with an explicit code-path fix and clear issue linkage; the other PRs are either unrelated or mainly test/infra changes.", + "best_pr_reason": "PR 46922 is the strongest representative for the cluster because it is the clearest and most complete concrete code fix among the candidates.", "soft_edge_verdicts": [ { "left": "pull_request:43291", "right": "pull_request:43665", "accept": false, - "reason": "Both are test-only PRs, but they touch different models and failure surfaces; no shared concrete bug path." + "reason": "Both are vague test-fix PRs, but they touch different models and different test files; no shared underlying bug." }, { "left": "pull_request:43291", "right": "pull_request:44815", "accept": false, - "reason": "Whisper test adjustments and fp8 dequant changes are unrelated subsystems and fixes." + "reason": "Whisper tokenizer/test changes are unrelated to the FP8 dequant/loading fix." }, { "left": "pull_request:43291", "right": "pull_request:46380", "accept": false, - "reason": "Whisper test fixes do not match the generation/watermarking config change." + "reason": "Whisper test cleanup and generation/watermarking logic are different code paths." }, { "left": "pull_request:43291", "right": "pull_request:46922", "accept": false, - "reason": "Whisper test changes and YOLOS device-mismatch fixes are different model bugs." + "reason": "Whisper tests versus YOLOS device-mismatch fix are unrelated." }, { "left": "pull_request:43665", "right": "pull_request:46539", "accept": false, - "reason": "Model test edits versus Phimoe CI/integration changes are not the same underlying issue." + "reason": "CLIP/ViT test changes are not the same issue as the Phimoe CI integration test fix." }, { "left": "pull_request:43665", "right": "pull_request:44815", "accept": false, - "reason": "Different models and different bug classes; the test changes do not indicate one shared fix." + "reason": "Generic test edits in vision models do not match the FP8 dequant/core loading change." }, { "left": "pull_request:43665", "right": "pull_request:46380", "accept": false, - "reason": "CLIP/ViT test noise is unrelated to generation config validation." + "reason": "Test-only changes for CLIP/ViT are unrelated to generation config and whisper generation fixes." }, { "left": "pull_request:43665", "right": "pull_request:46922", "accept": false, - "reason": "These are unrelated test edits versus a YOLOS runtime/device bug." + "reason": "Different models and different failure modes; only superficial similarity in being small fixes." }, { "left": "pull_request:44815", "right": "pull_request:46539", "accept": false, - "reason": "fp8 dequant/model-loading changes and Phimoe CI changes do not fix the same code-path problem." + "reason": "FP8 dequant/loading logic and Phimoe integration CI setup are different problems." }, { "left": "pull_request:44815", "right": "pull_request:46380", "accept": false, - "reason": "Different functionality: fine-grained FP8 dequantization versus generation watermarking validation." + "reason": "No shared concrete code path; FP8 quantization fix does not overlap with generation/watermarking and whisper changes." }, { "left": "pull_request:44815", "right": "pull_request:46922", "accept": false, - "reason": "FP8 quantization fixes and YOLOS device-placement fixes are unrelated." + "reason": "FP8 dequant fix and YOLOS device placement fix are unrelated." }, { "left": "pull_request:46380", "right": "pull_request:46539", "accept": false, - "reason": "Generation-config validation and Phimoe CI/device setup are separate issues." + "reason": "Generation config/watermarking and CI runner memory/device-map setup are different subsystems." }, { "left": "pull_request:46380", "right": "pull_request:46922", "accept": false, - "reason": "Watermarking-config validation and YOLOS device mismatch are different problems." + "reason": "Wholly different areas: generation config versus YOLOS model device handling." }, { "left": "pull_request:46539", "right": "pull_request:46922", "accept": false, - "reason": "CI runner memory/device-map setup is unrelated to the YOLOS model device mismatch." + "reason": "CI infrastructure for Phimoe is unrelated to the YOLOS device-mismatch bug." }, { "left": "pull_request:46886", "right": "pull_request:46922", "accept": true, - "reason": "Both PRs target the same YOLOS device-mismatch issue on the same file and are plausibly mergeable into one fix; 46922 simply adds the direct device transfer on top of the no-split-module adjustment." + "reason": "Same YOLOS device-mismatch/device_map='auto' bug; both target issue 46869 and the later PR adds the missing direct device fix alongside the same split-module guard." } ] }, "evaluator_result": { "accept": true, - "feedback": "The summary is grounded in the packet, and the single accepted soft-edge verdict is conservative: 46886 and 46922 share the same YOLOS issue target and file, with overlapping fixes that look mergeable into one PR. The other pairwise rejections are also appropriately cautious given the packet evidence." + "feedback": "The summary is well-grounded in the packet: only the two YOLOS PRs share the same file and issue target, and the rest are clearly different models/subsystems. The soft-edge verdicts are conservative, with the one true duplicate pair correctly accepted and the others reasonably rejected as not mergeable into a single bugfix PR." }, "evaluator_used": true, "retried": false, @@ -23755,52 +23846,52 @@ "pull_request:43488|pull_request:43532", "pull_request:43532|pull_request:46747" ], - "prepared_review_unit_hash": "caf7326dcb7065dcf86ac6c168a9f22262c84dc412b16bc48d4c1f8ed919602c", + "prepared_review_unit_hash": "f11a602ea787e5434e84f5d13bebb9907a19dc87bc22a8cf6ca02e5d2eee3521", "component_budget": { "node_count": 3, "item_count": 3, "soft_pair_count": 3, - "serialized_chars": 6348, - "estimated_input_tokens": 1587, - "estimated_eval_tokens": 3430 + "serialized_chars": 6350, + "estimated_input_tokens": 1588, + "estimated_eval_tokens": 3432 }, "budget": { "node_count": 3, "item_count": 3, "soft_pair_count": 3, - "serialized_chars": 6348, - "estimated_input_tokens": 1587, - "estimated_eval_tokens": 3430 + "serialized_chars": 6350, + "estimated_input_tokens": 1588, + "estimated_eval_tokens": 3432 }, "overflow_policy": "truncate_then_skip", "trimmed": false, "aggressively_trimmed": false, "split": false, "analyst_result": { - "summary": "These are three unrelated '[don't merge]'/test PRs touching different areas (packaging/versioning, workflow debug changes, and a security-gate workflow). They do not look like the same underlying bug or change, so none of the soft duplicate links should be accepted.", - "confidence": 0.93, + "summary": "All three pull requests are unrelated '[don't merge]' test or demo changes. They touch different files and different workflows/code paths, so none are duplicates of each other.", + "confidence": 0.96, "canonical_issue_reason": null, - "canonical_pr_reason": "PR 43532 is the broadest/most representative of the cluster in terms of files touched and context, but it is still just a dummy workflow/test PR rather than a true duplicate anchor.", + "canonical_pr_reason": "No single PR represents the same underlying change as the others; they are separate test PRs with different purposes. If a representative is needed, PR 46747 is the cleanest single-file workflow tweak.", "best_issue_reason": null, - "best_pr_reason": "PR 43532 has the most substantive patch surface and is the easiest representative to inspect, though it still should not be treated as the same change as the others.", + "best_pr_reason": "PR 46747 is the most self-contained and concrete change: one workflow file with a narrow trigger update, unlike the broader or multi-file test PRs.", "soft_edge_verdicts": [ { "left": "pull_request:43488", "right": "pull_request:46747", "accept": false, - "reason": "Different code paths and intents: 43488 is a packaging/version/dependency test PR, while 46747 only flips a security-gate workflow trigger from pull_request_target to pull_request. Shared 'don't merge' style is not enough to make them duplicates." + "reason": "Different changes: 43488 is a setup/version/dependency test PR, while 46747 alters a security-gate workflow trigger. They do not fix the same bug or change." }, { "left": "pull_request:43488", "right": "pull_request:43532", "accept": false, - "reason": "They overlap only on a Vit test file, but the actual changes are unrelated: 43488 edits setup/dependency metadata and a test, while 43532 rewrites several workflows and testing/notification helpers. Not the same underlying bug/change." + "reason": "Only superficial overlap via the vit test file; 43488 is about repo-bot/version formatting and dependencies, while 43532 is a workflow/diff-display test PR." }, { "left": "pull_request:43532", "right": "pull_request:46747", "accept": false, - "reason": "Both touch workflows, but the concrete edits are different and not mergeable as one fix: 43532 changes comment/scheduled/slack workflow plumbing plus tests, while 46747 only changes one security-gate workflow trigger." + "reason": "Both mention workflows, but they modify different workflow files for different behaviors. They are not the same underlying change and would not merge as one PR." } ] }, @@ -23830,12 +23921,12 @@ "soft_pairs": [ "pull_request:43612|pull_request:43613" ], - "prepared_review_unit_hash": "eb51d51c59c80742f8f1def282c7cd0e3859b9f41e9684675f952dc053d68dad", + "prepared_review_unit_hash": "c19ae61ce20962412cc97de9c6cea0cbe99d1b92839240722c1ecbbef8a6893e", "component_budget": { "node_count": 2, "item_count": 2, "soft_pair_count": 1, - "serialized_chars": 6082, + "serialized_chars": 6083, "estimated_input_tokens": 1521, "estimated_eval_tokens": 3298 }, @@ -23843,7 +23934,7 @@ "node_count": 2, "item_count": 2, "soft_pair_count": 1, - "serialized_chars": 6082, + "serialized_chars": 6083, "estimated_input_tokens": 1521, "estimated_eval_tokens": 3298 }, @@ -23852,18 +23943,18 @@ "aggressively_trimmed": false, "split": false, "analyst_result": { - "summary": "Related segmentation-pipeline additions, but they target different tasks/models and should not be merged as duplicates.", - "confidence": 0.97, + "summary": "These are closely related pipeline-adding PRs in the same area, but they implement different features and should not be merged as duplicates: one adds Promptable Concept Segmentation, the other adds Promptable Visual Segmentation.", + "confidence": 0.95, "canonical_issue_reason": null, - "canonical_pr_reason": "No canonical PR: these are separate feature additions rather than duplicate fixes.", + "canonical_pr_reason": "No true canonical duplicate PR here; if forced to pick one representative, PR 43613 is slightly broader and more central to the shared multimodal segmentation pipeline work.", "best_issue_reason": null, - "best_pr_reason": "PR 43613 is the broader and more complete representative of the visual-segmentation feature, with more backend/test coverage.", + "best_pr_reason": "PR 43613 is the better standalone representative because it covers a larger surface area and adds more model/test integration, but it is still a distinct feature from PR 43612.", "soft_edge_verdicts": [ { "left": "pull_request:43612", "right": "pull_request:43613", "accept": false, - "reason": "Different concrete features: Promptable Concept Segmentation vs Promptable Visual Segmentation. They share scaffold/docs patterns, but the core pipeline files, model integrations, and tests differ, so this is not the same underlying change." + "reason": "Reject: they target different concrete pipelines and APIs (concept segmentation vs visual segmentation). Shared docs, registry, and test scaffolding are not enough to treat them as the same underlying change." } ] }, @@ -23893,40 +23984,40 @@ "soft_pairs": [ "pull_request:43651|pull_request:43663" ], - "prepared_review_unit_hash": "1af5c15ac33ea462a8dcbbaa871f408f581b72152a1954526542ea725c08f8ce", + "prepared_review_unit_hash": "b7410709e84ca5c17e15c14e06dfcce691fe65753345b30a6cbe646e2efcad65", "component_budget": { "node_count": 2, "item_count": 2, "soft_pair_count": 1, - "serialized_chars": 4293, - "estimated_input_tokens": 1074, - "estimated_eval_tokens": 2404 + "serialized_chars": 4292, + "estimated_input_tokens": 1073, + "estimated_eval_tokens": 2402 }, "budget": { "node_count": 2, "item_count": 2, "soft_pair_count": 1, - "serialized_chars": 4293, - "estimated_input_tokens": 1074, - "estimated_eval_tokens": 2404 + "serialized_chars": 4292, + "estimated_input_tokens": 1073, + "estimated_eval_tokens": 2402 }, "overflow_policy": "truncate_then_skip", "trimmed": false, "aggressively_trimmed": false, "split": false, "analyst_result": { - "summary": "Both pull requests touch `trainer.py` and target custom-trainer extensibility, but they fix different problems: one changes gradient-accumulation loss scaling, the other refactors signature-column filtering. They are not the same underlying change and should not be merged as duplicates.", - "confidence": 0.97, + "summary": "These two PRs both modify `trainer.py`, but they address different trainer behaviors: one changes gradient-accumulation loss scaling, the other refactors signature-column filtering for column removal. They are not the same underlying change.", + "confidence": 0.98, "canonical_issue_reason": null, - "canonical_pr_reason": null, + "canonical_pr_reason": "PR #43651 is the clearer canonical item for the cluster because it makes a focused behavior change in training loss scaling and introduces a concrete hook for custom trainers.", "best_issue_reason": null, - "best_pr_reason": "PR 43651 is the more concrete standalone fix: it addresses a specific loss-scaling behavior in the training step. PR 43663 is also a refactor/extension-point change, but it is unrelated to loss scaling and not a duplicate of 43651.", + "best_pr_reason": "PR #43651 is the better representative because its purpose is narrowly defined and the diff directly changes the training-step loss-scaling path; PR #43663 is a separate refactor of input-column inspection.", "soft_edge_verdicts": [ { "left": "pull_request:43651", "right": "pull_request:43663", "accept": false, - "reason": "Different code paths and different user-facing problems: 43651 changes gradient-accumulation loss normalization, while 43663 extracts signature-column logic for input filtering. They share a file and theme but not the same underlying bug or change." + "reason": "Different code paths and different goals: loss scaling during training vs. signature column derivation for filtering. Shared filename alone is not enough to treat them as duplicates." } ] }, @@ -23956,7 +24047,7 @@ "soft_pairs": [ "pull_request:43656|pull_request:43842" ], - "prepared_review_unit_hash": "e747e152548f4cf66d892cde7d034f9d5ad0f7f1017eed6e4a4d4e1169909d6a", + "prepared_review_unit_hash": "275dd27bff1549111c6a1485377a8853ae865ad9e5bfd1edee82e0c01b045b44", "component_budget": { "node_count": 2, "item_count": 2, @@ -23978,18 +24069,18 @@ "aggressively_trimmed": false, "split": false, "analyst_result": { - "summary": "Two open PRs touch `src/transformers/cli/serve.py`, but they appear to fix different runtime typing/import problems rather than the same bug.", - "confidence": 0.77, + "summary": "Both PRs touch `src/transformers/cli/serve.py` and mention `TypeAdapter`, but they appear to fix different annotation/import-time problems rather than the same concrete bug. One is a broader `from __future__ import annotations` typing cleanup, while the other specifically handles `TypeAdapter` when `pydantic` is absent.", + "confidence": 0.82, "canonical_issue_reason": null, - "canonical_pr_reason": "PR 43842 is the more direct fix for the `TypeAdapter` NameError when pydantic is absent: it guards the import under `TYPE_CHECKING` and stringizes the annotation. PR 43656 is broader annotation cleanup/deferred-evaluation work and looks like a different fix path.", + "canonical_pr_reason": "PR 43842 is the more specific and better-scoped fix: it explicitly guards the `TypeAdapter` import under `TYPE_CHECKING` and quotes the annotation, directly addressing the optional-`pydantic` NameError path.", "best_issue_reason": null, - "best_pr_reason": "43842 best matches the cluster\u2019s likely root cause because it specifically addresses the `TypeAdapter` runtime NameError with minimal, targeted changes.", + "best_pr_reason": "PR 43842 is the strongest representative of this cluster because its title, issue target, and diff all align on the concrete `TypeAdapter`/missing-`pydantic` failure mode.", "soft_edge_verdicts": [ { "left": "pull_request:43656", "right": "pull_request:43842", "accept": false, - "reason": "Do not treat as duplicates: 43656 appears to fix broader import-time annotation evaluation in `serve.py`, while 43842 specifically fixes `TypeAdapter` when pydantic is not installed. Same file, but not the same concrete bug or mergeable change set." + "reason": "Related file and symbol, but the fixes target different runtime/type-annotation failure modes; they are not clearly the same bug and would not naturally be merged as one PR." } ] }, @@ -24117,8 +24208,8 @@ "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:44071", "pull_request:44066|pull_request:44068", "pull_request:44018|pull_request:44066", "pull_request:44019|pull_request:44722", @@ -24137,8 +24228,8 @@ "pull_request:44098|pull_request:44101", "pull_request:44085|pull_request:44086", "pull_request:44059|pull_request:44086", - "pull_request:44018|pull_request:44071", "pull_request:44059|pull_request:44722", + "pull_request:44018|pull_request:44071", "pull_request:44017|pull_request:44161", "pull_request:44017|pull_request:44073", "pull_request:44017|pull_request:44722", @@ -24182,355 +24273,415 @@ "pull_request:44024|pull_request:44076", "pull_request:44073|pull_request:44074" ], - "prepared_review_unit_hash": "2eddada8247216eb7d81a137b7853cd7958789315f3f13feeb560b19cf977067", + "prepared_review_unit_hash": "d9edcba9ca529763de101e1b6c708a577515a17856fa095ab9c9a44d26252512", "component_budget": { "node_count": 28, "item_count": 28, "soft_pair_count": 76, - "serialized_chars": 84113, - "estimated_input_tokens": 21029, - "estimated_eval_tokens": 42314 + "serialized_chars": 84111, + "estimated_input_tokens": 21028, + "estimated_eval_tokens": 42312 }, "budget": { "node_count": 28, "item_count": 28, "soft_pair_count": 72, - "serialized_chars": 76937, - "estimated_input_tokens": 19235, - "estimated_eval_tokens": 38726 + "serialized_chars": 76933, + "estimated_input_tokens": 19234, + "estimated_eval_tokens": 38724 }, "overflow_policy": "truncate_then_skip", "trimmed": false, "aggressively_trimmed": false, "split": true, "analyst_result": { - "summary": "This cluster is a mix of model-specific output-tracing refactors. The only true duplicate-like subclusters are GPT-J/CodeGen and GPT-Neo; most other PRs share only the broad standardized capture_outputs pattern and should stay separate.", - "confidence": 0.85, + "summary": "This cluster is mostly a family of model-specific refactors that standardize output tracing / capture_outputs across Transformers models. Most PRs are not duplicates of each other because they touch different model code paths; the only clear overlaps are within the GPT-J/CodeGen subcluster and the GPT-Neo pair.", + "confidence": 0.91, "canonical_issue_reason": null, - "canonical_pr_reason": "PR 44066 is the cleanest anchor: it is the most central GPT-J/CodeGen output-tracing refactor, explicitly tied to the tracking issue, and sits in the densest duplicate subcluster.", + "canonical_pr_reason": "PR 44722 is the strongest representative of the overlapping GPT-J/CodeGen subcluster: it targets the same output-tracing refactor as the other GPT-J PRs, touches the same concrete files, and is the most complete/latest instance among the clearly overlapping changes.", "best_issue_reason": null, - "best_pr_reason": "PR 44066 is the best overall representative because it covers the core GPT-J/CodeGen change family with explicit issue linkage and the broadest concrete scope in that subcluster.", + "best_pr_reason": "44722 is the best overall PR anchor for this cluster because it has the clearest concrete overlap with other accepted duplicates and the strongest explicit refactor framing; the rest of the cluster is too model-diverse to share one canonical PR cleanly.", "soft_edge_verdicts": [ { "left": "pull_request:43996", "right": "pull_request:44085", "accept": false, - "reason": "Different model families and concrete code paths; only the output-tracing refactor style overlaps." + "reason": "Different model families (CVT/FNet vs GPT-J); same refactor theme is too broad to be a duplicate." }, { "left": "pull_request:43996", "right": "pull_request:44044", "accept": false, - "reason": "CVT/FNet and DeBERTa are different model implementations; shared tracing cleanup is too broad." + "reason": "Different models and code paths (CVT/FNet vs DeBERTa V2)." }, { "left": "pull_request:44066", "right": "pull_request:44085", "accept": true, - "reason": "Same GPT-J/CodeGen output-tracing refactor family in the same code area; these look like duplicate iterations of one change." + "reason": "Both refactor GPT-J output tracing in the same file/code path; 44066 is a broader superset that also includes CodeGen." }, { "left": "pull_request:44007", "right": "pull_request:44072", "accept": false, - "reason": "ResNet/RT-DETR and EfficientNet are separate model implementations; the refactor pattern alone is not enough." + "reason": "RegNet/ResNet vs EfficientNet are distinct implementations, not the same underlying change." }, { "left": "pull_request:44072", "right": "pull_request:44722", "accept": false, - "reason": "EfficientNet and GPT-J/CodeGen are unrelated code paths." + "reason": "EfficientNet and GPT-J/CodeGen are unrelated model paths." }, { "left": "pull_request:44066", "right": "pull_request:44072", "accept": false, - "reason": "GPT-J/CodeGen and EfficientNet are different concrete model paths; only the generic tracing migration theme overlaps." + "reason": "GPT-J/CodeGen output tracing vs EfficientNet output tracing are different concrete changes." }, { "left": "pull_request:44013", "right": "pull_request:44044", "accept": false, - "reason": "MobileNetV2 and DeBERTa are unrelated model families." + "reason": "MobileNetV2 vs DeBERTa V2; only the refactor pattern is shared." }, { "left": "pull_request:44066", "right": "pull_request:44086", "accept": false, - "reason": "GPT-J/CodeGen and MGP-STR are unrelated concrete changes." - }, - { - "left": "pull_request:44066", - "right": "pull_request:44071", - "accept": false, - "reason": "GPT-J/CodeGen and MPT are separate model implementations." + "reason": "GPT-J/CodeGen refactor vs MGP-STR refactor; unrelated code paths." }, { "left": "pull_request:44018", "right": "pull_request:44068", "accept": true, - "reason": "Both are GPT-Neo output-tracing refactors in the same file family, and the diffs look like duplicate/revision iterations of the same change." + "reason": "Both target GPT-Neo output tracing in the same model/file, and appear to be alternate versions of the same refactor." + }, + { + "left": "pull_request:44066", + "right": "pull_request:44071", + "accept": false, + "reason": "GPT-J/CodeGen vs MPT; different model families." }, { "left": "pull_request:44066", "right": "pull_request:44068", "accept": false, - "reason": "GPT-J/CodeGen and GPT-Neo are different concrete model paths, despite the shared tracing refactor theme." + "reason": "GPT-J/CodeGen and GPT-Neo are different model code paths." }, { "left": "pull_request:44018", "right": "pull_request:44066", "accept": false, - "reason": "GPT-Neo and GPT-J/CodeGen are different model implementations." + "reason": "GPT-Neo vs GPT-J/CodeGen; same issue theme but not the same change." }, { "left": "pull_request:44019", "right": "pull_request:44722", "accept": false, - "reason": "ResNet and GPT-J/CodeGen are unrelated model families." + "reason": "ResNet output tracing is unrelated to GPT-J/CodeGen output tracing." }, { "left": "pull_request:44086", "right": "pull_request:44722", "accept": false, - "reason": "MGP-STR and GPT-J/CodeGen are unrelated concrete changes." + "reason": "MGP-STR and GPT-J/CodeGen are unrelated model implementations." }, { "left": "pull_request:44019", "right": "pull_request:44085", "accept": false, - "reason": "ResNet and GPT-J/CodeGen are only loosely related by the tracing refactor pattern." + "reason": "ResNet and GPT-J are different model paths." }, { "left": "pull_request:44018", "right": "pull_request:44019", "accept": false, - "reason": "GPT-Neo and ResNet are different models with different output implementations." + "reason": "GPT-Neo and ResNet are unrelated." }, { "left": "pull_request:44019", "right": "pull_request:44071", "accept": false, - "reason": "ResNet and MPT are unrelated code paths." + "reason": "ResNet vs MPT; same broad pattern but not the same concrete change." }, { "left": "pull_request:44018", - "right": "pull_request:44129", + "right": "pull_request:44116", "accept": false, - "reason": "GPT-Neo and SpeechT5 are unrelated model families." + "reason": "GPT-Neo vs Flaubert/XLM; different model families." }, { "left": "pull_request:44068", "right": "pull_request:44116", "accept": false, - "reason": "GPT-Neo and Flaubert/XLM are unrelated implementations." + "reason": "GPT-Neo vs Flaubert/XLM; only the decorator refactor theme overlaps." + }, + { + "left": "pull_request:44068", + "right": "pull_request:44722", + "accept": false, + "reason": "GPT-Neo and GPT-J/CodeGen are different model code paths." + }, + { + "left": "pull_request:44071", + "right": "pull_request:44722", + "accept": false, + "reason": "MPT and GPT-J/CodeGen are unrelated implementations." + }, + { + "left": "pull_request:44018", + "right": "pull_request:44085", + "accept": false, + "reason": "GPT-Neo vs GPT-J; different model-specific refactors." + }, + { + "left": "pull_request:44018", + "right": "pull_request:44086", + "accept": false, + "reason": "GPT-Neo vs MGP-STR, unrelated." + }, + { + "left": "pull_request:44068", + "right": "pull_request:44086", + "accept": false, + "reason": "GPT-Neo vs MGP-STR, unrelated." + }, + { + "left": "pull_request:44071", + "right": "pull_request:44085", + "accept": false, + "reason": "MPT vs GPT-J; not the same code-path problem." + }, + { + "left": "pull_request:44086", + "right": "pull_request:44085", + "accept": false, + "reason": "MGP-STR and GPT-J are different model families." + }, + { + "left": "pull_request:44059", + "right": "pull_request:44086", + "accept": false, + "reason": "GPT2 vs MGP-STR; unrelated changes." + }, + { + "left": "pull_request:44018", + "right": "pull_request:44071", + "accept": false, + "reason": "GPT-Neo vs MPT, different models." + }, + { + "left": "pull_request:44017", + "right": "pull_request:44161", + "accept": false, + "reason": "SegFormer vs LongT5; same standardized-output theme but not the same underlying change." }, { - "left": "pull_request:44068", - "right": "pull_request:44722", + "left": "pull_request:44017", + "right": "pull_request:44073", "accept": false, - "reason": "GPT-Neo and GPT-J/CodeGen are different model families." + "reason": "SegFormer vs VisualBERT are different model paths." }, { - "left": "pull_request:44071", + "left": "pull_request:44017", "right": "pull_request:44722", "accept": false, - "reason": "MPT and GPT-J/CodeGen are separate code paths." + "reason": "SegFormer output tracing is unrelated to GPT-J/CodeGen." }, { - "left": "pull_request:44018", - "right": "pull_request:44085", + "left": "pull_request:44068", + "right": "pull_request:44071", "accept": false, - "reason": "GPT-Neo and GPT-J/CodeGen are different model implementations." + "reason": "GPT-Neo vs MPT; different model-specific changes." }, { - "left": "pull_request:44018", - "right": "pull_request:44086", + "left": "pull_request:43997", + "right": "pull_request:44073", "accept": false, - "reason": "GPT-Neo and MGP-STR are unrelated changes." + "reason": "RegNet vs VisualBERT are unrelated models." }, { - "left": "pull_request:44068", - "right": "pull_request:44086", + "left": "pull_request:44017", + "right": "pull_request:44071", "accept": false, - "reason": "GPT-Neo and MGP-STR are unrelated code paths." + "reason": "SegFormer vs MPT; same refactor pattern only." }, { "left": "pull_request:44071", - "right": "pull_request:44085", + "right": "pull_request:44086", "accept": false, - "reason": "MPT and GPT-J are unrelated model families." + "reason": "MPT vs MGP-STR; unrelated model code paths." }, { - "left": "pull_request:44098", - "right": "pull_request:44101", + "left": "pull_request:44073", + "right": "pull_request:44161", "accept": false, - "reason": "ViLT and Flaubert/XLM are different model families." + "reason": "VisualBERT vs LongT5; not the same concrete change." }, { - "left": "pull_request:44085", - "right": "pull_request:44086", + "left": "pull_request:44073", + "right": "pull_request:44101", "accept": false, - "reason": "GPT-J and MGP-STR are unrelated concrete changes." + "reason": "VisualBERT vs Flaubert/XLM; different model families." }, { "left": "pull_request:44059", - "right": "pull_request:44086", + "right": "pull_request:44071", "accept": false, - "reason": "GPT2 and MGP-STR are unrelated implementations." + "reason": "GPT2 vs MPT are distinct model implementations." }, { - "left": "pull_request:44017", - "right": "pull_request:44161", + "left": "pull_request:44056", + "right": "pull_request:44101", "accept": false, - "reason": "SegFormer and LongT5 are different model families." + "reason": "MPNet vs Flaubert/XLM; unrelated." }, { - "left": "pull_request:44017", - "right": "pull_request:44073", + "left": "pull_request:44018", + "right": "pull_request:44059", "accept": false, - "reason": "SegFormer and VisualBert are unrelated code paths." + "reason": "GPT-Neo vs GPT2; same general refactor theme, different models." }, { - "left": "pull_request:44017", - "right": "pull_request:44722", + "left": "pull_request:44059", + "right": "pull_request:44066", "accept": false, - "reason": "SegFormer and GPT-J/CodeGen are unrelated." + "reason": "GPT2 vs GPT-J/CodeGen; unrelated concrete changes." }, { - "left": "pull_request:44068", - "right": "pull_request:44071", + "left": "pull_request:44101", + "right": "pull_request:44161", "accept": false, - "reason": "GPT-Neo and MPT are separate model implementations." + "reason": "Flaubert/XLM vs LongT5; different model paths." }, { "left": "pull_request:43997", - "right": "pull_request:44073", + "right": "pull_request:44017", "accept": false, - "reason": "RegNet and VisualBert are unrelated model families." + "reason": "RegNet vs SegFormer are unrelated implementations." }, { "left": "pull_request:43997", "right": "pull_request:44056", "accept": false, - "reason": "RegNet and MPNet are different concrete code paths." + "reason": "RegNet vs MPNet; same theme, different code paths." + }, + { + "left": "pull_request:44071", + "right": "pull_request:44076", + "accept": false, + "reason": "MPT vs ImageGPT; different models." }, { "left": "pull_request:44074", "right": "pull_request:44101", "accept": false, - "reason": "TextNet and Flaubert/XLM are unrelated models." + "reason": "TextNet vs Flaubert/XLM are unrelated." }, { "left": "pull_request:44056", "right": "pull_request:44161", "accept": false, - "reason": "MPNet and LongT5 are different model implementations." + "reason": "MPNet vs LongT5; different model-specific refactors." }, { "left": "pull_request:43997", "right": "pull_request:44018", "accept": false, - "reason": "RegNet and GPT-Neo are unrelated concrete changes." + "reason": "RegNet vs GPT-Neo; unrelated changes." }, { "left": "pull_request:43997", "right": "pull_request:44161", "accept": false, - "reason": "RegNet and LongT5 are unrelated model families." + "reason": "RegNet vs LongT5 are unrelated model families." }, { "left": "pull_request:44074", "right": "pull_request:44161", "accept": false, - "reason": "TextNet and LongT5 are unrelated code paths." + "reason": "TextNet vs LongT5; different code paths." }, { "left": "pull_request:44010", "right": "pull_request:44071", "accept": false, - "reason": "SqueezeBert and MPT are unrelated models." + "reason": "SqueezeBERT vs MPT; unrelated models." }, { "left": "pull_request:44018", "right": "pull_request:44129", "accept": false, - "reason": "GPT-Neo and SpeechT5 are different model families." + "reason": "GPT-Neo vs SpeechT5; different model families." }, { "left": "pull_request:44010", "right": "pull_request:44073", "accept": false, - "reason": "SqueezeBert and VisualBert are unrelated concrete changes." + "reason": "SqueezeBERT vs VisualBERT; same refactor theme only." }, { - "left": "pull_request:44044", - "right": "pull_request:44129", + "left": "pull_request:43997", + "right": "pull_request:44074", + "accept": false, + "reason": "RegNet vs TextNet; unrelated model implementations." + }, + { + "left": "pull_request:44010", + "right": "pull_request:44018", "accept": false, - "reason": "DeBERTa and SpeechT5 are different code paths." + "reason": "SqueezeBERT vs GPT-Neo are different code paths." }, { "left": "pull_request:44044", - "right": "pull_request:44154", + "right": "pull_request:44074", "accept": false, - "reason": "DeBERTa and Vits are unrelated model families." + "reason": "DeBERTa V2 vs TextNet; unrelated." }, { "left": "pull_request:44024", "right": "pull_request:44076", "accept": false, - "reason": "FocalNet and ImageGPT are separate implementations; only the refactor style matches." + "reason": "FocalNet vs ImageGPT; different model-specific changes." }, { "left": "pull_request:44073", "right": "pull_request:44074", "accept": false, - "reason": "VisualBert and TextNet are different models, so this is not the same concrete change." - }, - { - "left": "pull_request:44066", - "right": "pull_request:44722", - "accept": true, - "reason": "Same GPT-J/CodeGen output-tracing refactor family with the same concrete files; these are duplicate iterations of one change." - }, - { - "left": "pull_request:44085", - "right": "pull_request:44722", - "accept": true, - "reason": "Same GPT-J/CodeGen change family in the same file set; one is effectively a revision of the other." + "reason": "VisualBERT vs TextNet; not the same underlying change." }, { - "left": "pull_request:44085", + "left": "pull_request:44044", "right": "pull_request:44129", "accept": false, - "reason": "GPT-J and SpeechT5 are unrelated model implementations." + "reason": "DeBERTa V2 vs SpeechT5 are unrelated model paths." }, { - "left": "pull_request:43996", - "right": "pull_request:44007", + "left": "pull_request:44044", + "right": "pull_request:44154", "accept": false, - "reason": "CVT/FNet and ResNet/RT-DETR are unrelated model families." + "reason": "DeBERTa V2 vs VITS; different concrete refactors." }, { - "left": "pull_request:43996", - "right": "pull_request:44129", - "accept": false, - "reason": "CVT/FNet and SpeechT5 are unrelated code paths." + "left": "pull_request:44066", + "right": "pull_request:44722", + "accept": true, + "reason": "Same GPT-J/CodeGen output-tracing refactor family, with matching filenames and overlapping implementation changes." }, { - "left": "pull_request:44044", - "right": "pull_request:44154", - "accept": false, - "reason": "DeBERTa and Vits are unrelated model implementations." + "left": "pull_request:44085", + "right": "pull_request:44722", + "accept": true, + "reason": "Both are GPT-J output-tracing refactors on the same file and could plausibly be merged into one PR." } ] }, "evaluator_result": { - "accept": false, - "feedback": "The summary is mostly grounded, but one soft-edge verdict is not conservative/precise: `pull_request:44018` vs `pull_request:44068` is labeled false with the reason \u201cGPT-Neo vs GPT-J/CodeGen are different model implementations,\u201d but both PRs are GPT-Neo refactors. That rationale is factually wrong and makes the verdict unreliable. The rest looks broadly conservative, but this mismatch is enough to reject." + "accept": true, + "feedback": "The summary is grounded in the packet: it keeps the broad family-level refactor theme, but correctly narrows duplicate-like overlap to the same-model GPT-J/CodeGen and GPT-Neo subclusters. The accepted soft-edge pairs all look mergeable as single bugfix PRs, and the rejected pairs are conservatively separated by model family/code path. No obvious overstatement beyond the evidence." }, "evaluator_used": true, - "retried": true, + "retried": false, "accepted_nontrivial_soft_edge": true, "error_kind": null, "error_message": null @@ -24561,20 +24712,20 @@ "pull_request:43997|pull_request:44010", "pull_request:44010|pull_request:44129" ], - "prepared_review_unit_hash": "8404ba3d2b52c8eeda8589b8f117264683cfe370d4d90574a785fed526bceb74", + "prepared_review_unit_hash": "f093ec7de7c9c150b2b9270c1ee4944e41cabfcb005e4a790d1db838f884651e", "component_budget": { "node_count": 28, "item_count": 28, "soft_pair_count": 76, - "serialized_chars": 84113, - "estimated_input_tokens": 21029, - "estimated_eval_tokens": 42314 + "serialized_chars": 84111, + "estimated_input_tokens": 21028, + "estimated_eval_tokens": 42312 }, "budget": { "node_count": 6, "item_count": 6, "soft_pair_count": 4, - "serialized_chars": 12754, + "serialized_chars": 12756, "estimated_input_tokens": 3189, "estimated_eval_tokens": 6634 }, @@ -24583,36 +24734,36 @@ "aggressively_trimmed": false, "split": true, "analyst_result": { - "summary": "These are model-specific PRs that all refactor output tracing/capture patterns, but they affect different code paths and are not duplicates of one another. The soft pairs should all be rejected under a strict PR-duplicate policy.", - "confidence": 0.93, + "summary": "These PRs are all variants of the same broad refactor pattern\u2014migrating different models to standardized output tracing/capture decorators\u2014but they touch different model implementations and code paths, so they are not duplicates of one another.", + "confidence": 0.89, "canonical_issue_reason": null, - "canonical_pr_reason": "44129 is the most concrete representative: it has an explicit issue target, a focused SpeechT5 output-tracing refactor, and enough diff detail to anchor the cluster theme.", + "canonical_pr_reason": "No single PR is a true duplicate anchor, but PR #44010 is the most representative of the cluster theme: it explicitly refactors a model to standardized output collection decorators, has review activity, and is a clean example of the pattern seen across the others.", "best_issue_reason": null, - "best_pr_reason": "44129 is the best single PR representative because it is the most clearly scoped and tied to a tracked issue, while still matching the shared standardized output-capture refactor theme.", + "best_pr_reason": "PR #44129 is the strongest standalone candidate for triage because it includes an explicit issue target, a substantial output-tracing refactor in a concrete model path, and enough detail to understand the intended change without needing the rest of the cluster.", "soft_edge_verdicts": [ { "left": "pull_request:44056", "right": "pull_request:44073", "accept": false, - "reason": "Same refactor pattern, but for different models (MPNet vs VisualBert). They do not fix the same concrete bug and would not naturally merge into one PR." + "reason": "Both are standardized output-tracing refactors, but they target different models (MPNet vs VisualBert) and different internal layers. Same refactor theme, not the same underlying bug/change." }, { "left": "pull_request:44056", "right": "pull_request:44074", "accept": false, - "reason": "Both are output-tracing migrations, but they touch different model implementations (MPNet vs TextNet) and are separate code-path changes, not a duplicate fix." + "reason": "These touch different model implementations (MPNet vs TextNet). They share a refactor pattern, but not a single concrete code-path fix." }, { "left": "pull_request:43997", "right": "pull_request:44010", "accept": false, - "reason": "These are model-specific standardized-output refactors for RegNet and SqueezeBert. Similar theme, but not the same underlying change or bug." + "reason": "RegNet and SqueezeBert are unrelated model code paths. The similarity is only the migration pattern, so they should not be merged as duplicates." }, { "left": "pull_request:44010", "right": "pull_request:44129", "accept": false, - "reason": "SqueezeBert and SpeechT5 have different forward signatures and output-capture adaptations; they are distinct model migrations, not one mergeable fix." + "reason": "SqueezeBert and SpeechT5 are separate model stacks with different forward/output mechanics. Both are output-tracing refactors, but not the same underlying change." } ] }, @@ -24692,274 +24843,268 @@ "pull_request:44028|pull_request:44030", "pull_request:44029|pull_request:44030" ], - "prepared_review_unit_hash": "b0035a2e2974096fa460bb987d027b0db2648b99cb9f8041ea6f5d895c3cb88a", + "prepared_review_unit_hash": "609d579b637123ea9204aaf283ee90fc2880469c617046a01dd4e01ee4373da3", "component_budget": { "node_count": 13, "item_count": 13, "soft_pair_count": 40, - "serialized_chars": 40180, - "estimated_input_tokens": 10045, - "estimated_eval_tokens": 20346 + "serialized_chars": 40172, + "estimated_input_tokens": 10043, + "estimated_eval_tokens": 20342 }, "budget": { "node_count": 13, "item_count": 13, "soft_pair_count": 40, - "serialized_chars": 40180, - "estimated_input_tokens": 10045, - "estimated_eval_tokens": 20346 + "serialized_chars": 40172, + "estimated_input_tokens": 10043, + "estimated_eval_tokens": 20342 }, "overflow_policy": "truncate_then_skip", "trimmed": false, "aggressively_trimmed": false, "split": false, "analyst_result": { - "summary": "These PRs are all about the same general output-tracing/refactor theme, but they are separate model-specific changes rather than duplicates of one underlying fix.", - "confidence": 0.94, + "summary": "This cluster is a set of parallel, model-specific output-tracing refactors, not duplicate fixes. The PRs share a pattern and tooling, but each one targets a different architecture and code path, so they should stay separate.", + "confidence": 0.93, "canonical_issue_reason": null, - "canonical_pr_reason": "PR 44003 is the best representative of the cluster because it is the broadest and most substantive refactor, spanning two closely related Mamba-family files and adding shared output-capture plumbing.", + "canonical_pr_reason": "PR 44003 is the broadest representative in the set, spanning both `falcon_mamba` and `mamba` and showing the clearest instance of the shared output-tracing refactor pattern.", "best_issue_reason": null, - "best_pr_reason": "PR 44003 is the strongest cluster anchor: it covers a larger shared codepath than the single-model wrappers and is more representative of the refactor pattern than the narrower PRs.", + "best_pr_reason": "PR 44003 is the strongest representative of this cluster because it covers multiple related Mamba-family files and the core tracing machinery, making it the best single exemplar of the refactor series.", "soft_edge_verdicts": [ { "left": "pull_request:43998", "right": "pull_request:44002", "accept": false, - "reason": "Same refactor theme, but different model implementations and codepaths; not a single mergeable change." + "reason": "Different model families (`timm_backbone` vs `upernet`); same refactor theme but not the same bug/change." }, { "left": "pull_request:43998", "right": "pull_request:44000", "accept": false, - "reason": "Both are output-tracing refactors, but they target different models and files, so they are not duplicates." + "reason": "`timm_backbone` and `vision_text_dual_encoder` are unrelated code paths; both only adopt output-tracing helpers." }, { "left": "pull_request:43998", "right": "pull_request:44001", "accept": false, - "reason": "Separate model-specific refactors with no shared concrete bug or codepath." + "reason": "Separate models (`timm_backbone` vs `univnet`); no shared concrete defect." }, { "left": "pull_request:43998", "right": "pull_request:44003", "accept": false, - "reason": "Different model families; 44003 is Mamba-family while 43998 is timm_backbone, so this is only thematic similarity." + "reason": "Different architectures and files; this is a pattern match, not a duplicate fix." }, { "left": "pull_request:43998", "right": "pull_request:44029", "accept": false, - "reason": "Same broad output-refactor pattern, but unrelated model codepaths and no evidence of one shared fix." + "reason": "`timm_backbone` and `rwkv` are distinct model implementations with different tracing behavior." }, { "left": "pull_request:43999", "right": "pull_request:44001", "accept": false, - "reason": "Different models and different implementation targets; not the same underlying change." + "reason": "`mobilenet_v1` and `univnet` touch unrelated forward paths; same tooling theme only." }, { "left": "pull_request:43999", "right": "pull_request:44002", "accept": false, - "reason": "Both touch output tracing, but they are distinct model-specific refactors with separate merge surfaces." + "reason": "`mobilenet_v1` vs `upernet`; both refactor outputs, but not the same underlying change." }, { "left": "pull_request:43999", "right": "pull_request:44025", "accept": false, - "reason": "No shared concrete bug; these are separate model wrappers with different forward signatures and plumbing." + "reason": "Different vision models and different output contracts; not mergeable as one PR." }, { "left": "pull_request:43999", "right": "pull_request:44026", "accept": false, - "reason": "Different model families and files; similarity is only at the refactor pattern level." + "reason": "`mobilenet_v1` and `vision_encoder_decoder` are separate code paths and refactor scopes." }, { "left": "pull_request:43999", "right": "pull_request:44027", "accept": false, - "reason": "Separate encoder-decoder wrapper/model refactors, not one common fix." + "reason": "`mobilenet_v1` vs `speech_encoder_decoder`; parallel refactors, not duplicates." }, { "left": "pull_request:44000", "right": "pull_request:44001", "accept": false, - "reason": "Both are refactors for output handling, but they change different models and codepaths." + "reason": "`vision_text_dual_encoder` and `univnet` are unrelated model families; only the refactor pattern matches." }, { "left": "pull_request:44000", "right": "pull_request:44002", "accept": false, - "reason": "Same stylistic change, but distinct model implementations; not duplicate enough to merge together." + "reason": "Different subsystems (`vision_text_dual_encoder` vs `upernet`); no shared concrete bug." }, { "left": "pull_request:44000", "right": "pull_request:44003", "accept": false, - "reason": "Different subsystems and model codepaths; 44003's Mamba-family changes are unrelated to vision-text dual encoder." - }, - { - "left": "pull_request:44000", - "right": "pull_request:44004", - "accept": false, - "reason": "These are separate model-specific refactors; no shared concrete bug or identical fix." + "reason": "`vision_text_dual_encoder` and `falcon_mamba/mamba` change different forward implementations." }, { "left": "pull_request:44000", "right": "pull_request:44025", "accept": false, - "reason": "Different models and different forward/tracing plumbing; only the refactor style matches." + "reason": "Distinct model families and filenames; same output-tracing cleanup, not the same change." }, { "left": "pull_request:44001", "right": "pull_request:44002", "accept": false, - "reason": "Both are output-tracing cleanups, but they are independent model implementations." + "reason": "`univnet` and `upernet` are unrelated; only generic tracing refactor overlap." }, { "left": "pull_request:44001", "right": "pull_request:44003", "accept": false, - "reason": "UnivNet and Mamba are unrelated codepaths; same theme is not enough for duplication." - }, - { - "left": "pull_request:44001", - "right": "pull_request:44004", - "accept": false, - "reason": "Different model families and different forward signatures; no single underlying bug." + "reason": "Different architectures (`univnet` vs `mamba` family) and different code paths." }, { "left": "pull_request:44001", "right": "pull_request:44025", "accept": false, - "reason": "Separate model-specific refactors with no direct overlap in implementation." + "reason": "Separate model implementations; no evidence they fix one shared defect." }, { "left": "pull_request:44002", "right": "pull_request:44029", "accept": false, - "reason": "UperNet and RWKV are unrelated implementations; not the same code-path problem." + "reason": "`upernet` and `rwkv` are different models with different output semantics." }, { "left": "pull_request:44002", "right": "pull_request:44003", "accept": false, - "reason": "Different model families and different output-capture changes; only broad refactor similarity." - }, - { - "left": "pull_request:44002", - "right": "pull_request:44004", - "accept": false, - "reason": "Distinct models with separate changes; not plausible as one merged PR." + "reason": "`upernet` vs `mamba` family; same refactor style, not same bug." }, { "left": "pull_request:44003", "right": "pull_request:44028", "accept": false, - "reason": "Mamba-family refactor versus SuperPoint refactor; no shared concrete fix." + "reason": "`mamba` family and `superpoint` are unrelated implementations; no common code path." }, { "left": "pull_request:44003", "right": "pull_request:44004", "accept": false, - "reason": "CodeGen and Mamba are different model paths; similarity is only in the refactor theme." + "reason": "`mamba`/`falcon_mamba` and `codegen` are different architectures; not a single mergeable fix." }, { "left": "pull_request:43998", "right": "pull_request:44004", "accept": false, - "reason": "Timm backbone and CodeGen are unrelated implementations; not duplicates." + "reason": "`timm_backbone` and `codegen` touch unrelated model code; same theme only." + }, + { + "left": "pull_request:44001", + "right": "pull_request:44004", + "accept": false, + "reason": "`univnet` and `codegen` are different domains and output shapes; not duplicates." + }, + { + "left": "pull_request:44002", + "right": "pull_request:44004", + "accept": false, + "reason": "`upernet` and `codegen` are separate model paths with no shared defect." }, { "left": "pull_request:44025", "right": "pull_request:44026", "accept": false, - "reason": "Different model wrappers and separate refactors; no single bug or feature change spans both." + "reason": "Different composite models (`depth_anything` vs `vision_encoder_decoder`); not the same change." }, { "left": "pull_request:44025", "right": "pull_request:44027", "accept": false, - "reason": "These are distinct depth-anything versus speech-encoder-decoder changes, not one underlying fix." + "reason": "`depth_anything` and `speech_encoder_decoder` are unrelated refactors." }, { "left": "pull_request:44025", "right": "pull_request:44028", "accept": false, - "reason": "Different models and codepaths; shared output-tracing pattern only." + "reason": "`depth_anything` and `superpoint` are different vision models with different output mechanics." }, { "left": "pull_request:44025", "right": "pull_request:44029", "accept": false, - "reason": "DepthAnything and RWKV are unrelated model implementations; not mergeable as one duplicate." + "reason": "`depth_anything` vs `rwkv`; distinct code paths and model families." }, { "left": "pull_request:44025", "right": "pull_request:44030", "accept": false, - "reason": "Separate model-specific output-refactor work, not the same concrete change." + "reason": "`depth_anything` and `dpr` are unrelated model implementations." }, { "left": "pull_request:44026", - "right": "pull_request:44027", + "right": "pull_request:44029", "accept": false, - "reason": "Vision-encoder-decoder and speech-encoder-decoder are analogous wrappers, but they are independent codepaths and not duplicates." + "reason": "`vision_encoder_decoder` and `rwkv` are not the same bug/change." }, { "left": "pull_request:44026", - "right": "pull_request:44028", + "right": "pull_request:44027", "accept": false, - "reason": "Different model families with separate implementations; no shared underlying bug." + "reason": "`vision_encoder_decoder` vs `speech_encoder_decoder`; similar wrappers, but separate model code." }, { "left": "pull_request:44026", - "right": "pull_request:44029", + "right": "pull_request:44028", "accept": false, - "reason": "Unrelated models and no common fix beyond general tracing cleanup." + "reason": "`vision_encoder_decoder` and `superpoint` are different model types and output paths." }, { "left": "pull_request:44026", "right": "pull_request:44030", "accept": false, - "reason": "Different model codepaths; not a single concrete issue or change." + "reason": "`vision_encoder_decoder` and `dpr` are distinct implementations; no shared concrete fix." }, { "left": "pull_request:44027", "right": "pull_request:44028", "accept": false, - "reason": "Speech encoder-decoder and SuperPoint are unrelated; same refactor motif only." + "reason": "`speech_encoder_decoder` and `superpoint` are unrelated; only the helper refactor overlaps." }, { "left": "pull_request:44027", "right": "pull_request:44029", "accept": false, - "reason": "Different models and output plumbing; no evidence they fix the same bug." + "reason": "`speech_encoder_decoder` vs `rwkv`; different architecture and tracing behavior." }, { "left": "pull_request:44027", "right": "pull_request:44030", "accept": false, - "reason": "Separate model-specific changes, not one duplicate fix." + "reason": "`speech_encoder_decoder` and `dpr` are separate model paths, not duplicates." }, { "left": "pull_request:44028", "right": "pull_request:44029", "accept": false, - "reason": "SuperPoint and RWKV are unrelated implementations; not the same underlying change." + "reason": "`superpoint` and `rwkv` are unrelated changes despite similar output-capturing work." }, { "left": "pull_request:44028", "right": "pull_request:44030", "accept": false, - "reason": "Different model families and codepaths; too broad to treat as duplicates." + "reason": "`superpoint` and `dpr` target different models and output contracts." }, { "left": "pull_request:44029", "right": "pull_request:44030", "accept": false, - "reason": "RWKV and DPR are different model implementations with separate refactor scopes." + "reason": "`rwkv` and `dpr` are different model families; not the same underlying fix." } ] }, @@ -24989,7 +25134,7 @@ "soft_pairs": [ "pull_request:44662|pull_request:45401" ], - "prepared_review_unit_hash": "4e534b04bc2ba07aaa6d887dbf29776481a77e4955d96ba1349301e3d282eaff", + "prepared_review_unit_hash": "371c7ea1367aeb62edd3c82403e7d5ddcf72a50f1f80a602fc95eafedcb8c3f3", "component_budget": { "node_count": 2, "item_count": 2, @@ -25011,18 +25156,18 @@ "aggressively_trimmed": false, "split": false, "analyst_result": { - "summary": "Two separate model-implementation PRs that share framework plumbing (docs and auto registries) but target different models and code paths: PenguinVL vs Voxtral TTS. They are not duplicates or merge candidates.", - "confidence": 0.96, + "summary": "Two separate model-addition PRs: one for PenguinVL and one for Voxtral TTS. They share Transformers boilerplate and auto-registration files, but they implement different models with different codepaths, so they are not duplicates.", + "confidence": 0.98, "canonical_issue_reason": null, - "canonical_pr_reason": "PR 45401 is the better representative because it has an explicit issue target and a tightly scoped Voxtral TTS implementation, whereas 44662 is a different model addition with no linked issue.", + "canonical_pr_reason": "PR #45401 is the more canonical choice for this cluster because it is narrowly scoped to Voxtral TTS, explicitly targets issue #45027, and has a focused implementation for a single model release.", "best_issue_reason": null, - "best_pr_reason": "45401 is the strongest single PR in the cluster to keep as representative; 44662 is also substantial, but it is a separate model integration rather than the same change.", + "best_pr_reason": "PR #45401 is the best representative PR: it has an explicit issue link, a tighter scope, and clearly documents the intended Voxtral TTS support. PR #44662 is a separate PenguinVL implementation and should not be grouped as the same change.", "soft_edge_verdicts": [ { "left": "pull_request:44662", "right": "pull_request:45401", "accept": false, - "reason": "Different underlying additions for different models and architectures. The overlap is limited to shared docs/auto-registration files, which is expected for new model support and not evidence of the same fix or change." + "reason": "Different models and different underlying functionality. The overlap is only shared Transformers scaffolding (docs, auto-mappings, package registration), not the same bug or change, so they should not be merged as duplicates." } ] }, @@ -25052,7 +25197,7 @@ "soft_pairs": [ "pull_request:44979|pull_request:45363" ], - "prepared_review_unit_hash": "cb305809eb723991d93072a4072247de0890896a2cd1066f408aee9593f75327", + "prepared_review_unit_hash": "40d8d2ccda757175996a7b5a715655e22f8acdf96fd72630b3f66cb75fb9857b", "component_budget": { "node_count": 2, "item_count": 2, @@ -25074,18 +25219,18 @@ "aggressively_trimmed": false, "split": false, "analyst_result": { - "summary": "These are related to fusion, but they target different abstractions: one adds a general module-fusion API, the other adds n-to-1 kernel fusion through KernelConfig. They are not the same underlying change, so I would not merge them as duplicates.", - "confidence": 0.89, + "summary": "These are related by theme (module/kernel fusion) but they are not the same change: one introduces a generic module-fusion API, the other wires up n-to-1 kernel fusion through KernelConfig and integration code paths.", + "confidence": 0.91, "canonical_issue_reason": null, - "canonical_pr_reason": "PR 44979 is the broader, more foundational change: it introduces the Module Fusion API and tests the core abstraction directly.", + "canonical_pr_reason": "PR 45363 is the better representative of this cluster: it implements a concrete end-to-end fusion workflow across modeling/integration code, whereas PR 44979 is a more foundational API proposal in a separate code path.", "best_issue_reason": null, - "best_pr_reason": "PR 44979 is the best representative of the cluster because it defines the reusable fusion infrastructure rather than a narrower KernelConfig implementation.", + "best_pr_reason": "PR 45363 is the most user-facing and complete of the two, with explicit KernelConfig integration and example usage; PR 44979 is broader infrastructure rather than the specific fused behavior.", "soft_edge_verdicts": [ { "left": "pull_request:44979", "right": "pull_request:45363", "accept": false, - "reason": "Both involve fusion, but they implement different mechanisms and code paths: a new module-fusion API versus kernel fusion via KernelConfig. They are related conceptually, not duplicates of the same change." + "reason": "Different underlying changes: 44979 adds a generic module-fusion API and tests, while 45363 implements n-to-1 kernel fusion via KernelConfig in integrations/modeling utilities. They are conceptually adjacent but not the same fix/feature and would not be merged as one PR." } ] }, @@ -25115,12 +25260,12 @@ "soft_pairs": [ "pull_request:45086|pull_request:45317" ], - "prepared_review_unit_hash": "7a04e015d604d541a60fee21267c0612aad48b15e37473a969169428a91b5c3f", + "prepared_review_unit_hash": "63d856062941e5b4b484b436b5b1167241b394c51f01f25ca638bfc18b700584", "component_budget": { "node_count": 2, "item_count": 2, "soft_pair_count": 1, - "serialized_chars": 4414, + "serialized_chars": 4413, "estimated_input_tokens": 1104, "estimated_eval_tokens": 2464 }, @@ -25128,7 +25273,7 @@ "node_count": 2, "item_count": 2, "soft_pair_count": 1, - "serialized_chars": 4414, + "serialized_chars": 4413, "estimated_input_tokens": 1104, "estimated_eval_tokens": 2464 }, @@ -25137,18 +25282,18 @@ "aggressively_trimmed": false, "split": false, "analyst_result": { - "summary": "Both pull requests address the same Mistral regex patch bug in `src/transformers/tokenization_utils_tokenizers.py`; the later PR adds tests and is the more complete candidate.", - "confidence": 0.97, + "summary": "Both pull requests patch the same Mistral regex pre-tokenizer AttributeError in the same file and target the same issue; 45317 is the more complete version because it includes tests.", + "confidence": 0.98, "canonical_issue_reason": null, - "canonical_pr_reason": "PR 45317 is the better canonical PR because it fixes the same `AttributeError` as 45086 and also includes test coverage, making it the more complete change.", + "canonical_pr_reason": "pull_request:45317 is the better canonical PR: it fixes the same AttributeError as 45086 and also adds test coverage, making it the more complete change set.", "best_issue_reason": null, - "best_pr_reason": "PR 45317 is the best standalone PR because it contains the same code fix as 45086 plus tests, so it is the more merge-ready and representative implementation.", + "best_pr_reason": "pull_request:45317 is the strongest representative of the cluster because it covers the full fix plus tests, while 45086 is the narrower code-only variant.", "soft_edge_verdicts": [ { "left": "pull_request:45086", "right": "pull_request:45317", "accept": true, - "reason": "Same underlying bug and same code path in `_patch_mistral_regex`; 45317 is essentially the expanded version of 45086 with tests, so they are mergeable as one fix." + "reason": "Same concrete bug fix in _patch_mistral_regex on the same code path, same target issue, and same changed file; 45317 is essentially the tested version of 45086." } ] }, @@ -25178,7 +25323,7 @@ "soft_pairs": [ "pull_request:45115|pull_request:45569" ], - "prepared_review_unit_hash": "0dc92f688a935dd77ace69dcc163462b129782f72f6a9bea2e352f99640767de", + "prepared_review_unit_hash": "2527672c54bb54a30aa71cc017871ea3c3fc6ec18a12cfceafbeafc7fb9a5b2d", "component_budget": { "node_count": 2, "item_count": 2, @@ -25200,18 +25345,18 @@ "aggressively_trimmed": false, "split": false, "analyst_result": { - "summary": "These are related NemotronH workstreams, but they are not duplicates: one PR is a narrow conversion/mapping and inheritance refactor, while the other is a broader model-family split into NemotronHDense/NemotronHSparse with docs and tests.", - "confidence": 0.87, + "summary": "Two distinct NemotronH PRs: one refactors NemotronH to inherit GraniteMoeHybrid and add conversion mapping, while the other splits NemotronH into Dense/Sparse variants with docs, auto-mapping, and tests. They overlap on shared model files but are not the same change.", + "confidence": 0.95, "canonical_issue_reason": null, - "canonical_pr_reason": "pull_request:45569 is the better canonical PR because it represents the larger, more complete NemotronH family restructuring and includes the public-facing model/docs/test changes.", + "canonical_pr_reason": "PR #45569 is the better canonical PR because it represents the larger, more user-facing feature work: introducing NemotronHDense/NemotronHSparse, updating docs/auto mappings, and adding tests. PR #45115 is a narrower refactor/conversion-mapping adjustment.", "best_issue_reason": null, - "best_pr_reason": "pull_request:45569 is the strongest standalone PR to keep: it covers the concrete model split and API/docs surface, not just internal conversion glue.", + "best_pr_reason": "PR #45569 is the best representative of the cluster since it covers the main architectural split and packaging changes. PR #45115 is related but only addresses a smaller inheritance/conversion refactor.", "soft_edge_verdicts": [ { "left": "pull_request:45115", "right": "pull_request:45569", "accept": false, - "reason": "Different underlying changes: 45115 focuses on GraniteMoeHybrid-compatible conversion/inheritance wiring for NemotronH, while 45569 introduces a broader NemotronH dense/sparse model split plus docs/tests. Shared model files alone are not enough to treat them as the same fix." + "reason": "Shared filenames indicate both touch NemotronH internals, but the concrete changes differ: #45115 refactors inheritance and conversion key renaming, while #45569 adds separate Dense/Sparse model variants plus docs and tests. They are not the same underlying fix/change and would not plausibly merge as one PR." } ] }, @@ -25241,40 +25386,40 @@ "soft_pairs": [ "pull_request:45332|pull_request:45333" ], - "prepared_review_unit_hash": "2bd9d0ad7f2579b45422940138c1487a60b3bd484a48cffa4ed35a1557c0cdf7", + "prepared_review_unit_hash": "d0a6182857cc95b1722edf81f70281e688fee164d42955df0c3673facee35649", "component_budget": { "node_count": 2, "item_count": 2, "soft_pair_count": 1, - "serialized_chars": 6081, - "estimated_input_tokens": 1521, - "estimated_eval_tokens": 3298 + "serialized_chars": 6080, + "estimated_input_tokens": 1520, + "estimated_eval_tokens": 3296 }, "budget": { "node_count": 2, "item_count": 2, "soft_pair_count": 1, - "serialized_chars": 6081, - "estimated_input_tokens": 1521, - "estimated_eval_tokens": 3298 + "serialized_chars": 6080, + "estimated_input_tokens": 1520, + "estimated_eval_tokens": 3296 }, "overflow_policy": "truncate_then_skip", "trimmed": false, "aggressively_trimmed": false, "split": false, "analyst_result": { - "summary": "Two related but distinct PRs: one is a broad heterogeneous model support feature, while the other is a narrower per-layer configuration-only subset. They overlap heavily in files and topic, but they do not look like the same mergeable fix/change.", - "confidence": 0.91, + "summary": "These are related but not duplicates: PR #45332 is a broad heterogeneous model feature PR, while PR #45333 is a narrower configuration-only slice of the same overall theme.", + "confidence": 0.92, "canonical_issue_reason": null, - "canonical_pr_reason": "PR #45332 is the better canonical representative because it is the broader umbrella change (per-layer config plus modeling/integration updates) and subsumes the narrower configuration-only work in #45333.", + "canonical_pr_reason": "PR #45332 is the better canonical candidate because it covers the full heterogeneous model feature set, including modeling, cache/masking/integration changes, not just per-layer configuration.", "best_issue_reason": null, - "best_pr_reason": "PR #45332 is the best overall representative of the cluster since it covers the full heterogeneous model support effort rather than just the configuration portion.", + "best_pr_reason": "PR #45332 is the strongest standalone PR because it contains the broader end-to-end implementation; PR #45333 is a narrower subset focused on configuration utilities.", "soft_edge_verdicts": [ { "left": "pull_request:45332", "right": "pull_request:45333", "accept": false, - "reason": "Related but not duplicates: #45333 is a focused configuration-only PR, while #45332 adds broader heterogeneous model support including modeling, cache, masking, accelerate/executorch integration, and extra tests. They share infrastructure but do not fix the same concrete code-path problem." + "reason": "They share the same feature area and some files, but #45332 is a much broader model-support PR while #45333 is configuration-only. They are not the same concrete change and would not plausibly collapse into one duplicate PR." } ] }, @@ -25304,12 +25449,12 @@ "soft_pairs": [ "pull_request:45415|pull_request:45425" ], - "prepared_review_unit_hash": "5f4bb802dc23d5cab158feb826584c7d2139c22dda961b44e769fcbd00e6f336", + "prepared_review_unit_hash": "d177ba845645d9b4bad36a93d7351a2d2de2645f54528ed08ac9dc99c5169719", "component_budget": { "node_count": 2, "item_count": 2, "soft_pair_count": 1, - "serialized_chars": 5514, + "serialized_chars": 5513, "estimated_input_tokens": 1379, "estimated_eval_tokens": 3014 }, @@ -25317,7 +25462,7 @@ "node_count": 2, "item_count": 2, "soft_pair_count": 1, - "serialized_chars": 5514, + "serialized_chars": 5513, "estimated_input_tokens": 1379, "estimated_eval_tokens": 3014 }, @@ -25326,18 +25471,18 @@ "aggressively_trimmed": false, "split": false, "analyst_result": { - "summary": "Both PRs are typing-related, but they address different scopes: one is a broad type-checking sweep across many `src/transformers/*` files, while the other adds shared typing protocols/aliases and related updates. They overlap in a few touched files, but not in the same concrete change set or bug fix.", - "confidence": 0.87, + "summary": "Both PRs are typing-related, but they address different change scopes: one is a broad type-checking cleanup across many Transformer modules, while the other adds/adjusts typing protocols and targeted fixes around `modeling_utils`/PEFT integration.", + "confidence": 0.88, "canonical_issue_reason": null, - "canonical_pr_reason": "PR #45425 is the more focused and self-contained typing change, centered on reusable type aliases/protocols plus a small set of integration updates.", + "canonical_pr_reason": "PR 45425 is the better canonical representative because it is more focused and clearly centered on concrete typing API additions and related fixes, rather than a sweeping repo-wide type-checking pass.", "best_issue_reason": null, - "best_pr_reason": "PR #45415 is the broader repository-wide type-checking pass and is the best single representative of the larger typing cleanup effort in this cluster.", + "best_pr_reason": "PR 45425 is the best standalone PR to keep: it has a narrower, more coherent purpose and is easier to evaluate/merge than the much broader type-checking cleanup in 45415.", "soft_edge_verdicts": [ { "left": "pull_request:45415", "right": "pull_request:45425", "accept": false, - "reason": "Overlap is thematic rather than duplicative: #45415 enables type checking across many modules and silences checker warnings, while #45425 introduces typing helper protocols/aliases and updates specific integration points. They do not look like the same underlying bug or one mergeable fix." + "reason": "Reject: these are not the same underlying change. 45415 is a broad typing cleanup across many files plus pyproject type-ignore config, while 45425 introduces typing helpers/protocols and targeted fixes in `_typing.py`, `modeling_utils`, PEFT, and flash-attention code. Shared typing files do not make them duplicate PRs." } ] }, @@ -25367,12 +25512,12 @@ "soft_pairs": [ "pull_request:45548|pull_request:45662" ], - "prepared_review_unit_hash": "5f4ca1ccb73dc638f8a0d807efb026af38eb87066a1c76da79ae3400e5463cc4", + "prepared_review_unit_hash": "d1b9a1f4a9d79bf968021b5960a0045ba48f17bb42a1bc61108f86dfd6e7738e", "component_budget": { "node_count": 2, "item_count": 2, "soft_pair_count": 1, - "serialized_chars": 4507, + "serialized_chars": 4508, "estimated_input_tokens": 1127, "estimated_eval_tokens": 2510 }, @@ -25380,7 +25525,7 @@ "node_count": 2, "item_count": 2, "soft_pair_count": 1, - "serialized_chars": 4507, + "serialized_chars": 4508, "estimated_input_tokens": 1127, "estimated_eval_tokens": 2510 }, @@ -25389,18 +25534,18 @@ "aggressively_trimmed": false, "split": false, "analyst_result": { - "summary": "These PRs are related to expert-parallel loading, but they fix different concrete bugs: one is about serializing `distributed_config` and detecting EP for DeepSpeed ZeRO-3/accelerate launch, while the other prevents FSDP2/rank-0 broadcast from overwriting experts during sharded loading. They should not be treated as duplicates.", - "confidence": 0.97, + "summary": "Two open PRs both touch expert-parallel loading paths, but they address different bugs: one fixes EP + DeepSpeed ZeRO-3 loading via accelerate launch, the other fixes expert parameters being overwritten under EP + FSDP2/rank-0 broadcast. Shared modeling_utils edits are not enough to make them duplicates.", + "confidence": 0.91, "canonical_issue_reason": null, - "canonical_pr_reason": "No single canonical PR for a duplicate cluster: the two PRs target different failure modes and code paths, so neither cleanly subsumes the other.", + "canonical_pr_reason": "PR #45662 is the better canonical representative for the cluster because it targets the more concrete expert-parallel overwrite failure and spans multiple relevant loading/integration paths.", "best_issue_reason": null, - "best_pr_reason": "PR #45662 is the stronger representative of the EP loading-corruption class because it addresses a concrete overwrite bug in the sharded load path and includes the broader fixes needed across loading and MoE/tensor-parallel integration.", + "best_pr_reason": "PR #45662 is the strongest standalone fix: it addresses a specific distributed-loading bug with a clear failure mode and broader code impact than the narrower accelerate-launch loading patch.", "soft_edge_verdicts": [ { "left": "pull_request:45548", "right": "pull_request:45662", "accept": false, - "reason": "Different bugs and code paths: #45548 serializes `distributed_config` and adds `has_ep` for DeepSpeed ZeRO-3/accelerate launch, while #45662 fixes expert parameters being overwritten by rank-0 broadcast during FSDP2 loading. They are related, but not the same underlying change." + "reason": "Reject: they concern different underlying EP distributed-loading issues and would not plausibly be merged into one PR without conflating separate fixes." } ] }, @@ -25430,12 +25575,12 @@ "soft_pairs": [ "pull_request:45639|pull_request:45645" ], - "prepared_review_unit_hash": "c3d151659068c24b5387eb3142f058124528cfcc59dd6f09a2578b37157b6655", + "prepared_review_unit_hash": "6b5666e9a2bfe0bcfd0f9e191c3b55d053cba0c0d332cc510d4235abab565075", "component_budget": { "node_count": 2, "item_count": 2, "soft_pair_count": 1, - "serialized_chars": 4672, + "serialized_chars": 4671, "estimated_input_tokens": 1168, "estimated_eval_tokens": 2592 }, @@ -25443,7 +25588,7 @@ "node_count": 2, "item_count": 2, "soft_pair_count": 1, - "serialized_chars": 4672, + "serialized_chars": 4671, "estimated_input_tokens": 1168, "estimated_eval_tokens": 2592 }, @@ -25452,18 +25597,18 @@ "aggressively_trimmed": false, "split": false, "analyst_result": { - "summary": "Both PRs address the same xdist/captured_info collision problem around patched testing debug logs. PR 45645 is the more complete fix because it includes the core code change plus CI workflow support for multiple log files.", - "confidence": 0.91, + "summary": "Both PRs target the same xdist-related captured_info logging collision around the testing debug artifact path; PR 45645 is the broader, more complete fix and subsumes the narrower file-write change in 45639.", + "confidence": 0.92, "canonical_issue_reason": null, - "canonical_pr_reason": "PR 45645 is the broader, more complete implementation: it fixes the output-file collision and also updates CI to collect all captured_info artifacts, so it subsumes the narrower core change in PR 45639.", + "canonical_pr_reason": "PR 45645 is the best canonical representative because it addresses the core xdist collision in `testing_utils.py` and also updates CI log collection to read all `captured_info*.txt` files, making it the more complete end-to-end fix.", "best_issue_reason": null, - "best_pr_reason": "PR 45645 is the strongest single representative because it addresses the underlying bug and the downstream log-preservation path, making it the most complete and useful merge candidate.", + "best_pr_reason": "PR 45645 is the strongest single artifact in the cluster: it fixes the underlying output-file collision and the downstream CI consumption path, whereas 45639 only covers the write-side change.", "soft_edge_verdicts": [ { "left": "pull_request:45639", "right": "pull_request:45645", "accept": true, - "reason": "Same underlying xdist-safe captured_info collision fix in the same testing code path; 45645 extends 45639 with CI log aggregation and is plausibly mergeable as one PR." + "reason": "Same underlying bug: xdist-safe handling of `captured_info` debug logs. 45645 includes the same `testing_utils.py` fix plus the workflow change needed to consume per-worker files, so they are merge-compatible parts of one change." } ] }, @@ -25559,12 +25704,12 @@ "soft_pairs": [ "pull_request:46126|pull_request:46269" ], - "prepared_review_unit_hash": "873cb26ebc63ff1495804a3d07f2a46f6d2a90c8224e4153bba5d59d5179d8be", + "prepared_review_unit_hash": "7d153538887ba8549ee47106bfa9a1123ed7ce9de7c749ac4c0e8f21c46b2cd3", "component_budget": { "node_count": 2, "item_count": 2, "soft_pair_count": 1, - "serialized_chars": 5525, + "serialized_chars": 5526, "estimated_input_tokens": 1382, "estimated_eval_tokens": 3020 }, @@ -25572,7 +25717,7 @@ "node_count": 2, "item_count": 2, "soft_pair_count": 1, - "serialized_chars": 5525, + "serialized_chars": 5526, "estimated_input_tokens": 1382, "estimated_eval_tokens": 3020 }, @@ -25581,18 +25726,18 @@ "aggressively_trimmed": false, "split": false, "analyst_result": { - "summary": "The two PRs overlap on a few distributed filenames, but they are not the same change: #46126 is a focused EP/Trainer integration PR, while #46269 is a broad distributed branch-base PR touching many unrelated areas. I would not deduplicate them.", - "confidence": 0.92, + "summary": "Two PRs touch some of the same distributed/MoE files, but they are not the same change: #46126 is a focused EP + Trainer integration, while #46269 is a broad distributed branch base with many unrelated edits. They should not be deduplicated.", + "confidence": 0.93, "canonical_issue_reason": null, - "canonical_pr_reason": "#46126 is the more specific, self-contained change; #46269 is an umbrella distributed branch-base PR with many unrelated edits, so it is not a good canonical duplicate target.", + "canonical_pr_reason": "#46126 is the more specific, self-contained feature PR with a clear concrete goal (EP + Trainer integration). #46269 is a broad branch-base PR spanning many unrelated distributed changes, so it is a worse canonical representative.", "best_issue_reason": null, - "best_pr_reason": "#46126 is the best representative PR for this cluster because it has a clear, concrete feature scope. #46269 is too broad and mixed to serve as a standalone duplicate representative.", + "best_pr_reason": "#46126 is the best PR candidate because it cleanly represents the underlying EP/Trainer feature work without the noise of unrelated branch-base churn. It is easier to triage and understand as the canonical change.", "soft_edge_verdicts": [ { "left": "pull_request:46126", "right": "pull_request:46269", "accept": false, - "reason": "Shared distributed files alone are not enough: #46126 implements EP + Trainer integration, while #46269 is a broad base branch with many distributed and CI changes. They do not appear to fix the same concrete code-path problem and would not plausibly be merged into one PR." + "reason": "Reject: shared filenames and vague distributed-topic overlap are not enough. #46269 is a broad base branch with many unrelated modifications, not the same concrete bug fix or feature as #46126, so they would not plausibly be merged into one PR." } ] }, @@ -25660,220 +25805,220 @@ "pull_request:46504|pull_request:46716", "pull_request:46432|pull_request:46716" ], - "prepared_review_unit_hash": "2662268fbd465018a947141e2111f45bb08cbcaae829b3d9c133cdabb5685d33", + "prepared_review_unit_hash": "57d2e4c49391c93d78699344d82a8682d498cb85564def9c7ed170c6550304e5", "component_budget": { "node_count": 10, "item_count": 10, "soft_pair_count": 31, - "serialized_chars": 26690, - "estimated_input_tokens": 6673, - "estimated_eval_tokens": 13602 + "serialized_chars": 26665, + "estimated_input_tokens": 6667, + "estimated_eval_tokens": 13590 }, "budget": { "node_count": 10, "item_count": 10, "soft_pair_count": 31, - "serialized_chars": 26690, - "estimated_input_tokens": 6673, - "estimated_eval_tokens": 13602 + "serialized_chars": 26665, + "estimated_input_tokens": 6667, + "estimated_eval_tokens": 13590 }, "overflow_policy": "truncate_then_skip", "trimmed": false, "aggressively_trimmed": false, "split": false, "analyst_result": { - "summary": "This cluster is a set of small workflow-maintenance PRs, but they touch different files and different concerns. The apparent similarities are mostly stylistic; none looks like the same underlying fix/change.", - "confidence": 0.97, + "summary": "These are mostly unrelated workflow-maintenance PRs. Despite similar titles and some shared files, the diffs touch different jobs and different fixes, so none of the soft edges look like true duplicates.", + "confidence": 0.95, "canonical_issue_reason": null, - "canonical_pr_reason": "No canonical PR: these are distinct workflow changes in different files or different parts of the same file, not one duplicated fix.", + "canonical_pr_reason": "No single PR is a true canonical duplicate anchor here; the set spans separate CI/workflow changes rather than one underlying fix.", "best_issue_reason": null, - "best_pr_reason": "No best PR for dedupe; the cluster is heterogeneous and the nearest-looking pairs are only superficially similar.", + "best_pr_reason": "PR 46554 is the most substantive and security-focused change, but it is still not a duplicate of the others.", "soft_edge_verdicts": [ { "left": "pull_request:46138", "right": "pull_request:46554", "accept": false, - "reason": "Different workflows and problems: stdin handling in self-comment CI vs security-gate checks in PR CI." + "reason": "Different workflow files and different changes; not the same underlying bug or fix." }, { "left": "pull_request:46138", "right": "pull_request:46558", "accept": false, - "reason": "Unrelated workflow changes in different files; not the same bug or fix." + "reason": "Different workflows and unrelated updates; only broad CI similarity." }, { "left": "pull_request:46138", "right": "pull_request:46563", "accept": false, - "reason": "Different workflow areas; one updates comment CI, the other changes PR CI security-gate wiring." + "reason": "Different workflows and unrelated updates; not mergeable as one fix." }, { "left": "pull_request:46138", "right": "pull_request:46716", "accept": false, - "reason": "Distinct workflow fixes in different contexts; not mergeable as one duplicate change." + "reason": "Different workflows and unrelated CI updates." }, { "left": "pull_request:46138", "right": "pull_request:46504", "accept": false, - "reason": "Both use printf, but they fix different scripts in different workflows, so not the same underlying change." + "reason": "Different workflow files; both are CI tweaks but fix different script behavior." }, { "left": "pull_request:46432", "right": "pull_request:46503", "accept": false, - "reason": "SSH-runner hardening vs benchmark workflow dispatch support; unrelated changes." + "reason": "Different workflow files and unrelated changes." }, { "left": "pull_request:46432", "right": "pull_request:46501", "accept": false, - "reason": "Different workflows and different concerns; no shared code-path bug." + "reason": "Different workflows; one is SSH runner hardening, the other is anti-slop permission adjustment." }, { "left": "pull_request:46432", "right": "pull_request:46502", "accept": false, - "reason": "SSH runner changes vs assign-reviewers checkout ref; unrelated." + "reason": "Different workflows and different operational concerns; not the same fix." }, { "left": "pull_request:46432", "right": "pull_request:46554", "accept": false, - "reason": "Different workflow fixes; security hardening in SSH runner is not the same as PR CI gate logic." + "reason": "Different workflows and unrelated changes." }, { "left": "pull_request:46432", "right": "pull_request:46558", "accept": false, - "reason": "Unrelated workflow edits in different files and code paths." + "reason": "Different workflows and unrelated changes." }, { "left": "pull_request:46501", "right": "pull_request:46503", "accept": false, - "reason": "Different workflow files and behaviors; not the same fix." + "reason": "Different workflows and different purposes; only both are CI config changes." }, { "left": "pull_request:46501", "right": "pull_request:46502", "accept": false, - "reason": "Separate workflow updates with no shared underlying issue." + "reason": "Different workflow files and unrelated fixes." }, { "left": "pull_request:46501", "right": "pull_request:46504", "accept": false, - "reason": "Different workflow maintenance changes; only broad similarity." + "reason": "Different workflow files; not the same underlying bug or change." }, { "left": "pull_request:46501", "right": "pull_request:46554", "accept": false, - "reason": "Anti-slop permission placement vs PR CI security-gate logic; unrelated." + "reason": "Different workflow files and different concerns." }, { "left": "pull_request:46502", "right": "pull_request:46503", "accept": false, - "reason": "Different workflows and changes; not a duplicate fix." + "reason": "Different workflow files and unrelated behavior changes." }, { "left": "pull_request:46502", "right": "pull_request:46504", "accept": false, - "reason": "Distinct workflow updates; no common bug/change." + "reason": "Different workflow files and different fixes." }, { "left": "pull_request:46502", "right": "pull_request:46554", "accept": false, - "reason": "Unrelated workflow modifications in different files." + "reason": "Different workflow files and unrelated changes." }, { "left": "pull_request:46503", "right": "pull_request:46504", "accept": false, - "reason": "Benchmark dispatch logic vs build-image shell quoting; different problems." + "reason": "Different workflow files and different CI triggers/scripts; not one fix." }, { "left": "pull_request:46503", "right": "pull_request:46554", "accept": false, - "reason": "Benchmark workflow trigger change is unrelated to PR CI security-gate checks." + "reason": "Different workflow files and unrelated changes." }, { "left": "pull_request:46504", "right": "pull_request:46554", "accept": false, - "reason": "Build-image tag detection quoting vs PR CI security gate; different code paths." + "reason": "Different workflow files and unrelated changes." }, { "left": "pull_request:46504", "right": "pull_request:46558", "accept": false, - "reason": "Different workflow fixes; no shared underlying issue." + "reason": "Different workflow files and unrelated fixes." }, { "left": "pull_request:46554", "right": "pull_request:46558", "accept": false, - "reason": "Same file, but different PR CI security-gate edits (checks vs permissions); not the same change." + "reason": "Same file, but different changes: one adds checkout-and-diff security checks, the other adjusts permissions/approval flow." }, { "left": "pull_request:46554", "right": "pull_request:46563", "accept": false, - "reason": "Same workflow file, but one adds checks and the other changes reusable-workflow comments/wiring." + "reason": "Same file, but different changes: security-check logic versus reusable-workflow commentary." }, { "left": "pull_request:46554", "right": "pull_request:46716", "accept": false, - "reason": "Both touch the security-gate workflow, but they fix different aspects (checks vs secrets handling)." + "reason": "Same file, but different changes: workflow logic rewrite versus secrets handling." }, { "left": "pull_request:46558", "right": "pull_request:46563", "accept": false, - "reason": "Same file, different edits; not the same underlying bug or change." + "reason": "Same workflow file, but the patches address different aspects of the gate and are not one concrete fix." }, { "left": "pull_request:46558", "right": "pull_request:46716", "accept": false, - "reason": "Permissions/approval logic vs secrets inheritance; distinct fixes." + "reason": "Same file, but different changes: permission/approval handling versus secrets inheritance." }, { "left": "pull_request:46563", "right": "pull_request:46716", "accept": false, - "reason": "Different PR CI security-gate adjustments; not mergeable as one duplicate PR." + "reason": "Same file, but different changes and different code paths." }, { "left": "pull_request:46504", "right": "pull_request:46563", "accept": false, - "reason": "Build-image shell handling vs PR CI reusable-workflow note; unrelated." + "reason": "Different workflow files and different CI behavior changes." }, { "left": "pull_request:46432", "right": "pull_request:46563", "accept": false, - "reason": "SSH-runner hardening vs PR CI reusable-workflow note; different code paths." + "reason": "Different workflow files and unrelated changes." }, { "left": "pull_request:46504", "right": "pull_request:46716", "accept": false, - "reason": "Build-image workflow quoting fix vs PR CI secrets change; unrelated." + "reason": "Different workflow files and unrelated changes." }, { "left": "pull_request:46432", "right": "pull_request:46716", "accept": false, - "reason": "SSH-runner hardening and PR CI secrets handling are separate changes." + "reason": "Different workflow files and unrelated changes." } ] }, @@ -25966,40 +26111,40 @@ "soft_pairs": [ "pull_request:46281|pull_request:46890" ], - "prepared_review_unit_hash": "e41f98c109b25148b1ff3c34c6d682ecf9d4656cbb5e5eaac73d2259d35c60df", + "prepared_review_unit_hash": "b5a044bef26c6d139cb7c86e609cceaa9a4d19efec2aa0f819288c607b0f4cfc", "component_budget": { "node_count": 2, "item_count": 2, "soft_pair_count": 1, - "serialized_chars": 4279, - "estimated_input_tokens": 1070, - "estimated_eval_tokens": 2396 + "serialized_chars": 4271, + "estimated_input_tokens": 1068, + "estimated_eval_tokens": 2392 }, "budget": { "node_count": 2, "item_count": 2, "soft_pair_count": 1, - "serialized_chars": 4279, - "estimated_input_tokens": 1070, - "estimated_eval_tokens": 2396 + "serialized_chars": 4271, + "estimated_input_tokens": 1068, + "estimated_eval_tokens": 2392 }, "overflow_policy": "truncate_then_skip", "trimmed": false, "aggressively_trimmed": false, "split": false, "analyst_result": { - "summary": "Two security-focused PRs, but they fix different path-traversal entry points: one in sharded checkpoint index loading, the other in `transformers_weights` filename resolution. They share a theme, not the same underlying bug.", - "confidence": 0.95, + "summary": "Both PRs harden local checkpoint/file resolution against path traversal, but they protect different inputs and code paths: sharded checkpoint `weight_map` handling in `hub.py` vs `transformers_weights` / explicit filename resolution in `modeling_utils.py`.", + "confidence": 0.91, "canonical_issue_reason": null, - "canonical_pr_reason": "No canonical PR: the PRs target different code paths, different inputs, and different validation logic, so they are not duplicates.", + "canonical_pr_reason": "PR 46281 is the cleaner representative of the cluster theme: it directly fixes untrusted shard filenames in the sharded-checkpoint loader with a narrowly scoped change.", "best_issue_reason": null, - "best_pr_reason": "PR 46281 is the clearest representative of the cluster\u2019s theme because it directly hardens sharded-checkpoint loading against untrusted `weight_map` paths and adds a focused test.", + "best_pr_reason": "PR 46890 is the broader security-hardening change and has more review activity, but it is still a separate code path from 46281.", "soft_edge_verdicts": [ { "left": "pull_request:46281", "right": "pull_request:46890", "accept": false, - "reason": "Both are path-traversal hardening PRs, but they address different concrete bugs in different loaders (`get_checkpoint_shard_files` vs `_get_resolved_checkpoint_files`) and different source fields (`weight_map` vs `transformers_weights`). They are related in theme, not the same mergeable fix." + "reason": "Reject: both are path-traversal fixes, but they target different loaders and different untrusted fields. The diffs touch separate functions/files and would not naturally merge into one PR." } ] }, @@ -26021,48 +26166,48 @@ "source": "llm", "cache_hit": false, "model": "gpt-5.4-mini?service_tier=flex", - "cluster_id": "cluster-46574-2", + "cluster_id": "cluster-46310-2", "nodes": [ - "pull_request:46574", - "pull_request:46580" + "pull_request:46310", + "pull_request:47162" ], "soft_pairs": [ - "pull_request:46574|pull_request:46580" + "pull_request:46310|pull_request:47162" ], - "prepared_review_unit_hash": "b85d007832412a0198d47cfbf633a72efb723e2573929dffdd48c9db7b770f37", + "prepared_review_unit_hash": "6fd7d2767fbdc41eaf510d2b0823e1de7cd3f6c8c4944044a16c52a9bc8ae603", "component_budget": { "node_count": 2, "item_count": 2, "soft_pair_count": 1, - "serialized_chars": 4410, - "estimated_input_tokens": 1103, - "estimated_eval_tokens": 2462 + "serialized_chars": 4883, + "estimated_input_tokens": 1221, + "estimated_eval_tokens": 2698 }, "budget": { "node_count": 2, "item_count": 2, "soft_pair_count": 1, - "serialized_chars": 4410, - "estimated_input_tokens": 1103, - "estimated_eval_tokens": 2462 + "serialized_chars": 4883, + "estimated_input_tokens": 1221, + "estimated_eval_tokens": 2698 }, "overflow_policy": "truncate_then_skip", "trimmed": false, "aggressively_trimmed": false, "split": false, "analyst_result": { - "summary": "These are two unrelated open PRs: one fixes CodeLlama decode-time stripping of real leading spaces, the other fixes DeepSeek-Coder/DeepSeek-LLM being loaded with the wrong tokenizer class. They should not be clustered as duplicates.", - "confidence": 0.97, - "canonical_issue_reason": null, - "canonical_pr_reason": "No single canonical PR: the PRs address different model families and different code paths, so neither can stand in for the other.", - "best_issue_reason": null, - "best_pr_reason": "Neither PR is a good global best representative for the cluster because they are not variants of the same fix; each solves a separate tokenizer-loading bug.", + "summary": "Both PRs address the same off-by-one bug in MoonshineStreaming\u2019s right-side sliding window mask; PR 47162 is the broader, more complete fix and subsumes PR 46310.", + "confidence": 0.94, + "canonical_issue_reason": "Issue 46305 appears to be the underlying off-by-one right-window boundary bug in `sliding_window_mask_function` / `inner_mask`.", + "canonical_pr_reason": "PR 47162 is the better canonical PR because it fixes the same MoonshineStreaming bug, includes tests, and appears to supersede the narrower PR 46310 while also applying the same correction to Gemma4\u2019s analogous mask code.", + "best_issue_reason": "The shared issue target and identical boundary change strongly indicate a single underlying bug: the right window should include the boundary (`<=`) instead of excluding it (`<`).", + "best_pr_reason": "PR 47162 best represents the fix set because it contains the MoonshineStreaming correction plus validation coverage and broader related cleanup, whereas PR 46310 is only the narrower MoonshineStreaming-only version.", "soft_edge_verdicts": [ { - "left": "pull_request:46574", - "right": "pull_request:46580", - "accept": false, - "reason": "Different underlying problems: CodeLlama decoder stripping on decode vs. DeepSeek checkpoint/tokenizer-class routing. They do not look mergeable into one PR." + "left": "pull_request:46310", + "right": "pull_request:47162", + "accept": true, + "reason": "Both PRs fix the same concrete off-by-one in MoonshineStreaming\u2019s right-window mask; 47162 includes the same change and can plausibly subsume 46310." } ] }, @@ -26084,48 +26229,48 @@ "source": "llm", "cache_hit": false, "model": "gpt-5.4-mini?service_tier=flex", - "cluster_id": "cluster-46773-2", + "cluster_id": "cluster-46574-2", "nodes": [ - "pull_request:46773", - "pull_request:47087" + "pull_request:46574", + "pull_request:46580" ], "soft_pairs": [ - "pull_request:46773|pull_request:47087" + "pull_request:46574|pull_request:46580" ], - "prepared_review_unit_hash": "8e29d0dd48774fa7967515143dcf7d11298777d6756d25d875e89664c6b1fc7c", + "prepared_review_unit_hash": "c30efa054957a9a0124b77df056586692208c785459d72940037b4db9c221faf", "component_budget": { "node_count": 2, "item_count": 2, "soft_pair_count": 1, - "serialized_chars": 5432, - "estimated_input_tokens": 1358, - "estimated_eval_tokens": 2972 + "serialized_chars": 4410, + "estimated_input_tokens": 1103, + "estimated_eval_tokens": 2462 }, "budget": { "node_count": 2, "item_count": 2, "soft_pair_count": 1, - "serialized_chars": 5432, - "estimated_input_tokens": 1358, - "estimated_eval_tokens": 2972 + "serialized_chars": 4410, + "estimated_input_tokens": 1103, + "estimated_eval_tokens": 2462 }, "overflow_policy": "truncate_then_skip", "trimmed": false, "aggressively_trimmed": false, "split": false, "analyst_result": { - "summary": "These two PRs both touch masking logic, but they fix different bugs in different code paths: one is a batch-size-1 padding-mask condition in linear-attention model forwards, the other is recurrent-mask handling during continued forwards/chunked prefill. They are not duplicates and should not be merged.", - "confidence": 0.93, + "summary": "These two pull requests are unrelated fixes: one changes CodeLlama decode behavior to preserve real leading spaces, while the other adjusts tokenizer auto-selection for certain DeepSeek checkpoints. They should not be clustered as duplicates.", + "confidence": 0.98, "canonical_issue_reason": null, - "canonical_pr_reason": "PR #47087 is the stronger canonical PR candidate because it has an explicit issue target, a focused description of a single recurrent-mask bug, and its diff is centered in one utility path with supporting tests.", + "canonical_pr_reason": "No single canonical PR: the two PRs address different bugs in different code paths (CodeLlama decoding vs. DeepSeek tokenizer-class selection).", "best_issue_reason": null, - "best_pr_reason": "PR #47087 is the best representative of the cluster only in the sense that it is the more self-contained, explicitly scoped fix; however, it is still distinct from PR #46773 and should not be treated as the same change.", + "best_pr_reason": "Neither PR is a better representative of the other; they solve separate issues and are not mergeable into one fix.", "soft_edge_verdicts": [ { - "left": "pull_request:46773", - "right": "pull_request:47087", + "left": "pull_request:46574", + "right": "pull_request:46580", "accept": false, - "reason": "Both concern padding masks, but the concrete failures differ: #46773 fixes a batch-size-1 check in linear-attention model code, while #47087 fixes recurrent-layer masking on continued forwards via masking_utils. They would not plausibly be merged into one PR as the same bug fix." + "reason": "Different underlying problems and code paths: CodeLlama decoder stripping on decode versus DeepSeek auto-tokenizer misclassification. They do not appear to be the same change." } ] }, @@ -26147,56 +26292,48 @@ "source": "llm", "cache_hit": false, "model": "gpt-5.4-mini?service_tier=flex", - "cluster_id": "cluster-46781-3", + "cluster_id": "cluster-46773-2", "nodes": [ - "pull_request:46781", - "pull_request:46993", - "pull_request:46994" + "pull_request:46773", + "pull_request:47087" ], "soft_pairs": [ - "pull_request:46781|pull_request:46993", - "pull_request:46993|pull_request:46994" + "pull_request:46773|pull_request:47087" ], - "prepared_review_unit_hash": "1d0b2d6edbace5379f045855db33668f0714970d22fc7e9ce92c004a69e280cb", + "prepared_review_unit_hash": "68095d6ed89285c70a64f4ba626a139c6df548ebb72a6fecf54b4bd5c4d59561", "component_budget": { - "node_count": 3, - "item_count": 3, - "soft_pair_count": 2, - "serialized_chars": 6604, - "estimated_input_tokens": 1651, - "estimated_eval_tokens": 3558 + "node_count": 2, + "item_count": 2, + "soft_pair_count": 1, + "serialized_chars": 5432, + "estimated_input_tokens": 1358, + "estimated_eval_tokens": 2972 }, "budget": { - "node_count": 3, - "item_count": 3, - "soft_pair_count": 2, - "serialized_chars": 6604, - "estimated_input_tokens": 1651, - "estimated_eval_tokens": 3558 + "node_count": 2, + "item_count": 2, + "soft_pair_count": 1, + "serialized_chars": 5432, + "estimated_input_tokens": 1358, + "estimated_eval_tokens": 2972 }, "overflow_policy": "truncate_then_skip", "trimmed": false, "aggressively_trimmed": false, "split": false, "analyst_result": { - "summary": "These three pull requests are all in generation/cache code, but they fix different bugs: cache layer validation, prompt lookup decoding length handling, and empty-list validation for logits processors. They should not be deduplicated into one change.", + "summary": "These two PRs are related by theme (masking being skipped) but they fix different bugs in different code paths: 46773 targets batch-size-1 handling in linear-attention model forwards, while 47087 fixes recurrent-layer masking during continued/chunked forwards and cache continuation. They should not be deduplicated.", "confidence": 0.97, "canonical_issue_reason": null, - "canonical_pr_reason": "No single canonical PR: the three PRs target different code paths and bug classes, so there is no true duplicate representative among them.", + "canonical_pr_reason": "No single canonical PR: 46773 and 47087 address different masking regressions in different implementations, so neither can stand in for the other.", "best_issue_reason": null, - "best_pr_reason": "No best PR for dedupe purposes; each PR is a separate, self-contained fix and none subsumes the others.", + "best_pr_reason": "47087 is the stronger standalone PR: it has an explicit issue link, a focused utility-level fix, and broader test coverage around the recurrent-mask continuation bug.", "soft_edge_verdicts": [ { - "left": "pull_request:46781", - "right": "pull_request:46993", - "accept": false, - "reason": "Different bugs and code paths: QuantizedCache layer-type validation vs. prompt lookup decoding max_length accounting." - }, - { - "left": "pull_request:46993", - "right": "pull_request:46994", + "left": "pull_request:46773", + "right": "pull_request:47087", "accept": false, - "reason": "Both are generation-related, but one fixes prompt lookup candidate truncation while the other tightens validation for empty bad_words_ids/sequence_bias inputs." + "reason": "Same broad masking topic, but the underlying bugs and code paths differ: batch-size-1 linear-attention masking vs. recurrent continued-forward/cache-continuation masking." } ] }, @@ -26218,86 +26355,48 @@ "source": "llm", "cache_hit": false, "model": "gpt-5.4-mini?service_tier=flex", - "cluster_id": "cluster-47032-5", + "cluster_id": "cluster-47074-2", "nodes": [ - "pull_request:47032", - "pull_request:47051", - "pull_request:47060", "pull_request:47074", "pull_request:47076" ], "soft_pairs": [ - "pull_request:47032|pull_request:47051", - "pull_request:47051|pull_request:47074", - "pull_request:47051|pull_request:47060", - "pull_request:47060|pull_request:47074", - "pull_request:47060|pull_request:47076", "pull_request:47074|pull_request:47076" ], - "prepared_review_unit_hash": "6b69e2b13b7f825ffcb984f161ad72e47767f58a09f2eb60d1918709419ef1ab", + "prepared_review_unit_hash": "668ed5c6efe479c988678593021c5202e09f2218f8f2c6b0006d36060fead4e1", "component_budget": { - "node_count": 5, - "item_count": 5, - "soft_pair_count": 6, - "serialized_chars": 12764, - "estimated_input_tokens": 3191, - "estimated_eval_tokens": 6638 + "node_count": 2, + "item_count": 2, + "soft_pair_count": 1, + "serialized_chars": 4486, + "estimated_input_tokens": 1122, + "estimated_eval_tokens": 2500 }, "budget": { - "node_count": 5, - "item_count": 5, - "soft_pair_count": 6, - "serialized_chars": 12764, - "estimated_input_tokens": 3191, - "estimated_eval_tokens": 6638 + "node_count": 2, + "item_count": 2, + "soft_pair_count": 1, + "serialized_chars": 4486, + "estimated_input_tokens": 1122, + "estimated_eval_tokens": 2500 }, "overflow_policy": "truncate_then_skip", "trimmed": false, "aggressively_trimmed": false, "split": false, "analyst_result": { - "summary": "Mostly unrelated multimodal-generation fixes across different pipelines/models. There is some thematic overlap in generation-kwarg plumbing, but no pair clearly looks like the same underlying bug/change; the closest is 47032 vs 47051, yet they still address different fixes.", - "confidence": 0.9, + "summary": "These are two unrelated pull requests in different multimodal code paths: one fixes a key mismatch in AnyToAnyPipeline, the other adds Gemma4 placeholder validation for audio/text inputs. They are not duplicates or mergeable into one change.", + "confidence": 0.97, "canonical_issue_reason": null, - "canonical_pr_reason": "No single PR cleanly subsumes the cluster; if a representative is needed, 47051 is the broadest pipeline-level change in the common multimodal-generation area.", + "canonical_pr_reason": "No canonical PR: 47074 and 47076 fix different bugs in different components, with no shared file targets or concrete code-path overlap.", "best_issue_reason": null, - "best_pr_reason": "47051 is the best anchor only in a thematic sense because it spans multimodal generation pipelines, but it is not a duplicate umbrella for the other PRs.", + "best_pr_reason": "47076 is the stronger standalone fix in this cluster because it addresses a specific input-validation bug and includes companion test coverage; 47074 is a smaller pipeline key-normalization fix.", "soft_edge_verdicts": [ - { - "left": "pull_request:47032", - "right": "pull_request:47051", - "accept": false, - "reason": "Related image-text-to-text generation plumbing, but 47032 is a specific stop_sequence fix while 47051 is broader kwargs forwarding; not the same concrete bug." - }, - { - "left": "pull_request:47051", - "right": "pull_request:47074", - "accept": false, - "reason": "Both touch AnyToAnyPipeline, but one forwards generation kwargs and the other normalizes input keys; different code-path problems." - }, - { - "left": "pull_request:47051", - "right": "pull_request:47060", - "accept": false, - "reason": "Different subsystems and fixes: multimodal pipeline kwargs forwarding vs Qwen Omni batched post-processing." - }, - { - "left": "pull_request:47060", - "right": "pull_request:47074", - "accept": false, - "reason": "Qwen Omni decoding bug versus AnyToAnyPipeline input-key normalization; unrelated concrete changes." - }, - { - "left": "pull_request:47060", - "right": "pull_request:47076", - "accept": false, - "reason": "Different models and failure modes: Qwen post-processing versus Gemma4 placeholder validation." - }, { "left": "pull_request:47074", "right": "pull_request:47076", "accept": false, - "reason": "Different pipeline/model areas and distinct bugs; only broad multimodal similarity, not a duplicate change." + "reason": "Different underlying problems and code paths: 47074 normalizes the AnyToAnyPipeline payload key from images to videos, while 47076 validates Gemma4 audio placeholder batching. They do not appear to be the same bug or a single combined change." } ] }, @@ -26327,40 +26426,40 @@ "soft_pairs": [ "pull_request:47105|pull_request:47129" ], - "prepared_review_unit_hash": "2661c15230d6851013f935f92a5d1027acf6afdef297b8565c2268c09745a275", + "prepared_review_unit_hash": "27359829971ccb171755fa388e49d64d2b6c965d0a0a873579660fe8212f0b9d", "component_budget": { "node_count": 2, "item_count": 2, "soft_pair_count": 1, - "serialized_chars": 4171, - "estimated_input_tokens": 1043, - "estimated_eval_tokens": 2342 + "serialized_chars": 4182, + "estimated_input_tokens": 1046, + "estimated_eval_tokens": 2348 }, "budget": { "node_count": 2, "item_count": 2, "soft_pair_count": 1, - "serialized_chars": 4171, - "estimated_input_tokens": 1043, - "estimated_eval_tokens": 2342 + "serialized_chars": 4182, + "estimated_input_tokens": 1046, + "estimated_eval_tokens": 2348 }, "overflow_policy": "truncate_then_skip", "trimmed": false, "aggressively_trimmed": false, "split": false, "analyst_result": { - "summary": "These two pull requests are not duplicates: one re-enables a VITS model forward test, while the other updates the RT-DETR v2 weight converter for changed model key names. They touch different models, different code paths, and solve unrelated problems.", + "summary": "These two PRs are unrelated: one re-enables a VITS forward test, the other fixes an RT-DETR v2 weight converter mapping issue. The soft similarity is superficial and should not be merged as a duplicate cluster.", "confidence": 0.98, "canonical_issue_reason": null, - "canonical_pr_reason": null, + "canonical_pr_reason": "No canonical PR: the changes target different models/code paths and do not appear to be the same underlying fix.", "best_issue_reason": null, - "best_pr_reason": "If a single representative is needed, PR #47129 is the stronger standalone change because it fixes an actual converter/code-path compatibility issue rather than only adjusting test expectations.", + "best_pr_reason": "Neither PR is a good global representative for the cluster because they solve unrelated problems in different subsystems.", "soft_edge_verdicts": [ { "left": "pull_request:47105", "right": "pull_request:47129", "accept": false, - "reason": "Different subsystems and bug types: VITS test re-enable/shape assertions vs RT-DETR v2 converter key mapping. They do not appear mergeable into one PR or to address the same underlying issue." + "reason": "Different artifacts and different fixes: VITS test expectations vs RT-DETR v2 converter key mapping. They do not look mergeable into one PR." } ] }, @@ -26397,12 +26496,12 @@ "pull_request:47123|pull_request:47150", "pull_request:47124|pull_request:47150" ], - "prepared_review_unit_hash": "0356e660d12bcac6e74c1efdabb6c2bb591f24f6e0a4bae3b205bb1b5f0a76a5", + "prepared_review_unit_hash": "bbba2464a29c84a8a7f080bd010182fd91438f350ddf7586b2dbf2c3c2569d09", "component_budget": { "node_count": 4, "item_count": 4, "soft_pair_count": 6, - "serialized_chars": 9332, + "serialized_chars": 9331, "estimated_input_tokens": 2333, "estimated_eval_tokens": 4922 }, @@ -26410,7 +26509,7 @@ "node_count": 4, "item_count": 4, "soft_pair_count": 6, - "serialized_chars": 9332, + "serialized_chars": 9331, "estimated_input_tokens": 2333, "estimated_eval_tokens": 4922 }, @@ -26419,48 +26518,48 @@ "aggressively_trimmed": false, "split": false, "analyst_result": { - "summary": "These PRs are not duplicates of one another: they cover different models and different failure modes (Edgetam kwarg handling, MiniMax-M3-VL test loading, Falcon-Mamba expected outputs, and Whisper generation length handling). Only the Whisper and Edgetam items touch source code; the others are test-only fixes.", - "confidence": 0.94, + "summary": "All four PRs target different models and different failure modes, so they are not duplicates of one another. The only code-change PR is the Whisper fix; the others are test/environment updates for unrelated models.", + "confidence": 0.97, "canonical_issue_reason": null, - "canonical_pr_reason": "PR 47150 is the strongest standalone change in the set: it fixes a concrete Whisper generation code path by updating max_length during conditioned decoding. The others are model-specific test adjustments or unrelated fixes.", + "canonical_pr_reason": "PR 47150 is the strongest representative of the set because it fixes an actual library code path in Whisper, whereas the others are model-specific test or environment adjustments. It is not a duplicate of the others, but it is the most substantive single change here.", "best_issue_reason": null, - "best_pr_reason": "PR 47150 is the best representative of a real bug fix because it changes production code rather than test expectations or test setup, and it has a clear, self-contained behavior change.", + "best_pr_reason": "PR 47150 is the best standalone PR in this cluster: it changes core Whisper generation logic to address a concrete bug, while the other PRs only adjust tests or loading setup for different models.", "soft_edge_verdicts": [ { "left": "pull_request:47122", "right": "pull_request:47123", "accept": false, - "reason": "Different models and different problems: Edgetam is a runtime kwarg incompatibility in model code, while MiniMax-M3-VL is a test loading/offload setup issue." + "reason": "Different models and different fixes: Edgetam drops an unexpected `original_sizes` kwarg in model forward, while MiniMax-M3-VL adjusts test loading with `offload_folder`. Not the same bug or change." }, { "left": "pull_request:47122", "right": "pull_request:47124", "accept": false, - "reason": "Edgetam fixes a source-code kwarg crash; Falcon-Mamba only updates expected test outputs. These are not the same underlying bug." + "reason": "Different subsystems and failure types: Edgetam fixes a backbone kwarg mismatch; Falcon-Mamba only updates expected generation outputs in tests." }, { "left": "pull_request:47122", "right": "pull_request:47150", "accept": false, - "reason": "Both touch model code, but they are unrelated code paths in different models with different failure modes; not mergeable as one fix." + "reason": "Unrelated model bugs: Edgetam input-kwarg cleanup versus Whisper max-length handling for segmented generation." }, { "left": "pull_request:47123", "right": "pull_request:47124", "accept": false, - "reason": "Both are test-only PRs, but one addresses loading/offload setup and the other changes generation expectations for a different model." + "reason": "Both are test-side changes, but for different models and different failure causes (`load_error` vs `output_mismatch`). No shared underlying bug." }, { "left": "pull_request:47123", "right": "pull_request:47150", "accept": false, - "reason": "MiniMax-M3-VL test loading issue is unrelated to Whisper generation length logic." + "reason": "MiniMax-M3-VL test loading workaround is unrelated to Whisper generation-length logic." }, { "left": "pull_request:47124", "right": "pull_request:47150", "accept": false, - "reason": "Falcon-Mamba expected-output updates are not the same underlying bug as Whisper's max_length handling in generation." + "reason": "Falcon-Mamba expected-output refresh is a test expectation update, not the same as Whisper\u2019s source-code fix for max_length handling." } ] }, diff --git a/analysis/current/manifest.json b/analysis/current/manifest.json index da0eaec41f0d68654ef912ea73eae70450069ee4..ad287822751cd0dabd0f71d5080ef4333f92cb33 100644 --- a/analysis/current/manifest.json +++ b/analysis/current/manifest.json @@ -1,8 +1,8 @@ { - "analysis_id": "hybrid-model-20260708t060015z", + "analysis_id": "hybrid-model-20260708t120111z", "archived_artifacts": { - "hybrid": "snapshots/20260708T060015Z/analysis-runs/hybrid-model-20260708t060015z/analysis-report-hybrid.json", - "hybrid_reviews": "snapshots/20260708T060015Z/analysis-runs/hybrid-model-20260708t060015z/analysis-report-hybrid.llm-reviews.json" + "hybrid": "snapshots/20260708T120111Z/analysis-runs/hybrid-model-20260708t120111z/analysis-report-hybrid.json", + "hybrid_reviews": "snapshots/20260708T120111Z/analysis-runs/hybrid-model-20260708t120111z/analysis-report-hybrid.llm-reviews.json" }, "artifacts": { "hybrid": "analysis/current/analysis-report-hybrid.json", @@ -10,9 +10,9 @@ }, "channel": "canonical", "model": null, - "published_at": "2026-07-08T06:13:00Z", + "published_at": "2026-07-08T12:14:09Z", "repo": "huggingface/transformers", "schema_version": 1, - "snapshot_id": "20260708T060015Z", + "snapshot_id": "20260708T120111Z", "variant": "hybrid" } diff --git a/snapshots/20260708T120111Z/analysis-runs/hybrid-model-20260708t120111z/analysis-report-hybrid.json b/snapshots/20260708T120111Z/analysis-runs/hybrid-model-20260708t120111z/analysis-report-hybrid.json new file mode 100644 index 0000000000000000000000000000000000000000..667dfe84a5731abfff1f848f6939f9b75f6eab97 --- /dev/null +++ b/snapshots/20260708T120111Z/analysis-runs/hybrid-model-20260708t120111z/analysis-report-hybrid.json @@ -0,0 +1,2901 @@ +{ + "schema_version": "1.0", + "repo": "huggingface/transformers", + "snapshot_id": "20260708T120111Z", + "generated_at": "2026-07-08T12:14:07Z", + "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-43451-6", + "summary": "Cluster of 6 related pull requests linked by soft_similarity.", + "status": "open", + "confidence": 0.5, + "canonical_issue_number": null, + "canonical_pr_number": 43451, + "issue_numbers": [], + "pr_numbers": [ + 43451, + 46912, + 47035, + 47096, + 47163, + 47166 + ], + "evidence_types": [ + "soft_similarity" + ], + "pr_comparisons": [ + { + "left_pr_number": 43451, + "right_pr_number": 46912, + "code_similarity": 0.108, + "size_similarity": 0.441, + "file_overlap": 0.0, + "area_overlap": 0.0, + "patch_similarity": 0.134, + "shared_filenames": [], + "shared_file_areas": [] + }, + { + "left_pr_number": 43451, + "right_pr_number": 47035, + "code_similarity": 0.001, + "size_similarity": 0.002, + "file_overlap": 0.0, + "area_overlap": 0.0, + "patch_similarity": 0.001, + "shared_filenames": [], + "shared_file_areas": [] + }, + { + "left_pr_number": 43451, + "right_pr_number": 47096, + "code_similarity": 0.036, + "size_similarity": 0.099, + "file_overlap": 0.0, + "area_overlap": 0.0, + "patch_similarity": 0.108, + "shared_filenames": [], + "shared_file_areas": [] + }, + { + "left_pr_number": 43451, + "right_pr_number": 47163, + "code_similarity": 0.004, + "size_similarity": 0.005, + "file_overlap": 0.0, + "area_overlap": 0.0, + "patch_similarity": 0.02, + "shared_filenames": [], + "shared_file_areas": [] + }, + { + "left_pr_number": 43451, + "right_pr_number": 47166, + "code_similarity": 0.004, + "size_similarity": 0.011, + "file_overlap": 0.0, + "area_overlap": 0.0, + "patch_similarity": 0.013, + "shared_filenames": [], + "shared_file_areas": [] + }, + { + "left_pr_number": 46912, + "right_pr_number": 47035, + "code_similarity": 0.001, + "size_similarity": 0.004, + "file_overlap": 0.0, + "area_overlap": 0.0, + "patch_similarity": 0.002, + "shared_filenames": [], + "shared_file_areas": [] + }, + { + "left_pr_number": 46912, + "right_pr_number": 47096, + "code_similarity": 0.06, + "size_similarity": 0.224, + "file_overlap": 0.0, + "area_overlap": 0.0, + "patch_similarity": 0.1, + "shared_filenames": [], + "shared_file_areas": [] + }, + { + "left_pr_number": 46912, + "right_pr_number": 47163, + "code_similarity": 0.005, + "size_similarity": 0.01, + "file_overlap": 0.0, + "area_overlap": 0.0, + "patch_similarity": 0.022, + "shared_filenames": [], + "shared_file_areas": [] + }, + { + "left_pr_number": 46912, + "right_pr_number": 47166, + "code_similarity": 0.008, + "size_similarity": 0.025, + "file_overlap": 0.0, + "area_overlap": 0.0, + "patch_similarity": 0.017, + "shared_filenames": [], + "shared_file_areas": [] + }, + { + "left_pr_number": 47035, + "right_pr_number": 47096, + "code_similarity": 0.004, + "size_similarity": 0.019, + "file_overlap": 0.0, + "area_overlap": 0.0, + "patch_similarity": 0.004, + "shared_filenames": [], + "shared_file_areas": [] + }, + { + "left_pr_number": 47035, + "right_pr_number": 47163, + "code_similarity": 0.083, + "size_similarity": 0.414, + "file_overlap": 0.0, + "area_overlap": 0.0, + "patch_similarity": 0.0, + "shared_filenames": [], + "shared_file_areas": [] + }, + { + "left_pr_number": 47035, + "right_pr_number": 47166, + "code_similarity": 0.039, + "size_similarity": 0.171, + "file_overlap": 0.0, + "area_overlap": 0.0, + "patch_similarity": 0.033, + "shared_filenames": [], + "shared_file_areas": [] + }, + { + "left_pr_number": 47096, + "right_pr_number": 47163, + "code_similarity": 0.013, + "size_similarity": 0.046, + "file_overlap": 0.0, + "area_overlap": 0.0, + "patch_similarity": 0.023, + "shared_filenames": [], + "shared_file_areas": [] + }, + { + "left_pr_number": 47096, + "right_pr_number": 47166, + "code_similarity": 0.025, + "size_similarity": 0.112, + "file_overlap": 0.0, + "area_overlap": 0.0, + "patch_similarity": 0.016, + "shared_filenames": [], + "shared_file_areas": [] + }, + { + "left_pr_number": 47163, + "right_pr_number": 47166, + "code_similarity": 0.089, + "size_similarity": 0.414, + "file_overlap": 0.0, + "area_overlap": 0.0, + "patch_similarity": 0.043, + "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-46894-5", + "summary": "Cluster of 2 issues and 3 PRs centered on issue #46894.", + "status": "open", + "confidence": 0.8, + "canonical_issue_number": 46894, + "canonical_pr_number": 46895, + "issue_numbers": [ + 46894, + 46897 + ], + "pr_numbers": [ + 46895, + 46958, + 47090 + ], + "evidence_types": [ + "closing_reference", + "shared_issue_target", + "soft_similarity" + ], + "pr_comparisons": [ + { + "left_pr_number": 46895, + "right_pr_number": 46958, + "code_similarity": 0.779, + "size_similarity": 0.915, + "file_overlap": 1.0, + "area_overlap": 0.5, + "patch_similarity": 0.81, + "shared_filenames": [ + "src/transformers/models/cohere_asr/modeling_cohere_asr.py", + "src/transformers/models/cohere_asr/modular_cohere_asr.py", + "tests/models/cohere_asr/test_modeling_cohere_asr.py" + ], + "shared_file_areas": [ + { + "filename": "src/transformers/models/cohere_asr/modeling_cohere_asr.py", + "left_ranges": [ + [ + 637, + 650 + ] + ], + "right_ranges": [ + [ + 22, + 28 + ], + [ + 638, + 645 + ] + ] + }, + { + "filename": "src/transformers/models/cohere_asr/modular_cohere_asr.py", + "left_ranges": [ + [ + 500, + 513 + ] + ], + "right_ranges": [ + [ + 16, + 22 + ], + [ + 501, + 508 + ] + ] + }, + { + "filename": "tests/models/cohere_asr/test_modeling_cohere_asr.py", + "left_ranges": [ + [ + 35, + 41 + ], + [ + 176, + 214 + ] + ], + "right_ranges": [ + [ + 15, + 21 + ], + [ + 180, + 221 + ] + ] + } + ] + }, + { + "left_pr_number": 46895, + "right_pr_number": 47090, + "code_similarity": 0.819, + "size_similarity": 0.932, + "file_overlap": 1.0, + "area_overlap": 0.783, + "patch_similarity": 0.389, + "shared_filenames": [ + "src/transformers/models/cohere_asr/modeling_cohere_asr.py", + "src/transformers/models/cohere_asr/modular_cohere_asr.py", + "tests/models/cohere_asr/test_modeling_cohere_asr.py" + ], + "shared_file_areas": [ + { + "filename": "src/transformers/models/cohere_asr/modeling_cohere_asr.py", + "left_ranges": [ + [ + 637, + 650 + ] + ], + "right_ranges": [ + [ + 637, + 649 + ] + ] + }, + { + "filename": "src/transformers/models/cohere_asr/modular_cohere_asr.py", + "left_ranges": [ + [ + 500, + 513 + ] + ], + "right_ranges": [ + [ + 500, + 512 + ] + ] + }, + { + "filename": "tests/models/cohere_asr/test_modeling_cohere_asr.py", + "left_ranges": [ + [ + 35, + 41 + ], + [ + 176, + 214 + ] + ], + "right_ranges": [ + [ + 175, + 222 + ] + ] + } + ] + }, + { + "left_pr_number": 46958, + "right_pr_number": 47090, + "code_similarity": 0.742, + "size_similarity": 0.852, + "file_overlap": 1.0, + "area_overlap": 0.611, + "patch_similarity": 0.387, + "shared_filenames": [ + "src/transformers/models/cohere_asr/modeling_cohere_asr.py", + "src/transformers/models/cohere_asr/modular_cohere_asr.py", + "tests/models/cohere_asr/test_modeling_cohere_asr.py" + ], + "shared_file_areas": [ + { + "filename": "src/transformers/models/cohere_asr/modeling_cohere_asr.py", + "left_ranges": [ + [ + 22, + 28 + ], + [ + 638, + 645 + ] + ], + "right_ranges": [ + [ + 637, + 649 + ] + ] + }, + { + "filename": "src/transformers/models/cohere_asr/modular_cohere_asr.py", + "left_ranges": [ + [ + 16, + 22 + ], + [ + 501, + 508 + ] + ], + "right_ranges": [ + [ + 500, + 512 + ] + ] + }, + { + "filename": "tests/models/cohere_asr/test_modeling_cohere_asr.py", + "left_ranges": [ + [ + 15, + 21 + ], + [ + 180, + 221 + ] + ], + "right_ranges": [ + [ + 175, + 222 + ] + ] + } + ] + } + ] + }, + { + "cluster_id": "cluster-44779-5", + "summary": "Cluster of 3 issues and 2 PRs centered on issue #45488.", + "status": "open", + "confidence": 0.6, + "canonical_issue_number": 45488, + "canonical_pr_number": 47017, + "issue_numbers": [ + 44779, + 45488, + 46489 + ], + "pr_numbers": [ + 46580, + 47017 + ], + "evidence_types": [ + "closing_reference", + "soft_similarity" + ], + "pr_comparisons": [ + { + "left_pr_number": 46580, + "right_pr_number": 47017, + "code_similarity": 0.241, + "size_similarity": 0.352, + "file_overlap": 0.5, + "area_overlap": 0.0, + "patch_similarity": 0.136, + "shared_filenames": [ + "src/transformers/models/auto/tokenization_auto.py" + ], + "shared_file_areas": [ + { + "filename": "src/transformers/models/auto/tokenization_auto.py", + "left_ranges": [ + [ + 361, + 367 + ], + [ + 399, + 421 + ], + [ + 724, + 739 + ] + ], + "right_ranges": [ + [ + 428, + 475 + ], + [ + 807, + 825 + ] + ] + } + ] + } + ] + }, + { + "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-46693-3", + "summary": "Cluster of 1 issues and 2 PRs centered on issue #46693.", + "status": "open", + "confidence": 0.55, + "canonical_issue_number": 46693, + "canonical_pr_number": 47134, + "issue_numbers": [ + 46693 + ], + "pr_numbers": [ + 46696, + 47134 + ], + "evidence_types": [ + "closing_reference" + ], + "pr_comparisons": [ + { + "left_pr_number": 46696, + "right_pr_number": 47134, + "code_similarity": 0.464, + "size_similarity": 0.8, + "file_overlap": 1.0, + "area_overlap": 0.0, + "patch_similarity": 0.029, + "shared_filenames": [ + "src/transformers/modeling_flash_attention_utils.py" + ], + "shared_file_areas": [ + { + "filename": "src/transformers/modeling_flash_attention_utils.py", + "left_ranges": [ + [ + 667, + 692 + ] + ], + "right_ranges": [ + [ + 316, + 323 + ], + [ + 366, + 373 + ] + ] + } + ] + } + ] + }, + { + "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-46869-3", + "summary": "Cluster of 1 issues and 2 PRs centered on issue #46869.", + "status": "open", + "confidence": 0.75, + "canonical_issue_number": 46869, + "canonical_pr_number": 46886, + "issue_numbers": [ + 46869 + ], + "pr_numbers": [ + 46886, + 46922 + ], + "evidence_types": [ + "closing_reference", + "shared_issue_target" + ], + "pr_comparisons": [ + { + "left_pr_number": 46886, + "right_pr_number": 46922, + "code_similarity": 0.711, + "size_similarity": 0.857, + "file_overlap": 1.0, + "area_overlap": 0.5, + "patch_similarity": 0.429, + "shared_filenames": [ + "src/transformers/models/yolos/modeling_yolos.py" + ], + "shared_file_areas": [ + { + "filename": "src/transformers/models/yolos/modeling_yolos.py", + "left_ranges": [ + [ + 446, + 452 + ] + ], + "right_ranges": [ + [ + 434, + 440 + ], + [ + 446, + 452 + ] + ] + } + ] + } + ] + }, + { + "cluster_id": "cluster-46159-3", + "summary": "Cluster of 1 issues and 2 PRs centered on issue #46159.", + "status": "open", + "confidence": 0.55, + "canonical_issue_number": 46159, + "canonical_pr_number": 46174, + "issue_numbers": [ + 46159 + ], + "pr_numbers": [ + 46174, + 46780 + ], + "evidence_types": [ + "closing_reference" + ], + "pr_comparisons": [ + { + "left_pr_number": 46174, + "right_pr_number": 46780, + "code_similarity": 0.841, + "size_similarity": 0.87, + "file_overlap": 1.0, + "area_overlap": 0.767, + "patch_similarity": 0.657, + "shared_filenames": [ + "src/transformers/models/aimv2/modeling_aimv2.py", + "src/transformers/models/d_fine/modeling_d_fine.py", + "src/transformers/models/deimv2/modeling_deimv2.py", + "src/transformers/models/pp_doclayout_v2/modeling_pp_doclayout_v2.py", + "src/transformers/models/pp_doclayout_v3/modeling_pp_doclayout_v3.py", + "src/transformers/models/rt_detr/modeling_rt_detr.py", + "src/transformers/models/rt_detr_v2/modeling_rt_detr_v2.py", + "src/transformers/models/vit_mae/modeling_vit_mae.py", + "src/transformers/models/vit_mae/modular_vit_mae.py" + ], + "shared_file_areas": [ + { + "filename": "src/transformers/models/aimv2/modeling_aimv2.py", + "left_ranges": [ + [ + 134, + 141 + ], + [ + 144, + 154 + ], + [ + 157, + 165 + ] + ], + "right_ranges": [ + [ + 134, + 140 + ], + [ + 143, + 153 + ], + [ + 156, + 162 + ] + ] + }, + { + "filename": "src/transformers/models/d_fine/modeling_d_fine.py", + "left_ranges": [ + [ + 631, + 638 + ], + [ + 641, + 651 + ], + [ + 654, + 662 + ] + ], + "right_ranges": [ + [ + 631, + 637 + ], + [ + 640, + 650 + ], + [ + 653, + 659 + ] + ] + }, + { + "filename": "src/transformers/models/deimv2/modeling_deimv2.py", + "left_ranges": [ + [ + 703, + 710 + ], + [ + 713, + 723 + ], + [ + 726, + 734 + ] + ], + "right_ranges": [ + [ + 703, + 709 + ], + [ + 712, + 722 + ], + [ + 725, + 731 + ] + ] + }, + { + "filename": "src/transformers/models/pp_doclayout_v2/modeling_pp_doclayout_v2.py", + "left_ranges": [ + [ + 1551, + 1558 + ], + [ + 1561, + 1571 + ], + [ + 1574, + 1582 + ] + ], + "right_ranges": [ + [ + 1551, + 1557 + ], + [ + 1560, + 1570 + ], + [ + 1573, + 1579 + ] + ] + }, + { + "filename": "src/transformers/models/pp_doclayout_v3/modeling_pp_doclayout_v3.py", + "left_ranges": [ + [ + 804, + 811 + ], + [ + 814, + 824 + ], + [ + 827, + 835 + ] + ], + "right_ranges": [ + [ + 804, + 810 + ], + [ + 813, + 823 + ], + [ + 826, + 832 + ] + ] + }, + { + "filename": "src/transformers/models/rt_detr/modeling_rt_detr.py", + "left_ranges": [ + [ + 847, + 854 + ], + [ + 857, + 867 + ], + [ + 870, + 878 + ] + ], + "right_ranges": [ + [ + 847, + 853 + ], + [ + 856, + 866 + ], + [ + 869, + 875 + ] + ] + }, + { + "filename": "src/transformers/models/rt_detr_v2/modeling_rt_detr_v2.py", + "left_ranges": [ + [ + 976, + 983 + ], + [ + 986, + 996 + ], + [ + 999, + 1007 + ] + ], + "right_ranges": [ + [ + 976, + 982 + ], + [ + 985, + 995 + ], + [ + 998, + 1004 + ] + ] + }, + { + "filename": "src/transformers/models/vit_mae/modeling_vit_mae.py", + "left_ranges": [ + [ + 152, + 159 + ], + [ + 162, + 172 + ], + [ + 175, + 183 + ] + ], + "right_ranges": [ + [ + 152, + 158 + ], + [ + 161, + 171 + ], + [ + 174, + 180 + ] + ] + }, + { + "filename": "src/transformers/models/vit_mae/modular_vit_mae.py", + "left_ranges": [ + [ + 62, + 69 + ], + [ + 72, + 82 + ], + [ + 85, + 93 + ] + ], + "right_ranges": [ + [ + 62, + 68 + ], + [ + 71, + 81 + ], + [ + 84, + 90 + ] + ] + } + ] + } + ] + }, + { + "cluster_id": "cluster-45561-3", + "summary": "Cluster of 1 issues and 2 PRs centered on issue #45561.", + "status": "open", + "confidence": 0.8, + "canonical_issue_number": 45561, + "canonical_pr_number": 45645, + "issue_numbers": [ + 45561 + ], + "pr_numbers": [ + 45639, + 45645 + ], + "evidence_types": [ + "closing_reference", + "shared_issue_target", + "soft_similarity" + ], + "pr_comparisons": [ + { + "left_pr_number": 45639, + "right_pr_number": 45645, + "code_similarity": 0.721, + "size_similarity": 0.973, + "file_overlap": 0.75, + "area_overlap": 0.563, + "patch_similarity": 0.696, + "shared_filenames": [ + "src/transformers/testing_utils.py", + "tests/utils/test_testing_utils.py", + "utils/notification_service.py" + ], + "shared_file_areas": [ + { + "filename": "src/transformers/testing_utils.py", + "left_ranges": [ + [ + 3611, + 3618 + ], + [ + 3835, + 3869 + ] + ], + "right_ranges": [ + [ + 3525, + 3558 + ], + [ + 3782, + 3788 + ] + ] + }, + { + "filename": "tests/utils/test_testing_utils.py", + "left_ranges": [ + [ + 1, + 144 + ] + ], + "right_ranges": [ + [ + 1, + 114 + ] + ] + }, + { + "filename": "utils/notification_service.py", + "left_ranges": [ + [ + 935, + 967 + ] + ], + "right_ranges": [ + [ + 935, + 967 + ] + ] + } + ] + } + ] + }, + { + "cluster_id": "cluster-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-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": 46310, + "issue_numbers": [ + 46305 + ], + "pr_numbers": [ + 46310, + 47162 + ], + "evidence_types": [ + "closing_reference", + "shared_issue_target", + "soft_similarity" + ], + "pr_comparisons": [ + { + "left_pr_number": 46310, + "right_pr_number": 47162, + "code_similarity": 0.577, + "size_similarity": 0.338, + "file_overlap": 0.5, + "area_overlap": 1.0, + "patch_similarity": 0.062, + "shared_filenames": [ + "src/transformers/models/moonshine_streaming/modeling_moonshine_streaming.py", + "src/transformers/models/moonshine_streaming/modular_moonshine_streaming.py" + ], + "shared_file_areas": [ + { + "filename": "src/transformers/models/moonshine_streaming/modeling_moonshine_streaming.py", + "left_ranges": [ + [ + 356, + 362 + ] + ], + "right_ranges": [ + [ + 356, + 362 + ] + ] + }, + { + "filename": "src/transformers/models/moonshine_streaming/modular_moonshine_streaming.py", + "left_ranges": [ + [ + 269, + 275 + ] + ], + "right_ranges": [ + [ + 269, + 275 + ] + ] + } + ] + } + ] + }, + { + "cluster_id": "cluster-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": [ + { + "cluster_id": "cluster-44779-5", + "canonical_issue_number": 45488, + "duplicate_issue_numbers": [ + 44779, + 46489 + ], + "reason": "Issues in cluster-44779-5 are treated as duplicates because they share closing_reference, soft_similarity evidence." + }, + { + "cluster_id": "cluster-46894-5", + "canonical_issue_number": 46894, + "duplicate_issue_numbers": [ + 46897 + ], + "reason": "Issues in cluster-46894-5 are treated as duplicates because they share closing_reference, shared_issue_target, soft_similarity evidence." + } + ], + "duplicate_prs": [ + { + "cluster_id": "cluster-41211-3", + "canonical_pr_number": 44339, + "duplicate_pr_numbers": [ + 41356 + ], + "target_issue_number": 41211, + "reason": "PRs in cluster-41211-3 are treated as duplicates because they converge on issue #41211 with closing_reference evidence." + }, + { + "cluster_id": "cluster-43240-3", + "canonical_pr_number": 43251, + "duplicate_pr_numbers": [ + 43254 + ], + "target_issue_number": 43240, + "reason": "PRs in cluster-43240-3 are treated as duplicates because they converge on issue #43240 with closing_reference, shared_issue_target evidence." + }, + { + "cluster_id": "cluster-43451-6", + "canonical_pr_number": 43451, + "duplicate_pr_numbers": [ + 46912, + 47035, + 47096, + 47163, + 47166 + ], + "target_issue_number": null, + "reason": "PRs in cluster-43451-6 are treated as duplicates because they share soft_similarity evidence." + }, + { + "cluster_id": "cluster-43698-3", + "canonical_pr_number": 43779, + "duplicate_pr_numbers": [ + 43816 + ], + "target_issue_number": 43698, + "reason": "PRs in cluster-43698-3 are treated as duplicates because they converge on issue #43698 with closing_reference, shared_issue_target evidence." + }, + { + "cluster_id": "cluster-43979-11", + "canonical_pr_number": 44007, + "duplicate_pr_numbers": [ + 43996, + 44013, + 44044, + 44066, + 44072, + 44085, + 44129, + 44154, + 44722 + ], + "target_issue_number": 43979, + "reason": "PRs in cluster-43979-11 are treated as duplicates because they converge on issue #43979 with closing_reference, shared_issue_target, soft_similarity evidence." + }, + { + "cluster_id": "cluster-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-44779-5", + "canonical_pr_number": 47017, + "duplicate_pr_numbers": [ + 46580 + ], + "target_issue_number": 45488, + "reason": "PRs in cluster-44779-5 are treated as duplicates because they converge on issue #45488 with closing_reference, 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-46159-3", + "canonical_pr_number": 46174, + "duplicate_pr_numbers": [ + 46780 + ], + "target_issue_number": 46159, + "reason": "PRs in cluster-46159-3 are treated as duplicates because they converge on issue #46159 with closing_reference evidence." + }, + { + "cluster_id": "cluster-46305-3", + "canonical_pr_number": 46310, + "duplicate_pr_numbers": [ + 47162 + ], + "target_issue_number": 46305, + "reason": "PRs in cluster-46305-3 are treated as duplicates because they converge on issue #46305 with closing_reference, shared_issue_target, soft_similarity evidence." + }, + { + "cluster_id": "cluster-46693-3", + "canonical_pr_number": 47134, + "duplicate_pr_numbers": [ + 46696 + ], + "target_issue_number": 46693, + "reason": "PRs in cluster-46693-3 are treated as duplicates because they converge on issue #46693 with closing_reference evidence." + }, + { + "cluster_id": "cluster-46869-3", + "canonical_pr_number": 46886, + "duplicate_pr_numbers": [ + 46922 + ], + "target_issue_number": 46869, + "reason": "PRs in cluster-46869-3 are treated as duplicates because they converge on issue #46869 with closing_reference, shared_issue_target evidence." + }, + { + "cluster_id": "cluster-46894-5", + "canonical_pr_number": 46895, + "duplicate_pr_numbers": [ + 46958, + 47090 + ], + "target_issue_number": 46894, + "reason": "PRs in cluster-46894-5 are treated as duplicates because they converge on issue #46894 with closing_reference, shared_issue_target, soft_similarity evidence." + } + ], + "best_issue": { + "cluster_id": "cluster-44779-5", + "issue_number": 45488, + "reason": "Issue #45488 is the strongest global issue candidate because it is open, belongs to a cluster with 5 artifacts, and carries 16 discussion comments plus 4 inbound references.", + "score": 186.0 + }, + "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/20260708T120111Z/analysis-runs/hybrid-model-20260708t120111z/analysis-report-hybrid.llm-reviews.json b/snapshots/20260708T120111Z/analysis-runs/hybrid-model-20260708t120111z/analysis-report-hybrid.llm-reviews.json new file mode 100644 index 0000000000000000000000000000000000000000..f8a5bab6c1418b5e5bc1e65e1738b1521fefd571 --- /dev/null +++ b/snapshots/20260708T120111Z/analysis-runs/hybrid-model-20260708t120111z/analysis-report-hybrid.llm-reviews.json @@ -0,0 +1,26574 @@ +{ + "schema_version": "1.0", + "repo": "huggingface/transformers", + "snapshot_id": "20260708T120111Z", + "generated_at": "2026-07-08T12:14:07Z", + "model": "gpt-5.4-mini?service_tier=flex", + "reviews": [ + { + "label": "issue", + "component_index": 1, + "component_count": 27, + "review_unit_index": 1, + "review_unit_count": 63, + "status": "reviewed", + "reason": null, + "source": "llm", + "cache_hit": false, + "model": "gpt-5.4-mini?service_tier=flex", + "cluster_id": "cluster-10000-514", + "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", + "issue:46869" + ], + "soft_pairs": [ + "issue:44322|issue:45464", + "issue:45237|issue:45685", + "issue:43716|issue:45237", + "issue:43828|issue:45237", + "issue:45237|issue:46468", + "issue:45237|issue:45507", + "issue:30277|issue:35545", + "issue:35545|issue:42915", + "issue:35545|issue:41720", + "issue:45901|issue:46015", + "issue:35545|issue:36010", + "issue:44509|issue:44512", + "issue:44918|issue:46165", + "issue:46165|issue:46869", + "issue:44805|issue:45823", + "issue:43638|issue:44805", + "issue:43638|issue:46165", + "issue:43596|issue:44805", + "issue:44805|issue:46294", + "issue:44805|issue:45161", + "issue:46165|issue:46468", + "issue:44560|issue:46165", + "issue:34634|issue:35545", + "issue:30277|issue:43638", + "issue:43596|issue:43638", + "issue:43039|issue:45161", + "issue:43975|issue:44457", + "issue:44292|issue:44457", + "issue:44841|issue:46661", + "issue:43716|issue:46869", + "issue:24643|issue:30277", + "issue:34928|issue:44805", + "issue:45685|issue:46869", + "issue:45859|issue:45902", + "issue:34928|issue:36331", + "issue:43638|issue:45161", + "issue:43596|issue:45161", + "issue:43716|issue:45507", + "issue:44292|issue:45237", + "issue:45874|issue:46294", + "issue:39290|issue:45081", + "issue:45800|issue:46515", + "issue:45823|issue:45874", + "issue:34928|issue:41720", + "issue:43425|issue:44292", + "issue:43828|issue:45161", + "issue:43541|issue:43716", + "issue:43828|issue:44292", + "issue:43716|issue:45685", + "issue:45507|issue:45685", + "issue:44292|issue:44918", + "issue:43716|issue:44292", + "issue:44918|issue:45161", + "issue:43039|issue:44918", + "issue:41762|issue:43638", + "issue:43827|issue:44512", + "issue:43828|issue:45685", + "issue:43039|issue:43828", + "issue:30277|issue:36331", + "issue:34928|issue:36010", + "issue:45237|issue:46015", + "issue:43039|issue:44805" + ], + "prepared_review_unit_hash": "97619858853a9afd1a3f06970dcc6263fa045271b03d45966a259ee3c741e7ea", + "component_budget": { + "node_count": 514, + "item_count": 514, + "soft_pair_count": 1891, + "serialized_chars": 980895, + "estimated_input_tokens": 245224, + "estimated_eval_tokens": 490704 + }, + "budget": { + "node_count": 48, + "item_count": 48, + "soft_pair_count": 62, + "serialized_chars": 46274, + "estimated_input_tokens": 11569, + "estimated_eval_tokens": 23394 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": true, + "analyst_result": { + "summary": "This cluster is mostly a set of unrelated issues grouped by broad text similarity. The only clear duplicate families are the DeepSpeed ZeRO-3 \"index 0 is out of bounds\" reports, a torch-2.4.1 compatibility pair, and a Qwen3_5MoeTextRotaryEmbedding offload/materialization pair.", + "confidence": 0.74, + "canonical_issue_reason": "Issue 43596 is the most general representative of the DeepSpeed ZeRO-3 index-out-of-bounds family; 43638 and 41762 read like model-specific variants of the same zero-sized tensor/init bug.", + "canonical_pr_reason": null, + "best_issue_reason": "43596 is the best global issue anchor because it describes the shared failure mode most cleanly and broadly, while the other accepted duplicates are narrower reproductions of the same ZeRO-3 init problem.", + "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 failure." + }, + { + "left": "issue:45237", + "right": "issue:45685", + "accept": false, + "reason": "Unrelated hardware/runtime problems: AMD GPU support vs MPS histogram op support." + }, + { + "left": "issue:43716", + "right": "issue:45237", + "accept": false, + "reason": "Different bug classes and models; dtype mismatch in preprocessing is not the AMD GPU failure." + }, + { + "left": "issue:43828", + "right": "issue:45237", + "accept": false, + "reason": "Autocast dtype mismatch on Phi MoE is unrelated to GPT-OSS AMD GPU runtime failure." + }, + { + "left": "issue:45237", + "right": "issue:46468", + "accept": false, + "reason": "Different models and symptoms; Nemotron example breakage is not the GPT-OSS AMD GPU issue." + }, + { + "left": "issue:45237", + "right": "issue:45507", + "accept": false, + "reason": "Different concrete failures: AMD GPU runtime issue vs GraniteMoEHybrid invalid-method call." + }, + { + "left": "issue:30277", + "right": "issue:35545", + "accept": false, + "reason": "DeepSpeed collective mismatch and ONNX export error are unrelated." + }, + { + "left": "issue:35545", + "right": "issue:42915", + "accept": false, + "reason": "ONNX export failure and FineGrainedFP8 failure do not share the same code-path." + }, + { + "left": "issue:35545", + "right": "issue:41720", + "accept": false, + "reason": "ONNX export error vs Qwen3 auto-device-map CUDA assert are different issues." + }, + { + "left": "issue:45901", + "right": "issue:46015", + "accept": false, + "reason": "Both are docs/examples-related at a high level, but they target different content and fixes." + }, + { + "left": "issue:35545", + "right": "issue:36010", + "accept": false, + "reason": "Export-to-ONNX and GenerationMixin import errors are unrelated." + }, + { + "left": "issue:44509", + "right": "issue:44512", + "accept": false, + "reason": "Both are v5 docs cleanup items, but they reference different removed features and likely different edits." + }, + { + "left": "issue:44918", + "right": "issue:46165", + "accept": false, + "reason": "TRl SFT embedding unpacking and model resume/trust_remote_code are different training issues." + }, + { + "left": "issue:46165", + "right": "issue:46869", + "accept": false, + "reason": "Remote-code resume bug and YOLO multi-GPU device-map bug are unrelated." + }, + { + "left": "issue:44805", + "right": "issue:45823", + "accept": false, + "reason": "Mask-shape mismatch and Gemma4 device mismatch are different forward-path failures." + }, + { + "left": "issue:43638", + "right": "issue:44805", + "accept": false, + "reason": "Zero-3 index-out-of-bounds and mask/tensor shape mismatch are different failures." + }, + { + "left": "issue:43638", + "right": "issue:46165", + "accept": false, + "reason": "DeepSpeed ZeRO-3 init bug and model resume/trust_remote_code bug are unrelated." + }, + { + "left": "issue:43596", + "right": "issue:44805", + "accept": false, + "reason": "Different symptom and subsystem; this is not the same Zero-3 init bug." + }, + { + "left": "issue:44805", + "right": "issue:46294", + "accept": false, + "reason": "Mask/index tensor mismatch is unrelated to Qwen3.5 video token/feature mismatch." + }, + { + "left": "issue:44805", + "right": "issue:45161", + "accept": false, + "reason": "Mask-shape error and GPT-OSS TP-only failure are different bugs." + }, + { + "left": "issue:46165", + "right": "issue:46468", + "accept": false, + "reason": "Different areas: resuming remote-code training vs Nemotron example breakage." + }, + { + "left": "issue:44560", + "right": "issue:46165", + "accept": false, + "reason": "Video chat-template failure and remote-code resume failure are unrelated." + }, + { + "left": "issue:34634", + "right": "issue:35545", + "accept": false, + "reason": "BarkProcessor voice_preset bug is unrelated to ModernBERT ONNX export." + }, + { + "left": "issue:30277", + "right": "issue:43638", + "accept": false, + "reason": "Collective mismatch under ZeRO-3 is a different issue from the index-out-of-bounds init bug." + }, + { + "left": "issue:43596", + "right": "issue:43638", + "accept": true, + "reason": "Same DeepSpeed ZeRO-3 'index 0 is out of bounds' failure; BertModel vs non-pretrained Bert is just a repro variant." + }, + { + "left": "issue:43039", + "right": "issue:45161", + "accept": false, + "reason": "Liger Kernel cross_entropy being called is unrelated to GPT-OSS tensor-parallel failure." + }, + { + "left": "issue:43975", + "right": "issue:44457", + "accept": false, + "reason": "Wrong detokenization and LoRA merge/save mismatch are different issues." + }, + { + "left": "issue:44292", + "right": "issue:44457", + "accept": false, + "reason": "Qwen-3 NVFP4 runtime failure is unrelated to LoRA merge/save behavior." + }, + { + "left": "issue:44841", + "right": "issue:46661", + "accept": false, + "reason": "Processor failure for Voxtral is unrelated to remote-code model_class/config_class import handling." + }, + { + "left": "issue:43716", + "right": "issue:46869", + "accept": false, + "reason": "Mistral-3 dtype mismatch and YOLO multi-GPU device-map failure are different issues." + }, + { + "left": "issue:24643", + "right": "issue:30277", + "accept": false, + "reason": "Both involve DeepSpeed, but the concrete errors and code paths differ." + }, + { + "left": "issue:34928", + "right": "issue:44805", + "accept": false, + "reason": "Activation-checkpointing/FSDP recompute mismatch is not the same as a mask-shape index error." + }, + { + "left": "issue:45685", + "right": "issue:46869", + "accept": false, + "reason": "MPS histogram limitation and YOLO multi-GPU device-map failure are unrelated." + }, + { + "left": "issue:45859", + "right": "issue:45902", + "accept": true, + "reason": "Same Qwen3_5MoeTextRotaryEmbedding buffer/offload bug; one report is a forward/offload symptom and the other is a materialization symptom." + }, + { + "left": "issue:34928", + "right": "issue:36331", + "accept": false, + "reason": "FSDP activation-checkpointing mismatch and CustomTrainer API breakage are unrelated." + }, + { + "left": "issue:43638", + "right": "issue:45161", + "accept": false, + "reason": "DeepSpeed ZeRO-3 init bug and GPT-OSS TP-only failure are unrelated." + }, + { + "left": "issue:43596", + "right": "issue:45161", + "accept": false, + "reason": "DeepSpeed ZeRO-3 index error and GPT-OSS tensor-parallel bug are unrelated." + }, + { + "left": "issue:43716", + "right": "issue:44292", + "accept": false, + "reason": "Different models and failure modes; dtype mismatch vs NVFP4 runtime error." + }, + { + "left": "issue:44292", + "right": "issue:45237", + "accept": false, + "reason": "Qwen-3 NVFP4 failure and GPT-OSS AMD GPU failure are different issues." + }, + { + "left": "issue:45874", + "right": "issue:46294", + "accept": false, + "reason": "Gemma4 reverse-embedding lookup and Qwen3.5 video token mismatch are unrelated." + }, + { + "left": "issue:39290", + "right": "issue:45081", + "accept": false, + "reason": "Gemma3 config attribute error under vLLM and Mistral regex patch crash are different bugs." + }, + { + "left": "issue:45800", + "right": "issue:46515", + "accept": true, + "reason": "Both report the same torch 2.4.1 incompatibility; 45800 just names transformers 5.8.0 explicitly." + }, + { + "left": "issue:45823", + "right": "issue:45874", + "accept": false, + "reason": "Gemma4 device mismatch and Gemma4 reverse-embedding lookup are different forward-path problems." + }, + { + "left": "issue:34928", + "right": "issue:41720", + "accept": false, + "reason": "FSDP checkpoint recompute mismatch is unrelated to Qwen3 auto-device-map CUDA assert." + }, + { + "left": "issue:43425", + "right": "issue:44292", + "accept": false, + "reason": "Generic torch incompatibility and Qwen-3 NVFP4 failure are not the same bug." + }, + { + "left": "issue:43828", + "right": "issue:45161", + "accept": false, + "reason": "Phi autocast dtype mismatch and GPT-OSS TP-only failure are different issues." + }, + { + "left": "issue:43541", + "right": "issue:43716", + "accept": false, + "reason": "Mixtral grouped_mm tracing error is unrelated to Mistral-3 preprocessing dtype mismatch." + }, + { + "left": "issue:43828", + "right": "issue:44292", + "accept": false, + "reason": "Different models and failure modes: autocast dtype mismatch vs NVFP4 runtime failure." + }, + { + "left": "issue:43716", + "right": "issue:45685", + "accept": false, + "reason": "Mistral-3 dtype mismatch and MPS histogram limitation are unrelated." + }, + { + "left": "issue:45507", + "right": "issue:45685", + "accept": false, + "reason": "GraniteMoEHybrid invalid-method bug and MPS histogram limitation are unrelated." + } + ] + }, + "evaluator_result": { + "accept": true, + "feedback": "The summary is grounded in the packet and stays conservative. The three accepted duplicate families are supported by the titles, and the rejected soft edges are mostly cautious/non-overstated. I don\u2019t see an evidence stretch that would justify rejection." + }, + "evaluator_used": true, + "retried": false, + "accepted_nontrivial_soft_edge": true, + "error_kind": null, + "error_message": null + }, + { + "label": "issue", + "component_index": 1, + "component_count": 27, + "review_unit_index": 2, + "review_unit_count": 63, + "status": "reviewed", + "reason": null, + "source": "llm", + "cache_hit": false, + "model": "gpt-5.4-mini?service_tier=flex", + "cluster_id": "cluster-10000-514", + "nodes": [ + "issue:36010", + "issue:39290", + "issue:41093", + "issue:41720", + "issue:41762", + "issue:43039", + "issue:43296", + "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:44512", + "issue:44560", + "issue:44589", + "issue:44841", + "issue:44863", + "issue:45020", + "issue:45070", + "issue:45084", + "issue:45237", + "issue:45362", + "issue:45399", + "issue:45464", + "issue:45507", + "issue:45685", + "issue:45800", + "issue:45812", + "issue:45823", + "issue:45901", + "issue:45920", + "issue:46015", + "issue:46294", + "issue:46304", + "issue:46402", + "issue:46661", + "issue:46772", + "issue:46869" + ], + "soft_pairs": [ + "issue:45800|issue:46304", + "issue:43296|issue:46304", + "issue:43866|issue:46402", + "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:43866|issue:46772", + "issue:44863|issue:46015", + "issue:45020|issue:46304", + "issue:45812|issue:45920", + "issue:43872|issue:46304", + "issue:44292|issue:46015", + "issue:43866|issue:44863", + "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:43792|issue:45070", + "issue:45362|issue:45464", + "issue:43596|issue:45800", + "issue:43541|issue:46869", + "issue:44292|issue:45901", + "issue:39290|issue:45823", + "issue:45464|issue:45685", + "issue:36010|issue:39290", + "issue:43541|issue:45685", + "issue:43541|issue:43828", + "issue:44841|issue:45084", + "issue:45507|issue:46869", + "issue:43404|issue:43866", + "issue:41093|issue:41762", + "issue:45399|issue:46661", + "issue:43572|issue:45070", + "issue:45800|issue:45823" + ], + "prepared_review_unit_hash": "cd8bdc78eca6880ffc2eb8bb39920fcfc8a659b0cebe9a4f89b04adaedf4098b", + "component_budget": { + "node_count": 514, + "item_count": 514, + "soft_pair_count": 1891, + "serialized_chars": 980895, + "estimated_input_tokens": 245224, + "estimated_eval_tokens": 490704 + }, + "budget": { + "node_count": 48, + "item_count": 48, + "soft_pair_count": 42, + "serialized_chars": 40038, + "estimated_input_tokens": 10010, + "estimated_eval_tokens": 20276 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": true, + "analyst_result": { + "summary": "This cluster is mostly a grab-bag of unrelated model-loading/runtime regressions. The only clear duplicate pair is the AutoTokenizer v5 regression across multiple model families; most other soft pairs are merely similar symptom classes and should stay separate.", + "confidence": 0.62, + "canonical_issue_reason": "Issue 45920 is the broader, more representative report: it names multiple affected model families and frames the problem as a cross-model v5 AutoTokenizer regression, whereas 45812 is Granite-specific.", + "canonical_pr_reason": null, + "best_issue_reason": "45920 is the best issue to anchor the cluster because it is the most general and likely to capture the underlying tokenizer regression with the widest impact.", + "best_pr_reason": null, + "soft_edge_verdicts": [ + { + "left": "issue:45800", + "right": "issue:46304", + "accept": false, + "reason": "Both are version-compatibility complaints, but one is a torch/transformers incompatibility and the other is a specific USE_DISTRIBUTED=0 breakage; not the same concrete bug." + }, + { + "left": "issue:43296", + "right": "issue:46304", + "accept": false, + "reason": "Different failures in different stacks: PaddleOCR-VL loading in vLLM vs distributed support regression in PyTorch support." + }, + { + "left": "issue:43866", + "right": "issue:46402", + "accept": false, + "reason": "Both are model-load failures, but one is a corrupted checkpoint claim and the other is missing text-encoder weights; different underlying problems." + }, + { + "left": "issue:41762", + "right": "issue:43596", + "accept": false, + "reason": "Both mention Zero-3 index errors, but they involve different models and loading paths; too little evidence they are the same bug." + }, + { + "left": "issue:39290", + "right": "issue:41720", + "accept": false, + "reason": "Gemma3 config attribute error in vLLM and Qwen3 auto device-map cuda assert are unrelated runtime failures." + }, + { + "left": "issue:43901", + "right": "issue:44512", + "accept": false, + "reason": "Both are docs regressions, but they concern different deprecated docs behaviors and are not the same issue." + }, + { + "left": "issue:44589", + "right": "issue:46304", + "accept": false, + "reason": "Float8 storage lookup failure is unrelated to the PyTorch distributed support regression." + }, + { + "left": "issue:45237", + "right": "issue:45901", + "accept": false, + "reason": "AMD GPU execution failure and table-question-answering crash are different code paths and symptoms." + }, + { + "left": "issue:43039", + "right": "issue:44560", + "accept": false, + "reason": "Liger Kernel cross-entropy invocation and Qwen3-vl video StopIteration are unrelated issues." + }, + { + "left": "issue:45823", + "right": "issue:46294", + "accept": false, + "reason": "Both involve device/input handling, but one is a forward device mismatch and the other is video token/feature mismatch in chat templating; not the same bug." + }, + { + "left": "issue:43975", + "right": "issue:46015", + "accept": false, + "reason": "Tokenizer detokenization bug and outdated RAG docs are unrelated." + }, + { + "left": "issue:44841", + "right": "issue:45823", + "accept": false, + "reason": "Voxtral processor failure and Gemma4 device mismatch do not look like the same underlying problem." + }, + { + "left": "issue:43866", + "right": "issue:46772", + "accept": false, + "reason": "Checkpoint corruption and NVFP4/model-opt compatibility are different failure modes." + }, + { + "left": "issue:44863", + "right": "issue:46015", + "accept": false, + "reason": "NemotronH checkpoint-loading issue and RAG docs are unrelated." + }, + { + "left": "issue:45020", + "right": "issue:46304", + "accept": false, + "reason": "Remote-code model breakage and PyTorch distributed support regression affect different mechanisms." + }, + { + "left": "issue:45812", + "right": "issue:45920", + "accept": true, + "reason": "Same underlying tokenizer regression: both report wrong token IDs in v5 across Granite/other models, and one is clearly the broader umbrella report." + }, + { + "left": "issue:43872", + "right": "issue:46304", + "accept": false, + "reason": "bitsandbytes Int8Params constructor error is unrelated to the torch distributed compatibility issue." + }, + { + "left": "issue:44292", + "right": "issue:46015", + "accept": false, + "reason": "NVFP4 inference failure and outdated RAG docs are unrelated." + }, + { + "left": "issue:43866", + "right": "issue:44863", + "accept": false, + "reason": "Both are loading problems, but they concern different model families and different failure descriptions." + }, + { + "left": "issue:39290", + "right": "issue:43531", + "accept": false, + "reason": "Both mention sliding-window behavior, but the concrete failures differ too much: Gemma3 config attribute error vs Qwen3-MoE sliding_window issue." + }, + { + "left": "issue:41093", + "right": "issue:43638", + "accept": false, + "reason": "Shape mismatch in attention mask indexing is not the same as the Zero-3 empty-tensor index error." + }, + { + "left": "issue:43792", + "right": "issue:45901", + "accept": false, + "reason": "Whisper loading/running failure and table QA crash are unrelated." + }, + { + "left": "issue:43854", + "right": "issue:43866", + "accept": false, + "reason": "Unit-test loading failure for GLM-4.7-Flash and a corrupted Ovis2 checkpoint are different issues." + }, + { + "left": "issue:43425", + "right": "issue:45070", + "accept": false, + "reason": "Torch version incompatibility and a PretrainedConfig/pydantic regression are not the same concrete bug." + }, + { + "left": "issue:43975", + "right": "issue:45901", + "accept": false, + "reason": "DeepSeek detokenization bug and table-question-answering crash are unrelated." + }, + { + "left": "issue:43792", + "right": "issue:45070", + "accept": false, + "reason": "Whisper inference failure and pydantic config regression do not share the same code path." + }, + { + "left": "issue:45362", + "right": "issue:45464", + "accept": false, + "reason": "Both involve Qwen3.5 chat, but one is a generic chat crash and the other is a streaming chat/completions API failure; not clearly the same bug." + }, + { + "left": "issue:43596", + "right": "issue:45800", + "accept": false, + "reason": "Zero-3 Bert loading failure and torch/transformers version incompatibility are different classes of problems." + }, + { + "left": "issue:43541", + "right": "issue:46869", + "accept": false, + "reason": "Mixtral grouped_mm tracing failure and YOLO multi-GPU device-map failure are unrelated." + }, + { + "left": "issue:44292", + "right": "issue:45901", + "accept": false, + "reason": "NVFP4 model execution failure and table QA crash are unrelated." + }, + { + "left": "issue:39290", + "right": "issue:45823", + "accept": false, + "reason": "Gemma3 config/sliding-window failure and Gemma4 device mismatch are different bugs despite both being model-specific runtime issues." + }, + { + "left": "issue:45464", + "right": "issue:45685", + "accept": false, + "reason": "Chat/completions streaming failure and MPS histogram unsupported-for-Int error are unrelated." + }, + { + "left": "issue:36010", + "right": "issue:39290", + "accept": false, + "reason": "GenerationMixin import error and Gemma3 config attribute error are unrelated." + }, + { + "left": "issue:43541", + "right": "issue:45685", + "accept": false, + "reason": "Grouped_mm tracing failure and MPS histogram type error do not look like the same bug." + }, + { + "left": "issue:43541", + "right": "issue:43828", + "accept": false, + "reason": "Grouped_mm tracing failure and autocast dtype mismatch are different runtime issues." + }, + { + "left": "issue:44841", + "right": "issue:45084", + "accept": false, + "reason": "Voxtral processor failure and template compilation error are unrelated." + }, + { + "left": "issue:45507", + "right": "issue:46869", + "accept": false, + "reason": "GraniteMoEHybrid invalid method call and YOLO multi-GPU device-map failure are different model-specific bugs." + }, + { + "left": "issue:43404", + "right": "issue:43866", + "accept": false, + "reason": "Mistral3 lm_head tying bug and Ovis2 checkpoint corruption are unrelated model issues." + }, + { + "left": "issue:41093", + "right": "issue:41762", + "accept": false, + "reason": "Both are tensor/indexing failures, but the exact symptoms and contexts differ enough that they do not look like the same bug." + }, + { + "left": "issue:45399", + "right": "issue:46661", + "accept": false, + "reason": "Flash-attn2 fallback logic and remote-code config_class attribute errors are separate code paths." + }, + { + "left": "issue:43572", + "right": "issue:45070", + "accept": false, + "reason": "Missing pad_token_idx in StableLmConfig and PretrainedConfig pydantic regression are unrelated configuration issues." + }, + { + "left": "issue:45800", + "right": "issue:45823", + "accept": false, + "reason": "Torch/transformers compatibility and Gemma4 device mismatch are different failures." + } + ] + }, + "evaluator_result": { + "accept": true, + "feedback": "The cluster summary is grounded in the packet and stays conservative. The accepted pair 45812/45920 is well supported by the titles: both describe the same AutoTokenizer v5 token-ID regression, with 45920 clearly the broader umbrella report. The other soft-edge rejections are appropriately cautious given only symptom-level similarity. No mergeability issues apply here since this is an issue-clustering judgment, not a PR-pair merge decision." + }, + "evaluator_used": true, + "retried": false, + "accepted_nontrivial_soft_edge": true, + "error_kind": null, + "error_message": null + }, + { + "label": "issue", + "component_index": 1, + "component_count": 27, + "review_unit_index": 3, + "review_unit_count": 63, + "status": "reviewed", + "reason": null, + "source": "llm", + "cache_hit": false, + "model": "gpt-5.4-mini?service_tier=flex", + "cluster_id": "cluster-10000-514", + "nodes": [ + "issue:39186", + "issue:40990", + "issue:41093", + "issue:42915", + "issue:43054", + "issue:43296", + "issue:43366", + "issue:43404", + "issue:43531", + "issue:43541", + "issue:43596", + "issue:43645", + "issue:43716", + "issue:43828", + "issue:43873", + "issue:43994", + "issue:44291", + "issue:44360", + "issue:44387", + "issue:44479", + "issue:44492", + "issue:44512", + "issue:44530", + "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:46165", + "issue:46294", + "issue:46463", + "issue:46468", + "issue:46640", + "issue:46661", + "issue:46869" + ], + "soft_pairs": [ + "issue:43296|issue:43366", + "issue:44805|issue:44918", + "issue:43541|issue:45161", + "issue:43828|issue:45464", + "issue:43366|issue:45464", + "issue:44291|issue:45070", + "issue:44387|issue:45005", + "issue:43531|issue:45070", + "issue:41093|issue:43596", + "issue:43716|issue:46294", + "issue:45020|issue:45800", + "issue:44479|issue:46294", + "issue:43054|issue:43994", + "issue:46468|issue:46869", + "issue:43596|issue:46165", + "issue:45759|issue:45902", + "issue:46640|issue:46661", + "issue:43716|issue:44918", + "issue:39186|issue:44918", + "issue:43645|issue:45399", + "issue:40990|issue:44960", + "issue:45632|issue:45698", + "issue:42915|issue:45399", + "issue:43404|issue:45005", + "issue:44360|issue:44512", + "issue:43873|issue:45005", + "issue:44492|issue:44512", + "issue:44530|issue:45005", + "issue:44918|issue:45507", + "issue:45183|issue:46463", + "issue:44863|issue:45005" + ], + "prepared_review_unit_hash": "529d0e998713d73792452d9ca998fe2d0065aa9d10920213375f62ece4522669", + "component_budget": { + "node_count": 514, + "item_count": 514, + "soft_pair_count": 1891, + "serialized_chars": 980895, + "estimated_input_tokens": 245224, + "estimated_eval_tokens": 490704 + }, + "budget": { + "node_count": 47, + "item_count": 47, + "soft_pair_count": 31, + "serialized_chars": 36244, + "estimated_input_tokens": 9061, + "estimated_eval_tokens": 18378 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": true, + "analyst_result": { + "summary": "The cluster is mostly heterogeneous: several unrelated bug reports across model families, quantization, multimodal inputs, remote code, and validation. A few soft edges look like plausible duplicates, but there is no single global duplicate group or canonical issue/PR for the whole cluster.", + "confidence": 0.79, + "canonical_issue_reason": null, + "canonical_pr_reason": null, + "best_issue_reason": "No clear global best issue; the items split into multiple unrelated bug families, so picking one as canonical would be misleading.", + "best_pr_reason": "No PRs are present in this cluster.", + "soft_edge_verdicts": [ + { + "left": "issue:43296", + "right": "issue:43366", + "accept": false, + "reason": "Different failures: PaddleOCR-VL load error vs GGUF gpt-oss architecture support." + }, + { + "left": "issue:44805", + "right": "issue:44918", + "accept": false, + "reason": "Different code paths: mask/tensor shape mismatch vs Qwen3.5 embedding unpacking with SFT trainer." + }, + { + "left": "issue:43541", + "right": "issue:45161", + "accept": false, + "reason": "Different model/runtime problems: torch dynamo grouped_mm tracing vs TP support for GPT-OSS MoE." + }, + { + "left": "issue:43828", + "right": "issue:45464", + "accept": false, + "reason": "One is an autocast dtype mismatch in Phi-tiny-MoE; the other is a Qwen3.5 streaming chat/completions failure." + }, + { + "left": "issue:43366", + "right": "issue:45464", + "accept": false, + "reason": "GGUF gpt-oss support and Qwen3.5 streaming inference are unrelated issues." + }, + { + "left": "issue:44291", + "right": "issue:45070", + "accept": false, + "reason": "Different breakages: init_empty_weights argument handling vs PretrainedConfig/pydantic field changes." + }, + { + "left": "issue:44387", + "right": "issue:45005", + "accept": false, + "reason": "Quantization OOM under v5.x is not the same bug as tied-weights handling on translation models." + }, + { + "left": "issue:43531", + "right": "issue:45070", + "accept": false, + "reason": "Qwen3-MoE sliding_window behavior is unrelated to the PretrainedConfig pydantic regression." + }, + { + "left": "issue:41093", + "right": "issue:43596", + "accept": false, + "reason": "Both are index errors, but they affect different model/setup paths and do not look like the same bug." + }, + { + "left": "issue:43716", + "right": "issue:46294", + "accept": false, + "reason": "Different multimodal mismatches: image preprocessor dtype vs video feature/token mismatch." + }, + { + "left": "issue:45020", + "right": "issue:45800", + "accept": false, + "reason": "Remote-code model breakage and torch/transformers version incompatibility are distinct issues." + }, + { + "left": "issue:44479", + "right": "issue:46294", + "accept": true, + "reason": "Both target the Qwen3.5/Qwen-VL video-input path and describe the same kind of video token/feature regression." + }, + { + "left": "issue:43054", + "right": "issue:43994", + "accept": true, + "reason": "Same Siglip2 family with closely related bad-output symptoms; both point to the model/pipeline producing degraded or nonsensical embeddings." + }, + { + "left": "issue:46468", + "right": "issue:46869", + "accept": false, + "reason": "Official Nemotron example breakage is unrelated to YOLO multi-GPU device-map failure." + }, + { + "left": "issue:43596", + "right": "issue:46165", + "accept": false, + "reason": "Deepspeed ZeRO3/BertModel index failure is unrelated to resuming remote-code training." + }, + { + "left": "issue:45759", + "right": "issue:45902", + "accept": false, + "reason": "Composite multimodal config unwrapping and uninitialized rotary embedding memory are different bugs." + }, + { + "left": "issue:46640", + "right": "issue:46661", + "accept": false, + "reason": "Flash Attention 2 availability issue is unrelated to remote-code model_class/config_class loading." + }, + { + "left": "issue:43716", + "right": "issue:44918", + "accept": false, + "reason": "Image preprocessor dtype mismatch and SFT embedding unpacking are different multimodal/training issues." + }, + { + "left": "issue:39186", + "right": "issue:44918", + "accept": false, + "reason": "FSDP weight-shape RuntimeError is unrelated to Qwen3.5 embedding unpacking." + }, + { + "left": "issue:43645", + "right": "issue:45399", + "accept": false, + "reason": "Jupyter custom-model initialization and flash-attn2 fallback gating are not the same bug." + }, + { + "left": "issue:40990", + "right": "issue:44960", + "accept": false, + "reason": "High perplexity on gpt-oss-20b and GLM5 are unrelated model-quality reports." + }, + { + "left": "issue:45632", + "right": "issue:45698", + "accept": true, + "reason": "Both concern remote-code/custom-module resolution picking the wrong cached or loaded module for local/saved models." + }, + { + "left": "issue:42915", + "right": "issue:45399", + "accept": false, + "reason": "Qwen3-MoE FP8 config failure is unrelated to flash-attn2 fallback checks." + }, + { + "left": "issue:43404", + "right": "issue:45005", + "accept": false, + "reason": "Mistral3 tied lm_head weights and translation-model tied-weights regressions are different model families and failure modes." + }, + { + "left": "issue:44360", + "right": "issue:44512", + "accept": false, + "reason": "DSA indexer behavior and docs cleanup are unrelated." + }, + { + "left": "issue:43873", + "right": "issue:45005", + "accept": false, + "reason": "Quantization offloading and tied-weights handling do not describe the same underlying bug." + }, + { + "left": "issue:44492", + "right": "issue:44512", + "accept": false, + "reason": "A cache-strategy typo is not the same as outdated CLI docs." + }, + { + "left": "issue:44530", + "right": "issue:45005", + "accept": false, + "reason": "PagedAttentionCache linear_attention crash is unrelated to tied-weights regressions." + }, + { + "left": "issue:44918", + "right": "issue:45507", + "accept": false, + "reason": "Qwen3.5 embedding unpacking and GraniteMoEHybrid invalid method calls are different code paths." + }, + { + "left": "issue:45183", + "right": "issue:46463", + "accept": false, + "reason": "Broader input_ids/inputs_embeds XOR validation issue is not the same as T5 decoder-input error messaging." + }, + { + "left": "issue:44863", + "right": "issue:45005", + "accept": false, + "reason": "NemotronH checkpoint-loading failure is unrelated to tied-weights issues in translation models." + } + ] + }, + "evaluator_result": { + "accept": true, + "feedback": "The summary is grounded in the packet and stays conservative about cluster heterogeneity. The accepted soft edges are plausible duplicates and do not overstate certainty, and there are no PR pairs to assess." + }, + "evaluator_used": true, + "retried": false, + "accepted_nontrivial_soft_edge": true, + "error_kind": null, + "error_message": null + }, + { + "label": "issue", + "component_index": 1, + "component_count": 27, + "review_unit_index": 4, + "review_unit_count": 63, + "status": "reviewed", + "reason": null, + "source": "llm", + "cache_hit": false, + "model": "gpt-5.4-mini?service_tier=flex", + "cluster_id": "cluster-10000-514", + "nodes": [ + "issue:24643", + "issue:28530", + "issue:36331", + "issue:38175", + "issue:39692", + "issue:41553", + "issue:41762", + "issue:42915", + "issue:42947", + "issue:43054", + "issue:43493", + "issue:43643", + "issue:43644", + "issue:43782", + "issue:43824", + "issue:43828", + "issue:43856", + "issue:43872", + "issue:43883", + "issue:43950", + "issue:43975", + "issue:44292", + "issue:44387", + "issue:44403", + "issue:44451", + "issue:44488", + "issue:44534", + "issue:44589", + "issue:44910", + "issue:44928", + "issue:44960", + "issue:44991", + "issue:45020", + "issue:45399", + "issue:45520", + "issue:45663", + "issue:45685", + "issue:45701", + "issue:45823", + "issue:45859", + "issue:45902", + "issue:45920", + "issue:45987", + "issue:46489", + "issue:46491", + "issue:46858", + "issue:46899", + "issue:46932" + ], + "soft_pairs": [ + "issue:43975|issue:44292", + "issue:46489|issue:46491", + "issue:39692|issue:43054", + "issue:43950|issue:44534", + "issue:36331|issue:41762", + "issue:43644|issue:43950", + "issue:43872|issue:44589", + "issue:46899|issue:46932", + "issue:43828|issue:44928", + "issue:43054|issue:45701", + "issue:45920|issue:46489", + "issue:38175|issue:43054", + "issue:45987|issue:46858", + "issue:43950|issue:45902", + "issue:43054|issue:43493", + "issue:44910|issue:45859", + "issue:42947|issue:44387", + "issue:45663|issue:45902", + "issue:44403|issue:44451", + "issue:41553|issue:43883", + "issue:24643|issue:28530", + "issue:43975|issue:45520", + "issue:44403|issue:44488", + "issue:43643|issue:45020", + "issue:41762|issue:42915", + "issue:42947|issue:43856", + "issue:44991|issue:45399", + "issue:43824|issue:43872", + "issue:45685|issue:45823", + "issue:43782|issue:44960" + ], + "prepared_review_unit_hash": "5bdf928a82e2c4909bb620bdb7ff250696276a1886fdaa829b47c1327a2b9093", + "component_budget": { + "node_count": 514, + "item_count": 514, + "soft_pair_count": 1891, + "serialized_chars": 980895, + "estimated_input_tokens": 245224, + "estimated_eval_tokens": 490704 + }, + "budget": { + "node_count": 48, + "item_count": 48, + "soft_pair_count": 30, + "serialized_chars": 36903, + "estimated_input_tokens": 9226, + "estimated_eval_tokens": 18708 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": true, + "analyst_result": { + "summary": "This cluster is mostly a collection of unrelated Transformers bug reports. I only found three true duplicate groups: non-persistent buffer corruption on load, the StaticCache.get_seq_length() return-type bug, and the Gemma4 quantized vision/audio dtype-cast bug. The rest are only loosely similar by subsystem or model family, not the same underlying issue.", + "confidence": 0.87, + "canonical_issue_reason": "issue:43950 is the clearest and most complete report of the non-persistent-buffer corruption regression, and it has the strongest duplicate evidence from closely matching follow-up reports.", + "canonical_pr_reason": null, + "best_issue_reason": "issue:43950 is the best overall issue representative here because it precisely describes a concrete regression and is the center of one of the true duplicate groups.", + "best_pr_reason": null, + "soft_edge_verdicts": [ + { + "left": "issue:43975", + "right": "issue:44292", + "accept": false, + "reason": "Both are error reports, but they describe different failures on different models/code paths (DeepSeek tokenizer detokenization vs Qwen NVFP4 loading)." + }, + { + "left": "issue:46489", + "right": "issue:46491", + "accept": false, + "reason": "Both are tokenizer regressions, but the symptoms and affected models differ; no clear evidence they come from the same concrete bug." + }, + { + "left": "issue:39692", + "right": "issue:43054", + "accept": false, + "reason": "One is a broken docs example/model-processor mismatch and quantization failure; the other is degraded SigLIP2 text embeddings. Different problems." + }, + { + "left": "issue:43950", + "right": "issue:44534", + "accept": true, + "reason": "Same underlying regression: non-persistent buffers are being mishandled during from_pretrained/load in Transformers v5." + }, + { + "left": "issue:36331", + "right": "issue:41762", + "accept": false, + "reason": "Unrelated failures: a Trainer custom loss signature change vs a DeepSpeed ZeRO-3 model-loading index error." + }, + { + "left": "issue:43644", + "right": "issue:43950", + "accept": true, + "reason": "Both report the same non-persistent-buffer corruption bug in Transformers 5.x." + }, + { + "left": "issue:43872", + "right": "issue:44589", + "accept": false, + "reason": "Both are runtime/type errors, but one is bitsandbytes initialization incompatibility and the other is a float8 storage lookup issue." + }, + { + "left": "issue:46899", + "right": "issue:46932", + "accept": true, + "reason": "These describe the same Gemma4 quantized vision/audio embedder dtype-casting bug; 46932 explicitly says it is the same as another report." + }, + { + "left": "issue:43828", + "right": "issue:44928", + "accept": false, + "reason": "Different Qwen/MoE/autocast issues with different symptoms and code paths; not the same bug." + }, + { + "left": "issue:43054", + "right": "issue:45701", + "accept": false, + "reason": "One is a model quality discrepancy, the other a tokenizer-version behavior change. Too broad to merge as duplicates." + }, + { + "left": "issue:45920", + "right": "issue:46489", + "accept": false, + "reason": "Both concern tokenizer regressions in v5, but they affect different models and report different failure modes." + }, + { + "left": "issue:38175", + "right": "issue:43054", + "accept": false, + "reason": "Both mention SigLIP2, but one is zero probabilities and the other is worse text embeddings; not the same concrete defect." + }, + { + "left": "issue:45987", + "right": "issue:46858", + "accept": true, + "reason": "Same cache API bug: get_seq_length() returns a tensor instead of a plain int, causing downstream static-cache arithmetic issues." + }, + { + "left": "issue:43950", + "right": "issue:45902", + "accept": false, + "reason": "One is about buffer corruption on load; the other is uninitialized memory after meta/to_empty materialization. Related area, different bug." + }, + { + "left": "issue:43054", + "right": "issue:43493", + "accept": false, + "reason": "Both involve SigLIP2, but one is poor embedding quality and the other is an implementation discrepancy vs JAX. Not the same issue." + }, + { + "left": "issue:44910", + "right": "issue:45859", + "accept": false, + "reason": "Flash-attention packed-sequence misinterpretation vs CPU-offload incompatibility in rotary embedding are distinct problems." + }, + { + "left": "issue:42947", + "right": "issue:44387", + "accept": false, + "reason": "Gradient checkpointing/LoRA behavior and int4 quantization memory OOM are separate training issues." + }, + { + "left": "issue:45663", + "right": "issue:45902", + "accept": false, + "reason": "Different Gemma/Qwen bugs: FSDP2 shared_kv_states rebuilding vs uninitialized inv_freq after materialization." + }, + { + "left": "issue:44403", + "right": "issue:44451", + "accept": false, + "reason": "One is a noisy load message, the other is a hard model-loading failure. No evidence of the same root cause." + }, + { + "left": "issue:41553", + "right": "issue:43883", + "accept": false, + "reason": "AutoTokenizer error handling for Voxtral and Molmo weight-tying metadata are unrelated." + }, + { + "left": "issue:24643", + "right": "issue:28530", + "accept": false, + "reason": "DeepSpeed training weight-shape error and early-stopping metric selection are unrelated." + }, + { + "left": "issue:43975", + "right": "issue:45520", + "accept": false, + "reason": "DeepSeek detokenization regression and Python 3.13 flash_attn import error are unrelated." + }, + { + "left": "issue:44403", + "right": "issue:44488", + "accept": false, + "reason": "Both mention model loading, but there is no concrete evidence they share one bug beyond broad similarity." + }, + { + "left": "issue:43643", + "right": "issue:45020", + "accept": false, + "reason": "45020 is a broad remote_code regression bucket; 43643 is a specific missing-fields symptom. Too broad to call duplicates confidently." + }, + { + "left": "issue:41762", + "right": "issue:42915", + "accept": false, + "reason": "Gemma3 ZeRO-3 loading error and Qwen3 MoE FP8 failure are different models and failure modes." + }, + { + "left": "issue:42947", + "right": "issue:43856", + "accept": false, + "reason": "LoRA checkpointing inefficiency and Qwen3 MoE memory usage are separate performance issues." + }, + { + "left": "issue:44991", + "right": "issue:45399", + "accept": false, + "reason": "Tokenizer loading failure and flash-attn2 fallback checks are unrelated subsystems." + }, + { + "left": "issue:43824", + "right": "issue:43872", + "accept": false, + "reason": "Different incompatibilities: missing Qwen2_5_VL export vs bitsandbytes Int8Params keyword mismatch." + }, + { + "left": "issue:45685", + "right": "issue:45823", + "accept": false, + "reason": "MPS histogram/int issue in MoE and Gemma4 device-mapping mismatch are unrelated." + }, + { + "left": "issue:43782", + "right": "issue:44960", + "accept": false, + "reason": "Different model families and different failures; no evidence of the same underlying bug." + } + ] + }, + "evaluator_result": { + "accept": true, + "feedback": "The summary is grounded in the packet and the true-duplicate calls are conservative. The three accepted groups are well-supported by the titles/details: the non-persistent-buffer regression around `from_pretrained`, the `StaticCache.get_seq_length()` return-type bug, and the Gemma4 quantized vision/audio dtype-cast bug. The remaining pairs are appropriately treated as similar-but-distinct rather than duplicates." + }, + "evaluator_used": true, + "retried": false, + "accepted_nontrivial_soft_edge": true, + "error_kind": null, + "error_message": null + }, + { + "label": "issue", + "component_index": 1, + "component_count": 27, + "review_unit_index": 5, + "review_unit_count": 63, + "status": "reviewed", + "reason": null, + "source": "llm", + "cache_hit": false, + "model": "gpt-5.4-mini?service_tier=flex", + "cluster_id": "cluster-10000-514", + "nodes": [ + "issue:24643", + "issue:30064", + "issue:36331", + "issue:41093", + "issue:41553", + "issue:42491", + "issue:42947", + "issue:43257", + "issue:43404", + "issue:43792", + "issue:43827", + "issue:43828", + "issue:43854", + "issue:43872", + "issue:43881", + "issue:43883", + "issue:43940", + "issue:43950", + "issue:44220", + "issue:44291", + "issue:44368", + "issue:44479", + "issue:44509", + "issue:44560", + "issue:44661", + "issue:44805", + "issue:44863", + "issue:44928", + "issue:44991", + "issue:45020", + "issue:45081", + "issue:45305", + "issue:45313", + "issue:45356", + "issue:45399", + "issue:45488", + "issue:45588", + "issue:45715", + "issue:45759", + "issue:45823", + "issue:45923", + "issue:46315", + "issue:46489", + "issue:46661", + "issue:46869", + "issue:46894", + "issue:46897" + ], + "soft_pairs": [ + "issue:45081|issue:45399", + "issue:41553|issue:43881", + "issue:45020|issue:45823", + "issue:44661|issue:45020", + "issue:43828|issue:45823", + "issue:42491|issue:44368", + "issue:45020|issue:45356", + "issue:43881|issue:43883", + "issue:45715|issue:46315", + "issue:43872|issue:46869", + "issue:45759|issue:46661", + "issue:43872|issue:44291", + "issue:43827|issue:44509", + "issue:45488|issue:46489", + "issue:45020|issue:46661", + "issue:43404|issue:43950", + "issue:46894|issue:46897", + "issue:43792|issue:44220", + "issue:43257|issue:45305", + "issue:36331|issue:41093", + "issue:42947|issue:45923", + "issue:43940|issue:45313", + "issue:43854|issue:44863", + "issue:24643|issue:30064", + "issue:44479|issue:44560", + "issue:44928|issue:45588", + "issue:41093|issue:44805", + "issue:41553|issue:44991" + ], + "prepared_review_unit_hash": "361d85e7f0e4234114f8ded4236e537081af363a6722b16aa037c7e0af9d96d8", + "component_budget": { + "node_count": 514, + "item_count": 514, + "soft_pair_count": 1891, + "serialized_chars": 980895, + "estimated_input_tokens": 245224, + "estimated_eval_tokens": 490704 + }, + "budget": { + "node_count": 47, + "item_count": 47, + "soft_pair_count": 28, + "serialized_chars": 36178, + "estimated_input_tokens": 9045, + "estimated_eval_tokens": 18346 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": true, + "analyst_result": { + "summary": "Most pairs are distinct model- or subsystem-specific regressions, but a few are clear duplicates: the v5 docs pipeline reference issue, the `_is_hf_initialized` loading regression, the DeepSeek tokenizer regression, and the training-loss double-shift reports. The broad remote-code regression report is the best umbrella issue for the overall cluster.", + "confidence": 0.74, + "canonical_issue_reason": "Issue 45020 is the broadest, most central report of the recurring v5 compatibility regressions around model/tokenizer loading, so it best represents the cluster.", + "canonical_pr_reason": null, + "best_issue_reason": "45020 is the strongest umbrella issue for the shared regression theme; the other issues are narrower manifestations or adjacent edge cases.", + "best_pr_reason": null, + "soft_edge_verdicts": [ + { + "left": "issue:45081", + "right": "issue:45399", + "accept": false, + "reason": "Different bugs: a Mistral regex patch crash vs. a flash-attn fallback gating problem." + }, + { + "left": "issue:41553", + "right": "issue:43881", + "accept": false, + "reason": "Both are loading failures, but they concern different models and different failure modes." + }, + { + "left": "issue:45020", + "right": "issue:45823", + "accept": false, + "reason": "Remote-code loading regression vs. Gemma4 device-map mismatch are separate code paths." + }, + { + "left": "issue:44661", + "right": "issue:45020", + "accept": false, + "reason": "Tokenizer-mapping edge case vs. broad remote-code regression; not the same bug." + }, + { + "left": "issue:43828", + "right": "issue:45823", + "accept": false, + "reason": "Different runtime failures: dtype mismatch under autocast vs. device placement mismatch." + }, + { + "left": "issue:42491", + "right": "issue:44368", + "accept": false, + "reason": "Different LoRA/tie-weights behaviors; one is a compatibility break, the other is a warning about config flags." + }, + { + "left": "issue:45020", + "right": "issue:45356", + "accept": false, + "reason": "Remote-code model loading breakage and a tokenizer codec regression are unrelated." + }, + { + "left": "issue:43881", + "right": "issue:43883", + "accept": false, + "reason": "Different model-loading failures with different missing pieces in the implementation." + }, + { + "left": "issue:45715", + "right": "issue:46315", + "accept": false, + "reason": "Both are tokenizer performance regressions, but they hit different methods and likely different hot paths." + }, + { + "left": "issue:43872", + "right": "issue:46869", + "accept": false, + "reason": "Unexpected kwarg in parameter initialization vs. YOLO multi-GPU device-map failure are unrelated." + }, + { + "left": "issue:45759", + "right": "issue:46661", + "accept": false, + "reason": "Different AutoModel API problems: config unwrapping vs. missing config_class on remote-code models." + }, + { + "left": "issue:43872", + "right": "issue:44291", + "accept": true, + "reason": "Same `_is_hf_initialized` keyword-argument regression during weight initialization/loading; 43872 is a bitsandbytes manifestation." + }, + { + "left": "issue:43827", + "right": "issue:44509", + "accept": true, + "reason": "Same docs regression: stale v5 pipeline-task references remain in the documentation." + }, + { + "left": "issue:45488", + "right": "issue:46489", + "accept": true, + "reason": "Same tokenizer regression in v5 around preserving tokenizer.json pre-tokenizers; both break DeepSeek-family tokenization." + }, + { + "left": "issue:45020", + "right": "issue:46661", + "accept": false, + "reason": "Remote-code regressions and missing config_class are related only broadly, not the same concrete bug." + }, + { + "left": "issue:43404", + "right": "issue:43950", + "accept": false, + "reason": "Both are tied to weight handling, but one is a tied-weight bug and the other is a non-persistent buffer regression." + }, + { + "left": "issue:46894", + "right": "issue:46897", + "accept": true, + "reason": "Same training-loss double-shift bug pattern; only the affected model family differs." + }, + { + "left": "issue:43792", + "right": "issue:44220", + "accept": false, + "reason": "Whisper runtime failure and _torch_extract_fbank_features are too generic to treat as the same concrete bug." + }, + { + "left": "issue:43257", + "right": "issue:45305", + "accept": false, + "reason": "DeepSpeed weight-conversion failure vs. gradient averaging bug are different training issues." + }, + { + "left": "issue:36331", + "right": "issue:41093", + "accept": false, + "reason": "Unexpected compute_loss kwarg and mask shape mismatch are unrelated failures." + }, + { + "left": "issue:42947", + "right": "issue:45923", + "accept": false, + "reason": "Both involve gradient checkpointing, but one is ineffective LoRA checkpointing and the other is a missing capability flag." + }, + { + "left": "issue:43940", + "right": "issue:45313", + "accept": false, + "reason": "Both are ZeRO-3 loading failures, but the model/module specifics differ and the bugs do not clearly collapse to one fix." + }, + { + "left": "issue:43854", + "right": "issue:44863", + "accept": false, + "reason": "Different model-loading failures for different architectures." + }, + { + "left": "issue:24643", + "right": "issue:30064", + "accept": false, + "reason": "DeepSpeed training error and segmentation-map processing are unrelated." + }, + { + "left": "issue:44479", + "right": "issue:44560", + "accept": false, + "reason": "Both are video-input regressions, but the affected models and failure modes differ." + }, + { + "left": "issue:44928", + "right": "issue:45588", + "accept": false, + "reason": "Gradient-explosion RLHF bug vs. flash-attention integration crash are separate problems." + }, + { + "left": "issue:41093", + "right": "issue:44805", + "accept": false, + "reason": "Same error shape family, but not enough evidence they are the same masking bug." + }, + { + "left": "issue:41553", + "right": "issue:44991", + "accept": false, + "reason": "Tokenizer loading failures for different models; the error origins are different." + } + ] + }, + "evaluator_result": { + "accept": true, + "feedback": "The cluster summary is grounded in the packet, and the accepted soft-edge pairs are all conservative duplicates/near-duplicates based on the issue titles and descriptions. The rejected pairs are also framed cautiously as different bugs or only loosely related. The choice of 45020 as the umbrella issue is defensible as a broad remote-code regression anchor." + }, + "evaluator_used": true, + "retried": false, + "accepted_nontrivial_soft_edge": true, + "error_kind": null, + "error_message": null + }, + { + "label": "issue", + "component_index": 1, + "component_count": 27, + "review_unit_index": 6, + "review_unit_count": 63, + "status": "reviewed", + "reason": null, + "source": "llm", + "cache_hit": false, + "model": "gpt-5.4-mini?service_tier=flex", + "cluster_id": "cluster-10000-514", + "nodes": [ + "issue:36010", + "issue:38175", + "issue:41553", + "issue:42491", + "issue:42617", + "issue:42915", + "issue:43366", + "issue:43531", + "issue:43550", + "issue:43596", + "issue:43638", + "issue:43645", + "issue:43646", + "issue:43716", + "issue:43720", + "issue:43824", + "issue:43828", + "issue:43950", + "issue:43957", + "issue:43994", + "issue:44155", + "issue:44488", + "issue:44530", + "issue:44560", + "issue:44589", + "issue:44661", + "issue:44683", + "issue:44805", + "issue:44863", + "issue:44877", + "issue:44910", + "issue:44918", + "issue:45030", + "issue:45325", + "issue:45584", + "issue:45663", + "issue:45672", + "issue:45685", + "issue:45812", + "issue:45823", + "issue:45834", + "issue:45902", + "issue:46402", + "issue:46468", + "issue:46489", + "issue:46515", + "issue:46858", + "issue:46899" + ], + "soft_pairs": [ + "issue:43716|issue:46468", + "issue:44530|issue:44863", + "issue:43366|issue:43828", + "issue:44863|issue:46402", + "issue:43366|issue:43531", + "issue:43596|issue:46515", + "issue:45325|issue:45902", + "issue:41553|issue:44488", + "issue:43366|issue:44589", + "issue:43638|issue:46515", + "issue:44155|issue:45584", + "issue:45663|issue:45823", + "issue:42915|issue:43957", + "issue:43824|issue:43957", + "issue:43716|issue:44560", + "issue:43950|issue:45834", + "issue:42617|issue:43366", + "issue:43645|issue:43646", + "issue:44560|issue:45685", + "issue:44910|issue:46858", + "issue:36010|issue:42915", + "issue:44560|issue:44918", + "issue:43550|issue:44910", + "issue:44683|issue:46515", + "issue:45812|issue:46489", + "issue:45823|issue:46515", + "issue:44661|issue:45823", + "issue:44560|issue:44805", + "issue:38175|issue:43994", + "issue:44877|issue:45030", + "issue:43638|issue:44661", + "issue:43596|issue:44661", + "issue:45672|issue:46899", + "issue:42491|issue:43720" + ], + "prepared_review_unit_hash": "f141435a4385435e2a62d615d06d759ac6f0ad30f2d0ae0f340c9528e060e128", + "component_budget": { + "node_count": 514, + "item_count": 514, + "soft_pair_count": 1891, + "serialized_chars": 980895, + "estimated_input_tokens": 245224, + "estimated_eval_tokens": 490704 + }, + "budget": { + "node_count": 48, + "item_count": 48, + "soft_pair_count": 34, + "serialized_chars": 38047, + "estimated_input_tokens": 9512, + "estimated_eval_tokens": 19280 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": true, + "analyst_result": { + "summary": "The set is mostly unrelated model/regression reports with two clear duplicate-like pairs: custom model initialization in Transformers 5.0 and bad SigLIP2 inference output.", + "confidence": 0.84, + "canonical_issue_reason": "issue:43646 is the best canonical issue because it states the shared Transformers 5.0 custom-model initialization regression in the broadest, cleanest form; issue:43645 is just a notebook-specific variant.", + "canonical_pr_reason": null, + "best_issue_reason": "issue:43646 is the strongest representative of the cluster: concise, broad, and clearly about the underlying regression rather than an environment-specific symptom.", + "best_pr_reason": null, + "soft_edge_verdicts": [ + { + "left": "issue:43716", + "right": "issue:46468", + "accept": false, + "reason": "Different bugs and code paths: image preprocessor/model dtype mismatch vs a broken Nemotron example." + }, + { + "left": "issue:44530", + "right": "issue:44863", + "accept": false, + "reason": "Both are loading/runtime failures, but they affect different models and mechanisms (PagedAttentionCache vs NemotronH checkpoint loading)." + }, + { + "left": "issue:43366", + "right": "issue:43828", + "accept": false, + "reason": "Unrelated symptoms: GGUF architecture support vs autocast dtype mismatch on Phi-tiny-MoE." + }, + { + "left": "issue:44863", + "right": "issue:46402", + "accept": false, + "reason": "Different model-loading failures for different models; no shared concrete bug." + }, + { + "left": "issue:43366", + "right": "issue:43531", + "accept": false, + "reason": "Same general Qwen/GGUF space, but different underlying issues: model architecture support vs sliding-window behavior." + }, + { + "left": "issue:43596", + "right": "issue:46515", + "accept": false, + "reason": "The first is a specific Bert/deepspeed zero3 IndexError; the second is a broad torch version incompatibility report." + }, + { + "left": "issue:45325", + "right": "issue:45902", + "accept": false, + "reason": "Different mechanisms: Qwen2.5-VL rope scaling for still images vs Qwen3_5Moe rotary embedding materialization bug." + }, + { + "left": "issue:41553", + "right": "issue:44488", + "accept": false, + "reason": "Both are load failures, but for different tokenizer/model targets and likely different root causes." + }, + { + "left": "issue:43366", + "right": "issue:44589", + "accept": false, + "reason": "GGUF architecture support and a Float8 storage lookup error are unrelated." + }, + { + "left": "issue:43638", + "right": "issue:46515", + "accept": false, + "reason": "A Bert/deepspeed zero3 index error is not the same as a generic torch compatibility issue." + }, + { + "left": "issue:44155", + "right": "issue:45584", + "accept": false, + "reason": "Different product areas and failures: AudioFlamingo3 batching leak vs Whisper empty-transcription generation." + }, + { + "left": "issue:45663", + "right": "issue:45823", + "accept": false, + "reason": "Same model family, but training FSDP2 KeyError and inference device-map mismatch are different bugs." + }, + { + "left": "issue:42915", + "right": "issue:43957", + "accept": false, + "reason": "Qwen3Moe FP8 failure and meta-device loading regression are different code-path problems." + }, + { + "left": "issue:43824", + "right": "issue:43957", + "accept": false, + "reason": "Qwen2.5-VL import error is unrelated to the generic meta-device loading bug." + }, + { + "left": "issue:43716", + "right": "issue:44560", + "accept": false, + "reason": "Mistral-3 dtype mismatch and Qwen3-vl video StopIteration are different failures." + }, + { + "left": "issue:43950", + "right": "issue:45834", + "accept": false, + "reason": "Strict config validation for granite_speech and Kosmos2.5 OCR index error are unrelated." + }, + { + "left": "issue:42617", + "right": "issue:43366", + "accept": false, + "reason": "Not the same bug: running 3d_parallel.py vs GGUF gpt-oss architecture support." + }, + { + "left": "issue:43645", + "right": "issue:43646", + "accept": true, + "reason": "Same underlying Transformers 5.0 custom-model initialization regression; the notebook wording is just a narrower symptom." + }, + { + "left": "issue:44560", + "right": "issue:45685", + "accept": false, + "reason": "Qwen3-vl video StopIteration and an MPS histogram-not-implemented error are unrelated." + }, + { + "left": "issue:44910", + "right": "issue:46858", + "accept": false, + "reason": "Both involve cache/attention, but one is a flash-attention memory access bug and the other is a static-cache seq_length arithmetic issue." + }, + { + "left": "issue:36010", + "right": "issue:42915", + "accept": false, + "reason": "ImportError for GenerationMixin and Qwen3Moe FP8 failure are unrelated." + }, + { + "left": "issue:44560", + "right": "issue:44918", + "accept": false, + "reason": "Different symptoms and code paths: video StopIteration vs TRL SFT unpacking failure." + }, + { + "left": "issue:43550", + "right": "issue:44910", + "accept": false, + "reason": "Bamba torch.compile+SDPA and Qwen3.5 flash-attention memory access are different model/runtime bugs." + }, + { + "left": "issue:44683", + "right": "issue:46515", + "accept": false, + "reason": "Compiled flex_attention on torch>=2.9 is not the same as a broad torch 2.4.1 incompatibility report." + }, + { + "left": "issue:45812", + "right": "issue:46489", + "accept": false, + "reason": "Both are tokenizer regressions, but they affect different model families and likely different tokenizer code paths." + }, + { + "left": "issue:45823", + "right": "issue:46515", + "accept": false, + "reason": "Gemma4 device-map mismatch and a general torch version incompatibility are not the same issue." + }, + { + "left": "issue:44661", + "right": "issue:45823", + "accept": false, + "reason": "Tokenizer-mapping failure and Gemma4 device mismatch are unrelated." + }, + { + "left": "issue:44560", + "right": "issue:44805", + "accept": false, + "reason": "Qwen3-vl video StopIteration does not match the generic mask/tensor shape IndexError." + }, + { + "left": "issue:38175", + "right": "issue:43994", + "accept": true, + "reason": "Same model and same symptom family: incorrect SigLIP2 inference outputs, with one phrased as zero probabilities and the other as nonsensical pipeline/AutoModel results." + }, + { + "left": "issue:44877", + "right": "issue:45030", + "accept": false, + "reason": "Different config-validation problems affecting different model types." + }, + { + "left": "issue:43638", + "right": "issue:44661", + "accept": false, + "reason": "Bert/deepspeed zero3 index error and tokenizer-mapping changes are unrelated." + }, + { + "left": "issue:43596", + "right": "issue:44661", + "accept": false, + "reason": "A zero3/Bert IndexError is not the same as the add-new-model-like tokenizer mapping bug." + }, + { + "left": "issue:45672", + "right": "issue:46899", + "accept": false, + "reason": "Both are Gemma4 quantization-related, but they describe different dtype/weight-loading failures." + }, + { + "left": "issue:42491", + "right": "issue:43720", + "accept": false, + "reason": "LoRA/Qwen3_Moe HF4.x vs HF5.x incompatibility and BitNet packed-weight loading are different problems." + } + ] + }, + "evaluator_result": { + "accept": true, + "feedback": "Grounded overall: the canonical issue choice is supported by the titles, and the soft-edge decisions stay conservative. The two accepted pairs are plausibly duplicate-like from the packet, while the rest are reasonably rejected as different model/code-path failures. No mergeability issue for PR pairs since none are present." + }, + "evaluator_used": true, + "retried": false, + "accepted_nontrivial_soft_edge": true, + "error_kind": null, + "error_message": null + }, + { + "label": "issue", + "component_index": 1, + "component_count": 27, + "review_unit_index": 7, + "review_unit_count": 63, + "status": "reviewed", + "reason": null, + "source": "llm", + "cache_hit": false, + "model": "gpt-5.4-mini?service_tier=flex", + "cluster_id": "cluster-10000-514", + "nodes": [ + "issue:38175", + "issue:39692", + "issue:41720", + "issue:42491", + "issue:42915", + "issue:43257", + "issue:43317", + "issue:43632", + "issue:43644", + "issue:43645", + "issue:43708", + "issue:43854", + "issue:43856", + "issue:43873", + "issue:43931", + "issue:43994", + "issue:44075", + "issue:44164", + "issue:44291", + "issue:44387", + "issue:44451", + "issue:44488", + "issue:44534", + "issue:44661", + "issue:44756", + "issue:44805", + "issue:44863", + "issue:44960", + "issue:44977", + "issue:44991", + "issue:45276", + "issue:45292", + "issue:45313", + "issue:45335", + "issue:45357", + "issue:45362", + "issue:45406", + "issue:45488", + "issue:45507", + "issue:45812", + "issue:45834", + "issue:46050", + "issue:46459", + "issue:46468", + "issue:46512", + "issue:46846", + "issue:46897", + "issue:46901" + ], + "soft_pairs": [ + "issue:43632|issue:45834", + "issue:44661|issue:44805", + "issue:43317|issue:44756", + "issue:38175|issue:39692", + "issue:43644|issue:44534", + "issue:41720|issue:42915", + "issue:42915|issue:44756", + "issue:42491|issue:45276", + "issue:44451|issue:44488", + "issue:44960|issue:44977", + "issue:44863|issue:46846", + "issue:46459|issue:46512", + "issue:45313|issue:45357", + "issue:45507|issue:46468", + "issue:43644|issue:43645", + "issue:43257|issue:43854", + "issue:39692|issue:43994", + "issue:44488|issue:44991", + "issue:43856|issue:43873", + "issue:42491|issue:45292", + "issue:43854|issue:43931", + "issue:42491|issue:45335", + "issue:43708|issue:44075", + "issue:43856|issue:44387", + "issue:44960|issue:46050", + "issue:44164|issue:44291", + "issue:46897|issue:46901", + "issue:45276|issue:45335", + "issue:45488|issue:45812", + "issue:45362|issue:45406" + ], + "prepared_review_unit_hash": "6e4b0f4ad9cf603261136deb4743bffe64af41f4146ace0a5f9715010cab4d90", + "component_budget": { + "node_count": 514, + "item_count": 514, + "soft_pair_count": 1891, + "serialized_chars": 980895, + "estimated_input_tokens": 245224, + "estimated_eval_tokens": 490704 + }, + "budget": { + "node_count": 48, + "item_count": 48, + "soft_pair_count": 30, + "serialized_chars": 36522, + "estimated_input_tokens": 9131, + "estimated_eval_tokens": 18518 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": true, + "analyst_result": { + "summary": "This cluster is mostly heterogeneous; only one pair looks like a true duplicate. The clearest canonical issue is the generic Transformers v5 non-persistent-buffer regression.", + "confidence": 0.43, + "canonical_issue_reason": "issue:43644 is the broader, root-cause phrased report and exactly matches issue:44534\u2019s symptom; it is the better anchor than the more symptom-focused duplicate.", + "canonical_pr_reason": null, + "best_issue_reason": "issue:43644 is the best representative issue because it describes the underlying v5 regression in a general way and already aligns with the only clear duplicate pair.", + "best_pr_reason": null, + "soft_edge_verdicts": [ + { + "left": "issue:43632", + "right": "issue:45834", + "accept": false, + "reason": "Different failures in unrelated model paths; same version mention is not enough to treat them as one bug." + }, + { + "left": "issue:44661", + "right": "issue:44805", + "accept": false, + "reason": "One is a tokenizer-mapping edge case, the other is a mask/tensor shape error; no shared concrete code path is evident." + }, + { + "left": "issue:43317", + "right": "issue:44756", + "accept": false, + "reason": "GPU/CPU offload loading failure vs mmap OOM on Strix Halo are different issues and fixes." + }, + { + "left": "issue:38175", + "right": "issue:39692", + "accept": false, + "reason": "Both are about SigLIP2, but one is zero probabilities and the other is docs/quantization mistakes; not the same underlying bug." + }, + { + "left": "issue:43644", + "right": "issue:44534", + "accept": true, + "reason": "Same Transformers v5 non-persistent-buffer junk regression; the titles describe the same underlying bug." + }, + { + "left": "issue:41720", + "right": "issue:42915", + "accept": false, + "reason": "Both involve Qwen3/MoE, but one is device-mapping CUDA assert and the other is FineGrainedFP8 loading failure." + }, + { + "left": "issue:42915", + "right": "issue:44756", + "accept": false, + "reason": "Qwen3Moe FP8 failure is unrelated to mmap/OOM behavior on Strix Halo." + }, + { + "left": "issue:42491", + "right": "issue:45276", + "accept": false, + "reason": "LoRA loading on hf4.x vs resize_token_embeddings propagation in Gemma4 are unrelated bugs." + }, + { + "left": "issue:44451", + "right": "issue:44488", + "accept": false, + "reason": "Both are model loading complaints, but they name different models and do not clearly point to the same code-path defect." + }, + { + "left": "issue:44960", + "right": "issue:44977", + "accept": false, + "reason": "GLM5 and Qwen3.5 flash-attention failures are separate model-specific problems." + }, + { + "left": "issue:44863", + "right": "issue:46846", + "accept": false, + "reason": "NemotronH checkpoint loading vs Qwen3.5 tensor-parallel plan omissions are different issues." + }, + { + "left": "issue:46459", + "right": "issue:46512", + "accept": false, + "reason": "Both mention dtype, but one is composite checkpoint loading and the other is from_config bookkeeping; not the same bug." + }, + { + "left": "issue:45313", + "right": "issue:45357", + "accept": false, + "reason": "DeepSpeed ZeRO-3 weight loading and incorrect save_pretrained visual keys are different failure modes." + }, + { + "left": "issue:45507", + "right": "issue:46468", + "accept": false, + "reason": "Invalid method call in GraniteMoEHybrid is unrelated to the broken Nemotron example." + }, + { + "left": "issue:43644", + "right": "issue:43645", + "accept": false, + "reason": "Related Transformers v5 regressions, but one is junk in buffers and the other is notebook custom-model initialization; not the same concrete bug." + }, + { + "left": "issue:43257", + "right": "issue:43854", + "accept": false, + "reason": "Qwen3 MOE weight conversion with accelerate+deepspeed and GLM-4.7 Flash test loading are unrelated." + }, + { + "left": "issue:39692", + "right": "issue:43994", + "accept": false, + "reason": "Both involve SigLIP2, but one is a bad docs example while the other is a runtime wrong-output regression." + }, + { + "left": "issue:44488", + "right": "issue:44991", + "accept": false, + "reason": "Tokenizer loading failures for different models are too vague to merge as one concrete bug." + }, + { + "left": "issue:43856", + "right": "issue:43873", + "accept": false, + "reason": "Training-memory inefficiency and quantized offloading behavior are distinct problems." + }, + { + "left": "issue:42491", + "right": "issue:45292", + "accept": false, + "reason": "LoRA/model-version compatibility is unrelated to resize_token_embeddings not updating output embeddings." + }, + { + "left": "issue:43854", + "right": "issue:43931", + "accept": false, + "reason": "Different model load errors: unit-test load failure vs weight shape mismatch." + }, + { + "left": "issue:42491", + "right": "issue:45335", + "accept": false, + "reason": "Different model families and different affected embedding surfaces; not one shared bug." + }, + { + "left": "issue:43708", + "right": "issue:44075", + "accept": false, + "reason": "Trainer checkpoint-resume math and SGD args not being used are unrelated." + }, + { + "left": "issue:43856", + "right": "issue:44387", + "accept": false, + "reason": "Memory inefficiency during Qwen3 MoE training is not the same as int4 quantization reserved-memory OOM regression." + }, + { + "left": "issue:44960", + "right": "issue:46050", + "accept": false, + "reason": "GLM5 and sinq save/load breakage are unrelated model lifecycle issues." + }, + { + "left": "issue:44164", + "right": "issue:44291", + "accept": false, + "reason": "Both concern save/load internals, but extra_state handling and unexpected _is_hf_initialized arguments are separate defects." + }, + { + "left": "issue:46897", + "right": "issue:46901", + "accept": false, + "reason": "Both mention a double-shift loss bug, but they are model-specific reports with no evidence of the same shared code path." + }, + { + "left": "issue:45276", + "right": "issue:45335", + "accept": false, + "reason": "Both mention resize_token_embeddings, but they affect different model implementations and embedding targets; too weak to treat as one bug." + }, + { + "left": "issue:45488", + "right": "issue:45812", + "accept": false, + "reason": "Tokenizer-pretokenizer override for DeepSeek models is not the same as wrong token IDs for Granite models." + }, + { + "left": "issue:45362", + "right": "issue:45406", + "accept": false, + "reason": "Qwen3.5 chat crash and Gemma4Processor missing _tokenizer are different failures." + } + ] + }, + "evaluator_result": { + "accept": true, + "feedback": "Grounded overall and appropriately conservative. The single accepted pair (issue:43644 / issue:44534) is clearly duplicative from the packet, and the other rejections are generally well-justified by distinct titles/symptoms rather than superficial similarity. No evidence of overstatement in the canonical issue choice." + }, + "evaluator_used": true, + "retried": false, + "accepted_nontrivial_soft_edge": true, + "error_kind": null, + "error_message": null + }, + { + "label": "issue", + "component_index": 1, + "component_count": 27, + "review_unit_index": 8, + "review_unit_count": 63, + "status": "reviewed", + "reason": null, + "source": "llm", + "cache_hit": false, + "model": "gpt-5.4-mini?service_tier=flex", + "cluster_id": "cluster-10000-514", + "nodes": [ + "issue:36010", + "issue:39404", + "issue:39692", + "issue:41720", + "issue:42175", + "issue:42796", + "issue:42947", + "issue:43335", + "issue:43381", + "issue:43404", + "issue:43425", + "issue:43454", + "issue:43645", + "issue:43854", + "issue:43873", + "issue:43957", + "issue:44079", + "issue:44155", + "issue:44164", + "issue:44291", + "issue:44292", + "issue:44387", + "issue:44402", + "issue:44479", + "issue:44488", + "issue:44589", + "issue:44756", + "issue:44863", + "issue:44877", + "issue:44910", + "issue:44912", + "issue:44928", + "issue:44960", + "issue:45216", + "issue:45276", + "issue:45292", + "issue:45335", + "issue:45588", + "issue:45663", + "issue:45684", + "issue:45698", + "issue:45704", + "issue:45759", + "issue:45823", + "issue:45859", + "issue:46326", + "issue:46468", + "issue:46829" + ], + "soft_pairs": [ + "issue:43425|issue:43957", + "issue:44910|issue:45588", + "issue:44402|issue:44488", + "issue:39692|issue:44960", + "issue:41720|issue:44155", + "issue:44291|issue:44756", + "issue:45663|issue:45759", + "issue:43404|issue:43454", + "issue:42947|issue:45704", + "issue:43873|issue:44387", + "issue:44079|issue:46829", + "issue:45663|issue:45859", + "issue:44292|issue:44912", + "issue:42947|issue:43381", + "issue:36010|issue:39404", + "issue:43957|issue:44756", + "issue:44155|issue:45823", + "issue:44387|issue:44756", + "issue:43335|issue:45663", + "issue:44863|issue:46468", + "issue:45276|issue:45292", + "issue:44479|issue:44877", + "issue:43381|issue:44928", + "issue:44164|issue:44756", + "issue:42796|issue:46326", + "issue:45588|issue:45663", + "issue:45684|issue:45698", + "issue:43854|issue:45216", + "issue:45292|issue:45335", + "issue:43645|issue:44164", + "issue:42175|issue:43957", + "issue:44589|issue:44756" + ], + "prepared_review_unit_hash": "8225e142a5660d84f6bc4b7818e878d32b97e7988271ae4943710ff94fcb32d6", + "component_budget": { + "node_count": 514, + "item_count": 514, + "soft_pair_count": 1891, + "serialized_chars": 980895, + "estimated_input_tokens": 245224, + "estimated_eval_tokens": 490704 + }, + "budget": { + "node_count": 48, + "item_count": 48, + "soft_pair_count": 32, + "serialized_chars": 37385, + "estimated_input_tokens": 9347, + "estimated_eval_tokens": 18950 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": true, + "analyst_result": { + "summary": "This cluster is mostly a loose mix of unrelated bug reports. The only clear duplicate chain is the resize_token_embeddings regression across generic and model-specific reports; the rest are too model- or subsystem-specific to merge as the same underlying issue.", + "confidence": 0.79, + "canonical_issue_reason": "issue:45292 is the broadest, most model-agnostic description of the shared resize_token_embeddings regression, while issue:45276 and issue:45335 are specific manifestations of the same missing propagation to downstream embeddings.", + "canonical_pr_reason": null, + "best_issue_reason": "issue:45292 is the best umbrella issue because it states the core failure concisely and can absorb the more specific Gemma4 and t5gemma variants as duplicates.", + "best_pr_reason": null, + "soft_edge_verdicts": [ + { + "left": "issue:43425", + "right": "issue:43957", + "accept": false, + "reason": "Both involve model loading/runtime failures, but one is a Torch 2.10 compatibility complaint and the other is a meta-device loading regression; different code paths and likely different fixes." + }, + { + "left": "issue:44910", + "right": "issue:45588", + "accept": false, + "reason": "Both touch flash attention, but one is a Qwen3.5 packed-sequence/position_ids bug and the other is an s_aux=None crash in sink-less models; not the same underlying issue." + }, + { + "left": "issue:44402", + "right": "issue:44488", + "accept": false, + "reason": "Both mention model loading, but the titles point to different models and likely different tokenizer/vocab problems; too little evidence they share one bug." + }, + { + "left": "issue:39692", + "right": "issue:44960", + "accept": false, + "reason": "One is a SigLIP2 documentation/example error report and the other is a GLM5 issue; unrelated." + }, + { + "left": "issue:41720", + "right": "issue:44155", + "accept": false, + "reason": "Qwen3 device-mapping cudaErrorAssert and AudioFlamingo3 batched inference leakage are distinct failures in different models." + }, + { + "left": "issue:44291", + "right": "issue:44756", + "accept": false, + "reason": "Unexpected _is_hf_initialized argument handling and Strix Halo mmap OOM are unrelated initialization vs memory-management problems." + }, + { + "left": "issue:45663", + "right": "issue:45759", + "accept": false, + "reason": "Gemma4 FSDP2 shared_kv_states KeyError and Qwen 3.5/3.6 text_config unwrapping are different model/config code paths." + }, + { + "left": "issue:43404", + "right": "issue:43454", + "accept": false, + "reason": "Both are about lm_head tying, but they target different model families and different integration paths; not enough to treat as the same concrete bug." + }, + { + "left": "issue:42947", + "right": "issue:45704", + "accept": false, + "reason": "Gradient checkpointing/LoRA and apex.FusedRMSNorm memory leak are unrelated." + }, + { + "left": "issue:43873", + "right": "issue:44387", + "accept": false, + "reason": "Both mention quantization/OOM, but one is offloading behavior and the other is increased reserved memory under int4; too broad to merge." + }, + { + "left": "issue:44079", + "right": "issue:46829", + "accept": false, + "reason": "ModelOutput key assignment bug and VideoPrism output field mix-up are different manifestations with different affected classes." + }, + { + "left": "issue:45663", + "right": "issue:45859", + "accept": false, + "reason": "Different models and different failures: Gemma4 FSDP2 shared state vs Qwen3_5Moe CPU offload incompatibility." + }, + { + "left": "issue:44292", + "right": "issue:44912", + "accept": false, + "reason": "Different quantization formats and symptoms: NVFP4 runtime error vs MXFP4 load fallback to bf16." + }, + { + "left": "issue:42947", + "right": "issue:43381", + "accept": false, + "reason": "Both involve gradient checkpointing, but one is ineffective with PEFT LoRA and the other is an eval-mode restriction; not the same bug." + }, + { + "left": "issue:36010", + "right": "issue:39404", + "accept": false, + "reason": "ImportError for GenerationMixin and Whisper return_language regression are unrelated." + }, + { + "left": "issue:43957", + "right": "issue:44756", + "accept": false, + "reason": "Meta-device loading regressions and Strix Halo mmap OOM are different issues." + }, + { + "left": "issue:44155", + "right": "issue:45823", + "accept": false, + "reason": "AudioFlamingo3 batch leakage and Gemma4 device_map mismatch are different model-specific forward-path bugs." + }, + { + "left": "issue:44387", + "right": "issue:44756", + "accept": false, + "reason": "Both are memory-related, but one is int4 reserved-memory growth and the other is a platform-specific mmap OOM workaround; not the same root cause." + }, + { + "left": "issue:43335", + "right": "issue:45663", + "accept": false, + "reason": "SwitchTransformersConfig sparse-layer construction and Gemma4 FSDP2 shared_kv_states are unrelated." + }, + { + "left": "issue:44863", + "right": "issue:46468", + "accept": false, + "reason": "NemotronH checkpoint loading and the official Nemotron example being broken are related in theme but not clearly the same underlying bug." + }, + { + "left": "issue:45276", + "right": "issue:45292", + "accept": true, + "reason": "These describe the same resize_token_embeddings regression; 45276 is a model-specific instance and 45292 states the generic bug." + }, + { + "left": "issue:44479", + "right": "issue:44877", + "accept": false, + "reason": "Video-input regression for Qwen VL models and strict granite_speech config loading are different problems." + }, + { + "left": "issue:43381", + "right": "issue:44928", + "accept": false, + "reason": "Eval-mode gradient checkpointing restriction and Qwen3.5 RLHF NaN/SDPA fallback are unrelated." + }, + { + "left": "issue:44164", + "right": "issue:44756", + "accept": false, + "reason": "save/from_pretrained extra_state handling and Strix Halo mmap OOM are unrelated." + }, + { + "left": "issue:42796", + "right": "issue:46326", + "accept": false, + "reason": "Tokenizer BPE marker handling and continue_final_message behavior are both mistral_common-adjacent but not the same concrete bug." + }, + { + "left": "issue:45588", + "right": "issue:45663", + "accept": false, + "reason": "Flash-attention s_aux=None crash and Gemma4 FSDP2 shared_kv_states KeyError are distinct implementation failures." + }, + { + "left": "issue:45684", + "right": "issue:45698", + "accept": false, + "reason": "One is read-only permission propagation during save_pretrained; the other is loading the wrong custom module after save_pretrained. Related area, different bugs." + }, + { + "left": "issue:43854", + "right": "issue:45216", + "accept": false, + "reason": "GLM-4.7-Flash test loading issue and Qwen3.5 save_pretrained regression are different model-specific problems." + }, + { + "left": "issue:45292", + "right": "issue:45335", + "accept": true, + "reason": "Both report the same resize_token_embeddings bug; 45335 is just the t5gemma-specific symptom of the generic failure described in 45292." + }, + { + "left": "issue:43645", + "right": "issue:44164", + "accept": false, + "reason": "Jupyter custom-model initialization breakage and extra_state save/from_pretrained handling are different serialization/loading issues." + }, + { + "left": "issue:42175", + "right": "issue:43957", + "accept": false, + "reason": "Missing TensorFlow backend in a torch extra install and meta-device loading regressions are unrelated." + }, + { + "left": "issue:44589", + "right": "issue:44756", + "accept": false, + "reason": "Float8 storage lookup failure and Strix Halo mmap OOM are different low-level/runtime problems." + } + ] + }, + "evaluator_result": { + "accept": true, + "feedback": "The summary is grounded in the packet and stays conservative on the soft-edge pairs. The resize_token_embeddings chain (45292 with 45276/45335) is well supported by the titles, and the other refusals avoid overclaiming shared root cause from only thematic similarity." + }, + "evaluator_used": true, + "retried": false, + "accepted_nontrivial_soft_edge": true, + "error_kind": null, + "error_message": null + }, + { + "label": "issue", + "component_index": 1, + "component_count": 27, + "review_unit_index": 9, + "review_unit_count": 63, + "status": "reviewed", + "reason": null, + "source": "llm", + "cache_hit": false, + "model": "gpt-5.4-mini?service_tier=flex", + "cluster_id": "cluster-10000-514", + "nodes": [ + "issue:36032", + "issue:39404", + "issue:43425", + "issue:43452", + "issue:43596", + "issue:43611", + "issue:43645", + "issue:43646", + "issue:43708", + "issue:43716", + "issue:43828", + "issue:43856", + "issue:43901", + "issue:43906", + "issue:43994", + "issue:44079", + "issue:44164", + "issue:44265", + "issue:44402", + "issue:44451", + "issue:44479", + "issue:44509", + "issue:44530", + "issue:44589", + "issue:44671", + "issue:44683", + "issue:44756", + "issue:44805", + "issue:44898", + "issue:44912", + "issue:44991", + "issue:45081", + "issue:45084", + "issue:45479", + "issue:45701", + "issue:45759", + "issue:45812", + "issue:45823", + "issue:46050", + "issue:46315", + "issue:46326", + "issue:46491", + "issue:46665", + "issue:46821", + "issue:46869", + "issue:46894", + "issue:46899", + "issue:46901" + ], + "soft_pairs": [ + "issue:44402|issue:44451", + "issue:43646|issue:43708", + "issue:46894|issue:46901", + "issue:43425|issue:43828", + "issue:46315|issue:46491", + "issue:45812|issue:46315", + "issue:44530|issue:45084", + "issue:39404|issue:45701", + "issue:43611|issue:43645", + "issue:43425|issue:46869", + "issue:44683|issue:45823", + "issue:43425|issue:43716", + "issue:39404|issue:46665", + "issue:43425|issue:45823", + "issue:44589|issue:45084", + "issue:44479|issue:44805", + "issue:43901|issue:44509", + "issue:43994|issue:44079", + "issue:43906|issue:44530", + "issue:44164|issue:45759", + "issue:36032|issue:43452", + "issue:44265|issue:44898", + "issue:46050|issue:46326", + "issue:43906|issue:46326", + "issue:44991|issue:45081", + "issue:39404|issue:44912", + "issue:44671|issue:44991", + "issue:45823|issue:46899", + "issue:45479|issue:46894", + "issue:44164|issue:46821", + "issue:43856|issue:44756", + "issue:43596|issue:44683" + ], + "prepared_review_unit_hash": "6cfbee9f681fcec4e1bb7683f453e3fe026f8254bf75686f5125b12412ca7eb2", + "component_budget": { + "node_count": 514, + "item_count": 514, + "soft_pair_count": 1891, + "serialized_chars": 980895, + "estimated_input_tokens": 245224, + "estimated_eval_tokens": 490704 + }, + "budget": { + "node_count": 48, + "item_count": 48, + "soft_pair_count": 32, + "serialized_chars": 37541, + "estimated_input_tokens": 9386, + "estimated_eval_tokens": 19028 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": true, + "analyst_result": { + "summary": "The cluster is heterogeneous and I would reject all soft edges. The closest-looking pair, 44402/44451, still lacks evidence of the same underlying bug: 44402 reports a concrete tokenizer invariant violation (`pad`/`unk` outside `vocab_size`), while 44451 only reports that ScandiBERT cannot be loaded.", + "confidence": 0.94, + "canonical_issue_reason": "issue:44402 is the clearest standalone root-cause report because it identifies a specific tokenizer/vocab invariant violation rather than only a downstream loading symptom.", + "canonical_pr_reason": null, + "best_issue_reason": "issue:44402 is the most actionable issue in the set: it gives the concrete failure mode and is easiest to map to a fix.", + "best_pr_reason": null, + "soft_edge_verdicts": [ + { + "left": "issue:44402", + "right": "issue:44451", + "accept": false, + "reason": "Both are tokenizer/load failures, but the evidence is not enough to say they share the same underlying bug; 44402 is a specific vocab-size invariant violation, while 44451 only reports a ScandiBERT load failure." + }, + { + "left": "issue:43646", + "right": "issue:43708", + "accept": false, + "reason": "Different bugs: custom model initialization in notebooks versus Trainer resume/max_steps miscalculation." + }, + { + "left": "issue:46894", + "right": "issue:46901", + "accept": false, + "reason": "Same broad double-shift pattern, but the reports point to different models and code paths; not enough evidence they are the same bug." + }, + { + "left": "issue:43425", + "right": "issue:43828", + "accept": false, + "reason": "One is general Torch 2.10 incompatibility; the other is a specific autocast dtype mismatch on Phi-tiny-MoE. Not the same failure." + }, + { + "left": "issue:46315", + "right": "issue:46491", + "accept": false, + "reason": "Different tokenizer regressions: performance slowdown in `convert_tokens_to_ids` versus an encode/decode spacing regression." + }, + { + "left": "issue:45812", + "right": "issue:46315", + "accept": false, + "reason": "Wrong token IDs for Granite models is a functional tokenizer correctness bug, not the same as `convert_tokens_to_ids` slowness." + }, + { + "left": "issue:44530", + "right": "issue:45084", + "accept": false, + "reason": "Different code paths and symptoms: Qwen3.5 PagedAttentionCache crash versus a template compilation error." + }, + { + "left": "issue:39404", + "right": "issue:45701", + "accept": false, + "reason": "Whisper pipeline `return_language` behavior and a general tokenization-change report are unrelated." + }, + { + "left": "issue:43611", + "right": "issue:43645", + "accept": false, + "reason": "Both involve custom models, but one is about `base_model_prefix` loading and the other is notebook-defined model initialization." + }, + { + "left": "issue:43425", + "right": "issue:46869", + "accept": false, + "reason": "General Torch version incompatibility and YOLO multi-GPU device-map failure are different bugs." + }, + { + "left": "issue:44683", + "right": "issue:45823", + "accept": false, + "reason": "Compiled flex_attention on newer torch and Gemma4 device-mismatch during forward are unrelated failures." + }, + { + "left": "issue:43425", + "right": "issue:43716", + "accept": false, + "reason": "Torch 2.10 incompatibility is not the same as a Mistral-3 image-preprocessor/model dtype mismatch." + }, + { + "left": "issue:39404", + "right": "issue:46665", + "accept": false, + "reason": "Whisper pipeline language-return regression and `from_pretrained` inside Python classes are different issues." + }, + { + "left": "issue:43425", + "right": "issue:45823", + "accept": false, + "reason": "A broad Torch incompatibility report does not match the specific Gemma4 device-map bug." + }, + { + "left": "issue:44589", + "right": "issue:45084", + "accept": false, + "reason": "Missing Float8 storage type and non-template-node compilation failure are unrelated." + }, + { + "left": "issue:44479", + "right": "issue:44805", + "accept": false, + "reason": "Video-input regression in Qwen models and a mask-shape IndexError are different bugs." + }, + { + "left": "issue:43901", + "right": "issue:44509", + "accept": false, + "reason": "Both are documentation-related, but they concern different deprecated pipeline tasks and not the same change." + }, + { + "left": "issue:43994", + "right": "issue:44079", + "accept": false, + "reason": "Incorrect Siglip2 outputs and `ModelOutput` key assignment logic are different problems." + }, + { + "left": "issue:43906", + "right": "issue:44530", + "accept": false, + "reason": "An isolated reproduction of another issue is not the same as the Qwen3.5 PagedAttention crash itself." + }, + { + "left": "issue:44164", + "right": "issue:45759", + "accept": false, + "reason": "Serialization of `extra_state` and unwrapping `text_config` in multimodal configs are separate bugs." + }, + { + "left": "issue:36032", + "right": "issue:43452", + "accept": false, + "reason": "T5 tokenizer method-name conflict and gguf loading breakage are unrelated tokenizer/model-loading issues." + }, + { + "left": "issue:44265", + "right": "issue:44898", + "accept": false, + "reason": "`torch.export` failure with `torch_compilable_check` is different from the Perceiver interpolation failure." + }, + { + "left": "issue:46050", + "right": "issue:46326", + "accept": false, + "reason": "Quantized save/load breakage is unrelated to the `continue_final_message` bug on mistral_common tokenizers." + }, + { + "left": "issue:43906", + "right": "issue:46326", + "accept": false, + "reason": "A reproduction issue and a tokenizer continuation bug are not the same underlying defect." + }, + { + "left": "issue:44991", + "right": "issue:45081", + "accept": false, + "reason": "Both are tokenizer-loading regressions, but they involve different tokenizers and distinct failure mechanisms." + }, + { + "left": "issue:39404", + "right": "issue:44912", + "accept": false, + "reason": "Whisper pipeline behavior and MXFP4 quantization fallback are unrelated." + }, + { + "left": "issue:44671", + "right": "issue:44991", + "accept": false, + "reason": "CamemBERT masked-LM prediction regression and tokenizer load failure are different bugs." + }, + { + "left": "issue:45823", + "right": "issue:46899", + "accept": false, + "reason": "Gemma4 device mismatch during forward is different from 4-bit multimodal dtype-cast breakage." + }, + { + "left": "issue:45479", + "right": "issue:46894", + "accept": false, + "reason": "Sequence-classification zero-loss and ASR training-loss double shift are different failure modes." + }, + { + "left": "issue:44164", + "right": "issue:46821", + "accept": false, + "reason": "`extra_state` save/load handling and a TimesFM `window_size` attribute error are unrelated." + }, + { + "left": "issue:43856", + "right": "issue:44756", + "accept": false, + "reason": "Qwen3 MoE memory inefficiency and Strix Halo mmap OOM are different memory/performance problems." + }, + { + "left": "issue:43596", + "right": "issue:44683", + "accept": false, + "reason": "Deepspeed Zero3 BertModel indexing failure is unrelated to compiled flex_attention on torch >= 2.9." + } + ] + }, + "evaluator_result": { + "accept": false, + "feedback": "The set is mostly conservative, but the 44402/44451 soft-edge is overstated. In the packet, 44402 is a specific tokenizer invariant violation (`pad`/`unk` outside `vocab_size`), while 44451 only says ScandiBERT cannot be loaded; that is not enough evidence to call them the same underlying bug from the provided material. The rest of the pairwise rejections look appropriately cautious." + }, + "evaluator_used": true, + "retried": true, + "accepted_nontrivial_soft_edge": false, + "error_kind": null, + "error_message": null + }, + { + "label": "issue", + "component_index": 1, + "component_count": 27, + "review_unit_index": 10, + "review_unit_count": 63, + "status": "reviewed", + "reason": null, + "source": "llm", + "cache_hit": false, + "model": "gpt-5.4-mini?service_tier=flex", + "cluster_id": "cluster-10000-514", + "nodes": [ + "issue:32090", + "issue:34567", + "issue:36032", + "issue:39404", + "issue:40444", + "issue:40990", + "issue:42175", + "issue:42796", + "issue:42915", + "issue:43299", + "issue:43404", + "issue:43421", + "issue:43452", + "issue:43482", + "issue:43541", + "issue:43582", + "issue:43611", + "issue:43632", + "issue:43643", + "issue:43644", + "issue:43645", + "issue:43646", + "issue:43716", + "issue:43931", + "issue:44206", + "issue:44220", + "issue:44451", + "issue:44462", + "issue:44488", + "issue:44530", + "issue:44661", + "issue:44683", + "issue:44737", + "issue:44811", + "issue:44987", + "issue:44991", + "issue:45081", + "issue:45084", + "issue:45479", + "issue:45584", + "issue:45859", + "issue:46178", + "issue:46285", + "issue:46326", + "issue:46402", + "issue:46489", + "issue:46897", + "issue:47050" + ], + "soft_pairs": [ + "issue:43299|issue:43931", + "issue:45479|issue:46897", + "issue:44488|issue:44987", + "issue:43632|issue:47050", + "issue:44488|issue:45081", + "issue:43611|issue:43644", + "issue:36032|issue:42915", + "issue:44462|issue:46489", + "issue:42796|issue:43404", + "issue:43716|issue:45084", + "issue:46285|issue:46402", + "issue:44661|issue:44683", + "issue:32090|issue:39404", + "issue:43643|issue:44530", + "issue:44451|issue:44987", + "issue:43541|issue:45084", + "issue:42915|issue:44530", + "issue:44737|issue:45859", + "issue:40990|issue:43421", + "issue:34567|issue:40444", + "issue:43644|issue:43646", + "issue:42175|issue:46178", + "issue:40444|issue:46326", + "issue:43582|issue:43632", + "issue:36032|issue:43645", + "issue:44206|issue:44220", + "issue:43452|issue:43482", + "issue:44811|issue:45584", + "issue:44987|issue:44991" + ], + "prepared_review_unit_hash": "06d8079f1636a34ba2cbd5db9588a2f6729967b8d35a9216e7f1583db1557d7a", + "component_budget": { + "node_count": 514, + "item_count": 514, + "soft_pair_count": 1891, + "serialized_chars": 980895, + "estimated_input_tokens": 245224, + "estimated_eval_tokens": 490704 + }, + "budget": { + "node_count": 48, + "item_count": 48, + "soft_pair_count": 29, + "serialized_chars": 36516, + "estimated_input_tokens": 9129, + "estimated_eval_tokens": 18514 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": true, + "analyst_result": { + "summary": "Most pairs are separate regressions affecting different models/features. I\u2019d use the broad v5 custom-model initialization regression as the issue canonical, and only one soft edge looks like the same audio fbank extraction bug.", + "confidence": 0.69, + "canonical_issue_reason": "issue:43646 is the broadest umbrella among the cluster\u2019s v5 initialization/loading regressions and is the most plausible duplicate sink for nearby custom-model breakage reports.", + "canonical_pr_reason": null, + "best_issue_reason": "issue:43646 is the most generally phrased and centrally applicable issue in the set, making it the best global duplicate target for related model-initialization regressions.", + "best_pr_reason": null, + "soft_edge_verdicts": [ + { + "left": "issue:43299", + "right": "issue:43931", + "accept": false, + "reason": "Both are Qwen-related loading errors, but one is a Qwen3Moe FP8 config failure and the other is a Qwen3-VL shape mismatch; different concrete bugs." + }, + { + "left": "issue:45479", + "right": "issue:46897", + "accept": false, + "reason": "One is a sequence-classification zero-loss bug, the other is a training-loss double-shift issue in Florence2; the failure modes and code paths differ." + }, + { + "left": "issue:44488", + "right": "issue:44987", + "accept": false, + "reason": "Both mention model loading, but one is about tokenizer.json being ignored and the other is a specific transformers>=5.1.0 loading failure; not clearly the same bug." + }, + { + "left": "issue:43632", + "right": "issue:47050", + "accept": false, + "reason": "These are unrelated: `_is_hf_initialized` regression versus an ONNX fp16 attention dtype mismatch." + }, + { + "left": "issue:44488", + "right": "issue:45081", + "accept": false, + "reason": "Both are tokenizer-loading regressions, but the Mistral regex crash is a different specific path than ignoring tokenizer.json." + }, + { + "left": "issue:43611", + "right": "issue:43644", + "accept": false, + "reason": "Both are v5 breakages, but one is base_model_prefix loading and the other is non-persistent buffer initialization; different defects." + }, + { + "left": "issue:36032", + "right": "issue:42915", + "accept": false, + "reason": "T5 tokenizer method-name conflict and Qwen3Moe FP8 loading are unrelated." + }, + { + "left": "issue:44462", + "right": "issue:46489", + "accept": false, + "reason": "Both are tokenizer issues, but one is repository tokenizer.json selection and the other is DeepSeek-Coder output correctness after a prior fix; not the same bug." + }, + { + "left": "issue:42796", + "right": "issue:43404", + "accept": false, + "reason": "Tokenizer BPE-marker handling and Mistral3 lm_head weight tying are different problems." + }, + { + "left": "issue:43716", + "right": "issue:45084", + "accept": false, + "reason": "Image preprocessor dtype mismatch and template compilation failure are unrelated." + }, + { + "left": "issue:46285", + "right": "issue:46402", + "accept": false, + "reason": "CLIP structure inconsistency and Sam3Model text-encoder weight loading are different model-specific regressions." + }, + { + "left": "issue:44661", + "right": "issue:44683", + "accept": false, + "reason": "Model registration logic and torch>=2.9 flex_attention compilation are unrelated." + }, + { + "left": "issue:32090", + "right": "issue:39404", + "accept": false, + "reason": "Trainer GPU broadcast NoneType error and Whisper return_language regression are unrelated." + }, + { + "left": "issue:43643", + "right": "issue:44530", + "accept": false, + "reason": "Missing fields from AutoConfig with trust_remote_code and PagedAttentionCache invalid group type are different failures." + }, + { + "left": "issue:44451", + "right": "issue:44987", + "accept": false, + "reason": "Both are loading failures, but they concern different repos and likely different root causes." + }, + { + "left": "issue:43541", + "right": "issue:45084", + "accept": false, + "reason": "Mixtral grouped_mm tracing failure and non-template-node compilation error are not the same code-path bug." + }, + { + "left": "issue:42915", + "right": "issue:44530", + "accept": false, + "reason": "Both involve Qwen-family models, but one is FineGrainedFP8Config loading and the other is cache group-type handling; distinct issues." + }, + { + "left": "issue:44737", + "right": "issue:45859", + "accept": false, + "reason": "XLNet CPU tensor creation and Qwen3_5MoE CPU offload incompatibility are unrelated." + }, + { + "left": "issue:40990", + "right": "issue:43421", + "accept": false, + "reason": "Perplexity regression on gpt-oss and runtime post-processor updates for special tokens are different problems." + }, + { + "left": "issue:34567", + "right": "issue:40444", + "accept": false, + "reason": "TrainerState token counting and multi-image iterable-dataset finetuning failures are unrelated." + }, + { + "left": "issue:43644", + "right": "issue:43646", + "accept": false, + "reason": "Both are v5/custom-model regressions, but junk-filled non-persistent buffers is not the same bug as custom model initialization breaking." + }, + { + "left": "issue:42175", + "right": "issue:46178", + "accept": false, + "reason": "Backend packaging omission and BACKENDS_MAPPING validation are different dependency/backend issues." + }, + { + "left": "issue:40444", + "right": "issue:46326", + "accept": false, + "reason": "Multi-image Qwen2.5-VL finetuning and mistral_common continue_final_message behavior are unrelated." + }, + { + "left": "issue:43582", + "right": "issue:43632", + "accept": false, + "reason": "Apple Silicon allocator warmup TypeError and `_is_hf_initialized` regression are unrelated." + }, + { + "left": "issue:36032", + "right": "issue:43645", + "accept": false, + "reason": "T5 tokenizer method conflict and Jupyter custom model initialization breakage are different issues." + }, + { + "left": "issue:44206", + "right": "issue:44220", + "accept": true, + "reason": "Both point to the same audio fbank extraction path: 44206 names the unsupported center-arg crash in `_torch_extract_fbank_features`, and 44220 is the generic report on that function." + }, + { + "left": "issue:43452", + "right": "issue:43482", + "accept": false, + "reason": "gguf_file loading regressions and Qwen2.5-GGUF v5 loading failure are related only broadly; the specific failure is not clearly the same." + }, + { + "left": "issue:44811", + "right": "issue:45584", + "accept": false, + "reason": "Whisper batch_decode skip_special_tokens and empty-transcription generation after align_special_tokens are separate tokenizer/generation bugs." + }, + { + "left": "issue:44987", + "right": "issue:44991", + "accept": false, + "reason": "Both are tokenization/loading regressions on different repositories, but the titles indicate different concrete breakages." + } + ] + }, + "evaluator_result": { + "accept": true, + "feedback": "The summary is grounded in the packet, and the one accepted soft edge (44206/44220) is well supported by the titles. The rejections are conservative overall and do not overclaim identity beyond what the packet shows. The canonical issue choice is plausible given the nearby v5 custom-model initialization regressions, without asserting stronger evidence than available." + }, + "evaluator_used": true, + "retried": false, + "accepted_nontrivial_soft_edge": true, + "error_kind": null, + "error_message": null + }, + { + "label": "issue", + "component_index": 1, + "component_count": 27, + "review_unit_index": 11, + "review_unit_count": 63, + "status": "reviewed", + "reason": null, + "source": "llm", + "cache_hit": false, + "model": "gpt-5.4-mini?service_tier=flex", + "cluster_id": "cluster-10000-514", + "nodes": [ + "issue:28530", + "issue:34567", + "issue:36032", + "issue:42915", + "issue:43208", + "issue:43262", + "issue:43317", + "issue:43322", + "issue:43377", + "issue:43404", + "issue:43482", + "issue:43531", + "issue:43540", + "issue:43632", + "issue:43645", + "issue:43646", + "issue:43742", + "issue:43746", + "issue:43792", + "issue:43819", + "issue:43883", + "issue:43994", + "issue:44008", + "issue:44062", + "issue:44265", + "issue:44291", + "issue:44403", + "issue:44451", + "issue:44464", + "issue:44466", + "issue:44488", + "issue:44534", + "issue:44589", + "issue:44610", + "issue:44849", + "issue:44877", + "issue:44987", + "issue:45081", + "issue:45084", + "issue:45325", + "issue:45507", + "issue:45987", + "issue:46326", + "issue:46661", + "issue:46821", + "issue:46869", + "issue:46899", + "issue:47013" + ], + "soft_pairs": [ + "issue:45084|issue:45507", + "issue:43317|issue:43482", + "issue:43746|issue:46326", + "issue:43262|issue:43377", + "issue:43208|issue:47013", + "issue:43792|issue:46869", + "issue:44403|issue:45081", + "issue:42915|issue:45084", + "issue:43645|issue:46661", + "issue:43646|issue:44534", + "issue:43377|issue:43819", + "issue:44849|issue:45325", + "issue:43883|issue:44488", + "issue:44987|issue:45081", + "issue:28530|issue:34567", + "issue:43404|issue:44466", + "issue:42915|issue:44291", + "issue:43994|issue:44610", + "issue:36032|issue:43742", + "issue:44451|issue:45081", + "issue:44464|issue:45987", + "issue:44062|issue:44987", + "issue:44008|issue:46899", + "issue:42915|issue:44589", + "issue:44265|issue:44610", + "issue:43531|issue:44877", + "issue:43322|issue:43540", + "issue:43632|issue:46821" + ], + "prepared_review_unit_hash": "61c3cffde65db7c3f5f796b02cee37a466bb24d8dacf1e180ce15a494c1940da", + "component_budget": { + "node_count": 514, + "item_count": 514, + "soft_pair_count": 1891, + "serialized_chars": 980895, + "estimated_input_tokens": 245224, + "estimated_eval_tokens": 490704 + }, + "budget": { + "node_count": 48, + "item_count": 48, + "soft_pair_count": 28, + "serialized_chars": 36236, + "estimated_input_tokens": 9059, + "estimated_eval_tokens": 18374 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": true, + "analyst_result": { + "summary": "This cluster is a loose set of unrelated issue reports that only weakly resemble one another through broad \"Transformers v5 regression\" wording. They cover many distinct models, loaders, and runtime paths, so none of the soft pairs look like safe duplicates.", + "confidence": 0.86, + "canonical_issue_reason": "Issue #43646 is the broadest representative of the cluster\u2019s main recurring theme: Transformers v5 breaking model initialization/loading. It is more general than the model-specific reports, so it works best as the cluster\u2019s issue anchor.", + "canonical_pr_reason": null, + "best_issue_reason": "#43646 is the most central and general issue in this set, but it should be treated as a representative umbrella report rather than a true duplicate target for the others.", + "best_pr_reason": null, + "soft_edge_verdicts": [ + { + "left": "issue:45084", + "right": "issue:45507", + "accept": false, + "reason": "Different problems: template compilation error vs GraniteMoEHybrid calling an invalid method." + }, + { + "left": "issue:43317", + "right": "issue:43482", + "accept": false, + "reason": "Both are loading failures, but one is device_map/offload dequantization and the other is Qwen2.5-GGUF/Transformers v5 compatibility." + }, + { + "left": "issue:43746", + "right": "issue:46326", + "accept": false, + "reason": "PEFT local checkpoint loading for GraniteSpeech is unrelated to the mistral_common continue_final_message tokenizer bug." + }, + { + "left": "issue:43262", + "right": "issue:43377", + "accept": false, + "reason": "Audio chat-template sampling-rate mismatch is unrelated to MIMI encoder batch padding-mask behavior." + }, + { + "left": "issue:43208", + "right": "issue:47013", + "accept": false, + "reason": "Both concern xLSTM, but one reports multiple small-model training bugs and the other is a specific return_last_states unpacking crash." + }, + { + "left": "issue:43792", + "right": "issue:46869", + "accept": false, + "reason": "Whisper model runtime failure and YOLO multi-GPU device_map_auto failure are different code paths and model families." + }, + { + "left": "issue:44403", + "right": "issue:45081", + "accept": false, + "reason": "Unnecessary loading noise is not the same bug as Mistral regex patch crashing on tokenizer backend access." + }, + { + "left": "issue:42915", + "right": "issue:45084", + "accept": false, + "reason": "Qwen3MoE FP8 config failure and template-node compilation failure are unrelated." + }, + { + "left": "issue:43645", + "right": "issue:46661", + "accept": false, + "reason": "Notebook custom-model initialization breakage and remote-code config_class AttributeError are different initialization failures." + }, + { + "left": "issue:43646", + "right": "issue:44534", + "accept": false, + "reason": "Generic custom model initialization breakage is distinct from non-persistent buffer corruption/junk values." + }, + { + "left": "issue:43377", + "right": "issue:43819", + "accept": false, + "reason": "MIMI padding-mask discrepancy and DAC from_latents/forward mismatch are different model-specific numerical bugs." + }, + { + "left": "issue:44849", + "right": "issue:45325", + "accept": false, + "reason": "Qwen3.5 hidden_states bug and Qwen2.5-VL rope_index scaling bug affect different model variants and different internals." + }, + { + "left": "issue:43883", + "right": "issue:44488", + "accept": false, + "reason": "Missing all_tied_weights_keys on MolmoForCausalLM is unrelated to ScandiBERT loading failures." + }, + { + "left": "issue:44987", + "right": "issue:45081", + "accept": false, + "reason": "Physical-intelligence/fast loading regression and Mistral regex patch crash are different loaders and failures." + }, + { + "left": "issue:28530", + "right": "issue:34567", + "accept": false, + "reason": "Early stopping metric selection and num_input_tokens_seen not updating are unrelated Trainer behaviors." + }, + { + "left": "issue:43404", + "right": "issue:44466", + "accept": false, + "reason": "lm_head weight tying in Mistral3ForConditionalGeneration is unrelated to device-dependent lm_head serialization." + }, + { + "left": "issue:42915", + "right": "issue:44291", + "accept": false, + "reason": "Qwen3Moe FineGrainedFP8Config failure and init_empty_weights _is_hf_initialized argument issue are different v5 regressions." + }, + { + "left": "issue:43994", + "right": "issue:44610", + "accept": false, + "reason": "SigLIP2 nonsensical outputs with AutoModel/pipeline are not the same as OmDet-Turbo processor/model input-size mismatch." + }, + { + "left": "issue:36032", + "right": "issue:43742", + "accept": false, + "reason": "T5Tokenizer add_special_tokens collision and MobileLLM-125M key error are separate tokenizer/model loading issues." + }, + { + "left": "issue:44451", + "right": "issue:45081", + "accept": false, + "reason": "ScandiBERT loading failure is unrelated to the Mistral tokenizer regex backend crash." + }, + { + "left": "issue:44464", + "right": "issue:45987", + "accept": false, + "reason": "Chunked generation inconsistency with compiled forward and StaticCache.get_seq_length returning tensor shape are different generation/cache bugs." + }, + { + "left": "issue:44062", + "right": "issue:44987", + "accept": false, + "reason": "AddedToken special-arg duplication and physical-intelligence/fast loading regression do not share a concrete code path." + }, + { + "left": "issue:44008", + "right": "issue:46899", + "accept": false, + "reason": "Gemma 3n variable-name collision in forward is unrelated to Gemma4 unsafe dtype casting in multimodal embedders." + }, + { + "left": "issue:42915", + "right": "issue:44589", + "accept": false, + "reason": "Qwen3Moe FP8 config failure and missing Float8 storage type are different quantization/storage issues." + }, + { + "left": "issue:44265", + "right": "issue:44610", + "accept": false, + "reason": "torch.export/torch_compilable_check failure and OmDet-Turbo image-size mismatch are unrelated." + }, + { + "left": "issue:43531", + "right": "issue:44877", + "accept": false, + "reason": "Qwen3-MoE sliding_window behavior and strict config blocking granite_speech loading are different model/config issues." + }, + { + "left": "issue:43322", + "right": "issue:43540", + "accept": false, + "reason": "Llava Next loading segfault and Qwen3OmniMoe video input ValueError are different multimodal failures." + }, + { + "left": "issue:43632", + "right": "issue:46821", + "accept": false, + "reason": "_is_hf_initialized flag breakage and TimesFM window_size AttributeError are unrelated Transformers v5 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": 27, + "review_unit_index": 12, + "review_unit_count": 63, + "status": "reviewed", + "reason": null, + "source": "llm", + "cache_hit": false, + "model": "gpt-5.4-mini?service_tier=flex", + "cluster_id": "cluster-10000-514", + "nodes": [ + "issue:33666", + "issue:35707", + "issue:36032", + "issue:40444", + "issue:40990", + "issue:41720", + "issue:42175", + "issue:42796", + "issue:43089", + "issue:43296", + "issue:43381", + "issue:43421", + "issue:43425", + "issue:43482", + "issue:43525", + "issue:43531", + "issue:43611", + "issue:43632", + "issue:43716", + "issue:43746", + "issue:43792", + "issue:43883", + "issue:44079", + "issue:44220", + "issue:44291", + "issue:44403", + "issue:44479", + "issue:44534", + "issue:44568", + "issue:44589", + "issue:44683", + "issue:44756", + "issue:44811", + "issue:44991", + "issue:45081", + "issue:45216", + "issue:45325", + "issue:45357", + "issue:45538", + "issue:45542", + "issue:45663", + "issue:45812", + "issue:45923", + "issue:45987", + "issue:46315", + "issue:46491", + "issue:46710", + "issue:46899" + ], + "soft_pairs": [ + "issue:36032|issue:41720", + "issue:44479|issue:45325", + "issue:43089|issue:45987", + "issue:43425|issue:44403", + "issue:45216|issue:45357", + "issue:43296|issue:43531", + "issue:40444|issue:45325", + "issue:43611|issue:44534", + "issue:44991|issue:46315", + "issue:44220|issue:44683", + "issue:43883|issue:44991", + "issue:43716|issue:46899", + "issue:44403|issue:44991", + "issue:43381|issue:45923", + "issue:42796|issue:43746", + "issue:45812|issue:46491", + "issue:43792|issue:44589", + "issue:43632|issue:44756", + "issue:33666|issue:40444", + "issue:44291|issue:44589", + "issue:40990|issue:44811", + "issue:42796|issue:46710", + "issue:43482|issue:43525", + "issue:35707|issue:40444", + "issue:44079|issue:45663", + "issue:43531|issue:44291", + "issue:43421|issue:44568", + "issue:41720|issue:45081", + "issue:42175|issue:45542", + "issue:43482|issue:45538" + ], + "prepared_review_unit_hash": "284a006b07eebc40daf31b92a73027887225eafc3991fe7069aab113069a27fd", + "component_budget": { + "node_count": 514, + "item_count": 514, + "soft_pair_count": 1891, + "serialized_chars": 980895, + "estimated_input_tokens": 245224, + "estimated_eval_tokens": 490704 + }, + "budget": { + "node_count": 48, + "item_count": 48, + "soft_pair_count": 30, + "serialized_chars": 36827, + "estimated_input_tokens": 9207, + "estimated_eval_tokens": 18670 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": true, + "analyst_result": { + "summary": "These issues are largely unrelated; the only strong duplicate-like pair is the Qwen3.5 `save_pretrained` regression reports. Most other pairs share a broad subsystem or model family but describe different failure modes.", + "confidence": 0.88, + "canonical_issue_reason": "issue:45216 is the clearest representative: it describes a concrete Qwen3.5 `save_pretrained` regression with a version boundary, and it aligns closely with the later follow-up report in 45357.", + "canonical_pr_reason": null, + "best_issue_reason": "issue:45216 is the best issue candidate because it is specific, reproducible, and captures a user-visible regression rather than a vague compatibility or performance complaint.", + "best_pr_reason": null, + "soft_edge_verdicts": [ + { + "left": "issue:36032", + "right": "issue:41720", + "accept": false, + "reason": "Different models and failure modes: T5 tokenizer init conflict vs Qwen3 auto-device-map CUDA assert." + }, + { + "left": "issue:44479", + "right": "issue:45325", + "accept": false, + "reason": "Both involve Qwen VL, but one is a broad video-input regression and the other is a still-image RoPE/index scaling bug." + }, + { + "left": "issue:43089", + "right": "issue:45987", + "accept": false, + "reason": "Generation performance overhead vs a cache length return-type bug; not the same code-path issue." + }, + { + "left": "issue:43425", + "right": "issue:44403", + "accept": false, + "reason": "Torch-version incompatibility is unrelated to the loading-noise complaint." + }, + { + "left": "issue:45216", + "right": "issue:45357", + "accept": true, + "reason": "Both report the same Qwen3.5 `save_pretrained` checkpoint regression; 45357 reads like a later, narrower manifestation of the same bug." + }, + { + "left": "issue:43296", + "right": "issue:43531", + "accept": false, + "reason": "Different problems: model loading failure for PaddleOCR-VL vs Qwen3-MoE sliding-window behavior." + }, + { + "left": "issue:40444", + "right": "issue:45325", + "accept": false, + "reason": "IterableDataset/multi-image finetuning failure is different from a Qwen2.5-VL still-image position-id scaling bug." + }, + { + "left": "issue:43611", + "right": "issue:44534", + "accept": false, + "reason": "Base-model-prefix loading regression and non-persistent buffer junk are separate model-initialization issues." + }, + { + "left": "issue:44991", + "right": "issue:46315", + "accept": false, + "reason": "Tokenizer loading failure and slow `convert_tokens_to_ids` are different tokenizer regressions." + }, + { + "left": "issue:44220", + "right": "issue:44683", + "accept": false, + "reason": "Audio feature extraction issue vs compiled flex_attention on newer torch; unrelated code paths." + }, + { + "left": "issue:43883", + "right": "issue:44991", + "accept": false, + "reason": "Missing tied-weights attribute in Molmo is unrelated to tokenizer loading for EMBEDDIA/est-roberta." + }, + { + "left": "issue:43716", + "right": "issue:46899", + "accept": false, + "reason": "Both are multimodal dtype-related reports, but they affect different models and different processing stages." + }, + { + "left": "issue:44403", + "right": "issue:44991", + "accept": false, + "reason": "Loading noise is not the same as tokenizer loading failure." + }, + { + "left": "issue:43381", + "right": "issue:45923", + "accept": false, + "reason": "Eval-mode checkpointing restriction is distinct from a missing `supports_gradient_checkpointing` flag." + }, + { + "left": "issue:42796", + "right": "issue:43746", + "accept": false, + "reason": "Tokenizer BPE-marker handling and PEFT local checkpoint loading are unrelated." + }, + { + "left": "issue:45812", + "right": "issue:46491", + "accept": false, + "reason": "Granite token ID regression and CodeLlama leading-space round-trip regression are different tokenizer behaviors." + }, + { + "left": "issue:43792", + "right": "issue:44589", + "accept": false, + "reason": "Whisper runtime failure is not the same as the float8 storage lookup error." + }, + { + "left": "issue:43632", + "right": "issue:44756", + "accept": false, + "reason": "`_is_hf_initialized` flag regression and Strix Halo mmap OOM are unrelated." + }, + { + "left": "issue:33666", + "right": "issue:40444", + "accept": false, + "reason": "Both mention Qwen VL training, but one is generic multi-GPU training and the other is an IterableDataset multi-image finetuning failure." + }, + { + "left": "issue:44291", + "right": "issue:44589", + "accept": false, + "reason": "`init_empty_weights` argument handling and float8 storage resolution are separate v5 regressions." + }, + { + "left": "issue:40990", + "right": "issue:44811", + "accept": false, + "reason": "Perplexity regression on GPT-OSS is unrelated to Whisper batch decode special-token handling." + }, + { + "left": "issue:42796", + "right": "issue:46710", + "accept": false, + "reason": "Tokenizer BPE-marker bug and DeepSeek accuracy regression are different symptoms and likely different root causes." + }, + { + "left": "issue:43482", + "right": "issue:43525", + "accept": false, + "reason": "Qwen2.5-GGUF loading failure and missing `pad_token_id` on Llama4Config are unrelated model/config bugs." + }, + { + "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 iterable finetuning." + }, + { + "left": "issue:44079", + "right": "issue:45663", + "accept": false, + "reason": "ModelOutput key assignment and Gemma4 FSDP2 shared_kv_state rebuild are different internal issues." + }, + { + "left": "issue:43531", + "right": "issue:44291", + "accept": false, + "reason": "Qwen3-MoE sliding_window behavior and `init_empty_weights` type error do not match." + }, + { + "left": "issue:43421", + "right": "issue:44568", + "accept": false, + "reason": "Runtime post-processor updates for tokenizers are unrelated to missing BOS/EOS insertion in mdeberta-v3." + }, + { + "left": "issue:41720", + "right": "issue:45081", + "accept": false, + "reason": "CUDA assert during device-mapped Qwen3 load is unrelated to a Mistral tokenizer regex crash." + }, + { + "left": "issue:42175", + "right": "issue:45542", + "accept": false, + "reason": "Backend packaging omission with `.[torch]` is distinct from a tensorboard-without-TF backend error." + }, + { + "left": "issue:43482", + "right": "issue:45538", + "accept": false, + "reason": "Qwen2.5-GGUF loading failure and CLIPTokenizer max-length default are unrelated." + } + ] + }, + "evaluator_result": { + "accept": true, + "feedback": "The summary is grounded in the packet and stays conservative. The duplicate-like Qwen3.5 save_pretrained pair is the only clearly related one, and the rest of the soft-edge rejections are supported by the titles and issue scopes." + }, + "evaluator_used": true, + "retried": false, + "accepted_nontrivial_soft_edge": true, + "error_kind": null, + "error_message": null + }, + { + "label": "issue", + "component_index": 1, + "component_count": 27, + "review_unit_index": 13, + "review_unit_count": 63, + "status": "reviewed", + "reason": null, + "source": "llm", + "cache_hit": false, + "model": "gpt-5.4-mini?service_tier=flex", + "cluster_id": "cluster-10000-514", + "nodes": [ + "issue:40444", + "issue:40990", + "issue:41553", + "issue:42175", + "issue:42796", + "issue:42915", + "issue:43262", + "issue:43296", + "issue:43317", + "issue:43377", + "issue:43454", + "issue:43482", + "issue:43493", + "issue:43582", + "issue:43606", + "issue:43611", + "issue:43644", + "issue:43646", + "issue:43701", + "issue:43746", + "issue:43818", + "issue:43856", + "issue:43873", + "issue:43994", + "issue:44079", + "issue:44265", + "issue:44291", + "issue:44462", + "issue:44466", + "issue:44479", + "issue:44509", + "issue:44534", + "issue:44829", + "issue:44877", + "issue:45072", + "issue:45325", + "issue:45357", + "issue:45381", + "issue:45479", + "issue:45491", + "issue:45584", + "issue:45701", + "issue:45812", + "issue:45823", + "issue:45834", + "issue:46402", + "issue:46821", + "issue:46909" + ], + "soft_pairs": [ + "issue:44462|issue:45812", + "issue:43296|issue:46821", + "issue:44079|issue:44534", + "issue:42175|issue:43296", + "issue:43317|issue:43856", + "issue:44509|issue:46909", + "issue:43611|issue:43646", + "issue:43377|issue:45491", + "issue:43644|issue:44079", + "issue:43818|issue:46402", + "issue:40444|issue:45381", + "issue:44265|issue:45072", + "issue:43818|issue:45357", + "issue:43493|issue:43994", + "issue:45325|issue:45834", + "issue:45325|issue:45381", + "issue:44829|issue:45479", + "issue:43454|issue:44466", + "issue:44291|issue:44877", + "issue:42915|issue:43482", + "issue:43701|issue:45325", + "issue:43873|issue:45701", + "issue:42796|issue:45381", + "issue:41553|issue:43582", + "issue:43606|issue:44265", + "issue:40990|issue:45584", + "issue:44479|issue:45823", + "issue:43262|issue:43746" + ], + "prepared_review_unit_hash": "280237ba53cb1893388dab536f98af9ae75691fa408265f31a692778ed453c66", + "component_budget": { + "node_count": 514, + "item_count": 514, + "soft_pair_count": 1891, + "serialized_chars": 980895, + "estimated_input_tokens": 245224, + "estimated_eval_tokens": 490704 + }, + "budget": { + "node_count": 48, + "item_count": 48, + "soft_pair_count": 28, + "serialized_chars": 36290, + "estimated_input_tokens": 9073, + "estimated_eval_tokens": 18402 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": true, + "analyst_result": { + "summary": "This cluster is heterogeneous: it mixes unrelated tokenizer, multimodal, loading/offload, training, and documentation regressions. There is no clear duplicate set or single canonical bug across the items.", + "confidence": 0.94, + "canonical_issue_reason": null, + "canonical_pr_reason": null, + "best_issue_reason": "No issue is a strong global representative because the cluster spans many different models and code paths rather than one underlying bug.", + "best_pr_reason": null, + "soft_edge_verdicts": [ + { + "left": "issue:44462", + "right": "issue:45812", + "accept": false, + "reason": "Both involve AutoTokenizer/tokenization, but one is about ignoring tokenizer.json while the other is a Granite-specific token-id regression; different failure modes and fixes." + }, + { + "left": "issue:43296", + "right": "issue:46821", + "accept": false, + "reason": "Completely different models and bugs: PaddleOCR-VL load failure vs TimesFM window_size AttributeError." + }, + { + "left": "issue:44079", + "right": "issue:44534", + "accept": false, + "reason": "Different core issues: ModelOutput key assignment vs non-persistent buffer serialization in v5." + }, + { + "left": "issue:42175", + "right": "issue:43296", + "accept": false, + "reason": "Packaging/backend dependency problem vs a model-loading/runtime failure; not the same underlying bug." + }, + { + "left": "issue:43317", + "right": "issue:43856", + "accept": false, + "reason": "Offload/load failure with dequantized models vs Qwen3 MoE training memory usage; distinct code paths." + }, + { + "left": "issue:44509", + "right": "issue:46909", + "accept": false, + "reason": "Both are documentation-related, but they concern different docs content and are not the same concrete issue." + }, + { + "left": "issue:43611", + "right": "issue:43646", + "accept": false, + "reason": "Both mention Transformers v5 regressions, but one is about base_model_prefix loading and the other about custom model initialization." + }, + { + "left": "issue:43377", + "right": "issue:45491", + "accept": false, + "reason": "Both involve batching/padding behavior, but they are different models with different symptoms and likely different root causes." + }, + { + "left": "issue:43644", + "right": "issue:44079", + "accept": false, + "reason": "Non-persistent buffer serialization and ModelOutput key handling are unrelated internals." + }, + { + "left": "issue:43818", + "right": "issue:46402", + "accept": false, + "reason": "Different model families and failures: Video-LLaVA architecture issue vs Sam3Model text-encoder weight loading." + }, + { + "left": "issue:40444", + "right": "issue:45381", + "accept": false, + "reason": "Same model family, but one is finetuning with multiple images per prompt and the other is video position-id regression; not clearly the same bug." + }, + { + "left": "issue:44265", + "right": "issue:45072", + "accept": false, + "reason": "torch.export/torch_compilable_check failure is unrelated to bfloat16 dtype mismatches." + }, + { + "left": "issue:43818", + "right": "issue:45357", + "accept": false, + "reason": "Both are multimodal model issues, but video_tower temporal attention/weight sharing is unrelated to incorrect visual encoder keys on save_pretrained." + }, + { + "left": "issue:43493", + "right": "issue:43994", + "accept": false, + "reason": "Both are SigLIP2-related, but one reports HF-vs-JAX implementation discrepancy while the other reports bad AutoModel/pipeline outputs; not enough evidence they are the same defect." + }, + { + "left": "issue:45325", + "right": "issue:45834", + "accept": false, + "reason": "Different models and errors: Qwen2.5-VL rope-index scaling vs Kosmos2.5 long-OCR index error." + }, + { + "left": "issue:45325", + "right": "issue:45381", + "accept": false, + "reason": "Both are Qwen2.5-VL position-id bugs, but one is still-image temporal scaling and the other is video vision_position_ids; related area, but not clearly one duplicate." + }, + { + "left": "issue:44829", + "right": "issue:45479", + "accept": false, + "reason": "Different classification regressions across different model families; one is FA3 training degeneration, the other is zero-loss single-label classification behavior." + }, + { + "left": "issue:43454", + "right": "issue:44466", + "accept": false, + "reason": "AyaVision lm_head tying bug is unrelated to Granite lm_head serialization inconsistency." + }, + { + "left": "issue:44291", + "right": "issue:44877", + "accept": false, + "reason": "Both concern loading/config handling in v5, but one is init_empty_weights argument mismatch and the other is strict config rejection for granite_speech." + }, + { + "left": "issue:42915", + "right": "issue:43482", + "accept": false, + "reason": "Qwen3 MoE FP8 config failure and Qwen2.5-GGUF loading with transformers v5 are separate model/format compatibility issues." + }, + { + "left": "issue:43701", + "right": "issue:45325", + "accept": false, + "reason": "resume_from_checkpoint key mismatch is a training/checkpointing bug, not the same as Qwen2.5-VL rope index scaling." + }, + { + "left": "issue:43873", + "right": "issue:45701", + "accept": false, + "reason": "Quantization/offloading behavior and tokenization changes are different subsystems and different failure modes." + }, + { + "left": "issue:42796", + "right": "issue:45381", + "accept": false, + "reason": "Tokenizer BPE-marker handling and Qwen2.5-VL video position IDs are unrelated." + }, + { + "left": "issue:41553", + "right": "issue:43582", + "accept": false, + "reason": "AutoTokenizer error messaging for Voxtral and an Apple Silicon TypeError in caching_allocator_warmup are unrelated." + }, + { + "left": "issue:43606", + "right": "issue:44265", + "accept": false, + "reason": "CPU-offload device mismatch and torch.export export failure via torch_compilable_check are distinct problems." + }, + { + "left": "issue:40990", + "right": "issue:45584", + "accept": false, + "reason": "High perplexity on GPT-OSS with WikiText and Whisper empty-transcription generation failure are unrelated model behaviors." + }, + { + "left": "issue:44479", + "right": "issue:45823", + "accept": false, + "reason": "Both are multimodal regressions, but one is video-input behavior for Qwen variants and the other is a Gemma4 PLE device-mismatch issue." + }, + { + "left": "issue:43262", + "right": "issue:43746", + "accept": false, + "reason": "Audio processor sampling-rate defaulting and GraniteSpeech PEFT checkpoint loading are 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": 27, + "review_unit_index": 14, + "review_unit_count": 63, + "status": "reviewed", + "reason": null, + "source": "llm", + "cache_hit": false, + "model": "gpt-5.4-mini?service_tier=flex", + "cluster_id": "cluster-10000-514", + "nodes": [ + "issue:34634", + "issue:35707", + "issue:36331", + "issue:39692", + "issue:40444", + "issue:40990", + "issue:41720", + "issue:42617", + "issue:42796", + "issue:42915", + "issue:42947", + "issue:43257", + "issue:43317", + "issue:43322", + "issue:43377", + "issue:43454", + "issue:43550", + "issue:43572", + "issue:43606", + "issue:43720", + "issue:43749", + "issue:43819", + "issue:43828", + "issue:43873", + "issue:43881", + "issue:44186", + "issue:44265", + "issue:44387", + "issue:44488", + "issue:44534", + "issue:44568", + "issue:44610", + "issue:44877", + "issue:44898", + "issue:44977", + "issue:44991", + "issue:45072", + "issue:45701", + "issue:45704", + "issue:45759", + "issue:45823", + "issue:45920", + "issue:45923", + "issue:46050", + "issue:46209", + "issue:46305", + "issue:46315", + "issue:46821" + ], + "soft_pairs": [ + "issue:43828|issue:46821", + "issue:35707|issue:41720", + "issue:34634|issue:35707", + "issue:39692|issue:40444", + "issue:41720|issue:45823", + "issue:44186|issue:44265", + "issue:43606|issue:43819", + "issue:43454|issue:43550", + "issue:43257|issue:45923", + "issue:42796|issue:46209", + "issue:45823|issue:46821", + "issue:44568|issue:45920", + "issue:43377|issue:44610", + "issue:43454|issue:45072", + "issue:42617|issue:42915", + "issue:40990|issue:42947", + "issue:46305|issue:46315", + "issue:43572|issue:45759", + "issue:43377|issue:44898", + "issue:43454|issue:43720", + "issue:43749|issue:44534", + "issue:43317|issue:43873", + "issue:36331|issue:42617", + "issue:43322|issue:46050", + "issue:41720|issue:43322", + "issue:36331|issue:42915", + "issue:44877|issue:44991", + "issue:41720|issue:44977", + "issue:44387|issue:45704", + "issue:43881|issue:44488", + "issue:40990|issue:45701", + "issue:41720|issue:42617" + ], + "prepared_review_unit_hash": "c5abfe8aa4c104088ee1f73e831d6e77c2d634f82345db6ff76cd0b7216c9fe5", + "component_budget": { + "node_count": 514, + "item_count": 514, + "soft_pair_count": 1891, + "serialized_chars": 980895, + "estimated_input_tokens": 245224, + "estimated_eval_tokens": 490704 + }, + "budget": { + "node_count": 48, + "item_count": 48, + "soft_pair_count": 32, + "serialized_chars": 37224, + "estimated_input_tokens": 9306, + "estimated_eval_tokens": 18868 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": true, + "analyst_result": { + "summary": "This is a loose grab bag of unrelated Transformers bugs spanning tokenization, multimodal processors, device mapping/offloading, quantization, training APIs, and generation. None of the soft pairs look like the same underlying defect, so no duplicate merges are justified.", + "confidence": 0.89, + "canonical_issue_reason": "Issue 43873 is the best umbrella representative: it is open, discusses a broad offloading/quantization regression, and sits closest to the dominant runtime-loading theme in this cluster.", + "canonical_pr_reason": null, + "best_issue_reason": "Issue 43873 is the strongest standalone issue for triage because it is concrete, high-signal, and broadly representative of the cluster\u2019s most common failure mode (offloading/quantization/device-map regressions).", + "best_pr_reason": null, + "soft_edge_verdicts": [ + { + "left": "issue:43828", + "right": "issue:46821", + "accept": false, + "reason": "Autocast dtype mismatch vs TimesFM AttributeError: different model families and failure modes." + }, + { + "left": "issue:35707", + "right": "issue:41720", + "accept": false, + "reason": "Progressive generation cache bug vs Qwen3 CUDA assert; unrelated generation vs device-mapping crash." + }, + { + "left": "issue:34634", + "right": "issue:35707", + "accept": false, + "reason": "Bark voice_preset processor bug vs progressive generation inputs_embeds/past_key_values bug." + }, + { + "left": "issue:39692", + "right": "issue:40444", + "accept": false, + "reason": "SigLIP2 doc/model mismatch vs Qwen2.5-VL multi-image iterable dataset failure; different workflows." + }, + { + "left": "issue:41720", + "right": "issue:45823", + "accept": false, + "reason": "Both mention device mismatch, but these are different models and distinct code paths." + }, + { + "left": "issue:44186", + "right": "issue:44265", + "accept": false, + "reason": "Tokenizer padding/NER crash vs torch.export failure on torch_compilable_check; unrelated." + }, + { + "left": "issue:43606", + "right": "issue:43819", + "accept": false, + "reason": "CPU-offload device mismatch vs DAC STE mismatch; unrelated." + }, + { + "left": "issue:43454", + "right": "issue:43550", + "accept": false, + "reason": "Weight tying bug vs torch.compile/SDPA failure; unrelated." + }, + { + "left": "issue:43257", + "right": "issue:45923", + "accept": false, + "reason": "MoE weight conversion under accelerate/deepspeed vs missing gradient-checkpointing flag; unrelated." + }, + { + "left": "issue:42796", + "right": "issue:46209", + "accept": false, + "reason": "Tokenizer BPE-marker handling vs FP8 inference on B200; unrelated." + }, + { + "left": "issue:45823", + "right": "issue:46821", + "accept": false, + "reason": "Gemma4 device mismatch vs TimesFM window_size AttributeError; unrelated." + }, + { + "left": "issue:44568", + "right": "issue:45920", + "accept": false, + "reason": "Both are tokenizer regressions, but one is missing BOS/EOS insertion and the other is wrong token IDs across models; too broad." + }, + { + "left": "issue:43377", + "right": "issue:44610", + "accept": false, + "reason": "MIMI batching/padding-mask bug vs OmDet processor size mismatch; unrelated." + }, + { + "left": "issue:43454", + "right": "issue:45072", + "accept": false, + "reason": "lm_head weight tying bug vs bf16 dtype mismatch; unrelated." + }, + { + "left": "issue:42617", + "right": "issue:42915", + "accept": false, + "reason": "3d_parallel launch failure vs Qwen3MoE FineGrainedFP8Config failure; unrelated." + }, + { + "left": "issue:40990", + "right": "issue:42947", + "accept": false, + "reason": "High perplexity regression vs LoRA gradient-checkpointing ineffectiveness; unrelated." + }, + { + "left": "issue:46305", + "right": "issue:46315", + "accept": false, + "reason": "Streaming mask off-by-one vs slow convert_tokens_to_ids regression; unrelated." + }, + { + "left": "issue:43572", + "right": "issue:45759", + "accept": false, + "reason": "Missing pad_token_idx config vs text_config unwrapping bug; unrelated." + }, + { + "left": "issue:43377", + "right": "issue:44898", + "accept": false, + "reason": "MIMI encoder padding-mask bug vs Perceiver resolution handling; unrelated." + }, + { + "left": "issue:43454", + "right": "issue:43720", + "accept": false, + "reason": "Weight tying bug vs BitNet packed-weight unpacking bug; unrelated." + }, + { + "left": "issue:43749", + "right": "issue:44534", + "accept": false, + "reason": "FSDP CPU-RAM loading bug vs non-persistent buffer junk; unrelated." + }, + { + "left": "issue:43317", + "right": "issue:43873", + "accept": false, + "reason": "Both touch offloading/quantization, but one is a specific dequantized-model load failure and the other is a broader offloading complaint; not the same concrete bug." + }, + { + "left": "issue:36331", + "right": "issue:42617", + "accept": false, + "reason": "CustomTrainer API break vs 3d_parallel launch issue; unrelated." + }, + { + "left": "issue:43322", + "right": "issue:46050", + "accept": false, + "reason": "Llava Next load segfault vs sinq save/load regression; unrelated." + }, + { + "left": "issue:41720", + "right": "issue:43322", + "accept": false, + "reason": "Qwen3 device-map assert vs Llava Next segfault; unrelated." + }, + { + "left": "issue:36331", + "right": "issue:42915", + "accept": false, + "reason": "Trainer API break vs Qwen3MoE FP8 config failure; unrelated." + }, + { + "left": "issue:44877", + "right": "issue:44991", + "accept": false, + "reason": "Strict config loading vs tokenizer loading regression; unrelated." + }, + { + "left": "issue:41720", + "right": "issue:44977", + "accept": false, + "reason": "Same model family, but device-map crash and flash-attention generation bug are different failures." + }, + { + "left": "issue:44387", + "right": "issue:45704", + "accept": false, + "reason": "Int4 quantization memory growth vs apex RMSNorm memory leak; unrelated." + }, + { + "left": "issue:43881", + "right": "issue:44488", + "accept": false, + "reason": "GLM-4V load failure vs sleng-bert tokenizer load failure; unrelated." + }, + { + "left": "issue:40990", + "right": "issue:45701", + "accept": false, + "reason": "Perplexity regression vs tokenizer-version behavior change; related theme but not the same bug." + }, + { + "left": "issue:41720", + "right": "issue:42617", + "accept": false, + "reason": "Qwen3 auto-device-map crash vs 3d_parallel launch issue; 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": 27, + "review_unit_index": 15, + "review_unit_count": 63, + "status": "reviewed", + "reason": null, + "source": "llm", + "cache_hit": false, + "model": "gpt-5.4-mini?service_tier=flex", + "cluster_id": "cluster-10000-514", + "nodes": [ + "issue:26413", + "issue:28530", + "issue:33666", + "issue:36010", + "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:43883", + "issue:44008", + "issue:44186", + "issue:44423", + "issue:44451", + "issue:44466", + "issue:44610", + "issue:44734", + "issue:44756", + "issue:44898", + "issue:44991", + "issue:45072", + "issue:45161", + "issue:45357", + "issue:45406", + "issue:45685", + "issue:45701", + "issue:45823", + "issue:46402", + "issue:46612", + "issue:46665", + "issue:46821", + "issue:46858", + "issue:46869", + "issue:47020" + ], + "soft_pairs": [ + "issue:43450|issue:45701", + "issue:44466|issue:45357", + "issue:43065|issue:46402", + "issue:26413|issue:28530", + "issue:43873|issue:44756", + "issue:44186|issue:45072", + "issue:46612|issue:46858", + "issue:45685|issue:46821", + "issue:43295|issue:45701", + "issue:43550|issue:43720", + "issue:43577|issue:43708", + "issue:43646|issue:46665", + "issue:36010|issue:42617", + "issue:44451|issue:47020", + "issue:43550|issue:43606", + "issue:44186|issue:44610", + "issue:42915|issue:43317", + "issue:43404|issue:45357", + "issue:46821|issue:46869", + "issue:44610|issue:44898", + "issue:28530|issue:33666", + "issue:41720|issue:45161", + "issue:44734|issue:46858", + "issue:43881|issue:44991", + "issue:44008|issue:45823", + "issue:43531|issue:43572", + "issue:44423|issue:45406", + "issue:43526|issue:43883" + ], + "prepared_review_unit_hash": "361e15b453e83b57c677efadba9a8cd38a288436a97fef29bc08d11258b07893", + "component_budget": { + "node_count": 514, + "item_count": 514, + "soft_pair_count": 1891, + "serialized_chars": 980895, + "estimated_input_tokens": 245224, + "estimated_eval_tokens": 490704 + }, + "budget": { + "node_count": 48, + "item_count": 48, + "soft_pair_count": 28, + "serialized_chars": 36191, + "estimated_input_tokens": 9048, + "estimated_eval_tokens": 18352 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": true, + "analyst_result": { + "summary": "The proposed soft links are mostly broad topical similarity rather than true duplicates; none look like the same underlying bug/change. The set is dominated by unrelated Transformers regressions, and the closest pairs still differ in concrete failure mode and code path.", + "confidence": 0.67, + "canonical_issue_reason": "issue:43873 is the best representative open issue here: it is broad, active, and centered on offloading/quantization/device-map behavior, which is the closest recurring theme among the overbroad soft matches.", + "canonical_pr_reason": null, + "best_issue_reason": "issue:43873 is the strongest global issue candidate because it is the most general and reusable triage target among the set; several other items are much narrower or clearly unrelated.", + "best_pr_reason": null, + "soft_edge_verdicts": [ + { + "left": "issue:43450", + "right": "issue:45701", + "accept": false, + "reason": "Video processor batched-shape bug vs tokenizer/version-change report; different subsystem and failure mode." + }, + { + "left": "issue:44466", + "right": "issue:45357", + "accept": false, + "reason": "Both mention serialization-like behavior, but one is lm_head/tied weights and the other is visual encoder key saving; not the same bug." + }, + { + "left": "issue:43065", + "right": "issue:46402", + "accept": false, + "reason": "Both are Sam3-related, but one is a dummy Conv2d in the pixel decoder while the other is missing text encoder weight loading." + }, + { + "left": "issue:26413", + "right": "issue:28530", + "accept": false, + "reason": "Checkpoint resume/epoch iterator failure is unrelated to early stopping metric lookup." + }, + { + "left": "issue:43873", + "right": "issue:44756", + "accept": false, + "reason": "Both involve memory/offload concerns, but one is quantization offloading behavior and the other is mmap OOM avoidance on Strix Halo." + }, + { + "left": "issue:44186", + "right": "issue:45072", + "accept": false, + "reason": "Tokenizer padding/truncation crash is unrelated to bf16 dtype mismatch in model wrappers." + }, + { + "left": "issue:46612", + "right": "issue:46858", + "accept": false, + "reason": "Beam-search cache reorder omission and static-cache seq_length tensor arithmetic are different cache bugs." + }, + { + "left": "issue:45685", + "right": "issue:46821", + "accept": false, + "reason": "MPS histogram/int issue in MoE is unrelated to a TimesFM window_size attribute error." + }, + { + "left": "issue:43295", + "right": "issue:45701", + "accept": false, + "reason": "Regression around processor.tokenizer access is not the same as a tokenizer output changing across versions." + }, + { + "left": "issue:43550", + "right": "issue:43720", + "accept": false, + "reason": "One is a torch.compile/SDPA failure, the other is packed-weight unpacking during accelerate loading." + }, + { + "left": "issue:43577", + "right": "issue:43708", + "accept": false, + "reason": "Model dtype loading problem vs resume_from_checkpoint max_steps miscalculation; different code paths." + }, + { + "left": "issue:43646", + "right": "issue:46665", + "accept": false, + "reason": "Custom model init breakage and from_pretrained-inside-class inconsistency are related only at a very broad API level." + }, + { + "left": "issue:36010", + "right": "issue:42617", + "accept": false, + "reason": "ImportError for GenerationMixin is unrelated to not being able to run 3d_parallel.py." + }, + { + "left": "issue:44451", + "right": "issue:47020", + "accept": false, + "reason": "Both are tokenizer-loading issues, but one is a specific model load failure and the other is a Unigram constructor type mismatch." + }, + { + "left": "issue:43550", + "right": "issue:43606", + "accept": false, + "reason": "Both involve loading/dispatch behavior, but one is compile+SDPA and the other is CPU-offload device mismatch." + }, + { + "left": "issue:44186", + "right": "issue:44610", + "accept": false, + "reason": "Tokenizer crash on batched padding/truncation is not the same as a processor producing the wrong image size." + }, + { + "left": "issue:42915", + "right": "issue:43317", + "accept": false, + "reason": "Qwen3Moe FP8 config failure and dequantized model offload load failure are distinct quantization bugs." + }, + { + "left": "issue:43404", + "right": "issue:45357", + "accept": false, + "reason": "Untied lm_head weights in Mistral3 are unrelated to incorrect visual encoder keys in Qwen3.5 save_pretrained." + }, + { + "left": "issue:46821", + "right": "issue:46869", + "accept": false, + "reason": "TimesFM window_size attribute error and YOLO multi-GPU device-map failure are different models and code paths." + }, + { + "left": "issue:44610", + "right": "issue:44898", + "accept": false, + "reason": "Both concern image sizes, but one is an OmDet-Turbo processor/model mismatch and the other is Perceiver interpolation not working at non-default resolution." + }, + { + "left": "issue:28530", + "right": "issue:33666", + "accept": false, + "reason": "Early stopping metric handling does not match Qwen2-VL multi-GPU training." + }, + { + "left": "issue:41720", + "right": "issue:45161", + "accept": false, + "reason": "Auto device mapping cuda assert on Qwen3 and TP not working on GPT-OSS MoE are only loosely related by distributed inference." + }, + { + "left": "issue:44734", + "right": "issue:46858", + "accept": false, + "reason": "Both mention cache/prefill, but one is a responses/KV-continuation indexing crash and the other is a seq_length device-tensor arithmetic bug." + }, + { + "left": "issue:43881", + "right": "issue:44991", + "accept": false, + "reason": "glm-4v-9b load failure and tokenizer load failure for est-roberta are unrelated." + }, + { + "left": "issue:44008", + "right": "issue:45823", + "accept": false, + "reason": "Gemma3n variable-name collision and Gemma4 PLE device mismatch are different model bugs." + }, + { + "left": "issue:43531", + "right": "issue:43572", + "accept": false, + "reason": "Qwen3-MoE sliding_window behavior and missing pad_token_idx in StableLmConfig are unrelated regressions." + }, + { + "left": "issue:44423", + "right": "issue:45406", + "accept": false, + "reason": "Both are serve/processor-related crashes, but the failing attribute and model families are different." + }, + { + "left": "issue:43526", + "right": "issue:43883", + "accept": false, + "reason": "BeitImageProcessorFast reduce_labels and Molmo all_tied_weights_keys are unrelated issues in different components." + } + ] + }, + "evaluator_result": null, + "evaluator_used": false, + "retried": false, + "accepted_nontrivial_soft_edge": false, + "error_kind": null, + "error_message": null + }, + { + "label": "issue", + "component_index": 1, + "component_count": 27, + "review_unit_index": 16, + "review_unit_count": 63, + "status": "reviewed", + "reason": null, + "source": "llm", + "cache_hit": false, + "model": "gpt-5.4-mini?service_tier=flex", + "cluster_id": "cluster-10000-514", + "nodes": [ + "issue:28530", + "issue:29159", + "issue:30064", + "issue:33357", + "issue:34567", + "issue:34634", + "issue:35707", + "issue:36032", + "issue:36331", + "issue:41720", + "issue:42490", + "issue:42915", + "issue:43316", + "issue:43322", + "issue:43329", + "issue:43388", + "issue:43425", + "issue:43452", + "issue:43531", + "issue:43540", + "issue:43550", + "issue:43606", + "issue:43653", + "issue:43720", + "issue:43746", + "issue:43782", + "issue:43873", + "issue:44186", + "issue:44361", + "issue:44464", + "issue:44492", + "issue:44509", + "issue:44534", + "issue:44671", + "issue:44743", + "issue:44913", + "issue:45059", + "issue:45632", + "issue:45684", + "issue:45701", + "issue:45704", + "issue:45759", + "issue:46291", + "issue:46665", + "issue:46821", + "issue:46858", + "issue:47020", + "issue:47145" + ], + "soft_pairs": [ + "issue:42915|issue:46291", + "issue:36331|issue:41720", + "issue:44464|issue:46858", + "issue:44671|issue:44743", + "issue:45632|issue:45684", + "issue:44186|issue:44361", + "issue:43425|issue:46821", + "issue:28530|issue:30064", + "issue:43316|issue:44913", + "issue:45701|issue:47145", + "issue:42915|issue:43322", + "issue:33357|issue:35707", + "issue:29159|issue:36032", + "issue:30064|issue:34634", + "issue:43873|issue:45704", + "issue:43540|issue:43782", + "issue:43531|issue:45759", + "issue:44492|issue:44509", + "issue:43452|issue:47020", + "issue:43550|issue:43653", + "issue:44534|issue:46665", + "issue:41720|issue:43746", + "issue:43606|issue:43720", + "issue:42490|issue:43329", + "issue:34567|issue:35707", + "issue:43531|issue:46821", + "issue:43388|issue:45059" + ], + "prepared_review_unit_hash": "e36020859446628ec0d6c1c58e9e32085ec14d30a09ad6cb171eae0c896f3424", + "component_budget": { + "node_count": 514, + "item_count": 514, + "soft_pair_count": 1891, + "serialized_chars": 980895, + "estimated_input_tokens": 245224, + "estimated_eval_tokens": 490704 + }, + "budget": { + "node_count": 48, + "item_count": 48, + "soft_pair_count": 27, + "serialized_chars": 35871, + "estimated_input_tokens": 8968, + "estimated_eval_tokens": 18192 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": true, + "analyst_result": { + "summary": "This cluster is a loose collection of unrelated Transformers issues with a few superficial topical similarities (tokenizers, config loading, multimodal models, generation, offloading). None of the soft pairs look like the same underlying bug/change, so no duplicate merges are justified.", + "confidence": 0.88, + "canonical_issue_reason": "issue:43531 is the best anchor among a heterogeneous set because it is one of the more central nodes in the similarity graph and is a concrete, code-path-specific bug report; however, the cluster as a whole is not a strong duplicate set.", + "canonical_pr_reason": null, + "best_issue_reason": "issue:43531 is the least-bad representative for the small Qwen/config-related subcluster, but it should not be treated as a global duplicate umbrella for the rest of the items.", + "best_pr_reason": null, + "soft_edge_verdicts": [ + { + "left": "issue:42915", + "right": "issue:46291", + "accept": false, + "reason": "Different failures: Qwen3Moe FP8/load behavior vs pipelines import requiring kernels version/revision. Same broad release window, not the same bug." + }, + { + "left": "issue:36331", + "right": "issue:41720", + "accept": false, + "reason": "Trainer custom compute_loss signature mismatch vs Qwen3 auto device-mapping CUDA assert. Different layers and root causes." + }, + { + "left": "issue:44464", + "right": "issue:46858", + "accept": false, + "reason": "Both involve generation/cache, but one is inconsistent chunked generation under compiled forward and the other is a device-tensor seq_length arithmetic bug in static cache prefill. Not the same code path." + }, + { + "left": "issue:44671", + "right": "issue:44743", + "accept": false, + "reason": "CamemBERT masked LM prediction regression and Qwen recurrent-state reset are unrelated model-specific bugs." + }, + { + "left": "issue:45632", + "right": "issue:45684", + "accept": false, + "reason": "One is cache-path collision for trust_remote_code; the other is read-only permissions propagation in save_pretrained/register_for_auto_class. Different filesystem issues." + }, + { + "left": "issue:44186", + "right": "issue:44361", + "accept": false, + "reason": "Both are tokenizer bugs, but one crashes on NER/batched padding and the other raises an AttributeError on tasks. Distinct failures and code paths." + }, + { + "left": "issue:43425", + "right": "issue:46821", + "accept": false, + "reason": "Torch 2.10 compatibility issue vs TimesFM window_size AttributeError. Different APIs and failure modes." + }, + { + "left": "issue:28530", + "right": "issue:30064", + "accept": false, + "reason": "Early stopping metric selection vs image processor handling of void segmentation maps. No shared underlying bug." + }, + { + "left": "issue:43316", + "right": "issue:44913", + "accept": false, + "reason": "Gemma3TextConfig API discrepancy vs GPTNeoXConfig rotary_pct reload defaulting. Separate config serialization/deserialization bugs." + }, + { + "left": "issue:45701", + "right": "issue:47145", + "accept": false, + "reason": "Tokenizer behavior changes across versions vs FP8 compressed-tensors checkpoints missing weight_scale application. Different subsystems." + }, + { + "left": "issue:42915", + "right": "issue:43322", + "accept": false, + "reason": "Qwen3Moe FineGrainedFP8Config failure vs Llava Next segmentation fault. Different models and failure mechanisms." + }, + { + "left": "issue:33357", + "right": "issue:35707", + "accept": false, + "reason": "MacOS CLIP bus error vs progressive generation with inputs_embeds/past_key_values. Unrelated runtime crash vs generation semantics." + }, + { + "left": "issue:29159", + "right": "issue:36032", + "accept": false, + "reason": "Slow/fast tokenizer inconsistency vs T5Tokenizer add_special_tokens name conflict. Both tokenizer-related, but not the same defect." + }, + { + "left": "issue:30064", + "right": "issue:34634", + "accept": false, + "reason": "Void segmentation-map processing vs BarkProcessor voice_preset handling. Different processor code paths and symptoms." + }, + { + "left": "issue:43873", + "right": "issue:45704", + "accept": false, + "reason": "Quantization offloading issue vs T5 apex FusedRMSNorm memory leak. Different components and root causes." + }, + { + "left": "issue:43540", + "right": "issue:43782", + "accept": false, + "reason": "Qwen3OmniMoe video input ValueError vs Qwen3VL weight_only load error. Same family, but different failing paths." + }, + { + "left": "issue:43531", + "right": "issue:45759", + "accept": false, + "reason": "Both touch Qwen3 config handling, but one is a sliding_window bug and the other is from_config not unwrapping text_config for multimodal configs. Not the same bug." + }, + { + "left": "issue:44492", + "right": "issue:44509", + "accept": false, + "reason": "Both are docs-related, but they cover different removed/deprecated pipeline tasks. Not duplicates." + }, + { + "left": "issue:43452", + "right": "issue:47020", + "accept": false, + "reason": "gguf_file loading issue vs XLMRobertaTokenizer passing a dict to Unigram. Different loading stacks and errors." + }, + { + "left": "issue:43550", + "right": "issue:43653", + "accept": false, + "reason": "Bamba torch.compile/SDPA failure vs BigBirdTokenizer special-token registration bug. Unrelated model/runtime vs tokenizer issue." + }, + { + "left": "issue:44534", + "right": "issue:46665", + "accept": false, + "reason": "Non-persistent buffers filled with junk in v5 vs inconsistent from_pretrained behavior inside Python classes. Different mechanisms." + }, + { + "left": "issue:41720", + "right": "issue:43746", + "accept": false, + "reason": "Qwen3 auto device mapping CUDA assert vs GraniteSpeech PEFT local-checkpoint loading. Different model families and failure modes." + }, + { + "left": "issue:43606", + "right": "issue:43720", + "accept": false, + "reason": "CPU offload device mismatch vs BitNet packed weights not unpacked during accelerate loading. Related loading context, but distinct bugs." + }, + { + "left": "issue:42490", + "right": "issue:43329", + "accept": false, + "reason": "Trailing-slash local path import warning vs undefined video-branch helper names in multimodal token counting. Different concerns." + }, + { + "left": "issue:34567", + "right": "issue:35707", + "accept": false, + "reason": "TrainerState token counter not updating vs generation with inputs_embeds and past_key_values. No shared underlying bug." + }, + { + "left": "issue:43531", + "right": "issue:46821", + "accept": false, + "reason": "Both mention window-related settings, but one is Qwen3-MoE sliding_window behavior and the other is TimesFM window_size AttributeError. Different models and APIs." + }, + { + "left": "issue:43388", + "right": "issue:45059", + "accept": false, + "reason": "gather_for_metrics batch truncation bug vs SAM3 text+bounding-box odd behavior. Different evaluation and multimodal 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": 27, + "review_unit_index": 17, + "review_unit_count": 63, + "status": "reviewed", + "reason": null, + "source": "llm", + "cache_hit": false, + "model": "gpt-5.4-mini?service_tier=flex", + "cluster_id": "cluster-10000-514", + "nodes": [ + "issue:34634", + "issue:39692", + "issue:41720", + "issue:42886", + "issue:43381", + "issue:43421", + "issue:43526", + "issue:43531", + "issue:43540", + "issue:43606", + "issue:43644", + "issue:43653", + "issue:43746", + "issue:43844", + "issue:43873", + "issue:43881", + "issue:43927", + "issue:44062", + "issue:44077", + "issue:44368", + "issue:44451", + "issue:44514", + "issue:44545", + "issue:44589", + "issue:44610", + "issue:44625", + "issue:44843", + "issue:44898", + "issue:44918", + "issue:44936", + "issue:44987", + "issue:44991", + "issue:45042", + "issue:45072", + "issue:45357", + "issue:45406", + "issue:45464", + "issue:45479", + "issue:45704", + "issue:45759", + "issue:45859", + "issue:45923", + "issue:46050", + "issue:46402", + "issue:46655", + "issue:46665", + "issue:46693", + "issue:46858" + ], + "soft_pairs": [ + "issue:45406|issue:45464", + "issue:34634|issue:41720", + "issue:43606|issue:43653", + "issue:43531|issue:44589", + "issue:45759|issue:45859", + "issue:44451|issue:44991", + "issue:43540|issue:46858", + "issue:44077|issue:44625", + "issue:45357|issue:46402", + "issue:43644|issue:46665", + "issue:44987|issue:45042", + "issue:43526|issue:43881", + "issue:44062|issue:44451", + "issue:44898|issue:45072", + "issue:43381|issue:43844", + "issue:44625|issue:45479", + "issue:39692|issue:44610", + "issue:41720|issue:45704", + "issue:44368|issue:45357", + "issue:43746|issue:46050", + "issue:43381|issue:45704", + "issue:44918|issue:44936", + "issue:46655|issue:46693", + "issue:43421|issue:43927", + "issue:45759|issue:45923", + "issue:45704|issue:45923", + "issue:43746|issue:43873", + "issue:42886|issue:44843", + "issue:44514|issue:44545" + ], + "prepared_review_unit_hash": "fd654b47f520ad7c350704b7033611316dc6b87a012cfc57967f54a4e7eb8c5b", + "component_budget": { + "node_count": 514, + "item_count": 514, + "soft_pair_count": 1891, + "serialized_chars": 980895, + "estimated_input_tokens": 245224, + "estimated_eval_tokens": 490704 + }, + "budget": { + "node_count": 48, + "item_count": 48, + "soft_pair_count": 29, + "serialized_chars": 36606, + "estimated_input_tokens": 9152, + "estimated_eval_tokens": 18560 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": true, + "analyst_result": { + "summary": "Mostly false-positive similarity edges across unrelated bugs. Two pairs look like true duplicates: the HF_HUB_OFFLINE tokenizer/cache regression (42886\u219444843) and the exact Qwen2_5_VLProcessor batched padding=False crash (44514\u219444545). Everything else is only broad subsystem/model overlap, not the same underlying issue.", + "confidence": 0.79, + "canonical_issue_reason": "No single canonical issue cleanly represents the whole set because the soft edges split into at least two separate duplicate pairs plus many unrelated false positives.", + "canonical_pr_reason": null, + "best_issue_reason": "No global best issue across the full set; the items are heterogeneous, so any single representative would be misleading.", + "best_pr_reason": null, + "soft_edge_verdicts": [ + { + "left": "issue:45406", + "right": "issue:45464", + "accept": false, + "reason": "Different bugs: Gemma4Processor missing _tokenizer vs Qwen3.5 chat/streaming failure." + }, + { + "left": "issue:34634", + "right": "issue:41720", + "accept": false, + "reason": "Bark voice_preset behavior and Qwen3 auto device mapping cudaErrorAssert are unrelated." + }, + { + "left": "issue:43606", + "right": "issue:43653", + "accept": false, + "reason": "CPU offload device mismatch in Bark-small vs BigBirdTokenizer special-token decode issue." + }, + { + "left": "issue:43531", + "right": "issue:44589", + "accept": false, + "reason": "Qwen3-MoE sliding_window bug and Float8 storage lookup error are different code paths." + }, + { + "left": "issue:45759", + "right": "issue:45859", + "accept": false, + "reason": "Qwen multimodal config unwrapping vs Qwen3.5 rotary embedding CPU-offload incompatibility." + }, + { + "left": "issue:44451", + "right": "issue:44991", + "accept": false, + "reason": "Both are load failures, but for different models and different root causes." + }, + { + "left": "issue:43540", + "right": "issue:46858", + "accept": false, + "reason": "Qwen3OmniMoe video ValueError vs static-cache seq_length tensor arithmetic on non-CUDA backends." + }, + { + "left": "issue:44077", + "right": "issue:44625", + "accept": false, + "reason": "PatchTSTMixer post_init policy and Qwen3.5 num_labels propagation are unrelated." + }, + { + "left": "issue:45357", + "right": "issue:46402", + "accept": false, + "reason": "Qwen3.5 visual encoder save_pretrained keys vs Sam3Model text-encoder loading failure." + }, + { + "left": "issue:43644", + "right": "issue:46665", + "accept": false, + "reason": "Non-persistent buffer junk and from_pretrained-in-Python-class inconsistency are distinct issues." + }, + { + "left": "issue:44987", + "right": "issue:45042", + "accept": false, + "reason": "Loading physical-intelligence/fast and PIL processors requiring torchvision are different bugs." + }, + { + "left": "issue:43526", + "right": "issue:43881", + "accept": false, + "reason": "BeitImageProcessorFast reduce_labels bug vs glm-4v-9b loading failure." + }, + { + "left": "issue:44062", + "right": "issue:44451", + "accept": false, + "reason": "AddedToken special-kwarg TypeError is unrelated to ScandiBERT loading failure." + }, + { + "left": "issue:44898", + "right": "issue:45072", + "accept": false, + "reason": "Perceiver interpolation failure and SwitchTransformers/TimmWrapper dtype mismatch are not the same bug." + }, + { + "left": "issue:43381", + "right": "issue:43844", + "accept": false, + "reason": "Eval-mode gradient checkpointing restriction vs ZeRO-3 gradient blow-up are different problems." + }, + { + "left": "issue:44625", + "right": "issue:45479", + "accept": false, + "reason": "Qwen3.5 num_labels propagation and single-label zero-loss behavior are separate classification bugs." + }, + { + "left": "issue:39692", + "right": "issue:44610", + "accept": false, + "reason": "Documentation model/processor mismatch is different from OmDet-Turbo processor/model size mismatch." + }, + { + "left": "issue:41720", + "right": "issue:45704", + "accept": false, + "reason": "Qwen3 CUDA assert and T5 apex RMSNorm memory leak are unrelated." + }, + { + "left": "issue:44368", + "right": "issue:45357", + "accept": false, + "reason": "LoRA tie_word_embeddings warning is not the same as incorrect visual encoder keys on save." + }, + { + "left": "issue:43746", + "right": "issue:46050", + "accept": false, + "reason": "PEFT local-checkpoint loading and sinq save/load breakage are different checkpointing bugs." + }, + { + "left": "issue:43381", + "right": "issue:45704", + "accept": false, + "reason": "Gradient checkpointing eval-mode issue vs T5 apex memory leak are unrelated." + }, + { + "left": "issue:44918", + "right": "issue:44936", + "accept": false, + "reason": "TRL SFT embedding unpacking and trainer.evaluate-after-train failure are different trainer bugs." + }, + { + "left": "issue:46655", + "right": "issue:46693", + "accept": false, + "reason": "Video padding calculation bug vs FA2 long-context performance regression." + }, + { + "left": "issue:43421", + "right": "issue:43927", + "accept": false, + "reason": "Runtime special-token post-processor updates and config token-ID persistence are separate tokenizer issues." + }, + { + "left": "issue:45759", + "right": "issue:45923", + "accept": false, + "reason": "Config unwrapping for composite Qwen models vs missing supports_gradient_checkpointing flag." + }, + { + "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:43873", + "accept": false, + "reason": "PEFT local checkpoint loading vs quantization/offloading behavior with quantization." + }, + { + "left": "issue:42886", + "right": "issue:44843", + "accept": true, + "reason": "Same HF_HUB_OFFLINE regression: tokenizer cache load fails because AutoTokenizer unconditionally hits model_info() in _patch_mistral_regex." + }, + { + "left": "issue:44514", + "right": "issue:44545", + "accept": true, + "reason": "Verbatim duplicate of the same Qwen2_5_VLProcessor.apply_chat_template crash with padding=False." + } + ] + }, + "evaluator_result": { + "accept": true, + "feedback": "Grounded and conservative overall. The summary correctly treats the set as heterogeneous and only upgrades the two clearly duplicated pairs supported by the titles: 42886/44843 and 44514/44545. The remaining soft-edge rejections stay appropriately cautious and do not overstate evidence." + }, + "evaluator_used": true, + "retried": false, + "accepted_nontrivial_soft_edge": true, + "error_kind": null, + "error_message": null + }, + { + "label": "issue", + "component_index": 1, + "component_count": 27, + "review_unit_index": 18, + "review_unit_count": 63, + "status": "reviewed", + "reason": null, + "source": "llm", + "cache_hit": false, + "model": "gpt-5.4-mini?service_tier=flex", + "cluster_id": "cluster-10000-514", + "nodes": [ + "issue:30064", + "issue:36010", + "issue:36331", + "issue:39692", + "issue:41720", + "issue:43295", + "issue:43421", + "issue:43493", + "issue:43531", + "issue:43540", + "issue:43550", + "issue:43596", + "issue:43611", + "issue:43643", + "issue:43653", + "issue:43720", + "issue:43782", + "issue:43819", + "issue:43873", + "issue:43927", + "issue:43994", + "issue:44062", + "issue:44186", + "issue:44220", + "issue:44361", + "issue:44464", + "issue:44488", + "issue:44625", + "issue:44661", + "issue:44671", + "issue:44857", + "issue:44991", + "issue:44995", + "issue:45081", + "issue:45356", + "issue:45381", + "issue:45834", + "issue:45907", + "issue:45987", + "issue:46032", + "issue:46294", + "issue:46463", + "issue:46665", + "issue:46693", + "issue:46829", + "issue:46832", + "issue:46858", + "issue:47050" + ], + "soft_pairs": [ + "issue:30064|issue:36010", + "issue:39692|issue:43493", + "issue:43421|issue:43653", + "issue:43994|issue:44671", + "issue:44464|issue:47050", + "issue:46032|issue:46463", + "issue:43782|issue:45381", + "issue:44857|issue:47050", + "issue:45081|issue:45356", + "issue:43421|issue:44186", + "issue:45356|issue:45907", + "issue:44995|issue:46858", + "issue:39692|issue:43873", + "issue:39692|issue:47050", + "issue:43720|issue:43819", + "issue:43295|issue:45834", + "issue:43653|issue:43720", + "issue:43550|issue:43927", + "issue:43596|issue:44220", + "issue:46829|issue:46832", + "issue:43550|issue:44361", + "issue:43531|issue:43643", + "issue:44625|issue:47050", + "issue:44062|issue:44488", + "issue:43540|issue:46294", + "issue:45381|issue:46693", + "issue:45987|issue:46829", + "issue:43611|issue:46665", + "issue:30064|issue:36331", + "issue:41720|issue:44991", + "issue:44220|issue:44661" + ], + "prepared_review_unit_hash": "68fcff3788f1eede3b1a9c37eef128abf245bf7566349074708238ef6cb0c0c0", + "component_budget": { + "node_count": 514, + "item_count": 514, + "soft_pair_count": 1891, + "serialized_chars": 980895, + "estimated_input_tokens": 245224, + "estimated_eval_tokens": 490704 + }, + "budget": { + "node_count": 48, + "item_count": 48, + "soft_pair_count": 31, + "serialized_chars": 37263, + "estimated_input_tokens": 9316, + "estimated_eval_tokens": 18888 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": true, + "analyst_result": { + "summary": "These items are heterogeneous issues across unrelated code paths. The suggested soft edges mostly share broad keywords or subsystem names, but not the same underlying bug or fix.", + "confidence": 0.96, + "canonical_issue_reason": "No canonical issue: the set spans unrelated tokenizer, vision, training, quantization, cache, and loading regressions, so there is no single representative duplicate.", + "canonical_pr_reason": null, + "best_issue_reason": "No safe global best issue. Even the closest-looking pairs are only superficially related and would not be merged as the same bug.", + "best_pr_reason": null, + "soft_edge_verdicts": [ + { + "left": "issue:30064", + "right": "issue:36010", + "accept": false, + "reason": "ImageProcessor segmentation-map handling and a GenerationMixin import error are unrelated subsystems and failure modes." + }, + { + "left": "issue:39692", + "right": "issue:43493", + "accept": false, + "reason": "Both mention SigLIP2, but one is a docs/example mismatch and quantization failure while the other is an implementation discrepancy vs JAX." + }, + { + "left": "issue:43421", + "right": "issue:43653", + "accept": false, + "reason": "Special-token post-processor updates at runtime and BigBirdTokenizer mask-token registration causing empty decode are different tokenizer bugs." + }, + { + "left": "issue:43994", + "right": "issue:44671", + "accept": false, + "reason": "Different model families and symptoms: SigLIP2 nonsensical outputs vs CamemBERT masked-LM prediction regression." + }, + { + "left": "issue:44464", + "right": "issue:47050", + "accept": false, + "reason": "Compiled forward generation inconsistency and fp16 ONNX attention dtype mismatch are different execution paths and root causes." + }, + { + "left": "issue:46032", + "right": "issue:46463", + "accept": false, + "reason": "Mamba2Mixer cache misuse with seq_len > 1 is unrelated to T5 decoder input validation error messaging." + }, + { + "left": "issue:43782", + "right": "issue:45381", + "accept": false, + "reason": "Both are Qwen vision-language issues, but one is a weight_only loading failure and the other is a video vision_position_ids regression; not the same bug." + }, + { + "left": "issue:44857", + "right": "issue:47050", + "accept": false, + "reason": "AMP/CUDA crash in LwDetrImageLoss is unrelated to Canine ONNX export dtype mismatch in attention." + }, + { + "left": "issue:45081", + "right": "issue:45356", + "accept": false, + "reason": "Tokenizer backend attribute error in Mistral regex patching and Kimi tokenizer codec-handling regression are separate tokenizer issues." + }, + { + "left": "issue:43421", + "right": "issue:44186", + "accept": false, + "reason": "Runtime special-token post-processing and LayoutLMv2Tokenizer crashes on NER padding/truncation are different code paths and symptoms." + }, + { + "left": "issue:45356", + "right": "issue:45907", + "accept": false, + "reason": "A tokenizer regression is unrelated to list_repo_files not forwarding revision in CI." + }, + { + "left": "issue:44995", + "right": "issue:46858", + "accept": false, + "reason": "Both involve cache length, but one is a stale indexer cache on second forward and the other is a device-tensor return breaking static-cache arithmetic under compile." + }, + { + "left": "issue:39692", + "right": "issue:43873", + "accept": false, + "reason": "SigLIP2 docs/quantization issues are not the same as generic offloading not working with quantization." + }, + { + "left": "issue:39692", + "right": "issue:47050", + "accept": false, + "reason": "A SigLIP2 documentation/quantization report is not the same as a Canine fp16 ONNX attention dtype mismatch." + }, + { + "left": "issue:43720", + "right": "issue:43819", + "accept": false, + "reason": "BitNet packed-weight unpacking during accelerate loading and DAC from_latents missing STE are different model internals and bugs." + }, + { + "left": "issue:43295", + "right": "issue:45834", + "accept": false, + "reason": "Processor/tokenizer regression in v4.57.5 and Kosmos2.5 long OCR index error are unrelated multimodal bugs." + }, + { + "left": "issue:43653", + "right": "issue:43720", + "accept": false, + "reason": "BigBirdTokenizer special-token registration and BitNet packed-weight loading are different domains and code paths." + }, + { + "left": "issue:43550", + "right": "issue:43927", + "accept": false, + "reason": "Bamba torch.compile SDPA failure and DiaConfig custom token IDs lost after save/load are unrelated failures." + }, + { + "left": "issue:43596", + "right": "issue:44220", + "accept": false, + "reason": "Deepspeed ZeRO-3/BertModel index error and _torch_extract_fbank_features() are different subsystems." + }, + { + "left": "issue:46829", + "right": "issue:46832", + "accept": false, + "reason": "VideoPrism output field misassignment and RT-DETR forward TypeError are distinct model bugs." + }, + { + "left": "issue:43550", + "right": "issue:44361", + "accept": false, + "reason": "Bamba compile/SDPA and MLukeTokenizer AttributeError are unrelated issues." + }, + { + "left": "issue:43531", + "right": "issue:43643", + "accept": false, + "reason": "Qwen3 sliding_window handling and AutoConfig trust_remote_code missing fields are different configuration/loading problems." + }, + { + "left": "issue:44625", + "right": "issue:47050", + "accept": false, + "reason": "Qwen3.5 num_labels propagation and Canine ONNX export dtype mismatch are unrelated bugs." + }, + { + "left": "issue:44062", + "right": "issue:44488", + "accept": false, + "reason": "AddedToken special-argument conflict and failure to load a specific tokenizer are different tokenizer-loading problems." + }, + { + "left": "issue:43540", + "right": "issue:46294", + "accept": false, + "reason": "Both concern Qwen video inputs, but one is a processing ValueError in Qwen3OmniMoe and the other is chat-template feature/token mismatch in Qwen3.5-9B." + }, + { + "left": "issue:45381", + "right": "issue:46693", + "accept": false, + "reason": "Qwen2.5-VL video position IDs and FA2 long-context performance regression are unrelated issues." + }, + { + "left": "issue:45987", + "right": "issue:46829", + "accept": false, + "reason": "StaticCache sequence-length return type and VideoPrism hidden_states output are not the same bug." + }, + { + "left": "issue:43611", + "right": "issue:46665", + "accept": false, + "reason": "Base_model_prefix loading breakage and inconsistent from_pretrained usage inside Python classes are different loading semantics issues." + }, + { + "left": "issue:30064", + "right": "issue:36331", + "accept": false, + "reason": "Void segmentation-map processing and CustomTrainer.compute_loss signature mismatch are unrelated." + }, + { + "left": "issue:41720", + "right": "issue:44991", + "accept": false, + "reason": "Qwen3 auto device-mapping CUDA assert and tokenizer loading for EMBEDDIA/est-roberta are unrelated failures." + }, + { + "left": "issue:44220", + "right": "issue:44661", + "accept": false, + "reason": "Audio feature extraction and add-new-model-like failing inside TOKENIZER_MAPPING_NAMES are 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": 27, + "review_unit_index": 19, + "review_unit_count": 63, + "status": "reviewed", + "reason": null, + "source": "llm", + "cache_hit": false, + "model": "gpt-5.4-mini?service_tier=flex", + "cluster_id": "cluster-10000-514", + "nodes": [ + "issue:34634", + "issue:36010", + "issue:39692", + "issue:43322", + "issue:43335", + "issue:43421", + "issue:43540", + "issue:43550", + "issue:43577", + "issue:43606", + "issue:43618", + "issue:43643", + "issue:43644", + "issue:43653", + "issue:43697", + "issue:43749", + "issue:43756", + "issue:43818", + "issue:43819", + "issue:44008", + "issue:44186", + "issue:44442", + "issue:44448", + "issue:44479", + "issue:44488", + "issue:44561", + "issue:44671", + "issue:44811", + "issue:44821", + "issue:44843", + "issue:44877", + "issue:44898", + "issue:44991", + "issue:45020", + "issue:45072", + "issue:45081", + "issue:45084", + "issue:45183", + "issue:45479", + "issue:45584", + "issue:45685", + "issue:45834", + "issue:45987", + "issue:46291", + "issue:46858", + "issue:46932", + "issue:47020" + ], + "soft_pairs": [ + "issue:44186|issue:44898", + "issue:43653|issue:43819", + "issue:44488|issue:47020", + "issue:43550|issue:43819", + "issue:43335|issue:43756", + "issue:43618|issue:43697", + "issue:43643|issue:44877", + "issue:45987|issue:46932", + "issue:44821|issue:45084", + "issue:44008|issue:46858", + "issue:34634|issue:36010", + "issue:44843|issue:45081", + "issue:39692|issue:45834", + "issue:43540|issue:44008", + "issue:44186|issue:44442", + "issue:44448|issue:44671", + "issue:45584|issue:45834", + "issue:45183|issue:45479", + "issue:44991|issue:47020", + "issue:43322|issue:45685", + "issue:44991|issue:46291", + "issue:43644|issue:43749", + "issue:44561|issue:45020", + "issue:43606|issue:45072", + "issue:43421|issue:44811", + "issue:43550|issue:45072", + "issue:43577|issue:43818", + "issue:44479|issue:45834" + ], + "prepared_review_unit_hash": "6647d02ca62f5ce78c29802d8ee42cd6429c239b9658e3324c5a98c1a2ddaa72", + "component_budget": { + "node_count": 514, + "item_count": 514, + "soft_pair_count": 1891, + "serialized_chars": 980895, + "estimated_input_tokens": 245224, + "estimated_eval_tokens": 490704 + }, + "budget": { + "node_count": 47, + "item_count": 47, + "soft_pair_count": 28, + "serialized_chars": 35859, + "estimated_input_tokens": 8965, + "estimated_eval_tokens": 18186 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": true, + "analyst_result": { + "summary": "Mostly heterogeneous Transformers bug reports spanning tokenizer/config loading, remote_code regressions, multimodal/video failures, and dtype/cache/offload issues. Only one soft pair looks like the same underlying bug; the rest are surface-level similarities across different code paths.", + "confidence": 0.71, + "canonical_issue_reason": "Issue 45020 is the broadest umbrella report here, covering recent remote_code regressions and serving as the closest cluster-level representative, though the set is otherwise mixed.", + "canonical_pr_reason": null, + "best_issue_reason": "Issue 45020 has the widest apparent blast radius and is the most representative high-level issue among these reports.", + "best_pr_reason": null, + "soft_edge_verdicts": [ + { + "left": "issue:44186", + "right": "issue:44898", + "accept": false, + "reason": "Both are model/input handling bugs, but LayoutLMv2 tokenizer NER/padding crashes and Perceiver non-default-resolution failures are unrelated code paths." + }, + { + "left": "issue:43653", + "right": "issue:43819", + "accept": false, + "reason": "BigBirdTokenizer special-token registration and DAC from_latents/STE mismatch are different subsystems and failures." + }, + { + "left": "issue:44488", + "right": "issue:47020", + "accept": false, + "reason": "Both involve tokenizer loading, but one is a model-specific load failure and the other is an XLMRobertaTokenizer init bug with a different root cause." + }, + { + "left": "issue:43550", + "right": "issue:43819", + "accept": false, + "reason": "torch.compile/SDPA failure in Bamba is unrelated to DAC latent conversion mismatch." + }, + { + "left": "issue:43335", + "right": "issue:43756", + "accept": false, + "reason": "SwitchTransformers sparse-layer config and Smollm3 RoPE-layer dropout are separate architecture bugs." + }, + { + "left": "issue:43618", + "right": "issue:43697", + "accept": false, + "reason": "Missing CLIPOutput attentions and RTDetrV2 output drift in v5 are distinct model regressions." + }, + { + "left": "issue:43643", + "right": "issue:44877", + "accept": false, + "reason": "Both touch config/remote-code loading, but one is missing fields from trust_remote_code and the other is strict config blocking granite_speech; different failure modes." + }, + { + "left": "issue:45987", + "right": "issue:46932", + "accept": false, + "reason": "StaticCache seq-length type/shape bug and Gemma4 quantized-embedder dtype casting are unrelated." + }, + { + "left": "issue:44821", + "right": "issue:45084", + "accept": false, + "reason": "AutoImageProcessor URL loading and template-node compile errors do not share the same underlying bug." + }, + { + "left": "issue:44008", + "right": "issue:46858", + "accept": false, + "reason": "Gemma3n variable collision in forward() and cache.get_seq_length() returning a device tensor are different bugs." + }, + { + "left": "issue:34634", + "right": "issue:36010", + "accept": false, + "reason": "BarkProcessor voice_preset behavior and GenerationMixin import failure are unrelated." + }, + { + "left": "issue:44843", + "right": "issue:45081", + "accept": false, + "reason": "Both involve _patch_mistral_regex, but one is an offline model_info call and the other is a backend_tokenizer attribute crash; distinct defects." + }, + { + "left": "issue:39692", + "right": "issue:45834", + "accept": false, + "reason": "SigLIP2 doc-example errors and Kosmos2.5 long-OCR index errors are different models and code paths." + }, + { + "left": "issue:43540", + "right": "issue:44008", + "accept": false, + "reason": "Qwen3OmniMoe video-input ValueError and Gemma3n audio_mel_mask collision are unrelated multimodal bugs." + }, + { + "left": "issue:44186", + "right": "issue:44442", + "accept": false, + "reason": "LayoutLMv2 tokenizer crashes and AutoTokenizer failing to load FastSpeech2ConformerTokenizer are different tokenizer regressions." + }, + { + "left": "issue:44448", + "right": "issue:44671", + "accept": false, + "reason": "Pegasus v4/v5 output changes and CamemBERT masked-LM prediction drift are separate model regressions." + }, + { + "left": "issue:45584", + "right": "issue:45834", + "accept": false, + "reason": "Whisper empty-transcription handling and Kosmos2.5 OCR length errors are unrelated." + }, + { + "left": "issue:45183", + "right": "issue:45479", + "accept": false, + "reason": "Input validation XOR messaging and single-label zero-loss classification are different layers of behavior." + }, + { + "left": "issue:44991", + "right": "issue:47020", + "accept": true, + "reason": "These look like the same tokenizer-loading regression: the est-roberta load failure is a concrete manifestation of the XLMRobertaTokenizer Unigram-vocab bug." + }, + { + "left": "issue:43322", + "right": "issue:45685", + "accept": false, + "reason": "Llava Next load-time segfault and moe MPS histogram failure are unrelated." + }, + { + "left": "issue:44991", + "right": "issue:46291", + "accept": false, + "reason": "Tokenizer loading failure for est-roberta and kernels package version/revision import failure are different issues." + }, + { + "left": "issue:43644", + "right": "issue:43749", + "accept": false, + "reason": "Non-persistent buffer junk in v5 and broken FSDP CPU RAM efficient loading are separate loading regressions." + }, + { + "left": "issue:44561", + "right": "issue:45020", + "accept": false, + "reason": "Both mention remote_code regressions, but 44561 is specifically the is_torch_fx_available removal while 45020 is a broader umbrella report; not clearly the same concrete bug." + }, + { + "left": "issue:43606", + "right": "issue:45072", + "accept": false, + "reason": "CPU offload device mismatch for bark-small and bfloat16 dtype mismatch in SwitchTransformers/TimmWrapperModel are different failure modes." + }, + { + "left": "issue:43421", + "right": "issue:44811", + "accept": false, + "reason": "Runtime post-processor updates for special tokens and Whisper batch_decode skip_special_tokens handling are unrelated tokenizer/processor behaviors." + }, + { + "left": "issue:43550", + "right": "issue:45072", + "accept": false, + "reason": "Bamba torch.compile SDPA failure and bfloat16 dtype mismatches in other models are not the same bug." + }, + { + "left": "issue:43577", + "right": "issue:43818", + "accept": false, + "reason": "BLIP2 dtype propagation and Video-LLaVA temporal-attention/weight-sharing issues are unrelated." + }, + { + "left": "issue:44479", + "right": "issue:45834", + "accept": false, + "reason": "Qwen video-input regression and Kosmos2.5 long-OCR index error are different multimodal paths." + } + ] + }, + "evaluator_result": { + "accept": true, + "feedback": "The summary is grounded in the packet: issue 45020 is the clearest umbrella report, and the cluster is indeed mostly heterogeneous across tokenizer/config loading, remote_code, multimodal, and dtype/cache/offload bugs. The soft-edge calls are conservative overall, and the single accepted pair (44991 \u2194 47020) is plausibly the same tokenizer-loading regression." + }, + "evaluator_used": true, + "retried": false, + "accepted_nontrivial_soft_edge": true, + "error_kind": null, + "error_message": null + }, + { + "label": "issue", + "component_index": 1, + "component_count": 27, + "review_unit_index": 20, + "review_unit_count": 63, + "status": "reviewed", + "reason": null, + "source": "llm", + "cache_hit": false, + "model": "gpt-5.4-mini?service_tier=flex", + "cluster_id": "cluster-10000-514", + "nodes": [ + "issue:28218", + "issue:29159", + "issue:34634", + "issue:36331", + "issue:42175", + "issue:43262", + "issue:43388", + "issue:43540", + "issue:43577", + "issue:43582", + "issue:43645", + "issue:43653", + "issue:43716", + "issue:43720", + "issue:43749", + "issue:43782", + "issue:43819", + "issue:43825", + "issue:43827", + "issue:43931", + "issue:43994", + "issue:44008", + "issue:44291", + "issue:44610", + "issue:44671", + "issue:44805", + "issue:44857", + "issue:44912", + "issue:44936", + "issue:44977", + "issue:45003", + "issue:45072", + "issue:45081", + "issue:45325", + "issue:45381", + "issue:45701", + "issue:46032", + "issue:46175", + "issue:46285", + "issue:46294", + "issue:46463", + "issue:46512", + "issue:46665", + "issue:46697", + "issue:46728", + "issue:46846", + "issue:46932" + ], + "soft_pairs": [ + "issue:43582|issue:46512", + "issue:43782|issue:44977", + "issue:44805|issue:44936", + "issue:43825|issue:43994", + "issue:43388|issue:46032", + "issue:46294|issue:46463", + "issue:43716|issue:44008", + "issue:43931|issue:46846", + "issue:43540|issue:45381", + "issue:28218|issue:29159", + "issue:44610|issue:45072", + "issue:43749|issue:44291", + "issue:43720|issue:45072", + "issue:43645|issue:46665", + "issue:43825|issue:43827", + "issue:34634|issue:36331", + "issue:42175|issue:45081", + "issue:44671|issue:45701", + "issue:43577|issue:43582", + "issue:45003|issue:46285", + "issue:46175|issue:46932", + "issue:43782|issue:45325", + "issue:46697|issue:46728", + "issue:43262|issue:43819", + "issue:43653|issue:44912", + "issue:43582|issue:44857" + ], + "prepared_review_unit_hash": "14891676e0305ff7e3528f63512b3e1d480cf3b1ffc1588329350ab5fc10b015", + "component_budget": { + "node_count": 514, + "item_count": 514, + "soft_pair_count": 1891, + "serialized_chars": 980895, + "estimated_input_tokens": 245224, + "estimated_eval_tokens": 490704 + }, + "budget": { + "node_count": 47, + "item_count": 47, + "soft_pair_count": 26, + "serialized_chars": 35469, + "estimated_input_tokens": 8868, + "estimated_eval_tokens": 17992 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": true, + "analyst_result": { + "summary": "This set is mostly heterogeneous: tokenizer, multimodal, training, loading, and backend issues are all mixed together. I don\u2019t see any soft pair that clearly describes the same underlying bug/change, so the cluster looks like false-positive similarity rather than true duplicates. #28218 is only the closest local representative for the tokenizer-related subset.", + "confidence": 0.72, + "canonical_issue_reason": "#28218 is the clearest, most concrete symptom report in the set and the oldest issue with some inbound references; if a single issue must represent the only loosely related tokenizer subset, this is the best fit.", + "canonical_pr_reason": null, + "best_issue_reason": "No issue is a strong cluster-wide canonical because the items span unrelated subsystems; #28218 is merely the best representative for the narrow tokenizer-space symptom, not the whole cluster.", + "best_pr_reason": null, + "soft_edge_verdicts": [ + { + "left": "issue:43582", + "right": "issue:46512", + "accept": false, + "reason": "Different dtype/state bugs: Apple Silicon warmup TypeError vs from_config leaving config.dtype stale." + }, + { + "left": "issue:43782", + "right": "issue:44977", + "accept": false, + "reason": "Unrelated failures: Qwen3VL from_pretrained weight_only error vs Qwen3.5 flash-attention generation issues." + }, + { + "left": "issue:44805", + "right": "issue:44936", + "accept": false, + "reason": "Mask/index shape mismatch bug is unrelated to trainer.evaluate() failing after trainer.train()." + }, + { + "left": "issue:43825", + "right": "issue:43994", + "accept": false, + "reason": "One is a pipeline v5 error-message issue; the other is a model output correctness issue for SigLIP2." + }, + { + "left": "issue:43388", + "right": "issue:46032", + "accept": false, + "reason": "Different code paths: gather_for_metrics dropping tuple labels vs Mamba2 use_cache correctness." + }, + { + "left": "issue:46294", + "right": "issue:46463", + "accept": false, + "reason": "Both mention input validation, but Qwen3.5 video/chat-template mismatch is not the same bug as T5 decoder-input error messaging." + }, + { + "left": "issue:43716", + "right": "issue:44008", + "accept": false, + "reason": "Both are multimodal dtype/forward-path bugs, but they affect different models and different concrete failures." + }, + { + "left": "issue:43931", + "right": "issue:46846", + "accept": false, + "reason": "Qwen3-VL weight-shape mismatch and Qwen3.5 tensor-parallel plan omissions are different loading/parallelism problems." + }, + { + "left": "issue:43540", + "right": "issue:45381", + "accept": false, + "reason": "Video-input processing failure in Qwen3OmniMoe is not the same as Qwen2.5-VL rope/position-id scaling." + }, + { + "left": "issue:28218", + "right": "issue:29159", + "accept": false, + "reason": "Tokenizer-space behavior and slow/fast tokenizer inconsistency may be related, but the evidence is too broad to conclude the same underlying bug." + }, + { + "left": "issue:44610", + "right": "issue:45072", + "accept": false, + "reason": "Different models and failures: processor size mismatch vs bfloat16 dtype mismatch in inference wrappers." + }, + { + "left": "issue:43749", + "right": "issue:44291", + "accept": false, + "reason": "FSDP CPU RAM loading breakage is a different path from init_empty_weights rejecting an unexpected argument." + }, + { + "left": "issue:43720", + "right": "issue:45072", + "accept": false, + "reason": "BitNet packed-weight loading and dtype mismatch inference issues are unrelated." + }, + { + "left": "issue:43645", + "right": "issue:46665", + "accept": false, + "reason": "Both involve custom model initialization, but the notebook regression and from_pretrained-in-classes inconsistency are not the same concrete bug." + }, + { + "left": "issue:43825", + "right": "issue:43827", + "accept": false, + "reason": "Related to the same v5 pipeline removals, but one is an error message problem and the other is stale docs; not the same artifact change." + }, + { + "left": "issue:34634", + "right": "issue:36331", + "accept": false, + "reason": "BarkProcessor voice_preset and CustomTrainer.compute_loss are unrelated bugs in different subsystems." + }, + { + "left": "issue:42175", + "right": "issue:45081", + "accept": false, + "reason": "Backend dependency packaging and Mistral regex patch crashes are different problems." + }, + { + "left": "issue:44671", + "right": "issue:45701", + "accept": false, + "reason": "CamemBERT masked-LM prediction regression and generic tokenization version changes are not the same issue." + }, + { + "left": "issue:43577", + "right": "issue:43582", + "accept": false, + "reason": "BLIP2 dtype mismatch and Apple Silicon allocator warmup TypeError are unrelated." + }, + { + "left": "issue:45003", + "right": "issue:46285", + "accept": false, + "reason": "Unsafe sys.modules access is not the same bug as CLIP text-model structural inconsistency." + }, + { + "left": "issue:46175", + "right": "issue:46932", + "accept": false, + "reason": "Gemma4 one-hot memory blowup and Gemma4 quantized embedders dtype casting are different issues." + }, + { + "left": "issue:43782", + "right": "issue:45325", + "accept": false, + "reason": "Qwen3VL weight_only loading error and Qwen2.5-VL still-image rope scaling are different model/version bugs." + }, + { + "left": "issue:46697", + "right": "issue:46728", + "accept": false, + "reason": "Swin2SR over-padding and Aria patch-count miscalculation are both image-processing bugs, but not the same concrete failure." + }, + { + "left": "issue:43262", + "right": "issue:43819", + "accept": false, + "reason": "Audio apply_chat_template sampling-rate defaulting and DAC from_latents/STE mismatch are unrelated." + }, + { + "left": "issue:43653", + "right": "issue:44912", + "accept": false, + "reason": "BigBirdTokenizer special-token registration and MXFP4 quantization fallback are different bugs." + }, + { + "left": "issue:43582", + "right": "issue:44857", + "accept": false, + "reason": "Apple Silicon caching warmup TypeError is unrelated to LwDetrImageLoss float16 AMP/CUDA crashes." + } + ] + }, + "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": 27, + "review_unit_index": 21, + "review_unit_count": 63, + "status": "reviewed", + "reason": null, + "source": "llm", + "cache_hit": false, + "model": "gpt-5.4-mini?service_tier=flex", + "cluster_id": "cluster-10000-514", + "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:43643", + "issue:43645", + "issue:43696", + "issue:43697", + "issue:43742", + "issue:43746", + "issue:43819", + "issue:43854", + "issue:43906", + "issue:43927", + "issue:44060", + "issue:44368", + "issue:44554", + "issue:44589", + "issue:44610", + "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:45759", + "issue:45859", + "issue:46209", + "issue:46291", + "issue:46326", + "issue:47050" + ], + "soft_pairs": [ + "issue:44589|issue:45042", + "issue:43742|issue:45020", + "issue:45372|issue:46326", + "issue:43550|issue:44912", + "issue:36010|issue:43645", + "issue:36010|issue:36331", + "issue:43257|issue:43329", + "issue:43906|issue:46209", + "issue:43531|issue:45042", + "issue:43526|issue:43697", + "issue:43638|issue:44936", + "issue:43452|issue:44991", + "issue:42175|issue:46291", + "issue:45081|issue:45406", + "issue:43819|issue:43927", + "issue:44060|issue:44368", + "issue:43329|issue:45600", + "issue:45084|issue:45538", + "issue:43526|issue:44554", + "issue:42490|issue:43257", + "issue:43746|issue:44811", + "issue:43643|issue:45759", + "issue:43854|issue:44977", + "issue:45859|issue:47050", + "issue:38175|issue:44610", + "issue:39692|issue:44912", + "issue:43262|issue:44811", + "issue:43257|issue:43696" + ], + "prepared_review_unit_hash": "fd8eb8f2161879218a0ed25ca9e42fa5c0e1d183c3cd360a5050b8918261d47e", + "component_budget": { + "node_count": 514, + "item_count": 514, + "soft_pair_count": 1891, + "serialized_chars": 980895, + "estimated_input_tokens": 245224, + "estimated_eval_tokens": 490704 + }, + "budget": { + "node_count": 48, + "item_count": 48, + "soft_pair_count": 28, + "serialized_chars": 36518, + "estimated_input_tokens": 9130, + "estimated_eval_tokens": 18516 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": true, + "analyst_result": { + "summary": "The cluster is mostly a grab bag of unrelated Transformers bugs across loading, tokenizers, processors, training, quantization, and model-specific regressions. I did not find any soft pair that clearly shares the same underlying bug or fix path.", + "confidence": 0.91, + "canonical_issue_reason": "issue:45020 is the broadest regression-style report in the set and the closest thing to an umbrella issue, though the cluster is not a true duplicate group.", + "canonical_pr_reason": null, + "best_issue_reason": "issue:45020 is the most general issue here for broad load-time regressions, but it is still too vague to serve as a strong canonical for the whole cluster.", + "best_pr_reason": null, + "soft_edge_verdicts": [ + { + "left": "issue:44589", + "right": "issue:45042", + "accept": false, + "reason": "Different bugs: FP8 storage lookup vs PIL/torchvision backend requirement." + }, + { + "left": "issue:43742", + "right": "issue:45020", + "accept": false, + "reason": "Both are loading-related, but one is a specific MobileLLM key error and the other is a broad remote_code regression." + }, + { + "left": "issue:45372", + "right": "issue:46326", + "accept": false, + "reason": "Both mention mistral_common, but one is an import failure and the other is a tokenizer continue_final_message behavior bug." + }, + { + "left": "issue:43550", + "right": "issue:44912", + "accept": false, + "reason": "Unrelated: SDPA torch.compile failure for Bamba vs MXFP4 quantization fallback for git-oss." + }, + { + "left": "issue:36010", + "right": "issue:43645", + "accept": false, + "reason": "Different failure modes: GenerationMixin import path vs custom model init in Jupyter notebooks." + }, + { + "left": "issue:36010", + "right": "issue:36331", + "accept": false, + "reason": "ImportError for GenerationMixin is unrelated to the CustomTrainer compute_loss signature change." + }, + { + "left": "issue:43257", + "right": "issue:43329", + "accept": false, + "reason": "Qwen3 MoE weight conversion/loading issue vs undefined video-token helper variables in multimodal token counting." + }, + { + "left": "issue:43906", + "right": "issue:46209", + "accept": false, + "reason": "Isolated reproduction of an older issue is not the same bug as FP8 inference failure on B200." + }, + { + "left": "issue:43531", + "right": "issue:45042", + "accept": false, + "reason": "Different subsystems: Qwen3-MoE sliding window vs PIL image processor torchvision dependency." + }, + { + "left": "issue:43526", + "right": "issue:43697", + "accept": false, + "reason": "BEiT label reduction bug is unrelated to RTDetrV2 output differences." + }, + { + "left": "issue:43638", + "right": "issue:44936", + "accept": false, + "reason": "Different training problems: ZeRO-3/non-pretrained Bert IndexError vs trainer.evaluate() after train()." + }, + { + "left": "issue:43452", + "right": "issue:44991", + "accept": false, + "reason": "Both involve model/tokenizer loading, but gguf_file handling and est-roberta tokenizer loading are distinct issues." + }, + { + "left": "issue:42175", + "right": "issue:46291", + "accept": false, + "reason": "Packaging/backends issue vs kernels package version/revision import failure." + }, + { + "left": "issue:45081", + "right": "issue:45406", + "accept": false, + "reason": "Mistral regex patch crash and Gemma4Processor missing _tokenizer are different code paths and symptoms." + }, + { + "left": "issue:43819", + "right": "issue:43927", + "accept": false, + "reason": "DAC latent reconstruction mismatch is unrelated to DiaConfig losing custom token IDs." + }, + { + "left": "issue:44060", + "right": "issue:44368", + "accept": false, + "reason": "Both mention tied-weights warnings in Qwen3, but one is a wrong weight-tying warning and the other is a repeated fine-tuning warning; not the same concrete bug." + }, + { + "left": "issue:43329", + "right": "issue:45600", + "accept": false, + "reason": "Multimodal video token counting bug vs auto_mappings.py CI breakage from removed configs." + }, + { + "left": "issue:45084", + "right": "issue:45538", + "accept": false, + "reason": "Template compilation error is unrelated to CLIPTokenizer model_max_length behavior." + }, + { + "left": "issue:43526", + "right": "issue:44554", + "accept": false, + "reason": "BEiT fast processor label reduction and MPS attention correctness are different bugs." + }, + { + "left": "issue:42490", + "right": "issue:43257", + "accept": false, + "reason": "Local-path import warning with trailing slash is unrelated to Qwen3 MoE weight conversion under accelerate+deepspeed." + }, + { + "left": "issue:43746", + "right": "issue:44811", + "accept": false, + "reason": "PEFT checkpoint loading for GraniteSpeech is unrelated to Whisper batch_decode skipping special tokens." + }, + { + "left": "issue:43643", + "right": "issue:45759", + "accept": false, + "reason": "Both involve config handling, but trust_remote_code missing fields and from_config text_config unwrapping are different failures." + }, + { + "left": "issue:43854", + "right": "issue:44977", + "accept": false, + "reason": "Model loading in unit tests for GLM-4.7-Flash is unrelated to Qwen3.5 flash-attention generation issues." + }, + { + "left": "issue:45859", + "right": "issue:47050", + "accept": false, + "reason": "CPU offload incompatibility in Qwen3_5MoeTextRotaryEmbedding is unrelated to Canine fp16 ONNX dtype mismatch." + }, + { + "left": "issue:38175", + "right": "issue:44610", + "accept": false, + "reason": "SigLIP2 zero-probabilities and OmDet-Turbo processor/model size mismatch are separate vision bugs." + }, + { + "left": "issue:39692", + "right": "issue:44912", + "accept": false, + "reason": "SigLIP2 docs/model-processor mismatch and git-oss MXFP4 quantization fallback are different issues." + }, + { + "left": "issue:43262", + "right": "issue:44811", + "accept": false, + "reason": "Audio chat-template sample-rate defaulting and Whisper batch_decode skip_special_tokens are unrelated processor behaviors." + }, + { + "left": "issue:43257", + "right": "issue:43696", + "accept": false, + "reason": "Qwen3 MoE loading/conversion bug is unrelated to GPT-oss CUDA OOM." + } + ] + }, + "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": 27, + "review_unit_index": 22, + "review_unit_count": 63, + "status": "reviewed", + "reason": null, + "source": "llm", + "cache_hit": false, + "model": "gpt-5.4-mini?service_tier=flex", + "cluster_id": "cluster-10000-514", + "nodes": [ + "issue:28282", + "issue:29159", + "issue:29870", + "issue:33357", + "issue:34567", + "issue:34689", + "issue:39379", + "issue:40990", + "issue:42175", + "issue:42915", + "issue:43257", + "issue:43329", + "issue:43425", + "issue:43450", + "issue:43454", + "issue:43525", + "issue:43540", + "issue:43577", + "issue:43650", + "issue:43697", + "issue:43742", + "issue:43749", + "issue:43824", + "issue:43906", + "issue:43931", + "issue:44291", + "issue:44387", + "issue:44479", + "issue:44554", + "issue:44568", + "issue:44617", + "issue:44756", + "issue:44821", + "issue:44857", + "issue:44912", + "issue:44987", + "issue:45042", + "issue:45356", + "issue:45357", + "issue:45459", + "issue:45538", + "issue:45803", + "issue:46291", + "issue:46421", + "issue:46424", + "issue:46701", + "issue:46832", + "issue:47145" + ], + "soft_pairs": [ + "issue:43329|issue:43450", + "issue:43329|issue:44479", + "issue:42915|issue:43742", + "issue:43257|issue:43931", + "issue:44912|issue:47145", + "issue:44291|issue:45042", + "issue:34567|issue:40990", + "issue:42915|issue:44821", + "issue:44387|issue:44617", + "issue:45459|issue:45803", + "issue:29159|issue:34567", + "issue:43650|issue:43906", + "issue:46421|issue:46424", + "issue:43525|issue:46291", + "issue:28282|issue:29870", + "issue:43749|issue:44756", + "issue:44987|issue:45538", + "issue:43425|issue:46291", + "issue:28282|issue:34689", + "issue:43540|issue:46701", + "issue:43454|issue:43577", + "issue:44568|issue:45356", + "issue:33357|issue:34567", + "issue:39379|issue:44554", + "issue:42175|issue:43824", + "issue:43697|issue:45357", + "issue:44857|issue:46832" + ], + "prepared_review_unit_hash": "fc3552733b4f21e70dae1accb347fe8de82f9d3dcd65d83824f7f991060d3539", + "component_budget": { + "node_count": 514, + "item_count": 514, + "soft_pair_count": 1891, + "serialized_chars": 980895, + "estimated_input_tokens": 245224, + "estimated_eval_tokens": 490704 + }, + "budget": { + "node_count": 48, + "item_count": 48, + "soft_pair_count": 27, + "serialized_chars": 35833, + "estimated_input_tokens": 8959, + "estimated_eval_tokens": 18174 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": true, + "analyst_result": { + "summary": "This cluster is mostly a collection of unrelated issue reports across loaders, tokenizers, multimodal/video, quantization, cache, and training utilities. Only one soft edge looks like a likely true duplicate around the Qwen video-input regression path.", + "confidence": 0.69, + "canonical_issue_reason": "issue:44479 is the best canonical issue here because it is the broader, user-facing regression report spanning multiple Qwen VL variants and a specific release window, and it likely subsumes the more specific video-token bug described in issue:43329.", + "canonical_pr_reason": null, + "best_issue_reason": "issue:44479 is the strongest representative issue overall: it clearly states the affected models, regression scope, and concrete failure mode, making it the most useful anchor if the video regression cluster is deduplicated.", + "best_pr_reason": null, + "soft_edge_verdicts": [ + { + "left": "issue:43329", + "right": "issue:43450", + "accept": false, + "reason": "Both involve video inputs, but one is a tokenizer/token-counting bug and the other is a batched processor shape issue; different code paths and symptoms." + }, + { + "left": "issue:43329", + "right": "issue:44479", + "accept": true, + "reason": "Both point to the same Qwen VLM video-input regression; 43329 identifies the likely root cause in multimodal token counting, while 44479 reports the broader user-facing breakage." + }, + { + "left": "issue:42915", + "right": "issue:43742", + "accept": false, + "reason": "Different models and failure modes: FP8 config handling vs a key error during MobileLLM loading." + }, + { + "left": "issue:43257", + "right": "issue:43931", + "accept": false, + "reason": "Both are Qwen loading problems, but one is accelerate/deepspeed MOE conversion and the other is a VL weight-shape mismatch; not the same bug." + }, + { + "left": "issue:44912", + "right": "issue:47145", + "accept": false, + "reason": "Both are quantization/loading correctness issues, but they involve different formats and code paths (MXFP4 fallback vs FP8 weight_scale application)." + }, + { + "left": "issue:44291", + "right": "issue:45042", + "accept": false, + "reason": "Unrelated subsystems: init_empty_weights loading failure vs an image processor torchvision dependency regression." + }, + { + "left": "issue:34567", + "right": "issue:40990", + "accept": false, + "reason": "TrainerState token counting is unrelated to the GPT-OSS perplexity regression." + }, + { + "left": "issue:42915", + "right": "issue:44821", + "accept": false, + "reason": "No concrete overlap: Qwen3Moe FP8 config vs AutoImageProcessor URL loading." + }, + { + "left": "issue:44387", + "right": "issue:44617", + "accept": false, + "reason": "Both mention memory/quantization-like stress, but one is int4 reserved-memory growth and the other is a SAM video OOM report; different bugs." + }, + { + "left": "issue:45459", + "right": "issue:45803", + "accept": false, + "reason": "Both involve exception handling, but they are separate tokenizer/data conversion code paths with different triggers and fixes." + }, + { + "left": "issue:29159", + "right": "issue:34567", + "accept": false, + "reason": "Tokenizer consistency and TrainerState token accounting are unrelated." + }, + { + "left": "issue:43650", + "right": "issue:43906", + "accept": false, + "reason": "A vague request and an isolated reproduction are not the same bug." + }, + { + "left": "issue:46421", + "right": "issue:46424", + "accept": false, + "reason": "Both are cache-related, but one is lazy init recompilation and the other is max_cache_length behavior for quantized cache only." + }, + { + "left": "issue:43525", + "right": "issue:46291", + "accept": false, + "reason": "Config attribute missing vs pipelines import failure due to kernels dependency metadata; unrelated." + }, + { + "left": "issue:28282", + "right": "issue:29870", + "accept": false, + "reason": "PyTorch missing ImportError vs Bart generate UnboundLocalError are unrelated." + }, + { + "left": "issue:43749", + "right": "issue:44756", + "accept": false, + "reason": "Both concern memory/loading, but one is FSDP CPU RAM-efficient loading and the other is mmap on Strix Halo; different fixes." + }, + { + "left": "issue:44987", + "right": "issue:45538", + "accept": false, + "reason": "Loading a specific model under transformers>=5.1.0 is unrelated to CLIPTokenizer's max_length constant." + }, + { + "left": "issue:43425", + "right": "issue:46291", + "accept": false, + "reason": "Torch version incompatibility is not the same as kernels version/revision requirements for pipelines import." + }, + { + "left": "issue:28282", + "right": "issue:34689", + "accept": false, + "reason": "Generic PyTorch missing-library error vs Llama 3.2 Vision loading regression are not the same issue." + }, + { + "left": "issue:43540", + "right": "issue:46701", + "accept": false, + "reason": "Different multimodal model families and different failure modes (video processing ValueError vs image-text pipeline output handling)." + }, + { + "left": "issue:43454", + "right": "issue:43577", + "accept": false, + "reason": "Weight tying/generation corruption in AyaVision is unrelated to Blip2 dtype propagation." + }, + { + "left": "issue:44568", + "right": "issue:45356", + "accept": false, + "reason": "Both are tokenizer regressions, but they affect different models and behaviors; not the same concrete bug." + }, + { + "left": "issue:33357", + "right": "issue:34567", + "accept": false, + "reason": "MacOS CLIP bus error is unrelated to TrainerState token tracking." + }, + { + "left": "issue:39379", + "right": "issue:44554", + "accept": false, + "reason": "mllama multimodal mask handling and MPS attention correctness are different bugs." + }, + { + "left": "issue:42175", + "right": "issue:43824", + "accept": false, + "reason": "Backend installation metadata issue vs missing exported model class; unrelated." + }, + { + "left": "issue:43697", + "right": "issue:45357", + "accept": false, + "reason": "RT-DETR output differences and Qwen3.5 save_pretrained key mapping are distinct issues." + }, + { + "left": "issue:44857", + "right": "issue:46832", + "accept": false, + "reason": "Both touch RT-DETR-ish code, but one is an AMP/CUDA loss crash and the other is a forward shape/type error when feature levels exceed backbone outputs." + } + ] + }, + "evaluator_result": { + "accept": true, + "feedback": "The summary is grounded in the packet and stays conservative. The only accepted soft edge, issue:43329 \u2194 issue:44479, is supported by the issue descriptions and the broader vs. specific video-regression framing is appropriately cautious. The remaining pairs are rejected with plausible distinctions rather than overclaiming duplication." + }, + "evaluator_used": true, + "retried": false, + "accepted_nontrivial_soft_edge": true, + "error_kind": null, + "error_message": null + }, + { + "label": "issue", + "component_index": 1, + "component_count": 27, + "review_unit_index": 23, + "review_unit_count": 63, + "status": "reviewed", + "reason": null, + "source": "llm", + "cache_hit": false, + "model": "gpt-5.4-mini?service_tier=flex", + "cluster_id": "cluster-10000-514", + "nodes": [ + "issue:24643", + "issue:27957", + "issue:28282", + "issue:29127", + "issue:29159", + "issue:34567", + "issue:36331", + "issue:38175", + "issue:41720", + "issue:42175", + "issue:42915", + "issue:43404", + "issue:43452", + "issue:43540", + "issue:43577", + "issue:43582", + "issue:43742", + "issue:43749", + "issue:44062", + "issue:44291", + "issue:44479", + "issue:44488", + "issue:44560", + "issue:44568", + "issue:44792", + "issue:44821", + "issue:44849", + "issue:44898", + "issue:44936", + "issue:44964", + "issue:44987", + "issue:44991", + "issue:45042", + "issue:45081", + "issue:45127", + "issue:45137", + "issue:45381", + "issue:45405", + "issue:45488", + "issue:45538", + "issue:45701", + "issue:45907", + "issue:46512", + "issue:46693", + "issue:46821", + "issue:46869", + "issue:47020" + ], + "soft_pairs": [ + "issue:45137|issue:45405", + "issue:43404|issue:45127", + "issue:45081|issue:47020", + "issue:28282|issue:29127", + "issue:43742|issue:44291", + "issue:44936|issue:46821", + "issue:42915|issue:45538", + "issue:44479|issue:46693", + "issue:45538|issue:46869", + "issue:44062|issue:44991", + "issue:24643|issue:29159", + "issue:43452|issue:43582", + "issue:27957|issue:28282", + "issue:42915|issue:43749", + "issue:43577|issue:46512", + "issue:41720|issue:43540", + "issue:44849|issue:45701", + "issue:42175|issue:44488", + "issue:43540|issue:44560", + "issue:44987|issue:45907", + "issue:43540|issue:44479", + "issue:44291|issue:44821", + "issue:44479|issue:45381", + "issue:34567|issue:36331", + "issue:38175|issue:44898", + "issue:44792|issue:44964", + "issue:44062|issue:45081", + "issue:44568|issue:45488", + "issue:45042|issue:45081" + ], + "prepared_review_unit_hash": "e8dd45ad15ae06cf75063db789668692ea6cb54800e8a910804d1c9801637394", + "component_budget": { + "node_count": 514, + "item_count": 514, + "soft_pair_count": 1891, + "serialized_chars": 980895, + "estimated_input_tokens": 245224, + "estimated_eval_tokens": 490704 + }, + "budget": { + "node_count": 47, + "item_count": 47, + "soft_pair_count": 29, + "serialized_chars": 35832, + "estimated_input_tokens": 8958, + "estimated_eval_tokens": 18172 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": true, + "analyst_result": { + "summary": "This cluster is mostly a loose semantic group of unrelated Transformers bug reports. Only the Qwen2.5-VL / Qwen3-VL video-input regression pair looks like the same underlying issue; the rest are distinct bugs in different models/code paths.", + "confidence": 0.84, + "canonical_issue_reason": "issue:44479 is the broadest and clearest report of the shared video-input regression, covering multiple Qwen multimodal variants; issue:45381 appears to be a narrower symptom/report of the same problem.", + "canonical_pr_reason": null, + "best_issue_reason": "issue:44479 is the best representative issue because it captures the regression at the right abstraction level and scope, making it the most useful duplicate target.", + "best_pr_reason": null, + "soft_edge_verdicts": [ + { + "left": "issue:45137", + "right": "issue:45405", + "accept": false, + "reason": "DeepSpeed ZeRO3 deque/index failure and a PEFT version bump are unrelated." + }, + { + "left": "issue:43404", + "right": "issue:45127", + "accept": false, + "reason": "Both mention tied weights, but one is a Mistral lm_head tying bug while the other is LoRA merge collapse with extended vocab; different code paths." + }, + { + "left": "issue:45081", + "right": "issue:47020", + "accept": false, + "reason": "Different tokenizer classes and different failures: Mistral regex patch crash vs XLMRoberta Unigram init error." + }, + { + "left": "issue:28282", + "right": "issue:29127", + "accept": false, + "reason": "PyTorch missing import error is unrelated to LayoutLMV3's error-message clarity issue." + }, + { + "left": "issue:43742", + "right": "issue:44291", + "accept": false, + "reason": "MobileLLM loading key error and init_empty_weights unexpected argument are separate loading bugs." + }, + { + "left": "issue:44936", + "right": "issue:46821", + "accept": false, + "reason": "Trainer evaluate-after-train failure is unrelated to TimesFM window_size AttributeError." + }, + { + "left": "issue:42915", + "right": "issue:45538", + "accept": false, + "reason": "FineGrainedFP8Config failure and CLIPTokenizer max-length behavior are unrelated." + }, + { + "left": "issue:44479", + "right": "issue:46693", + "accept": false, + "reason": "Qwen video-input regression and FA2 long-context performance regression are different problems." + }, + { + "left": "issue:45538", + "right": "issue:46869", + "accept": false, + "reason": "Tokenizer max-length handling is unrelated to YOLO multi-GPU device-map failure." + }, + { + "left": "issue:44062", + "right": "issue:44991", + "accept": false, + "reason": "AddedToken special-argument collision and est-roberta tokenizer loading failure are not the same bug." + }, + { + "left": "issue:24643", + "right": "issue:29159", + "accept": false, + "reason": "DeepSpeed training weight-shape error is unrelated to tokenizer slow/fast inconsistency." + }, + { + "left": "issue:43452", + "right": "issue:43582", + "accept": false, + "reason": "gguf_file loading and Apple Silicon caching_allocator_warmup TypeError are unrelated." + }, + { + "left": "issue:27957", + "right": "issue:28282", + "accept": false, + "reason": "Flash Attention 2 with XLMRoberta is unrelated to the PyTorch-not-found import error." + }, + { + "left": "issue:42915", + "right": "issue:43749", + "accept": false, + "reason": "Qwen3Moe FP8 config failure and FSDP CPU RAM efficient loading are different training/loading bugs." + }, + { + "left": "issue:43577", + "right": "issue:46512", + "accept": false, + "reason": "Both involve dtype, but one is BLIP2 load-time dtype propagation and the other is stale config dtype from from_config; not the same concrete bug." + }, + { + "left": "issue:41720", + "right": "issue:43540", + "accept": false, + "reason": "Auto device-mapping cudaErrorAssert and Qwen3OmniMoe video ValueError are different failures." + }, + { + "left": "issue:44849", + "right": "issue:45701", + "accept": false, + "reason": "Qwen3.5 hidden-states bug and version-dependent tokenization changes are unrelated." + }, + { + "left": "issue:42175", + "right": "issue:44488", + "accept": false, + "reason": "Packaging/backend dependency issue and loading a specific model are unrelated." + }, + { + "left": "issue:43540", + "right": "issue:44560", + "accept": false, + "reason": "Both are video-input issues in Qwen models, but they affect different models and error paths; not enough evidence of the same bug." + }, + { + "left": "issue:44987", + "right": "issue:45907", + "accept": false, + "reason": "Model loading failure and list_repo_files revision forwarding are unrelated." + }, + { + "left": "issue:43540", + "right": "issue:44479", + "accept": false, + "reason": "Different Qwen multimodal models and different failures; too broad to treat as the same bug." + }, + { + "left": "issue:44291", + "right": "issue:44821", + "accept": false, + "reason": "init_empty_weights constructor compatibility and AutoImageProcessor URL loading are separate issues." + }, + { + "left": "issue:44479", + "right": "issue:45381", + "accept": true, + "reason": "Both report the same Qwen2.5-VL/Qwen3-VL video-input regression, with 45381 describing the wrong vision_position_ids symptom of the broader 44479 issue." + }, + { + "left": "issue:34567", + "right": "issue:36331", + "accept": false, + "reason": "TrainerState token-count tracking and compute_loss signature incompatibility are unrelated trainer bugs." + }, + { + "left": "issue:38175", + "right": "issue:44898", + "accept": false, + "reason": "Unexpected zero probabilities and Perceiver interpolation failure are different model-specific inference bugs." + }, + { + "left": "issue:44792", + "right": "issue:44964", + "accept": false, + "reason": "Janus image-generation test failure and Phi-4 multimodal loading failure are unrelated." + }, + { + "left": "issue:44062", + "right": "issue:45081", + "accept": false, + "reason": "AddedToken special handling and Mistral regex patch crash are different tokenizer regressions." + }, + { + "left": "issue:44568", + "right": "issue:45488", + "accept": false, + "reason": "MDeBERTa special-token insertion and LlamaTokenizer ByteLevel/Metaspace override are distinct tokenizer regressions." + }, + { + "left": "issue:45042", + "right": "issue:45081", + "accept": false, + "reason": "PIL/torchvision dependency behavior is unrelated to Mistral tokenizer regex patching." + } + ] + }, + "evaluator_result": { + "accept": true, + "feedback": "The summary is well grounded in the packet and stays conservative. The canonical issue choice is supported by the titles: 44479 is the broader v5.3.0 video-input regression report, while 45381 is a narrower qwen2.5-vl symptom. The other soft-pair rejections are appropriately cautious given only title-level evidence." + }, + "evaluator_used": true, + "retried": false, + "accepted_nontrivial_soft_edge": true, + "error_kind": null, + "error_message": null + }, + { + "label": "issue", + "component_index": 1, + "component_count": 27, + "review_unit_index": 24, + "review_unit_count": 63, + "status": "reviewed", + "reason": null, + "source": "llm", + "cache_hit": false, + "model": "gpt-5.4-mini?service_tier=flex", + "cluster_id": "cluster-10000-514", + "nodes": [ + "issue:24643", + "issue:28282", + "issue:29870", + "issue:32090", + "issue:43257", + "issue:43452", + "issue:43526", + "issue:43531", + "issue:43540", + "issue:43577", + "issue:43582", + "issue:43697", + "issue:43716", + "issue:43742", + "issue:43756", + "issue:43906", + "issue:44488", + "issue:44568", + "issue:44589", + "issue:44625", + "issue:44737", + "issue:44821", + "issue:44877", + "issue:44913", + "issue:45072", + "issue:45216", + "issue:45325", + "issue:45357", + "issue:45405", + "issue:45479", + "issue:45588", + "issue:45600", + "issue:45663", + "issue:45672", + "issue:45674", + "issue:45823", + "issue:45920", + "issue:46121", + "issue:46399", + "issue:46463", + "issue:46468", + "issue:46491", + "issue:46693", + "issue:46701", + "issue:46821", + "issue:46829", + "issue:46846", + "issue:46899" + ], + "soft_pairs": [ + "issue:43257|issue:43756", + "issue:43742|issue:44589", + "issue:43582|issue:46821", + "issue:43531|issue:44821", + "issue:43756|issue:46399", + "issue:43756|issue:45325", + "issue:45405|issue:45600", + "issue:44737|issue:46463", + "issue:43716|issue:46701", + "issue:43906|issue:46693", + "issue:45823|issue:46463", + "issue:43577|issue:45072", + "issue:44568|issue:46491", + "issue:24643|issue:29870", + "issue:45672|issue:45674", + "issue:44625|issue:44877", + "issue:43452|issue:44488", + "issue:43540|issue:46693", + "issue:43906|issue:46468", + "issue:28282|issue:32090", + "issue:43697|issue:46829", + "issue:44913|issue:46121", + "issue:45663|issue:46846", + "issue:43526|issue:45479", + "issue:44737|issue:45588", + "issue:46463|issue:46899", + "issue:43697|issue:45216", + "issue:45357|issue:45920" + ], + "prepared_review_unit_hash": "c2621cb83b2c6171bd50bb6ab21de5b6642411dcad80b72e6cb3c649bdeb5b3f", + "component_budget": { + "node_count": 514, + "item_count": 514, + "soft_pair_count": 1891, + "serialized_chars": 980895, + "estimated_input_tokens": 245224, + "estimated_eval_tokens": 490704 + }, + "budget": { + "node_count": 48, + "item_count": 48, + "soft_pair_count": 28, + "serialized_chars": 36581, + "estimated_input_tokens": 9146, + "estimated_eval_tokens": 18548 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": true, + "analyst_result": { + "summary": "This is a heterogeneous issue cluster with superficial lexical overlap (mostly around model loading, dtype, rope, and config terms), but none of the soft pairs look like true duplicates of the same underlying bug.", + "confidence": 0.9, + "canonical_issue_reason": null, + "canonical_pr_reason": null, + "best_issue_reason": "issue:43756 is the most connected node in the soft-similarity graph, but that connectivity is spurious; the cluster is too mixed for a strong canonical issue.", + "best_pr_reason": null, + "soft_edge_verdicts": [ + { + "left": "issue:43257", + "right": "issue:43756", + "accept": false, + "reason": "Qwen3 MOE weight-conversion/loading bug vs. Smollm3 RoPE-layer mismatch; different models and failure modes." + }, + { + "left": "issue:43742", + "right": "issue:44589", + "accept": false, + "reason": "MobileLLM key-loading error vs. Float8 storage lookup TypeError; unrelated load paths." + }, + { + "left": "issue:43582", + "right": "issue:46821", + "accept": false, + "reason": "Apple Silicon warmup TypeError vs. TimesFM window_size AttributeError; different APIs and devices." + }, + { + "left": "issue:43531", + "right": "issue:44821", + "accept": false, + "reason": "Qwen3-MoE sliding-window behavior vs. AutoImageProcessor URL loading; no shared code-path bug." + }, + { + "left": "issue:43756", + "right": "issue:46399", + "accept": false, + "reason": "Smollm3 RoPE-layer dropping vs. Qwen3.5 apply_rotary_pos_emb registration error; both rope-related but not the same bug." + }, + { + "left": "issue:43756", + "right": "issue:45325", + "accept": false, + "reason": "RoPE layer-count mismatch vs. Qwen2.5-VL temporal position scaling; distinct model-specific rotary issues." + }, + { + "left": "issue:45405", + "right": "issue:45600", + "accept": false, + "reason": "PEFT version-pin problem vs. stale auto_mappings reference breakage; unrelated maintenance issues." + }, + { + "left": "issue:44737", + "right": "issue:46463", + "accept": false, + "reason": "XLNet CPU tensor allocation bug vs. T5 decoder input validation message; different subsystems entirely." + }, + { + "left": "issue:43716", + "right": "issue:46701", + "accept": false, + "reason": "Mistral-3 image dtype mismatch vs. DiffusionGemma pipeline output bug; different multimodal paths." + }, + { + "left": "issue:43906", + "right": "issue:46693", + "accept": false, + "reason": "A reproduction of an older issue vs. a FA2 long-context performance regression; not the same concrete defect." + }, + { + "left": "issue:45823", + "right": "issue:46463", + "accept": false, + "reason": "Gemma4 PLE device-map mismatch vs. T5 decoder validation error; unrelated model families." + }, + { + "left": "issue:43577", + "right": "issue:45072", + "accept": false, + "reason": "BLIP2 dtype propagation issue vs. SwitchTransformers/TimmWrapper bfloat16 mismatch; similar theme, different code paths." + }, + { + "left": "issue:44568", + "right": "issue:46491", + "accept": false, + "reason": "Tokenizer special-token regression vs. CodeLlama space round-trip regression; both tokenizer bugs but not the same behavior." + }, + { + "left": "issue:24643", + "right": "issue:29870", + "accept": false, + "reason": "DeepSpeed training weight-shape RuntimeError vs. Bart generate unbound variable; distinct training/generation failures." + }, + { + "left": "issue:45672", + "right": "issue:45674", + "accept": false, + "reason": "Gemma4 4-bit audio/vision dtype cast bug vs. BitsAndBytes skip-module exclusion bug; different quantization failures." + }, + { + "left": "issue:44625", + "right": "issue:44877", + "accept": false, + "reason": "Qwen3.5 num_labels propagation vs. strict config blocking granite_speech; different configuration validations." + }, + { + "left": "issue:43452", + "right": "issue:44488", + "accept": false, + "reason": "gguf_file loading regression vs. sleng-bert model loading failure; overlapping symptom only, not same root cause." + }, + { + "left": "issue:43540", + "right": "issue:46693", + "accept": false, + "reason": "Qwen3OmniMoe video-input ValueError vs. FA2 long-context generation slowdown; unrelated bugs." + }, + { + "left": "issue:43906", + "right": "issue:46468", + "accept": false, + "reason": "Reproduction of issue 38071 vs. broken Nemotron example; not enough evidence of same defect." + }, + { + "left": "issue:28282", + "right": "issue:32090", + "accept": false, + "reason": "Missing PyTorch import error vs. NoneType broadcast TypeError in Trainer; different layers of the stack." + }, + { + "left": "issue:43697", + "right": "issue:46829", + "accept": false, + "reason": "RTDetrV2 output divergence vs. VideoPrism output field mismatch; different models and outputs." + }, + { + "left": "issue:44913", + "right": "issue:46121", + "accept": false, + "reason": "GPTNeoX config reload issue vs. rope-parameter JSON list TypeError; unrelated config serialization bugs." + }, + { + "left": "issue:45663", + "right": "issue:46846", + "accept": false, + "reason": "Gemma4 FSDP2 shared_kv_states KeyError vs. Qwen3.5 tensor-parallel plan omission; different failure modes." + }, + { + "left": "issue:43526", + "right": "issue:45479", + "accept": false, + "reason": "BEiT image processor label reduction bug vs. sequence-classification zero-loss bug; not the same code path." + }, + { + "left": "issue:44737", + "right": "issue:45588", + "accept": false, + "reason": "XLNet CPU allocation issue vs. flash-attention None-handling crash; different components." + }, + { + "left": "issue:46463", + "right": "issue:46899", + "accept": false, + "reason": "T5 decoder validation message bug vs. Gemma4 multimodal 4-bit dtype cast; unrelated models and failures." + }, + { + "left": "issue:43697", + "right": "issue:45216", + "accept": false, + "reason": "RTDetrV2 output mismatch vs. Qwen3.5 save_pretrained regression; no shared underlying bug." + }, + { + "left": "issue:45357", + "right": "issue:45920", + "accept": false, + "reason": "Qwen3.5 incorrect visual-encoder checkpoint saving vs. wrong token IDs in AutoTokenizer; separate serialization/tokenization 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": 27, + "review_unit_index": 25, + "review_unit_count": 63, + "status": "reviewed", + "reason": null, + "source": "llm", + "cache_hit": false, + "model": "gpt-5.4-mini?service_tier=flex", + "cluster_id": "cluster-10000-514", + "nodes": [ + "issue:29159", + "issue:33357", + "issue:38175", + "issue:41762", + "issue:43295", + "issue:43452", + "issue:43525", + "issue:43531", + "issue:43643", + "issue:43646", + "issue:43653", + "issue:43688", + "issue:43720", + "issue:43742", + "issue:43749", + "issue:43756", + "issue:43761", + "issue:43819", + "issue:43994", + "issue:44361", + "issue:44442", + "issue:44466", + "issue:44493", + "issue:44568", + "issue:44811", + "issue:44843", + "issue:44857", + "issue:44898", + "issue:44936", + "issue:45005", + "issue:45200", + "issue:45356", + "issue:45372", + "issue:45375", + "issue:45542", + "issue:45701", + "issue:45704", + "issue:46050", + "issue:46170", + "issue:46178", + "issue:46402", + "issue:46463", + "issue:46701", + "issue:46829", + "issue:46832", + "issue:46932", + "issue:47000" + ], + "soft_pairs": [ + "issue:41762|issue:44936", + "issue:43720|issue:44898", + "issue:44466|issue:45005", + "issue:45200|issue:46463", + "issue:43643|issue:45375", + "issue:44493|issue:46050", + "issue:44361|issue:44442", + "issue:43688|issue:43749", + "issue:38175|issue:43295", + "issue:43653|issue:44361", + "issue:43452|issue:44843", + "issue:43531|issue:43742", + "issue:43994|issue:44898", + "issue:43819|issue:44811", + "issue:45372|issue:46170", + "issue:43761|issue:46829", + "issue:44857|issue:47000", + "issue:43525|issue:46701", + "issue:29159|issue:33357", + "issue:45542|issue:46178", + "issue:46832|issue:46932", + "issue:43295|issue:45356", + "issue:43688|issue:45704", + "issue:43653|issue:44568", + "issue:43688|issue:45701", + "issue:43646|issue:43749", + "issue:43756|issue:46402" + ], + "prepared_review_unit_hash": "0c11754a2a6b92c4e089c9267e8696e84b37b35b3604bf9b773e9024af8e251a", + "component_budget": { + "node_count": 514, + "item_count": 514, + "soft_pair_count": 1891, + "serialized_chars": 980895, + "estimated_input_tokens": 245224, + "estimated_eval_tokens": 490704 + }, + "budget": { + "node_count": 47, + "item_count": 47, + "soft_pair_count": 27, + "serialized_chars": 35499, + "estimated_input_tokens": 8875, + "estimated_eval_tokens": 18006 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": true, + "analyst_result": { + "summary": "The soft pairs are mostly false positives: they share broad subsystems (tokenizers, loading, vision models, losses) but not the same concrete bug or code path. Overall this looks like a heterogeneous set of unrelated issues rather than a duplicate cluster.", + "confidence": 0.96, + "canonical_issue_reason": "No single issue cleanly represents a true duplicate set here; the items are heterogeneous. If one must be chosen, #43646 is the broadest umbrella-style regression report about v5 breaking custom model initialization, though it does not subsume the others.", + "canonical_pr_reason": null, + "best_issue_reason": "#43646 is the most general and representative issue among the set, but only as a broad regression summary rather than a duplicate canonical.", + "best_pr_reason": null, + "soft_edge_verdicts": [ + { + "left": "issue:41762", + "right": "issue:44936", + "accept": false, + "reason": "Different failures: Gemma3 load-time ZeRO-3 index error vs trainer.evaluate() after training. No shared code path." + }, + { + "left": "issue:43720", + "right": "issue:44898", + "accept": false, + "reason": "Unrelated models and bugs: BitNet packed-weight loading vs Perceiver non-default resolution handling." + }, + { + "left": "issue:44466", + "right": "issue:45005", + "accept": false, + "reason": "Both mention tied weights, but one is device-dependent lm_head serialization and the other is translation-model tied-weight issues; too broad to treat as the same bug." + }, + { + "left": "issue:45200", + "right": "issue:46463", + "accept": false, + "reason": "Different model families and symptoms: Gemma 4 mm_token_type_ids defaulting vs T5 decoder-input validation message." + }, + { + "left": "issue:43643", + "right": "issue:45375", + "accept": false, + "reason": "Both are config-field related, but one is a generic AutoConfig trust_remote_code issue and the other is a specific strict-drop of a Gemma/Qwen vision field." + }, + { + "left": "issue:44493", + "right": "issue:46050", + "accept": false, + "reason": "State-dict/key mismatch vs quantized save/load breakage; similar theme but not the same underlying defect." + }, + { + "left": "issue:44361", + "right": "issue:44442", + "accept": false, + "reason": "Both are tokenizer-loading failures, but they affect different tokenizers and different failure modes." + }, + { + "left": "issue:43688", + "right": "issue:43749", + "accept": false, + "reason": "Aux-loss normalization bug vs FSDP CPU RAM efficient loading regression; unrelated subsystems." + }, + { + "left": "issue:38175", + "right": "issue:43295", + "accept": false, + "reason": "Different problems: SigLIP2 zero probabilities vs processor.tokenizer/custom image-tokenizer regression." + }, + { + "left": "issue:43653", + "right": "issue:44361", + "accept": false, + "reason": "BigBird special-token registration bug and MLukeTokenizer AttributeError are separate tokenizer issues." + }, + { + "left": "issue:43452", + "right": "issue:44843", + "accept": false, + "reason": "Both involve AutoTokenizer.from_pretrained, but the causes differ: gguf_file loading vs an unconditional model_info() call breaking offline mode." + }, + { + "left": "issue:43531", + "right": "issue:43742", + "accept": false, + "reason": "Qwen3-MoE sliding_window behavior and MobileLLM key error are unrelated." + }, + { + "left": "issue:43994", + "right": "issue:44898", + "accept": false, + "reason": "Both are vision-related inference issues, but with different models and symptoms; not the same bug." + }, + { + "left": "issue:43819", + "right": "issue:44811", + "accept": false, + "reason": "DAC forward/from_latents mismatch is unrelated to Whisper batch_decode skip_special_tokens handling." + }, + { + "left": "issue:45372", + "right": "issue:46170", + "accept": false, + "reason": "Different import/dependency failures on different platforms and libraries." + }, + { + "left": "issue:43761", + "right": "issue:46829", + "accept": false, + "reason": "Both concern hidden_states outputs, but they are different models and different return-path bugs." + }, + { + "left": "issue:44857", + "right": "issue:47000", + "accept": false, + "reason": "Different detection-loss crashes with different error conditions; not one concrete defect." + }, + { + "left": "issue:43525", + "right": "issue:46701", + "accept": false, + "reason": "Llama4Config missing pad_token_id is unrelated to DiffusionGemma pipeline failure." + }, + { + "left": "issue:29159", + "right": "issue:33357", + "accept": false, + "reason": "Tokenizer consistency issue vs MacOS bus error in a CLIP community model; unrelated." + }, + { + "left": "issue:45542", + "right": "issue:46178", + "accept": false, + "reason": "Both concern optional TensorFlow-related backends, but they are distinct backend-resolution bugs with different offending packages." + }, + { + "left": "issue:46832", + "right": "issue:46932", + "accept": false, + "reason": "RT-DETR feature-level mismatch and Gemma4 dtype casting for quantized embedders are unrelated." + }, + { + "left": "issue:43295", + "right": "issue:45356", + "accept": false, + "reason": "Both are regressions around tokenizers/processors, but the concrete breakage and affected models differ." + }, + { + "left": "issue:43688", + "right": "issue:45704", + "accept": false, + "reason": "Aux-loss normalization and T5 apex.FusedRMSNorm memory leak are unrelated." + }, + { + "left": "issue:43653", + "right": "issue:44568", + "accept": false, + "reason": "Both are special-token issues, but they affect different tokenizers and different token types." + }, + { + "left": "issue:43688", + "right": "issue:45701", + "accept": false, + "reason": "Aux-loss normalization bug is not the same as a broad tokenization-version-change report." + }, + { + "left": "issue:43646", + "right": "issue:43749", + "accept": false, + "reason": "Generic v5 custom-model initialization regression vs FSDP loading breakage; different code paths." + }, + { + "left": "issue:43756", + "right": "issue:46402", + "accept": false, + "reason": "Smollm3 RoPE-layer handling and Sam3Model text encoder weight loading 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": 27, + "review_unit_index": 26, + "review_unit_count": 63, + "status": "reviewed", + "reason": null, + "source": "llm", + "cache_hit": false, + "model": "gpt-5.4-mini?service_tier=flex", + "cluster_id": "cluster-10000-514", + "nodes": [ + "issue:32090", + "issue:33666", + "issue:36010", + "issue:36331", + "issue:41720", + "issue:42898", + "issue:43262", + "issue:43278", + "issue:43295", + "issue:43452", + "issue:43504", + "issue:43550", + "issue:43577", + "issue:43606", + "issue:43611", + "issue:43653", + "issue:43688", + "issue:43720", + "issue:43761", + "issue:43782", + "issue:43927", + "issue:44464", + "issue:44466", + "issue:44479", + "issue:44484", + "issue:44514", + "issue:44554", + "issue:44610", + "issue:44625", + "issue:44704", + "issue:44734", + "issue:44743", + "issue:44821", + "issue:44843", + "issue:44977", + "issue:44991", + "issue:45290", + "issue:45440", + "issue:45491", + "issue:46002", + "issue:46050", + "issue:46491", + "issue:46612", + "issue:46693", + "issue:46821", + "issue:46846", + "issue:47013" + ], + "soft_pairs": [ + "issue:44464|issue:44743", + "issue:43653|issue:43927", + "issue:44821|issue:46821", + "issue:42898|issue:46491", + "issue:43452|issue:43577", + "issue:44843|issue:44991", + "issue:44479|issue:44625", + "issue:43761|issue:47013", + "issue:33666|issue:36010", + "issue:44484|issue:44704", + "issue:32090|issue:33666", + "issue:44466|issue:46050", + "issue:43688|issue:45440", + "issue:44977|issue:46693", + "issue:41720|issue:44464", + "issue:44734|issue:46612", + "issue:44464|issue:44977", + "issue:43504|issue:43550", + "issue:33666|issue:36331", + "issue:43504|issue:43720", + "issue:43278|issue:43688", + "issue:43262|issue:43295", + "issue:43782|issue:46693", + "issue:45491|issue:46002", + "issue:43720|issue:43927", + "issue:43577|issue:44554", + "issue:43611|issue:46846", + "issue:43606|issue:43927", + "issue:43504|issue:44610", + "issue:44514|issue:45290" + ], + "prepared_review_unit_hash": "889eaa35b3e971e30337e7b1557019daf59d33f754a221e674f459fe30a34ab9", + "component_budget": { + "node_count": 514, + "item_count": 514, + "soft_pair_count": 1891, + "serialized_chars": 980895, + "estimated_input_tokens": 245224, + "estimated_eval_tokens": 490704 + }, + "budget": { + "node_count": 47, + "item_count": 47, + "soft_pair_count": 30, + "serialized_chars": 36445, + "estimated_input_tokens": 9112, + "estimated_eval_tokens": 18480 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": true, + "analyst_result": { + "summary": "This is a loose themed collection of Transformers issues, but the proposed soft links mostly connect different bugs in different models or code paths (loading, generation, tokenization, dtype, device placement, etc.). I would reject all soft edges as duplicates.", + "confidence": 0.94, + "canonical_issue_reason": "No canonical issue: the set fragments into many unrelated bugs, so no single issue cleanly represents the cluster.", + "canonical_pr_reason": null, + "best_issue_reason": "If forced to pick a loose anchor, issue 44514 is the closest to a repeated apply_chat_template crash pattern, but it still only covers a small slice of the set and is not a true duplicate hub.", + "best_pr_reason": null, + "soft_edge_verdicts": [ + { + "left": "issue:44464", + "right": "issue:44743", + "accept": false, + "reason": "Both involve generation/cache behavior, but one is a compiled-forward inconsistency and the other is a recurrent-state unpacking bug in a different model path." + }, + { + "left": "issue:43653", + "right": "issue:43927", + "accept": false, + "reason": "BigBird tokenizer special-token decoding and DiaConfig token-ID persistence are different failure modes and different subsystems." + }, + { + "left": "issue:44821", + "right": "issue:46821", + "accept": false, + "reason": "AutoImageProcessor URL loading and TimesFM window_size AttributeError are unrelated loader/model issues." + }, + { + "left": "issue:42898", + "right": "issue:46491", + "accept": false, + "reason": "Tokenization-space cleanup regression and CodeLlama leading-space round-trip regression are similar only at a very high level, not the same bug." + }, + { + "left": "issue:43452", + "right": "issue:43577", + "accept": false, + "reason": "GGUF loading breakage and BLIP2 dtype propagation are distinct loading vs dtype issues." + }, + { + "left": "issue:44843", + "right": "issue:44991", + "accept": false, + "reason": "Offline model_info() behavior in tokenizer patching and a specific tokenizer load failure for EMBEDDIA/est-roberta are not the same root cause." + }, + { + "left": "issue:44479", + "right": "issue:44625", + "accept": false, + "reason": "Qwen video-input regression and Qwen3.5 num_labels propagation affect different code paths and symptoms." + }, + { + "left": "issue:43761", + "right": "issue:47013", + "accept": false, + "reason": "CLIPVision hidden_states handling and xLSTM train-mode state unpacking are unrelated model bugs." + }, + { + "left": "issue:33666", + "right": "issue:36010", + "accept": false, + "reason": "Qwen2-VL multi-GPU training and GenerationMixin import failure are completely different issues." + }, + { + "left": "issue:44484", + "right": "issue:44704", + "accept": false, + "reason": "save_pretrained shard-size behavior and AutoProcessor cached_file kwarg forwarding are unrelated API problems." + }, + { + "left": "issue:32090", + "right": "issue:33666", + "accept": false, + "reason": "Trainer gpu_broadcast NoneType error and Qwen2-VL multi-GPU training are not the same defect." + }, + { + "left": "issue:44466", + "right": "issue:46050", + "accept": false, + "reason": "Serialization inconsistency of lm_head.weight and sinq quantized save/load breakage are different persistence bugs." + }, + { + "left": "issue:43688", + "right": "issue:45440", + "accept": false, + "reason": "Auxiliary-loss normalization and DeepSeekV3MoE divergence are separate model-behavior issues." + }, + { + "left": "issue:44977", + "right": "issue:46693", + "accept": false, + "reason": "Qwen3.5 flash-attention generation failure and a long-context FA2 performance regression are different symptoms and likely different causes." + }, + { + "left": "issue:41720", + "right": "issue:44464", + "accept": false, + "reason": "Auto device mapping cuda assert on Qwen3 and chunked-generation compiled-forward inconsistency are unrelated generation/runtime bugs." + }, + { + "left": "issue:44734", + "right": "issue:46612", + "accept": false, + "reason": "Serving KV-cache continuation indexing and beam-search cache reorder being skipped are both cache-related, but they are distinct code paths and bugs." + }, + { + "left": "issue:44464", + "right": "issue:44977", + "accept": false, + "reason": "Compiled-forward chunked generation inconsistency and Qwen3.5 flash-attention output failure are not the same underlying problem." + }, + { + "left": "issue:43504", + "right": "issue:43550", + "accept": false, + "reason": "Legacy field loading for Beit and Bamba torch.compile+SDPA failure are unrelated model-loading/runtime bugs." + }, + { + "left": "issue:33666", + "right": "issue:36331", + "accept": false, + "reason": "Qwen2-VL multi-GPU training and CustomTrainer.compute_loss signature mismatch are different Trainer-side issues." + }, + { + "left": "issue:43504", + "right": "issue:43720", + "accept": false, + "reason": "Beit legacy-field load failure and BitNet packed-weight unpacking during accelerate loading are separate loading problems." + }, + { + "left": "issue:43278", + "right": "issue:43688", + "accept": false, + "reason": "Embedding dtype drift in eval and auxiliary-loss normalization are different model numerics issues." + }, + { + "left": "issue:43262", + "right": "issue:43295", + "accept": false, + "reason": "Audio chat-template sampling-rate default and processor.tokenizer/image-handling regression are unrelated processor bugs." + }, + { + "left": "issue:43782", + "right": "issue:46693", + "accept": false, + "reason": "Qwen3VL weight_only loading error and FA2 long-context performance regression do not share a concrete code-path failure." + }, + { + "left": "issue:45491", + "right": "issue:46002", + "accept": false, + "reason": "Gemma3 mixed-length NaN embeddings and classification-head NaNs on checkout/init are similar only by symptom, not root cause." + }, + { + "left": "issue:43720", + "right": "issue:43927", + "accept": false, + "reason": "BitNet packed-weight loading and DiaConfig token-ID persistence are unrelated." + }, + { + "left": "issue:43577", + "right": "issue:44554", + "accept": false, + "reason": "BLIP2 dtype mismatch and an MPS attention correctness issue are different bugs in different execution paths." + }, + { + "left": "issue:43611", + "right": "issue:46846", + "accept": false, + "reason": "base_model_prefix loading regression and missing linear_attn TP-plan entries are separate model-compatibility issues." + }, + { + "left": "issue:43606", + "right": "issue:43927", + "accept": false, + "reason": "CPU offload device mismatch for bark-small and DiaConfig token-ID loss are unrelated." + }, + { + "left": "issue:43504", + "right": "issue:44610", + "accept": false, + "reason": "Beit legacy-field loading and OmDet processor/model-size mismatch are not the same underlying bug." + }, + { + "left": "issue:44514", + "right": "issue:45290", + "accept": false, + "reason": "Both hit apply_chat_template, but one is batched padding=False and the other is assistant tool-call/no-content handling; these are distinct 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": 27, + "review_unit_index": 27, + "review_unit_count": 63, + "status": "reviewed", + "reason": null, + "source": "llm", + "cache_hit": false, + "model": "gpt-5.4-mini?service_tier=flex", + "cluster_id": "cluster-10000-514", + "nodes": [ + "issue:28218", + "issue:33357", + "issue:33666", + "issue:39692", + "issue:42796", + "issue:43278", + "issue:43295", + "issue:43452", + "issue:43493", + "issue:43525", + "issue:43572", + "issue:43575", + "issue:43643", + "issue:43646", + "issue:43701", + "issue:43720", + "issue:43825", + "issue:43844", + "issue:43927", + "issue:43994", + "issue:44291", + "issue:44361", + "issue:44451", + "issue:44466", + "issue:44488", + "issue:44554", + "issue:44561", + "issue:44568", + "issue:44589", + "issue:44625", + "issue:44792", + "issue:44843", + "issue:44849", + "issue:44964", + "issue:44977", + "issue:45405", + "issue:45520", + "issue:45584", + "issue:45812", + "issue:45823", + "issue:46050", + "issue:46192", + "issue:46209", + "issue:46365", + "issue:46489", + "issue:46710", + "issue:46821", + "issue:46832" + ], + "soft_pairs": [ + "issue:43525|issue:43572", + "issue:43643|issue:44561", + "issue:43720|issue:44361", + "issue:44964|issue:45405", + "issue:44625|issue:45584", + "issue:43493|issue:43825", + "issue:44977|issue:46209", + "issue:44554|issue:46365", + "issue:28218|issue:33357", + "issue:43452|issue:44451", + "issue:45823|issue:46050", + "issue:39692|issue:43295", + "issue:42796|issue:44977", + "issue:43994|issue:46192", + "issue:43493|issue:43844", + "issue:43278|issue:43844", + "issue:43278|issue:43701", + "issue:43646|issue:44466", + "issue:46209|issue:46710", + "issue:44568|issue:46489", + "issue:33357|issue:33666", + "issue:43927|issue:44361", + "issue:44849|issue:46209", + "issue:44568|issue:45812", + "issue:44792|issue:45405", + "issue:43575|issue:45405", + "issue:43572|issue:44291", + "issue:44488|issue:44843", + "issue:44589|issue:45520", + "issue:46821|issue:46832" + ], + "prepared_review_unit_hash": "e9051c3688997d75a744d1b8a4f7f258a9cd5dc8afbfab9ef3fa98f6a715768f", + "component_budget": { + "node_count": 514, + "item_count": 514, + "soft_pair_count": 1891, + "serialized_chars": 980895, + "estimated_input_tokens": 245224, + "estimated_eval_tokens": 490704 + }, + "budget": { + "node_count": 48, + "item_count": 48, + "soft_pair_count": 30, + "serialized_chars": 36734, + "estimated_input_tokens": 9184, + "estimated_eval_tokens": 18624 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": true, + "analyst_result": { + "summary": "The set is mostly a collection of unrelated standalone issues rather than one duplicate cluster. I would not merge any of the soft pairs; no PRs are present.", + "confidence": 0.97, + "canonical_issue_reason": null, + "canonical_pr_reason": null, + "best_issue_reason": "issue:45812 is the strongest standalone bug report here: it is concrete, broad in impact, and describes a clear v4\u2192v5 tokenizer regression affecting all Granite models.", + "best_pr_reason": null, + "soft_edge_verdicts": [ + { + "left": "issue:43525", + "right": "issue:43572", + "accept": false, + "reason": "Both mention missing config token fields after v5, but they affect different models and different fields (`pad_token_id` vs `pad_token_idx`), so this does not look like the same bug." + }, + { + "left": "issue:43643", + "right": "issue:44561", + "accept": false, + "reason": "Both involve `trust_remote_code`, but one is about `AutoConfig` returning missing fields while the other is about removal of `is_torch_fx_available`; these are different regressions and code paths." + }, + { + "left": "issue:43720", + "right": "issue:44361", + "accept": false, + "reason": "Completely different subsystems: BitNet loading vs MLukeTokenizer task AttributeError." + }, + { + "left": "issue:44964", + "right": "issue:45405", + "accept": false, + "reason": "A model-loading bug versus a dependency-version bump issue; no shared underlying defect." + }, + { + "left": "issue:44625", + "right": "issue:45584", + "accept": false, + "reason": "Different models and failures: `num_labels` propagation vs Whisper empty-transcription generation." + }, + { + "left": "issue:43493", + "right": "issue:43825", + "accept": false, + "reason": "Both are around transformers behavior, but one is SigLIP2 correctness versus a pipeline error message regression; not the same underlying bug." + }, + { + "left": "issue:44977", + "right": "issue:46209", + "accept": false, + "reason": "Unrelated failures: Qwen3.5 flash-attention generation vs FP8 inference on B200." + }, + { + "left": "issue:44554", + "right": "issue:46365", + "accept": false, + "reason": "Different areas entirely: MPS attention correctness vs import-time Trainer/datasets issue." + }, + { + "left": "issue:28218", + "right": "issue:33357", + "accept": false, + "reason": "Tokenizer spacing bug versus a MacOS bus error in CLIP loading; no common code-path bug." + }, + { + "left": "issue:43452", + "right": "issue:44451", + "accept": false, + "reason": "Both are model/tokenizer loading complaints, but one is about `gguf_file` handling and the other is a specific BERT model load failure; not the same defect." + }, + { + "left": "issue:45823", + "right": "issue:46050", + "accept": false, + "reason": "Device mismatch in Gemma4 forward pass versus broken quantized model save/load; unrelated." + }, + { + "left": "issue:39692", + "right": "issue:43295", + "accept": false, + "reason": "Documentation example errors versus a regression in custom model/processor code handling." + }, + { + "left": "issue:42796", + "right": "issue:44977", + "accept": false, + "reason": "Tokenizer BPE-marker handling versus flash-attention generation; different bugs." + }, + { + "left": "issue:43994", + "right": "issue:46192", + "accept": false, + "reason": "SigLIP2 inference outputs versus `return_tensors` being ignored; unrelated API bugs." + }, + { + "left": "issue:43493", + "right": "issue:43844", + "accept": false, + "reason": "Model implementation discrepancy versus a DeepSpeed gradient increase issue; different failures." + }, + { + "left": "issue:43278", + "right": "issue:43844", + "accept": false, + "reason": "Embedding dtype change during evaluate is unrelated to the ZeRO-3 gradient regression." + }, + { + "left": "issue:43278", + "right": "issue:43701", + "accept": false, + "reason": "Different problems: dtype mismatch versus checkpoint resume key mismatch." + }, + { + "left": "issue:43646", + "right": "issue:44466", + "accept": false, + "reason": "Custom model initialization breakage versus inconsistent serialization of tied weights; not the same bug." + }, + { + "left": "issue:46209", + "right": "issue:46710", + "accept": false, + "reason": "FP8 inference failure on B200 and an accuracy regression after upgrading Transformers are different issues." + }, + { + "left": "issue:44568", + "right": "issue:46489", + "accept": false, + "reason": "Both concern tokenizer behavior in v5+, but one is special-token insertion for mDeBERTa and the other is incorrect output for DeepSeek-Coder; too different to treat as one bug." + }, + { + "left": "issue:33357", + "right": "issue:33666", + "accept": false, + "reason": "A runtime bus error on MacOS versus a multi-GPU training question; unrelated." + }, + { + "left": "issue:43927", + "right": "issue:44361", + "accept": false, + "reason": "DiaConfig save/load token ID loss is unrelated to MLukeTokenizer task AttributeError." + }, + { + "left": "issue:44849", + "right": "issue:46209", + "accept": false, + "reason": "Qwen3.5 hidden-state behavior versus FP8 inference failure; different code paths and symptoms." + }, + { + "left": "issue:44568", + "right": "issue:45812", + "accept": false, + "reason": "Both are tokenizer regressions, but one is about missing BOS/EOS insertion and the other is wrong token IDs for Granite models; they are not the same underlying bug." + }, + { + "left": "issue:44792", + "right": "issue:45405", + "accept": false, + "reason": "Janus image-generation test failure is unrelated to the PEFT version bump." + }, + { + "left": "issue:43575", + "right": "issue:45405", + "accept": false, + "reason": "OOM when loading a large model with tensor parallelism is unrelated to a dependency version bump." + }, + { + "left": "issue:43572", + "right": "issue:44291", + "accept": false, + "reason": "Missing config field after update versus `init_empty_weights` rejecting an unexpected argument; unrelated regressions." + }, + { + "left": "issue:44488", + "right": "issue:44843", + "accept": false, + "reason": "A specific model load failure is not the same as an offline-mode failure caused by an unconditional `model_info()` call." + }, + { + "left": "issue:44589", + "right": "issue:45520", + "accept": false, + "reason": "Different import/storage problems: missing Float8 storage type versus a Python 3.13 `flash_attn` import-utils KeyError." + }, + { + "left": "issue:46821", + "right": "issue:46832", + "accept": false, + "reason": "TimesFM `window_size` AttributeError and RT-DETR feature-level mismatch are unrelated model 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": 27, + "review_unit_index": 28, + "review_unit_count": 63, + "status": "reviewed", + "reason": null, + "source": "llm", + "cache_hit": false, + "model": "gpt-5.4-mini?service_tier=flex", + "cluster_id": "cluster-10000-514", + "nodes": [ + "issue:28218", + "issue:29159", + "issue:29942", + "issue:30333", + "issue:32090", + "issue:34567", + "issue:42490", + "issue:43257", + "issue:43299", + "issue:43334", + "issue:43504", + "issue:43572", + "issue:43606", + "issue:43653", + "issue:43696", + "issue:43824", + "issue:43844", + "issue:43873", + "issue:43937", + "issue:43994", + "issue:44112", + "issue:44464", + "issue:44493", + "issue:44589", + "issue:44849", + "issue:44869", + "issue:44964", + "issue:45072", + "issue:45084", + "issue:45200", + "issue:45229", + "issue:45245", + "issue:45357", + "issue:45520", + "issue:45584", + "issue:45588", + "issue:45759", + "issue:45823", + "issue:46050", + "issue:46285", + "issue:46399", + "issue:46701", + "issue:46710", + "issue:46752", + "issue:46832", + "issue:46846", + "issue:47145" + ], + "soft_pairs": [ + "issue:43334|issue:45759", + "issue:43824|issue:44589", + "issue:44849|issue:45357", + "issue:43504|issue:43606", + "issue:43334|issue:43572", + "issue:45245|issue:46710", + "issue:43257|issue:43299", + "issue:44493|issue:46285", + "issue:43504|issue:43653", + "issue:43606|issue:44112", + "issue:45072|issue:46832", + "issue:44849|issue:47145", + "issue:46399|issue:46846", + "issue:44464|issue:45584", + "issue:43844|issue:44964", + "issue:43873|issue:46050", + "issue:43994|issue:44493", + "issue:29159|issue:32090", + "issue:44869|issue:46752", + "issue:43937|issue:46701", + "issue:45200|issue:45229", + "issue:42490|issue:43696", + "issue:45229|issue:45823", + "issue:45084|issue:45520", + "issue:45588|issue:46050", + "issue:29942|issue:30333", + "issue:28218|issue:34567" + ], + "prepared_review_unit_hash": "ac34198e6274b1e865f2ec5fd1c15ceccce3043f58e80b3ea5c4993d8bd66e50", + "component_budget": { + "node_count": 514, + "item_count": 514, + "soft_pair_count": 1891, + "serialized_chars": 980895, + "estimated_input_tokens": 245224, + "estimated_eval_tokens": 490704 + }, + "budget": { + "node_count": 47, + "item_count": 47, + "soft_pair_count": 27, + "serialized_chars": 35288, + "estimated_input_tokens": 8822, + "estimated_eval_tokens": 17900 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": true, + "analyst_result": { + "summary": "This cluster is mostly a loose semantic collection of unrelated bug reports. The soft pairs share broad themes like Qwen/Gemma/model-loading or tokenizer issues, but not the same concrete failure mode, so none should be merged as duplicates.", + "confidence": 0.62, + "canonical_issue_reason": "issue:43334 is a concrete, reproducible loading regression with a specific stack trace and model family; among the set it is the clearest actionable bug report.", + "canonical_pr_reason": null, + "best_issue_reason": "issue:43334 has the cleanest failure signal and most specific diagnosis, making it the strongest representative issue in this set.", + "best_pr_reason": null, + "soft_edge_verdicts": [ + { + "left": "issue:43334", + "right": "issue:45759", + "accept": false, + "reason": "Both involve Qwen multimodal loading, but one is a pad_token_id attribute error and the other is AutoModelForCausalLM config unwrapping; different code paths and fixes." + }, + { + "left": "issue:43824", + "right": "issue:44589", + "accept": false, + "reason": "Importing a missing class and a Float8 storage lookup failure are unrelated bugs." + }, + { + "left": "issue:44849", + "right": "issue:45357", + "accept": false, + "reason": "Both are Qwen3.5 regressions, but one affects output_hidden_states and the other save_pretrained visual encoder keys; not the same issue." + }, + { + "left": "issue:43504", + "right": "issue:43606", + "accept": false, + "reason": "Different models and failure modes: legacy field loading vs CPU offload device mismatch." + }, + { + "left": "issue:43334", + "right": "issue:43572", + "accept": false, + "reason": "Both mention missing token id fields after a version update, but they affect different configs and model families." + }, + { + "left": "issue:45245", + "right": "issue:46710", + "accept": false, + "reason": "A category-count runtime limit and an accuracy regression in a different model are unrelated." + }, + { + "left": "issue:43257", + "right": "issue:43299", + "accept": false, + "reason": "Both concern Qwen MOE models, but one is weight conversion with accelerate/deepspeed and the other is a loading breakage in dev; different concrete problems." + }, + { + "left": "issue:44493", + "right": "issue:46285", + "accept": false, + "reason": "Unexpected position-id keys and CLIP text-model structure changes are unrelated bugs." + }, + { + "left": "issue:43504", + "right": "issue:43653", + "accept": false, + "reason": "Legacy field loading for BEiT is unrelated to BigBirdTokenizer special-token registration." + }, + { + "left": "issue:43606", + "right": "issue:44112", + "accept": false, + "reason": "A real inference device mismatch and a stale CI test failure are not the same issue." + }, + { + "left": "issue:45072", + "right": "issue:46832", + "accept": false, + "reason": "Different model subsystems and failures: dtype mismatch vs RTDetr feature-level indexing." + }, + { + "left": "issue:44849", + "right": "issue:47145", + "accept": false, + "reason": "Qwen3.5 hidden-state handling and FP8 weight_scale application are unrelated." + }, + { + "left": "issue:46399", + "right": "issue:46846", + "accept": false, + "reason": "Both are Qwen3.5-related, but one is kernel registration and the other is tensor-parallel planning; not mergeable as one fix." + }, + { + "left": "issue:44464", + "right": "issue:45584", + "accept": false, + "reason": "Compiled-forward chunked generation mismatch and Whisper empty-transcription decoding are different code paths." + }, + { + "left": "issue:43844", + "right": "issue:44964", + "accept": false, + "reason": "ZeRO-3 gradient behavior and Phi-4 multimodal loading are unrelated." + }, + { + "left": "issue:43873", + "right": "issue:46050", + "accept": false, + "reason": "Quantization/offloading broadly overlap, but one is runtime offload behavior and the other is save/load corruption after sinq quantization." + }, + { + "left": "issue:43994", + "right": "issue:44493", + "accept": false, + "reason": "SigLIP2 nonsensical outputs and unexpected position-id keys are different bugs." + }, + { + "left": "issue:29159", + "right": "issue:32090", + "accept": false, + "reason": "Tokenizer fast/slow inconsistency and NoneType GPU broadcast are unrelated." + }, + { + "left": "issue:44869", + "right": "issue:46752", + "accept": false, + "reason": "Both are edge-case crashes, but in different APIs and failure mechanisms." + }, + { + "left": "issue:43937", + "right": "issue:46701", + "accept": false, + "reason": "GLM generation config validation and DiffusionGemma pipeline failure are unrelated." + }, + { + "left": "issue:45200", + "right": "issue:45229", + "accept": false, + "reason": "Same model family, but text-only fine-tuning defaults and multi-GPU OOM are different issues." + }, + { + "left": "issue:42490", + "right": "issue:43696", + "accept": false, + "reason": "A path-format import warning is unrelated to CUDA OOM." + }, + { + "left": "issue:45229", + "right": "issue:45823", + "accept": false, + "reason": "Both are Gemma4 issues, but one is OOM and the other is a device-map forward mismatch; not the same bug." + }, + { + "left": "issue:45084", + "right": "issue:45520", + "accept": false, + "reason": "Compile-time non-template-node error and a Python 3.13 import-utils KeyError are unrelated." + }, + { + "left": "issue:45588", + "right": "issue:46050", + "accept": false, + "reason": "Flash-attention crash on sink-less models and sinq checkpoint save/load corruption are different code paths." + }, + { + "left": "issue:29942", + "right": "issue:30333", + "accept": false, + "reason": "Flash Attention test failures and MLflow reporting after job failure are unrelated." + }, + { + "left": "issue:28218", + "right": "issue:34567", + "accept": false, + "reason": "Tokenizer spacing behavior and TrainerState token-count tracking are different concerns." + } + ] + }, + "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": 27, + "review_unit_index": 29, + "review_unit_count": 63, + "status": "reviewed", + "reason": null, + "source": "llm", + "cache_hit": false, + "model": "gpt-5.4-mini?service_tier=flex", + "cluster_id": "cluster-10000-514", + "nodes": [ + "issue:15354", + "issue:29870", + "issue:29942", + "issue:37428", + "issue:38175", + "issue:39692", + "issue:41720", + "issue:42490", + "issue:42915", + "issue:43065", + "issue:43295", + "issue:43450", + "issue:43493", + "issue:43540", + "issue:43550", + "issue:43572", + "issue:43575", + "issue:43701", + "issue:43824", + "issue:43844", + "issue:43873", + "issue:43957", + "issue:44186", + "issue:44423", + "issue:44521", + "issue:44554", + "issue:44610", + "issue:44734", + "issue:44977", + "issue:44987", + "issue:45072", + "issue:45198", + "issue:45200", + "issue:45290", + "issue:45406", + "issue:45446", + "issue:45520", + "issue:45588", + "issue:45759", + "issue:45823", + "issue:46121", + "issue:46178", + "issue:46192", + "issue:46369", + "issue:46701", + "issue:46710", + "issue:46821" + ], + "soft_pairs": [ + "issue:43065|issue:46369", + "issue:15354|issue:38175", + "issue:44554|issue:45446", + "issue:43493|issue:43701", + "issue:41720|issue:46710", + "issue:44977|issue:46701", + "issue:44423|issue:44734", + "issue:37428|issue:43957", + "issue:43295|issue:43450", + "issue:44977|issue:46710", + "issue:45072|issue:45198", + "issue:45406|issue:46121", + "issue:45588|issue:45759", + "issue:43540|issue:45200", + "issue:43572|issue:46821", + "issue:43295|issue:43873", + "issue:43295|issue:44610", + "issue:44186|issue:46192", + "issue:45823|issue:46192", + "issue:43550|issue:43844", + "issue:43572|issue:44987", + "issue:43450|issue:43575", + "issue:29870|issue:29942", + "issue:45520|issue:46178", + "issue:37428|issue:43824", + "issue:45200|issue:45823", + "issue:42490|issue:43493", + "issue:39692|issue:46192", + "issue:44521|issue:45290", + "issue:43295|issue:43493", + "issue:42915|issue:43824" + ], + "prepared_review_unit_hash": "478f9f7fc72c594908b058e90e7d3c8c61f02d85e1edb62887cc8a4587f3d1bf", + "component_budget": { + "node_count": 514, + "item_count": 514, + "soft_pair_count": 1891, + "serialized_chars": 980895, + "estimated_input_tokens": 245224, + "estimated_eval_tokens": 490704 + }, + "budget": { + "node_count": 47, + "item_count": 47, + "soft_pair_count": 31, + "serialized_chars": 36583, + "estimated_input_tokens": 9146, + "estimated_eval_tokens": 18548 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": true, + "analyst_result": { + "summary": "This cluster is a loose similarity set across many unrelated Transformers issues (multimodal processors, flash-attention, serving, quantization, configs, and model-specific regressions). None of the soft pairs look like true duplicates of the same underlying bug/change.", + "confidence": 0.93, + "canonical_issue_reason": "Issue 43295 is the broadest representative anchor: a recent regression in processor/tokenizer handling that affects custom multimodal code. The rest of the cluster is too heterogeneous for a single true duplicate canonical.", + "canonical_pr_reason": "No pull requests are present in this cluster.", + "best_issue_reason": "Issue 43295 is the best global issue anchor because it describes a concrete regression with broad multimodal impact and is the most generally representative of the set.", + "best_pr_reason": "No pull requests are present in this cluster.", + "soft_edge_verdicts": [ + { + "left": "issue:43065", + "right": "issue:46369", + "accept": false, + "reason": "Same model family (Sam3), but different bugs: one is a dummy Conv2d implementation detail, the other is a memory overflow in the litetext model." + }, + { + "left": "issue:15354", + "right": "issue:38175", + "accept": false, + "reason": "Both involve vision models, but one is a torch.jit.script export limitation and the other is an inference/probability issue in SigLIP2." + }, + { + "left": "issue:44554", + "right": "issue:45446", + "accept": false, + "reason": "Different code paths: MPS attention correctness vs a PyTorch version-check/import bug in flex_attention." + }, + { + "left": "issue:43493", + "right": "issue:43701", + "accept": false, + "reason": "Unrelated symptoms: SigLIP2 implementation discrepancy vs resume-from-checkpoint key mismatch." + }, + { + "left": "issue:41720", + "right": "issue:46710", + "accept": false, + "reason": "Different models and failures: Qwen3 auto device-mapping CUDA assert vs an accuracy regression in DeepSeek-R1-Distill-Llama-8B." + }, + { + "left": "issue:44977", + "right": "issue:46701", + "accept": false, + "reason": "Both are generation-related, but one is a flash-attention bug in Qwen3.5 and the other is an image-text-to-text pipeline failure in DiffusionGemma." + }, + { + "left": "issue:44423", + "right": "issue:44734", + "accept": false, + "reason": "Both hit serve, but the failures are distinct: multimodal input conversion on strings vs KV-cache continuation tensor indexing." + }, + { + "left": "issue:37428", + "right": "issue:43957", + "accept": false, + "reason": "Import-time flash-attention symbol error vs model loading breakage with torch.device(\"meta\"); different root causes." + }, + { + "left": "issue:43295", + "right": "issue:43450", + "accept": false, + "reason": "Both are multimodal processor issues, but one is a processor/tokenizer regression and the other is a batched video-shape bug." + }, + { + "left": "issue:44977", + "right": "issue:46710", + "accept": false, + "reason": "Different bugs: flash-attention generation failure in Qwen3.5 vs a model accuracy regression after version upgrade." + }, + { + "left": "issue:45072", + "right": "issue:45198", + "accept": false, + "reason": "One is a bfloat16 dtype mismatch in inference, the other is save_pretrained/tokenization failure; not the same bug." + }, + { + "left": "issue:45406", + "right": "issue:46121", + "accept": false, + "reason": "Gemma4Processor _tokenizer attribute crash and rope-parameter JSON list typing are unrelated." + }, + { + "left": "issue:45588", + "right": "issue:45759", + "accept": false, + "reason": "Different paths: flash-attention sink-less model crash vs AutoModelForCausalLM.from_config not unwrapping text_config." + }, + { + "left": "issue:43540", + "right": "issue:45200", + "accept": false, + "reason": "Both are multimodal-model issues, but one is a video input processing ValueError and the other is a Gemma4 fine-tuning token-ID defaulting problem." + }, + { + "left": "issue:43572", + "right": "issue:46821", + "accept": false, + "reason": "Different config/attribute failures in unrelated model families (StableLm vs TimesFM)." + }, + { + "left": "issue:43295", + "right": "issue:43873", + "accept": false, + "reason": "Processor/tokenizer regression vs offloading/quantization behavior; they do not share the same code path." + }, + { + "left": "issue:43295", + "right": "issue:44610", + "accept": false, + "reason": "Both involve processors, but one is a tokenizer access regression and the other is a wrong image size output in OmDet-Turbo." + }, + { + "left": "issue:44186", + "right": "issue:46192", + "accept": false, + "reason": "LayoutLMv2 tokenizer crashes on NER/padding are not the same as return_tensors being ignored when text_kwargs is passed." + }, + { + "left": "issue:45823", + "right": "issue:46192", + "accept": false, + "reason": "Gemma4 device mismatch during forward is unrelated to return_tensors being ignored in multimodal preprocessing." + }, + { + "left": "issue:43550", + "right": "issue:43844", + "accept": false, + "reason": "torch.compile+SDPA failure and ZeRO-3 gradient growth are distinct training/runtime issues." + }, + { + "left": "issue:43572", + "right": "issue:44987", + "accept": false, + "reason": "StableLm missing config field and loading physical-intelligence/fast on transformers>=5.1.0 are unrelated regression types." + }, + { + "left": "issue:43450", + "right": "issue:43575", + "accept": false, + "reason": "Video processor shape bug vs tensor-parallel loading OOM; different subsystems and symptoms." + }, + { + "left": "issue:29870", + "right": "issue:29942", + "accept": false, + "reason": "Bart generate unbound-local error and Flash Attention 2 test failures are unrelated." + }, + { + "left": "issue:45520", + "right": "issue:46178", + "accept": false, + "reason": "Both are import/backend related, but one is a flash_attn KeyError in import_utils and the other is a missing BACKENDS_MAPPING entry for tensorflow_text." + }, + { + "left": "issue:37428", + "right": "issue:43824", + "accept": false, + "reason": "Flash-attention utility import failure and missing Qwen2_5_VL export are unrelated import errors." + }, + { + "left": "issue:45200", + "right": "issue:45823", + "accept": false, + "reason": "Gemma4 mm_token_type_ids defaulting for text-only fine-tuning is a different bug from Gemma4 PLE device mismatch under device_map=auto." + }, + { + "left": "issue:39692", + "right": "issue:46192", + "accept": false, + "reason": "SigLIP2 documentation example errors are unrelated to return_tensors being silently ignored." + }, + { + "left": "issue:44521", + "right": "issue:45290", + "accept": false, + "reason": "Both touch apply_chat_template, but one is assistant-mask generation for multimodal inputs while the other is a crash on tool-call assistant messages with no content." + }, + { + "left": "issue:43295", + "right": "issue:43493", + "accept": false, + "reason": "Processor/tokenizer regression and SigLIP2 HF-vs-JAX discrepancy are not the same underlying bug." + }, + { + "left": "issue:42915", + "right": "issue:43824", + "accept": false, + "reason": "Qwen3Moe FineGrainedFP8Config failure and Qwen2_5_VL import 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": 27, + "review_unit_index": 30, + "review_unit_count": 63, + "status": "reviewed", + "reason": null, + "source": "llm", + "cache_hit": false, + "model": "gpt-5.4-mini?service_tier=flex", + "cluster_id": "cluster-10000-514", + "nodes": [ + "issue:28218", + "issue:35141", + "issue:41720", + "issue:43262", + "issue:43278", + "issue:43334", + "issue:43335", + "issue:43388", + "issue:43450", + "issue:43452", + "issue:43525", + "issue:43526", + "issue:43550", + "issue:43606", + "issue:43653", + "issue:43819", + "issue:43824", + "issue:43827", + "issue:43856", + "issue:43901", + "issue:43950", + "issue:43975", + "issue:44112", + "issue:44442", + "issue:44466", + "issue:44561", + "issue:44610", + "issue:44737", + "issue:44792", + "issue:44857", + "issue:44898", + "issue:45083", + "issue:45127", + "issue:45198", + "issue:45200", + "issue:45325", + "issue:45381", + "issue:45412", + "issue:45563", + "issue:45593", + "issue:46192", + "issue:46209", + "issue:46234", + "issue:46285", + "issue:46512", + "issue:46894", + "issue:46899", + "issue:46932" + ], + "soft_pairs": [ + "issue:45381|issue:46209", + "issue:28218|issue:35141", + "issue:44466|issue:45127", + "issue:44737|issue:46894", + "issue:43278|issue:43856", + "issue:43653|issue:44112", + "issue:43452|issue:46512", + "issue:43824|issue:43975", + "issue:43262|issue:43450", + "issue:43950|issue:46285", + "issue:43334|issue:43525", + "issue:44442|issue:45198", + "issue:43526|issue:46192", + "issue:45200|issue:46899", + "issue:44898|issue:45412", + "issue:45200|issue:46932", + "issue:44792|issue:45563", + "issue:43819|issue:44112", + "issue:45200|issue:45325", + "issue:43335|issue:43550", + "issue:41720|issue:46209", + "issue:44561|issue:46234", + "issue:43827|issue:43901", + "issue:44610|issue:44857", + "issue:43819|issue:45083", + "issue:43388|issue:43526", + "issue:44857|issue:45593", + "issue:43606|issue:45083" + ], + "prepared_review_unit_hash": "f14e2af978cb6715a846377477108e0cd41050b7e2135a3f08792b47f1944a27", + "component_budget": { + "node_count": 514, + "item_count": 514, + "soft_pair_count": 1891, + "serialized_chars": 980895, + "estimated_input_tokens": 245224, + "estimated_eval_tokens": 490704 + }, + "budget": { + "node_count": 48, + "item_count": 48, + "soft_pair_count": 28, + "serialized_chars": 36570, + "estimated_input_tokens": 9143, + "estimated_eval_tokens": 18542 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": true, + "analyst_result": { + "summary": "This cluster is mostly heterogeneous and does not appear to contain broader duplicates. I found one strong duplicate-like pair around Gemma4 quantized multimodal embedders; the rest are separate bugs/issues. No PRs are present.", + "confidence": 0.89, + "canonical_issue_reason": "issue:46899 is the clearest representative of the only true duplicate-like problem in the set: Gemma4 quantized audio/vision embedders failing due to unsafe dtype casting. issue:46932 reads like a closely related follow-up/fix note rather than a separate underlying bug.", + "canonical_pr_reason": null, + "best_issue_reason": "issue:46899 is the best global issue anchor because it cleanly captures the shared Gemma4 quantized-embedder dtype-cast defect; issue:46932 is effectively the nearest duplicate candidate.", + "best_pr_reason": null, + "soft_edge_verdicts": [ + { + "left": "issue:45381", + "right": "issue:46209", + "accept": false, + "reason": "Different failures: Qwen2.5-VL vision position IDs vs an FP8 inference failure on B200." + }, + { + "left": "issue:28218", + "right": "issue:35141", + "accept": false, + "reason": "Tokenizer spacing after added tokens is unrelated to embedding reinitialization after resize." + }, + { + "left": "issue:44466", + "right": "issue:45127", + "accept": false, + "reason": "Both involve tied-weight behavior, but one is serialization/device-dependent while the other is LoRA merge collapse with extended vocab; not the same bug." + }, + { + "left": "issue:44737", + "right": "issue:46894", + "accept": false, + "reason": "XLNet CPU tensor creation and CohereASR double-shift loss are unrelated code paths." + }, + { + "left": "issue:43278", + "right": "issue:43856", + "accept": false, + "reason": "Different problems: eval-time dtype drift vs Qwen3 MoE memory usage." + }, + { + "left": "issue:43653", + "right": "issue:44112", + "accept": false, + "reason": "Tokenizer special-token registration bug is unrelated to a GraniteSpeech CI test failure." + }, + { + "left": "issue:43452", + "right": "issue:46512", + "accept": false, + "reason": "GGUF loading failures and stale config dtype after from_config are different issues." + }, + { + "left": "issue:43824", + "right": "issue:43975", + "accept": false, + "reason": "Import error for a Qwen2.5-VL class is unrelated to DeepSeek detokenization in v5." + }, + { + "left": "issue:43262", + "right": "issue:43450", + "accept": false, + "reason": "Audio chat-template sampling-rate default and batched video shape handling are different processor bugs." + }, + { + "left": "issue:43950", + "right": "issue:46285", + "accept": false, + "reason": "Non-persistent buffer corruption is unrelated to CLIP class structure inconsistencies." + }, + { + "left": "issue:43334", + "right": "issue:43525", + "accept": false, + "reason": "Same symptom class (missing pad_token_id), but different model configs and likely different fixes." + }, + { + "left": "issue:44442", + "right": "issue:45198", + "accept": false, + "reason": "AutoTokenizer loading failure and Wav2Vec2 save/tokenization failure are separate tokenizer issues." + }, + { + "left": "issue:43526", + "right": "issue:46192", + "accept": false, + "reason": "Label reduction in BeitImageProcessorFast is unrelated to return_tensors being ignored with text_kwargs." + }, + { + "left": "issue:45200", + "right": "issue:46899", + "accept": false, + "reason": "Gemma4 token-type-id defaults and Gemma4 quantized embedder dtype casting are different bugs." + }, + { + "left": "issue:44898", + "right": "issue:45412", + "accept": false, + "reason": "Perceiver interpolation failure and RT-DETR memory leak are unrelated." + }, + { + "left": "issue:45200", + "right": "issue:46932", + "accept": false, + "reason": "Different Gemma4 problems: default mm_token_type_ids vs quantized vision/audio dtype casting." + }, + { + "left": "issue:44792", + "right": "issue:45563", + "accept": false, + "reason": "Janus image-generation test failure is unrelated to a stale generate() warning." + }, + { + "left": "issue:43819", + "right": "issue:44112", + "accept": false, + "reason": "DAC latent mismatch and GraniteSpeech CI device override are unrelated." + }, + { + "left": "issue:45200", + "right": "issue:45325", + "accept": false, + "reason": "Token-type-id defaults and Qwen2.5-VL rope-index scaling are different model bugs." + }, + { + "left": "issue:43335", + "right": "issue:43550", + "accept": false, + "reason": "Sparse-layer construction in SwitchTransformers is unrelated to Bamba torch.compile SDPA failure." + }, + { + "left": "issue:41720", + "right": "issue:46209", + "accept": false, + "reason": "Different hardware/model-specific inference failures; not the same underlying bug." + }, + { + "left": "issue:44561", + "right": "issue:46234", + "accept": false, + "reason": "Removal of a utility function breaking trust_remote_code is unrelated to a missing import_utils symbol." + }, + { + "left": "issue:43827", + "right": "issue:43901", + "accept": false, + "reason": "Both are docs regressions, but they reference different API changes and different pages." + }, + { + "left": "issue:44610", + "right": "issue:44857", + "accept": false, + "reason": "Processor input-size mismatch and AMP CUDA crash are separate issues." + }, + { + "left": "issue:43819", + "right": "issue:45083", + "accept": false, + "reason": "DAC forward/from_latents mismatch and qwen3_omni_moe helper behavior are unrelated." + }, + { + "left": "issue:43388", + "right": "issue:43526", + "accept": false, + "reason": "Both concern labels, but one is gather_for_metrics truncation and the other is BeitImageProcessorFast label reduction." + }, + { + "left": "issue:44857", + "right": "issue:45593", + "accept": false, + "reason": "Float16 AMP crash and missing auxiliary losses are different D-FINE issues." + }, + { + "left": "issue:43606", + "right": "issue:45083", + "accept": false, + "reason": "CPU-offload device mismatch and qwen3_omni_moe output-length helper behavior 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": 27, + "review_unit_index": 31, + "review_unit_count": 63, + "status": "reviewed", + "reason": null, + "source": "llm", + "cache_hit": false, + "model": "gpt-5.4-mini?service_tier=flex", + "cluster_id": "cluster-10000-514", + "nodes": [ + "issue:28282", + "issue:29127", + "issue:29942", + "issue:32090", + "issue:33357", + "issue:34689", + "issue:35141", + "issue:37428", + "issue:38175", + "issue:39692", + "issue:42490", + "issue:43278", + "issue:43335", + "issue:43425", + "issue:43504", + "issue:43526", + "issue:43550", + "issue:43611", + "issue:43653", + "issue:43819", + "issue:43825", + "issue:43854", + "issue:44112", + "issue:44206", + "issue:44351", + "issue:44448", + "issue:44466", + "issue:44493", + "issue:44514", + "issue:44521", + "issue:44743", + "issue:44855", + "issue:45083", + "issue:45290", + "issue:45412", + "issue:45520", + "issue:45584", + "issue:45588", + "issue:45593", + "issue:45759", + "issue:45859", + "issue:46170", + "issue:46285", + "issue:46326", + "issue:46424", + "issue:46512", + "issue:46821", + "issue:46858" + ], + "soft_pairs": [ + "issue:37428|issue:43425", + "issue:46512|issue:46821", + "issue:44206|issue:45584", + "issue:43335|issue:43653", + "issue:45759|issue:46285", + "issue:38175|issue:45412", + "issue:46424|issue:46858", + "issue:29942|issue:32090", + "issue:28282|issue:33357", + "issue:43611|issue:44466", + "issue:43335|issue:43504", + "issue:44855|issue:45083", + "issue:43550|issue:44112", + "issue:45520|issue:46170", + "issue:37428|issue:44351", + "issue:44448|issue:44743", + "issue:43335|issue:45083", + "issue:45290|issue:46326", + "issue:43335|issue:44112", + "issue:43335|issue:45593", + "issue:43278|issue:43825", + "issue:44514|issue:44521", + "issue:34689|issue:35141", + "issue:42490|issue:43854", + "issue:45588|issue:45859", + "issue:29127|issue:32090", + "issue:39692|issue:43526", + "issue:43819|issue:44493" + ], + "prepared_review_unit_hash": "65794a7ec90cb34f2f42a15796433cd7bc6aec92415937a39f6904842d34dea0", + "component_budget": { + "node_count": 514, + "item_count": 514, + "soft_pair_count": 1891, + "serialized_chars": 980895, + "estimated_input_tokens": 245224, + "estimated_eval_tokens": 490704 + }, + "budget": { + "node_count": 48, + "item_count": 48, + "soft_pair_count": 28, + "serialized_chars": 36519, + "estimated_input_tokens": 9130, + "estimated_eval_tokens": 18516 + }, + "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 Transformers themes but not the same underlying bug. I would not merge any of the suggested pairs as duplicates.", + "confidence": 0.93, + "canonical_issue_reason": null, + "canonical_pr_reason": null, + "best_issue_reason": "issue:44521 is the strongest standalone representative: it has active discussion, inbound references, and a clear reproducible failure in a common API (`apply_chat_template`) affecting multimodal inputs.", + "best_pr_reason": null, + "soft_edge_verdicts": [ + { + "left": "issue:37428", + "right": "issue:43425", + "accept": false, + "reason": "Both are import/version problems, but one is a missing internal flash-attention symbol and the other is Torch 2.10 incompatibility; different root causes and fixes." + }, + { + "left": "issue:46512", + "right": "issue:46821", + "accept": false, + "reason": "`from_config(dtype=...)` leaving `model.config.dtype` stale is unrelated to TimesFM `window_size` raising `AttributeError`; different APIs and failure modes." + }, + { + "left": "issue:44206", + "right": "issue:45584", + "accept": false, + "reason": "`LasrFeatureExtractor` rejecting an unsupported `center` arg is unrelated to Whisper empty-transcription behavior after `align_special_tokens`." + }, + { + "left": "issue:43335", + "right": "issue:43653", + "accept": false, + "reason": "SwitchTransformers sparse-layer creation and BigBirdTokenizer special-token registration are separate bugs in different components." + }, + { + "left": "issue:45759", + "right": "issue:46285", + "accept": false, + "reason": "Composite Qwen multimodal config unwrapping is not the same bug as CLIP text-model structure divergence." + }, + { + "left": "issue:38175", + "right": "issue:45412", + "accept": false, + "reason": "Unexpected zero probabilities in SigLIP2 inference is unrelated to RT-DETR memory not being released on deletion." + }, + { + "left": "issue:46424", + "right": "issue:46858", + "accept": false, + "reason": "Ignoring `max_cache_length` for quantized cache and returning a device tensor from `get_seq_length()` are different cache bugs." + }, + { + "left": "issue:29942", + "right": "issue:32090", + "accept": false, + "reason": "Failing Flash Attention 2 tests and a Trainer `_gpu_broadcast_one` `NoneType` error are unrelated code paths." + }, + { + "left": "issue:28282", + "right": "issue:33357", + "accept": false, + "reason": "Missing PyTorch import for `AutoModel` is unrelated to the MacOS bus error loading a community CLIP model." + }, + { + "left": "issue:43611", + "right": "issue:44466", + "accept": false, + "reason": "Both touch v5 model loading/serialization, but `base_model_prefix` loading failure and `lm_head.weight` serialization inconsistency are distinct bugs." + }, + { + "left": "issue:43335", + "right": "issue:43504", + "accept": false, + "reason": "SwitchTransformers sparse-layer creation with `num_sparse_encoder_layers=0` is unrelated to BEiT legacy-field load failure." + }, + { + "left": "issue:44855", + "right": "issue:45083", + "accept": false, + "reason": "Python 3.13 `torch.jit` parsing/indentation failure is unrelated to `_get_feat_extract_output_lengths` behavior in qwen3_omni_moe." + }, + { + "left": "issue:43550", + "right": "issue:44112", + "accept": false, + "reason": "`torch.compile`/SDPA failure in Bamba-9B-v2 is unrelated to the stale device-override CI test in GraniteSpeech." + }, + { + "left": "issue:45520", + "right": "issue:46170", + "accept": false, + "reason": "Python 3.13 `flash_attn` `KeyError` is a different issue from Windows ROCm import/type errors involving `torch.distributed`." + }, + { + "left": "issue:37428", + "right": "issue:44351", + "accept": false, + "reason": "Missing `_flash_supports_window_size` and missing `HybridCache` are different import/export regressions." + }, + { + "left": "issue:44448", + "right": "issue:44743", + "accept": false, + "reason": "Pegasus output differences across versions do not match the recurrent-state reset bug in modular_qwen3_5 cache handling." + }, + { + "left": "issue:43335", + "right": "issue:45083", + "accept": false, + "reason": "SwitchTransformers sparse-layer creation is unrelated to the qwen3_omni_moe helper-length bug." + }, + { + "left": "issue:45290", + "right": "issue:46326", + "accept": false, + "reason": "Both involve chat templating, but one crashes on assistant tool-call messages with no content while the other is a `continue_final_message` bug in a mistral_common tokenizer path." + }, + { + "left": "issue:43335", + "right": "issue:44112", + "accept": false, + "reason": "SwitchTransformers config bug and GraniteSpeech CI device-override test failure are unrelated." + }, + { + "left": "issue:43335", + "right": "issue:45593", + "accept": false, + "reason": "Sparse-layer construction in SwitchTransformers is unrelated to D-FINE losing auxiliary losses when denoising is off." + }, + { + "left": "issue:43278", + "right": "issue:43825", + "accept": false, + "reason": "A dtype mismatch in `from_config` is unrelated to the pipeline translation-task warning regression." + }, + { + "left": "issue:44514", + "right": "issue:44521", + "accept": false, + "reason": "Both are `apply_chat_template` issues, but one is a batched `padding=False` crash and the other is all-zero assistant masks for multimodal inputs; different bugs." + }, + { + "left": "issue:34689", + "right": "issue:35141", + "accept": false, + "reason": "Llama 3.2 Vision model-loading regression and output-embedding reinitialization after resizing token embeddings are different problems." + }, + { + "left": "issue:42490", + "right": "issue:43854", + "accept": false, + "reason": "Trailing-slash local-path import warning is unrelated to the GLM-4.7-Flash unit-test loading failure." + }, + { + "left": "issue:45588", + "right": "issue:45859", + "accept": false, + "reason": "`s_aux=None` crash in flash_attention integration and CPU-offload incompatibility in Qwen3_5MoeTextRotaryEmbedding are different code paths." + }, + { + "left": "issue:29127", + "right": "issue:32090", + "accept": false, + "reason": "LayoutLMv3 error-message clarity and Trainer GPU broadcast `NoneType` failure are unrelated." + }, + { + "left": "issue:39692", + "right": "issue:43526", + "accept": false, + "reason": "SigLIP2 documentation-example problems are not the same as the BeitImageProcessorFast `reduce_labels` bug." + }, + { + "left": "issue:43819", + "right": "issue:44493", + "accept": false, + "reason": "DAC forward/from_latents mismatch is unrelated to the unexpected position-id key issue in models loading." + } + ] + }, + "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": 27, + "review_unit_index": 32, + "review_unit_count": 63, + "status": "reviewed", + "reason": null, + "source": "llm", + "cache_hit": false, + "model": "gpt-5.4-mini?service_tier=flex", + "cluster_id": "cluster-10000-514", + "nodes": [ + "issue:10000", + "issue:15354", + "issue:26413", + "issue:27957", + "issue:28218", + "issue:29870", + "issue:29942", + "issue:31501", + "issue:32090", + "issue:33357", + "issue:34689", + "issue:35141", + "issue:38175", + "issue:39692", + "issue:42490", + "issue:43278", + "issue:43329", + "issue:43550", + "issue:43575", + "issue:43606", + "issue:43701", + "issue:43720", + "issue:44112", + "issue:44206", + "issue:44448", + "issue:44464", + "issue:44493", + "issue:44617", + "issue:44779", + "issue:44857", + "issue:45072", + "issue:45083", + "issue:45127", + "issue:45245", + "issue:45290", + "issue:45310", + "issue:45381", + "issue:45459", + "issue:45478", + "issue:45593", + "issue:45823", + "issue:46234", + "issue:46424", + "issue:46431", + "issue:46566", + "issue:46697", + "issue:47000" + ], + "soft_pairs": [ + "issue:43606|issue:44857", + "issue:46431|issue:46566", + "issue:43720|issue:44112", + "issue:28218|issue:32090", + "issue:45310|issue:45478", + "issue:39692|issue:44493", + "issue:44448|issue:44779", + "issue:43701|issue:45083", + "issue:44206|issue:45823", + "issue:26413|issue:31501", + "issue:44617|issue:46697", + "issue:45127|issue:45245", + "issue:45459|issue:46234", + "issue:32090|issue:33357", + "issue:42490|issue:43575", + "issue:44857|issue:45072", + "issue:45593|issue:47000", + "issue:43550|issue:45083", + "issue:44464|issue:46424", + "issue:27957|issue:34689", + "issue:38175|issue:44493", + "issue:29942|issue:35141", + "issue:29870|issue:32090", + "issue:45290|issue:45381", + "issue:43278|issue:43329", + "issue:10000|issue:15354" + ], + "prepared_review_unit_hash": "417a639bf6ccc9b47783fa8223db35ffa7a2c2dab5d6561f069e59e4136ec211", + "component_budget": { + "node_count": 514, + "item_count": 514, + "soft_pair_count": 1891, + "serialized_chars": 980895, + "estimated_input_tokens": 245224, + "estimated_eval_tokens": 490704 + }, + "budget": { + "node_count": 47, + "item_count": 47, + "soft_pair_count": 26, + "serialized_chars": 34960, + "estimated_input_tokens": 8740, + "estimated_eval_tokens": 17736 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": true, + "analyst_result": { + "summary": "This cluster is mostly a grab bag of unrelated issue reports. The only clear duplicate pair is the two identical Qwen3.5 MoE from_pretrained errors; everything else looks like distinct bugs or regressions.", + "confidence": 0.89, + "canonical_issue_reason": "issue:45310 is the earliest report of the only clear duplicate pair and has the stronger cluster anchor role for the identical Qwen3.5 MoE from_pretrained error.", + "canonical_pr_reason": null, + "best_issue_reason": "issue:45310 is the best issue representative because it is the clearest duplicate candidate and the earliest, most canonical report among the items.", + "best_pr_reason": null, + "soft_edge_verdicts": [ + { + "left": "issue:43606", + "right": "issue:44857", + "accept": false, + "reason": "Different failures and code paths: CPU offload device mismatch in bark-small vs float16 AMP crash in LwDetrImageLoss." + }, + { + "left": "issue:46431", + "right": "issue:46566", + "accept": false, + "reason": "One is missing chat template support; the other is a device_map/meta-device loading issue in DiffusionGemma." + }, + { + "left": "issue:43720", + "right": "issue:44112", + "accept": false, + "reason": "Both are CI/loading-related, but the root causes differ: unpacking packed weights vs a stale device override in GraniteSpeech." + }, + { + "left": "issue:28218", + "right": "issue:32090", + "accept": false, + "reason": "Tokenizer spacing after added tokens and Trainer _gpu_broadcast_one NoneType errors are unrelated bugs." + }, + { + "left": "issue:45310", + "right": "issue:45478", + "accept": true, + "reason": "Same title and same concrete failure: transformers>=5.4.0 breaking Qwen3.5 MoE from_pretrained." + }, + { + "left": "issue:39692", + "right": "issue:44493", + "accept": false, + "reason": "SigLIP2 docs/example errors do not match the broad unexpected position-id key regressions." + }, + { + "left": "issue:44448", + "right": "issue:44779", + "accept": false, + "reason": "Pegasus v4/v5 output differences and Deepseek tokenizer regressions are different subsystems and symptoms." + }, + { + "left": "issue:43701", + "right": "issue:45083", + "accept": false, + "reason": "resume_from_checkpoint key mismatch is a checkpoint/resume bug; _get_feat_extract_output_lengths in qwen3_omni_moe is unrelated." + }, + { + "left": "issue:44206", + "right": "issue:45823", + "accept": false, + "reason": "Unsupported center arg in LasrFeatureExtractor and Gemma4 PLE device mismatch are distinct issues." + }, + { + "left": "issue:26413", + "right": "issue:31501", + "accept": false, + "reason": "Both affect training start/resume, but the epoch_iterator empty error and FineWeb training startup failure are different underlying problems." + }, + { + "left": "issue:44617", + "right": "issue:46697", + "accept": false, + "reason": "CUDA OOM in Sam3Video and over-padding in Swin2SR image processing are unrelated." + }, + { + "left": "issue:45127", + "right": "issue:45245", + "accept": false, + "reason": "LoRA merge/model collapse and category-count RuntimeError are entirely different bugs." + }, + { + "left": "issue:45459", + "right": "issue:46234", + "accept": false, + "reason": "Tokenizer protobuf error masking and a missing import_utils symbol are unrelated maintenance bugs." + }, + { + "left": "issue:32090", + "right": "issue:33357", + "accept": false, + "reason": "Trainer NoneType broadcast failure and MacOS CLIP bus error are different failure modes." + }, + { + "left": "issue:42490", + "right": "issue:43575", + "accept": false, + "reason": "Local-path import warning and tensor-parallel OOM on Qwen2-57B are unrelated." + }, + { + "left": "issue:44857", + "right": "issue:45072", + "accept": false, + "reason": "Both mention dtype-ish failures, but one is an AMP crash in a loss and the other is a bfloat16 inference mismatch; not the same code-path issue." + }, + { + "left": "issue:45593", + "right": "issue:47000", + "accept": false, + "reason": "Different detection-loss bugs: denoising-off auxiliary loss omission vs RTDetr cost matrix infeasibility." + }, + { + "left": "issue:43550", + "right": "issue:45083", + "accept": false, + "reason": "torch.compile/SDPA failure in Bamba is unrelated to qwen3_omni_moe feature-length behavior." + }, + { + "left": "issue:44464", + "right": "issue:46424", + "accept": false, + "reason": "Chunked generation inconsistency under compiled forward is a different bug from cache max length only applying to Quantized Cache." + }, + { + "left": "issue:27957", + "right": "issue:34689", + "accept": false, + "reason": "XLMRoberta Flash Attention 2 support and Llama 3.2 Vision model loading regression are distinct issues." + }, + { + "left": "issue:38175", + "right": "issue:44493", + "accept": false, + "reason": "siglip2 zero probabilities and unexpected position-id keys are not the same failure." + }, + { + "left": "issue:29942", + "right": "issue:35141", + "accept": false, + "reason": "Failing FA2 tests and embedding reinitialization after resize are different bugs." + }, + { + "left": "issue:29870", + "right": "issue:32090", + "accept": false, + "reason": "Bart generate decoder_attention_mask bug and Trainer broadcast NoneType bug are unrelated." + }, + { + "left": "issue:45290", + "right": "issue:45381", + "accept": false, + "reason": "Chat template/tool-call crash and qwen2.5-vl video position-id issue are both multimodal-related but clearly different paths." + }, + { + "left": "issue:43278", + "right": "issue:43329", + "accept": false, + "reason": "Embedding dtype drift in evaluate and an undefined video-branch multimodal token helper are unrelated." + }, + { + "left": "issue:10000", + "right": "issue:15354", + "accept": false, + "reason": "German DistilBertModel issue and GeneratorExp/JIT export failure are unrelated model-specific problems." + } + ] + }, + "evaluator_result": { + "accept": true, + "feedback": "Grounded and conservative overall. The only accepted soft edge, issue:45310 \u2194 issue:45478, is well supported by the identical title and clear duplicate signal, and choosing issue:45310 as the canonical/earliest representative is reasonable. The remaining pairs are rejected cautiously with distinctions that track the packet titles, and the summary does not overclaim broader duplication beyond the one clear pair." + }, + "evaluator_used": true, + "retried": false, + "accepted_nontrivial_soft_edge": true, + "error_kind": null, + "error_message": null + }, + { + "label": "issue", + "component_index": 1, + "component_count": 27, + "review_unit_index": 33, + "review_unit_count": 63, + "status": "reviewed", + "reason": null, + "source": "llm", + "cache_hit": false, + "model": "gpt-5.4-mini?service_tier=flex", + "cluster_id": "cluster-10000-514", + "nodes": [ + "issue:25251", + "issue:29127", + "issue:29870", + "issue:32090", + "issue:33290", + "issue:33357", + "issue:34567", + "issue:37428", + "issue:39692", + "issue:40444", + "issue:41720", + "issue:42175", + "issue:42490", + "issue:43278", + "issue:43381", + "issue:43526", + "issue:43575", + "issue:43701", + "issue:43873", + "issue:43927", + "issue:43976", + "issue:44186", + "issue:44360", + "issue:44361", + "issue:44448", + "issue:44466", + "issue:44561", + "issue:44568", + "issue:44779", + "issue:44855", + "issue:44871", + "issue:44977", + "issue:45198", + "issue:45229", + "issue:45290", + "issue:45440", + "issue:45491", + "issue:45588", + "issue:45698", + "issue:45803", + "issue:45823", + "issue:45920", + "issue:46192", + "issue:46399", + "issue:46421", + "issue:46655", + "issue:46858" + ], + "soft_pairs": [ + "issue:37428|issue:42175", + "issue:45290|issue:46655", + "issue:44448|issue:44977", + "issue:29870|issue:33290", + "issue:44360|issue:44561", + "issue:43278|issue:43381", + "issue:44855|issue:46399", + "issue:40444|issue:43526", + "issue:45920|issue:46192", + "issue:32090|issue:33290", + "issue:43927|issue:45698", + "issue:44977|issue:45229", + "issue:46421|issue:46858", + "issue:42490|issue:43873", + "issue:25251|issue:32090", + "issue:43873|issue:45229", + "issue:43976|issue:45440", + "issue:44448|issue:44466", + "issue:43381|issue:43701", + "issue:44871|issue:44977", + "issue:29127|issue:34567", + "issue:33357|issue:39692", + "issue:43278|issue:43575", + "issue:44561|issue:44779", + "issue:45588|issue:45823", + "issue:44568|issue:45491", + "issue:41720|issue:45229", + "issue:44186|issue:45198", + "issue:44361|issue:45198", + "issue:45198|issue:45803" + ], + "prepared_review_unit_hash": "b827f2565d0d9de0c34183ff609a60fa9d06a476455c0a4caaf220e3a991941d", + "component_budget": { + "node_count": 514, + "item_count": 514, + "soft_pair_count": 1891, + "serialized_chars": 980895, + "estimated_input_tokens": 245224, + "estimated_eval_tokens": 490704 + }, + "budget": { + "node_count": 47, + "item_count": 47, + "soft_pair_count": 30, + "serialized_chars": 36452, + "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 largely heterogeneous: the suggested soft pairs mostly share only broad subsystem labels (tokenization, cache, training, serialization) but not the same concrete bug. I would not merge any of the proposed pairs as duplicates.", + "confidence": 0.94, + "canonical_issue_reason": null, + "canonical_pr_reason": null, + "best_issue_reason": null, + "best_pr_reason": null, + "soft_edge_verdicts": [ + { + "left": "issue:37428", + "right": "issue:42175", + "accept": false, + "reason": "Different problems in unrelated areas: flash-attention import error vs missing TensorFlow backend in a torch install." + }, + { + "left": "issue:45290", + "right": "issue:46655", + "accept": false, + "reason": "Both are processing-related, but one is a chat template crash and the other is a video padding bug; different code paths." + }, + { + "left": "issue:44448", + "right": "issue:44977", + "accept": false, + "reason": "Generation mismatch for Pegasus v4/v5 is unrelated to Qwen3.5 flash-attention failures." + }, + { + "left": "issue:29870", + "right": "issue:33290", + "accept": false, + "reason": "Bart generation UnboundLocalError and DeepSpeed Adafactor OOM are separate training/inference failures." + }, + { + "left": "issue:44360", + "right": "issue:44561", + "accept": false, + "reason": "A DSA ReLU discussion and a v5 trust_remote_code regression are not the same bug." + }, + { + "left": "issue:43278", + "right": "issue:43381", + "accept": false, + "reason": "dtype drift between train/eval is different from checkpointing being disallowed in eval mode." + }, + { + "left": "issue:44855", + "right": "issue:46399", + "accept": false, + "reason": "Both mention decorators, but they fail in different models and for different reasons." + }, + { + "left": "issue:40444", + "right": "issue:43526", + "accept": false, + "reason": "Multi-image VL dataset finetuning failure is unrelated to BeitImageProcessorFast label reduction." + }, + { + "left": "issue:45920", + "right": "issue:46192", + "accept": false, + "reason": "Wrong token IDs in AutoTokenizer is a different bug from return_tensors being ignored with text_kwargs." + }, + { + "left": "issue:32090", + "right": "issue:33290", + "accept": false, + "reason": "Trainer GPU broadcast NoneType error is not the same as Adafactor/DeepSpeed OOM." + }, + { + "left": "issue:43927", + "right": "issue:45698", + "accept": false, + "reason": "Both involve save/load behavior, but one loses custom token IDs while the other loads the wrong custom module." + }, + { + "left": "issue:44977", + "right": "issue:45229", + "accept": false, + "reason": "Flash-attention generation bug and multi-GPU CUDA OOM are different failure modes." + }, + { + "left": "issue:46421", + "right": "issue:46858", + "accept": false, + "reason": "Cache lazy-init recompilation and device-tensor seq-length arithmetic breakage are distinct cache/compile issues." + }, + { + "left": "issue:42490", + "right": "issue:43873", + "accept": false, + "reason": "Local-path import warning is unrelated to quantization offloading not working." + }, + { + "left": "issue:25251", + "right": "issue:32090", + "accept": false, + "reason": "Pipeline top_k nesting behavior is unrelated to Trainer GPU broadcast errors." + }, + { + "left": "issue:43873", + "right": "issue:45229", + "accept": false, + "reason": "Quantization offloading failure is not the same as Gemma4 multi-GPU OOM." + }, + { + "left": "issue:43976", + "right": "issue:45440", + "accept": false, + "reason": "Python version support regression and DeepSeekV3MoE implementation divergence are unrelated." + }, + { + "left": "issue:44448", + "right": "issue:44466", + "accept": false, + "reason": "Output differences for Pegasus and lm_head serialization inconsistency are separate bugs." + }, + { + "left": "issue:43381", + "right": "issue:43701", + "accept": false, + "reason": "Eval-mode gradient checkpointing and resume_from_checkpoint key mismatch are different problems." + }, + { + "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:29127", + "right": "issue:34567", + "accept": false, + "reason": "LayoutLMv3 box-validation messaging and TrainerState token counting are unrelated." + }, + { + "left": "issue:33357", + "right": "issue:39692", + "accept": false, + "reason": "MacOS bus error on CLIP is not the same as SigLIP2 documentation/example errors." + }, + { + "left": "issue:43278", + "right": "issue:43575", + "accept": false, + "reason": "Embedding dtype drift and Qwen2-57B tensor-parallel OOM are unrelated." + }, + { + "left": "issue:44561", + "right": "issue:44779", + "accept": false, + "reason": "Removal of is_torch_fx_available breaking trust_remote_code is a different regression from DeepSeek tokenizer output changes." + }, + { + "left": "issue:45588", + "right": "issue:45823", + "accept": false, + "reason": "Flash-attention sink-less model crash and Gemma4 device-map mismatch are different forward-path bugs." + }, + { + "left": "issue:44568", + "right": "issue:45491", + "accept": false, + "reason": "Tokenizer special-token insertion bug is unrelated to Gemma3 NaN embeddings from sliding-window batching." + }, + { + "left": "issue:41720", + "right": "issue:45229", + "accept": false, + "reason": "Qwen3 auto device mapping assert and Gemma4 multi-GPU OOM are different device-placement failures." + }, + { + "left": "issue:44186", + "right": "issue:45198", + "accept": false, + "reason": "LayoutLMv2Tokenizer NER/padding crash and Wav2Vec2 save/tokenization failure are separate tokenizer issues." + }, + { + "left": "issue:44361", + "right": "issue:45198", + "accept": false, + "reason": "MLukeTokenizer AttributeError is unrelated to Wav2Vec2 save_pretrained/tokenization failure." + }, + { + "left": "issue:45198", + "right": "issue:45803", + "accept": false, + "reason": "Wav2Vec2 tokenizer/save bug and Fuyu's bare except swallowing errors are different defects." + } + ] + }, + "evaluator_result": null, + "evaluator_used": false, + "retried": false, + "accepted_nontrivial_soft_edge": false, + "error_kind": null, + "error_message": null + }, + { + "label": "issue", + "component_index": 1, + "component_count": 27, + "review_unit_index": 34, + "review_unit_count": 63, + "status": "reviewed", + "reason": null, + "source": "llm", + "cache_hit": false, + "model": "gpt-5.4-mini?service_tier=flex", + "cluster_id": "cluster-10000-514", + "nodes": [ + "issue:29127", + "issue:29870", + "issue:29942", + "issue:30064", + "issue:33357", + "issue:34689", + "issue:35141", + "issue:42490", + "issue:42915", + "issue:42994", + "issue:43065", + "issue:43066", + "issue:43381", + "issue:43412", + "issue:43421", + "issue:43526", + "issue:43575", + "issue:43611", + "issue:43653", + "issue:43701", + "issue:43756", + "issue:43994", + "issue:44206", + "issue:44265", + "issue:44387", + "issue:44442", + "issue:44462", + "issue:44479", + "issue:44743", + "issue:44977", + "issue:45072", + "issue:45198", + "issue:45245", + "issue:45372", + "issue:45406", + "issue:45412", + "issue:45440", + "issue:45459", + "issue:45491", + "issue:45563", + "issue:45588", + "issue:45593", + "issue:45712", + "issue:45834", + "issue:46121", + "issue:46369", + "issue:46693", + "issue:46701" + ], + "soft_pairs": [ + "issue:43611|issue:46121", + "issue:44977|issue:45563", + "issue:29870|issue:34689", + "issue:29942|issue:30064", + "issue:43653|issue:44743", + "issue:44206|issue:44479", + "issue:42915|issue:42994", + "issue:43756|issue:45440", + "issue:43653|issue:44442", + "issue:43065|issue:43994", + "issue:44265|issue:45198", + "issue:43065|issue:45412", + "issue:43575|issue:43701", + "issue:29127|issue:33357", + "issue:45491|issue:46693", + "issue:45372|issue:45588", + "issue:43412|issue:45588", + "issue:42490|issue:43381", + "issue:29127|issue:35141", + "issue:45245|issue:45563", + "issue:45459|issue:45712", + "issue:45440|issue:45593", + "issue:43994|issue:45412", + "issue:45372|issue:45406", + "issue:45406|issue:46701", + "issue:43421|issue:45491", + "issue:43575|issue:44387", + "issue:43066|issue:44462", + "issue:45412|issue:46369", + "issue:45372|issue:45459", + "issue:44442|issue:45072", + "issue:43526|issue:45834" + ], + "prepared_review_unit_hash": "2c2b63fa9e8f73b6f11e9ac9cf16cdd2960d118ea5c4fdd17572efcb7966dd17", + "component_budget": { + "node_count": 514, + "item_count": 514, + "soft_pair_count": 1891, + "serialized_chars": 980895, + "estimated_input_tokens": 245224, + "estimated_eval_tokens": 490704 + }, + "budget": { + "node_count": 48, + "item_count": 48, + "soft_pair_count": 32, + "serialized_chars": 37361, + "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 topical matches rather than true duplicates. I would not merge any soft pair. If a representative issue is needed, issue 46693 is the strongest standalone regression report by activity and references.", + "confidence": 0.79, + "canonical_issue_reason": "No clear duplicate cluster exists; issue 46693 is the most active, concrete regression report and is the least arbitrary anchor among the set.", + "canonical_pr_reason": null, + "best_issue_reason": "Issue 46693 has the richest discussion and inbound references, and it describes a user-visible regression with a well-defined failure mode.", + "best_pr_reason": null, + "soft_edge_verdicts": [ + { + "left": "issue:43611", + "right": "issue:46121", + "accept": false, + "reason": "Different bugs: model loading via base_model_prefix vs rope-validation TypeError from JSON-loaded lists." + }, + { + "left": "issue:44977", + "right": "issue:45563", + "accept": false, + "reason": "Different failure modes: flash-attention generation breakage vs a stale warning in paged generate()." + }, + { + "left": "issue:29870", + "right": "issue:34689", + "accept": false, + "reason": "Unrelated subsystems: BART generate() decoder mask bug vs Llama 3.2 Vision model loading regression." + }, + { + "left": "issue:29942", + "right": "issue:30064", + "accept": false, + "reason": "One is Flash Attention 2 test failures; the other is image processor handling of void segmentation maps." + }, + { + "left": "issue:43653", + "right": "issue:44743", + "accept": false, + "reason": "Tokenizer special-token decode bug vs recurrent-state reset in modular_qwen3_5 cache logic." + }, + { + "left": "issue:44206", + "right": "issue:44479", + "accept": false, + "reason": "Feature-extractor arg crash vs video-input regression across Qwen VL variants." + }, + { + "left": "issue:42915", + "right": "issue:42994", + "accept": false, + "reason": "Both involve quantization broadly, but one is a FineGrainedFP8 config failure and the other is quantized model saving." + }, + { + "left": "issue:43756", + "right": "issue:45440", + "accept": false, + "reason": "Different models and defects: missing RoPE layers in Smollm3 vs divergence in DeepseekV3MoE implementation." + }, + { + "left": "issue:43653", + "right": "issue:44442", + "accept": false, + "reason": "Tokenizer decode-output bug vs AutoTokenizer failing to load a specific tokenizer class." + }, + { + "left": "issue:43065", + "right": "issue:43994", + "accept": false, + "reason": "Different vision-model issues: dummy Conv2d in Sam3PixelDecoder vs nonsensical SigLIP2 inference." + }, + { + "left": "issue:44265", + "right": "issue:45198", + "accept": false, + "reason": "torch.export failure from torch_compilable_check is unrelated to Wav2Vec2 save_pretrained/tokenization failures." + }, + { + "left": "issue:43065", + "right": "issue:45412", + "accept": false, + "reason": "Sam3PixelDecoder dummy layer issue is unrelated to RT-DETR memory not being released on deletion." + }, + { + "left": "issue:43575", + "right": "issue:43701", + "accept": false, + "reason": "OOM on tensor-parallel model loading is unrelated to resume_from_checkpoint key mismatch." + }, + { + "left": "issue:29127", + "right": "issue:33357", + "accept": false, + "reason": "LayoutLMv3 box-information error and MacOS CLIP bus error are different models, symptoms, and code paths." + }, + { + "left": "issue:45491", + "right": "issue:46693", + "accept": false, + "reason": "Both mention generation context, but one is a Gemma3 NaN-embedding bug and the other is a FA2 long-context performance regression." + }, + { + "left": "issue:45372", + "right": "issue:45588", + "accept": false, + "reason": "Gemma 4 processor import failure is unrelated to flash_attention.py crashing on s_aux=None." + }, + { + "left": "issue:43412", + "right": "issue:45588", + "accept": false, + "reason": "Executorch export guard/erfinv support issue is a different code path from the flash_attention.py AttributeError." + }, + { + "left": "issue:42490", + "right": "issue:43381", + "accept": false, + "reason": "Local-path import warning is unrelated to gradient checkpointing being disallowed in eval mode." + }, + { + "left": "issue:29127", + "right": "issue:35141", + "accept": false, + "reason": "LayoutLMv3 error messaging and embedding reinitialization after resize_embeddings are separate bugs." + }, + { + "left": "issue:45245", + "right": "issue:45563", + "accept": false, + "reason": "Category-count RuntimeError is unrelated to the paged generate stale warning." + }, + { + "left": "issue:45459", + "right": "issue:45712", + "accept": false, + "reason": "Tokenizer protobuf error handling is unrelated to leftover dummy PT classes leaking into dir(transformers)." + }, + { + "left": "issue:45440", + "right": "issue:45593", + "accept": false, + "reason": "DeepseekV3MoE divergence is unrelated to D-FINE auxiliary-loss behavior." + }, + { + "left": "issue:43994", + "right": "issue:45412", + "accept": false, + "reason": "SigLIP2 inference quality issue and RT-DETR memory leak are different subsystems and symptoms." + }, + { + "left": "issue:45372", + "right": "issue:45406", + "accept": false, + "reason": "Gemma 4 processor loading import error is unrelated to transformers serve lacking Gemma4Processor._tokenizer." + }, + { + "left": "issue:45406", + "right": "issue:46701", + "accept": false, + "reason": "Missing _tokenizer in serve is unrelated to DiffusionGemmaGenerationOutput pipeline breakage." + }, + { + "left": "issue:43421", + "right": "issue:45491", + "accept": false, + "reason": "Runtime post-processor update behavior is unrelated to Gemma3 NaN embeddings on mixed-length batches." + }, + { + "left": "issue:43575", + "right": "issue:44387", + "accept": false, + "reason": "Model-loading OOM for Qwen2-57B is not the same as increased CUDA reserved memory under int4 quantization." + }, + { + "left": "issue:43066", + "right": "issue:44462", + "accept": false, + "reason": "Wrong tokenizer decoder type in v5 and ignoring tokenizer.json are related to tokenizers, but they are distinct loading bugs." + }, + { + "left": "issue:45412", + "right": "issue:46369", + "accept": false, + "reason": "RT-DETR deletion/GC memory leak is different from SAM3 litetext memory overflow during model use." + }, + { + "left": "issue:45372", + "right": "issue:45459", + "accept": false, + "reason": "Gemma 4 processor import failure is unrelated to the protobuf decode-error masking bug." + }, + { + "left": "issue:44442", + "right": "issue:45072", + "accept": false, + "reason": "FastSpeech2ConformerTokenizer loading failure is unrelated to bfloat16 dtype mismatches in SwitchTransformers/TimmWrapperModel." + }, + { + "left": "issue:43526", + "right": "issue:45834", + "accept": false, + "reason": "BeitImageProcessorFast reduce_labels bug is unrelated to Kosmos2.5 OCR index errors." + } + ] + }, + "evaluator_result": null, + "evaluator_used": false, + "retried": false, + "accepted_nontrivial_soft_edge": false, + "error_kind": null, + "error_message": null + }, + { + "label": "issue", + "component_index": 1, + "component_count": 27, + "review_unit_index": 35, + "review_unit_count": 63, + "status": "reviewed", + "reason": null, + "source": "llm", + "cache_hit": false, + "model": "gpt-5.4-mini?service_tier=flex", + "cluster_id": "cluster-10000-514", + "nodes": [ + "issue:27957", + "issue:29127", + "issue:29870", + "issue:33290", + "issue:34689", + "issue:35141", + "issue:42175", + "issue:42490", + "issue:42915", + "issue:43262", + "issue:43316", + "issue:43334", + "issue:43381", + "issue:43525", + "issue:43526", + "issue:43531", + "issue:43646", + "issue:43720", + "issue:43761", + "issue:43824", + "issue:43873", + "issue:43931", + "issue:43994", + "issue:44077", + "issue:44186", + "issue:44206", + "issue:44279", + "issue:44442", + "issue:44514", + "issue:44743", + "issue:44811", + "issue:44857", + "issue:44898", + "issue:44977", + "issue:45003", + "issue:45245", + "issue:45250", + "issue:45381", + "issue:45440", + "issue:45588", + "issue:45859", + "issue:46121", + "issue:46234", + "issue:46369", + "issue:46399", + "issue:46640", + "issue:46693", + "issue:46821" + ], + "soft_pairs": [ + "issue:27957|issue:33290", + "issue:27957|issue:29127", + "issue:43824|issue:44279", + "issue:43526|issue:46693", + "issue:33290|issue:34689", + "issue:33290|issue:35141", + "issue:44857|issue:45588", + "issue:43994|issue:46369", + "issue:43526|issue:44186", + "issue:42915|issue:43931", + "issue:43334|issue:43531", + "issue:45250|issue:46640", + "issue:43381|issue:43873", + "issue:45588|issue:46821", + "issue:43646|issue:44077", + "issue:46121|issue:46399", + "issue:44514|issue:45381", + "issue:43761|issue:45859", + "issue:44977|issue:45245", + "issue:44743|issue:44811", + "issue:44442|issue:44898", + "issue:43262|issue:44206", + "issue:42490|issue:44743", + "issue:42175|issue:46234", + "issue:43720|issue:44442", + "issue:27957|issue:29870", + "issue:43316|issue:43525", + "issue:45003|issue:45440" + ], + "prepared_review_unit_hash": "6ca5e913b3bae805e06fbc2af5cbb9b7513da3986fc6625509ce951710e60d1b", + "component_budget": { + "node_count": 514, + "item_count": 514, + "soft_pair_count": 1891, + "serialized_chars": 980895, + "estimated_input_tokens": 245224, + "estimated_eval_tokens": 490704 + }, + "budget": { + "node_count": 48, + "item_count": 48, + "soft_pair_count": 28, + "serialized_chars": 35948, + "estimated_input_tokens": 8987, + "estimated_eval_tokens": 18230 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": true, + "analyst_result": { + "summary": "Mostly a heterogeneous set of unrelated Transformers issues. The only recurring theme is Flash Attention 2, but the paired items describe different failures and code paths, so no soft pair should be merged.", + "confidence": 0.86, + "canonical_issue_reason": "Issue 46640 is the clearest current anchor for the Flash Attention 2 subgroup: it is explicitly about FA2, open, and broad enough to collect related reports.", + "canonical_pr_reason": null, + "best_issue_reason": "Issue 46640 is the best representative issue in this set because it is the most directly on-topic and general within the only visible cluster theme (FA2).", + "best_pr_reason": null, + "soft_edge_verdicts": [ + { + "left": "issue:27957", + "right": "issue:33290", + "accept": false, + "reason": "XLMRoberta/Flash Attention 2 and deepspeed Adafactor OOM are different bugs in different paths." + }, + { + "left": "issue:27957", + "right": "issue:29127", + "accept": false, + "reason": "FA2 model support vs LayoutLMv3 error messaging; unrelated failure modes." + }, + { + "left": "issue:43824", + "right": "issue:44279", + "accept": false, + "reason": "Importing a missing Qwen2.5-VL class is not the same as a generic dependency issue." + }, + { + "left": "issue:43526", + "right": "issue:46693", + "accept": false, + "reason": "BeitImageProcessorFast label reduction bug and FA2 long-context performance regression are unrelated." + }, + { + "left": "issue:33290", + "right": "issue:34689", + "accept": false, + "reason": "Adafactor/deepspeed OOM and Llama 3.2 Vision load break are distinct problems." + }, + { + "left": "issue:33290", + "right": "issue:35141", + "accept": false, + "reason": "Optimizer memory usage and embedding reinitialization during post_init are unrelated code paths." + }, + { + "left": "issue:44857", + "right": "issue:45588", + "accept": false, + "reason": "AMP/CUDA loss crash and flash_attention.py s_aux=None crash do not describe the same bug." + }, + { + "left": "issue:43994", + "right": "issue:46369", + "accept": false, + "reason": "Siglip2 bad outputs and sam3-litetext memory overflow are different issues." + }, + { + "left": "issue:43526", + "right": "issue:44186", + "accept": false, + "reason": "ImageProcessorFast label reduction and LayoutLMv2Tokenizer NER/padding crashes are unrelated." + }, + { + "left": "issue:42915", + "right": "issue:43931", + "accept": false, + "reason": "Qwen3Moe FineGrainedFP8Config failure and Qwen3-VL weight-shape mismatch are separate model-loading issues." + }, + { + "left": "issue:43334", + "right": "issue:43531", + "accept": false, + "reason": "Missing pad_token_id on Qwen3-VL text config is unrelated to sliding_window behavior in Qwen3-MoE." + }, + { + "left": "issue:45250", + "right": "issue:46640", + "accept": false, + "reason": "Both mention Flash Attention 2, but 45250 is a generic topic/title while 46640 is a specific crash report; not enough evidence of the same bug." + }, + { + "left": "issue:43381", + "right": "issue:43873", + "accept": false, + "reason": "Gradient checkpointing in eval mode and quantization offloading misbehavior are different problems." + }, + { + "left": "issue:45588", + "right": "issue:46821", + "accept": false, + "reason": "flash_attention.py AttributeError on sink-less models and TimesFM window_size AttributeError are unrelated." + }, + { + "left": "issue:43646", + "right": "issue:44077", + "accept": false, + "reason": "Custom model initialization break in Transformers 5.0 and patchtsmixer post_init policy are not the same change." + }, + { + "left": "issue:46121", + "right": "issue:46399", + "accept": false, + "reason": "JSON-loaded list handling in rope validation and Qwen3.5 rotary embedding registration are different code paths." + }, + { + "left": "issue:44514", + "right": "issue:45381", + "accept": false, + "reason": "Batched chat-template padding crash and wrong video vision_position_ids are distinct Qwen2.5-VL issues." + }, + { + "left": "issue:43761", + "right": "issue:45859", + "accept": false, + "reason": "CLIPVisionModel hidden_states regression and Qwen3_5MoeTextRotaryEmbedding CPU offload incompatibility are unrelated." + }, + { + "left": "issue:44977", + "right": "issue:45245", + "accept": false, + "reason": "Qwen3.5 flash-attention generation bug and category-count overflow are unrelated." + }, + { + "left": "issue:44743", + "right": "issue:44811", + "accept": false, + "reason": "Qwen3.5 recurrent-state reset and Whisper batch_decode skip_special_tokens are different components and symptoms." + }, + { + "left": "issue:44442", + "right": "issue:44898", + "accept": false, + "reason": "FastSpeech2ConformerTokenizer loading and Perceiver non-default resolution failure are unrelated." + }, + { + "left": "issue:43262", + "right": "issue:44206", + "accept": false, + "reason": "Audio chat-template sample-rate default and LasrFeatureExtractor center-arg crash are different audio issues." + }, + { + "left": "issue:42490", + "right": "issue:44743", + "accept": false, + "reason": "Local path trailing-slash import warning and Qwen3.5 recurrent-state reset are unrelated." + }, + { + "left": "issue:42175", + "right": "issue:46234", + "accept": false, + "reason": "Missing TensorFlow backend in extra install and missing import_utils helper are not the same dependency bug." + }, + { + "left": "issue:43720", + "right": "issue:44442", + "accept": false, + "reason": "BitNet packed-weight loading and tokenizer load failure are different failures." + }, + { + "left": "issue:27957", + "right": "issue:29870", + "accept": false, + "reason": "FA2 support for XLMRoberta and Bart generate decoder_attention_mask UnboundLocalError are unrelated." + }, + { + "left": "issue:43316", + "right": "issue:43525", + "accept": false, + "reason": "Gemma3TextConfig API discrepancy and Llama4Config missing pad_token_id are separate config issues." + }, + { + "left": "issue:45003", + "right": "issue:45440", + "accept": false, + "reason": "Unsafe sys.modules access in modeling_utils and DeepseekV3MoE divergence are unrelated bugs." + } + ] + }, + "evaluator_result": null, + "evaluator_used": false, + "retried": false, + "accepted_nontrivial_soft_edge": false, + "error_kind": null, + "error_message": null + }, + { + "label": "issue", + "component_index": 1, + "component_count": 27, + "review_unit_index": 36, + "review_unit_count": 63, + "status": "reviewed", + "reason": null, + "source": "llm", + "cache_hit": false, + "model": "gpt-5.4-mini?service_tier=flex", + "cluster_id": "cluster-10000-514", + "nodes": [ + "issue:16998", + "issue:22355", + "issue:25251", + "issue:27957", + "issue:28282", + "issue:32090", + "issue:42617", + "issue:42907", + "issue:43232", + "issue:43316", + "issue:43388", + "issue:43493", + "issue:43550", + "issue:43575", + "issue:43606", + "issue:43643", + "issue:43701", + "issue:43742", + "issue:43749", + "issue:43761", + "issue:43827", + "issue:43873", + "issue:43931", + "issue:44062", + "issue:44077", + "issue:44079", + "issue:44188", + "issue:44190", + "issue:44442", + "issue:44479", + "issue:44496", + "issue:44855", + "issue:44871", + "issue:44913", + "issue:44987", + "issue:45137", + "issue:45216", + "issue:45325", + "issue:45698", + "issue:45759", + "issue:45803", + "issue:45907", + "issue:46178", + "issue:46506", + "issue:46697", + "issue:46829", + "issue:46932" + ], + "soft_pairs": [ + "issue:44077|issue:44479", + "issue:43493|issue:43827", + "issue:43873|issue:44871", + "issue:44062|issue:45907", + "issue:45325|issue:46697", + "issue:44855|issue:45137", + "issue:43643|issue:44913", + "issue:43316|issue:45759", + "issue:43749|issue:43931", + "issue:44987|issue:46178", + "issue:44190|issue:44496", + "issue:43931|issue:45216", + "issue:43701|issue:46506", + "issue:44442|issue:45803", + "issue:42617|issue:43575", + "issue:43232|issue:43388", + "issue:43550|issue:43701", + "issue:43575|issue:43742", + "issue:43761|issue:44079", + "issue:43761|issue:44188", + "issue:27957|issue:32090", + "issue:43606|issue:43701", + "issue:25251|issue:28282", + "issue:46829|issue:46932", + "issue:42907|issue:45698", + "issue:16998|issue:22355" + ], + "prepared_review_unit_hash": "c455cdbc834ebb07e8bfa03a525abea5fe9f81de7207058c4eb40a9709de9efd", + "component_budget": { + "node_count": 514, + "item_count": 514, + "soft_pair_count": 1891, + "serialized_chars": 980895, + "estimated_input_tokens": 245224, + "estimated_eval_tokens": 490704 + }, + "budget": { + "node_count": 47, + "item_count": 47, + "soft_pair_count": 26, + "serialized_chars": 34974, + "estimated_input_tokens": 8744, + "estimated_eval_tokens": 17744 + }, + "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 superficial semantic similarity; I do not see a true duplicate group here.", + "confidence": 0.95, + "canonical_issue_reason": null, + "canonical_pr_reason": null, + "best_issue_reason": null, + "best_pr_reason": null, + "soft_edge_verdicts": [ + { + "left": "issue:44077", + "right": "issue:44479", + "accept": false, + "reason": "Different bugs: one is about `post_init` API policy for patchtsmixer, the other is a v5 video-input regression across Qwen-VL models." + }, + { + "left": "issue:43493", + "right": "issue:43827", + "accept": false, + "reason": "SigLIP2 implementation mismatch vs documentation still referencing removed `pipeline()` calls; unrelated topics and code paths." + }, + { + "left": "issue:43873", + "right": "issue:44871", + "accept": false, + "reason": "Quantization/offloading behavior versus Gemma-3 eos token config inconsistency; not the same underlying defect." + }, + { + "left": "issue:44062", + "right": "issue:45907", + "accept": false, + "reason": "Tokenizer `AddedToken` argument collision and `list_repo_files` not forwarding `revision` are unrelated API bugs." + }, + { + "left": "issue:45325", + "right": "issue:46697", + "accept": false, + "reason": "Qwen2.5-VL rope/position-id scaling bug and Swin2SR over-padding bug affect different models and functions." + }, + { + "left": "issue:44855", + "right": "issue:45137", + "accept": false, + "reason": "Python 3.13 `torch.jit.script` parsing issue in DebertaV2 import is unrelated to DeepSpeed ZeRO3 deque handling." + }, + { + "left": "issue:43643", + "right": "issue:44913", + "accept": false, + "reason": "`trust_remote_code` config field loss and GPTNeoX `rotary_pct` not persisting on reload are different config-loading problems." + }, + { + "left": "issue:43316", + "right": "issue:45759", + "accept": false, + "reason": "Both involve config APIs, but one is a Gemma3TextConfig discrepancy and the other is unwrapping `text_config` for composite Qwen multimodal configs; not the same bug." + }, + { + "left": "issue:43749", + "right": "issue:43931", + "accept": false, + "reason": "FSDP RAM-efficient loading regression and Qwen3-VL weight-shape mismatch are distinct model-loading failures." + }, + { + "left": "issue:44987", + "right": "issue:46178", + "accept": false, + "reason": "Loading `physical-intelligence/fast` failing in newer Transformers is unrelated to an unsupported backend entry for `tensorflow_text`." + }, + { + "left": "issue:44190", + "right": "issue:44496", + "accept": false, + "reason": "Local dataset loading in a training script and unrecognized model config are separate issues." + }, + { + "left": "issue:43931", + "right": "issue:45216", + "accept": false, + "reason": "One is a load-time shape mismatch for Qwen3-VL, the other is a save-pretrained checkpoint regression for Qwen3.5; different directions and code paths." + }, + { + "left": "issue:43701", + "right": "issue:46506", + "accept": false, + "reason": "Resume-from-checkpoint key mismatch and CPU/CUDA divergence in PPDocLayoutV3 are unrelated." + }, + { + "left": "issue:44442", + "right": "issue:45803", + "accept": false, + "reason": "FastSpeech2Conformer tokenizer loading failure and a bare `except` in Fuyu tensor conversion are not the same bug." + }, + { + "left": "issue:42617", + "right": "issue:43575", + "accept": false, + "reason": "`3d_parallel.py` run failure and Qwen2-57B-A14B-Instruct tensor-parallel OOM are different runtime problems." + }, + { + "left": "issue:43232", + "right": "issue:43388", + "accept": false, + "reason": "Generation kwargs sync issue and `gather_for_metrics` dropping tuple labels affect different training/generation code paths." + }, + { + "left": "issue:43550", + "right": "issue:43701", + "accept": false, + "reason": "Bamba torch.compile/SDPA failure is unrelated to resume-from-checkpoint key mismatch." + }, + { + "left": "issue:43575", + "right": "issue:43742", + "accept": false, + "reason": "Large-model OOM on load and MobileLLM key error are different model-loading failures." + }, + { + "left": "issue:43761", + "right": "issue:44079", + "accept": false, + "reason": "CLIPVisionModel not returning hidden states and ModelOutput key assignment when a key was previously `None` are related in theme but not clearly the same underlying bug." + }, + { + "left": "issue:43761", + "right": "issue:44188", + "accept": false, + "reason": "Hidden-states regression in CLIPVisionModel is unrelated to attention-kernel divergence under `torch.compile`." + }, + { + "left": "issue:27957", + "right": "issue:32090", + "accept": false, + "reason": "Flash Attention 2 support for XLMRoberta and a Trainer `_gpu_broadcast_one` NoneType error are unrelated." + }, + { + "left": "issue:43606", + "right": "issue:43701", + "accept": false, + "reason": "CPU-offload device mismatch for bark-small and resume-from-checkpoint key mismatch do not share the same cause." + }, + { + "left": "issue:25251", + "right": "issue:28282", + "accept": false, + "reason": "Pipeline `top_k` output nesting and missing PyTorch for AutoModel are unrelated." + }, + { + "left": "issue:46829", + "right": "issue:46932", + "accept": false, + "reason": "VideoPrism hidden_states misassignment and Gemma4 quantized embedder dtype casting are distinct bugs." + }, + { + "left": "issue:42907", + "right": "issue:45698", + "accept": false, + "reason": "Saving dequantized models and loading the wrong custom module after save_pretrained are both serialization-related, but not the same concrete defect." + }, + { + "left": "issue:16998", + "right": "issue:22355", + "accept": false, + "reason": "A question about `model_max_length` and a missing `transformers.onnx` module 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": 27, + "review_unit_index": 37, + "review_unit_count": 63, + "status": "reviewed", + "reason": null, + "source": "llm", + "cache_hit": false, + "model": "gpt-5.4-mini?service_tier=flex", + "cluster_id": "cluster-10000-514", + "nodes": [ + "issue:43065", + "issue:43295", + "issue:43316", + "issue:43352", + "issue:43404", + "issue:43441", + "issue:43452", + "issue:43504", + "issue:43540", + "issue:43572", + "issue:43646", + "issue:43650", + "issue:43716", + "issue:43825", + "issue:44077", + "issue:44162", + "issue:44242", + "issue:44464", + "issue:44496", + "issue:44514", + "issue:44534", + "issue:44610", + "issue:44617", + "issue:44792", + "issue:44871", + "issue:44913", + "issue:44977", + "issue:45127", + "issue:45200", + "issue:45542", + "issue:45584", + "issue:45600", + "issue:45632", + "issue:45698", + "issue:45803", + "issue:46050", + "issue:46170", + "issue:46178", + "issue:46202", + "issue:46234", + "issue:46291", + "issue:46365", + "issue:46369", + "issue:46421", + "issue:46424", + "issue:46640", + "issue:46821", + "issue:46894" + ], + "soft_pairs": [ + "issue:43295|issue:43825", + "issue:43295|issue:44871", + "issue:44871|issue:45584", + "issue:46234|issue:46291", + "issue:43316|issue:43716", + "issue:44077|issue:44534", + "issue:43316|issue:46821", + "issue:44464|issue:46421", + "issue:43452|issue:44496", + "issue:43646|issue:44913", + "issue:44871|issue:46050", + "issue:45803|issue:46365", + "issue:44162|issue:44792", + "issue:43404|issue:43441", + "issue:44242|issue:46894", + "issue:43065|issue:44617", + "issue:43540|issue:44514", + "issue:44610|issue:46369", + "issue:45632|issue:46424", + "issue:43572|issue:46178", + "issue:45542|issue:46234", + "issue:45803|issue:46170", + "issue:44792|issue:45600", + "issue:44792|issue:44977", + "issue:43504|issue:45698", + "issue:43650|issue:46202", + "issue:43352|issue:46640", + "issue:45127|issue:45200" + ], + "prepared_review_unit_hash": "15a013321fbc76ce26218c8cd8142b2d979182412b9abe118c048ba4c12fabd4", + "component_budget": { + "node_count": 514, + "item_count": 514, + "soft_pair_count": 1891, + "serialized_chars": 980895, + "estimated_input_tokens": 245224, + "estimated_eval_tokens": 490704 + }, + "budget": { + "node_count": 48, + "item_count": 48, + "soft_pair_count": 28, + "serialized_chars": 35927, + "estimated_input_tokens": 8982, + "estimated_eval_tokens": 18220 + }, + "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 many unrelated regressions, model-specific bugs, and import/load-path issues. None of the soft pairs look like the same underlying bug or change.", + "confidence": 0.95, + "canonical_issue_reason": "No single canonical duplicate exists; if one issue must anchor the cluster, #43646 is the broadest v5/custom-initialization regression.", + "canonical_pr_reason": null, + "best_issue_reason": "#43646 is the best representative because it is the broadest umbrella for the few adjacent custom-model/config regressions, though the cluster is still heterogeneous.", + "best_pr_reason": null, + "soft_edge_verdicts": [ + { + "left": "issue:43295", + "right": "issue:43825", + "accept": false, + "reason": "Different bugs: processor.tokenizer/image-input regression vs pipeline translation-task error messaging." + }, + { + "left": "issue:43295", + "right": "issue:44871", + "accept": false, + "reason": "Custom processor/tokenizer regression and Gemma eos_token_id mismatch are unrelated model/config issues." + }, + { + "left": "issue:44871", + "right": "issue:45584", + "accept": false, + "reason": "Gemma eos_token_id inconsistency vs Whisper empty-transcription crash after align_special_tokens; different code paths." + }, + { + "left": "issue:46234", + "right": "issue:46291", + "accept": false, + "reason": "Missing import helper vs kernels dependency import failure; both are import issues but not the same defect." + }, + { + "left": "issue:43316", + "right": "issue:43716", + "accept": false, + "reason": "Gemma3TextConfig API mismatch vs Mistral-3 dtype mismatch; different models and failure modes." + }, + { + "left": "issue:44077", + "right": "issue:44534", + "accept": false, + "reason": "Optional post_init restriction and junk non-persistent buffers are unrelated transformer-v5 behaviors." + }, + { + "left": "issue:43316", + "right": "issue:46821", + "accept": false, + "reason": "Different model families and bugs: config API discrepancy vs TimesFM window_size AttributeError." + }, + { + "left": "issue:44464", + "right": "issue:46421", + "accept": false, + "reason": "Chunked generation inconsistency with compiled forward is not the same as lazy cache initialization recompilation." + }, + { + "left": "issue:43452", + "right": "issue:44496", + "accept": false, + "reason": "Both touch model loading, but one is gguf_file/from_pretrained breakage and the other is unrecognized model config.json; not the same root cause." + }, + { + "left": "issue:43646", + "right": "issue:44913", + "accept": false, + "reason": "Broad custom model initialization regression vs GPTNeoX rotary_pct reload persistence bug; different issues." + }, + { + "left": "issue:44871", + "right": "issue:46050", + "accept": false, + "reason": "Gemma eos_token_id mismatch and sinq quantized save/load corruption are unrelated." + }, + { + "left": "issue:45803", + "right": "issue:46365", + "accept": false, + "reason": "Bare except in tensor conversion vs Trainer import failure due to datasets; unrelated problems." + }, + { + "left": "issue:44162", + "right": "issue:44792", + "accept": false, + "reason": "ESM2 breakage and janus image-generation test failure are different models and symptoms." + }, + { + "left": "issue:43404", + "right": "issue:43441", + "accept": false, + "reason": "lm_head weight tying bug vs FlashAttention-2 unsupported error; not the same code-path problem." + }, + { + "left": "issue:44242", + "right": "issue:46894", + "accept": false, + "reason": "Load-balancing loss omission and ASR double-shift bug are distinct training issues." + }, + { + "left": "issue:43065", + "right": "issue:44617", + "accept": false, + "reason": "Dummy Conv2d in Sam3PixelDecoder vs Sam3Video CUDA OOM; unrelated Sam3 problems." + }, + { + "left": "issue:43540", + "right": "issue:44514", + "accept": false, + "reason": "Qwen3OmniMoe video-processing ValueError and Qwen2_5_VLProcessor batching crash are different processors and failures." + }, + { + "left": "issue:44610", + "right": "issue:46369", + "accept": false, + "reason": "OmDet-Turbo input-size mismatch vs sam3-litetext memory overflow; not the same bug." + }, + { + "left": "issue:45632", + "right": "issue:46424", + "accept": false, + "reason": "Remote-code cache-path collision vs max_cache_length only honored for Quantized Cache; separate cache bugs." + }, + { + "left": "issue:43572", + "right": "issue:46178", + "accept": false, + "reason": "Missing StableLmConfig pad_token_idx and tensorflow_text backend mapping error are unrelated." + }, + { + "left": "issue:45542", + "right": "issue:46234", + "accept": false, + "reason": "TensorBoard-without-TF backend error and missing Falcon Mamba availability helper are different dependency issues." + }, + { + "left": "issue:45803", + "right": "issue:46170", + "accept": false, + "reason": "Swallowed KeyboardInterrupt in Fuyu conversion vs Windows ROCm torch.distributed import failure; unrelated." + }, + { + "left": "issue:44792", + "right": "issue:45600", + "accept": false, + "reason": "Janus test failure vs stale auto_mappings reference to removed Sam3LiteText configs; different code paths." + }, + { + "left": "issue:44792", + "right": "issue:44977", + "accept": false, + "reason": "Janus generation test failure and Qwen3.5 flash-attention behavior are unrelated model-specific bugs." + }, + { + "left": "issue:43504", + "right": "issue:45698", + "accept": false, + "reason": "Legacy preset field in BEiT segmentation and wrong custom module after save_pretrained are both load-related but not the same defect." + }, + { + "left": "issue:43650", + "right": "issue:46202", + "accept": false, + "reason": "Placeholder issue text and Gemma pipeline roundtrip failure are clearly unrelated." + }, + { + "left": "issue:43352", + "right": "issue:46640", + "accept": false, + "reason": "Specific unsupported FlashAttention-2 error for NemotronH vs a generic open Flash Attention 2 issue; not the same concrete bug." + }, + { + "left": "issue:45127", + "right": "issue:45200", + "accept": false, + "reason": "LoRA merge collapse with tied embeddings and Gemma 4 mm_token_type_ids defaulting are different model behaviors." + } + ] + }, + "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": 27, + "review_unit_index": 38, + "review_unit_count": 63, + "status": "reviewed", + "reason": null, + "source": "llm", + "cache_hit": false, + "model": "gpt-5.4-mini?service_tier=flex", + "cluster_id": "cluster-10000-514", + "nodes": [ + "issue:25251", + "issue:27957", + "issue:29870", + "issue:35141", + "issue:41950", + "issue:42490", + "issue:43335", + "issue:43388", + "issue:43475", + "issue:43525", + "issue:43531", + "issue:43550", + "issue:43644", + "issue:43673", + "issue:43688", + "issue:43827", + "issue:43931", + "issue:43994", + "issue:44060", + "issue:44077", + "issue:44190", + "issue:44242", + "issue:44534", + "issue:44617", + "issue:44734", + "issue:44743", + "issue:44855", + "issue:44869", + "issue:44877", + "issue:44913", + "issue:44964", + "issue:45072", + "issue:45245", + "issue:45381", + "issue:45406", + "issue:45563", + "issue:45584", + "issue:45588", + "issue:45632", + "issue:46082", + "issue:46369", + "issue:46421", + "issue:46617", + "issue:46697", + "issue:46832", + "issue:46858", + "issue:46897", + "issue:46932" + ], + "soft_pairs": [ + "issue:43525|issue:43531", + "issue:44534|issue:44913", + "issue:41950|issue:43475", + "issue:43673|issue:44743", + "issue:35141|issue:42490", + "issue:43388|issue:43688", + "issue:44190|issue:44964", + "issue:44855|issue:45588", + "issue:29870|issue:35141", + "issue:44869|issue:46832", + "issue:45632|issue:46617", + "issue:45072|issue:46932", + "issue:46082|issue:46697", + "issue:45245|issue:46858", + "issue:44734|issue:45406", + "issue:43827|issue:43994", + "issue:43644|issue:44913", + "issue:44617|issue:45584", + "issue:25251|issue:27957", + "issue:43335|issue:44855", + "issue:43931|issue:44060", + "issue:44242|issue:46897", + "issue:44077|issue:44877", + "issue:43673|issue:46421", + "issue:45381|issue:46369", + "issue:45563|issue:45584", + "issue:43550|issue:44190" + ], + "prepared_review_unit_hash": "3fbf127c1021af26959296009c24d2565a289eabc2b529bbc65224a561840fd5", + "component_budget": { + "node_count": 514, + "item_count": 514, + "soft_pair_count": 1891, + "serialized_chars": 980895, + "estimated_input_tokens": 245224, + "estimated_eval_tokens": 490704 + }, + "budget": { + "node_count": 48, + "item_count": 48, + "soft_pair_count": 27, + "serialized_chars": 36027, + "estimated_input_tokens": 9007, + "estimated_eval_tokens": 18270 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": true, + "analyst_result": { + "summary": "This set is heterogeneous: the items touch unrelated models, pipelines, cache paths, config behaviors, and decoding bugs. None of the soft pairs look like the same underlying issue/change; most only share a broad subsystem keyword.", + "confidence": 0.96, + "canonical_issue_reason": "No canonical issue emerges: the items cover many unrelated bugs across different models, pipelines, and subsystems.", + "canonical_pr_reason": null, + "best_issue_reason": "No single issue is a better global representative because there is no coherent duplicate cluster here.", + "best_pr_reason": null, + "soft_edge_verdicts": [ + { + "left": "issue:43525", + "right": "issue:43531", + "accept": false, + "reason": "Different models and failures: Llama4 pad_token_id attribute error vs Qwen3-MoE sliding_window behavior." + }, + { + "left": "issue:44534", + "right": "issue:44913", + "accept": false, + "reason": "Both are config/runtime regressions in Transformers v5, but they concern unrelated mechanics: non-persistent buffers vs GPTNeoX rotary_pct reload." + }, + { + "left": "issue:41950", + "right": "issue:43475", + "accept": false, + "reason": "A pipeline processor-selection bug is not the same as a Sam3VisionEncoderOutput missing-attribute error." + }, + { + "left": "issue:43673", + "right": "issue:44743", + "accept": false, + "reason": "Both involve cache handling, but one is missing GenerationMixin cache in chunked prefill while the other resets recurrent states in Qwen3 modular code." + }, + { + "left": "issue:35141", + "right": "issue:42490", + "accept": false, + "reason": "Completely different areas: tied embeddings/post_init reinitialization vs local-path import warning." + }, + { + "left": "issue:43388", + "right": "issue:43688", + "accept": false, + "reason": "Different training/eval bugs: gather_for_metrics tuple truncation vs auxiliary loss normalization in MoE models." + }, + { + "left": "issue:44190", + "right": "issue:44964", + "accept": false, + "reason": "Unrelated failures: a local dataset-loading script issue vs a multimodal model loading problem." + }, + { + "left": "issue:44855", + "right": "issue:45588", + "accept": false, + "reason": "Python 3.13 decorator parsing/import error is unrelated to a flash_attention crash on s_aux=None." + }, + { + "left": "issue:29870", + "right": "issue:35141", + "accept": false, + "reason": "Bart generate() decoder_attention_mask initialization bug is unrelated to embedding reinitialization in post_init." + }, + { + "left": "issue:44869", + "right": "issue:46832", + "accept": false, + "reason": "Whisper trailing-replacement-character decoding crash and RT-DETR feature-level mismatch are different code paths." + }, + { + "left": "issue:45632", + "right": "issue:46617", + "accept": false, + "reason": "Both concern trust_remote_code/cache, but one is a leaf-directory cache collision and the other is a FileNotFoundError from cache loading." + }, + { + "left": "issue:45072", + "right": "issue:46932", + "accept": false, + "reason": "Different model families and causes: bfloat16 dtype mismatches in inference vs quantized Gemma4 vision/audio embedder casting." + }, + { + "left": "issue:46082", + "right": "issue:46697", + "accept": false, + "reason": "LlamaConfig head_dim validation and Swin2SR over-padding are unrelated." + }, + { + "left": "issue:45245", + "right": "issue:46858", + "accept": false, + "reason": "Category-count overflow and cache.get_seq_length() returning a device tensor are unrelated bugs." + }, + { + "left": "issue:44734", + "right": "issue:45406", + "accept": false, + "reason": "KV-cache continuation indexing crash and Gemma4Processor missing _tokenizer are different serve failures." + }, + { + "left": "issue:43827", + "right": "issue:43994", + "accept": false, + "reason": "Docs still referencing removed pipeline APIs is not the same as a model producing bad results through AutoModel/pipeline usage." + }, + { + "left": "issue:43644", + "right": "issue:44913", + "accept": false, + "reason": "Non-persistent buffer initialization junk and GPTNeoXConfig rotary_pct reload regression are distinct issues." + }, + { + "left": "issue:44617", + "right": "issue:45584", + "accept": false, + "reason": "CUDA OOM in Sam3Video is unrelated to Whisper empty-transcription generation failure." + }, + { + "left": "issue:25251", + "right": "issue:27957", + "accept": false, + "reason": "A pipeline output-shape change and XLMRoberta Flash Attention 2 support are unrelated." + }, + { + "left": "issue:43335", + "right": "issue:44855", + "accept": false, + "reason": "SwitchTransformers sparse-layer construction and DebertaV2 import parsing are unrelated." + }, + { + "left": "issue:43931", + "right": "issue:44060", + "accept": false, + "reason": "Both mention Qwen models, but weight-shape loading errors and tied-weights warning misbinding are different bugs." + }, + { + "left": "issue:44242", + "right": "issue:46897", + "accept": false, + "reason": "MoE load-balancing loss gating and Florence2 training-loss double-shift are not the same bug." + }, + { + "left": "issue:44077", + "right": "issue:44877", + "accept": false, + "reason": "A model-specific post_init policy issue is unrelated to strict config blocking granite_speech loading." + }, + { + "left": "issue:43673", + "right": "issue:46421", + "accept": false, + "reason": "Both involve chunked prefill/cache, but one is missing cache state and the other is spurious recompilation from lazy cache init." + }, + { + "left": "issue:45381", + "right": "issue:46369", + "accept": false, + "reason": "Qwen2.5-VL video position IDs and Sam3-litetext memory overflow are unrelated." + }, + { + "left": "issue:45563", + "right": "issue:45584", + "accept": false, + "reason": "Paged-generate stale warning and Whisper empty-transcription failure are different generation bugs." + }, + { + "left": "issue:43550", + "right": "issue:44190", + "accept": false, + "reason": "Bamba torch.compile+SDPA failure is unrelated to the local dataset-loading script 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": 27, + "review_unit_index": 39, + "review_unit_count": 63, + "status": "reviewed", + "reason": null, + "source": "llm", + "cache_hit": false, + "model": "gpt-5.4-mini?service_tier=flex", + "cluster_id": "cluster-10000-514", + "nodes": [ + "issue:25251", + "issue:29870", + "issue:39692", + "issue:41720", + "issue:43065", + "issue:43295", + "issue:43335", + "issue:43526", + "issue:43540", + "issue:43643", + "issue:43854", + "issue:44112", + "issue:44190", + "issue:44246", + "issue:44462", + "issue:44464", + "issue:44479", + "issue:44496", + "issue:44521", + "issue:44610", + "issue:44617", + "issue:44734", + "issue:44779", + "issue:44792", + "issue:44933", + "issue:44938", + "issue:44964", + "issue:45003", + "issue:45127", + "issue:45245", + "issue:45276", + "issue:45290", + "issue:45357", + "issue:45362", + "issue:45381", + "issue:45542", + "issue:45563", + "issue:45584", + "issue:45588", + "issue:45600", + "issue:45907", + "issue:46369", + "issue:46489", + "issue:46655", + "issue:46697", + "issue:46752", + "issue:47013" + ], + "soft_pairs": [ + "issue:41720|issue:45245", + "issue:44933|issue:45362", + "issue:43065|issue:44610", + "issue:43335|issue:44190", + "issue:45588|issue:47013", + "issue:44479|issue:44734", + "issue:43854|issue:44792", + "issue:45381|issue:46655", + "issue:44617|issue:46369", + "issue:39692|issue:44190", + "issue:45127|issue:45357", + "issue:43295|issue:45276", + "issue:44464|issue:45563", + "issue:44779|issue:46489", + "issue:45276|issue:45584", + "issue:45290|issue:46752", + "issue:43540|issue:44734", + "issue:44112|issue:45600", + "issue:25251|issue:29870", + "issue:44521|issue:45381", + "issue:44462|issue:45907", + "issue:43643|issue:44496", + "issue:43526|issue:46697", + "issue:44938|issue:45542", + "issue:43854|issue:44964", + "issue:44246|issue:45003" + ], + "prepared_review_unit_hash": "b9e35787215526ca67c2fc01e77a0e3e09fd9482a1acec614a47efe659504f83", + "component_budget": { + "node_count": 514, + "item_count": 514, + "soft_pair_count": 1891, + "serialized_chars": 980895, + "estimated_input_tokens": 245224, + "estimated_eval_tokens": 490704 + }, + "budget": { + "node_count": 47, + "item_count": 47, + "soft_pair_count": 26, + "serialized_chars": 35009, + "estimated_input_tokens": 8753, + "estimated_eval_tokens": 17762 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": true, + "analyst_result": { + "summary": "Most pairs are clearly unrelated despite some superficial overlap in subsystems. The only strong duplicate-like pair is the DeepSeek tokenizer regression reports, which describe the same v5+ tokenizer breakage.", + "confidence": 0.89, + "canonical_issue_reason": "issue:46489 is the best canonical anchor because it is the most specific DeepSeek tokenizer regression report, explicitly ties the problem to v5+, and notes it fills a gap in PR #44801\u2019s fix.", + "canonical_pr_reason": null, + "best_issue_reason": "issue:46489 is the strongest issue to represent the cluster\u2019s one real duplicate pair: concrete failure, clear affected scope, and direct linkage to the prior fix.", + "best_pr_reason": null, + "soft_edge_verdicts": [ + { + "left": "issue:41720", + "right": "issue:45245", + "accept": false, + "reason": "Different failures and subsystems: Qwen3 device-mapping CUDA assert vs category-count overflow." + }, + { + "left": "issue:44933", + "right": "issue:45362", + "accept": false, + "reason": "Unrelated: missing import in image_utils vs Qwen3.5 chat crash." + }, + { + "left": "issue:43065", + "right": "issue:44610", + "accept": false, + "reason": "Both involve vision models, but one is a dummy Conv2d in Sam3PixelDecoder and the other is an OmDet-Turbo processor/model image-size mismatch." + }, + { + "left": "issue:43335", + "right": "issue:44190", + "accept": false, + "reason": "Different code paths: SwitchTransformers config layer construction vs local dataset loading in an example script." + }, + { + "left": "issue:45588", + "right": "issue:47013", + "accept": false, + "reason": "Different bugs: flash-attention sinkless-model s_aux handling vs xLSTM train-mode unpacking." + }, + { + "left": "issue:44479", + "right": "issue:44734", + "accept": false, + "reason": "Both are generation bugs, but one is a video-input regression in Qwen VL variants and the other is a serve KV-cache indexing crash." + }, + { + "left": "issue:43854", + "right": "issue:44792", + "accept": false, + "reason": "Different models and failures: GLM-4.7 load/test issue vs Janus image-generation test failure." + }, + { + "left": "issue:45381", + "right": "issue:46655", + "accept": false, + "reason": "Both are video-processing issues, but the reported symptoms and models differ (vision_position_ids vs padding calculation)." + }, + { + "left": "issue:44617", + "right": "issue:46369", + "accept": false, + "reason": "Both mention memory pressure in Sam3 variants, but one is Sam3Video CUDA OOM and the other is Sam3-LiteText memory overflow; not clearly the same bug." + }, + { + "left": "issue:39692", + "right": "issue:44190", + "accept": false, + "reason": "Doc-example model/processor mismatch and quantization problems are unrelated to local dataset loading in an example script." + }, + { + "left": "issue:45127", + "right": "issue:45357", + "accept": false, + "reason": "Different issues: LoRA merge with extended vocab causing collapse vs incorrect visual encoder keys saved by Qwen3.5." + }, + { + "left": "issue:43295", + "right": "issue:45276", + "accept": false, + "reason": "Processor/tokenizer API regression is not the same as gemma4 resize_token_embeddings not propagating to embeddings." + }, + { + "left": "issue:44462", + "right": "issue:45563", + "accept": false, + "reason": "Tokenizer-file precedence bug is unrelated to a stale generate warning about num_return_sequences." + }, + { + "left": "issue:44779", + "right": "issue:46489", + "accept": true, + "reason": "Same underlying bug: DeepSeek tokenizer output is wrong on v5/v5+ and works on v4; 46489 is a more specific report of the same regression with an explicit gap in PR #44801\u2019s fix." + }, + { + "left": "issue:45276", + "right": "issue:45584", + "accept": false, + "reason": "Resize-token-embedding propagation bug is unrelated to Whisper empty-transcription handling after align_special_tokens." + }, + { + "left": "issue:45290", + "right": "issue:46752", + "accept": false, + "reason": "Both hit apply_chat_template edge cases, but one is assistant tool-call/no-content handling and the other is an empty-list crash; too different to merge." + }, + { + "left": "issue:43540", + "right": "issue:44734", + "accept": false, + "reason": "Different failures: Qwen3OmniMoe video input ValueError vs serve KV-cache continuation tensor indexing." + }, + { + "left": "issue:44112", + "right": "issue:45600", + "accept": false, + "reason": "Both are CI-related, but the stale GraniteSpeech device-override test and removed Sam3LiteText auto-mapping configs are different problems." + }, + { + "left": "issue:25251", + "right": "issue:29870", + "accept": false, + "reason": "Unrelated: pipeline top_k nesting change vs Bart generate decoder_attention_mask assignment error." + }, + { + "left": "issue:44521", + "right": "issue:45381", + "accept": false, + "reason": "Both involve multimodal chat/video inputs, but one is assistant mask generation and the other is video position IDs." + }, + { + "left": "issue:44462", + "right": "issue:45907", + "accept": false, + "reason": "Tokenizer resolution bug is unrelated to list_repo_files not forwarding revision in a cookbook CI path." + }, + { + "left": "issue:43643", + "right": "issue:44496", + "accept": false, + "reason": "Both are config-loading failures, but missing fields from trust_remote_code and missing model_type are not the same bug." + }, + { + "left": "issue:43526", + "right": "issue:46697", + "accept": false, + "reason": "Different image-processor bugs: reduce_labels truncation vs over-padding when dimensions are already divisible." + }, + { + "left": "issue:44938", + "right": "issue:45542", + "accept": false, + "reason": "Unrelated import/load failures: Python 3.14 compatibility vs tensorboard-only undefined tf backend." + }, + { + "left": "issue:43854", + "right": "issue:44964", + "accept": false, + "reason": "Both are model-loading issues, but they affect different models and failure modes." + }, + { + "left": "issue:44246", + "right": "issue:45003", + "accept": false, + "reason": "Import slowness and unsafe sys.modules access are not the same concrete bug." + } + ] + }, + "evaluator_result": { + "accept": true, + "feedback": "Grounded overall. The only clearly duplicate-like pair in the packet is issue:44779 \u2194 issue:46489, and the other soft-edge rejections stay conservative by distinguishing different models, code paths, or failure modes. The canonical anchor choice is also 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": 27, + "review_unit_index": 40, + "review_unit_count": 63, + "status": "reviewed", + "reason": null, + "source": "llm", + "cache_hit": false, + "model": "gpt-5.4-mini?service_tier=flex", + "cluster_id": "cluster-10000-514", + "nodes": [ + "issue:27957", + "issue:32090", + "issue:35141", + "issue:36246", + "issue:38617", + "issue:41628", + "issue:41950", + "issue:42175", + "issue:42796", + "issue:42907", + "issue:43295", + "issue:43404", + "issue:43441", + "issue:43526", + "issue:43531", + "issue:43582", + "issue:43673", + "issue:43901", + "issue:43976", + "issue:44016", + "issue:44112", + "issue:44220", + "issue:44279", + "issue:44295", + "issue:44336", + "issue:44496", + "issue:44521", + "issue:44683", + "issue:44737", + "issue:44964", + "issue:44987", + "issue:45042", + "issue:45083", + "issue:45216", + "issue:45335", + "issue:45446", + "issue:45600", + "issue:45632", + "issue:45759", + "issue:45803", + "issue:46082", + "issue:46121", + "issue:46234", + "issue:46424", + "issue:46506", + "issue:46858", + "issue:46932" + ], + "soft_pairs": [ + "issue:43673|issue:46858", + "issue:44737|issue:45803", + "issue:27957|issue:35141", + "issue:44964|issue:45600", + "issue:44112|issue:44964", + "issue:42175|issue:44279", + "issue:45632|issue:46234", + "issue:42796|issue:44521", + "issue:32090|issue:35141", + "issue:36246|issue:41628", + "issue:36246|issue:38617", + "issue:43404|issue:45216", + "issue:43531|issue:45600", + "issue:44295|issue:46506", + "issue:46082|issue:46121", + "issue:41950|issue:44016", + "issue:43673|issue:46424", + "issue:43295|issue:45335", + "issue:45083|issue:46506", + "issue:44496|issue:45042", + "issue:43976|issue:44987", + "issue:43582|issue:46932", + "issue:43526|issue:43901", + "issue:43673|issue:44336", + "issue:43901|issue:44496", + "issue:44683|issue:45446", + "issue:44496|issue:45759", + "issue:44016|issue:44220", + "issue:42907|issue:43441" + ], + "prepared_review_unit_hash": "04924de299411dc032dc9040dc1cccec6dad6766b6d8acf1958880b34870142a", + "component_budget": { + "node_count": 514, + "item_count": 514, + "soft_pair_count": 1891, + "serialized_chars": 980895, + "estimated_input_tokens": 245224, + "estimated_eval_tokens": 490704 + }, + "budget": { + "node_count": 47, + "item_count": 47, + "soft_pair_count": 29, + "serialized_chars": 36050, + "estimated_input_tokens": 9013, + "estimated_eval_tokens": 18282 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": true, + "analyst_result": { + "summary": "These issues span many unrelated Transformers bugs and regressions; none of the proposed soft pairs looks like the same underlying defect, so all candidate edges should be rejected.", + "confidence": 0.94, + "canonical_issue_reason": null, + "canonical_pr_reason": null, + "best_issue_reason": "If a representative issue is needed, #44521 is the strongest cluster anchor: it is open, actively discussed, has inbound references, and concerns a broadly impactful multimodal API bug.", + "best_pr_reason": null, + "soft_edge_verdicts": [ + { + "left": "issue:43673", + "right": "issue:46858", + "accept": false, + "reason": "Both involve cache behavior, but one is about a missing cache in chunked prefill and the other is about a device-tensor seq_length breaking arithmetic in static cache; different bugs." + }, + { + "left": "issue:44737", + "right": "issue:45803", + "accept": false, + "reason": "XLNet CPU placement bug and Fuyu exception-swallowing are unrelated code paths and failure modes." + }, + { + "left": "issue:27957", + "right": "issue:35141", + "accept": false, + "reason": "Flash Attention 2 support for XLM-Roberta is unrelated to token embedding reinitialization when tie_word_embedding is false." + }, + { + "left": "issue:44964", + "right": "issue:45600", + "accept": false, + "reason": "Phi-4 multimodal loading failure and stale auto_mappings references are different issues in different loader paths." + }, + { + "left": "issue:44112", + "right": "issue:44964", + "accept": false, + "reason": "A GraniteSpeech CI test failure is not the same as a Phi-4 multimodal model loading regression." + }, + { + "left": "issue:42175", + "right": "issue:44279", + "accept": false, + "reason": "Package extras/backend dependency complaints are too broad here; these reports do not describe the same concrete dependency bug." + }, + { + "left": "issue:45632", + "right": "issue:46234", + "accept": false, + "reason": "A local trust_remote_code cache-path collision is unrelated to a missing import_utils availability helper." + }, + { + "left": "issue:42796", + "right": "issue:44521", + "accept": false, + "reason": "Tokenizer BPE-marker handling and multimodal assistant-mask generation are different tokenizer behaviors." + }, + { + "left": "issue:32090", + "right": "issue:35141", + "accept": false, + "reason": "Trainer GPU broadcast NoneType error and embedding resize/post_init reinitialization are unrelated." + }, + { + "left": "issue:36246", + "right": "issue:41628", + "accept": false, + "reason": "Both are import errors, but for different symbols and different package surfaces; they are not the same missing-export bug." + }, + { + "left": "issue:36246", + "right": "issue:38617", + "accept": false, + "reason": "Different missing imports from different modules; no shared underlying fix." + }, + { + "left": "issue:43404", + "right": "issue:45216", + "accept": false, + "reason": "Mistral3 lm_head tying is unrelated to the Qwen3.5 save_pretrained checkpoint regression." + }, + { + "left": "issue:43531", + "right": "issue:45600", + "accept": false, + "reason": "Qwen3-MoE sliding-window behavior and stale auto_mappings config references are separate bugs." + }, + { + "left": "issue:44295", + "right": "issue:46506", + "accept": false, + "reason": "Reading a registered buffer incorrectly and CPU/CUDA output divergence are not the same defect." + }, + { + "left": "issue:46082", + "right": "issue:46121", + "accept": false, + "reason": "Llama head_dim validation and rope-parameter JSON-list handling are distinct config-validation problems." + }, + { + "left": "issue:41950", + "right": "issue:44016", + "accept": false, + "reason": "Video-classification processor lookup and a notebook syntax error are unrelated." + }, + { + "left": "issue:43673", + "right": "issue:46424", + "accept": false, + "reason": "Generation cache missing in chunked prefill and max_cache_length being honored only for quantized cache are different cache semantics." + }, + { + "left": "issue:43295", + "right": "issue:45335", + "accept": false, + "reason": "Processor.tokenizer regression and resize_token_embeddings not updating decoder.embed_tokens are separate model/processor bugs." + }, + { + "left": "issue:45083", + "right": "issue:46506", + "accept": false, + "reason": "A helper returning unexpected lengths in qwen3_omni_moe is unrelated to PPDocLayoutV3 CPU/CUDA divergence." + }, + { + "left": "issue:44496", + "right": "issue:45042", + "accept": false, + "reason": "Unrecognized-model config loading and PIL backend requiring torchvision are different failure modes." + }, + { + "left": "issue:43976", + "right": "issue:44987", + "accept": false, + "reason": "Python version compatibility and loading a specific model checkpoint are unrelated." + }, + { + "left": "issue:43582", + "right": "issue:46932", + "accept": false, + "reason": "Apple Silicon warmup TypeError and Gemma4 quantized-embedder dtype casting are separate issues." + }, + { + "left": "issue:43526", + "right": "issue:43901", + "accept": false, + "reason": "BeitImageProcessorFast label reduction and TextClassificationPipeline docs mismatch are not the same bug." + }, + { + "left": "issue:43673", + "right": "issue:44336", + "accept": false, + "reason": "Generation cache behavior and ANSI codes emitted by loading_report are unrelated." + }, + { + "left": "issue:43901", + "right": "issue:44496", + "accept": false, + "reason": "A docs/behavior mismatch for return_all_scores is unrelated to an unrecognized-model config error." + }, + { + "left": "issue:44683", + "right": "issue:45446", + "accept": false, + "reason": "Both touch flex_attention, but one is a compile failure on torch>=2.9 and the other is an incorrect version check for AuxRequest import; different concrete problems." + }, + { + "left": "issue:44496", + "right": "issue:45759", + "accept": false, + "reason": "Unrecognized model config and from_config failing to unwrap text_config are different model-loading issues." + }, + { + "left": "issue:44016", + "right": "issue:44220", + "accept": false, + "reason": "A notebook syntax error and a feature-extraction helper issue are unrelated." + }, + { + "left": "issue:42907", + "right": "issue:43441", + "accept": false, + "reason": "Saving dequantized Ministral/Devstral models and FlashAttention failures in Ministral are separate 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": 27, + "review_unit_index": 41, + "review_unit_count": 63, + "status": "reviewed", + "reason": null, + "source": "llm", + "cache_hit": false, + "model": "gpt-5.4-mini?service_tier=flex", + "cluster_id": "cluster-10000-514", + "nodes": [ + "issue:39692", + "issue:41950", + "issue:42796", + "issue:43295", + "issue:43299", + "issue:43381", + "issue:43441", + "issue:43479", + "issue:43577", + "issue:43650", + "issue:43824", + "issue:43901", + "issue:43976", + "issue:44016", + "issue:44062", + "issue:44075", + "issue:44220", + "issue:44279", + "issue:44327", + "issue:44351", + "issue:44368", + "issue:44466", + "issue:44521", + "issue:44779", + "issue:44792", + "issue:44933", + "issue:45020", + "issue:45216", + "issue:45276", + "issue:45278", + "issue:45335", + "issue:45375", + "issue:45405", + "issue:45584", + "issue:45588", + "issue:45593", + "issue:45812", + "issue:46082", + "issue:46123", + "issue:46369", + "issue:46399", + "issue:46424", + "issue:46515", + "issue:46614", + "issue:46693", + "issue:46752", + "issue:47000", + "issue:47085" + ], + "soft_pairs": [ + "issue:43381|issue:45593", + "issue:43650|issue:44220", + "issue:44016|issue:44062", + "issue:44792|issue:46424", + "issue:43976|issue:44279", + "issue:43441|issue:45588", + "issue:43824|issue:44933", + "issue:44521|issue:46693", + "issue:44279|issue:46515", + "issue:43479|issue:45375", + "issue:45335|issue:45584", + "issue:43299|issue:45216", + "issue:46123|issue:46369", + "issue:44327|issue:46752", + "issue:46614|issue:47085", + "issue:44279|issue:45405", + "issue:43295|issue:44779", + "issue:39692|issue:43901", + "issue:42796|issue:43650", + "issue:44075|issue:44368", + "issue:43577|issue:45216", + "issue:44351|issue:45278", + "issue:41950|issue:45020", + "issue:43976|issue:45405", + "issue:44779|issue:45812", + "issue:46082|issue:46399", + "issue:43381|issue:47000", + "issue:44466|issue:45276" + ], + "prepared_review_unit_hash": "a4f346ee05595b957b3282a43119b39f96547ba812b68d12c137a1afc22ee581", + "component_budget": { + "node_count": 514, + "item_count": 514, + "soft_pair_count": 1891, + "serialized_chars": 980895, + "estimated_input_tokens": 245224, + "estimated_eval_tokens": 490704 + }, + "budget": { + "node_count": 48, + "item_count": 48, + "soft_pair_count": 28, + "serialized_chars": 36041, + "estimated_input_tokens": 9011, + "estimated_eval_tokens": 18278 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": true, + "analyst_result": { + "summary": "Mostly unrelated issue reports across different Transformer subsystems. The cluster is dominated by v5 regressions, but the soft pairs are generally distinct enough that none should be merged as duplicates.", + "confidence": 0.43, + "canonical_issue_reason": "Issue 45278 is the broadest umbrella-style report in the set, covering widespread breakage after a version upgrade; it is more suitable as the cluster anchor than the narrower model- or tokenizer-specific bugs.", + "canonical_pr_reason": null, + "best_issue_reason": "Issue 45278 best represents the overall cluster because it summarizes the common theme of post-upgrade regressions, while the others are mostly specific manifestations in single models, tokenizers, or pipelines.", + "best_pr_reason": null, + "soft_edge_verdicts": [ + { + "left": "issue:43381", + "right": "issue:45593", + "accept": false, + "reason": "Different bugs: eval-mode gradient checkpointing behavior vs D-FINE auxiliary-loss handling." + }, + { + "left": "issue:43650", + "right": "issue:44220", + "accept": false, + "reason": "No clear overlap; one is a vague data request, the other is an fbank feature extraction issue." + }, + { + "left": "issue:44016", + "right": "issue:44062", + "accept": false, + "reason": "Unrelated failures: notebook syntax error vs AddedToken keyword argument conflict." + }, + { + "left": "issue:44792", + "right": "issue:46424", + "accept": false, + "reason": "Different code paths: janus test failure vs cache max length handling." + }, + { + "left": "issue:43976", + "right": "issue:44279", + "accept": false, + "reason": "One is a Python version compatibility bug; the other is a generic dependency issue." + }, + { + "left": "issue:43441", + "right": "issue:45588", + "accept": false, + "reason": "Both mention flash attention, but they are different model-specific crashes with different root causes." + }, + { + "left": "issue:43824", + "right": "issue:44933", + "accept": false, + "reason": "Different import problems for different modules; not the same missing symbol." + }, + { + "left": "issue:44521", + "right": "issue:46693", + "accept": false, + "reason": "Multimodal chat-template masking bug is unrelated to the FA2 long-context performance regression." + }, + { + "left": "issue:44279", + "right": "issue:46515", + "accept": false, + "reason": "Both are dependency-related, but they concern different external packages and failures." + }, + { + "left": "issue:43479", + "right": "issue:45375", + "accept": false, + "reason": "Different config classes and different missing/defaulted fields." + }, + { + "left": "issue:45335", + "right": "issue:45584", + "accept": false, + "reason": "Resize-token-embeddings propagation bug is unrelated to Whisper empty-transcription generation failure." + }, + { + "left": "issue:43299", + "right": "issue:45216", + "accept": false, + "reason": "Both are Qwen-related regressions, but one is loading failure and the other is checkpoint saving behavior." + }, + { + "left": "issue:46123", + "right": "issue:46369", + "accept": false, + "reason": "Pipeline batch-tracking bug and SAM memory overflow are unrelated." + }, + { + "left": "issue:44327", + "right": "issue:46752", + "accept": false, + "reason": "Different crashes in different APIs: QA span decoding vs empty chat-template input." + }, + { + "left": "issue:46614", + "right": "issue:47085", + "accept": false, + "reason": "Tokenizer backend language-state bug is unrelated to tokenizer thread-safety regression." + }, + { + "left": "issue:44279", + "right": "issue:45405", + "accept": false, + "reason": "Generic dependency issue vs a specific PEFT minimum-version bump; not the same bug." + }, + { + "left": "issue:43295", + "right": "issue:44779", + "accept": false, + "reason": "Both are tokenizer regressions, but they affect different APIs and models with different failure modes." + }, + { + "left": "issue:39692", + "right": "issue:43901", + "accept": false, + "reason": "Both are documentation-related, but they describe different examples and different behaviors." + }, + { + "left": "issue:42796", + "right": "issue:43650", + "accept": false, + "reason": "Tokenizer BPE-marker bug is unrelated to the data request." + }, + { + "left": "issue:44075", + "right": "issue:44368", + "accept": false, + "reason": "SGD optimizer args not being used is unrelated to the tie_word_embeddings warning." + }, + { + "left": "issue:43577", + "right": "issue:45216", + "accept": false, + "reason": "Blip2 dtype propagation bug is unrelated to Qwen3.5 checkpoint serialization." + }, + { + "left": "issue:44351", + "right": "issue:45278", + "accept": false, + "reason": "45278 is a broad import-regression report, but 44351 is a single specific missing import and not clearly the same underlying issue." + }, + { + "left": "issue:41950", + "right": "issue:45020", + "accept": false, + "reason": "Video-classification processor lookup bug is different from remote_code model breakage." + }, + { + "left": "issue:43976", + "right": "issue:45405", + "accept": false, + "reason": "Python support regression vs unreleased PEFT dependency bump; unrelated." + }, + { + "left": "issue:44779", + "right": "issue:45812", + "accept": false, + "reason": "Different tokenizer regressions affecting different model families and different symptoms." + }, + { + "left": "issue:46082", + "right": "issue:46399", + "accept": false, + "reason": "LlamaConfig validation issue is unrelated to the Qwen3.5 rotary-pos-emb registration crash." + }, + { + "left": "issue:43381", + "right": "issue:47000", + "accept": false, + "reason": "Gradient checkpointing eval-mode issue is unrelated to RTDetr loss infeasibility." + }, + { + "left": "issue:44466", + "right": "issue:45276", + "accept": false, + "reason": "Both touch tied/embedding weights, but one is serialization and the other is resize propagation; 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": 27, + "review_unit_index": 42, + "review_unit_count": 63, + "status": "reviewed", + "reason": null, + "source": "llm", + "cache_hit": false, + "model": "gpt-5.4-mini?service_tier=flex", + "cluster_id": "cluster-10000-514", + "nodes": [ + "issue:33290", + "issue:41950", + "issue:42175", + "issue:42994", + "issue:43299", + "issue:43317", + "issue:43441", + "issue:43479", + "issue:43525", + "issue:43644", + "issue:43650", + "issue:43653", + "issue:43854", + "issue:43873", + "issue:43901", + "issue:43931", + "issue:43976", + "issue:44016", + "issue:44242", + "issue:44246", + "issue:44261", + "issue:44263", + "issue:44291", + "issue:44336", + "issue:44462", + "issue:44623", + "issue:44683", + "issue:44749", + "issue:45081", + "issue:45103", + "issue:45200", + "issue:45357", + "issue:45372", + "issue:45397", + "issue:45542", + "issue:45588", + "issue:45593", + "issue:45656", + "issue:45823", + "issue:45834", + "issue:45859", + "issue:45907", + "issue:46082", + "issue:46209", + "issue:46512", + "issue:46851", + "issue:47000", + "issue:47145" + ], + "soft_pairs": [ + "issue:43299|issue:45357", + "issue:46082|issue:46512", + "issue:43653|issue:45593", + "issue:41950|issue:45823", + "issue:43650|issue:46209", + "issue:44242|issue:45200", + "issue:43644|issue:47145", + "issue:43299|issue:43854", + "issue:44462|issue:45081", + "issue:43931|issue:47145", + "issue:44683|issue:45542", + "issue:45588|issue:47000", + "issue:44016|issue:44291", + "issue:43479|issue:43525", + "issue:45081|issue:45907", + "issue:33290|issue:45656", + "issue:42994|issue:43317", + "issue:44336|issue:46851", + "issue:43441|issue:45834", + "issue:44623|issue:44749", + "issue:43873|issue:47145", + "issue:45103|issue:45372", + "issue:45859|issue:46851", + "issue:44261|issue:44263", + "issue:44623|issue:45397", + "issue:42175|issue:43976", + "issue:43901|issue:44246" + ], + "prepared_review_unit_hash": "b21a9762f9e0570fc3f3fd298ce5365ba5de7f3db77c19addc56c65e9dbd5557", + "component_budget": { + "node_count": 514, + "item_count": 514, + "soft_pair_count": 1891, + "serialized_chars": 980895, + "estimated_input_tokens": 245224, + "estimated_eval_tokens": 490704 + }, + "budget": { + "node_count": 48, + "item_count": 48, + "soft_pair_count": 27, + "serialized_chars": 35740, + "estimated_input_tokens": 8935, + "estimated_eval_tokens": 18126 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": true, + "analyst_result": { + "summary": "This is a loose cluster of mostly unrelated issue reports; the nearest shared theme is loading/offload/quantization regressions. I would use issue #43873 as the canonical issue because it is the broadest representative of that theme, but none of the soft pairs look like true duplicates or the same underlying fix.", + "confidence": 0.87, + "canonical_issue_reason": "#43873 is the broadest, most central issue in the set: it covers quantization/offloading behavior in a way that best matches the cluster\u2019s recurring loading/offload reports.", + "canonical_pr_reason": null, + "best_issue_reason": "#43873 is the best umbrella issue for this cluster because it is broadly scoped and closest to the shared loading/offload/quantization thread seen in the other nearby reports.", + "best_pr_reason": null, + "soft_edge_verdicts": [ + { + "left": "issue:43299", + "right": "issue:45357", + "accept": false, + "reason": "Both involve Qwen/VL models, but one is a loading failure and the other is a save_pretrained regression with different symptoms and code paths." + }, + { + "left": "issue:46082", + "right": "issue:46512", + "accept": false, + "reason": "Different config bugs: explicit head_dim validation versus stale dtype metadata after from_config(dtype=...)." + }, + { + "left": "issue:43653", + "right": "issue:45593", + "accept": false, + "reason": "Tokenizer special-token decode bug and D-FINE auxiliary-loss behavior are unrelated." + }, + { + "left": "issue:41950", + "right": "issue:45823", + "accept": false, + "reason": "Different subsystems and failures: video pipeline processor lookup versus Gemma4 device_map forward mismatch." + }, + { + "left": "issue:43650", + "right": "issue:46209", + "accept": false, + "reason": "No clear shared bug; one is a vague data request, the other is an FP8 inference failure on B200." + }, + { + "left": "issue:44242", + "right": "issue:45200", + "accept": false, + "reason": "One is MoE load-balancing loss gating, the other is Gemma 4 multimodal token defaults; not the same defect." + }, + { + "left": "issue:43644", + "right": "issue:47145", + "accept": false, + "reason": "Both mention loading/corruption, but one is non-persistent buffers filled with junk while the other is FP8 weight_scale not applied on frozen checkpoints." + }, + { + "left": "issue:43299", + "right": "issue:43854", + "accept": false, + "reason": "Different model families and failures: Qwen3VL MoE loading versus GLM-4.7-Flash unit-test loading." + }, + { + "left": "issue:44462", + "right": "issue:45081", + "accept": false, + "reason": "Tokenizer loading issues, but one ignores tokenizer.json and the other crashes patching Mistral regex due to backend_tokenizer access." + }, + { + "left": "issue:43931", + "right": "issue:47145", + "accept": false, + "reason": "A weight-shape mismatch for Qwen3-VL is unrelated to FP8 compressed-tensors weight_scale handling." + }, + { + "left": "issue:44683", + "right": "issue:45542", + "accept": false, + "reason": "Completely different problems: compiled flex_attention on torch>=2.9 versus a missing TensorFlow backend when only tensorboard is installed." + }, + { + "left": "issue:45588", + "right": "issue:47000", + "accept": false, + "reason": "Different failure modes in different code paths: flash_attention s_aux=None crash versus RTDetrLoss infeasible cost matrix." + }, + { + "left": "issue:44016", + "right": "issue:44291", + "accept": false, + "reason": "A notebook syntax error is unrelated to the init_empty_weights TypeError regression." + }, + { + "left": "issue:43479", + "right": "issue:43525", + "accept": false, + "reason": "Both are config-related, but one concerns multimodal default sub-configs and the other a missing pad_token_id attribute." + }, + { + "left": "issue:45081", + "right": "issue:45907", + "accept": false, + "reason": "Tokenizer patching crash and list_repo_files revision forwarding are unrelated APIs." + }, + { + "left": "issue:33290", + "right": "issue:45656", + "accept": false, + "reason": "Both mention deepspeed, but one is Adafactor OOM and the other is a double optimizer-step bug; not the same underlying issue." + }, + { + "left": "issue:42994", + "right": "issue:43317", + "accept": false, + "reason": "Both are quantization/offload adjacent, but one is save_pretrained failure and the other is load failure with gpu/cpu offload." + }, + { + "left": "issue:44336", + "right": "issue:46851", + "accept": false, + "reason": "ANSI escape handling in loading_report is unrelated to Qwen3.5 variable-length position_ids support." + }, + { + "left": "issue:43441", + "right": "issue:45834", + "accept": false, + "reason": "Different model-specific bugs: FlashAttention failure for Ministral-3 versus an OCR index error in Kosmos2.5." + }, + { + "left": "issue:44623", + "right": "issue:44749", + "accept": false, + "reason": "processor.save_pretrained missing files and a slowdown report after upgrading Transformers are not the same bug." + }, + { + "left": "issue:43873", + "right": "issue:47145", + "accept": false, + "reason": "Both are in the broad offload/quantization area, but one is a general offloading behavior issue and the other is a specific FP8 checkpoint scaling bug." + }, + { + "left": "issue:45103", + "right": "issue:45372", + "accept": false, + "reason": "Auto-docstring annotation handling and Gemma 4 processor import failure are unrelated." + }, + { + "left": "issue:45859", + "right": "issue:46851", + "accept": false, + "reason": "Both are Qwen3.5 issues, but one is CPU offload incompatibility and the other is variable-length position_ids support; different code paths." + }, + { + "left": "issue:44261", + "right": "issue:44263", + "accept": false, + "reason": "Different model components and defects: MLA RMS norm epsilon precision issue versus torch.split return values in a MoE indexer." + }, + { + "left": "issue:44623", + "right": "issue:45397", + "accept": false, + "reason": "Missing processor save files and Gemma-4 zero3 from_pretrained are separate loading/saving problems." + }, + { + "left": "issue:42175", + "right": "issue:43976", + "accept": false, + "reason": "Package dependency installation and Python-version compatibility are different release/build issues." + }, + { + "left": "issue:43901", + "right": "issue:44246", + "accept": false, + "reason": "A docs mismatch about return_all_scores is unrelated to intermittent import latency." + } + ] + }, + "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": 27, + "review_unit_index": 43, + "review_unit_count": 63, + "status": "reviewed", + "reason": null, + "source": "llm", + "cache_hit": false, + "model": "gpt-5.4-mini?service_tier=flex", + "cluster_id": "cluster-10000-514", + "nodes": [ + "issue:41628", + "issue:42796", + "issue:43441", + "issue:43525", + "issue:43526", + "issue:43540", + "issue:43575", + "issue:43643", + "issue:43673", + "issue:43723", + "issue:43761", + "issue:43784", + "issue:43824", + "issue:43881", + "issue:43976", + "issue:44016", + "issue:44162", + "issue:44186", + "issue:44188", + "issue:44246", + "issue:44297", + "issue:44336", + "issue:44373", + "issue:44496", + "issue:44589", + "issue:44671", + "issue:44779", + "issue:44871", + "issue:44987", + "issue:45003", + "issue:45083", + "issue:45325", + "issue:45362", + "issue:45397", + "issue:45542", + "issue:45561", + "issue:45563", + "issue:45600", + "issue:45712", + "issue:45823", + "issue:46291", + "issue:46399", + "issue:46655", + "issue:46726", + "issue:46752", + "issue:46851", + "issue:47000", + "issue:47145" + ], + "soft_pairs": [ + "issue:45600|issue:45712", + "issue:44671|issue:44779", + "issue:43525|issue:43881", + "issue:43673|issue:43761", + "issue:44016|issue:45823", + "issue:43575|issue:47145", + "issue:44186|issue:47000", + "issue:44188|issue:44336", + "issue:43824|issue:43976", + "issue:44987|issue:45542", + "issue:43525|issue:43643", + "issue:43723|issue:44162", + "issue:42796|issue:43441", + "issue:44336|issue:44373", + "issue:44589|issue:45362", + "issue:45325|issue:46851", + "issue:44246|issue:44496", + "issue:46291|issue:46399", + "issue:44297|issue:45397", + "issue:44871|issue:47145", + "issue:45083|issue:46851", + "issue:41628|issue:43976", + "issue:46655|issue:46726", + "issue:43784|issue:45542", + "issue:46399|issue:46851", + "issue:43526|issue:45003", + "issue:45561|issue:45563", + "issue:43540|issue:46752" + ], + "prepared_review_unit_hash": "6e46cd65ae32f3db0c7e69495de98c5921090c42673b1e9bf60f5803f6ee5007", + "component_budget": { + "node_count": 514, + "item_count": 514, + "soft_pair_count": 1891, + "serialized_chars": 980895, + "estimated_input_tokens": 245224, + "estimated_eval_tokens": 490704 + }, + "budget": { + "node_count": 48, + "item_count": 48, + "soft_pair_count": 28, + "serialized_chars": 36038, + "estimated_input_tokens": 9010, + "estimated_eval_tokens": 18276 + }, + "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 (tokenizer, model loading, position_ids, kernels, processors) but describe different concrete failures, models, or code paths. I do not see a true duplicate family dominating the cluster, so there is no strong canonical issue candidate here. All soft edges are false positives.", + "confidence": 0.9, + "canonical_issue_reason": null, + "canonical_pr_reason": null, + "best_issue_reason": null, + "best_pr_reason": null, + "soft_edge_verdicts": [ + { + "left": "issue:45600", + "right": "issue:45712", + "accept": false, + "reason": "Both are repo-maintenance/regression issues, but one is stale auto-mappings for Sam3LiteText configs and the other is leftover dummy classes leaking into exports; different files and fixes." + }, + { + "left": "issue:44671", + "right": "issue:44779", + "accept": false, + "reason": "Both concern incorrect model behavior in v5, but CamemBERT MLM predictions and Deepseek tokenizer outputs are different model families and code paths." + }, + { + "left": "issue:43525", + "right": "issue:43881", + "accept": false, + "reason": "Separate loading failures: missing pad_token_id on Llama4Config vs a GLM-4V-9B load failure." + }, + { + "left": "issue:43673", + "right": "issue:43761", + "accept": false, + "reason": "Chunked prefill cache regression in GenerationMixin is unrelated to CLIPVisionModel not returning hidden_states." + }, + { + "left": "issue:44016", + "right": "issue:45823", + "accept": false, + "reason": "Notebook syntax error and Gemma4 device-mismatch forward bug are unrelated." + }, + { + "left": "issue:43575", + "right": "issue:47145", + "accept": false, + "reason": "Qwen2 OOM on tensor parallel loading and FP8 checkpoint weight_scale loading are different problems." + }, + { + "left": "issue:44186", + "right": "issue:47000", + "accept": false, + "reason": "LayoutLMv2 tokenizer NER/padding crash is unrelated to RTDetrLoss infeasible cost matrix." + }, + { + "left": "issue:44188", + "right": "issue:44336", + "accept": false, + "reason": "Torch.compile attention branching divergence and ANSI code leakage in loading_report are unrelated." + }, + { + "left": "issue:43824", + "right": "issue:43976", + "accept": false, + "reason": "Import error for a missing Qwen2.5-VL class is unrelated to the Python 3.9/3.10 compatibility issue." + }, + { + "left": "issue:44987", + "right": "issue:45542", + "accept": false, + "reason": "One is a model-loading failure for physical-intelligence/fast; the other is a TensorBoard/TensorFlow backend import issue." + }, + { + "left": "issue:43525", + "right": "issue:43643", + "accept": false, + "reason": "Both involve config fields, but they are different bugs: missing pad_token_id vs trust_remote_code returning incomplete AutoConfig fields." + }, + { + "left": "issue:43723", + "right": "issue:44162", + "accept": false, + "reason": "Generic tokenizer loading regression and ESM2-specific breakage are not the same underlying bug." + }, + { + "left": "issue:42796", + "right": "issue:43441", + "accept": false, + "reason": "Ministral tokenizer BPE-marker handling and Ministral flash-attention failure are different components." + }, + { + "left": "issue:44336", + "right": "issue:44373", + "accept": false, + "reason": "ANSI code emission in loading_report is unrelated to a wrong docstring for position_ids." + }, + { + "left": "issue:44589", + "right": "issue:45362", + "accept": false, + "reason": "Float8 storage lookup failure and Qwen3.5 chat crash are unrelated." + }, + { + "left": "issue:45325", + "right": "issue:46851", + "accept": false, + "reason": "Both mention position_ids, but one is Qwen2.5-VL temporal scaling for still images and the other is Qwen3.5 variable-length input handling." + }, + { + "left": "issue:44246", + "right": "issue:44496", + "accept": false, + "reason": "Slow import and unrecognized model/config load failure are unrelated." + }, + { + "left": "issue:46291", + "right": "issue:46399", + "accept": false, + "reason": "Both mention kernels, but one fails during transformers.pipelines import due to kernels package metadata and the other is a Qwen3.5 kernel-registration decorator issue." + }, + { + "left": "issue:44297", + "right": "issue:45397", + "accept": false, + "reason": "Tokenizer save_pretrained metadata mismatch is unrelated to Gemma-4 zero3 from_pretrained." + }, + { + "left": "issue:44871", + "right": "issue:47145", + "accept": false, + "reason": "Gemma eos_token_id inconsistency and FP8 weight_scale loading are different model/config bugs." + }, + { + "left": "issue:45083", + "right": "issue:46851", + "accept": false, + "reason": "Unexpected feature-extract output lengths in qwen3_omni_moe is not the same as Qwen3.5 variable-length position_ids handling." + }, + { + "left": "issue:41628", + "right": "issue:43976", + "accept": false, + "reason": "AutoImageProcessor import failure and Python version compatibility regression are unrelated." + }, + { + "left": "issue:46655", + "right": "issue:46726", + "accept": false, + "reason": "Both involve processors, but one is a video padding-calculation bug and the other is an AutoProcessor error when a vision backend is missing." + }, + { + "left": "issue:43784", + "right": "issue:45542", + "accept": false, + "reason": "sentence-transformers import NameError and TensorBoard/TensorFlow backend error are unrelated." + }, + { + "left": "issue:46399", + "right": "issue:46851", + "accept": false, + "reason": "Qwen3.5 kernel registration crash and Qwen3.5 variable-length position_ids are different code paths." + }, + { + "left": "issue:43526", + "right": "issue:45003", + "accept": false, + "reason": "BEiT label reduction bug and unsafe sys.modules access are unrelated." + }, + { + "left": "issue:45561", + "right": "issue:45563", + "accept": false, + "reason": "pytest-xdist race on captured_info.txt and a stale num_return_sequences warning are unrelated." + }, + { + "left": "issue:43540", + "right": "issue:46752", + "accept": false, + "reason": "Qwen3OmniMoe video-input ValueError and empty-list apply_chat_template IndexError 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": 27, + "review_unit_index": 44, + "review_unit_count": 63, + "status": "reviewed", + "reason": null, + "source": "llm", + "cache_hit": false, + "model": "gpt-5.4-mini?service_tier=flex", + "cluster_id": "cluster-10000-514", + "nodes": [ + "issue:36246", + "issue:41950", + "issue:42491", + "issue:42757", + "issue:42886", + "issue:43257", + "issue:43262", + "issue:43299", + "issue:43531", + "issue:43540", + "issue:43550", + "issue:43901", + "issue:44162", + "issue:44220", + "issue:44242", + "issue:44246", + "issue:44295", + "issue:44336", + "issue:44351", + "issue:44451", + "issue:44479", + "issue:44485", + "issue:44496", + "issue:44749", + "issue:44779", + "issue:44871", + "issue:45003", + "issue:45072", + "issue:45103", + "issue:45137", + "issue:45200", + "issue:45292", + "issue:45362", + "issue:45375", + "issue:45397", + "issue:45588", + "issue:45593", + "issue:45759", + "issue:46132", + "issue:46399", + "issue:46506", + "issue:46614", + "issue:46655", + "issue:46665", + "issue:46752", + "issue:46821", + "issue:46829", + "issue:46851" + ], + "soft_pairs": [ + "issue:45103|issue:46399", + "issue:45072|issue:46829", + "issue:44295|issue:45397", + "issue:44246|issue:44871", + "issue:43531|issue:45362", + "issue:45588|issue:46851", + "issue:46614|issue:46665", + "issue:42491|issue:43299", + "issue:44336|issue:44485", + "issue:45003|issue:46821", + "issue:44242|issue:45593", + "issue:43550|issue:46506", + "issue:42886|issue:46132", + "issue:44779|issue:45292", + "issue:36246|issue:42757", + "issue:36246|issue:44351", + "issue:43901|issue:45003", + "issue:43257|issue:45137", + "issue:44496|issue:45003", + "issue:41950|issue:44220", + "issue:43540|issue:46655", + "issue:44451|issue:46752", + "issue:43262|issue:44246", + "issue:45375|issue:45759", + "issue:44479|issue:46655", + "issue:44295|issue:44749", + "issue:44162|issue:45200" + ], + "prepared_review_unit_hash": "cd8dddf845478ea7a94744ac0a1dfd3e00510399fe0f877429d5c7cb565ef7fc", + "component_budget": { + "node_count": 514, + "item_count": 514, + "soft_pair_count": 1891, + "serialized_chars": 980895, + "estimated_input_tokens": 245224, + "estimated_eval_tokens": 490704 + }, + "budget": { + "node_count": 48, + "item_count": 48, + "soft_pair_count": 27, + "serialized_chars": 35864, + "estimated_input_tokens": 8966, + "estimated_eval_tokens": 18188 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": true, + "analyst_result": { + "summary": "These items are mostly distinct regressions or documentation/config issues that only share broad subsystem/model-family overlap. No soft pair looks like the same concrete bug/change, so I rejected all duplicate candidates.", + "confidence": 0.77, + "canonical_issue_reason": "issue:44479 is the broadest concrete regression in the set: it describes a multi-model v5.3.0 video-input breakage with a clear common failure mode, making it the best single anchor if one issue must represent this cluster.", + "canonical_pr_reason": null, + "best_issue_reason": "issue:44479 is the most representative issue because it spans multiple related Qwen video models and names a specific regression rather than a narrow, model-specific symptom.", + "best_pr_reason": null, + "soft_edge_verdicts": [ + { + "left": "issue:45103", + "right": "issue:46399", + "accept": false, + "reason": "Both mention loading/registration errors, but one is an auto-docstring crash from future annotations and the other is a Qwen3.5 kernel-registration problem; different code paths and bugs." + }, + { + "left": "issue:45072", + "right": "issue:46829", + "accept": false, + "reason": "These are both output/hidden-state mismatches, but they affect different models and different inference paths; not the same underlying defect." + }, + { + "left": "issue:44295", + "right": "issue:45397", + "accept": false, + "reason": "A buffer registration access error is unrelated to the Gemma-4 ZeRO3 from_pretrained failure; only generic training/loading context overlaps." + }, + { + "left": "issue:44246", + "right": "issue:44871", + "accept": false, + "reason": "One is an import-performance complaint and the other is an eos_token_id config inconsistency; no shared concrete bug." + }, + { + "left": "issue:43531", + "right": "issue:45362", + "accept": false, + "reason": "Both are Qwen-related, but one concerns sliding_window handling in MoE models while the other is a chat crash for Qwen3.5; different failure modes." + }, + { + "left": "issue:45588", + "right": "issue:46851", + "accept": false, + "reason": "Flash-attention sink handling and variable-length position_ids are separate code paths and symptoms." + }, + { + "left": "issue:46614", + "right": "issue:46665", + "accept": false, + "reason": "Stale phonemizer language state is unrelated to from_pretrained behavior inside Python classes." + }, + { + "left": "issue:42491", + "right": "issue:43299", + "accept": false, + "reason": "Both involve Qwen MoE loading, but one is LoRA compatibility across hf4.x/hf5.x while the other is a Qwen3VL MoE load regression; not the same bug." + }, + { + "left": "issue:44336", + "right": "issue:44485", + "accept": false, + "reason": "ANSI escape output in loading_report is unrelated to GLM-5 RoPE implementation." + }, + { + "left": "issue:45003", + "right": "issue:46821", + "accept": false, + "reason": "Unsafe sys.modules access in modeling_utils and a TimesFM window_size AttributeError are different issues with no shared code path." + }, + { + "left": "issue:44242", + "right": "issue:45593", + "accept": false, + "reason": "Both concern losses, but one is router/load-balancing loss gating and the other is D-FINE auxiliary loss suppression; distinct models and logic." + }, + { + "left": "issue:43550", + "right": "issue:46506", + "accept": false, + "reason": "torch.compile+SDPA on Bamba is unrelated to CPU/CUDA output differences in PPDocLayoutV3." + }, + { + "left": "issue:42886", + "right": "issue:46132", + "accept": false, + "reason": "Offline cache loading and AttentionInterface.register behavior changes are unrelated." + }, + { + "left": "issue:44779", + "right": "issue:45292", + "accept": false, + "reason": "Deepseek tokenizer regression and resize_token_embeddings not updating output embeddings are different functional bugs." + }, + { + "left": "issue:36246", + "right": "issue:42757", + "accept": false, + "reason": "These are both import errors, but they involve different missing symbols from different packages and unrelated fixes." + }, + { + "left": "issue:36246", + "right": "issue:44351", + "accept": false, + "reason": "Both are import/export issues, but Qwen2_5_VLImageProcessor and HybridCache are different symbols with different causes." + }, + { + "left": "issue:43901", + "right": "issue:45003", + "accept": false, + "reason": "Docs mentioning return_all_scores versus unsafe sys.modules access are not the same issue." + }, + { + "left": "issue:43257", + "right": "issue:45137", + "accept": false, + "reason": "DeepSpeed ZeRO3 deque pop error and Qwen3 MoE weight conversion under accelerate+deepspeed are both training-stack issues but not the same concrete bug." + }, + { + "left": "issue:44496", + "right": "issue:45003", + "accept": false, + "reason": "Unrecognized model config and unsafe sys.modules access are unrelated loading problems." + }, + { + "left": "issue:41950", + "right": "issue:44220", + "accept": false, + "reason": "Video-classification pipeline image-processor lookup and _torch_extract_fbank_features are separate media pipeline bugs." + }, + { + "left": "issue:43540", + "right": "issue:46655", + "accept": false, + "reason": "Both touch video processing, but one is a Qwen3OmniMoe value error and the other is a SmolVLM padding-calculation bug; different code paths." + }, + { + "left": "issue:44451", + "right": "issue:46752", + "accept": false, + "reason": "Model loading failure for ScandiBERT is unrelated to apply_chat_template on an empty list." + }, + { + "left": "issue:43262", + "right": "issue:44246", + "accept": false, + "reason": "Audio processor chat-template sampling rate defaults and slow imports are unrelated." + }, + { + "left": "issue:45375", + "right": "issue:45759", + "accept": false, + "reason": "Both involve Qwen 3.5/3.6 multimodal config handling, but one is a strict-config field drop and the other is from_config failing to unwrap text_config; different underlying bugs." + }, + { + "left": "issue:44479", + "right": "issue:46655", + "accept": false, + "reason": "Both are video-input regressions, but they affect different models and likely different preprocessing logic; too broad to treat as one duplicate." + }, + { + "left": "issue:44295", + "right": "issue:44749", + "accept": false, + "reason": "A buffer/position_ids access issue is unrelated to the Chinese-language performance regression after upgrading transformers." + }, + { + "left": "issue:44162", + "right": "issue:45200", + "accept": false, + "reason": "ESM2 breakage and Gemma 4 mm_token_type_ids defaults are different model-specific 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": 27, + "review_unit_index": 45, + "review_unit_count": 63, + "status": "reviewed", + "reason": null, + "source": "llm", + "cache_hit": false, + "model": "gpt-5.4-mini?service_tier=flex", + "cluster_id": "cluster-10000-514", + "nodes": [ + "issue:29870", + "issue:30333", + "issue:36246", + "issue:42199", + "issue:43065", + "issue:43295", + "issue:43352", + "issue:43408", + "issue:43475", + "issue:43526", + "issue:43596", + "issue:43653", + "issue:44038", + "issue:44060", + "issue:44077", + "issue:44162", + "issue:44242", + "issue:44368", + "issue:44466", + "issue:44479", + "issue:44623", + "issue:44857", + "issue:44877", + "issue:44977", + "issue:44987", + "issue:45003", + "issue:45137", + "issue:45250", + "issue:45292", + "issue:45335", + "issue:45375", + "issue:45397", + "issue:45412", + "issue:45459", + "issue:45561", + "issue:45600", + "issue:45656", + "issue:45859", + "issue:46123", + "issue:46170", + "issue:46369", + "issue:46399", + "issue:46506", + "issue:46726", + "issue:46851", + "issue:46894", + "issue:46897", + "issue:47000" + ], + "soft_pairs": [ + "issue:45003|issue:45561", + "issue:45292|issue:47000", + "issue:45137|issue:45600", + "issue:45859|issue:46399", + "issue:43295|issue:45292", + "issue:36246|issue:43596", + "issue:43475|issue:46369", + "issue:44466|issue:45335", + "issue:44623|issue:46726", + "issue:44877|issue:45375", + "issue:44242|issue:44857", + "issue:44038|issue:46506", + "issue:43526|issue:46897", + "issue:44987|issue:45459", + "issue:29870|issue:30333", + "issue:43352|issue:45250", + "issue:43526|issue:46894", + "issue:44479|issue:46170", + "issue:42199|issue:46851", + "issue:44060|issue:44077", + "issue:44368|issue:44977", + "issue:43653|issue:44162", + "issue:45397|issue:45656", + "issue:45412|issue:46123", + "issue:43065|issue:43408", + "issue:44877|issue:46170" + ], + "prepared_review_unit_hash": "a9f1c0428a4e1e6bc71c92dd075158cc02a20a895f46f90efcb9e8285b7f7051", + "component_budget": { + "node_count": 514, + "item_count": 514, + "soft_pair_count": 1891, + "serialized_chars": 980895, + "estimated_input_tokens": 245224, + "estimated_eval_tokens": 490704 + }, + "budget": { + "node_count": 48, + "item_count": 48, + "soft_pair_count": 26, + "serialized_chars": 35612, + "estimated_input_tokens": 8903, + "estimated_eval_tokens": 18062 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": true, + "analyst_result": { + "summary": "This cluster is highly heterogeneous: it spans unrelated bugs in generation, deepspeed, processors/tokenizers, SAM3, Qwen, DETR, config handling, and CI. None of the soft edges look like the same underlying defect or a single mergeable PR.", + "confidence": 0.98, + "canonical_issue_reason": "No clear canonical issue: the items do not share one underlying bug, so there is no meaningful duplicate hub for the cluster.", + "canonical_pr_reason": null, + "best_issue_reason": "No strong representative issue. The closest generic-looking item is 45292, but it is still unrelated to most others and should not be treated as the cluster center.", + "best_pr_reason": null, + "soft_edge_verdicts": [ + { + "left": "issue:45003", + "right": "issue:45561", + "accept": false, + "reason": "Different failures: unsafe sys.modules access in modeling_utils vs a pytest-xdist race on captured_info.txt in testing utilities." + }, + { + "left": "issue:45292", + "right": "issue:47000", + "accept": false, + "reason": "Unrelated bugs: token embedding resize not updating output embeddings vs RT-DETR loss crashing with infeasible cost matrix." + }, + { + "left": "issue:45137", + "right": "issue:45600", + "accept": false, + "reason": "Different code paths and symptoms: DeepSpeed ZeRO3 deque underflow vs stale auto_mappings references to removed Sam3LiteText configs." + }, + { + "left": "issue:45859", + "right": "issue:46399", + "accept": false, + "reason": "Both involve Qwen3.5-family models, but one is CPU offload incompatibility in rotary embedding and the other is a missing registration decorator crash with use_kernels=True." + }, + { + "left": "issue:43295", + "right": "issue:45292", + "accept": false, + "reason": "Different regressions: processor/tokenizer API breakage vs resize_token_embeddings not propagating to output embeddings." + }, + { + "left": "issue:36246", + "right": "issue:43596", + "accept": false, + "reason": "Completely different problems: missing Qwen2_5_VLImageProcessor import vs DeepSpeed ZeRO3/BertModel IndexError." + }, + { + "left": "issue:43475", + "right": "issue:46369", + "accept": false, + "reason": "Same broad SAM3 area, but distinct bugs: missing fpn_position_embeddings attribute vs memory overflow when calling sam3-litetext." + }, + { + "left": "issue:44466", + "right": "issue:45335", + "accept": false, + "reason": "Related to tied embeddings/resize behavior, but not the same bug: device-dependent lm_head serialization vs t5gemma decoder.embed_tokens not updating." + }, + { + "left": "issue:44623", + "right": "issue:46726", + "accept": false, + "reason": "Different processor issues: save_pretrained missing files vs AutoProcessor complaining about missing vision backend." + }, + { + "left": "issue:44877", + "right": "issue:45375", + "accept": false, + "reason": "Both mention strict config, but the concrete defects differ: granite_speech loading blocked vs Qwen3_5MoeVisionConfig field being silently dropped." + }, + { + "left": "issue:44242", + "right": "issue:44857", + "accept": false, + "reason": "Unrelated modeling bugs: MoE load-balancing loss omission vs LwDetrImageLoss crashing under float16 AMP on CUDA." + }, + { + "left": "issue:44038", + "right": "issue:46506", + "accept": false, + "reason": "Different models and failures: Qwen3-VL-Moe / transformers 5.0 breakage vs CPU/CUDA output mismatch in PPDocLayoutV3." + }, + { + "left": "issue:43526", + "right": "issue:46897", + "accept": false, + "reason": "Different label-processing bug vs ASR training-loss double-shift bug; same general ML area but not the same defect." + }, + { + "left": "issue:44987", + "right": "issue:45459", + "accept": false, + "reason": "Loading regression for physical-intelligence/fast vs tokenizer error masking when protobuf is missing; no shared code-path bug." + }, + { + "left": "issue:29870", + "right": "issue:30333", + "accept": false, + "reason": "Generation-time Bart unbound local error vs MLflow reporting not updated after training failure; unrelated functionality." + }, + { + "left": "issue:43352", + "right": "issue:45250", + "accept": false, + "reason": "One is a model-specific unsupported Flash Attention 2 guard, the other is a generic Flash Attention 2 topic; not the same concrete bug." + }, + { + "left": "issue:43526", + "right": "issue:46894", + "accept": false, + "reason": "Different issues despite similar 'double' wording: BEiT label reduction bug vs CohereASR training-loss double-shift bug." + }, + { + "left": "issue:44479", + "right": "issue:46170", + "accept": false, + "reason": "Unrelated: video-input regression for Qwen vision-language models vs Windows/ROCm import/type dependency problems." + }, + { + "left": "issue:42199", + "right": "issue:46851", + "accept": false, + "reason": "Different model families and failure modes: DETR cardinality/background-class error vs Qwen3.5 variable-length position_ids behavior." + }, + { + "left": "issue:44060", + "right": "issue:44077", + "accept": false, + "reason": "Different issues in the Qwen/patchtsmixer ecosystem: tied-weights warning vs optional post_init policy change." + }, + { + "left": "issue:44368", + "right": "issue:44977", + "accept": false, + "reason": "One is a tied-embeddings warning in LoRA fine-tuning, the other is a flash-attention generation failure for Qwen3.5." + }, + { + "left": "issue:43653", + "right": "issue:44162", + "accept": false, + "reason": "Different tokenizer/model bugs: BigBirdTokenizer special-token decode problem vs ESM2 being broken." + }, + { + "left": "issue:45397", + "right": "issue:45656", + "accept": false, + "reason": "Different DeepSpeed problems: gemma-4 from_pretrained failure vs optimizer.step being called twice." + }, + { + "left": "issue:45412", + "right": "issue:46123", + "accept": false, + "reason": "Memory/resource management in RT-DETR vs a MaskGenerationPipeline partial-batch bookkeeping bug; not the same defect." + }, + { + "left": "issue:43065", + "right": "issue:43408", + "accept": false, + "reason": "Different SAM3 issues: dummy Conv2d in Sam3PixelDecoder vs wrong model type warning when loading sam3_video into sam3_tracker." + }, + { + "left": "issue:44877", + "right": "issue:46170", + "accept": false, + "reason": "Strict config loading issue vs Windows/ROCm torch.distributed dependency failure; unrelated 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": 27, + "review_unit_index": 46, + "review_unit_count": 63, + "status": "reviewed", + "reason": null, + "source": "llm", + "cache_hit": false, + "model": "gpt-5.4-mini?service_tier=flex", + "cluster_id": "cluster-10000-514", + "nodes": [ + "issue:42886", + "issue:43065", + "issue:43475", + "issue:43540", + "issue:43646", + "issue:43701", + "issue:43723", + "issue:43906", + "issue:43931", + "issue:43994", + "issue:44162", + "issue:44188", + "issue:44297", + "issue:44361", + "issue:44368", + "issue:44483", + "issue:44623", + "issue:44749", + "issue:44871", + "issue:44964", + "issue:44995", + "issue:45103", + "issue:45137", + "issue:45200", + "issue:45229", + "issue:45375", + "issue:45478", + "issue:45563", + "issue:45588", + "issue:45593", + "issue:45656", + "issue:45685", + "issue:45987", + "issue:46082", + "issue:46123", + "issue:46132", + "issue:46170", + "issue:46202", + "issue:46421", + "issue:46431", + "issue:46710", + "issue:46726", + "issue:46752", + "issue:46832", + "issue:46858", + "issue:47020", + "issue:47068", + "issue:47145" + ], + "soft_pairs": [ + "issue:43475|issue:43540", + "issue:46170|issue:46832", + "issue:42886|issue:45229", + "issue:43723|issue:46710", + "issue:43931|issue:44368", + "issue:44188|issue:46202", + "issue:44368|issue:45563", + "issue:44995|issue:45200", + "issue:46132|issue:47145", + "issue:44483|issue:46202", + "issue:43906|issue:46202", + "issue:44188|issue:46421", + "issue:45593|issue:47068", + "issue:44995|issue:45588", + "issue:45987|issue:46082", + "issue:45685|issue:46431", + "issue:44361|issue:44749", + "issue:45103|issue:46858", + "issue:44297|issue:44623", + "issue:43065|issue:46123", + "issue:43994|issue:46726", + "issue:44749|issue:45656", + "issue:43701|issue:44749", + "issue:43646|issue:44162", + "issue:44623|issue:45656", + "issue:46752|issue:47020", + "issue:44871|issue:46132", + "issue:44964|issue:45137", + "issue:45375|issue:45478" + ], + "prepared_review_unit_hash": "6cd7e63e15c22fd5aef32fab4a7934487e4176fbf51590d5e25e89b176676ec5", + "component_budget": { + "node_count": 514, + "item_count": 514, + "soft_pair_count": 1891, + "serialized_chars": 980895, + "estimated_input_tokens": 245224, + "estimated_eval_tokens": 490704 + }, + "budget": { + "node_count": 48, + "item_count": 48, + "soft_pair_count": 29, + "serialized_chars": 36618, + "estimated_input_tokens": 9155, + "estimated_eval_tokens": 18566 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": true, + "analyst_result": { + "summary": "This cluster is mostly heterogeneous: most pairs are different bugs across unrelated models, pipelines, or subsystems. The only clear duplicate-looking pair is the Qwen3.5 MoE config/load failure; everything else should stay separate.", + "confidence": 0.79, + "canonical_issue_reason": "issue:45375 is the best canonical issue because it identifies the likely root cause directly: Qwen3_5MoeVisionConfig is missing deepstack_visual_indexes and @strict drops it, which explains the broader from_pretrained failure.", + "canonical_pr_reason": null, + "best_issue_reason": "issue:45375 is the strongest representative issue in the set: it is specific, root-cause oriented, and directly describes the underlying config/schema regression.", + "best_pr_reason": null, + "soft_edge_verdicts": [ + { + "left": "issue:43475", + "right": "issue:43540", + "accept": false, + "reason": "Different models and failures: SAM 3 video attribute error vs Qwen3OmniMoe video input validation." + }, + { + "left": "issue:46170", + "right": "issue:46832", + "accept": false, + "reason": "Unrelated subsystems: Windows/ROCm import dependency error vs RT-DETR forward shape/type error." + }, + { + "left": "issue:42886", + "right": "issue:45229", + "accept": false, + "reason": "Both are transformer bugs, but one is offline tokenizer cache loading and the other is Gemma multi-GPU OOM." + }, + { + "left": "issue:43723", + "right": "issue:46710", + "accept": false, + "reason": "Tokenizer loading regression vs model accuracy regression after upgrade; not the same bug." + }, + { + "left": "issue:43931", + "right": "issue:44368", + "accept": false, + "reason": "Weight-shape mismatch on Qwen3-VL load is unrelated to a warning about tie_word_embeddings in fine-tuning." + }, + { + "left": "issue:44188", + "right": "issue:46202", + "accept": false, + "reason": "Different code paths: torch.compile attention-kernel divergence vs Gemma pipeline roundtrip/tokenization issue." + }, + { + "left": "issue:44368", + "right": "issue:45563", + "accept": false, + "reason": "A config warning about tied embeddings is unrelated to a stale generate() warning on num_return_sequences." + }, + { + "left": "issue:44995", + "right": "issue:45200", + "accept": false, + "reason": "Stale MoE indexer cache on second forward is not the same as Gemma 4 missing mm_token_type_ids defaults." + }, + { + "left": "issue:46132", + "right": "issue:47145", + "accept": false, + "reason": "AttentionInterface registration behavior change is unrelated to FP8 checkpoint loading missing weight_scale." + }, + { + "left": "issue:44483", + "right": "issue:46202", + "accept": false, + "reason": "Chat/completions API rejection and Gemma pipeline roundtrip failure are different bugs." + }, + { + "left": "issue:43906", + "right": "issue:46202", + "accept": false, + "reason": "An isolated reproduction of another issue is not the same as a Gemma tokenizer roundtrip failure." + }, + { + "left": "issue:44188", + "right": "issue:46421", + "accept": false, + "reason": "torch.compile attention divergence and cache lazy-init recompilation are distinct problems." + }, + { + "left": "issue:45593", + "right": "issue:47068", + "accept": false, + "reason": "D-FINE auxiliary-loss logic is unrelated to a deepspeed loss host-sync concern." + }, + { + "left": "issue:44995", + "right": "issue:45588", + "accept": false, + "reason": "Stale MoE cache on second forward is unrelated to flash_attention crashing on s_aux=None." + }, + { + "left": "issue:45987", + "right": "issue:46082", + "accept": false, + "reason": "Cache return-type contract bug and LlamaConfig head_dim validation are different issues." + }, + { + "left": "issue:45685", + "right": "issue:46431", + "accept": false, + "reason": "MPS histogram failure in MoE has no overlap with missing chat-template support in a processor." + }, + { + "left": "issue:44361", + "right": "issue:44749", + "accept": false, + "reason": "Tokenizer AttributeError on tasks is unrelated to a performance regression after upgrade." + }, + { + "left": "issue:45103", + "right": "issue:46858", + "accept": false, + "reason": "auto_docstring crashes with future annotations; this is unrelated to static cache seq-length behavior under torch.compile." + }, + { + "left": "issue:44297", + "right": "issue:44623", + "accept": false, + "reason": "Tokenizer save_pretrained metadata mismatch and processor save_pretrained missing files are related only superficially." + }, + { + "left": "issue:43065", + "right": "issue:46123", + "accept": false, + "reason": "SAM3 pixel decoder dummy Conv2d is unrelated to MaskGenerationPipeline dropping the final partial batch." + }, + { + "left": "issue:43994", + "right": "issue:46726", + "accept": false, + "reason": "Siglip2 nonsensical outputs and missing-vision-backend processor detection are different failure modes." + }, + { + "left": "issue:44749", + "right": "issue:45656", + "accept": false, + "reason": "Performance slowdown after upgrade is not the same as deepspeed calling optimizer.step() twice." + }, + { + "left": "issue:43701", + "right": "issue:44749", + "accept": false, + "reason": "Checkpoint key mismatch and post-upgrade slowdown do not share a concrete code-path bug." + }, + { + "left": "issue:43646", + "right": "issue:44162", + "accept": false, + "reason": "Custom model initialization breakage and ESM2 breakage are both upgrade regressions but not the same defect." + }, + { + "left": "issue:44623", + "right": "issue:45656", + "accept": false, + "reason": "Processor serialization missing files is unrelated to deepspeed double optimizer stepping." + }, + { + "left": "issue:46752", + "right": "issue:47020", + "accept": false, + "reason": "Empty chat-template input IndexError and XLMRobertaTokenizer Unigram construction error are unrelated." + }, + { + "left": "issue:44871", + "right": "issue:46132", + "accept": false, + "reason": "Gemma eos_token_id inconsistency does not match AttentionInterface.register behavior change." + }, + { + "left": "issue:44964", + "right": "issue:45137", + "accept": false, + "reason": "Phi-4 multimodal load failure and DeepSpeed ZeRO3 deque underflow are unrelated." + }, + { + "left": "issue:45375", + "right": "issue:45478", + "accept": true, + "reason": "Likely same Qwen3.5 MoE loading regression: the config issue explains the broader from_pretrained error." + } + ] + }, + "evaluator_result": { + "accept": true, + "feedback": "The summary is grounded in the packet and stays conservative. The canonical pair 45375/45478 is well supported by the titles and looks mergeable as the same Qwen3.5 MoE config/loading fix. The other soft-edge rejections are cautious and consistent with the visible evidence; no overstatement stands out." + }, + "evaluator_used": true, + "retried": false, + "accepted_nontrivial_soft_edge": true, + "error_kind": null, + "error_message": null + }, + { + "label": "issue", + "component_index": 1, + "component_count": 27, + "review_unit_index": 47, + "review_unit_count": 63, + "status": "reviewed", + "reason": null, + "source": "llm", + "cache_hit": false, + "model": "gpt-5.4-mini?service_tier=flex", + "cluster_id": "cluster-10000-514", + "nodes": [ + "issue:28282", + "issue:30990", + "issue:36010", + "issue:39692", + "issue:41628", + "issue:42371", + "issue:42994", + "issue:43232", + "issue:43475", + "issue:43482", + "issue:43525", + "issue:43643", + "issue:43673", + "issue:43723", + "issue:43784", + "issue:43824", + "issue:43873", + "issue:43986", + "issue:44038", + "issue:44188", + "issue:44351", + "issue:44373", + "issue:44623", + "issue:44661", + "issue:44857", + "issue:44869", + "issue:44933", + "issue:44945", + "issue:45092", + "issue:45137", + "issue:45230", + "issue:45310", + "issue:45356", + "issue:45375", + "issue:45397", + "issue:45406", + "issue:45561", + "issue:45753", + "issue:46202", + "issue:46326", + "issue:46399", + "issue:46463", + "issue:46726", + "issue:46821", + "issue:46851", + "issue:46877", + "issue:47000", + "issue:47059" + ], + "soft_pairs": [ + "issue:42994|issue:43482", + "issue:43475|issue:46821", + "issue:43475|issue:43643", + "issue:45230|issue:45397", + "issue:44188|issue:46851", + "issue:39692|issue:46726", + "issue:43232|issue:43673", + "issue:28282|issue:30990", + "issue:44373|issue:46851", + "issue:45310|issue:45375", + "issue:43986|issue:46726", + "issue:44038|issue:44623", + "issue:46877|issue:47000", + "issue:41628|issue:44933", + "issue:46463|issue:46877", + "issue:42994|issue:43525", + "issue:44869|issue:45356", + "issue:43723|issue:46326", + "issue:42371|issue:45092", + "issue:43784|issue:46399", + "issue:43873|issue:44945", + "issue:45137|issue:47000", + "issue:41628|issue:44351", + "issue:36010|issue:41628", + "issue:42994|issue:43824", + "issue:44945|issue:45753", + "issue:45561|issue:47059", + "issue:44038|issue:45310", + "issue:44857|issue:46877", + "issue:44661|issue:47059", + "issue:45406|issue:46202" + ], + "prepared_review_unit_hash": "08fb64bae1a853ebd56a7044612aa6d2f2592ffa2d4247288ebfce24e54727ae", + "component_budget": { + "node_count": 514, + "item_count": 514, + "soft_pair_count": 1891, + "serialized_chars": 980895, + "estimated_input_tokens": 245224, + "estimated_eval_tokens": 490704 + }, + "budget": { + "node_count": 48, + "item_count": 48, + "soft_pair_count": 31, + "serialized_chars": 36646, + "estimated_input_tokens": 9162, + "estimated_eval_tokens": 18580 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": true, + "analyst_result": { + "summary": "This cluster is a loose mix of transformer import, generation, tokenizer, and model-specific regression reports. The soft links are mostly superficial wording overlap rather than the same underlying bug, so I would not merge any of them as duplicates.", + "confidence": 0.84, + "canonical_issue_reason": "issue:41628 is the broadest and most central import-regression report in the set, and it best represents the small import-error subset of this loose cluster.", + "canonical_pr_reason": null, + "best_issue_reason": "issue:41628 is the most generally framed user-facing failure here and the closest thing to a cluster anchor, though the overall set is still heterogeneous.", + "best_pr_reason": null, + "soft_edge_verdicts": [ + { + "left": "issue:42994", + "right": "issue:43482", + "accept": false, + "reason": "Save failure for quantized models and GGUF loading failure are different code paths and different operations." + }, + { + "left": "issue:43475", + "right": "issue:46821", + "accept": false, + "reason": "Different models and different missing attributes; SAM3 video output bug is unrelated to TimesFM window_size." + }, + { + "left": "issue:43475", + "right": "issue:43643", + "accept": false, + "reason": "One is a model output attribute bug, the other is a generic remote-code config field loss issue." + }, + { + "left": "issue:45230", + "right": "issue:45397", + "accept": false, + "reason": "Generic bug report versus a specific gemma-4 ZeRO3 from_pretrained failure; not the same underlying issue." + }, + { + "left": "issue:44188", + "right": "issue:46851", + "accept": false, + "reason": "torch.compile attention-kernel divergence is unrelated to Qwen3.5 variable-length position_ids handling." + }, + { + "left": "issue:39692", + "right": "issue:46726", + "accept": false, + "reason": "Documentation example errors are separate from runtime missing-vision-backend processor behavior." + }, + { + "left": "issue:43232", + "right": "issue:43673", + "accept": false, + "reason": "Both mention generation, but they target different failures: sync_gpus kwargs handling vs chunked_prefill cache access." + }, + { + "left": "issue:28282", + "right": "issue:30990", + "accept": false, + "reason": "Missing PyTorch dependency and sentence-transformers hanging on load are different problems." + }, + { + "left": "issue:44373", + "right": "issue:46851", + "accept": false, + "reason": "A wrong docstring is not the same as a runtime bug in variable-length position_ids handling." + }, + { + "left": "issue:45310", + "right": "issue:45375", + "accept": false, + "reason": "Related model family, but one is a from_pretrained error and the other is a silent config-field drop." + }, + { + "left": "issue:43986", + "right": "issue:46726", + "accept": false, + "reason": "Both mention vision backends, but they describe different loaders and different missing-backend failures." + }, + { + "left": "issue:44038", + "right": "issue:44623", + "accept": false, + "reason": "A Qwen3-VL-Moe bug and processor.save_pretrained missing files are unrelated failures." + }, + { + "left": "issue:46877", + "right": "issue:47000", + "accept": false, + "reason": "Import-time doc-lint output and an RTDetr loss crash are entirely unrelated." + }, + { + "left": "issue:41628", + "right": "issue:44933", + "accept": false, + "reason": "Both are import errors, but they are for different symbols and different modules." + }, + { + "left": "issue:46463", + "right": "issue:46877", + "accept": false, + "reason": "T5 decoder-input validation messaging has nothing to do with doc-lint output at import time." + }, + { + "left": "issue:42994", + "right": "issue:43525", + "accept": false, + "reason": "Quantized saving failure and missing pad_token_id on Llama4Config are different bugs." + }, + { + "left": "issue:44869", + "right": "issue:45356", + "accept": false, + "reason": "Whisper timestamp decoding crash and Gemma/Kimi tokenizer regression are separate tokenizer issues." + }, + { + "left": "issue:43723", + "right": "issue:46326", + "accept": false, + "reason": "Tokenizer loading in v5 and mistral_common continue_final_message are different tokenizer paths." + }, + { + "left": "issue:42371", + "right": "issue:45092", + "accept": false, + "reason": "TF32 API guidance and InternVL2 meta-initialization incompatibility are unrelated." + }, + { + "left": "issue:43784", + "right": "issue:46399", + "accept": false, + "reason": "Sentence-transformers import NameError and Qwen3.5 kernel registration failure are unrelated." + }, + { + "left": "issue:43873", + "right": "issue:44945", + "accept": false, + "reason": "Quantization offloading behavior and pipeline-parallel output correctness are different issues." + }, + { + "left": "issue:45137", + "right": "issue:47000", + "accept": false, + "reason": "DeepSpeed ZeRO3 deque underflow and RTDetr loss infeasibility are unrelated crashes." + }, + { + "left": "issue:41628", + "right": "issue:44351", + "accept": false, + "reason": "Different missing imports: AutoImageProcessor versus HybridCache." + }, + { + "left": "issue:36010", + "right": "issue:41628", + "accept": false, + "reason": "GenerationMixin import failure and AutoImageProcessor import failure are distinct compatibility issues." + }, + { + "left": "issue:42994", + "right": "issue:43824", + "accept": false, + "reason": "Quantized saving failure and missing Qwen2_5_VLForConditionalGeneration import are unrelated." + }, + { + "left": "issue:44945", + "right": "issue:45753", + "accept": false, + "reason": "Pipeline-parallel output errors and a Qwen3.5 infinite-loop case are not the same bug." + }, + { + "left": "issue:45561", + "right": "issue:47059", + "accept": false, + "reason": "pytest-xdist file-race in testing utils and a huggingface_hub CLI attribute error are unrelated." + }, + { + "left": "issue:44038", + "right": "issue:45310", + "accept": false, + "reason": "Same model family, but different versions and different failure modes; not mergeable as one bug." + }, + { + "left": "issue:44857", + "right": "issue:46877", + "accept": false, + "reason": "CUDA AMP loss crash and doc-lint stdout at import time are completely different problems." + }, + { + "left": "issue:44661", + "right": "issue:47059", + "accept": false, + "reason": "Model-generation mapping logic and CLI completion compatibility are unrelated." + }, + { + "left": "issue:45406", + "right": "issue:46202", + "accept": false, + "reason": "Gemma4 processor missing _tokenizer and Gemma pipeline roundtrip issues are different defects." + } + ] + }, + "evaluator_result": null, + "evaluator_used": false, + "retried": false, + "accepted_nontrivial_soft_edge": false, + "error_kind": null, + "error_message": null + }, + { + "label": "issue", + "component_index": 1, + "component_count": 27, + "review_unit_index": 48, + "review_unit_count": 63, + "status": "reviewed", + "reason": null, + "source": "llm", + "cache_hit": false, + "model": "gpt-5.4-mini?service_tier=flex", + "cluster_id": "cluster-10000-514", + "nodes": [ + "issue:30333", + "issue:30990", + "issue:33290", + "issue:35141", + "issue:39692", + "issue:42175", + "issue:42673", + "issue:42831", + "issue:42898", + "issue:43232", + "issue:43479", + "issue:43704", + "issue:43723", + "issue:43976", + "issue:44016", + "issue:44079", + "issue:44186", + "issue:44222", + "issue:44496", + "issue:44568", + "issue:44623", + "issue:44743", + "issue:44779", + "issue:44792", + "issue:44869", + "issue:44933", + "issue:45081", + "issue:45103", + "issue:45310", + "issue:45405", + "issue:45561", + "issue:45593", + "issue:45600", + "issue:45712", + "issue:45812", + "issue:46077", + "issue:46123", + "issue:46202", + "issue:46231", + "issue:46693", + "issue:46701", + "issue:46829", + "issue:46877", + "issue:46909", + "issue:47059", + "issue:47068", + "issue:47085" + ], + "soft_pairs": [ + "issue:45593|issue:46877", + "issue:44222|issue:45310", + "issue:45081|issue:45103", + "issue:46202|issue:46701", + "issue:43232|issue:44792", + "issue:42673|issue:43704", + "issue:42831|issue:44222", + "issue:42175|issue:44933", + "issue:44496|issue:45561", + "issue:45405|issue:46877", + "issue:30333|issue:33290", + "issue:44743|issue:47068", + "issue:43976|issue:44933", + "issue:30990|issue:35141", + "issue:45561|issue:45600", + "issue:43479|issue:46829", + "issue:42898|issue:44568", + "issue:42175|issue:45712", + "issue:44568|issue:47085", + "issue:44623|issue:45310", + "issue:42673|issue:46693", + "issue:44186|issue:47085", + "issue:45600|issue:47059", + "issue:46077|issue:46123", + "issue:42898|issue:45812", + "issue:43723|issue:44568", + "issue:43479|issue:44869", + "issue:44016|issue:46231", + "issue:39692|issue:46909", + "issue:43479|issue:44079", + "issue:42898|issue:44779" + ], + "prepared_review_unit_hash": "fceeb80d80ee08c311bd8abfa74948b00f997a653319e2e83a9d46f7de96b2be", + "component_budget": { + "node_count": 514, + "item_count": 514, + "soft_pair_count": 1891, + "serialized_chars": 980895, + "estimated_input_tokens": 245224, + "estimated_eval_tokens": 490704 + }, + "budget": { + "node_count": 47, + "item_count": 47, + "soft_pair_count": 31, + "serialized_chars": 36553, + "estimated_input_tokens": 9139, + "estimated_eval_tokens": 18534 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": true, + "analyst_result": { + "summary": "This set is mostly heterogeneous; the only clear duplicate pair is the Qwen3ForCausalLM VRAM leak report. No pull requests are present.", + "confidence": 0.95, + "canonical_issue_reason": "issue:42673 is the stronger canonical issue because it is the more detailed, earlier report of the same Qwen3ForCausalLM VRAM leak in multi-dataloader-thread usage; issue:43704 is a later near-identical repeat.", + "canonical_pr_reason": null, + "best_issue_reason": "issue:42673 is the best representative issue for the only confirmed duplicate subcluster and has the richest original description.", + "best_pr_reason": null, + "soft_edge_verdicts": [ + { + "left": "issue:45593", + "right": "issue:46877", + "accept": false, + "reason": "Different bugs: FP8/auxiliary-loss behavior vs auto_docstring import-time stderr logging." + }, + { + "left": "issue:44222", + "right": "issue:45310", + "accept": false, + "reason": "Both mention MoE/FP8, but one is save_pretrained behavior and the other is from_pretrained loading failure; not the same bug." + }, + { + "left": "issue:45081", + "right": "issue:45103", + "accept": false, + "reason": "Both are AttributeError crashes, but in unrelated code paths (Mistral regex patching vs auto_docstring kwargs processing)." + }, + { + "left": "issue:46202", + "right": "issue:46701", + "accept": false, + "reason": "Both are Gemma-related, but one is tokenizer roundtrip on \u2581 and the other is an image-text pipeline output issue." + }, + { + "left": "issue:43232", + "right": "issue:44792", + "accept": false, + "reason": "Unrelated generation-state bug vs a janus test failure." + }, + { + "left": "issue:42673", + "right": "issue:43704", + "accept": true, + "reason": "Same issue: identical title and same Qwen3ForCausalLM VRAM leak when used across multiple dataloader threads." + }, + { + "left": "issue:42831", + "right": "issue:44222", + "accept": false, + "reason": "Different FP8 problems: accuracy regression vs save_pretrained on MoE." + }, + { + "left": "issue:42175", + "right": "issue:44933", + "accept": false, + "reason": "Different subsystems: backend extras installation vs image_utils import regression." + }, + { + "left": "issue:44496", + "right": "issue:45561", + "accept": false, + "reason": "Unrelated: model loading/config error vs pytest-xdist file race in testing utils." + }, + { + "left": "issue:45405", + "right": "issue:46877", + "accept": false, + "reason": "Unrelated packaging/version pin issue vs auto_docstring import-time diagnostics." + }, + { + "left": "issue:30333", + "right": "issue:33290", + "accept": false, + "reason": "Different failure modes: MLflow reporting after training job failure vs Deepspeed/Adafactor OOM." + }, + { + "left": "issue:44743", + "right": "issue:47068", + "accept": false, + "reason": "Different bugs in different generation/training paths." + }, + { + "left": "issue:43976", + "right": "issue:44933", + "accept": false, + "reason": "Python version compatibility issue vs nonexistent image_utils import." + }, + { + "left": "issue:30990", + "right": "issue:35141", + "accept": false, + "reason": "Unrelated: loading hang vs embedding resize/post_init reinitialization." + }, + { + "left": "issue:45561", + "right": "issue:45600", + "accept": false, + "reason": "Different CI/test problems: xdist file contention vs removed Sam3LiteText config references." + }, + { + "left": "issue:43479", + "right": "issue:46829", + "accept": false, + "reason": "Different model-output/config initialization bug vs output field mismatch in VideoPrism classification." + }, + { + "left": "issue:42898", + "right": "issue:44568", + "accept": false, + "reason": "Both tokenizer regressions, but one is clean-up spacing behavior and the other is special-token addition for a specific tokenizer." + }, + { + "left": "issue:42175", + "right": "issue:45712", + "accept": false, + "reason": "Different installation/backend extras issue vs dummy_pt_objects leakage into dir(transformers)." + }, + { + "left": "issue:44568", + "right": "issue:47085", + "accept": false, + "reason": "Different tokenizer behaviors; no shared concrete code-path failure." + }, + { + "left": "issue:44623", + "right": "issue:45310", + "accept": false, + "reason": "Different loading/saving bugs: processor.save_pretrained missing files vs Qwen3.5 MoE from_pretrained failure." + }, + { + "left": "issue:42673", + "right": "issue:46693", + "accept": false, + "reason": "Both mention performance, but one is a VRAM leak in dataloader-thread usage and the other is a long-context FA2 regression." + }, + { + "left": "issue:44186", + "right": "issue:47085", + "accept": false, + "reason": "Different tokenizer issues; no evidence they share the same underlying thread-safety bug." + }, + { + "left": "issue:45600", + "right": "issue:47059", + "accept": false, + "reason": "Different CI failures: removed config references vs huggingface_hub CLI incompatibility." + }, + { + "left": "issue:46077", + "right": "issue:46123", + "accept": false, + "reason": "Different pipeline behaviors: bidirectional mask construction vs partial-batch result dropping." + }, + { + "left": "issue:42898", + "right": "issue:45812", + "accept": false, + "reason": "Both are tokenizer regressions, but different concrete failures and models." + }, + { + "left": "issue:43723", + "right": "issue:44568", + "accept": false, + "reason": "Different v5 tokenizer loading issue vs special-token/BOS-EOS behavior on one tokenizer." + }, + { + "left": "issue:43479", + "right": "issue:44869", + "accept": false, + "reason": "Different bugs: config default initialization vs Whisper word-timestamp decoding crash." + }, + { + "left": "issue:44016", + "right": "issue:46231", + "accept": false, + "reason": "Notebook syntax error vs documentation parameter-name mismatch." + }, + { + "left": "issue:39692", + "right": "issue:46909", + "accept": false, + "reason": "Both are broken docs examples, but for different models and different failure modes." + }, + { + "left": "issue:43479", + "right": "issue:44079", + "accept": false, + "reason": "Different issues: multimodal config defaults vs ModelOutput key assignment semantics." + }, + { + "left": "issue:42898", + "right": "issue:44779", + "accept": false, + "reason": "Both are tokenizer regressions, but different models and different incorrect outputs." + } + ] + }, + "evaluator_result": { + "accept": true, + "feedback": "The summary is grounded in the packet. The only confirmed duplicate pair is correctly identified as issue:42673 / issue:43704, with conservative reasoning and the earlier, more detailed report chosen as canonical. The soft-edge verdicts stay cautious and are supported by the titles and metadata. No pull requests are present, matching 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": 27, + "review_unit_index": 49, + "review_unit_count": 63, + "status": "reviewed", + "reason": null, + "source": "llm", + "cache_hit": false, + "model": "gpt-5.4-mini?service_tier=flex", + "cluster_id": "cluster-10000-514", + "nodes": [ + "issue:22355", + "issue:27957", + "issue:28282", + "issue:30333", + "issue:34689", + "issue:35141", + "issue:41628", + "issue:42222", + "issue:42673", + "issue:42757", + "issue:42796", + "issue:42831", + "issue:42898", + "issue:43116", + "issue:43295", + "issue:43479", + "issue:43644", + "issue:43723", + "issue:43784", + "issue:43824", + "issue:43874", + "issue:44060", + "issue:44075", + "issue:44190", + "issue:44222", + "issue:44295", + "issue:44297", + "issue:44360", + "issue:44373", + "issue:44448", + "issue:44485", + "issue:44623", + "issue:44734", + "issue:44843", + "issue:44933", + "issue:45092", + "issue:45103", + "issue:45125", + "issue:45200", + "issue:45245", + "issue:45356", + "issue:45405", + "issue:45656", + "issue:45712", + "issue:46132", + "issue:46231", + "issue:46762", + "issue:47085" + ], + "soft_pairs": [ + "issue:43784|issue:43824", + "issue:44843|issue:47085", + "issue:42673|issue:47085", + "issue:27957|issue:30333", + "issue:30333|issue:34689", + "issue:43874|issue:45092", + "issue:42898|issue:43295", + "issue:44295|issue:45656", + "issue:42222|issue:45200", + "issue:44734|issue:45103", + "issue:42898|issue:45356", + "issue:30333|issue:35141", + "issue:22355|issue:28282", + "issue:44222|issue:44297", + "issue:44075|issue:46231", + "issue:43116|issue:44190", + "issue:45405|issue:45712", + "issue:44297|issue:45656", + "issue:42831|issue:44623", + "issue:44933|issue:45712", + "issue:44448|issue:46132", + "issue:44222|issue:44623", + "issue:44060|issue:45125", + "issue:42796|issue:43723", + "issue:43479|issue:43644", + "issue:45245|issue:46132", + "issue:44373|issue:44485", + "issue:44360|issue:46762", + "issue:41628|issue:42757" + ], + "prepared_review_unit_hash": "fb4aed9b77579f719f1630f1684b7f2767cd2550b045cfcf6d6e48c39fb97c85", + "component_budget": { + "node_count": 514, + "item_count": 514, + "soft_pair_count": 1891, + "serialized_chars": 980895, + "estimated_input_tokens": 245224, + "estimated_eval_tokens": 490704 + }, + "budget": { + "node_count": 48, + "item_count": 48, + "soft_pair_count": 29, + "serialized_chars": 36332, + "estimated_input_tokens": 9083, + "estimated_eval_tokens": 18422 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": true, + "analyst_result": { + "summary": "This is a heterogeneous grab bag of unrelated Transformers issues (imports, tokenizer/model regressions, training/runtime bugs, docs, and config problems). None of the soft-similarity pairs look like the same underlying bug or change, so all proposed duplicate links should be rejected.", + "confidence": 0.91, + "canonical_issue_reason": "issue:30333 is the least-arbitrary anchor only because it has the most soft links, but the cluster is too heterogeneous for a true canonical duplicate.", + "canonical_pr_reason": null, + "best_issue_reason": "issue:30333 is the best available representative by graph connectivity, though it is still not a genuine duplicate hub.", + "best_pr_reason": null, + "soft_edge_verdicts": [ + { + "left": "issue:43784", + "right": "issue:43824", + "accept": false, + "reason": "Both are import errors, but one is about sentence-transformers/nn and the other about a missing Qwen VL class; unrelated failures." + }, + { + "left": "issue:44843", + "right": "issue:47085", + "accept": false, + "reason": "Offline tokenizer patching vs Rust tokenizer thread-safety are different code paths and bugs." + }, + { + "left": "issue:42673", + "right": "issue:47085", + "accept": false, + "reason": "VRAM leak with Qwen3-Next in dataloader threads is not the same as a tokenizer thread-safety regression." + }, + { + "left": "issue:27957", + "right": "issue:30333", + "accept": false, + "reason": "Flash Attention 2 on XLMRoberta and MLFlow reporting after failed training are unrelated." + }, + { + "left": "issue:30333", + "right": "issue:34689", + "accept": false, + "reason": "Training-job MLFlow reporting and Llama 3.2 Vision model loading are different issues." + }, + { + "left": "issue:43874", + "right": "issue:45092", + "accept": false, + "reason": "Missing image-processor method in GLM46V is unrelated to InternVL2 meta-init incompatibility." + }, + { + "left": "issue:42898", + "right": "issue:43295", + "accept": false, + "reason": "Both are v5 regressions, but one is tokenizer cleanup behavior and the other is processor/tokenizer image handling." + }, + { + "left": "issue:44295", + "right": "issue:45656", + "accept": false, + "reason": "Buffer registration access error and double optimizer step with deepspeed are unrelated." + }, + { + "left": "issue:42222", + "right": "issue:45200", + "accept": false, + "reason": "VitPose model breakage and Gemma 4 mm_token_type_ids defaults are different multimodal bugs." + }, + { + "left": "issue:44734", + "right": "issue:45103", + "accept": false, + "reason": "Serve/KV-cache tensor indexing crash and auto_docstring future-annotations crash are unrelated." + }, + { + "left": "issue:42898", + "right": "issue:45356", + "accept": false, + "reason": "Tokenizer cleanup behavior change and Kimi-K2.5 codec/regression warning are different tokenizer issues." + }, + { + "left": "issue:30333", + "right": "issue:35141", + "accept": false, + "reason": "MLFlow reporting on failed jobs and embedding resize/post_init reinitialization are unrelated." + }, + { + "left": "issue:22355", + "right": "issue:28282", + "accept": false, + "reason": "Missing transformers.onnx module and missing PyTorch dependency are distinct import problems." + }, + { + "left": "issue:44222", + "right": "issue:44297", + "accept": false, + "reason": "FP8 MoE save_pretrained behavior and tokenizer_class metadata mismatch are different save_pretrained issues." + }, + { + "left": "issue:44075", + "right": "issue:46231", + "accept": false, + "reason": "SGD optimizer arguments being ignored is not the same as a documentation parameter-name typo." + }, + { + "left": "issue:43116", + "right": "issue:44190", + "accept": false, + "reason": "run_classification.py multi-label output bug and local dataset loading in image classification are unrelated example-script issues." + }, + { + "left": "issue:45405", + "right": "issue:45712", + "accept": false, + "reason": "PEFT version pinning and leftover dummy classes are completely different repository maintenance problems." + }, + { + "left": "issue:44297", + "right": "issue:45656", + "accept": false, + "reason": "Tokenizer config save mismatch and double optimizer stepping are unrelated." + }, + { + "left": "issue:42831", + "right": "issue:44623", + "accept": false, + "reason": "FineGrainedFP8 accuracy issues and missing processor save files are different bugs." + }, + { + "left": "issue:44933", + "right": "issue:45712", + "accept": false, + "reason": "Bad import from image_utils and stray dummy classes in dummy_pt_objects.py are unrelated cleanup/import issues." + }, + { + "left": "issue:44448", + "right": "issue:46132", + "accept": false, + "reason": "Different outputs across v4/v5 for Pegasus and AttentionInterface.register changing behavior are not the same fix." + }, + { + "left": "issue:44222", + "right": "issue:44623", + "accept": false, + "reason": "FP8 MoE save_pretrained bug and processor.save_pretrained missing files are different serialization problems." + }, + { + "left": "issue:44060", + "right": "issue:45125", + "accept": false, + "reason": "Wrong tied-weights warning in Qwen3-Next and missing _tp_plan in Qwen3_5Moe are unrelated Qwen model bugs." + }, + { + "left": "issue:42796", + "right": "issue:43723", + "accept": false, + "reason": "Tokenizer BPE-marker handling and AutoTokenizer.from_pretrained v5 loading are different tokenizer failures." + }, + { + "left": "issue:43479", + "right": "issue:43644", + "accept": false, + "reason": "Default multimodal config initialization and non-persistent buffer junk filling are unrelated v5 issues." + }, + { + "left": "issue:45245", + "right": "issue:46132", + "accept": false, + "reason": "Category cardinality limit and AttentionInterface.register behavior are not the same defect." + }, + { + "left": "issue:44373", + "right": "issue:44485", + "accept": false, + "reason": "Docstring typo for position_ids and GLM-5 RoPE implementation are unrelated." + }, + { + "left": "issue:44360", + "right": "issue:46762", + "accept": false, + "reason": "DSA indexer lacking ReLU and Minimax attention numerical differences are different low-level math issues." + }, + { + "left": "issue:41628", + "right": "issue:42757", + "accept": false, + "reason": "Missing AutoImageProcessor import and missing is_offline_mode import are different API/export 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": 27, + "review_unit_index": 50, + "review_unit_count": 63, + "status": "reviewed", + "reason": null, + "source": "llm", + "cache_hit": false, + "model": "gpt-5.4-mini?service_tier=flex", + "cluster_id": "cluster-10000-514", + "nodes": [ + "issue:22355", + "issue:27957", + "issue:30990", + "issue:38175", + "issue:41628", + "issue:42831", + "issue:42907", + "issue:43122", + "issue:43232", + "issue:43475", + "issue:43504", + "issue:43723", + "issue:43824", + "issue:44038", + "issue:44190", + "issue:44222", + "issue:44261", + "issue:44263", + "issue:44279", + "issue:44295", + "issue:44297", + "issue:44360", + "issue:44373", + "issue:44484", + "issue:44623", + "issue:44829", + "issue:44964", + "issue:44977", + "issue:45071", + "issue:45092", + "issue:45310", + "issue:45357", + "issue:45834", + "issue:46002", + "issue:46050", + "issue:46082", + "issue:46209", + "issue:46305", + "issue:46328", + "issue:46369", + "issue:46424", + "issue:46431", + "issue:46614", + "issue:46655", + "issue:46909", + "issue:46932", + "issue:47059" + ], + "soft_pairs": [ + "issue:44297|issue:45310", + "issue:45071|issue:46082", + "issue:43504|issue:47059", + "issue:43723|issue:44977", + "issue:44222|issue:44484", + "issue:43122|issue:46209", + "issue:46614|issue:46932", + "issue:44263|issue:45092", + "issue:44295|issue:44623", + "issue:46209|issue:46431", + "issue:43232|issue:46424", + "issue:43232|issue:45834", + "issue:44263|issue:44373", + "issue:30990|issue:38175", + "issue:44190|issue:46082", + "issue:43475|issue:46305", + "issue:44263|issue:44360", + "issue:46305|issue:46655", + "issue:46369|issue:46909", + "issue:42831|issue:45310", + "issue:22355|issue:27957", + "issue:42907|issue:46050", + "issue:44279|issue:46328", + "issue:45092|issue:45357", + "issue:42831|issue:44038", + "issue:44829|issue:46002", + "issue:44261|issue:45092", + "issue:41628|issue:43824", + "issue:44964|issue:47059" + ], + "prepared_review_unit_hash": "09b3f77fc2a6aaab6370321b916b1a50799ca25e1ba22d3d64d2ede9fe9d50c8", + "component_budget": { + "node_count": 514, + "item_count": 514, + "soft_pair_count": 1891, + "serialized_chars": 980895, + "estimated_input_tokens": 245224, + "estimated_eval_tokens": 490704 + }, + "budget": { + "node_count": 47, + "item_count": 47, + "soft_pair_count": 29, + "serialized_chars": 35451, + "estimated_input_tokens": 8863, + "estimated_eval_tokens": 17982 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": true, + "analyst_result": { + "summary": "The set is heterogeneous: most pairs share only a broad transformers-v5 / model-loading / serialization theme, but not the same concrete bug or fix path. I would not merge any of the soft-edge candidates.", + "confidence": 0.72, + "canonical_issue_reason": "If one issue must represent the cluster, issue 45092 is the best fit: it is a concrete, user-facing v5 compatibility regression with a clear failure mode (remote-code checkpoints and meta initialization) and stronger signal than the narrower or unrelated reports.", + "canonical_pr_reason": null, + "best_issue_reason": "Issue 45092 is the strongest canonical candidate because it is specific, reproducible, and central to the recurring v5 loading-compatibility theme; the other issues are either narrower symptoms or clearly orthogonal.", + "best_pr_reason": null, + "soft_edge_verdicts": [ + { + "left": "issue:44297", + "right": "issue:45310", + "accept": false, + "reason": "Both involve model/tokenizer loading, but one is a tokenizer_class save-pretrained mismatch and the other is a Qwen3.5 MoE from_pretrained failure; different bug surfaces and likely different fixes." + }, + { + "left": "issue:45071", + "right": "issue:46082", + "accept": false, + "reason": "These are different config-validation regressions: PretrainedConfig type checking versus LlamaConfig head_dim handling." + }, + { + "left": "issue:43504", + "right": "issue:47059", + "accept": false, + "reason": "BEiT pretrained loading with a legacy field and a CLI completion error from huggingface_hub are unrelated." + }, + { + "left": "issue:43723", + "right": "issue:44977", + "accept": false, + "reason": "Generic AutoTokenizer loading in v5 and Qwen3.5 flash-attention generation issues are different code paths." + }, + { + "left": "issue:44222", + "right": "issue:44484", + "accept": false, + "reason": "FP8 save_pretrained for MoE and the default max_shard_size question are not the same bug." + }, + { + "left": "issue:43122", + "right": "issue:46209", + "accept": false, + "reason": "Tokenizer behavior regression and FP8 inference failure on B200 are unrelated." + }, + { + "left": "issue:46614", + "right": "issue:46932", + "accept": false, + "reason": "Stale phonemizer language state and Gemma4 dtype casting for quantized embedders are different subsystems." + }, + { + "left": "issue:44263", + "right": "issue:45092", + "accept": false, + "reason": "torch.split return semantics in a DSA indexer and remote-code checkpoint meta-init incompatibility are not the same concrete bug." + }, + { + "left": "issue:44295", + "right": "issue:44623", + "accept": false, + "reason": "A position_ids buffer read error and missing files from processor.save_pretrained are different save/load problems." + }, + { + "left": "issue:46209", + "right": "issue:46431", + "accept": false, + "reason": "FP8 inference on B200 and apply_chat_template failing because a processor has no chat template are unrelated." + }, + { + "left": "issue:43232", + "right": "issue:46424", + "accept": false, + "reason": "Generation kwargs after sync_gpus and max_cache_length honoring only quantized cache are separate generation/cache behaviors." + }, + { + "left": "issue:43232", + "right": "issue:45834", + "accept": false, + "reason": "Generation kwargs handling and a Kosmos2.5 OCR index error are unrelated." + }, + { + "left": "issue:44263", + "right": "issue:44373", + "accept": false, + "reason": "A torch.split / DSA indexer issue and a wrong docstring for position_ids are not duplicates." + }, + { + "left": "issue:30990", + "right": "issue:38175", + "accept": false, + "reason": "Sentence Transformers loading stuck and zero probabilities in siglip2 are different bugs." + }, + { + "left": "issue:44190", + "right": "issue:46082", + "accept": false, + "reason": "A local dataset loading failure and a LlamaConfig head_dim validation issue are unrelated." + }, + { + "left": "issue:43475", + "right": "issue:46305", + "accept": false, + "reason": "SAM3VisionEncoderOutput missing an attribute and a streaming mask off-by-one are different code paths." + }, + { + "left": "issue:44263", + "right": "issue:44360", + "accept": false, + "reason": "Both touch the DSA indexer, but one is about torch.split outputs and the other about a missing ReLU; too different to be the same bug." + }, + { + "left": "issue:46305", + "right": "issue:46655", + "accept": false, + "reason": "A sliding-window off-by-one in streaming and incorrect video padding calculation are distinct." + }, + { + "left": "issue:46369", + "right": "issue:46909", + "accept": false, + "reason": "Memory overflow in sam3-litetext and a broken DIA documentation example are unrelated." + }, + { + "left": "issue:42831", + "right": "issue:45310", + "accept": false, + "reason": "FineGrainedFP8 accuracy loss and Qwen3.5 Moe from_pretrained errors are not the same bug." + }, + { + "left": "issue:22355", + "right": "issue:27957", + "accept": false, + "reason": "Missing transformers.onnx module and XLMRoberta flash-attention behavior are unrelated." + }, + { + "left": "issue:42907", + "right": "issue:46050", + "accept": false, + "reason": "Dequantized model save failure and sinq save/load breakage are both serialization-related but not the same concrete code-path problem." + }, + { + "left": "issue:44279", + "right": "issue:46328", + "accept": false, + "reason": "A generic dependency issue and a SequenceFeatureExtractor.pad performance issue are unrelated." + }, + { + "left": "issue:45092", + "right": "issue:45357", + "accept": false, + "reason": "Old InternVL2 meta-init incompatibility and Qwen3.5 incorrect visual encoder keys in save_pretrained are separate model-specific regressions." + }, + { + "left": "issue:43824", + "right": "issue:44038", + "accept": false, + "reason": "Missing import for Qwen2_5_VLForConditionalGeneration and a Qwen3-VL-Moe bug are different failures." + }, + { + "left": "issue:44829", + "right": "issue:46002", + "accept": false, + "reason": "Degenerate training with flash_attention_3 and NaNs in classification heads upon checkout/init are different symptoms with different likely causes." + }, + { + "left": "issue:44261", + "right": "issue:45092", + "accept": false, + "reason": "Missing rms_norm_eps in MLA q_a_layernorm and remote-code checkpoint meta-init incompatibility are unrelated." + }, + { + "left": "issue:41628", + "right": "issue:43824", + "accept": false, + "reason": "Both are import errors, but they concern different symbols and likely different export/registration causes." + }, + { + "left": "issue:44964", + "right": "issue:47059", + "accept": false, + "reason": "Phi-4 multimodal loading on latest transformers and a hub CLI completion attribute 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": 27, + "review_unit_index": 51, + "review_unit_count": 63, + "status": "reviewed", + "reason": null, + "source": "llm", + "cache_hit": false, + "model": "gpt-5.4-mini?service_tier=flex", + "cluster_id": "cluster-10000-514", + "nodes": [ + "issue:22355", + "issue:27957", + "issue:28218", + "issue:30990", + "issue:33357", + "issue:39379", + "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:43329", + "issue:43408", + "issue:43540", + "issue:43576", + "issue:43644", + "issue:43646", + "issue:43784", + "issue:43976", + "issue:44188", + "issue:44263", + "issue:44295", + "issue:44297", + "issue:44327", + "issue:44373", + "issue:44483", + "issue:44568", + "issue:44945", + "issue:44964", + "issue:45278", + "issue:45381", + "issue:45406", + "issue:45593", + "issue:45753", + "issue:46002", + "issue:46172", + "issue:46305", + "issue:46369", + "issue:46402", + "issue:46909", + "issue:47059" + ], + "soft_pairs": [ + "issue:43408|issue:46402", + "issue:45381|issue:46305", + "issue:44483|issue:45406", + "issue:39379|issue:46172", + "issue:42831|issue:45753", + "issue:43644|issue:46305", + "issue:22355|issue:28218", + "issue:44327|issue:47059", + "issue:39401|issue:44568", + "issue:44188|issue:44373", + "issue:42757|issue:45278", + "issue:43408|issue:46369", + "issue:42886|issue:42994", + "issue:42886|issue:43066", + "issue:27957|issue:30990", + "issue:42796|issue:43122", + "issue:43232|issue:45593", + "issue:45381|issue:46909", + "issue:43540|issue:46909", + "issue:43646|issue:46002", + "issue:42907|issue:44945", + "issue:44295|issue:44297", + "issue:30990|issue:33357", + "issue:43576|issue:43784", + "issue:43976|issue:44945", + "issue:42907|issue:44964", + "issue:42371|issue:44263", + "issue:41628|issue:45278", + "issue:43329|issue:47059" + ], + "prepared_review_unit_hash": "ba232d7ba95fd0c34d39ffa24f963dfa433f5919393bafab133241ee68c5ad90", + "component_budget": { + "node_count": 514, + "item_count": 514, + "soft_pair_count": 1891, + "serialized_chars": 980895, + "estimated_input_tokens": 245224, + "estimated_eval_tokens": 490704 + }, + "budget": { + "node_count": 48, + "item_count": 48, + "soft_pair_count": 29, + "serialized_chars": 35873, + "estimated_input_tokens": 8969, + "estimated_eval_tokens": 18194 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": true, + "analyst_result": { + "summary": "The cluster is a grab bag of unrelated bug reports. None of the soft-similarity pairs look like the same underlying bug or change, so all soft edges are rejected.", + "confidence": 0.91, + "canonical_issue_reason": "issue:45278 is the broadest, clearest regression report in the set: multiple import errors after upgrading from 4.57.0 to 5.5.0. It is more representative than the narrower, single-symbol import bugs.", + "canonical_pr_reason": null, + "best_issue_reason": "issue:45278 is the best global issue candidate because it describes a concrete, user-facing regression with a clear version jump and broad impact, making it a stronger canonical representative than the other isolated reports.", + "best_pr_reason": null, + "soft_edge_verdicts": [ + { + "left": "issue:43408", + "right": "issue:46402", + "accept": false, + "reason": "Different problems: a sam3_model/sam3_tracker type mismatch warning vs text encoder weights not loading in Sam3Model." + }, + { + "left": "issue:45381", + "right": "issue:46305", + "accept": false, + "reason": "Both involve vision/video, but one is Qwen2.5-VL position IDs and the other is a MoonshineStreaming sliding-window off-by-one; different code paths and fixes." + }, + { + "left": "issue:44483", + "right": "issue:45406", + "accept": false, + "reason": "API request handling for /v1/chat/completions is unrelated to Gemma4Processor missing _tokenizer." + }, + { + "left": "issue:39379", + "right": "issue:46172", + "accept": false, + "reason": "mllama masked-out-mask handling and deepseek-v4 tensor transpose are unrelated implementation bugs." + }, + { + "left": "issue:42831", + "right": "issue:45753", + "accept": false, + "reason": "FineGrainedFP8 accuracy regression and a Qwen3_5 image infinite loop are not the same bug." + }, + { + "left": "issue:43644", + "right": "issue:46305", + "accept": false, + "reason": "Non-persistent buffer junk in v5.0.0 is unrelated to a sliding_window_mask off-by-one in MoonshineStreaming." + }, + { + "left": "issue:22355", + "right": "issue:28218", + "accept": false, + "reason": "Missing transformers.onnx module and tokenizer adds an extra space after added tokens are different issues." + }, + { + "left": "issue:44327", + "right": "issue:47059", + "accept": false, + "reason": "QA decode_spans top-k bounds error and a broken CLI completion attribute are unrelated failures." + }, + { + "left": "issue:39401", + "right": "issue:44568", + "accept": false, + "reason": "Qwen3 tokenizer offset_mapping is a tokenizer alignment bug; add_special_tokens not adding BOS/EOS is a different tokenizer behavior regression." + }, + { + "left": "issue:44188", + "right": "issue:44373", + "accept": false, + "reason": "Diverging attention kernels under torch.compile and a wrong docstring are not the same code-path problem." + }, + { + "left": "issue:42757", + "right": "issue:45278", + "accept": false, + "reason": "Both are import failures, but one is missing is_offline_mode from huggingface_hub while the other is many import errors after a version upgrade; too broad and not the same root cause." + }, + { + "left": "issue:43408", + "right": "issue:46369", + "accept": false, + "reason": "A model-type warning and a memory overflow in sam3-litetext are different symptoms and likely different fixes." + }, + { + "left": "issue:42886", + "right": "issue:42994", + "accept": false, + "reason": "Offline tokenizer cache loading and quantized model saving are unrelated." + }, + { + "left": "issue:42886", + "right": "issue:43066", + "accept": false, + "reason": "Tokenizer cache/offline loading and wrong tokenizer decoder type in v5 are distinct tokenizer issues." + }, + { + "left": "issue:27957", + "right": "issue:30990", + "accept": false, + "reason": "Flash Attention 2 support for XLMRoberta and Sentence Transformers getting stuck loading are unrelated." + }, + { + "left": "issue:42796", + "right": "issue:43122", + "accept": false, + "reason": "Both concern tokenization, but one is Ministral BPE marker handling and the other is a version-to-version tokenization change; different models and likely different causes." + }, + { + "left": "issue:43232", + "right": "issue:45593", + "accept": false, + "reason": "Generation kwargs after sync_gpus and D-FINE auxiliary-loss behavior are unrelated." + }, + { + "left": "issue:45381", + "right": "issue:46909", + "accept": false, + "reason": "Qwen2.5-VL video position IDs and a broken DIA documentation example are not the same bug." + }, + { + "left": "issue:43540", + "right": "issue:46909", + "accept": false, + "reason": "Qwen3OmniMoe video input ValueError and a documentation example issue are unrelated." + }, + { + "left": "issue:43646", + "right": "issue:46002", + "accept": false, + "reason": "Custom model initialization breakage and NaNs in classification heads are different failures." + }, + { + "left": "issue:42907", + "right": "issue:44945", + "accept": false, + "reason": "Saving dequantized Ministrals and incorrect pipeline-parallel LLM output do not look like the same underlying bug." + }, + { + "left": "issue:44295", + "right": "issue:44297", + "accept": false, + "reason": "Reading a buffer-backed position_ids tensor and tokenizer_config tokenizer_class mismatch are separate serialization issues." + }, + { + "left": "issue:30990", + "right": "issue:33357", + "accept": false, + "reason": "Sentence Transformers loading hang and a MacOS bus error with a CLIP model are different runtime failures." + }, + { + "left": "issue:43576", + "right": "issue:43784", + "accept": false, + "reason": "Broken transformers env CLI and sentence-transformers import NameError are unrelated." + }, + { + "left": "issue:43976", + "right": "issue:44945", + "accept": false, + "reason": "Python version support regression and pipeline-parallel output corruption are not the same issue." + }, + { + "left": "issue:42907", + "right": "issue:44964", + "accept": false, + "reason": "Saving dequantized models and loading Phi-4 multimodal are different model workflow failures." + }, + { + "left": "issue:42371", + "right": "issue:44263", + "accept": false, + "reason": "TF32 configuration guidance and a torch.split return-value issue are unrelated." + }, + { + "left": "issue:41628", + "right": "issue:45278", + "accept": false, + "reason": "Cannot import AutoImageProcessor and many import errors after upgrade share a broad theme, but they are different concrete import breakages with different likely fixes." + }, + { + "left": "issue:43329", + "right": "issue:47059", + "accept": false, + "reason": "Undefined variables in multimodal token counting and a CLI completion attribute 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": 27, + "review_unit_index": 52, + "review_unit_count": 63, + "status": "reviewed", + "reason": null, + "source": "llm", + "cache_hit": false, + "model": "gpt-5.4-mini?service_tier=flex", + "cluster_id": "cluster-10000-514", + "nodes": [ + "issue:22355", + "issue:29127", + "issue:29159", + "issue:36296", + "issue:39379", + "issue:39401", + "issue:42175", + "issue:42757", + "issue:42907", + "issue:43116", + "issue:43122", + "issue:43232", + "issue:43278", + "issue:43408", + "issue:43452", + "issue:43475", + "issue:43493", + "issue:43756", + "issue:43784", + "issue:43824", + "issue:43937", + "issue:44038", + "issue:44060", + "issue:44117", + "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:45200", + "issue:45216", + "issue:45290", + "issue:45478", + "issue:45600", + "issue:45753", + "issue:46172", + "issue:46326", + "issue:46614", + "issue:46762" + ], + "soft_pairs": [ + "issue:43756|issue:44060", + "issue:44117|issue:44704", + "issue:42907|issue:43937", + "issue:43784|issue:45600", + "issue:39401|issue:44488", + "issue:39401|issue:44521", + "issue:44625|issue:44829", + "issue:46172|issue:46762", + "issue:44360|issue:44485", + "issue:44704|issue:45290", + "issue:43122|issue:46326", + "issue:43408|issue:43475", + "issue:43122|issue:44521", + "issue:44038|issue:44297", + "issue:42907|issue:43452", + "issue:44038|issue:44295", + "issue:22355|issue:29159", + "issue:44038|issue:44222", + "issue:45216|issue:45478", + "issue:43824|issue:44351", + "issue:43116|issue:45200", + "issue:39401|issue:46614", + "issue:43122|issue:44779", + "issue:42757|issue:44351", + "issue:36296|issue:43116", + "issue:43278|issue:44945", + "issue:22355|issue:29127", + "issue:42175|issue:44938", + "issue:39379|issue:43232", + "issue:43493|issue:45753" + ], + "prepared_review_unit_hash": "6580a57c72761926cfc74bd6ab70357e21e3a4b015a772379b90232eb1bfe04b", + "component_budget": { + "node_count": 514, + "item_count": 514, + "soft_pair_count": 1891, + "serialized_chars": 980895, + "estimated_input_tokens": 245224, + "estimated_eval_tokens": 490704 + }, + "budget": { + "node_count": 48, + "item_count": 48, + "soft_pair_count": 30, + "serialized_chars": 36431, + "estimated_input_tokens": 9108, + "estimated_eval_tokens": 18472 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": true, + "analyst_result": { + "summary": "The cluster is highly heterogeneous: most items are unrelated model-specific bugs, tokenizer regressions, import errors, or generation issues. None of the soft pairs look like the same underlying defect, so I would not merge any of them as duplicates.", + "confidence": 0.95, + "canonical_issue_reason": "No single canonical issue stands out; the items span unrelated bug families across tokenizers, loaders, multimodal handling, generation, and model-specific regressions.", + "canonical_pr_reason": null, + "best_issue_reason": "No issue is globally suitable as a canonical duplicate representative for this cluster because it is not one coherent bug family.", + "best_pr_reason": null, + "soft_edge_verdicts": [ + { + "left": "issue:43756", + "right": "issue:44060", + "accept": false, + "reason": "Different problems: Smollm3 RoPE layer selection vs Qwen3-Next tied-weights warning." + }, + { + "left": "issue:44117", + "right": "issue:44704", + "accept": false, + "reason": "Tokenizer mapping fallback and AutoProcessor kwargs forwarding are different code paths." + }, + { + "left": "issue:42907", + "right": "issue:43937", + "accept": false, + "reason": "Model save/dequantization bug vs invalid GenerationConfig validation." + }, + { + "left": "issue:43784", + "right": "issue:45600", + "accept": false, + "reason": "Missing nn import in sentence-transformers integration vs stale auto_mappings references." + }, + { + "left": "issue:39401", + "right": "issue:44488", + "accept": false, + "reason": "Qwen3 offset_mapping bug vs failure to load a specific model repository." + }, + { + "left": "issue:39401", + "right": "issue:44521", + "accept": false, + "reason": "Tokenizer offset mapping issue and multimodal assistant_mask bug are different defects." + }, + { + "left": "issue:44625", + "right": "issue:44829", + "accept": false, + "reason": "num_labels propagation mismatch vs flash_attention_3 training degeneration." + }, + { + "left": "issue:46172", + "right": "issue:46762", + "accept": false, + "reason": "DeepSeek-V4 dtype/transpose bug is unrelated to Minimax attention numerical differences." + }, + { + "left": "issue:44360", + "right": "issue:44485", + "accept": false, + "reason": "DSA indexer activation discussion is not the same as GLM-5 RoPE implementation." + }, + { + "left": "issue:44704", + "right": "issue:45290", + "accept": false, + "reason": "Processor file-resolution kwargs vs chat-template tool-call crash are different failures." + }, + { + "left": "issue:43122", + "right": "issue:46326", + "accept": false, + "reason": "Different tokenizer bugs in different subsystems; not the same underlying change." + }, + { + "left": "issue:43408", + "right": "issue:43475", + "accept": false, + "reason": "Model-type warning on SAM3 tracker vs missing fpn_position_embeddings attribute." + }, + { + "left": "issue:44038", + "right": "issue:44297", + "accept": false, + "reason": "Qwen3-VL-Moe loading bug vs tokenizer save_pretrained metadata mismatch." + }, + { + "left": "issue:42907", + "right": "issue:43452", + "accept": false, + "reason": "Saving a dequantized model and gguf_file load handling are distinct issues." + }, + { + "left": "issue:44038", + "right": "issue:44295", + "accept": false, + "reason": "Qwen3-VL-Moe regression vs position_ids buffer read error; different code paths." + }, + { + "left": "issue:22355", + "right": "issue:29159", + "accept": false, + "reason": "Importing transformers.onnx and slow/fast tokenizer inconsistency are unrelated." + }, + { + "left": "issue:44038", + "right": "issue:44222", + "accept": false, + "reason": "Qwen3-VL-Moe load bug and FP8 save_pretrained bug do not match." + }, + { + "left": "issue:45216", + "right": "issue:45478", + "accept": false, + "reason": "Both involve Qwen3.5, but one is save_pretrained corruption and the other is from_pretrained failure; not the same fix." + }, + { + "left": "issue:43824", + "right": "issue:44351", + "accept": false, + "reason": "Missing Qwen2_5_VLForConditionalGeneration export vs missing HybridCache export." + }, + { + "left": "issue:43116", + "right": "issue:45200", + "accept": false, + "reason": "Example script multi-label output bug vs Gemma 4 mm_token_type_ids defaulting issue." + }, + { + "left": "issue:39401", + "right": "issue:46614", + "accept": false, + "reason": "Tokenizer offset mapping bug vs stale phonemizer language backend state." + }, + { + "left": "issue:43122", + "right": "issue:44779", + "accept": false, + "reason": "Different model tokenization regressions; not one concrete bug." + }, + { + "left": "issue:42757", + "right": "issue:44351", + "accept": false, + "reason": "Different missing symbols from different packages/modules; too broad to deduplicate." + }, + { + "left": "issue:36296", + "right": "issue:43116", + "accept": false, + "reason": "Tensor parallel training bug is unrelated to the classification example returning empty results." + }, + { + "left": "issue:43278", + "right": "issue:44945", + "accept": false, + "reason": "BF16->FP32 dtype drift in eval and pipeline-parallel output corruption are different defects." + }, + { + "left": "issue:22355", + "right": "issue:29127", + "accept": false, + "reason": "Import failure for transformers.onnx vs unclear LayoutLMv3 error messaging." + }, + { + "left": "issue:42175", + "right": "issue:44938", + "accept": false, + "reason": "Missing tensorflow backend in extras vs Python 3.14 load failure are not the same bug." + }, + { + "left": "issue:39379", + "right": "issue:43232", + "accept": false, + "reason": "mllama mask handling and generation kwargs after sync_gpus are unrelated." + }, + { + "left": "issue:43493", + "right": "issue:45753", + "accept": false, + "reason": "SigLIP2 implementation discrepancy vs Qwen3_5 image-specific infinite loop." + }, + { + "left": "issue:44360", + "right": "issue:44485", + "accept": false, + "reason": "DSA indexer ReLU discussion does not match GLM-5 RoPE implementation." + } + ] + }, + "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": 27, + "review_unit_index": 53, + "review_unit_count": 63, + "status": "reviewed", + "reason": null, + "source": "llm", + "cache_hit": false, + "model": "gpt-5.4-mini?service_tier=flex", + "cluster_id": "cluster-10000-514", + "nodes": [ + "issue:28348", + "issue:38617", + "issue:42371", + "issue:42673", + "issue:42796", + "issue:42898", + "issue:43066", + "issue:43116", + "issue:43493", + "issue:43611", + "issue:43618", + "issue:43825", + "issue:43827", + "issue:43874", + "issue:43937", + "issue:43976", + "issue:43992", + "issue:44038", + "issue:44075", + "issue:44188", + "issue:44355", + "issue:44485", + "issue:44610", + "issue:44625", + "issue:44704", + "issue:44779", + "issue:44936", + "issue:44995", + "issue:45072", + "issue:45083", + "issue:45278", + "issue:45357", + "issue:45446", + "issue:45478", + "issue:46002", + "issue:46231", + "issue:46328", + "issue:46614", + "issue:46640", + "issue:46697", + "issue:46710", + "issue:46726", + "issue:46829", + "issue:46877", + "issue:46909", + "issue:46981", + "issue:47116" + ], + "soft_pairs": [ + "issue:44188|issue:46328", + "issue:46231|issue:46877", + "issue:43937|issue:44485", + "issue:42898|issue:46614", + "issue:38617|issue:45278", + "issue:43976|issue:45446", + "issue:43611|issue:46002", + "issue:44038|issue:45478", + "issue:43066|issue:44779", + "issue:42673|issue:45072", + "issue:44355|issue:46231", + "issue:45357|issue:45478", + "issue:44625|issue:46002", + "issue:43116|issue:43493", + "issue:45083|issue:45446", + "issue:44610|issue:46909", + "issue:42371|issue:43874", + "issue:44075|issue:44936", + "issue:43618|issue:46829", + "issue:46726|issue:46981", + "issue:46328|issue:47116", + "issue:43992|issue:44704", + "issue:43066|issue:46710", + "issue:28348|issue:46640", + "issue:43116|issue:43825", + "issue:43827|issue:45446", + "issue:42673|issue:42796", + "issue:44995|issue:46697" + ], + "prepared_review_unit_hash": "14ef2d1104fc5d22140662a67451cc7d4e82f36f2d1d4cafdacaffed97f927aa", + "component_budget": { + "node_count": 514, + "item_count": 514, + "soft_pair_count": 1891, + "serialized_chars": 980895, + "estimated_input_tokens": 245224, + "estimated_eval_tokens": 490704 + }, + "budget": { + "node_count": 47, + "item_count": 47, + "soft_pair_count": 28, + "serialized_chars": 35236, + "estimated_input_tokens": 8809, + "estimated_eval_tokens": 17874 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": true, + "analyst_result": { + "summary": "This cluster is very heterogeneous: most items are unrelated issues that only share broad Transformers/v5/model-family vocabulary. None of the soft pairs look like true duplicates of the same concrete bug or change, so all soft edges should be rejected.", + "confidence": 0.34, + "canonical_issue_reason": "Issue 45278 is the best umbrella representative because it captures the broad v5 regression/import-breakage theme that appears most often across the set, even though the cluster is not a true duplicate group.", + "canonical_pr_reason": null, + "best_issue_reason": "Issue 45278 is the most representative global issue candidate due to its broad cross-version breakage framing and overlap with several v5-related regression reports.", + "best_pr_reason": null, + "soft_edge_verdicts": [ + { + "left": "issue:44188", + "right": "issue:46328", + "accept": false, + "reason": "Different bugs: attention-kernel divergence under torch.compile vs a SequenceFeatureExtractor padding performance issue." + }, + { + "left": "issue:46231", + "right": "issue:46877", + "accept": false, + "reason": "One is a docs parameter name mismatch; the other is import-time docstring lint output. No shared code-path bug." + }, + { + "left": "issue:43937", + "right": "issue:44485", + "accept": false, + "reason": "Both mention GLM-5, but one is invalid GenerationConfig handling and the other is a RoPE implementation concern." + }, + { + "left": "issue:42898", + "right": "issue:46614", + "accept": false, + "reason": "Both are tokenizer-related regressions, but one is clean-up spacing behavior in v5 and the other is stale phonemizer language state." + }, + { + "left": "issue:38617", + "right": "issue:45278", + "accept": false, + "reason": "45378 is a broad import-error umbrella; 38617 is a specific missing symbol import. Related context, but not the same concrete bug." + }, + { + "left": "issue:43976", + "right": "issue:45446", + "accept": false, + "reason": "Different platform/version compatibility problems: Python version support vs a PyTorch version guard for an import." + }, + { + "left": "issue:43611", + "right": "issue:46002", + "accept": false, + "reason": "Model loading with base_model_prefix is unrelated to NaNs in classification heads." + }, + { + "left": "issue:44038", + "right": "issue:45478", + "accept": false, + "reason": "Same broad Qwen3 MoE family, but different versions/models and different failure modes for from_pretrained." + }, + { + "left": "issue:43066", + "right": "issue:44779", + "accept": false, + "reason": "Both are v5 tokenizer regressions, but they concern different tokenizers and distinct symptoms." + }, + { + "left": "issue:42673", + "right": "issue:45072", + "accept": false, + "reason": "VRAM leakage in Qwen3ForCausalLM threads is unrelated to dtype mismatches in inference models." + }, + { + "left": "issue:44355", + "right": "issue:46231", + "accept": false, + "reason": "Compiled Python file errors have nothing to do with a documentation parameter typo." + }, + { + "left": "issue:45357", + "right": "issue:45478", + "accept": false, + "reason": "Both involve Qwen3.5 Moe, but one is save_pretrained key corruption and the other is a from_pretrained load error." + }, + { + "left": "issue:44625", + "right": "issue:46002", + "accept": false, + "reason": "num_labels propagation in Qwen3.5 is a config bug; NaNs in classification heads are a different training/init failure." + }, + { + "left": "issue:43116", + "right": "issue:43493", + "accept": false, + "reason": "A classification example script returning empty results is unrelated to a SigLIP2 implementation discrepancy." + }, + { + "left": "issue:45083", + "right": "issue:45446", + "accept": false, + "reason": "Feature-length computation in qwen3_omni_moe is unrelated to a PyTorch version check in flex_attention." + }, + { + "left": "issue:44610", + "right": "issue:46909", + "accept": false, + "reason": "An image processor sizing bug is unrelated to a broken documentation example." + }, + { + "left": "issue:42371", + "right": "issue:43874", + "accept": false, + "reason": "TF32 API guidance and a missing multimodal-token helper method are different issues." + }, + { + "left": "issue:44075", + "right": "issue:44936", + "accept": false, + "reason": "Optimizer argument handling and trainer train/evaluate sequencing are separate bugs." + }, + { + "left": "issue:43618", + "right": "issue:46829", + "accept": false, + "reason": "Both are output-field assignment issues, but they affect different model classes and outputs, so they are not the same concrete bug." + }, + { + "left": "issue:46726", + "right": "issue:46981", + "accept": false, + "reason": "AutoProcessor backend-missing error handling is unrelated to retry behavior on HTTP 429 responses." + }, + { + "left": "issue:46328", + "right": "issue:47116", + "accept": false, + "reason": "A padding performance issue is unrelated to a streaming pipeline regression." + }, + { + "left": "issue:43992", + "right": "issue:44704", + "accept": false, + "reason": "Missing embed_tokens weight loading is a model checkpoint issue; passing kwargs to cached_file is a separate AutoProcessor bug." + }, + { + "left": "issue:43066", + "right": "issue:46710", + "accept": false, + "reason": "Tokenizer decoder type regression and DeepSeek accuracy regression after upgrade are not the same bug." + }, + { + "left": "issue:28348", + "right": "issue:46640", + "accept": false, + "reason": "Flash Attention 2 support request for GPT2LMHeadModel is not the same as a generic flash-attention issue." + }, + { + "left": "issue:43116", + "right": "issue:43825", + "accept": false, + "reason": "Example-script multi-label output failure is unrelated to a pipeline error message regression." + }, + { + "left": "issue:43827", + "right": "issue:45446", + "accept": false, + "reason": "Docs still referencing pipeline() and a PyTorch version check bug are unrelated." + }, + { + "left": "issue:42673", + "right": "issue:42796", + "accept": false, + "reason": "VRAM leakage in multithreaded inference is unrelated to tokenizer BPE-marker handling." + }, + { + "left": "issue:44995", + "right": "issue:46697", + "accept": false, + "reason": "A second-forward stale-cache crash is unrelated to image over-padding in Swin2SR." + } + ] + }, + "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": 27, + "review_unit_index": 54, + "review_unit_count": 63, + "status": "reviewed", + "reason": null, + "source": "llm", + "cache_hit": false, + "model": "gpt-5.4-mini?service_tier=flex", + "cluster_id": "cluster-10000-514", + "nodes": [ + "issue:33453", + "issue:39379", + "issue:39401", + "issue:42222", + "issue:42617", + "issue:43064", + "issue:43408", + "issue:43441", + "issue:43450", + "issue:43673", + "issue:43688", + "issue:43742", + "issue:43824", + "issue:43937", + "issue:43986", + "issue:44117", + "issue:44263", + "issue:44291", + "issue:44380", + "issue:44410", + "issue:44466", + "issue:44704", + "issue:44779", + "issue:44829", + "issue:44977", + "issue:45071", + "issue:45081", + "issue:45092", + "issue:45216", + "issue:45310", + "issue:45341", + "issue:45405", + "issue:45446", + "issue:45561", + "issue:45563", + "issue:45588", + "issue:45632", + "issue:46002", + "issue:46077", + "issue:46328", + "issue:46512", + "issue:46752", + "issue:46762", + "issue:46897", + "issue:46981", + "issue:47000", + "issue:47059", + "issue:47110" + ], + "soft_pairs": [ + "issue:45092|issue:46762", + "issue:33453|issue:39401", + "issue:44263|issue:46762", + "issue:43824|issue:46981", + "issue:43064|issue:43673", + "issue:43673|issue:46328", + "issue:39379|issue:43688", + "issue:42222|issue:43441", + "issue:42222|issue:47000", + "issue:44380|issue:46077", + "issue:42222|issue:46752", + "issue:43937|issue:45563", + "issue:45405|issue:45446", + "issue:44117|issue:45588", + "issue:44291|issue:45071", + "issue:44466|issue:46002", + "issue:39379|issue:46897", + "issue:44410|issue:45310", + "issue:39401|issue:44779", + "issue:43450|issue:43986", + "issue:43937|issue:44977", + "issue:44704|issue:47110", + "issue:45071|issue:47059", + "issue:45341|issue:46328", + "issue:45071|issue:46512", + "issue:44829|issue:45216", + "issue:45341|issue:45561", + "issue:43408|issue:43742", + "issue:45632|issue:46981", + "issue:44117|issue:45081", + "issue:42617|issue:45341" + ], + "prepared_review_unit_hash": "238445268bd15ceb72341a7a4c7aec402af5eb70c8d434d2810f219d2eb152b9", + "component_budget": { + "node_count": 514, + "item_count": 514, + "soft_pair_count": 1891, + "serialized_chars": 980895, + "estimated_input_tokens": 245224, + "estimated_eval_tokens": 490704 + }, + "budget": { + "node_count": 48, + "item_count": 48, + "soft_pair_count": 31, + "serialized_chars": 37011, + "estimated_input_tokens": 9253, + "estimated_eval_tokens": 18762 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": true, + "analyst_result": { + "summary": "The cluster is loosely centered on Transformers issues, but the soft links mostly connect unrelated bugs. The broad tokenizer-loading regression is the best umbrella issue; no PRs are present.", + "confidence": 0.67, + "canonical_issue_reason": "Issue 33453 is the broadest and most generic: it describes a tokenizer-loading regression without narrowing to a specific model or code path, making it the best umbrella for nearby tokenizer-related reports.", + "canonical_pr_reason": null, + "best_issue_reason": "Issue 33453 is the strongest representative for this cluster because it captures the general failure mode shared by several tokenizer/loading complaints, while being broad enough to serve as the canonical issue.", + "best_pr_reason": null, + "soft_edge_verdicts": [ + { + "left": "issue:45092", + "right": "issue:46762", + "accept": false, + "reason": "Different failures: remote-code meta-initialization incompatibility vs Minimax MSA attention numerical differences." + }, + { + "left": "issue:33453", + "right": "issue:39401", + "accept": false, + "reason": "Both are tokenizer-related, but one is a broad loading regression and the other is a Qwen3 offset_mapping bug; different underlying defects." + }, + { + "left": "issue:44263", + "right": "issue:46762", + "accept": false, + "reason": "GlmMoeDsaIndexer torch.split behavior is unrelated to Minimax attention numerical drift." + }, + { + "left": "issue:43824", + "right": "issue:46981", + "accept": false, + "reason": "ImportError for a missing Qwen2.5-VL class is unrelated to AutoProcessor retry behavior on HTTP 429." + }, + { + "left": "issue:43064", + "right": "issue:43673", + "accept": false, + "reason": "FSDP2/PEFT optimizer-state corruption is a training/distributed bug, not the GenerationMixin cache issue." + }, + { + "left": "issue:43673", + "right": "issue:46328", + "accept": false, + "reason": "Generation cache missing and SequenceFeatureExtractor.pad performance are different code paths and symptoms." + }, + { + "left": "issue:39379", + "right": "issue:43688", + "accept": false, + "reason": "mllama mask handling and auxiliary-loss normalization are distinct model-specific bugs." + }, + { + "left": "issue:42222", + "right": "issue:43441", + "accept": false, + "reason": "vitpose model breakage and Ministral FlashAttention failure do not share the same root cause." + }, + { + "left": "issue:42222", + "right": "issue:47000", + "accept": false, + "reason": "vitpose breakage is unrelated to RTDetrLoss cost-matrix infeasibility." + }, + { + "left": "issue:44380", + "right": "issue:46077", + "accept": false, + "reason": "Both involve attention behavior, but GPT2 attention scaling and Gemma4 bidirectional-mask construction are different bugs." + }, + { + "left": "issue:42222", + "right": "issue:46752", + "accept": false, + "reason": "vitpose breakage is unrelated to apply_chat_template crashing on an empty list." + }, + { + "left": "issue:43937", + "right": "issue:45563", + "accept": false, + "reason": "GLM-5 GenerationConfig validation and paged-generate stale warning are separate issues." + }, + { + "left": "issue:45405", + "right": "issue:45446", + "accept": false, + "reason": "A dependency version bump being too high is unrelated to the flex_attention PyTorch version check." + }, + { + "left": "issue:44117", + "right": "issue:45588", + "accept": false, + "reason": "TOKENIZER_MAPPING_NAMES returning None and flash_attention crashing on s_aux=None are different failure sources." + }, + { + "left": "issue:44291", + "right": "issue:45071", + "accept": false, + "reason": "Both are v5-related regressions, but one is about init_empty_weights arguments and the other about PretrainedConfig type checking." + }, + { + "left": "issue:44466", + "right": "issue:46002", + "accept": false, + "reason": "lm_head serialization/tied-weights behavior is unrelated to NaNs in classification heads." + }, + { + "left": "issue:39379", + "right": "issue:46897", + "accept": false, + "reason": "mllama mask handling is unrelated to Florence2 training-loss double-shift behavior." + }, + { + "left": "issue:44410", + "right": "issue:45310", + "accept": false, + "reason": "Missing qwen3next projections and Qwen3.5 MoE from_pretrained errors are different model-specific problems." + }, + { + "left": "issue:39401", + "right": "issue:44779", + "accept": false, + "reason": "Both are tokenizer regressions, but they affect different models and appear to come from different tokenizer logic." + }, + { + "left": "issue:43450", + "right": "issue:43986", + "accept": false, + "reason": "Batched video processor shape issues are unrelated to the AutoProcessor/torchvision import crash." + }, + { + "left": "issue:43937", + "right": "issue:44977", + "accept": false, + "reason": "Invalid GenerationConfig for GLM-5 is unrelated to Qwen3.5 flash-attention generation failures." + }, + { + "left": "issue:44704", + "right": "issue:47110", + "accept": false, + "reason": "Passing kwargs to cached_file and special_tokens being a list are unrelated API bugs." + }, + { + "left": "issue:45071", + "right": "issue:47059", + "accept": false, + "reason": "PretrainedConfig type checking regression is unrelated to the hf hub CLI completion attribute breakage." + }, + { + "left": "issue:45341", + "right": "issue:46328", + "accept": false, + "reason": "A testing_utils bug and a SequenceFeatureExtractor.pad performance issue are not the same defect." + }, + { + "left": "issue:45071", + "right": "issue:46512", + "accept": false, + "reason": "PretrainedConfig type checking and stale config.dtype after from_config are separate configuration issues." + }, + { + "left": "issue:44829", + "right": "issue:45216", + "accept": false, + "reason": "FlashAttention-3 training degeneration and Qwen3.5 checkpoint serialization are different bugs." + }, + { + "left": "issue:45341", + "right": "issue:45561", + "accept": false, + "reason": "The xdist captured_info.txt race is more specific than the unrelated general testing_utils bug." + }, + { + "left": "issue:43408", + "right": "issue:43742", + "accept": false, + "reason": "sam3_tracker warning from sam3_video and MobileLLM-125M KeyError are unrelated loading issues." + }, + { + "left": "issue:45632", + "right": "issue:46981", + "accept": false, + "reason": "Cache-path collisions for trust_remote_code and repeated 429 retries are unrelated loader/network problems." + }, + { + "left": "issue:44117", + "right": "issue:45081", + "accept": false, + "reason": "Both hit tokenizer loading, but one is a mapping-name assumption and the other is a Mistral regex patch crash." + }, + { + "left": "issue:42617", + "right": "issue:45341", + "accept": false, + "reason": "3d_parallel.py execution failure is unrelated to the testing_utils 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": 27, + "review_unit_index": 55, + "review_unit_count": 63, + "status": "reviewed", + "reason": null, + "source": "llm", + "cache_hit": false, + "model": "gpt-5.4-mini?service_tier=flex", + "cluster_id": "cluster-10000-514", + "nodes": [ + "issue:27957", + "issue:38617", + "issue:39401", + "issue:41628", + "issue:42222", + "issue:42371", + "issue:43066", + "issue:43097", + "issue:43122", + "issue:43208", + "issue:43352", + "issue:43673", + "issue:43874", + "issue:43931", + "issue:43937", + "issue:44075", + "issue:44077", + "issue:44220", + "issue:44261", + "issue:44263", + "issue:44360", + "issue:44410", + "issue:44483", + "issue:44484", + "issue:44485", + "issue:44493", + "issue:44496", + "issue:44877", + "issue:44933", + "issue:45071", + "issue:45278", + "issue:45310", + "issue:45341", + "issue:45464", + "issue:45478", + "issue:45656", + "issue:46172", + "issue:46178", + "issue:46209", + "issue:46231", + "issue:46298", + "issue:46617", + "issue:46752", + "issue:46762", + "issue:46894", + "issue:46981", + "issue:47116" + ], + "soft_pairs": [ + "issue:38617|issue:41628", + "issue:43122|issue:46298", + "issue:42222|issue:46178", + "issue:44877|issue:46231", + "issue:39401|issue:46209", + "issue:42371|issue:44261", + "issue:43937|issue:44496", + "issue:44484|issue:45310", + "issue:44261|issue:46762", + "issue:44933|issue:45278", + "issue:42371|issue:46617", + "issue:44410|issue:45478", + "issue:44075|issue:45656", + "issue:38617|issue:44220", + "issue:44263|issue:44485", + "issue:43066|issue:43122", + "issue:43208|issue:46894", + "issue:43097|issue:44493", + "issue:46752|issue:47116", + "issue:44077|issue:46981", + "issue:43673|issue:45341", + "issue:43931|issue:44410", + "issue:41628|issue:45071", + "issue:44483|issue:45464", + "issue:43874|issue:44261", + "issue:44484|issue:45478", + "issue:27957|issue:43352", + "issue:44263|issue:46172", + "issue:44360|issue:46172" + ], + "prepared_review_unit_hash": "9cd87b2350cc1f9a7b7ecb270d96e24f174b409a237e847c84f380d0b29b6875", + "component_budget": { + "node_count": 514, + "item_count": 514, + "soft_pair_count": 1891, + "serialized_chars": 980895, + "estimated_input_tokens": 245224, + "estimated_eval_tokens": 490704 + }, + "budget": { + "node_count": 47, + "item_count": 47, + "soft_pair_count": 29, + "serialized_chars": 35140, + "estimated_input_tokens": 8785, + "estimated_eval_tokens": 17826 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": true, + "analyst_result": { + "summary": "These issues are largely heterogeneous regressions across imports, tokenizers, model loading, training, and API behavior. None of the proposed soft pairs look like the same concrete bug/change, so I would reject them all. If a representative issue must be named, #45278 is the broadest umbrella report, but it is not a true canonical duplicate target for this set.", + "confidence": 0.88, + "canonical_issue_reason": "The set is too heterogeneous to have one true canonical issue; the proposed pairs do not share a single concrete failure mode.", + "canonical_pr_reason": null, + "best_issue_reason": "#45278 is the broadest umbrella report here ('many import errors after update from 4.57.0 to 5.5.0'), so it is the best representative issue if one must be picked.", + "best_pr_reason": null, + "soft_edge_verdicts": [ + { + "left": "issue:38617", + "right": "issue:41628", + "accept": false, + "reason": "Both are import errors, but they are about different missing symbols and likely different code paths." + }, + { + "left": "issue:43122", + "right": "issue:46298", + "accept": false, + "reason": "Tokenizer regression vs noisy logging; unrelated bugs." + }, + { + "left": "issue:42222", + "right": "issue:46178", + "accept": false, + "reason": "Model-specific breakage vs backend mapping/availability error; not the same fix." + }, + { + "left": "issue:44877", + "right": "issue:46231", + "accept": false, + "reason": "Config-loading strictness and a doc parameter typo are unrelated." + }, + { + "left": "issue:39401", + "right": "issue:46209", + "accept": false, + "reason": "Tokenizer offset mapping bug vs FP8/B200 inference failure; different layers." + }, + { + "left": "issue:42371", + "right": "issue:44261", + "accept": false, + "reason": "TF32 settings/API issue vs MLA layernorm epsilon mismatch; different problems." + }, + { + "left": "issue:43937", + "right": "issue:44496", + "accept": false, + "reason": "GLM-5 generation config validation vs model-type recognition on load; different failure modes." + }, + { + "left": "issue:44261", + "right": "issue:46762", + "accept": false, + "reason": "Both mention numerical differences, but they affect different models and different attention paths." + }, + { + "left": "issue:44933", + "right": "issue:45278", + "accept": false, + "reason": "A single missing import is not the same bug as a broad multi-import regression." + }, + { + "left": "issue:42371", + "right": "issue:46617", + "accept": false, + "reason": "TF32 behavior control and trust_remote_code cache loading are unrelated." + }, + { + "left": "issue:44410", + "right": "issue:45478", + "accept": false, + "reason": "Different Qwen-family loading bugs in different modules; not one concrete code-path problem." + }, + { + "left": "issue:44075", + "right": "issue:45656", + "accept": false, + "reason": "Unused SGD args and duplicate optimizer steps are different optimizer bugs." + }, + { + "left": "issue:38617", + "right": "issue:44220", + "accept": false, + "reason": "ImportError vs audio feature extraction failure; unrelated." + }, + { + "left": "issue:44263", + "right": "issue:44485", + "accept": false, + "reason": "Torch split return handling vs RoPE implementation; different GLM issues." + }, + { + "left": "issue:43066", + "right": "issue:43122", + "accept": false, + "reason": "Both are tokenizer-related regressions, but one is decoder-type mismatch and the other is changed tokenization output; not clearly the same bug." + }, + { + "left": "issue:43208", + "right": "issue:46894", + "accept": false, + "reason": "Different training blockers in different model families; no shared concrete fix." + }, + { + "left": "issue:43097", + "right": "issue:44493", + "accept": false, + "reason": "Removed embedding-tie API vs unexpected position-id keys; unrelated regressions." + }, + { + "left": "issue:46752", + "right": "issue:47116", + "accept": false, + "reason": "Empty chat-template crash vs streaming pipeline regression are different failures." + }, + { + "left": "issue:44077", + "right": "issue:46981", + "accept": false, + "reason": "PatchTSMixer init policy vs HTTP 429 retry behavior; not the same underlying issue." + }, + { + "left": "issue:43673", + "right": "issue:45341", + "accept": false, + "reason": "Missing generation cache in v5 chunked prefill vs a testing_utils bug; unrelated." + }, + { + "left": "issue:43931", + "right": "issue:44410", + "accept": false, + "reason": "Qwen3-VL weight-shape mismatch and missing Qwen3Next projections are separate model-loading defects." + }, + { + "left": "issue:41628", + "right": "issue:45071", + "accept": false, + "reason": "Missing AutoImageProcessor import and PreTrainedConfig type checking are different compatibility issues." + }, + { + "left": "issue:44483", + "right": "issue:45464", + "accept": false, + "reason": "General chat/completions rejection vs streaming-specific Qwen3.5 failure; not one concrete bug." + }, + { + "left": "issue:43874", + "right": "issue:44261", + "accept": false, + "reason": "Different multimodal/token-count vs layernorm-epsilon bugs." + }, + { + "left": "issue:44484", + "right": "issue:45478", + "accept": false, + "reason": "Save-pretrained shard size question vs Qwen3.5 load failure; unrelated." + }, + { + "left": "issue:27957", + "right": "issue:43352", + "accept": false, + "reason": "Both mention Flash Attention 2 support, but they concern different models and likely different implementations." + }, + { + "left": "issue:44263", + "right": "issue:46172", + "accept": false, + "reason": "Different GLM/DeepSeek code paths and different tensor-shape bugs." + }, + { + "left": "issue:44360", + "right": "issue:46172", + "accept": false, + "reason": "Missing ReLU in the DSA indexer is unrelated to the DeepSeek transpose 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": 27, + "review_unit_index": 56, + "review_unit_count": 63, + "status": "reviewed", + "reason": null, + "source": "llm", + "cache_hit": false, + "model": "gpt-5.4-mini?service_tier=flex", + "cluster_id": "cluster-10000-514", + "nodes": [ + "issue:38617", + "issue:42757", + "issue:42886", + "issue:42890", + "issue:43012", + "issue:43064", + "issue:43097", + "issue:43208", + "issue:43441", + "issue:43452", + "issue:43531", + "issue:43577", + "issue:43673", + "issue:43873", + "issue:43874", + "issue:43931", + "issue:43986", + "issue:44038", + "issue:44062", + "issue:44117", + "issue:44261", + "issue:44263", + "issue:44315", + "issue:44351", + "issue:44410", + "issue:44484", + "issue:44485", + "issue:44829", + "issue:44987", + "issue:45003", + "issue:45341", + "issue:45372", + "issue:45375", + "issue:45440", + "issue:45588", + "issue:45753", + "issue:46050", + "issue:46123", + "issue:46172", + "issue:46298", + "issue:46369", + "issue:46655", + "issue:46664", + "issue:46701", + "issue:46752", + "issue:46821", + "issue:46897", + "issue:46981" + ], + "soft_pairs": [ + "issue:43441|issue:44829", + "issue:43012|issue:43577", + "issue:43931|issue:44038", + "issue:43097|issue:43452", + "issue:43986|issue:45341", + "issue:44261|issue:44485", + "issue:45003|issue:45341", + "issue:42886|issue:46981", + "issue:45753|issue:46369", + "issue:43531|issue:45341", + "issue:44117|issue:44987", + "issue:43874|issue:44263", + "issue:43097|issue:45588", + "issue:43673|issue:46123", + "issue:44410|issue:45375", + "issue:45440|issue:46172", + "issue:44038|issue:44410", + "issue:38617|issue:42757", + "issue:43986|issue:45372", + "issue:43097|issue:46050", + "issue:43064|issue:44315", + "issue:43208|issue:46897", + "issue:43986|issue:44484", + "issue:46655|issue:46664", + "issue:46298|issue:46752", + "issue:38617|issue:44351", + "issue:42890|issue:46298", + "issue:44062|issue:44117", + "issue:43097|issue:43873", + "issue:43012|issue:46821", + "issue:44117|issue:46701" + ], + "prepared_review_unit_hash": "d5f2931e60060c4d371253c04616d3b11b8503b4b2182a026f6a0c71b12f0e9e", + "component_budget": { + "node_count": 514, + "item_count": 514, + "soft_pair_count": 1891, + "serialized_chars": 980895, + "estimated_input_tokens": 245224, + "estimated_eval_tokens": 490704 + }, + "budget": { + "node_count": 48, + "item_count": 48, + "soft_pair_count": 31, + "serialized_chars": 36924, + "estimated_input_tokens": 9231, + "estimated_eval_tokens": 18718 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": true, + "analyst_result": { + "summary": "This cluster is noisy: the issues span many unrelated regressions, import errors, and model-specific bugs. I don\u2019t see a single underlying duplicate bug/change tying them together, so no soft edges should be merged.", + "confidence": 0.92, + "canonical_issue_reason": "No single canonical duplicate: the items are mostly distinct failures across different models, processors, and API surfaces rather than one shared root cause.", + "canonical_pr_reason": null, + "best_issue_reason": "Issue 43097 is the best representative if one must be chosen: it\u2019s a clear, user-facing breaking-change report with broader API impact and some discussion/reference activity.", + "best_pr_reason": null, + "soft_edge_verdicts": [ + { + "left": "issue:43441", + "right": "issue:44829", + "accept": false, + "reason": "Both mention FlashAttention, but one is a Ministral-3 failure and the other is degenerate training with flash_attention_3; different code paths and symptoms." + }, + { + "left": "issue:43012", + "right": "issue:43577", + "accept": false, + "reason": "Both are dtype/precision complaints, but one is a PyTorch compile warning and the other is BLIP2 loading not casting qformer; not the same bug." + }, + { + "left": "issue:43931", + "right": "issue:44038", + "accept": false, + "reason": "Different Qwen3-VL/MoE loading problems with different failure modes; too broad to be the same underlying issue." + }, + { + "left": "issue:43097", + "right": "issue:43452", + "accept": false, + "reason": "One is a removed API/compatibility complaint, the other is a gguf_file preload regression; unrelated breakages." + }, + { + "left": "issue:43986", + "right": "issue:45341", + "accept": false, + "reason": "A processor import/dependency crash versus a testing_utils bug; no shared underlying code-path." + }, + { + "left": "issue:44261", + "right": "issue:44485", + "accept": false, + "reason": "Different model internals: missing rms_norm_eps precision issue versus GLM-5 RoPE implementation concerns." + }, + { + "left": "issue:45003", + "right": "issue:45341", + "accept": false, + "reason": "Unsafe sys.modules access in modeling_utils is unrelated to the testing_utils issue." + }, + { + "left": "issue:42886", + "right": "issue:46981", + "accept": false, + "reason": "Both involve hub/network behavior, but offline cache loading and repeated 429 retries are different failure modes." + }, + { + "left": "issue:45753", + "right": "issue:46369", + "accept": false, + "reason": "Different model families and symptoms: infinite loop on a specific image versus memory overflow." + }, + { + "left": "issue:43531", + "right": "issue:45341", + "accept": false, + "reason": "Sliding_window behavior in Qwen3-MoE has no relation to the testing utility bug." + }, + { + "left": "issue:44117", + "right": "issue:44987", + "accept": false, + "reason": "One is a tokenizer mapping assumption bug, the other is a model-loading failure for a specific repo; not the same root cause." + }, + { + "left": "issue:43874", + "right": "issue:44263", + "accept": false, + "reason": "Both touch GLM-related code, but image-patch counting and torch.split return values are separate bugs." + }, + { + "left": "issue:43097", + "right": "issue:45588", + "accept": false, + "reason": "Removed API visibility and flash_attention.py crashing on sink-less models are unrelated." + }, + { + "left": "issue:43673", + "right": "issue:46123", + "accept": false, + "reason": "Cache handling in GenerationMixin and final-batch is_last behavior in MaskGenerationPipeline are different code paths." + }, + { + "left": "issue:44410", + "right": "issue:45375", + "accept": false, + "reason": "Both are Qwen3-family issues, but missing projections and a missing config field are different problems." + }, + { + "left": "issue:45440", + "right": "issue:46172", + "accept": false, + "reason": "Different DeepSeek variants and different bugs; no shared concrete fix." + }, + { + "left": "issue:44038", + "right": "issue:44410", + "accept": false, + "reason": "Qwen3-VL-Moe failure versus qwen3next missing projections; same broad family only, not the same bug." + }, + { + "left": "issue:38617", + "right": "issue:42757", + "accept": false, + "reason": "Both are import errors, but they concern different symbols/modules and likely different compatibility issues." + }, + { + "left": "issue:43986", + "right": "issue:45372", + "accept": false, + "reason": "Missing torchvision dependency and mistral_common ReasoningEffort import failure are distinct processor-loading issues." + }, + { + "left": "issue:43097", + "right": "issue:46050", + "accept": false, + "reason": "API-removal complaint versus quantized model save/load corruption; unrelated." + }, + { + "left": "issue:43064", + "right": "issue:44315", + "accept": false, + "reason": "Different training-stack problems: FSDP2/PEFT optimizer state corruption versus Liger Kernel not applied with model_init." + }, + { + "left": "issue:43208", + "right": "issue:46897", + "accept": false, + "reason": "xLSTM training blockers and Florence2 double-shift loss bug are different model-specific issues." + }, + { + "left": "issue:43986", + "right": "issue:44484", + "accept": false, + "reason": "Processor crash without torchvision is unrelated to the max_shard_size question." + }, + { + "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; not the same concrete bug." + }, + { + "left": "issue:46298", + "right": "issue:46752", + "accept": false, + "reason": "Noisy logging and apply_chat_template crashing on empty input are unrelated." + }, + { + "left": "issue:38617", + "right": "issue:44351", + "accept": false, + "reason": "Two different missing-import symbols from different modules; same symptom class, not the same issue." + }, + { + "left": "issue:42890", + "right": "issue:46298", + "accept": false, + "reason": "A flaky test due to missing set_seed versus a logging message issue; unrelated." + }, + { + "left": "issue:44062", + "right": "issue:44117", + "accept": false, + "reason": "Tokenizer AddedToken argument duplication and TOKENIZER_MAPPING_NAMES returning None are different tokenizer-loading bugs." + }, + { + "left": "issue:43097", + "right": "issue:43873", + "accept": false, + "reason": "Breaking API removal and quantization offloading behavior are not the same bug." + }, + { + "left": "issue:43012", + "right": "issue:46821", + "accept": false, + "reason": "Precision warning during compile and TimesFM window_size AttributeError are unrelated." + }, + { + "left": "issue:44117", + "right": "issue:46701", + "accept": false, + "reason": "Tokenizer mapping assumptions and DiffusionGemma pipeline failure do not share a concrete fix path." + } + ] + }, + "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": 27, + "review_unit_index": 57, + "review_unit_count": 63, + "status": "reviewed", + "reason": null, + "source": "llm", + "cache_hit": false, + "model": "gpt-5.4-mini?service_tier=flex", + "cluster_id": "cluster-10000-514", + "nodes": [ + "issue:36683", + "issue:42757", + "issue:42890", + "issue:42907", + "issue:43012", + "issue:43064", + "issue:43066", + "issue:43097", + "issue:43208", + "issue:43352", + "issue:43502", + "issue:43550", + "issue:43582", + "issue:43824", + "issue:43825", + "issue:43827", + "issue:43906", + "issue:43976", + "issue:43986", + "issue:43992", + "issue:44016", + "issue:44077", + "issue:44162", + "issue:44222", + "issue:44368", + "issue:44380", + "issue:44451", + "issue:44484", + "issue:44655", + "issue:44704", + "issue:44734", + "issue:44936", + "issue:44938", + "issue:45103", + "issue:45478", + "issue:45479", + "issue:45584", + "issue:45712", + "issue:45753", + "issue:46132", + "issue:46172", + "issue:46178", + "issue:46298", + "issue:46365", + "issue:46421", + "issue:46710", + "issue:47059", + "issue:47116" + ], + "soft_pairs": [ + "issue:46172|issue:46710", + "issue:43502|issue:43992", + "issue:44704|issue:45103", + "issue:44936|issue:46365", + "issue:44451|issue:47116", + "issue:44380|issue:46421", + "issue:44938|issue:46178", + "issue:43986|issue:44016", + "issue:43064|issue:47116", + "issue:42907|issue:45753", + "issue:42907|issue:44655", + "issue:43012|issue:43550", + "issue:45712|issue:46298", + "issue:43012|issue:44368", + "issue:43992|issue:44484", + "issue:43827|issue:47116", + "issue:43208|issue:43550", + "issue:43352|issue:43976", + "issue:43352|issue:44380", + "issue:43066|issue:43906", + "issue:43825|issue:44655", + "issue:44734|issue:47116", + "issue:43097|issue:46298", + "issue:45584|issue:46365", + "issue:44380|issue:46132", + "issue:44162|issue:46172", + "issue:44077|issue:46365", + "issue:42890|issue:47059", + "issue:44222|issue:45478", + "issue:44380|issue:45479", + "issue:42757|issue:43824", + "issue:36683|issue:43582", + "issue:43064|issue:44380" + ], + "prepared_review_unit_hash": "ae360f47890d4d0efe43570df9688e951bcbf4500247495116fc4532d6cacfa9", + "component_budget": { + "node_count": 514, + "item_count": 514, + "soft_pair_count": 1891, + "serialized_chars": 980895, + "estimated_input_tokens": 245224, + "estimated_eval_tokens": 490704 + }, + "budget": { + "node_count": 48, + "item_count": 48, + "soft_pair_count": 33, + "serialized_chars": 37467, + "estimated_input_tokens": 9367, + "estimated_eval_tokens": 18990 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": true, + "analyst_result": { + "summary": "The set is a collection of mostly unrelated Transformers issues across imports, model configs, training, docs, and backend behavior. None of the soft pairs look like true duplicates or the same concrete change, and there are no PRs in the cluster.", + "confidence": 0.97, + "canonical_issue_reason": "No canonical issue: the items span many unrelated bugs and feature regressions, so there is no single underlying issue to dedupe against.", + "canonical_pr_reason": null, + "best_issue_reason": "Issue 47116 is the broadest/recentest regression report and has inbound references, so it is the closest thing to a representative issue, but it is still not a true canonical for this mixed cluster.", + "best_pr_reason": null, + "soft_edge_verdicts": [ + { + "left": "issue:46172", + "right": "issue:46710", + "accept": false, + "reason": "Different bugs: a DeepSeek-v4 tensor expression issue vs an accuracy regression after upgrading Transformers." + }, + { + "left": "issue:43502", + "right": "issue:43992", + "accept": false, + "reason": "Different code paths: offline-file fetching vs missing UMT5 encoder weights on load." + }, + { + "left": "issue:44704", + "right": "issue:45103", + "accept": false, + "reason": "Both involve loading/documentation plumbing, but one is cached_file kwargs in AutoProcessor while the other is an auto_docstring annotation crash." + }, + { + "left": "issue:44936", + "right": "issue:46365", + "accept": false, + "reason": "Trainer.evaluate failure after training is unrelated to importing Trainer being broken by installed datasets." + }, + { + "left": "issue:44451", + "right": "issue:47116", + "accept": false, + "reason": "Model load failure for a specific checkpoint vs a streaming pipeline regression; not the same bug." + }, + { + "left": "issue:44380", + "right": "issue:46421", + "accept": false, + "reason": "GPT2 attention scaling ignored in SDPA/FlashAttention is unrelated to cache lazy initialization recompilation." + }, + { + "left": "issue:44938", + "right": "issue:46178", + "accept": false, + "reason": "Python 3.14 import/load failure is unrelated to a backend mapping ValueError for tensorflow_text." + }, + { + "left": "issue:43986", + "right": "issue:44016", + "accept": false, + "reason": "AutoProcessor/torchvision crash is unrelated to a notebook syntax error." + }, + { + "left": "issue:43064", + "right": "issue:47116", + "accept": false, + "reason": "Trainer/FSDP2/PEFT optimizer-state corruption is unrelated to streaming pipeline regressions." + }, + { + "left": "issue:42907", + "right": "issue:45753", + "accept": false, + "reason": "Dequantized save failure for Ministral/Devstral is unrelated to a Qwen3.5 infinite-loop inference bug." + }, + { + "left": "issue:42907", + "right": "issue:44655", + "accept": false, + "reason": "Saving dequantized models is not the same as saving Pipeline objects with save_pretrained." + }, + { + "left": "issue:43012", + "right": "issue:43550", + "accept": false, + "reason": "bfloat16 compile warnings and Bamba SDPA compile failure are different backend problems." + }, + { + "left": "issue:45712", + "right": "issue:46298", + "accept": false, + "reason": "Left is dummy PT object cleanup; right is a noisy logging message." + }, + { + "left": "issue:43012", + "right": "issue:44368", + "accept": false, + "reason": "bfloat16 compilation warning is unrelated to tie_word_embeddings warning spam during Qwen3.5 LoRA fine-tuning." + }, + { + "left": "issue:43992", + "right": "issue:44484", + "accept": false, + "reason": "Missing encoder weights on load is unrelated to save_pretrained shard size defaults." + }, + { + "left": "issue:43827", + "right": "issue:47116", + "accept": false, + "reason": "Docs still referencing pipeline() is separate from a runtime streaming pipeline regression." + }, + { + "left": "issue:43208", + "right": "issue:43550", + "accept": false, + "reason": "xLSTM training bugs are unrelated to Bamba torch.compile+SDPA failure." + }, + { + "left": "issue:43352", + "right": "issue:43976", + "accept": false, + "reason": "Flash Attention support for NemotronH is unrelated to a Python version compatibility issue." + }, + { + "left": "issue:43352", + "right": "issue:44380", + "accept": false, + "reason": "Nemotron FlashAttention support gating is unrelated to GPT2 attention-scaling being ignored." + }, + { + "left": "issue:43066", + "right": "issue:43906", + "accept": false, + "reason": "Wrong tokenizer decoder type in v5 is unrelated to an isolated reproduction of another issue." + }, + { + "left": "issue:43825", + "right": "issue:44655", + "accept": false, + "reason": "pipeline() translation messaging is unrelated to inability to save Pipeline objects." + }, + { + "left": "issue:44734", + "right": "issue:47116", + "accept": false, + "reason": "KV-cache continuation tensor indexing in serve is a different runtime path from streaming pipeline regression." + }, + { + "left": "issue:43097", + "right": "issue:46298", + "accept": false, + "reason": "Removed tie_embeddings_and_encoder_decoder API is unrelated to a noisy logging message." + }, + { + "left": "issue:45584", + "right": "issue:46365", + "accept": false, + "reason": "Whisper empty-transcription failure is unrelated to importing Trainer being broken by datasets." + }, + { + "left": "issue:44380", + "right": "issue:46132", + "accept": false, + "reason": "GPT2 SDPA/FlashAttention scaling bug is unrelated to AttentionInterface.register behavior changes." + }, + { + "left": "issue:44162", + "right": "issue:46172", + "accept": false, + "reason": "ESM2 breakage is unrelated to a DeepSeek-v4 tensor operation bug." + }, + { + "left": "issue:44077", + "right": "issue:46365", + "accept": false, + "reason": "patchtsmixer post_init validation is unrelated to the Trainer import conflict." + }, + { + "left": "issue:42890", + "right": "issue:47059", + "accept": false, + "reason": "Missing set_seed in a SAM HQ test is unrelated to a CLI breakage with huggingface_hub 1.22.0." + }, + { + "left": "issue:44222", + "right": "issue:45478", + "accept": false, + "reason": "FP8 save_pretrained for MoE and Qwen3.5 MoE from_pretrained are related to different save/load paths and not the same bug." + }, + { + "left": "issue:44380", + "right": "issue:45479", + "accept": false, + "reason": "Attention scaling ignored in backends is unrelated to degenerate zero-loss classification behavior." + }, + { + "left": "issue:42757", + "right": "issue:43824", + "accept": false, + "reason": "Import failure for is_offline_mode from huggingface_hub is unrelated to missing Qwen2_5_VL exports from transformers." + }, + { + "left": "issue:36683", + "right": "issue:43582", + "accept": false, + "reason": "Gemma3Config missing vocab_size is unrelated to an Apple Silicon caching_allocator_warmup TypeError." + }, + { + "left": "issue:43064", + "right": "issue:44380", + "accept": false, + "reason": "FSDP2/PEFT optimizer-state corruption is unrelated to GPT2 attention scaling being ignored in SDPA/FlashAttention." + } + ] + }, + "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": 27, + "review_unit_index": 58, + "review_unit_count": 63, + "status": "reviewed", + "reason": null, + "source": "llm", + "cache_hit": false, + "model": "gpt-5.4-mini?service_tier=flex", + "cluster_id": "cluster-10000-514", + "nodes": [ + "issue:29870", + "issue:36010", + "issue:36296", + "issue:36683", + "issue:38617", + "issue:42175", + "issue:42757", + "issue:42994", + "issue:43012", + "issue:43064", + "issue:43208", + "issue:43299", + "issue:43352", + "issue:43475", + "issue:43502", + "issue:43525", + "issue:43526", + "issue:43611", + "issue:43618", + "issue:43756", + "issue:43761", + "issue:43867", + "issue:43937", + "issue:43976", + "issue:44038", + "issue:44077", + "issue:44230", + "issue:44355", + "issue:44380", + "issue:44483", + "issue:44485", + "issue:44492", + "issue:44514", + "issue:44625", + "issue:44655", + "issue:44683", + "issue:44938", + "issue:45216", + "issue:45276", + "issue:45290", + "issue:46050", + "issue:46077", + "issue:46082", + "issue:46365", + "issue:46877", + "issue:47059", + "issue:47110" + ], + "soft_pairs": [ + "issue:42175|issue:46365", + "issue:43012|issue:44683", + "issue:44483|issue:44514", + "issue:38617|issue:43502", + "issue:43618|issue:44380", + "issue:43756|issue:44485", + "issue:43618|issue:43761", + "issue:45276|issue:46077", + "issue:36010|issue:42757", + "issue:43208|issue:43526", + "issue:44038|issue:44230", + "issue:42994|issue:44655", + "issue:43299|issue:43867", + "issue:44483|issue:45290", + "issue:44625|issue:46365", + "issue:43352|issue:43937", + "issue:44625|issue:46077", + "issue:43611|issue:44655", + "issue:44355|issue:46877", + "issue:36683|issue:43475", + "issue:44655|issue:46050", + "issue:43976|issue:44938", + "issue:44938|issue:47059", + "issue:36683|issue:44077", + "issue:43867|issue:45216", + "issue:36683|issue:47110", + "issue:46077|issue:46082", + "issue:43618|issue:45216", + "issue:36683|issue:43525", + "issue:29870|issue:36296", + "issue:43064|issue:44492" + ], + "prepared_review_unit_hash": "69313572dbed1b267fe50b41222360a5b46106694280795a4b69b44427d43ff9", + "component_budget": { + "node_count": 514, + "item_count": 514, + "soft_pair_count": 1891, + "serialized_chars": 980895, + "estimated_input_tokens": 245224, + "estimated_eval_tokens": 490704 + }, + "budget": { + "node_count": 47, + "item_count": 47, + "soft_pair_count": 31, + "serialized_chars": 36109, + "estimated_input_tokens": 9028, + "estimated_eval_tokens": 18312 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": true, + "analyst_result": { + "summary": "No strong duplicate cluster emerges here: the items span unrelated import, config, saving, training, and model-specific regressions. The soft edges look like broad thematic similarity only, so I would not merge any of the candidate pairs.", + "confidence": 0.95, + "canonical_issue_reason": null, + "canonical_pr_reason": null, + "best_issue_reason": null, + "best_pr_reason": null, + "soft_edge_verdicts": [ + { + "left": "issue:42175", + "right": "issue:46365", + "accept": false, + "reason": "Both mention environment/backend setup, but one is about missing TensorFlow in a torch install and the other about Trainer import breaking with datasets." + }, + { + "left": "issue:43012", + "right": "issue:44683", + "accept": false, + "reason": "Both involve compiled/precision-related warnings or failures, but they are different code paths and different failure modes." + }, + { + "left": "issue:44483", + "right": "issue:44514", + "accept": false, + "reason": "Both concern request/chat handling, but one is an API accept/reject issue and the other is a batched chat template crash." + }, + { + "left": "issue:38617", + "right": "issue:43502", + "accept": false, + "reason": "Both are import/offline-related at a high level, but one is an ImportError and the other is network calls despite local_files_only=True." + }, + { + "left": "issue:43618", + "right": "issue:44380", + "accept": false, + "reason": "Both are model backend regressions, but one is missing CLIP attentions and the other is GPT2 attention scaling being ignored." + }, + { + "left": "issue:43756", + "right": "issue:44485", + "accept": false, + "reason": "Both mention RoPE/model behavior, but they target different models and different implementation bugs." + }, + { + "left": "issue:43618", + "right": "issue:43761", + "accept": false, + "reason": "Both are CLIP regressions, but one is missing attentions and the other is hidden_states not returned; separate concrete bugs." + }, + { + "left": "issue:45276", + "right": "issue:46077", + "accept": false, + "reason": "Both are Gemma4 issues, but one is about token embedding resize propagation and the other about bidirectional attention masks." + }, + { + "left": "issue:36010", + "right": "issue:42757", + "accept": false, + "reason": "Both are import errors, but they concern different symbols from different dependencies and unrelated breakages." + }, + { + "left": "issue:43208", + "right": "issue:43526", + "accept": false, + "reason": "These are unrelated: xLSTM training bugs versus a BeitImageProcessorFast label-reduction bug." + }, + { + "left": "issue:44038", + "right": "issue:44230", + "accept": false, + "reason": "Both are Qwen3-VL/MoE/fp8-related, but one is a general loading bug and the other is fp8 inference support." + }, + { + "left": "issue:42994", + "right": "issue:44655", + "accept": false, + "reason": "Both involve save_pretrained, but one is quantized model saving and the other is Pipeline serialization." + }, + { + "left": "issue:43299", + "right": "issue:43867", + "accept": false, + "reason": "Both are model loading failures, but the roots differ: Qwen3VL MoE loading versus sorted state_dict handling." + }, + { + "left": "issue:44483", + "right": "issue:45290", + "accept": false, + "reason": "Both affect chat/completions-style input handling, but one is API request validation and the other is tool-call chat templating." + }, + { + "left": "issue:44625", + "right": "issue:46365", + "accept": false, + "reason": "These are unrelated configuration propagation and import problems." + }, + { + "left": "issue:43352", + "right": "issue:43937", + "accept": false, + "reason": "Different model-specific issues: Flash Attention support for Nemotron versus GenerationConfig validation for GLM-5." + }, + { + "left": "issue:44625", + "right": "issue:46077", + "accept": false, + "reason": "Same model family, but different bugs: num_labels propagation versus attention-mask construction." + }, + { + "left": "issue:43611", + "right": "issue:44655", + "accept": false, + "reason": "Both touch model/pipeline APIs, but one is a v5 loading regression and the other is pipeline save_pretrained serialization." + }, + { + "left": "issue:44355", + "right": "issue:46877", + "accept": false, + "reason": "These are unrelated: compiled Python file errors versus import-time docstring lint output." + }, + { + "left": "issue:36683", + "right": "issue:43475", + "accept": false, + "reason": "Different models and failures: Gemma3Config missing vocab_size versus SAM 3 Video missing fpn_position_embeddings." + }, + { + "left": "issue:44655", + "right": "issue:46050", + "accept": false, + "reason": "Both are save/load issues, but one is Pipeline serialization and the other is sinq quantized model round-trip breakage." + }, + { + "left": "issue:43976", + "right": "issue:44938", + "accept": false, + "reason": "Both are version/platform compatibility reports, but the Python version ranges and failure details do not match." + }, + { + "left": "issue:44938", + "right": "issue:47059", + "accept": false, + "reason": "Both are environment compatibility problems, but one is Python 3.14 loading and the other is a huggingface_hub CLI regression." + }, + { + "left": "issue:36683", + "right": "issue:44077", + "accept": false, + "reason": "Both are model/config regressions, but one is a missing config attribute and the other is an invalid optional post_init rule." + }, + { + "left": "issue:43867", + "right": "issue:45216", + "accept": false, + "reason": "Both mention loading/saving checkpoints, but the specific bugs are different and model-specific." + }, + { + "left": "issue:36683", + "right": "issue:47110", + "accept": false, + "reason": "The titles indicate different failures: missing vocab_size versus special_tokens being a list instead of a mapping." + }, + { + "left": "issue:46077", + "right": "issue:46082", + "accept": false, + "reason": "Both are Gemma4 config issues, but one is attention-mask semantics and the other is a head_dim validation constraint." + }, + { + "left": "issue:43618", + "right": "issue:45216", + "accept": false, + "reason": "Different regressions in different model paths: CLIP attentions versus Qwen3.5 checkpoint saving." + }, + { + "left": "issue:36683", + "right": "issue:43525", + "accept": false, + "reason": "Both are missing-config-attribute bugs, but they affect different models and different attributes." + }, + { + "left": "issue:29870", + "right": "issue:36296", + "accept": false, + "reason": "Both are training/generation bugs, but one is a BART generate() unbound variable and the other is a tensor-parallel training issue." + }, + { + "left": "issue:43064", + "right": "issue:44492", + "accept": false, + "reason": "These are unrelated training-state versus typo/documentation 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": 27, + "review_unit_index": 59, + "review_unit_count": 63, + "status": "reviewed", + "reason": null, + "source": "llm", + "cache_hit": false, + "model": "gpt-5.4-mini?service_tier=flex", + "cluster_id": "cluster-10000-514", + "nodes": [ + "issue:28282", + "issue:29870", + "issue:33453", + "issue:36296", + "issue:41669", + "issue:42890", + "issue:42898", + "issue:43010", + "issue:43065", + "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:44393", + "issue:44485", + "issue:44556", + "issue:44734", + "issue:44855", + "issue:44908", + "issue:44964", + "issue:45125", + "issue:45278", + "issue:45310", + "issue:45357", + "issue:45375", + "issue:45440", + "issue:45478", + "issue:45561", + "issue:45850", + "issue:46202", + "issue:46285", + "issue:46365", + "issue:46617", + "issue:46664", + "issue:46851", + "issue:47000" + ], + "soft_pairs": [ + "issue:43010|issue:45278", + "issue:43992|issue:45310", + "issue:44075|issue:46365", + "issue:44908|issue:46851", + "issue:44393|issue:47000", + "issue:29870|issue:33453", + "issue:43867|issue:46202", + "issue:44355|issue:44855", + "issue:43010|issue:45850", + "issue:36296|issue:43575", + "issue:43867|issue:45375", + "issue:41669|issue:44246", + "issue:42890|issue:45561", + "issue:43618|issue:44077", + "issue:43010|issue:43976", + "issue:43867|issue:45357", + "issue:43992|issue:45478", + "issue:28282|issue:33453", + "issue:43618|issue:46285", + "issue:44360|issue:45440", + "issue:42890|issue:44964", + "issue:42890|issue:44112", + "issue:43502|issue:44062", + "issue:42898|issue:46664", + "issue:44734|issue:46617", + "issue:44222|issue:45125", + "issue:43723|issue:44556", + "issue:43502|issue:44485", + "issue:43010|issue:43065", + "issue:43874|issue:46617" + ], + "prepared_review_unit_hash": "e1736205b64d12d596a06a35b47bf06cf28a27f16e28d0750e2a18ed72c08683", + "component_budget": { + "node_count": 514, + "item_count": 514, + "soft_pair_count": 1891, + "serialized_chars": 980895, + "estimated_input_tokens": 245224, + "estimated_eval_tokens": 490704 + }, + "budget": { + "node_count": 48, + "item_count": 48, + "soft_pair_count": 30, + "serialized_chars": 36307, + "estimated_input_tokens": 9077, + "estimated_eval_tokens": 18410 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": true, + "analyst_result": { + "summary": "Most pairs are unrelated and should stay separate. The only plausible duplicate is the import-time slowdown pair: one issue reports that `import transformers` is slow, and the other points to `import *` usage as the likely cause/fix.", + "confidence": 0.77, + "canonical_issue_reason": "issue:44246 is the clearest user-facing report of the import slowdown and is broad enough to serve as the anchor for the related performance regression discussion; issue:41669 is the more specific root-cause/fix-oriented counterpart.", + "canonical_pr_reason": null, + "best_issue_reason": "issue:44246 is the best overall issue candidate among the set because it captures the symptom most directly and has the strongest discussion/activity around the import-performance problem.", + "best_pr_reason": null, + "soft_edge_verdicts": [ + { + "left": "issue:43010", + "right": "issue:45278", + "accept": false, + "reason": "Different bugs: cache/layer `update()` no-grad semantics vs broad import errors after upgrade." + }, + { + "left": "issue:43992", + "right": "issue:45310", + "accept": false, + "reason": "Different models and failure modes; one is UMT5 encoder weights, the other is Qwen3.5 MoE loading." + }, + { + "left": "issue:44075", + "right": "issue:46365", + "accept": false, + "reason": "Optimizer argument handling is unrelated to `Trainer` import failure." + }, + { + "left": "issue:44908", + "right": "issue:46851", + "accept": false, + "reason": "Scheduler kwargs bug is unrelated to Qwen3.5 variable-length position-id handling." + }, + { + "left": "issue:44393", + "right": "issue:47000", + "accept": false, + "reason": "Both are vision/detection-adjacent, but they are different concrete failures and code paths." + }, + { + "left": "issue:29870", + "right": "issue:33453", + "accept": false, + "reason": "Bart generate-time decoder mask bug is unrelated to tokenizer loading regression." + }, + { + "left": "issue:43867", + "right": "issue:46202", + "accept": false, + "reason": "Different loading/roundtrip problems; no clear shared underlying bug." + }, + { + "left": "issue:44355", + "right": "issue:44855", + "accept": false, + "reason": "Both involve import/compile-time failures, but the causes are different (`compiled` files vs `torch.jit.script` parsing on Python 3.13)." + }, + { + "left": "issue:43010", + "right": "issue:45850", + "accept": false, + "reason": "No common code path; one is cache update semantics, the other is a broad complaints issue." + }, + { + "left": "issue:36296", + "right": "issue:43575", + "accept": false, + "reason": "Tensor-parallel training bug vs OOM when loading a specific large model; not the same defect." + }, + { + "left": "issue:43867", + "right": "issue:45375", + "accept": false, + "reason": "Unrelated model-loading issue vs missing Qwen3.5 vision config field." + }, + { + "left": "issue:41669", + "right": "issue:44246", + "accept": true, + "reason": "Same import-performance regression: `import transformers` is slow, and `import *` usage is presented as the underlying cause/fix for that slowdown." + }, + { + "left": "issue:42890", + "right": "issue:45561", + "accept": false, + "reason": "Test flakiness from missing seeds is unrelated to xdist race on `captured_info.txt`." + }, + { + "left": "issue:43618", + "right": "issue:44077", + "accept": false, + "reason": "Different model/API issues: CLIP attentions assignment vs patchtsmixer `post_init` policy." + }, + { + "left": "issue:43010", + "right": "issue:43976", + "accept": false, + "reason": "No shared bug; cache/layer updates vs Python-version compatibility." + }, + { + "left": "issue:43867", + "right": "issue:45357", + "accept": false, + "reason": "Generic model load error vs Qwen3.5 save-pretrained visual-key regression." + }, + { + "left": "issue:43992", + "right": "issue:45478", + "accept": false, + "reason": "Different issues despite both being Qwen-related; `UMT5Encoder` weight loading is not the same as Qwen3.5 MoE `from_pretrained` failure." + }, + { + "left": "issue:28282", + "right": "issue:33453", + "accept": false, + "reason": "PyTorch-missing ImportError is unrelated to tokenizer-loading regression." + }, + { + "left": "issue:43618", + "right": "issue:46285", + "accept": false, + "reason": "Different CLIP structure/attention issues; not the same underlying bug." + }, + { + "left": "issue:44360", + "right": "issue:45440", + "accept": false, + "reason": "A discussion about the DSA indexer is unrelated to DeepSeekV3MoE implementation divergence." + }, + { + "left": "issue:42890", + "right": "issue:44964", + "accept": false, + "reason": "Unrelated: SamHQ test seed flakiness vs Phi-4 multimodal loading failure." + }, + { + "left": "issue:42890", + "right": "issue:44112", + "accept": false, + "reason": "Test-seeding flakiness is unrelated to the GraniteSpeech device-override CI failure." + }, + { + "left": "issue:43502", + "right": "issue:44062", + "accept": false, + "reason": "Local-files-only network request bug is unrelated to `AddedToken` argument duplication." + }, + { + "left": "issue:42898", + "right": "issue:46664", + "accept": false, + "reason": "Tokenization cleanup behavior change is unrelated to SmolVLM video `num_frames` semantics." + }, + { + "left": "issue:44734", + "right": "issue:46617", + "accept": false, + "reason": "KV-cache continuation indexing crash is unrelated to trust_remote_code cache FileNotFoundError." + }, + { + "left": "issue:44222", + "right": "issue:45125", + "accept": false, + "reason": "Different Qwen3.5/MoE topics: FP8 save_pretrained vs missing tensor-parallel plan." + }, + { + "left": "issue:43723", + "right": "issue:44556", + "accept": false, + "reason": "Both are loading-related, but one is tokenizer-specific and the other is checkpoint reload after version upgrade; not the same concrete bug." + }, + { + "left": "issue:43502", + "right": "issue:44485", + "accept": false, + "reason": "local_files_only API bug is unrelated to GLM-5 RoPE implementation discussion." + }, + { + "left": "issue:43010", + "right": "issue:43065", + "accept": false, + "reason": "Cache/layer update semantics are unrelated to Sam3PixelDecoder's dummy Conv2d." + }, + { + "left": "issue:43874", + "right": "issue:46617", + "accept": false, + "reason": "ImageProcessorFast multimodal token counting bug is unrelated to cache FileNotFoundError." + } + ] + }, + "evaluator_result": { + "accept": true, + "feedback": "Grounded overall. The only accepted soft edge is the import-performance pair (41669/44246), which is well supported by the titles and activity; the other verdicts are conservatively rejected and do not overstate shared bugs." + }, + "evaluator_used": true, + "retried": false, + "accepted_nontrivial_soft_edge": true, + "error_kind": null, + "error_message": null + }, + { + "label": "issue", + "component_index": 1, + "component_count": 27, + "review_unit_index": 60, + "review_unit_count": 63, + "status": "reviewed", + "reason": null, + "source": "llm", + "cache_hit": false, + "model": "gpt-5.4-mini?service_tier=flex", + "cluster_id": "cluster-10000-514", + "nodes": [ + "issue:29870", + "issue:30333", + "issue:30990", + "issue:31795", + "issue:33453", + "issue:34689", + "issue:36296", + "issue:38617", + "issue:41628", + "issue:42548", + "issue:42757", + "issue:43010", + "issue:43232", + "issue:43299", + "issue:43352", + "issue:43475", + "issue:43502", + "issue:43576", + "issue:43824", + "issue:43827", + "issue:43867", + "issue:43901", + "issue:44222", + "issue:44230", + "issue:44315", + "issue:44351", + "issue:44355", + "issue:44360", + "issue:44387", + "issue:44485", + "issue:44556", + "issue:44749", + "issue:44961", + "issue:45071", + "issue:45125", + "issue:45375", + "issue:45468", + "issue:45478", + "issue:45850", + "issue:46132", + "issue:46143", + "issue:46144", + "issue:46209", + "issue:46617", + "issue:46664", + "issue:46726", + "issue:47059", + "issue:47110" + ], + "soft_pairs": [ + "issue:43824|issue:46617", + "issue:45850|issue:47110", + "issue:43576|issue:46132", + "issue:44387|issue:44556", + "issue:45125|issue:45478", + "issue:30333|issue:33453", + "issue:42548|issue:44351", + "issue:44355|issue:44749", + "issue:30990|issue:31795", + "issue:45125|issue:45375", + "issue:43502|issue:44222", + "issue:43502|issue:44961", + "issue:43827|issue:44961", + "issue:46664|issue:46726", + "issue:43010|issue:43232", + "issue:30333|issue:36296", + "issue:43299|issue:44556", + "issue:29870|issue:31795", + "issue:33453|issue:34689", + "issue:44355|issue:45071", + "issue:42757|issue:43576", + "issue:43576|issue:45071", + "issue:43576|issue:47059", + "issue:31795|issue:38617", + "issue:43576|issue:43824", + "issue:44961|issue:45468", + "issue:44315|issue:44360", + "issue:44485|issue:44961", + "issue:44351|issue:46617", + "issue:43352|issue:44961", + "issue:36296|issue:43901", + "issue:43475|issue:46617", + "issue:46143|issue:46144", + "issue:41628|issue:42548", + "issue:45375|issue:47110", + "issue:44222|issue:44230", + "issue:44315|issue:46209", + "issue:43867|issue:47110" + ], + "prepared_review_unit_hash": "6df9452d9cf54538036f3907f51009574a8abfa7c4c6a26fa0ad599cb272f926", + "component_budget": { + "node_count": 514, + "item_count": 514, + "soft_pair_count": 1891, + "serialized_chars": 980895, + "estimated_input_tokens": 245224, + "estimated_eval_tokens": 490704 + }, + "budget": { + "node_count": 48, + "item_count": 48, + "soft_pair_count": 38, + "serialized_chars": 38508, + "estimated_input_tokens": 9627, + "estimated_eval_tokens": 19510 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": true, + "analyst_result": { + "summary": "Most candidate pairs are only loosely related by subsystem or version and do not look like the same underlying bug. The only plausible duplicate pair is the RoFormer attention/kwargs issue.", + "confidence": 0.84, + "canonical_issue_reason": "issue:46144 is the most concrete and code-path-specific RoFormer report; it names the attention-interface gap that likely explains the kwargs passthrough symptom in issue:46143.", + "canonical_pr_reason": null, + "best_issue_reason": "issue:46144 is the best representative issue here because it is specific, actionable, and points to the likely root cause rather than a downstream symptom.", + "best_pr_reason": null, + "soft_edge_verdicts": [ + { + "left": "issue:43824", + "right": "issue:46617", + "accept": false, + "reason": "Different failures: missing exported Qwen2_5_VL symbol vs trust_remote_code cache FileNotFoundError." + }, + { + "left": "issue:45850", + "right": "issue:47110", + "accept": false, + "reason": "One is a broad complaint about v5 regressions; the other is a specific special_tokens shape error." + }, + { + "left": "issue:43576", + "right": "issue:46132", + "accept": false, + "reason": "CLI/env command breakage is unrelated to AttentionInterface.register behavior changes." + }, + { + "left": "issue:44387", + "right": "issue:44556", + "accept": false, + "reason": "Int4 quantization OOM is a different problem from checkpoint reload incompatibility across versions." + }, + { + "left": "issue:45125", + "right": "issue:45478", + "accept": false, + "reason": "Both involve Qwen3.5 MoE, but one is tensor-parallel planning and the other is a from_pretrained loading error; not enough evidence of the same bug." + }, + { + "left": "issue:30333", + "right": "issue:33453", + "accept": false, + "reason": "MLFlow reporting on failed jobs is unrelated to tokenizer loading regression." + }, + { + "left": "issue:42548", + "right": "issue:44351", + "accept": false, + "reason": "Both are import errors, but for different symbols and likely different causes." + }, + { + "left": "issue:44355", + "right": "issue:44749", + "accept": false, + "reason": "Compiled-file errors are unrelated to the data-filtering slowdown after upgrade." + }, + { + "left": "issue:30990", + "right": "issue:31795", + "accept": false, + "reason": "Loading hang and forward() documentation confusion do not describe the same bug." + }, + { + "left": "issue:45125", + "right": "issue:45375", + "accept": false, + "reason": "Tensor-parallel plan support and a missing config field are distinct issues." + }, + { + "left": "issue:43502", + "right": "issue:44222", + "accept": false, + "reason": "Local-files-only network requests are unrelated to FP8 save_pretrained behavior." + }, + { + "left": "issue:43502", + "right": "issue:44961", + "accept": false, + "reason": "Local-files-only bug and an unrelated issue titled racoon are clearly different." + }, + { + "left": "issue:43827", + "right": "issue:44961", + "accept": false, + "reason": "Pipeline-removal docs cleanup is unrelated to racoon." + }, + { + "left": "issue:46664", + "right": "issue:46726", + "accept": false, + "reason": "Video num_frames semantics and missing-vision-backend image processor handling are different bugs." + }, + { + "left": "issue:43010", + "right": "issue:43232", + "accept": false, + "reason": "Both are generation-related, but one is about no_grad decoration of cache updates and the other about sync_gpus/model kwargs update ordering." + }, + { + "left": "issue:30333", + "right": "issue:36296", + "accept": false, + "reason": "MLFlow reporting is unrelated to tensor-parallel training failures." + }, + { + "left": "issue:43299", + "right": "issue:44556", + "accept": false, + "reason": "Qwen3VL MoE loading regression and general checkpoint reload regression are too broad and not clearly the same bug." + }, + { + "left": "issue:29870", + "right": "issue:31795", + "accept": false, + "reason": "Bart generate() UnboundLocalError is unrelated to forward() documentation wording." + }, + { + "left": "issue:33453", + "right": "issue:34689", + "accept": false, + "reason": "Tokenizer loading regression and Llama 3.2 vision model loading breakage are different concrete failures." + }, + { + "left": "issue:44355", + "right": "issue:45071", + "accept": false, + "reason": "Compiled-file execution errors are unrelated to PretrainedConfig type checking." + }, + { + "left": "issue:42757", + "right": "issue:43576", + "accept": false, + "reason": "huggingface_hub import failure and transformers env command breakage are different layers and symptoms." + }, + { + "left": "issue:43576", + "right": "issue:45071", + "accept": false, + "reason": "CLI command breakage is unrelated to type-checking behavior changes." + }, + { + "left": "issue:43576", + "right": "issue:47059", + "accept": false, + "reason": "Both mention CLI, but one is the env command in v5 and the other is a hub version-specific completion error." + }, + { + "left": "issue:31795", + "right": "issue:38617", + "accept": false, + "reason": "Documentation confusion and missing layer_type_validation import are unrelated." + }, + { + "left": "issue:43576", + "right": "issue:43824", + "accept": false, + "reason": "CLI env issues and missing model-class import are not the same bug." + }, + { + "left": "issue:44961", + "right": "issue:45468", + "accept": false, + "reason": "Unrelated issue titles; no shared underlying bug is evident." + }, + { + "left": "issue:44315", + "right": "issue:44360", + "accept": false, + "reason": "Liger Kernel application and DSA indexer ReLU are completely different areas." + }, + { + "left": "issue:44485", + "right": "issue:44961", + "accept": false, + "reason": "GLM-5 RoPE implementation discussion is unrelated to racoon." + }, + { + "left": "issue:44351", + "right": "issue:46617", + "accept": false, + "reason": "HybridCache import failure and cached trust_remote_code FileNotFoundError are different import/load problems." + }, + { + "left": "issue:43352", + "right": "issue:44961", + "accept": false, + "reason": "Flash Attention 2 support error and racoon are unrelated." + }, + { + "left": "issue:36296", + "right": "issue:43901", + "accept": false, + "reason": "Tensor-parallel training bug and TextClassificationPipeline docs issue are not the same bug." + }, + { + "left": "issue:43475", + "right": "issue:46617", + "accept": false, + "reason": "SAM 3 Video attribute error and trust_remote_code cache FileNotFoundError are different failures." + }, + { + "left": "issue:46143", + "right": "issue:46144", + "accept": true, + "reason": "Both describe the same RoFormer attention-path problem: kwargs not propagating because the attention implementation does not use the attention interface." + }, + { + "left": "issue:41628", + "right": "issue:42548", + "accept": false, + "reason": "AutoImageProcessor and PreTrainedModel import failures are different missing exports." + }, + { + "left": "issue:45375", + "right": "issue:47110", + "accept": false, + "reason": "Missing deepstack_visual_indexes config field and special_tokens list/dict error are unrelated." + }, + { + "left": "issue:44222", + "right": "issue:44230", + "accept": false, + "reason": "FP8 save_pretrained support and FP8 inference support are related broadly, but not the same concrete bug." + }, + { + "left": "issue:44315", + "right": "issue:46209", + "accept": false, + "reason": "model_init Liger Kernel application and FP8 inference on B200 are unrelated." + }, + { + "left": "issue:43867", + "right": "issue:47110", + "accept": false, + "reason": "State-dict sorting load error and special_tokens key access error are different bugs." + } + ] + }, + "evaluator_result": { + "accept": true, + "feedback": "The summary is well grounded in the packet, and the only accepted duplicate pair (issue:46143 / issue:46144) is supported by the titles and the stated RoFormer attention-interface linkage. The remaining soft-edge rejections are appropriately conservative and do not overstate evidence." + }, + "evaluator_used": true, + "retried": false, + "accepted_nontrivial_soft_edge": true, + "error_kind": null, + "error_message": null + }, + { + "label": "issue", + "component_index": 1, + "component_count": 27, + "review_unit_index": 61, + "review_unit_count": 63, + "status": "reviewed", + "reason": null, + "source": "llm", + "cache_hit": false, + "model": "gpt-5.4-mini?service_tier=flex", + "cluster_id": "cluster-10000-514", + "nodes": [ + "issue:31795", + "issue:33453", + "issue:41669", + "issue:42371", + "issue:42890", + "issue:43352", + "issue:43408", + "issue:43525", + "issue:43572", + "issue:43576", + "issue:43673", + "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:44485", + "issue:44492", + "issue:44556", + "issue:44623", + "issue:44704", + "issue:44861", + "issue:44908", + "issue:45003", + "issue:45071", + "issue:45092", + "issue:45125", + "issue:45230", + "issue:45341", + "issue:45362", + "issue:45468", + "issue:45542", + "issue:45753", + "issue:45850", + "issue:46095", + "issue:46218", + "issue:46328", + "issue:46664", + "issue:47059", + "issue:47110" + ], + "soft_pairs": [ + "issue:43525|issue:47110", + "issue:45362|issue:45850", + "issue:43673|issue:44492", + "issue:44485|issue:46664", + "issue:44273|issue:45542", + "issue:44623|issue:45230", + "issue:44038|issue:45125", + "issue:44273|issue:45003", + "issue:44360|issue:46664", + "issue:41669|issue:43673", + "issue:44393|issue:45071", + "issue:44393|issue:45753", + "issue:43576|issue:44556", + "issue:43525|issue:45850", + "issue:44556|issue:45092", + "issue:44230|issue:46095", + "issue:43572|issue:45850", + "issue:44492|issue:45850", + "issue:45850|issue:47059", + "issue:42371|issue:44492", + "issue:41669|issue:46218", + "issue:43408|issue:44315", + "issue:44246|issue:44273", + "issue:42890|issue:44908", + "issue:43576|issue:44908", + "issue:43352|issue:44393", + "issue:44295|issue:45230", + "issue:43976|issue:44393", + "issue:44297|issue:45230", + "issue:44908|issue:45071", + "issue:44492|issue:46328", + "issue:41669|issue:44351", + "issue:44704|issue:44908", + "issue:41669|issue:45341", + "issue:43976|issue:46218", + "issue:44077|issue:44861", + "issue:31795|issue:33453", + "issue:43576|issue:44273", + "issue:44360|issue:45468" + ], + "prepared_review_unit_hash": "ca1cec2966b21f7439247281d99da6c246a1dbfe81fef01a51f402466fbeceb1", + "component_budget": { + "node_count": 514, + "item_count": 514, + "soft_pair_count": 1891, + "serialized_chars": 980895, + "estimated_input_tokens": 245224, + "estimated_eval_tokens": 490704 + }, + "budget": { + "node_count": 47, + "item_count": 47, + "soft_pair_count": 39, + "serialized_chars": 38165, + "estimated_input_tokens": 9542, + "estimated_eval_tokens": 19340 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": true, + "analyst_result": { + "summary": "These issues are mostly separate regressions or feature bugs across different models/subsystems; the soft links are topic-similar but not true duplicates. The broad v5 regression meta-report is the closest representative, but there is no strong single canonical bug here.", + "confidence": 0.84, + "canonical_issue_reason": "issue:45850 is the broadest umbrella report about many Transformers v5 breakages, so it best represents the cluster overall.", + "canonical_pr_reason": null, + "best_issue_reason": "issue:45850 is the most general and central issue in this set; the rest are narrower, distinct failures.", + "best_pr_reason": null, + "soft_edge_verdicts": [ + { + "left": "issue:43525", + "right": "issue:47110", + "accept": false, + "reason": "Different failures: missing pad_token_id on Llama4Config vs a special_tokens structure error." + }, + { + "left": "issue:45362", + "right": "issue:45850", + "accept": false, + "reason": "One is a Qwen3.5 chat crash; the other is a broad v5 complaint, not the same bug." + }, + { + "left": "issue:43673", + "right": "issue:44492", + "accept": false, + "reason": "Chunked-prefill cache regression and a cache-strategy typo are unrelated." + }, + { + "left": "issue:44485", + "right": "issue:46664", + "accept": false, + "reason": "GLM-5 RoPE implementation discussion vs SmolVLM video num_frames semantics." + }, + { + "left": "issue:44273", + "right": "issue:45542", + "accept": false, + "reason": "Lazy loading malfunction and a TensorFlow/tensorboard backend error are different issues." + }, + { + "left": "issue:44623", + "right": "issue:45230", + "accept": false, + "reason": "A specific processor.save_pretrained missing-files bug vs a generic placeholder report." + }, + { + "left": "issue:44038", + "right": "issue:45125", + "accept": false, + "reason": "Qwen3-VL-Moe bug report vs missing _tp_plan for tensor parallelism." + }, + { + "left": "issue:44273", + "right": "issue:45003", + "accept": false, + "reason": "Lazy loading failure and unsafe sys.modules access are distinct code paths." + }, + { + "left": "issue:44360", + "right": "issue:46664", + "accept": false, + "reason": "DSA indexer activation issue vs SmolVLM video frame handling." + }, + { + "left": "issue:41669", + "right": "issue:43673", + "accept": false, + "reason": "Import-* performance regression and missing GenerationMixin cache are not the same problem." + }, + { + "left": "issue:44393", + "right": "issue:45071", + "accept": false, + "reason": "Qwen3-VL bounding-box output bug vs PretrainedConfig type-checking regression." + }, + { + "left": "issue:44393", + "right": "issue:45753", + "accept": false, + "reason": "Bounding-box hallucination/error and an infinite-loop-on-image bug are different failures." + }, + { + "left": "issue:43576", + "right": "issue:44556", + "accept": false, + "reason": "Broken `transformers env` command and checkpoint reload incompatibility are separate v5 regressions." + }, + { + "left": "issue:43525", + "right": "issue:45850", + "accept": false, + "reason": "Specific Llama4 config attribute bug vs a broad multi-model v5 breakage report." + }, + { + "left": "issue:44556", + "right": "issue:45092", + "accept": false, + "reason": "Checkpoint reload after upgrade and remote-code meta-init incompatibility are different load-path issues." + }, + { + "left": "issue:44230", + "right": "issue:46095", + "accept": false, + "reason": "FP8 inference support for Qwen3-VL/Qwen3.5 vs DeepSeekV4 weight conversion tooling." + }, + { + "left": "issue:43572", + "right": "issue:45850", + "accept": false, + "reason": "StableLmConfig pad_token_idx regression is a specific config bug, not the general v5 umbrella report." + }, + { + "left": "issue:44492", + "right": "issue:45850", + "accept": false, + "reason": "A cache-strategy typo is not the same as the broad v5 breakage meta-issue." + }, + { + "left": "issue:45850", + "right": "issue:47059", + "accept": false, + "reason": "Broad v5 breakage report vs a CLI incompatibility with huggingface_hub 1.22.0." + }, + { + "left": "issue:42371", + "right": "issue:44492", + "accept": false, + "reason": "TF32 settings deprecation and cache-strategy typo are unrelated." + }, + { + "left": "issue:41669", + "right": "issue:46218", + "accept": false, + "reason": "Import-* slowdown and failure to import Qwen3_5ForCausalLM are different symptoms." + }, + { + "left": "issue:43408", + "right": "issue:44315", + "accept": false, + "reason": "sam3_tracker model-type warning and Liger Kernel not applied with model_init are unrelated." + }, + { + "left": "issue:44246", + "right": "issue:44273", + "accept": false, + "reason": "General import slowdown and lazy loading failure are both performance/loading themes but not the same bug." + }, + { + "left": "issue:42890", + "right": "issue:44908", + "accept": false, + "reason": "Flaky SamHQ tests from missing seeds vs scheduler kwargs being ignored." + }, + { + "left": "issue:43576", + "right": "issue:44908", + "accept": false, + "reason": "Broken env command and inverse_sqrt kwargs handling are different regressions." + }, + { + "left": "issue:43352", + "right": "issue:44393", + "accept": false, + "reason": "Flash Attention 2 support error for Nemotron vs Qwen3-VL bounding-box output bug." + }, + { + "left": "issue:44295", + "right": "issue:45230", + "accept": false, + "reason": "position_ids buffer access error vs a generic placeholder issue." + }, + { + "left": "issue:43976", + "right": "issue:44393", + "accept": false, + "reason": "Python version compatibility break vs Qwen3-VL bbox hallucination." + }, + { + "left": "issue:44297", + "right": "issue:45230", + "accept": false, + "reason": "tokenizer_class mismatch on save_pretrained vs a generic bug report." + }, + { + "left": "issue:44908", + "right": "issue:45071", + "accept": false, + "reason": "Scheduler kwargs bug and PretrainedConfig type checking are unrelated." + }, + { + "left": "issue:44492", + "right": "issue:46328", + "accept": false, + "reason": "Cache-strategy typo and SequenceFeatureExtractor.pad performance issue are unrelated." + }, + { + "left": "issue:41669", + "right": "issue:44351", + "accept": false, + "reason": "Import-* slowdown and missing HybridCache import are different import problems." + }, + { + "left": "issue:44704", + "right": "issue:44908", + "accept": false, + "reason": "AutoProcessor cached_file kwargs propagation and scheduler kwargs handling are unrelated." + }, + { + "left": "issue:41669", + "right": "issue:45341", + "accept": false, + "reason": "Import-* slowdown and a testing_utils bug are not the same issue." + }, + { + "left": "issue:43976", + "right": "issue:46218", + "accept": false, + "reason": "Python 3.9/3.10 compatibility break and missing Qwen3_5ForCausalLM import are different." + }, + { + "left": "issue:44077", + "right": "issue:44861", + "accept": false, + "reason": "patchtsmixer post_init policy and _get_tied_weight_keys AttributeError are unrelated." + }, + { + "left": "issue:31795", + "right": "issue:33453", + "accept": false, + "reason": "Documentation confusion in model.forward vs tokenizer loading regression are different problems." + }, + { + "left": "issue:43576", + "right": "issue:44273", + "accept": false, + "reason": "Broken env command and lazy loading malfunction are both v5-related but not the same code-path bug." + }, + { + "left": "issue:44360", + "right": "issue:45468", + "accept": false, + "reason": "DSA indexer activation issue and Gemma-4 audio positional encoding bug 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": 27, + "review_unit_index": 62, + "review_unit_count": 63, + "status": "reviewed", + "reason": null, + "source": "llm", + "cache_hit": false, + "model": "gpt-5.4-mini?service_tier=flex", + "cluster_id": "cluster-10000-514", + "nodes": [ + "issue:28348", + "issue:29870", + "issue:30333", + "issue:31515", + "issue:31795", + "issue:34689", + "issue:36296", + "issue:38617", + "issue:41669", + "issue:42371", + "issue:42913", + "issue:43066", + "issue:43298", + "issue:43352", + "issue:43502", + "issue:43519", + "issue:43576", + "issue:43618", + "issue:43704", + "issue:44230", + "issue:44261", + "issue:44263", + "issue:44273", + "issue:44315", + "issue:44351", + "issue:44485", + "issue:44492", + "issue:44623", + "issue:44655", + "issue:44829", + "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", + "issue:46894", + "issue:47059", + "issue:47110" + ], + "soft_pairs": [ + "issue:43519|issue:46655", + "issue:44315|issue:44829", + "issue:44273|issue:44908", + "issue:45230|issue:45310", + "issue:43519|issue:44485", + "issue:45468|issue:46894", + "issue:43576|issue:46218", + "issue:44655|issue:44861", + "issue:43519|issue:44908", + "issue:41669|issue:43704", + "issue:31795|issue:34689", + "issue:44485|issue:45468", + "issue:43704|issue:44485", + "issue:46018|issue:46172", + "issue:43502|issue:43519", + "issue:42371|issue:43704", + "issue:45850|issue:46218", + "issue:43519|issue:44623", + "issue:44263|issue:45468", + "issue:44261|issue:45468", + "issue:43298|issue:46095", + "issue:43519|issue:43704", + "issue:43352|issue:46218", + "issue:44351|issue:46218", + "issue:44861|issue:47110", + "issue:42913|issue:43066", + "issue:44861|issue:47059", + "issue:28348|issue:45250", + "issue:44908|issue:44998", + "issue:44492|issue:44998", + "issue:44998|issue:45850", + "issue:43618|issue:44998", + "issue:44938|issue:44998", + "issue:44861|issue:45290", + "issue:31515|issue:38617", + "issue:44230|issue:44623", + "issue:28348|issue:43352", + "issue:29870|issue:31515", + "issue:31515|issue:36296", + "issue:30333|issue:31515", + "issue:31515|issue:31795", + "issue:44230|issue:45310", + "issue:43298|issue:44230" + ], + "prepared_review_unit_hash": "1fbe5c4aa7ad398858be3714edb9dc3f4843ec27bd86de6fba39536afc5ab5b1", + "component_budget": { + "node_count": 514, + "item_count": 514, + "soft_pair_count": 1891, + "serialized_chars": 980895, + "estimated_input_tokens": 245224, + "estimated_eval_tokens": 490704 + }, + "budget": { + "node_count": 48, + "item_count": 48, + "soft_pair_count": 43, + "serialized_chars": 39821, + "estimated_input_tokens": 9956, + "estimated_eval_tokens": 20168 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": true, + "analyst_result": { + "summary": "This set is mostly heterogeneous Transformers issues, and the soft-similarity links look like broad topical overlap rather than true duplicates. I would not merge any of the candidate pairs; the only reasonable canonical choice is a concrete, well-scoped bug report rather than one of the vague or generic issues.", + "confidence": 0.61, + "canonical_issue_reason": "issue:43519 is the most concrete and actionable report in the set, and it also sits at the center of several soft-similarity candidates, so it is the best representative issue if one must be chosen.", + "canonical_pr_reason": null, + "best_issue_reason": "issue:43519 has a specific failure mode, clear subsystem ownership, and enough detail to serve as the most useful anchor issue in this otherwise mixed group.", + "best_pr_reason": null, + "soft_edge_verdicts": [ + { + "left": "issue:43519", + "right": "issue:46655", + "accept": false, + "reason": "Both are processor bugs, but one is timestamp math in Qwen3VL and the other is padding in SmolVLM video processing; different code paths and failure modes." + }, + { + "left": "issue:44315", + "right": "issue:44829", + "accept": false, + "reason": "Liger Kernel application during model_init is unrelated to flash_attention_3 degenerating training; different mechanisms and fixes." + }, + { + "left": "issue:44273", + "right": "issue:44908", + "accept": false, + "reason": "Lazy loading behavior and inverse_sqrt scheduler kwargs are unrelated subsystems." + }, + { + "left": "issue:45230", + "right": "issue:45310", + "accept": false, + "reason": "Generic bug report versus a specific Qwen3.5 MoE from_pretrained regression; no shared underlying defect." + }, + { + "left": "issue:43519", + "right": "issue:44485", + "accept": false, + "reason": "Both mention model internals, but one is timestamp calculation and the other is RoPE implementation for GLM-5; not the same bug." + }, + { + "left": "issue:45468", + "right": "issue:46894", + "accept": false, + "reason": "Gemma-4 audio positional encoding and CohereASR training-loss shift bug are different models and different failures." + }, + { + "left": "issue:43576", + "right": "issue:46218", + "accept": false, + "reason": "Broken CLI env command and inability to import Qwen3_5ForCausalLM are unrelated." + }, + { + "left": "issue:44655", + "right": "issue:44861", + "accept": false, + "reason": "Saving Pipeline objects and _get_tied_weight_keys AttributeError are different save/load code paths." + }, + { + "left": "issue:43519", + "right": "issue:44908", + "accept": false, + "reason": "Timestamp calculation in a processor versus scheduler kwarg handling are unrelated." + }, + { + "left": "issue:41669", + "right": "issue:43704", + "accept": false, + "reason": "Import-star cleanup/performance issue is unrelated to VRAM leakage in Qwen3ForCausalLM." + }, + { + "left": "issue:31795", + "right": "issue:34689", + "accept": false, + "reason": "Documentation confusion around forward arguments is not the same as a Llama 3.2 90B Vision model-loading regression." + }, + { + "left": "issue:44485", + "right": "issue:45468", + "accept": false, + "reason": "GLM-5 RoPE implementation and Gemma-4 audio positional encoding are distinct model-specific issues." + }, + { + "left": "issue:43704", + "right": "issue:44485", + "accept": false, + "reason": "VRAM leak in Qwen3ForCausalLM is not the same problem as GLM-5 RoPE behavior." + }, + { + "left": "issue:46018", + "right": "issue:46172", + "accept": false, + "reason": "DeepSeek-V4 gating logic and tensor transpose behavior are unrelated." + }, + { + "left": "issue:43502", + "right": "issue:43519", + "accept": false, + "reason": "local_files_only network leakage is a loading/download issue, not a processor timestamp bug." + }, + { + "left": "issue:42371", + "right": "issue:43704", + "accept": false, + "reason": "TF32 settings guidance and VRAM leakage are unrelated." + }, + { + "left": "issue:45850", + "right": "issue:46218", + "accept": false, + "reason": "Broad complaint about version 5 breaking models does not match a specific Qwen3.5 import failure." + }, + { + "left": "issue:43519", + "right": "issue:44623", + "accept": false, + "reason": "Processor timestamp math and missing save_pretrained files are different bugs in different APIs." + }, + { + "left": "issue:44263", + "right": "issue:45468", + "accept": false, + "reason": "torch.split return values in GlmMoeDsaIndexer and Gemma-4 audio positional encoding are unrelated." + }, + { + "left": "issue:44261", + "right": "issue:45468", + "accept": false, + "reason": "Missing rms_norm_eps in MLA q_a_layernorm and Gemma-4 audio positional encoding are not the same issue." + }, + { + "left": "issue:43298", + "right": "issue:46095", + "accept": false, + "reason": "Continuous batching audio support and DeepSeek-V4 weight conversion are different feature requests." + }, + { + "left": "issue:43519", + "right": "issue:43704", + "accept": false, + "reason": "Processor timestamp calculation and VRAM leakage are different bug classes." + }, + { + "left": "issue:43352", + "right": "issue:46218", + "accept": false, + "reason": "NemotronHForCausalLM flash-attn support and Qwen3_5 import failure are separate model-specific issues." + }, + { + "left": "issue:44351", + "right": "issue:46218", + "accept": false, + "reason": "HybridCache import failure and Qwen3_5 module import failure are both import-related, but not the same missing symbol or code path." + }, + { + "left": "issue:44861", + "right": "issue:47110", + "accept": false, + "reason": "Tied-weight key handling and special_tokens list/dict mismatch are different attribute/type errors." + }, + { + "left": "issue:42913", + "right": "issue:43066", + "accept": false, + "reason": "Both concern tokenizer v5 behavior, but one is a broad regression report and the other is a specific decoder-type bug; not clearly the same underlying defect." + }, + { + "left": "issue:44861", + "right": "issue:47059", + "accept": false, + "reason": "AttributeError in tied weights logic and CLI completion breakage from huggingface_hub are unrelated." + }, + { + "left": "issue:28348", + "right": "issue:45250", + "accept": false, + "reason": "GPT2 flash attention 2.0 support request and generic flash attention 2.0 issue are too broad to be the same concrete bug." + }, + { + "left": "issue:44908", + "right": "issue:44998", + "accept": false, + "reason": "Scheduler kwarg handling and an unrelated low-information issue title do not indicate the same bug." + }, + { + "left": "issue:44492", + "right": "issue:44998", + "accept": false, + "reason": "Cache strategy typo and the unrelated generic issue are not duplicates." + }, + { + "left": "issue:44998", + "right": "issue:45850", + "accept": false, + "reason": "Generic 'Unemployment' issue and broad '5.0 breaks too many models' complaint are not the same underlying problem." + }, + { + "left": "issue:43618", + "right": "issue:44998", + "accept": false, + "reason": "CLIPOutput attentions regression is unrelated to the generic issue." + }, + { + "left": "issue:44938", + "right": "issue:44998", + "accept": false, + "reason": "Python 3.14 load failure is unrelated to the generic issue." + }, + { + "left": "issue:44861", + "right": "issue:45290", + "accept": false, + "reason": "Tied weight key crash and chat template tool-call crash are different code paths." + }, + { + "left": "issue:31515", + "right": "issue:38617", + "accept": false, + "reason": "Slow from_pretrained checkpoint loading and missing layer_type_validation import are unrelated." + }, + { + "left": "issue:44230", + "right": "issue:44623", + "accept": false, + "reason": "FP8 support request for Qwen3-VL/Qwen3.5 MoE is not the same as processor.save_pretrained missing files." + }, + { + "left": "issue:28348", + "right": "issue:43352", + "accept": false, + "reason": "Both mention flash attention 2.0, but they target different models and different support gaps." + }, + { + "left": "issue:29870", + "right": "issue:31515", + "accept": false, + "reason": "BART generate() decoder_attention_mask bug and slow checkpoint loading are unrelated." + }, + { + "left": "issue:31515", + "right": "issue:36296", + "accept": false, + "reason": "Checkpoint load slowness and tensor parallel training bug are different subsystems." + }, + { + "left": "issue:30333", + "right": "issue:31515", + "accept": false, + "reason": "MLFlow reporting on failed training and checkpoint loading slowness are unrelated." + }, + { + "left": "issue:31515", + "right": "issue:31795", + "accept": false, + "reason": "Load performance issue and forward-API documentation confusion are not the same bug." + }, + { + "left": "issue:44230", + "right": "issue:45310", + "accept": false, + "reason": "FP8 support request and Qwen3.5 MoE from_pretrained error are different reports." + }, + { + "left": "issue:43298", + "right": "issue:44230", + "accept": false, + "reason": "Continuous batching support for audio models and FP8 support for Qwen3 VL/MoE 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": 27, + "review_unit_index": 63, + "review_unit_count": 63, + "status": "reviewed", + "reason": null, + "source": "llm", + "cache_hit": false, + "model": "gpt-5.4-mini?service_tier=flex", + "cluster_id": "cluster-10000-514", + "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:43668|issue:43813", + "issue:43240|issue:43668", + "issue:37934|issue:43668", + "issue:42936|issue:43668" + ], + "prepared_review_unit_hash": "7cffb2df2d35e5ca4cdaa220cb5526b42e91319e1799a96559fb9dbe8854eef8", + "component_budget": { + "node_count": 514, + "item_count": 514, + "soft_pair_count": 1891, + "serialized_chars": 980895, + "estimated_input_tokens": 245224, + "estimated_eval_tokens": 490704 + }, + "budget": { + "node_count": 9, + "item_count": 9, + "soft_pair_count": 12, + "serialized_chars": 8608, + "estimated_input_tokens": 2152, + "estimated_eval_tokens": 4560 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": true, + "analyst_result": { + "summary": "This cluster is heterogeneous: the issues cover unrelated bugs in different models/components (norms, tokenizers, loss kwargs, quantization typos, attention interfaces). None of the soft-similarity pairs look like true duplicates or the same underlying fix.", + "confidence": 0.97, + "canonical_issue_reason": "No clear canonical issue exists; the items are about distinct code paths and models rather than one shared bug.", + "canonical_pr_reason": null, + "best_issue_reason": "No single issue is a better global representative than the others because the cluster does not center on one problem.", + "best_pr_reason": null, + "soft_edge_verdicts": [ + { + "left": "issue:43717", + "right": "issue:46144", + "accept": false, + "reason": "Different model subsystems and failure modes: Mamba/Mamba-2 init_weights behavior vs RoFormer attention interface usage." + }, + { + "left": "issue:42936", + "right": "issue:46018", + "accept": false, + "reason": "Mask2Former ignore_value handling is unrelated to DeepSeek-V4 shared-expert gating logic." + }, + { + "left": "issue:43668", + "right": "issue:44031", + "accept": false, + "reason": "A config type-hint bug in ModernBERT is unrelated to tokenizer regex warning behavior." + }, + { + "left": "issue:42936", + "right": "issue:43240", + "accept": false, + "reason": "One is about Mask2Former's ignore_value being unused; the other is about kwargs not reaching a loss function." + }, + { + "left": "issue:43240", + "right": "issue:46144", + "accept": false, + "reason": "Loss-function kwargs plumbing and RoFormer attention interface usage are different code paths and not the same bug." + }, + { + "left": "issue:42936", + "right": "issue:44031", + "accept": false, + "reason": "Mask2Former ignore_value handling does not match tokenizer regex warning regressions." + }, + { + "left": "issue:42936", + "right": "issue:46144", + "accept": false, + "reason": "These concern unrelated areas: segmentation ignore_value vs RoFormer attention implementation." + }, + { + "left": "issue:42936", + "right": "issue:43813", + "accept": false, + "reason": "Mask2Former loss/ignore_value handling is unrelated to a quantization attribute typo." + }, + { + "left": "issue:43668", + "right": "issue:43813", + "accept": false, + "reason": "ModernBERT norm_eps typing and orig_conversion.quantization_operation typo are separate issues." + }, + { + "left": "issue:43240", + "right": "issue:43668", + "accept": false, + "reason": "Loss kwargs propagation is unrelated to a ModernBERTConfig type-hint correction." + }, + { + "left": "issue:37934", + "right": "issue:43668", + "accept": false, + "reason": "Llama4TextL2Norm implementation question is unrelated to ModernBERT norm_eps typing." + }, + { + "left": "issue:42936", + "right": "issue:43668", + "accept": false, + "reason": "Mask2Former ignore_value usage and ModernBERT config typing do not share the same underlying defect." + } + ] + }, + "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": 27, + "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": "c9c803e0465eee2115af7207519ccad5e8e469eaf27b2847e6a1dfb5c6ef924d", + "component_budget": { + "node_count": 3, + "item_count": 3, + "soft_pair_count": 3, + "serialized_chars": 2873, + "estimated_input_tokens": 719, + "estimated_eval_tokens": 1694 + }, + "budget": { + "node_count": 3, + "item_count": 3, + "soft_pair_count": 3, + "serialized_chars": 2873, + "estimated_input_tokens": 719, + "estimated_eval_tokens": 1694 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": false, + "analyst_result": { + "summary": "These three issues are unrelated: tokenizer truncation behavior, learning-rate scheduler functionality, and Wav2Vec2 word boundary timestamps. No duplicate cluster is present.", + "confidence": 0.99, + "canonical_issue_reason": null, + "canonical_pr_reason": null, + "best_issue_reason": null, + "best_pr_reason": null, + "soft_edge_verdicts": [ + { + "left": "issue:10519", + "right": "issue:11307", + "accept": false, + "reason": "Different domains and requests: truncation strategy vs. Wav2Vec2 word time offsets; no shared bug or code path." + }, + { + "left": "issue:10519", + "right": "issue:11005", + "accept": false, + "reason": "One is about tokenizer truncation options, the other about ReduceLROnPlateau-like scheduling behavior; unrelated functionality." + }, + { + "left": "issue:11005", + "right": "issue:11307", + "accept": false, + "reason": "Learning-rate scheduler feature request and speech-model word offset timing 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": 3, + "component_count": 27, + "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": "75c108dea984b9477a33df43276a20d456a4de6a332c92836d6c9d9db894b0d8", + "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 issues do not appear to be true duplicates. Issue 16618 is a broad model-loading failure report, while 15980 is specifically about an unhelpful error when downloading a private model without login; 13244 is about Tapas tokenization and is unrelated.", + "confidence": 0.86, + "canonical_issue_reason": "Issue 16618 is the broadest and most central report in the cluster, but it still appears to be a distinct model-loading problem rather than the same bug as the other issues.", + "canonical_pr_reason": null, + "best_issue_reason": "16618 is the best candidate only because its title is the most general and likely to capture the core user-facing failure; however, it should not be treated as a duplicate of the others.", + "best_pr_reason": null, + "soft_edge_verdicts": [ + { + "left": "issue:13244", + "right": "issue:16618", + "accept": false, + "reason": "Tapas tokenization mismatches TensorFlow code is a separate preprocessing/tokenization issue, not a model-loading failure." + }, + { + "left": "issue:15980", + "right": "issue:16618", + "accept": false, + "reason": "Both mention model loading, but 15980 is specifically about a bad private-model auth/download error message, while 16618 is a different generic inability-to-load-model report." + } + ] + }, + "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": 27, + "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": "05b21069aa6759391557ab4327bc0537e58136d1223c57bcabed4e1482ddd03d", + "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 separate open community feature requests: one for Flash Attention 2 support across more architectures, and one for GGUF support across more architectures. They share a similar framing but target different capabilities and code paths, so they are not duplicates.", + "confidence": 0.97, + "canonical_issue_reason": null, + "canonical_pr_reason": null, + "best_issue_reason": "Neither issue is a good canonical representative for the other because they request different features; if forced to choose within this pair, neither should subsume the other.", + "best_pr_reason": null, + "soft_edge_verdicts": [ + { + "left": "issue:26350", + "right": "issue:33260", + "accept": false, + "reason": "Different feature requests (Flash Attention 2 vs GGUF support) with different implementation targets; similar wording alone is not enough to treat them as 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": 5, + "component_count": 27, + "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-27088-9", + "nodes": [ + "issue:27088", + "issue:28936", + "issue:38435", + "issue:43834", + "issue:44365", + "issue:44370", + "issue:46222", + "issue:46226", + "issue:46931" + ], + "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:43834|issue:46931", + "issue:44365|issue:44370", + "issue:44365|issue:46222", + "issue:44365|issue:46226", + "issue:44365|issue:46931", + "issue:44370|issue:46222", + "issue:44370|issue:46226", + "issue:44370|issue:46931", + "issue:46222|issue:46226", + "issue:46222|issue:46931", + "issue:46226|issue:46931" + ], + "prepared_review_unit_hash": "425f9c6e2531644cc11f4bdd9edbdd812c250eed2dc553101a54150fcf333a12", + "component_budget": { + "node_count": 9, + "item_count": 9, + "soft_pair_count": 29, + "serialized_chars": 15976, + "estimated_input_tokens": 3994, + "estimated_eval_tokens": 8244 + }, + "budget": { + "node_count": 9, + "item_count": 9, + "soft_pair_count": 29, + "serialized_chars": 15976, + "estimated_input_tokens": 3994, + "estimated_eval_tokens": 8244 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": false, + "analyst_result": { + "summary": "This cluster is mostly a set of separate docs-translation request issues for different languages. I only see duplicate-like evidence inside the exact same-title subclusters (the generic placeholder issues and the repeated Thai issue); the cross-language pairs should not be merged.", + "confidence": 0.81, + "canonical_issue_reason": "Issue 43834 is the cleanest representative of the exact-duplicate generic-template subcluster: it has the same title/body pattern as 44365 and 44370 and appears to be the earliest copy.", + "canonical_pr_reason": null, + "best_issue_reason": "43834 is the best issue to anchor the duplicate subcluster because it is one of the exact-title matches and is the most canonical/earliest-looking instance among them.", + "best_pr_reason": null, + "soft_edge_verdicts": [ + { + "left": "issue:27088", + "right": "issue:44370", + "accept": false, + "reason": "Different language request (Turkish vs generic template); not the same underlying issue." + }, + { + "left": "issue:27088", + "right": "issue:44365", + "accept": false, + "reason": "Different language request; similarity is only from the translation-template wording." + }, + { + "left": "issue:27088", + "right": "issue:43834", + "accept": false, + "reason": "Specific Turkish request vs generic placeholder issue; no evidence they are the same report." + }, + { + "left": "issue:27088", + "right": "issue:38435", + "accept": false, + "reason": "Different language target (Turkish vs Romanian); distinct translation requests." + }, + { + "left": "issue:27088", + "right": "issue:28936", + "accept": false, + "reason": "Different language target (Turkish vs Spanish); distinct requests despite similar template text." + }, + { + "left": "issue:28936", + "right": "issue:43834", + "accept": false, + "reason": "Spanish request vs generic placeholder issue; same template, not the same language request." + }, + { + "left": "issue:28936", + "right": "issue:44365", + "accept": false, + "reason": "Different language target; similarity comes from the shared docs-translation template." + }, + { + "left": "issue:28936", + "right": "issue:38435", + "accept": false, + "reason": "Different language targets (Spanish vs Romanian); not duplicates." + }, + { + "left": "issue:28936", + "right": "issue:44370", + "accept": false, + "reason": "Different language target; no evidence of one underlying issue." + }, + { + "left": "issue:38435", + "right": "issue:44370", + "accept": false, + "reason": "Romanian request vs generic template issue; not the same concrete translation request." + }, + { + "left": "issue:38435", + "right": "issue:44365", + "accept": false, + "reason": "Romanian request vs generic template issue; similarity is superficial." + }, + { + "left": "issue:38435", + "right": "issue:43834", + "accept": false, + "reason": "Romanian request vs generic placeholder issue; different underlying request." + }, + { + "left": "issue:38435", + "right": "issue:46222", + "accept": false, + "reason": "Romanian vs Thai request; clearly different issues." + }, + { + "left": "issue:38435", + "right": "issue:46226", + "accept": false, + "reason": "Romanian vs Thai request; clearly different issues." + }, + { + "left": "issue:43834", + "right": "issue:44370", + "accept": true, + "reason": "Exact same placeholder title/body pattern; looks like a duplicate copy of the same templated issue." + }, + { + "left": "issue:43834", + "right": "issue:44365", + "accept": true, + "reason": "Exact same placeholder title/body pattern; duplicate of the same templated issue." + }, + { + "left": "issue:43834", + "right": "issue:46222", + "accept": false, + "reason": "Different title/language text; no strong evidence they are the same issue." + }, + { + "left": "issue:43834", + "right": "issue:46226", + "accept": false, + "reason": "Different title/language text; not the same underlying request." + }, + { + "left": "issue:43834", + "right": "issue:46931", + "accept": false, + "reason": "Different language/text and no shared concrete request." + }, + { + "left": "issue:44365", + "right": "issue:44370", + "accept": true, + "reason": "Exact same placeholder title/body pattern; likely duplicate issue reports." + }, + { + "left": "issue:44365", + "right": "issue:46222", + "accept": false, + "reason": "Generic placeholder issue vs Thai-language request; not the same request." + }, + { + "left": "issue:44365", + "right": "issue:46226", + "accept": false, + "reason": "Generic placeholder issue vs Thai-language request; not the same request." + }, + { + "left": "issue:44365", + "right": "issue:46931", + "accept": false, + "reason": "Generic placeholder issue vs different Thai request; not duplicates." + }, + { + "left": "issue:44370", + "right": "issue:46222", + "accept": false, + "reason": "Different language/text; only the broad template is shared." + }, + { + "left": "issue:44370", + "right": "issue:46226", + "accept": false, + "reason": "Different language/text; no evidence of same underlying issue." + }, + { + "left": "issue:44370", + "right": "issue:46931", + "accept": false, + "reason": "Different language/text; not the same request." + }, + { + "left": "issue:46222", + "right": "issue:46226", + "accept": true, + "reason": "Identical Thai title/body pattern and near-identical metadata; strong duplicate signal." + }, + { + "left": "issue:46222", + "right": "issue:46931", + "accept": false, + "reason": "Different Thai titles; not enough to treat as the same issue." + }, + { + "left": "issue:46226", + "right": "issue:46931", + "accept": false, + "reason": "Different Thai titles; similarity is not sufficient for duplication." + } + ] + }, + "evaluator_result": { + "accept": true, + "feedback": "The assessment is grounded in the packet and stays conservative overall. The exact-title generic-template subcluster (43834/44365/44370) and the identical Thai pair (46222/46226) are the only places with strong duplicate evidence, while the cross-language pairs are appropriately rejected. The only mild caution is that some reasons could have been phrased a bit more cautiously, but nothing here rises to the level of overstating the evidence." + }, + "evaluator_used": true, + "retried": false, + "accepted_nontrivial_soft_edge": true, + "error_kind": null, + "error_message": null + }, + { + "label": "issue", + "component_index": 6, + "component_count": 27, + "review_unit_index": 1, + "review_unit_count": 1, + "status": "reviewed", + "reason": null, + "source": "llm", + "cache_hit": false, + "model": "gpt-5.4-mini?service_tier=flex", + "cluster_id": "cluster-36718-21", + "nodes": [ + "issue:36718", + "issue:36817", + "issue:36879", + "issue:37015", + "issue:41115", + "issue:41211", + "issue:42503", + "issue:42738", + "issue:42954", + "issue:43671", + "issue:43837", + "issue:43909", + "issue:44121", + "issue:44205", + "issue:45027", + "issue:45177", + "issue:45322", + "issue:46133", + "issue:46645", + "issue:46653", + "issue:47004" + ], + "soft_pairs": [ + "issue:36718|issue:47004", + "issue:36718|issue:41115", + "issue:36817|issue:41211", + "issue:36817|issue:43671", + "issue:36817|issue:36879", + "issue:36817|issue:42954", + "issue:36879|issue:41211", + "issue:36879|issue:43909", + "issue:36879|issue:44121", + "issue:36879|issue:43837", + "issue:36879|issue:43671", + "issue:37015|issue:43837", + "issue:37015|issue:44121", + "issue:37015|issue:43909", + "issue: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:42954", + "issue:42738|issue:45027", + "issue:42503|issue:42738", + "issue:42738|issue:42954", + "issue:42738|issue:46653", + "issue:42954|issue:44205", + "issue:42954|issue:43671", + "issue:42954|issue:45177", + "issue:42954|issue:46653", + "issue:43671|issue:43837", + "issue:43671|issue:45027", + "issue:43671|issue:45177", + "issue:43671|issue:44121", + "issue:43837|issue:43909", + "issue:43837|issue:45322", + "issue:43837|issue:44121", + "issue:43909|issue:44205", + "issue:43909|issue:46645", + "issue:44121|issue:45322", + "issue: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": "be9be1871bfb9ba7ba01a55ff65a8fa742c68de0d448b2ece0f22adf02505cce", + "component_budget": { + "node_count": 21, + "item_count": 21, + "soft_pair_count": 48, + "serialized_chars": 28619, + "estimated_input_tokens": 7155, + "estimated_eval_tokens": 14566 + }, + "budget": { + "node_count": 21, + "item_count": 21, + "soft_pair_count": 48, + "serialized_chars": 28619, + "estimated_input_tokens": 7155, + "estimated_eval_tokens": 14566 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": false, + "analyst_result": { + "summary": "This cluster is mostly a set of superficially similar model-request issues, not true duplicates. The soft edges connect different models, tasks, and modalities, so none should be merged as the same underlying change.", + "confidence": 0.96, + "canonical_issue_reason": "No single canonical duplicate issue stands out; the items are distinct model requests that only share broad wording.", + "canonical_pr_reason": null, + "best_issue_reason": "If a representative must be chosen, issue 36879 is the most generic model-add request, but it still does not represent the rest well enough to be a true canonical.", + "best_pr_reason": null, + "soft_edge_verdicts": [ + { + "left": "issue:36718", + "right": "issue:47004", + "accept": false, + "reason": "Both are keypoint-related, but RoMa matcher vs LoMa detector+matcher are different concrete requests and code paths." + }, + { + "left": "issue:36718", + "right": "issue:41115", + "accept": false, + "reason": "RoMa keypoint matcher and MiniCPM3 architecture are unrelated model requests." + }, + { + "left": "issue:36817", + "right": "issue:41211", + "accept": false, + "reason": "EuroBert config and DEIMv2 are different model families with no shared change." + }, + { + "left": "issue:36817", + "right": "issue:43671", + "accept": false, + "reason": "EuroBert config and Qwen3-TTS support are different modalities and fixes." + }, + { + "left": "issue:36817", + "right": "issue:36879", + "accept": false, + "reason": "EuroBert config and RF-DETR model addition are unrelated requests." + }, + { + "left": "issue:36817", + "right": "issue:42954", + "accept": false, + "reason": "EuroBert config and XiaoMi MiMo are distinct model-add requests." + }, + { + "left": "issue:36879", + "right": "issue:41211", + "accept": false, + "reason": "RF-DETR and DEIMv2 are different detectors; not the same concrete addition." + }, + { + "left": "issue:36879", + "right": "issue:43909", + "accept": false, + "reason": "RF-DETR vision support and LFM2.5 Audio 1.5B are different modalities." + }, + { + "left": "issue:36879", + "right": "issue:44121", + "accept": false, + "reason": "RF-DETR model support and a weight-sparse transformer proposal are unrelated." + }, + { + "left": "issue:36879", + "right": "issue:43837", + "accept": false, + "reason": "RF-DETR object detection and Qwen3-ASR audio transcription are different code paths." + }, + { + "left": "issue:36879", + "right": "issue:43671", + "accept": false, + "reason": "RF-DETR and Qwen3-TTS are unrelated model requests." + }, + { + "left": "issue:37015", + "right": "issue:43837", + "accept": false, + "reason": "NeoBERT and Qwen3-ASR do not target the same model or feature." + }, + { + "left": "issue:37015", + "right": "issue:44121", + "accept": false, + "reason": "NeoBERT support and the weight-sparse transformer request are different changes." + }, + { + "left": "issue:37015", + "right": "issue:43909", + "accept": false, + "reason": "NeoBERT and LFM2.5 Audio 1.5B are unrelated model requests." + }, + { + "left": "issue:36879", + "right": "issue:37015", + "accept": false, + "reason": "RF-DETR and NeoBERT are different models with separate implementation needs." + }, + { + "left": "issue:37015", + "right": "issue:42503", + "accept": false, + "reason": "NeoBERT and ModernVBERT are both BERT-like, but they are still distinct model families and not one duplicate request." + }, + { + "left": "issue:41115", + "right": "issue:43837", + "accept": false, + "reason": "MiniCPM3 architecture and Qwen3-ASR support are unrelated." + }, + { + "left": "issue:41115", + "right": "issue:44121", + "accept": false, + "reason": "MiniCPM3 architecture and weight-sparse transformer support are different concrete changes." + }, + { + "left": "issue:41115", + "right": "issue:42503", + "accept": false, + "reason": "MiniCPM3 architecture and ModernVBERT models are different model requests." + }, + { + "left": "issue:41115", + "right": "issue:45027", + "accept": false, + "reason": "MiniCPM3 architecture and Voxtral TTS support target different model paths." + }, + { + "left": "issue:41211", + "right": "issue:43671", + "accept": false, + "reason": "DEIMv2 and Qwen3-TTS are unrelated model additions." + }, + { + "left": "issue:42503", + "right": "issue:43837", + "accept": false, + "reason": "ModernVBERT models and Qwen3-ASR support are different tasks and models." + }, + { + "left": "issue:42503", + "right": "issue:43671", + "accept": false, + "reason": "ModernVBERT models and Qwen3-TTS support are unrelated." + }, + { + "left": "issue:42503", + "right": "issue:42954", + "accept": false, + "reason": "ModernVBERT models and XiaoMi MiMo are different requests." + }, + { + "left": "issue:42738", + "right": "issue:45027", + "accept": false, + "reason": "BERT-like models with RoPE and Voxtral TTS support are unrelated." + }, + { + "left": "issue:42503", + "right": "issue:42738", + "accept": false, + "reason": "Both are BERT-related, but ModernVBERT models and BERT-like models with RoPE are not the same concrete change." + }, + { + "left": "issue:42738", + "right": "issue:42954", + "accept": false, + "reason": "BERT-like models with RoPE and XiaoMi MiMo are distinct requests." + }, + { + "left": "issue:42738", + "right": "issue:46653", + "accept": false, + "reason": "A model architecture request and a person-named issue are clearly unrelated." + }, + { + "left": "issue:42954", + "right": "issue:44205", + "accept": false, + "reason": "XiaoMi MiMo and SAM3-LiteText are different models and modalities." + }, + { + "left": "issue:42954", + "right": "issue:43671", + "accept": false, + "reason": "MiMo and Qwen3-TTS are unrelated model requests." + }, + { + "left": "issue:42954", + "right": "issue:45177", + "accept": false, + "reason": "MiMo and DeepSeek-OCR2 are different model families." + }, + { + "left": "issue:42954", + "right": "issue:46653", + "accept": false, + "reason": "MiMo support and a person-named issue have no shared change." + }, + { + "left": "issue:43671", + "right": "issue:43837", + "accept": false, + "reason": "Qwen3-TTS and Qwen3-ASR are the same vendor family but different modalities and code paths, so they are not one duplicate fix." + }, + { + "left": "issue:43671", + "right": "issue:45027", + "accept": false, + "reason": "Qwen3-TTS and Voxtral TTS are different model families." + }, + { + "left": "issue:43671", + "right": "issue:44121", + "accept": false, + "reason": "Qwen3-TTS support and the weight-sparse transformer proposal are unrelated." + }, + { + "left": "issue:43837", + "right": "issue:43909", + "accept": false, + "reason": "Qwen3-ASR and LFM2.5 Audio 1.5B are both audio-related, but they are different concrete models and implementations." + }, + { + "left": "issue:43837", + "right": "issue:45322", + "accept": false, + "reason": "Qwen3-ASR and EUPE are unrelated model requests." + }, + { + "left": "issue:43837", + "right": "issue:44121", + "accept": false, + "reason": "Qwen3-ASR support and the weight-sparse transformer request are not the same fix." + }, + { + "left": "issue:43837", + "right": "issue:45027", + "accept": false, + "reason": "Qwen3-ASR and Voxtral TTS are different tasks, not a shared underlying bug or change." + }, + { + "left": "issue:37015", + "right": "issue:45027", + "accept": false, + "reason": "NeoBERT and Voxtral TTS are unrelated model requests." + }, + { + "left": "issue:43837", + "right": "issue:45177", + "accept": false, + "reason": "ASR support and DeepSeek-OCR2 are different modalities and code paths." + }, + { + "left": "issue:45322", + "right": "issue:46133", + "accept": false, + "reason": "EUPE and TIPSv2 are both vision-language themed, but they are distinct model additions and not one mergeable change." + }, + { + "left": "issue:42503", + "right": "issue:46653", + "accept": false, + "reason": "ModernVBERT models and a person-named issue are unrelated." + }, + { + "left": "issue:45027", + "right": "issue:46653", + "accept": false, + "reason": "Voxtral TTS support and a person-named issue 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": 7, + "component_count": 27, + "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": "1b784d7d1f6c281e5d561e3bc85c9f3ed43a1a4b684a297c9500f0d020b5aa8b", + "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 documentation request to add Pruna as an optimization framework, while the other is a Vision RoPE refactor. They do not describe the same underlying bug or change, so the soft duplicate edge should be rejected.", + "confidence": 0.98, + "canonical_issue_reason": "issue:46443 is the better representative issue because it is a substantive technical refactor with a clear code-change scope, whereas issue:38740 is a docs-only request.", + "canonical_pr_reason": null, + "best_issue_reason": "issue:46443 is the strongest issue in the set for canonicalization since it concerns an actual implementation change rather than a documentation addition.", + "best_pr_reason": null, + "soft_edge_verdicts": [ + { + "left": "issue:38740", + "right": "issue:46443", + "accept": false, + "reason": "Reject: these are about different topics and code paths; a docs request for an optimization framework is not the same as a Vision RoPE refactor." + } + ] + }, + "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": 27, + "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": "272f6d7d76045de2c1e4e4dffdbeb05e6eede6fe5ccaf2572d0d7310193f8bdb", + "component_budget": { + "node_count": 2, + "item_count": 2, + "soft_pair_count": 1, + "serialized_chars": 1480, + "estimated_input_tokens": 370, + "estimated_eval_tokens": 996 + }, + "budget": { + "node_count": 2, + "item_count": 2, + "soft_pair_count": 1, + "serialized_chars": 1480, + "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, while the other requests DeepSeek v3.2 support. They target different features and there is no shared implementation or explicit linkage.", + "confidence": 0.98, + "canonical_issue_reason": null, + "canonical_pr_reason": null, + "best_issue_reason": null, + "best_pr_reason": null, + "soft_edge_verdicts": [ + { + "left": "issue:41084", + "right": "issue:43037", + "accept": false, + "reason": "Different underlying requests with no overlapping code path or explicit relation: decoding behavior vs model/version support." + } + ] + }, + "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": 27, + "review_unit_index": 1, + "review_unit_count": 1, + "status": "reviewed", + "reason": null, + "source": "llm", + "cache_hit": false, + "model": "gpt-5.4-mini?service_tier=flex", + "cluster_id": "cluster-41306-2", + "nodes": [ + "issue:41306", + "issue:41826" + ], + "soft_pairs": [ + "issue:41306|issue:41826" + ], + "prepared_review_unit_hash": "08662c766402b314aa15b337c074af765dcb94face98ad64016c01bd1e56cb5c", + "component_budget": { + "node_count": 2, + "item_count": 2, + "soft_pair_count": 1, + "serialized_chars": 1585, + "estimated_input_tokens": 397, + "estimated_eval_tokens": 1050 + }, + "budget": { + "node_count": 2, + "item_count": 2, + "soft_pair_count": 1, + "serialized_chars": 1585, + "estimated_input_tokens": 397, + "estimated_eval_tokens": 1050 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": false, + "analyst_result": { + "summary": "These two issues are related only at a high level (memory-efficient tiled layers), but they describe different features and code paths: one is about exposing a TiledFusedLogitsLoss/FLCE flag in `from_pretrained()`, the other about adding TiledMLP for lower memory use. Not duplicates.", + "confidence": 0.95, + "canonical_issue_reason": "No clear canonical issue: the reports target different components and changes, so they should not be merged as one underlying bug or feature request.", + "canonical_pr_reason": null, + "best_issue_reason": "If one must be prioritized, #41306 is the more concrete and self-contained request because it specifies a precise `from_pretrained()` integration point and exact feature flag behavior.", + "best_pr_reason": null, + "soft_edge_verdicts": [ + { + "left": "issue:41306", + "right": "issue:41826", + "accept": false, + "reason": "Different requested changes: FLCE/TiledFusedLogitsLoss flag support in `from_pretrained()` versus integrating TiledMLP. They share a broad memory-optimization theme but not the same underlying bug or implementation." + } + ] + }, + "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": 27, + "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": "ba450d8b0c9274afde801856d5273eeb15c7d80f42539319be1004283991c4d6", + "component_budget": { + "node_count": 2, + "item_count": 2, + "soft_pair_count": 1, + "serialized_chars": 1584, + "estimated_input_tokens": 396, + "estimated_eval_tokens": 1048 + }, + "budget": { + "node_count": 2, + "item_count": 2, + "soft_pair_count": 1, + "serialized_chars": 1584, + "estimated_input_tokens": 396, + "estimated_eval_tokens": 1048 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": false, + "analyst_result": { + "summary": "These two closed issues are related only at a very broad Trainer-customization level, but they describe different underlying problems: column pruning customizability vs FLOPS estimation for custom trainers. They should not be treated as duplicates.", + "confidence": 0.93, + "canonical_issue_reason": "issue:43598 is the broader, more general Trainer customizability request and is the better representative of the cluster theme.", + "canonical_pr_reason": null, + "best_issue_reason": "issue:43598 is the better standalone issue because it frames a clearer, wider Trainer behavior change; issue:43600 is a narrower performance-metrics concern.", + "best_pr_reason": null, + "soft_edge_verdicts": [ + { + "left": "issue:43598", + "right": "issue:43600", + "accept": false, + "reason": "Different code-path problems: one is about `remove_unused_column` behavior in Trainer, the other is about FLOPS estimation for custom trainers. They are related in subsystem but not the same bug/change." + } + ] + }, + "evaluator_result": null, + "evaluator_used": false, + "retried": false, + "accepted_nontrivial_soft_edge": false, + "error_kind": null, + "error_message": null + }, + { + "label": "issue", + "component_index": 11, + "component_count": 27, + "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": "e559a5ffa6d39d1c36f87267a112e745f34c0895ca4cbe79f65017f63a0928c7", + "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 closed issues are both about expanding Qwen model support, but they target different model families and different AutoModel heads, so they do not look like duplicates.", + "confidence": 0.89, + "canonical_issue_reason": "issue #43684 is the broader, more general request and would be the better representative if this cluster were being summarized as a single issue topic.", + "canonical_pr_reason": null, + "best_issue_reason": "#43684 is the clearest canonical issue candidate because it describes a broader registration gap across two AutoModel classes, whereas #44405 is a narrower task-specific support request for a different model/config family.", + "best_pr_reason": null, + "soft_edge_verdicts": [ + { + "left": "issue:43684", + "right": "issue:44405", + "accept": false, + "reason": "Different underlying changes: Qwen3-Omni AutoModel/AutoModelForConditionalGeneration registration vs Qwen3.5 sequence-classification support for Qwen3_5Config. Same broad theme, but not 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": 12, + "component_count": 27, + "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": "5e885f5c0696f9e66b4e0269adfc40865e6a19833b484e4909ecc4dc333ea7ac", + "component_budget": { + "node_count": 4, + "item_count": 4, + "soft_pair_count": 6, + "serialized_chars": 4259, + "estimated_input_tokens": 1065, + "estimated_eval_tokens": 2386 + }, + "budget": { + "node_count": 4, + "item_count": 4, + "soft_pair_count": 6, + "serialized_chars": 4259, + "estimated_input_tokens": 1065, + "estimated_eval_tokens": 2386 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": false, + "analyst_result": { + "summary": "Three issues share the exact same generic title \u201chuggingface \u201d and were created within about a day of each other, so they likely duplicate the same report; the \u201cFett\u201d issue looks unrelated and should stay separate.", + "confidence": 0.71, + "canonical_issue_reason": "Issue 43845 is the earliest of the near-identical \u201chuggingface \u201d reports and is a reasonable representative for that duplicate set.", + "canonical_pr_reason": null, + "best_issue_reason": "Issue 43845 is the best issue candidate because it is an exact-title match with the other likely duplicates and is the earliest of the group.", + "best_pr_reason": null, + "soft_edge_verdicts": [ + { + "left": "issue:43835", + "right": "issue:43845", + "accept": false, + "reason": "Different titles with no shared substantive evidence; \u201cFett\u201d does not look like the same underlying issue as \u201chuggingface \u201d." + }, + { + "left": "issue:43835", + "right": "issue:43846", + "accept": false, + "reason": "Different titles with no shared substantive evidence; this does not look like a duplicate pair." + }, + { + "left": "issue:43835", + "right": "issue:43859", + "accept": false, + "reason": "Different titles with no shared substantive evidence; likely unrelated." + }, + { + "left": "issue:43845", + "right": "issue:43846", + "accept": true, + "reason": "Exact same title, very similar metadata, and close creation times strongly suggest the same underlying issue." + }, + { + "left": "issue:43845", + "right": "issue:43859", + "accept": true, + "reason": "Exact same title and close timing; these look like duplicate reports of the same issue." + }, + { + "left": "issue:43846", + "right": "issue:43859", + "accept": true, + "reason": "Exact same title and close timing; likely duplicate reports of the same issue." + } + ] + }, + "evaluator_result": { + "accept": true, + "feedback": "The summary is grounded in the packet: issues 43845, 43846, and 43859 share the exact same title and are created close together, while 43835 has a different title and no other shared evidence. The representative selection of 43845 as the earliest of the same-title group is reasonable, and the duplicate verdicts are phrased conservatively as likely rather than certain." + }, + "evaluator_used": true, + "retried": false, + "accepted_nontrivial_soft_edge": true, + "error_kind": null, + "error_message": null + }, + { + "label": "issue", + "component_index": 13, + "component_count": 27, + "review_unit_index": 1, + "review_unit_count": 1, + "status": "reviewed", + "reason": null, + "source": "llm", + "cache_hit": false, + "model": "gpt-5.4-mini?service_tier=flex", + "cluster_id": "cluster-44371-6", + "nodes": [ + "issue:44371", + "issue:44418", + "issue:44572", + "issue:44573", + "issue:44937", + "issue:47036" + ], + "soft_pairs": [ + "issue:44371|issue:44572", + "issue:44371|issue:44573", + "issue:44371|issue:44418", + "issue:44371|issue:47036", + "issue:44371|issue:44937", + "issue:44418|issue:44572", + "issue:44418|issue:44573", + "issue:44572|issue:44573", + "issue:44572|issue:47036", + "issue:44572|issue:44937", + "issue:44573|issue:47036", + "issue:44573|issue:44937" + ], + "prepared_review_unit_hash": "fc90d0a800bebf954041a1d5a2704d54281bda435c992099bf8dc5aa99dab37f", + "component_budget": { + "node_count": 6, + "item_count": 6, + "soft_pair_count": 12, + "serialized_chars": 7355, + "estimated_input_tokens": 1839, + "estimated_eval_tokens": 3934 + }, + "budget": { + "node_count": 6, + "item_count": 6, + "soft_pair_count": 12, + "serialized_chars": 7355, + "estimated_input_tokens": 1839, + "estimated_eval_tokens": 3934 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": false, + "analyst_result": { + "summary": "These issues do not look like a real duplicate set; they are mostly spammy or placeholder titles with no substantive body evidence. If a representative issue must be chosen, #47036 is the least spam-like, but there is no strong canonical duplicate here.", + "confidence": 0.94, + "canonical_issue_reason": "#47036 is the most specific and least spam-like item in the set, with at least one discussion activity point. Even so, it is not clearly a substantive canonical bug report.", + "canonical_pr_reason": null, + "best_issue_reason": "#47036 is the best available representative only by elimination; the others are generic spam/placeholder titles with no useful details.", + "best_pr_reason": null, + "soft_edge_verdicts": [ + { + "left": "issue:44371", + "right": "issue:44572", + "accept": false, + "reason": "Both are spam placeholders; no shared bug, code path, or concrete change to unify." + }, + { + "left": "issue:44371", + "right": "issue:44573", + "accept": false, + "reason": "Both look like unrelated spam entries, not duplicates of a real issue." + }, + { + "left": "issue:44371", + "right": "issue:44418", + "accept": false, + "reason": "One is spam and the other is a generic documentation suggestion; too little evidence of the same underlying issue." + }, + { + "left": "issue:44371", + "right": "issue:47036", + "accept": false, + "reason": "No shared concrete problem; only generic low-signal titles." + }, + { + "left": "issue:44371", + "right": "issue:44937", + "accept": false, + "reason": "Unrelated spam versus a specific quoted title; no evidence they describe the same issue." + }, + { + "left": "issue:44418", + "right": "issue:44572", + "accept": false, + "reason": "Generic documentation suggestion and spam placeholder are not the same underlying bug/change." + }, + { + "left": "issue:44418", + "right": "issue:44573", + "accept": false, + "reason": "A docs suggestion and a spam issue do not match on a concrete problem statement." + }, + { + "left": "issue:44572", + "right": "issue:44573", + "accept": false, + "reason": "Both are spammy placeholders, but there is no actual shared bug/change evidence to merge." + }, + { + "left": "issue:44572", + "right": "issue:47036", + "accept": false, + "reason": "No concrete overlap beyond generic similarity signals." + }, + { + "left": "issue:44572", + "right": "issue:44937", + "accept": false, + "reason": "The titles are unrelated and there is no issue-body evidence of the same underlying problem." + }, + { + "left": "issue:44573", + "right": "issue:47036", + "accept": false, + "reason": "No shared bug, code path, or explicit linkage; just low-signal similarity." + }, + { + "left": "issue:44573", + "right": "issue:44937", + "accept": false, + "reason": "Spam placeholder and quoted-title issue are not plausibly duplicates." + } + ] + }, + "evaluator_result": null, + "evaluator_used": false, + "retried": false, + "accepted_nontrivial_soft_edge": false, + "error_kind": null, + "error_message": null + }, + { + "label": "issue", + "component_index": 14, + "component_count": 27, + "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": "2c661c03e57095f24a6ad276c1ee75acce601a1580da44b11e3962e2d942925a", + "component_budget": { + "node_count": 2, + "item_count": 2, + "soft_pair_count": 1, + "serialized_chars": 1644, + "estimated_input_tokens": 411, + "estimated_eval_tokens": 1078 + }, + "budget": { + "node_count": 2, + "item_count": 2, + "soft_pair_count": 1, + "serialized_chars": 1644, + "estimated_input_tokens": 411, + "estimated_eval_tokens": 1078 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": false, + "analyst_result": { + "summary": "The two closed issues have the same title and very similar bodies, with no differentiating targets or files, so they appear to be duplicate reports of the same underlying energy-optimization topic. Issue #45145 is the better canonical choice because it was opened earlier and has an inbound reference.", + "confidence": 0.99, + "canonical_issue_reason": "#45145 is earlier and already has an inbound reference, while #45151 adds no distinct scope or identifiers.", + "canonical_pr_reason": null, + "best_issue_reason": "#45145 is the best issue to keep as the canonical duplicate because it is the earliest instance of the same report and has slightly more discussion linkage.", + "best_pr_reason": null, + "soft_edge_verdicts": [ + { + "left": "issue:45145", + "right": "issue:45151", + "accept": true, + "reason": "Same title and highly similar body with no distinct issue targets, filenames, or other differentiating evidence; these look like duplicate reports of the same issue." + } + ] + }, + "evaluator_result": { + "accept": true, + "feedback": "Grounded and conservative: both items share the exact title, have similar body sizes, no filenames or explicit targets, and only soft-similarity evidence is present. The canonical choice of #45145 is supported by earlier creation time and an inbound reference." + }, + "evaluator_used": true, + "retried": false, + "accepted_nontrivial_soft_edge": true, + "error_kind": null, + "error_message": null + }, + { + "label": "issue", + "component_index": 15, + "component_count": 27, + "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": 16, + "component_count": 27, + "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": "b44917473bddedc1ab0ea264ae70723ba4d972aa137f4bcd526739f6ee702e00", + "component_budget": { + "node_count": 2, + "item_count": 2, + "soft_pair_count": 1, + "serialized_chars": 1476, + "estimated_input_tokens": 369, + "estimated_eval_tokens": 994 + }, + "budget": { + "node_count": 2, + "item_count": 2, + "soft_pair_count": 1, + "serialized_chars": 1476, + "estimated_input_tokens": 369, + "estimated_eval_tokens": 994 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": false, + "analyst_result": { + "summary": "These two issues are not duplicates; they appear to be separate, unrelated personal-name/date entries with no shared bug report context or actionable technical content.", + "confidence": 0.96, + "canonical_issue_reason": null, + "canonical_pr_reason": null, + "best_issue_reason": null, + "best_pr_reason": null, + "soft_edge_verdicts": [ + { + "left": "issue:45762", + "right": "issue:45763", + "accept": false, + "reason": "Only superficial similarity in the titles; the names and dates differ and there is no shared issue context indicating 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": 17, + "component_count": 27, + "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": "7eb00c8d765d382c58e73d7189567008d3bbeb11114ce44a4a1be6d582e62684", + "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": "Two closed issues with unrelated personal-name/date titles and no body or code evidence; they do not look like duplicates of the same underlying bug.", + "confidence": 0.98, + "canonical_issue_reason": "Issue 45764 is the earliest-created item, so it is the only reasonable cluster representative if one must be chosen; however, it is not meaningfully better than 45765.", + "canonical_pr_reason": null, + "best_issue_reason": "Issue 45764 is marginally preferable as the cluster anchor only because it was created first and has the same low-information shape as 45765.", + "best_pr_reason": null, + "soft_edge_verdicts": [ + { + "left": "issue:45764", + "right": "issue:45765", + "accept": false, + "reason": "They only share a superficial formatting pattern in the titles; there is no shared problem statement, code path, or other evidence that they describe 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": 18, + "component_count": 27, + "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": "aba1664f0911fd1ff3a821cccfa763892c44ea5e4f818ba3fc81ef24195f12fc", + "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": "The two issues are unrelated: one requests a new Qwen3_5 token-classification/value-model class, while the other asks for FP8 training support. They do not describe the same underlying bug or feature.", + "confidence": 0.99, + "canonical_issue_reason": null, + "canonical_pr_reason": null, + "best_issue_reason": null, + "best_pr_reason": null, + "soft_edge_verdicts": [ + { + "left": "issue:45810", + "right": "issue:46295", + "accept": false, + "reason": "Different requests in different areas: model-class addition for token classification vs FP8 training support. No shared code path or concrete implementation overlap." + } + ] + }, + "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": 27, + "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": "16a18fb13f06f6a06b2f25033f00dfc7f5669d81a3c6d5aa012d90bc73016aa1", + "component_budget": { + "node_count": 5, + "item_count": 5, + "soft_pair_count": 8, + "serialized_chars": 5728, + "estimated_input_tokens": 1432, + "estimated_eval_tokens": 3120 + }, + "budget": { + "node_count": 5, + "item_count": 5, + "soft_pair_count": 8, + "serialized_chars": 5728, + "estimated_input_tokens": 1432, + "estimated_eval_tokens": 3120 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": false, + "analyst_result": { + "summary": "There are two clear duplicate issue pairs: 45870/45871 and 45872/45873. Issue 45869 is a separate dataset-promo topic and should not be merged with the DataForge example/service requests.", + "confidence": 0.96, + "canonical_issue_reason": "45870 is a clean representative of the exact-duplicate pair 45870/45871 and is the earliest item in that pair.", + "canonical_pr_reason": null, + "best_issue_reason": "45870 is the best single issue to keep as a representative, but note the cluster also contains a second, separate duplicate pair (45872/45873).", + "best_pr_reason": null, + "soft_edge_verdicts": [ + { + "left": "issue:45869", + "right": "issue:45870", + "accept": false, + "reason": "Different artifact types/topics: a dataset promotion versus an example request; not the same underlying issue." + }, + { + "left": "issue:45869", + "right": "issue:45871", + "accept": false, + "reason": "Different artifact types/topics: a dataset promotion versus an example request; not the same underlying issue." + }, + { + "left": "issue:45870", + "right": "issue:45871", + "accept": true, + "reason": "Exact title match and strong duplicate signals; same underlying issue." + }, + { + "left": "issue:45870", + "right": "issue:45872", + "accept": false, + "reason": "Related financial/DataForge theme, but the titles describe different requests (example prediction vs scraping service); not the same issue." + }, + { + "left": "issue:45870", + "right": "issue:45873", + "accept": false, + "reason": "Related financial/DataForge theme, but the titles describe different requests (example prediction vs scraping service); not the same issue." + }, + { + "left": "issue:45871", + "right": "issue:45872", + "accept": false, + "reason": "Related theme only; these are different requests with different scopes, so they should not be deduplicated." + }, + { + "left": "issue:45871", + "right": "issue:45873", + "accept": false, + "reason": "Related theme only; these are different requests with different scopes, so they should not be deduplicated." + }, + { + "left": "issue:45872", + "right": "issue:45873", + "accept": true, + "reason": "Exact title match and strong duplicate signals; same underlying issue." + } + ] + }, + "evaluator_result": { + "accept": true, + "feedback": "The summary is grounded in the packet: 45870/45871 and 45872/45873 are clear duplicate pairs based on identical titles and matching body lengths, and 45869 is reasonably kept separate as a different dataset/sample topic. The soft-edge verdicts are conservative and do not overstate cross-pair duplication." + }, + "evaluator_used": true, + "retried": false, + "accepted_nontrivial_soft_edge": true, + "error_kind": null, + "error_message": null + }, + { + "label": "issue", + "component_index": 20, + "component_count": 27, + "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": "6e8649b0f572a8725686f3a63a6d5c377b9c66c969ec0e6d94a117598b997c14", + "component_budget": { + "node_count": 2, + "item_count": 2, + "soft_pair_count": 1, + "serialized_chars": 1640, + "estimated_input_tokens": 410, + "estimated_eval_tokens": 1076 + }, + "budget": { + "node_count": 2, + "item_count": 2, + "soft_pair_count": 1, + "serialized_chars": 1640, + "estimated_input_tokens": 410, + "estimated_eval_tokens": 1076 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": false, + "analyst_result": { + "summary": "Both issues concern `cache_implementation=\"static\"` and `linear_attention`, but they describe different failure modes: one is shape corruption during `early_initialization`, the other is a `KeyError` for `linear_attention` in Qwen3.5. They are related, but not clearly the same bug.", + "confidence": 0.76, + "canonical_issue_reason": "Issue 46439 is the better canonical issue because it describes a broader, more root-cause-like problem in static cache `early_initialization` affecting linear-attention layers, rather than a model-specific crash symptom.", + "canonical_pr_reason": null, + "best_issue_reason": "Issue 46439 is the best overall issue candidate because it is more general and appears to capture the underlying corruption mechanism, while 46441 reads more like a specific downstream manifestation or compatibility failure.", + "best_pr_reason": null, + "soft_edge_verdicts": [ + { + "left": "issue:46439", + "right": "issue:46441", + "accept": false, + "reason": "Related subsystem, but not the same concrete bug: 46439 reports wrong-shape/runtime corruption in `early_initialization`, whereas 46441 reports a `KeyError: 'linear_attention'` for Qwen3.5. The failure modes and likely fixes differ." + } + ] + }, + "evaluator_result": null, + "evaluator_used": false, + "retried": false, + "accepted_nontrivial_soft_edge": false, + "error_kind": null, + "error_message": null + }, + { + "label": "issue", + "component_index": 21, + "component_count": 27, + "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": "e1fe5644e1faed07e4494b6e7a61fc0b5888d5d1d1d290c4ec380234cd842c8d", + "component_budget": { + "node_count": 2, + "item_count": 2, + "soft_pair_count": 1, + "serialized_chars": 1707, + "estimated_input_tokens": 427, + "estimated_eval_tokens": 1110 + }, + "budget": { + "node_count": 2, + "item_count": 2, + "soft_pair_count": 1, + "serialized_chars": 1707, + "estimated_input_tokens": 427, + "estimated_eval_tokens": 1110 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": false, + "analyst_result": { + "summary": "The two issues are effectively identical reports of the same Wav2Vec2CTCTokenizer v5 special-token regression; they should be clustered together. The open issue is the better canonical tracker because it is newer and already has inbound references.", + "confidence": 0.99, + "canonical_issue_reason": "issue:46552 is the active/open report with inbound references, so it is the better canonical issue to keep as the cluster representative.", + "canonical_pr_reason": null, + "best_issue_reason": "issue:46552 is the best single issue because it is open, current, and already attracting references, making it the most useful tracker for the bug.", + "best_pr_reason": null, + "soft_edge_verdicts": [ + { + "left": "issue:46551", + "right": "issue:46552", + "accept": true, + "reason": "Same title and same underlying bug report; the issues are duplicate descriptions of the same tokenizer regression." + } + ] + }, + "evaluator_result": { + "accept": true, + "feedback": "The summary is grounded in the packet: the two issues have identical titles and body lengths, and the pair is marked with soft_similarity. The canonical choice of issue:46552 is also supported because it is open and has inbound references, while issue:46551 is closed. The soft-edge duplicate verdict is conservative and reasonable here." + }, + "evaluator_used": true, + "retried": false, + "accepted_nontrivial_soft_edge": true, + "error_kind": null, + "error_message": null + }, + { + "label": "issue", + "component_index": 22, + "component_count": 27, + "review_unit_index": 1, + "review_unit_count": 1, + "status": "reviewed", + "reason": null, + "source": "llm", + "cache_hit": false, + "model": "gpt-5.4-mini?service_tier=flex", + "cluster_id": "cluster-46657-2", + "nodes": [ + "issue:46657", + "issue:46692" + ], + "soft_pairs": [ + "issue:46657|issue:46692" + ], + "prepared_review_unit_hash": "00a3d2d76d9bed7b372649690cacfe2a97cc34e6ceb8a42a124f12a5df24d627", + "component_budget": { + "node_count": 2, + "item_count": 2, + "soft_pair_count": 1, + "serialized_chars": 1569, + "estimated_input_tokens": 393, + "estimated_eval_tokens": 1042 + }, + "budget": { + "node_count": 2, + "item_count": 2, + "soft_pair_count": 1, + "serialized_chars": 1569, + "estimated_input_tokens": 393, + "estimated_eval_tokens": 1042 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": false, + "analyst_result": { + "summary": "The two issues are both training-related, but they describe different failure modes: a tensor shape mismatch in object detection training vs. an eventual out-of-memory condition in RT-DETR v2. They should not be treated as duplicates.", + "confidence": 0.93, + "canonical_issue_reason": "Issue #46657 is the cleaner representative if one must be chosen: it states a concrete, specific bug symptom (tensor size mismatch) and is already closed. #46692 is a different problem with a different failure mode.", + "canonical_pr_reason": null, + "best_issue_reason": "#46657 is the best issue candidate because it is narrowly scoped and clearly describes a distinct code-path failure. #46692 appears unrelated enough that it should not be grouped under the same canonical issue.", + "best_pr_reason": null, + "soft_edge_verdicts": [ + { + "left": "issue:46657", + "right": "issue:46692", + "accept": false, + "reason": "Do not merge: the symptoms and likely root causes differ. A tensor dimension mismatch during object detection training is not the same underlying bug as memory growth leading to OOM after epochs in RT-DETR v2." + } + ] + }, + "evaluator_result": null, + "evaluator_used": false, + "retried": false, + "accepted_nontrivial_soft_edge": false, + "error_kind": null, + "error_message": null + }, + { + "label": "issue", + "component_index": 23, + "component_count": 27, + "review_unit_index": 1, + "review_unit_count": 1, + "status": "reviewed", + "reason": null, + "source": "llm", + "cache_hit": false, + "model": "gpt-5.4-mini?service_tier=flex", + "cluster_id": "cluster-46822-2", + "nodes": [ + "issue:46822", + "issue:46823" + ], + "soft_pairs": [ + "issue:46822|issue:46823" + ], + "prepared_review_unit_hash": "0bf2f8392e7576114b7f50c1a2997d46d6f4d8165d4f1612ca0a761aab22a7f6", + "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": "These are two different bug reports, not duplicates: one is about ZeRO-3 zero.Init failing to partition a composite model submodule and causing OOM, the other is about 4-bit + auto device mapping incorrectly sending a fitting model to CPU and raising ValueError.", + "confidence": 0.98, + "canonical_issue_reason": "No single canonical issue: the titles describe unrelated loading/dispatch bugs in different code paths.", + "canonical_pr_reason": null, + "best_issue_reason": "No clear best issue for the cluster because the items are not the same underlying problem.", + "best_pr_reason": null, + "soft_edge_verdicts": [ + { + "left": "issue:46822", + "right": "issue:46823", + "accept": false, + "reason": "Only superficial similarity around model loading/memory placement; one is a ZeRO-3 partitioning/OOM problem, the other is a 4-bit device_map/max_memory dispatch error. Different root causes and different fixes." + } + ] + }, + "evaluator_result": null, + "evaluator_used": false, + "retried": false, + "accepted_nontrivial_soft_edge": false, + "error_kind": null, + "error_message": null + }, + { + "label": "issue", + "component_index": 24, + "component_count": 27, + "review_unit_index": 1, + "review_unit_count": 1, + "status": "reviewed", + "reason": null, + "source": "llm", + "cache_hit": false, + "model": "gpt-5.4-mini?service_tier=flex", + "cluster_id": "cluster-46906-12", + "nodes": [ + "issue:46906", + "issue:46940", + "issue:46948", + "issue:46969", + "issue:46985", + "issue:46998", + "issue:47037", + "issue:47049", + "issue:47053", + "issue:47091", + "issue:47102", + "issue:47120" + ], + "soft_pairs": [ + "issue:46906|issue:46940", + "issue:46906|issue:46948", + "issue:46906|issue:46969", + "issue:46906|issue:46985", + "issue:46906|issue:47037", + "issue:46940|issue:46948", + "issue:46940|issue:46969", + "issue:46940|issue:46985", + "issue:46940|issue:47049", + "issue:46948|issue:46969", + "issue:46948|issue:46985", + "issue:46948|issue:46998", + "issue:46969|issue:46985", + "issue:46969|issue:47037", + "issue:46969|issue:47091", + "issue:46985|issue:47037", + "issue:46985|issue:47049", + "issue:46998|issue:47053", + "issue:46998|issue:47037", + "issue:46998|issue:47049", + "issue:46985|issue:46998", + "issue:46998|issue:47091", + "issue:47037|issue:47049", + "issue:47037|issue:47091", + "issue:47037|issue:47053", + "issue:47049|issue:47053", + "issue:47049|issue:47091", + "issue:47053|issue:47091", + "issue:47053|issue:47102", + "issue:47091|issue:47102", + "issue:46985|issue:47091", + "issue:47037|issue:47102", + "issue:46985|issue:47102", + "issue:47049|issue:47102", + "issue:47102|issue:47120", + "issue:47053|issue:47120", + "issue:47049|issue:47120", + "issue:46998|issue:47120", + "issue:47091|issue:47120" + ], + "prepared_review_unit_hash": "6ea204b2d3853f0db4fdcda3031475bae1f8f419968a354a3b92735806966a94", + "component_budget": { + "node_count": 12, + "item_count": 12, + "soft_pair_count": 39, + "serialized_chars": 20779, + "estimated_input_tokens": 5195, + "estimated_eval_tokens": 10646 + }, + "budget": { + "node_count": 12, + "item_count": 12, + "soft_pair_count": 39, + "serialized_chars": 20779, + "estimated_input_tokens": 5195, + "estimated_eval_tokens": 10646 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": false, + "analyst_result": { + "summary": "These are date-stamped Serde/Serge integration triage reports with highly similar template text, but they do not look like the same underlying bug report. They read more like separate daily tracking issues, so the soft duplicate edges should be rejected.", + "confidence": 0.93, + "canonical_issue_reason": "Issue 47120 is the best anchor if one must be chosen: it is the latest open triage thread and has the most recent inbound references, but the set is not actually duplicate-heavy.", + "canonical_pr_reason": null, + "best_issue_reason": "Issue 47120 is the most useful single issue for follow-up because it is current, open, and appears to supersede the earlier daily triage entries.", + "best_pr_reason": null, + "soft_edge_verdicts": [ + { + "left": "issue:46906", + "right": "issue:46940", + "accept": false, + "reason": "Same triage template and adjacent dates, but this looks like a separate daily report rather than one bug duplicated." + }, + { + "left": "issue:46906", + "right": "issue:46948", + "accept": false, + "reason": "Highly similar wording, but the date-stamped triage entries suggest distinct daily logs, not the same issue." + }, + { + "left": "issue:46906", + "right": "issue:46969", + "accept": false, + "reason": "Shared template only; no evidence these describe the same concrete failure." + }, + { + "left": "issue:46906", + "right": "issue:46985", + "accept": false, + "reason": "These are separate date-specific triage issues, not a single duplicate bug." + }, + { + "left": "issue:46906", + "right": "issue:47037", + "accept": false, + "reason": "Similarity comes from the repeated title format, but the issues are different daily reports." + }, + { + "left": "issue:46940", + "right": "issue:46948", + "accept": false, + "reason": "Adjacent triage threads with the same template, but likely distinct day-by-day reports." + }, + { + "left": "issue:46940", + "right": "issue:46969", + "accept": false, + "reason": "No concrete shared failure is shown beyond the recurring triage wording." + }, + { + "left": "issue:46940", + "right": "issue:46985", + "accept": false, + "reason": "Similar title pattern only; not enough to treat as one duplicated issue." + }, + { + "left": "issue:46940", + "right": "issue:47049", + "accept": false, + "reason": "Different dated triage entries, so they should not be merged as duplicates." + }, + { + "left": "issue:46948", + "right": "issue:46969", + "accept": false, + "reason": "Same recurring triage format, but these appear to be separate daily log issues." + }, + { + "left": "issue:46948", + "right": "issue:46985", + "accept": false, + "reason": "The overlap is in template text, not a shared underlying bug report." + }, + { + "left": "issue:46948", + "right": "issue:46998", + "accept": false, + "reason": "Date-stamped triage threads are not strong duplicate evidence by themselves." + }, + { + "left": "issue:46969", + "right": "issue:46985", + "accept": false, + "reason": "They look like consecutive tracking reports, not one duplicate issue." + }, + { + "left": "issue:46969", + "right": "issue:47037", + "accept": false, + "reason": "Same subsystem and template, but no sign of the same concrete failure." + }, + { + "left": "issue:46969", + "right": "issue:47091", + "accept": false, + "reason": "These are separate daily triage entries rather than duplicates of one bug." + }, + { + "left": "issue:46985", + "right": "issue:46998", + "accept": false, + "reason": "Recurring title pattern only; no concrete duplicate bug evidence." + }, + { + "left": "issue:46985", + "right": "issue:47037", + "accept": false, + "reason": "Looks like another daily triage note, not the same issue instance." + }, + { + "left": "issue:46985", + "right": "issue:47049", + "accept": false, + "reason": "Date-based tracking issues should stay separate unless they share the same failure details." + }, + { + "left": "issue:46998", + "right": "issue:47053", + "accept": false, + "reason": "The common wording is from the triage template, not a shared defect." + }, + { + "left": "issue:46998", + "right": "issue:47037", + "accept": false, + "reason": "Separate daily triage reports; not enough to merge as duplicates." + }, + { + "left": "issue:46998", + "right": "issue:47049", + "accept": false, + "reason": "These appear to document different days of integration failures, not one bug." + }, + { + "left": "issue:46985", + "right": "issue:46998", + "accept": false, + "reason": "Duplicate-looking title alone is insufficient; these are distinct date-stamped triage issues." + }, + { + "left": "issue:46998", + "right": "issue:47091", + "accept": false, + "reason": "Same template, but no evidence of the same underlying problem." + }, + { + "left": "issue:47037", + "right": "issue:47049", + "accept": false, + "reason": "Consecutive triage reports with similar wording are not duplicates by default." + }, + { + "left": "issue:47037", + "right": "issue:47091", + "accept": false, + "reason": "These are separate daily triage threads, not one merged bug report." + }, + { + "left": "issue:47037", + "right": "issue:47053", + "accept": false, + "reason": "Shared phrasing only; no concrete identical failure is shown." + }, + { + "left": "issue:47049", + "right": "issue:47053", + "accept": false, + "reason": "Different dates and no specific shared bug details, so reject duplicate merge." + }, + { + "left": "issue:47049", + "right": "issue:47091", + "accept": false, + "reason": "Looks like another daily tracking entry rather than the same issue." + }, + { + "left": "issue:47053", + "right": "issue:47091", + "accept": false, + "reason": "Template similarity is expected here and does not indicate a duplicate." + }, + { + "left": "issue:47053", + "right": "issue:47102", + "accept": false, + "reason": "Separate date-stamped triage issues; not the same underlying change or bug." + }, + { + "left": "issue:47091", + "right": "issue:47102", + "accept": false, + "reason": "No shared concrete failure beyond the repeated triage title." + }, + { + "left": "issue:46985", + "right": "issue:47091", + "accept": false, + "reason": "Recurring triage format only; these should remain distinct issues." + }, + { + "left": "issue:47037", + "right": "issue:47102", + "accept": false, + "reason": "They read as independent daily reports, not duplicates." + }, + { + "left": "issue:46985", + "right": "issue:47102", + "accept": false, + "reason": "Same subsystem and template, but no same-bug evidence." + }, + { + "left": "issue:47049", + "right": "issue:47102", + "accept": false, + "reason": "Different dated tracking issues; do not collapse them as duplicates." + }, + { + "left": "issue:47102", + "right": "issue:47120", + "accept": false, + "reason": "Adjacent daily triage threads, but still separate reports rather than one duplicate issue." + }, + { + "left": "issue:47053", + "right": "issue:47120", + "accept": false, + "reason": "The overlap is the shared triage template, not a shared concrete bug." + }, + { + "left": "issue:47049", + "right": "issue:47120", + "accept": false, + "reason": "These look like sequential daily integration reports, not duplicate issues." + }, + { + "left": "issue:46998", + "right": "issue:47120", + "accept": false, + "reason": "No evidence of the same underlying failure beyond the repeated title pattern." + }, + { + "left": "issue:47091", + "right": "issue:47120", + "accept": false, + "reason": "Date-stamped triage entries should not be merged without identical bug details." + } + ] + }, + "evaluator_result": null, + "evaluator_used": false, + "retried": false, + "accepted_nontrivial_soft_edge": false, + "error_kind": null, + "error_message": null + }, + { + "label": "issue", + "component_index": 25, + "component_count": 27, + "review_unit_index": 1, + "review_unit_count": 1, + "status": "reviewed", + "reason": null, + "source": "llm", + "cache_hit": false, + "model": "gpt-5.4-mini?service_tier=flex", + "cluster_id": "cluster-46983-2", + "nodes": [ + "issue:46983", + "issue:46984" + ], + "soft_pairs": [ + "issue:46983|issue:46984" + ], + "prepared_review_unit_hash": "41f98382c30e7a786d511356165bd3ffea9d76d23a67bab147758d17f6031a11", + "component_budget": { + "node_count": 2, + "item_count": 2, + "soft_pair_count": 1, + "serialized_chars": 1492, + "estimated_input_tokens": 373, + "estimated_eval_tokens": 1002 + }, + "budget": { + "node_count": 2, + "item_count": 2, + "soft_pair_count": 1, + "serialized_chars": 1492, + "estimated_input_tokens": 373, + "estimated_eval_tokens": 1002 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": false, + "analyst_result": { + "summary": "The two closed issues appear to be duplicates: same title, very similar body length, and high textual similarity. Issue #46983 is the better canonical representative because it was opened first.", + "confidence": 0.98, + "canonical_issue_reason": "#46983 and #46984 have the same title and closely matching descriptions; #46983 was created first, so it is the cleaner canonical issue.", + "canonical_pr_reason": null, + "best_issue_reason": "#46983 is the best issue to keep as the canonical record because it is the earliest instance of the duplicate report and there is no evidence that #46984 adds distinct scope.", + "best_pr_reason": null, + "soft_edge_verdicts": [ + { + "left": "issue:46983", + "right": "issue:46984", + "accept": true, + "reason": "Same issue title and highly similar content indicate the same underlying request; this is a clear duplicate pair." + } + ] + }, + "evaluator_result": { + "accept": true, + "feedback": "The conclusion is grounded in the packet: both issues have the same title, the soft-similarity score is high (Jaccard 0.837), and #46983 was created slightly earlier, which supports it as the canonical record. The duplicate verdict is conservative enough given the available evidence." + }, + "evaluator_used": true, + "retried": false, + "accepted_nontrivial_soft_edge": true, + "error_kind": null, + "error_message": null + }, + { + "label": "issue", + "component_index": 26, + "component_count": 27, + "review_unit_index": 1, + "review_unit_count": 1, + "status": "reviewed", + "reason": null, + "source": "llm", + "cache_hit": false, + "model": "gpt-5.4-mini?service_tier=flex", + "cluster_id": "cluster-47007-2", + "nodes": [ + "issue:47007", + "issue:47015" + ], + "soft_pairs": [ + "issue:47007|issue:47015" + ], + "prepared_review_unit_hash": "b1876cf596406f166329b82bd403b412872158a5fd60328c84575d8ee76b5622", + "component_budget": { + "node_count": 2, + "item_count": 2, + "soft_pair_count": 1, + "serialized_chars": 1625, + "estimated_input_tokens": 407, + "estimated_eval_tokens": 1070 + }, + "budget": { + "node_count": 2, + "item_count": 2, + "soft_pair_count": 1, + "serialized_chars": 1625, + "estimated_input_tokens": 407, + "estimated_eval_tokens": 1070 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": false, + "analyst_result": { + "summary": "The two issues are topically related around AI-driven vulnerability discovery, but they do not look like the same underlying report. They read more like separate writeups/proposals with different emphasis and no explicit linkage.", + "confidence": 0.87, + "canonical_issue_reason": "No clear canonical duplicate: the titles point to different angles (lessons from LLM usage vs. a novel transformer-based cybersecurity system) and there is no shared evidence tying them to one issue.", + "canonical_pr_reason": null, + "best_issue_reason": "If one issue must represent the cluster, 47007 is slightly more concrete and scoped, but it still is not a true duplicate of 47015.", + "best_pr_reason": null, + "soft_edge_verdicts": [ + { + "left": "issue:47007", + "right": "issue:47015", + "accept": false, + "reason": "Same broad theme, but not the same underlying bug or change; the titles suggest different content and intent, with no explicit issue links or shared artifacts." + } + ] + }, + "evaluator_result": null, + "evaluator_used": false, + "retried": false, + "accepted_nontrivial_soft_edge": false, + "error_kind": null, + "error_message": null + }, + { + "label": "issue", + "component_index": 27, + "component_count": 27, + "review_unit_index": 1, + "review_unit_count": 1, + "status": "reviewed", + "reason": null, + "source": "llm", + "cache_hit": false, + "model": "gpt-5.4-mini?service_tier=flex", + "cluster_id": "cluster-47138-2", + "nodes": [ + "issue:47138", + "issue:47139" + ], + "soft_pairs": [ + "issue:47138|issue:47139" + ], + "prepared_review_unit_hash": "71cb8b08b9036beee9481427d0067b51238c11f29ce42550103b65a64494125a", + "component_budget": { + "node_count": 2, + "item_count": 2, + "soft_pair_count": 1, + "serialized_chars": 1591, + "estimated_input_tokens": 398, + "estimated_eval_tokens": 1052 + }, + "budget": { + "node_count": 2, + "item_count": 2, + "soft_pair_count": 1, + "serialized_chars": 1591, + "estimated_input_tokens": 398, + "estimated_eval_tokens": 1052 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": false, + "analyst_result": { + "summary": "The two issues are effectively duplicates: same title, same apparent body length, and no distinguishing metadata beyond inbound references. Issue 47139 is the better canonical/best representative because it already has inbound references and likely serves as the more connected thread.", + "confidence": 0.99, + "canonical_issue_reason": "Issue 47139 is the better canonical issue because it has inbound references, making it the more established discussion thread for the same reported bug.", + "canonical_pr_reason": null, + "best_issue_reason": "Issue 47139 is the best issue candidate for the same reason: it is the more referenced duplicate and likely the preferred tracking thread.", + "best_pr_reason": null, + "soft_edge_verdicts": [ + { + "left": "issue:47138", + "right": "issue:47139", + "accept": true, + "reason": "Same title and no visible differences in scope; they appear to report the same underlying bug, so this is a clear duplicate pair." + } + ] + }, + "evaluator_result": { + "accept": true, + "feedback": "The assessment is grounded in the packet: the two issues have identical titles and body lengths, and issue 47139 has the stronger inbound-reference signal, so choosing it as the canonical/best representative is conservative and reasonable." + }, + "evaluator_used": true, + "retried": false, + "accepted_nontrivial_soft_edge": true, + "error_kind": null, + "error_message": null + }, + { + "label": "pull_request", + "component_index": 1, + "component_count": 33, + "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-40546-42", + "nodes": [ + "pull_request:43451", + "pull_request:44517", + "pull_request:46196", + "pull_request:46266", + "pull_request:46658", + "pull_request:46696", + "pull_request:46707", + "pull_request:46836", + "pull_request:46895", + "pull_request:46903", + "pull_request:46912", + "pull_request:46954", + "pull_request:46957", + "pull_request:46958", + "pull_request:46974", + "pull_request:46995", + "pull_request:47034", + "pull_request:47035", + "pull_request:47069", + "pull_request:47072", + "pull_request:47077", + "pull_request:47078", + "pull_request:47082", + "pull_request:47089", + "pull_request:47090", + "pull_request:47093", + "pull_request:47096", + "pull_request:47100", + "pull_request:47108", + "pull_request:47113", + "pull_request:47114", + "pull_request:47134", + "pull_request:47135", + "pull_request:47142", + "pull_request:47143", + "pull_request:47156", + "pull_request:47157", + "pull_request:47161", + "pull_request:47163", + "pull_request:47166" + ], + "soft_pairs": [ + "pull_request:46696|pull_request:47134", + "pull_request:46903|pull_request:47090", + "pull_request:46903|pull_request:46958", + "pull_request:47156|pull_request:47157", + "pull_request:46954|pull_request:46958", + "pull_request:46958|pull_request:46974", + "pull_request:47072|pull_request:47089", + "pull_request:47077|pull_request:47157", + "pull_request:47077|pull_request:47078", + "pull_request:46266|pull_request:46957", + "pull_request:47072|pull_request:47134", + "pull_request:47114|pull_request:47135", + "pull_request:44517|pull_request:47082", + "pull_request:47034|pull_request:47035", + "pull_request:46995|pull_request:47163", + "pull_request:47069|pull_request:47089", + "pull_request:46196|pull_request:46995", + "pull_request:47135|pull_request:47163", + "pull_request:47093|pull_request:47114", + "pull_request:46707|pull_request:47108", + "pull_request:46707|pull_request:47114", + "pull_request:47089|pull_request:47113", + "pull_request:47093|pull_request:47161", + "pull_request:46266|pull_request:46707", + "pull_request:46995|pull_request:47143", + "pull_request:47082|pull_request:47108", + "pull_request:47108|pull_request:47114", + "pull_request:47078|pull_request:47156", + "pull_request:47077|pull_request:47114", + "pull_request:47093|pull_request:47143", + "pull_request:47093|pull_request:47163", + "pull_request:47134|pull_request:47135", + "pull_request:46995|pull_request:47161", + "pull_request:46707|pull_request:46957", + "pull_request:46658|pull_request:47114", + "pull_request:46707|pull_request:47082", + "pull_request:46266|pull_request:47114", + "pull_request:47114|pull_request:47156", + "pull_request:46912|pull_request:47082", + "pull_request:46912|pull_request:47078", + "pull_request:47134|pull_request:47161", + "pull_request:43451|pull_request:47082", + "pull_request:46895|pull_request:47134", + "pull_request:47113|pull_request:47161", + "pull_request:47100|pull_request:47134", + "pull_request:47035|pull_request:47089", + "pull_request:43451|pull_request:46995", + "pull_request:47100|pull_request:47142", + "pull_request:47156|pull_request:47166", + "pull_request:47035|pull_request:47161", + "pull_request:46266|pull_request:47034", + "pull_request:47034|pull_request:47108", + "pull_request:47035|pull_request:47113", + "pull_request:47035|pull_request:47069", + "pull_request:47100|pull_request:47113", + "pull_request:46836|pull_request:47135", + "pull_request:46957|pull_request:47108", + "pull_request:47100|pull_request:47161", + "pull_request:47069|pull_request:47113", + "pull_request:46196|pull_request:47161", + "pull_request:47069|pull_request:47082", + "pull_request:47089|pull_request:47134", + "pull_request:43451|pull_request:46266", + "pull_request:43451|pull_request:47078", + "pull_request:47078|pull_request:47096", + "pull_request:47135|pull_request:47161", + "pull_request:43451|pull_request:44517", + "pull_request:47100|pull_request:47143", + "pull_request:46912|pull_request:47113", + "pull_request:46912|pull_request:47142", + "pull_request:44517|pull_request:47143", + "pull_request:44517|pull_request:47163" + ], + "prepared_review_unit_hash": "9adb1c334be1f69b501c576278b8e1d63e4954df1016345375d33531a285adb6", + "component_budget": { + "node_count": 42, + "item_count": 42, + "soft_pair_count": 126, + "serialized_chars": 150672, + "estimated_input_tokens": 37668, + "estimated_eval_tokens": 75592 + }, + "budget": { + "node_count": 40, + "item_count": 40, + "soft_pair_count": 72, + "serialized_chars": 115711, + "estimated_input_tokens": 28928, + "estimated_eval_tokens": 58112 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": true, + "analyst_result": { + "summary": "This cluster is mostly unrelated PRs spanning new model additions, docs, exporter/runtime tweaks, and small bug fixes. Only one soft edge looks like a true duplicate: the two CohereASR training-loss double-shift fixes.", + "confidence": 0.92, + "canonical_issue_reason": null, + "canonical_pr_reason": "PR 47090 is the clearest representative of the only true duplicate pair: it fixes the same CohereASR loss double-shift bug as 46958, on the same files and issue target, with the most explicit rationale.", + "best_issue_reason": null, + "best_pr_reason": "PR 47090 is the best overall PR candidate because it is the most explicit and complete version of the CohereASR double-shift fix among the near-duplicates.", + "soft_edge_verdicts": [ + { + "left": "pull_request:46696", + "right": "pull_request:47134", + "accept": false, + "reason": "Same flash-attention area and issue target, but different concrete regressions: tracing/int sync vs unpadding max-seqlen performance." + }, + { + "left": "pull_request:46903", + "right": "pull_request:47090", + "accept": false, + "reason": "Shared training-loss pattern, but different model families and fixes: PPFormulaNet plus CohereASR versus CohereASR only." + }, + { + "left": "pull_request:46903", + "right": "pull_request:46958", + "accept": false, + "reason": "Both touch CohereASR code, but 46903 is a broader multi-model change and not the same concrete bug/fix as 46958." + }, + { + "left": "pull_request:47156", + "right": "pull_request:47157", + "accept": false, + "reason": "Both are docs PRs, but they translate different pages in different locales and are unrelated changes." + }, + { + "left": "pull_request:46954", + "right": "pull_request:46958", + "accept": false, + "reason": "Pytest fp32-keep behavior change versus CohereASR loss-shift fix; no shared underlying bug." + }, + { + "left": "pull_request:46958", + "right": "pull_request:46974", + "accept": false, + "reason": "Unrelated fixes in different models: CohereASR training loss versus a MusicGen generate typo." + }, + { + "left": "pull_request:47072", + "right": "pull_request:47089", + "accept": false, + "reason": "Chat-template JSON schema parsing fix and AqlmConfig error-message wording are unrelated." + }, + { + "left": "pull_request:47077", + "right": "pull_request:47157", + "accept": false, + "reason": "Different Japanese docs updates in different pages; not the same underlying change." + }, + { + "left": "pull_request:47077", + "right": "pull_request:47078", + "accept": false, + "reason": "Both are Japanese docs PRs, but one is a translation effort and the other is wording cleanup." + }, + { + "left": "pull_request:46266", + "right": "pull_request:46957", + "accept": false, + "reason": "DINO conversion mapping work and FSDP2 migration are unrelated subsystems and changes." + }, + { + "left": "pull_request:47072", + "right": "pull_request:47134", + "accept": false, + "reason": "Chat schema enum handling and flash-attention performance work are unrelated." + }, + { + "left": "pull_request:47114", + "right": "pull_request:47089", + "accept": false, + "reason": "Docs on custom-code conversion mapping and AqlmConfig error messages do not fix the same bug or feature." + }, + { + "left": "pull_request:46196", + "right": "pull_request:46995", + "accept": false, + "reason": "Both broaden attention/runtime support, but they target different backends and code paths." + }, + { + "left": "pull_request:47135", + "right": "pull_request:47163", + "accept": false, + "reason": "Docs quote-rendering cleanup and XPU paged-attention tests are unrelated." + }, + { + "left": "pull_request:47093", + "right": "pull_request:47114", + "accept": false, + "reason": "OpenPangu-V2 model addition versus documentation about custom-code conversion mapping; unrelated." + }, + { + "left": "pull_request:46707", + "right": "pull_request:47108", + "accept": false, + "reason": "Native FSDP2 migration and a doc-builder comment tweak are completely unrelated." + }, + { + "left": "pull_request:46707", + "right": "pull_request:47114", + "accept": false, + "reason": "FSDP2 migration versus docs on conversion mapping lookup rules; not the same change." + }, + { + "left": "pull_request:47089", + "right": "pull_request:47113", + "accept": false, + "reason": "AqlmConfig message wording and audio feature/proc requirements fixes are unrelated." + }, + { + "left": "pull_request:47093", + "right": "pull_request:47161", + "accept": false, + "reason": "Model addition versus exporter min-version enforcement; no shared underlying change." + }, + { + "left": "pull_request:46266", + "right": "pull_request:46707", + "accept": false, + "reason": "DINO modernization and FSDP2 integration are different features with different targets." + }, + { + "left": "pull_request:46995", + "right": "pull_request:47143", + "accept": false, + "reason": "Flash-linear-attention backend support and NeuCodec model support are unrelated." + }, + { + "left": "pull_request:47135", + "right": "pull_request:47161", + "accept": false, + "reason": "Docs quote rendering and exporter version gating are unrelated." + }, + { + "left": "pull_request:47035", + "right": "pull_request:47161", + "accept": false, + "reason": "Neuron compile backend defaulting and exporter version requirements are unrelated." + }, + { + "left": "pull_request:46196", + "right": "pull_request:47161", + "accept": false, + "reason": "Attention backend dispatch support and exporter version enforcement do not fix the same thing." + }, + { + "left": "pull_request:47034", + "right": "pull_request:47035", + "accept": false, + "reason": "The draft Neuron experiment script and torch-compile backend default are adjacent but not the same concrete change." + }, + { + "left": "pull_request:47069", + "right": "pull_request:47089", + "accept": false, + "reason": "Num_labels None typing/behavior and AqlmConfig error-message wording are unrelated." + }, + { + "left": "pull_request:47069", + "right": "pull_request:47082", + "accept": false, + "reason": "Num_labels property typing fix and RoMa model addition are unrelated." + }, + { + "left": "pull_request:47077", + "right": "pull_request:47114", + "accept": false, + "reason": "Japanese docs translation and RoMa model registration are unrelated." + }, + { + "left": "pull_request:47078", + "right": "pull_request:47156", + "accept": false, + "reason": "Japanese wording cleanup and Korean Optimum translation/typo fix are unrelated docs updates." + }, + { + "left": "pull_request:47077", + "right": "pull_request:47156", + "accept": false, + "reason": "Different locales and different pages; not the same underlying documentation change." + }, + { + "left": "pull_request:47089", + "right": "pull_request:47134", + "accept": false, + "reason": "AqlmConfig validation messages and flash-attention performance regression are unrelated." + }, + { + "left": "pull_request:46912", + "right": "pull_request:47082", + "accept": false, + "reason": "RoPE test expansion and RoMa model addition are unrelated." + }, + { + "left": "pull_request:46912", + "right": "pull_request:47078", + "accept": false, + "reason": "RoPE tests and Japanese docs wording cleanup are unrelated." + }, + { + "left": "pull_request:47134", + "right": "pull_request:47161", + "accept": false, + "reason": "Flash-attention unpadding performance fix and exporter version requirement changes are unrelated." + }, + { + "left": "pull_request:43451", + "right": "pull_request:47082", + "accept": false, + "reason": "Molmo2 model addition and RoMa keypoint matcher addition are unrelated model-introduction PRs." + }, + { + "left": "pull_request:46895", + "right": "pull_request:47134", + "accept": false, + "reason": "CohereASR loss-shift fix and flash-attention performance regression are unrelated bugs." + }, + { + "left": "pull_request:47113", + "right": "pull_request:47161", + "accept": false, + "reason": "Audio requirements cleanup and exporter minimum-version enforcement are unrelated." + }, + { + "left": "pull_request:47093", + "right": "pull_request:47143", + "accept": false, + "reason": "OpenPangu-V2 model addition and NeuCodec support are different model-introduction PRs." + }, + { + "left": "pull_request:47093", + "right": "pull_request:47163", + "accept": false, + "reason": "OpenPangu-V2 model addition and XPU paged-attention tests are unrelated." + }, + { + "left": "pull_request:47134", + "right": "pull_request:47135", + "accept": false, + "reason": "Flash-attention performance fix and docs quote rendering cleanup are unrelated." + }, + { + "left": "pull_request:43451", + "right": "pull_request:46995", + "accept": false, + "reason": "Molmo2 model addition and flash-linear-attention backend support are unrelated." + }, + { + "left": "pull_request:47100", + "right": "pull_request:47143", + "accept": false, + "reason": "Hub-kernels API cleanup and NeuCodec support are unrelated." + }, + { + "left": "pull_request:46912", + "right": "pull_request:47113", + "accept": false, + "reason": "RoPE test coverage and audio requirements cleanup are unrelated." + }, + { + "left": "pull_request:46912", + "right": "pull_request:47142", + "accept": false, + "reason": "RoPE tests and pipelines generator cleanup are unrelated." + }, + { + "left": "pull_request:47156", + "right": "pull_request:47166", + "accept": false, + "reason": "Korean docs translation and quantization test updates are unrelated." + }, + { + "left": "pull_request:46266", + "right": "pull_request:47034", + "accept": false, + "reason": "DINO conversion modernization and a Neuron experiment script are unrelated." + }, + { + "left": "pull_request:47034", + "right": "pull_request:47108", + "accept": false, + "reason": "Neuron experiment code and doc-builder comment tweak are unrelated." + }, + { + "left": "pull_request:47035", + "right": "pull_request:47113", + "accept": false, + "reason": "Torch-compile backend defaulting and audio requirements cleanup are unrelated." + }, + { + "left": "pull_request:47035", + "right": "pull_request:47069", + "accept": false, + "reason": "Neuron compile backend behavior and num_labels typing are unrelated." + }, + { + "left": "pull_request:47035", + "right": "pull_request:47082", + "accept": false, + "reason": "Neuron compile backend defaulting and RoMa model addition are unrelated." + }, + { + "left": "pull_request:43451", + "right": "pull_request:46266", + "accept": false, + "reason": "Molmo2 model addition and DINO modern-standards work are unrelated." + }, + { + "left": "pull_request:47100", + "right": "pull_request:47142", + "accept": false, + "reason": "Hub-kernels error-path cleanup and pipeline generator handling are unrelated." + }, + { + "left": "pull_request:47069", + "right": "pull_request:47161", + "accept": false, + "reason": "Num_labels None behavior and exporter version requirements are unrelated." + }, + { + "left": "pull_request:47082", + "right": "pull_request:47108", + "accept": false, + "reason": "RoMa model addition and doc-builder comment tweak are unrelated." + }, + { + "left": "pull_request:44517", + "right": "pull_request:47143", + "accept": false, + "reason": "Qwen3-TTS model support and NeuCodec support are unrelated model additions." + }, + { + "left": "pull_request:44517", + "right": "pull_request:47082", + "accept": false, + "reason": "Qwen3-TTS model support and RoMa keypoint matcher addition are unrelated." + }, + { + "left": "pull_request:46957", + "right": "pull_request:47108", + "accept": false, + "reason": "Finegrained FP8 generation changes and a doc-builder comment tweak are unrelated." + }, + { + "left": "pull_request:46995", + "right": "pull_request:47143", + "accept": false, + "reason": "Flash-linear-attention support and NeuCodec support are unrelated." + }, + { + "left": "pull_request:47093", + "right": "pull_request:47114", + "accept": false, + "reason": "OpenPangu-V2 model support and docs about conversion lookup rules are unrelated." + }, + { + "left": "pull_request:47090", + "right": "pull_request:46958", + "accept": true, + "reason": "Same CohereASR training-loss double-shift bug, same files, same issue target, and the fixes are direct variants of 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": 1, + "component_count": 33, + "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-40546-42", + "nodes": [ + "pull_request:40546", + "pull_request:43451", + "pull_request:44517", + "pull_request:46196", + "pull_request:46658", + "pull_request:46707", + "pull_request:46836", + "pull_request:46895", + "pull_request:46912", + "pull_request:46957", + "pull_request:46958", + "pull_request:47003", + "pull_request:47034", + "pull_request:47035", + "pull_request:47069", + "pull_request:47077", + "pull_request:47090", + "pull_request:47096", + "pull_request:47142", + "pull_request:47143", + "pull_request:47156", + "pull_request:47161", + "pull_request:47163", + "pull_request:47166" + ], + "soft_pairs": [ + "pull_request:44517|pull_request:46836", + "pull_request:46957|pull_request:47161", + "pull_request:43451|pull_request:46957", + "pull_request:46957|pull_request:47096", + "pull_request:46658|pull_request:46895", + "pull_request:46912|pull_request:47077", + "pull_request:43451|pull_request:47142", + "pull_request:47096|pull_request:47142", + "pull_request:43451|pull_request:47069", + "pull_request:43451|pull_request:46895", + "pull_request:47003|pull_request:47142", + "pull_request:46895|pull_request:47090", + "pull_request:46958|pull_request:47090", + "pull_request:46895|pull_request:46958", + "pull_request:47077|pull_request:47156", + "pull_request:47163|pull_request:47166", + "pull_request:46707|pull_request:47034", + "pull_request:46658|pull_request:46836", + "pull_request:43451|pull_request:46836", + "pull_request:46658|pull_request:47143", + "pull_request:46836|pull_request:47143", + "pull_request:46836|pull_request:47163", + "pull_request:46196|pull_request:47143", + "pull_request:47143|pull_request:47163", + "pull_request:46196|pull_request:47163", + "pull_request:46836|pull_request:47161", + "pull_request:43451|pull_request:47077", + "pull_request:43451|pull_request:47161", + "pull_request:43451|pull_request:46196", + "pull_request:40546|pull_request:43451", + "pull_request:47143|pull_request:47161", + "pull_request:47161|pull_request:47163", + "pull_request:43451|pull_request:46658", + "pull_request:46658|pull_request:47096", + "pull_request:43451|pull_request:47143", + "pull_request:43451|pull_request:47163", + "pull_request:46912|pull_request:47163", + "pull_request:43451|pull_request:47096", + "pull_request:43451|pull_request:47003", + "pull_request:47003|pull_request:47096", + "pull_request:43451|pull_request:47034", + "pull_request:47034|pull_request:47096", + "pull_request:43451|pull_request:47166", + "pull_request:47096|pull_request:47166", + "pull_request:43451|pull_request:47035", + "pull_request:47003|pull_request:47034", + "pull_request:47003|pull_request:47166", + "pull_request:47096|pull_request:47143", + "pull_request:47003|pull_request:47143", + "pull_request:47096|pull_request:47163", + "pull_request:47003|pull_request:47163", + "pull_request:47143|pull_request:47166", + "pull_request:46836|pull_request:47096", + "pull_request:46836|pull_request:47003" + ], + "prepared_review_unit_hash": "7fc23bb3cc9747df1bc11de291a2355edbfaeeaf90e2dac86838377435f4fcf9", + "component_budget": { + "node_count": 42, + "item_count": 42, + "soft_pair_count": 126, + "serialized_chars": 150672, + "estimated_input_tokens": 37668, + "estimated_eval_tokens": 75592 + }, + "budget": { + "node_count": 24, + "item_count": 24, + "soft_pair_count": 54, + "serialized_chars": 77931, + "estimated_input_tokens": 19483, + "estimated_eval_tokens": 39222 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": true, + "analyst_result": { + "summary": "This set is mostly unrelated PRs spanning new model additions, docs, exporters, tests, and infra. The only clear duplicate cluster is the CohereASR training-loss double-shift fixes; the rest share broad subsystems at most, not the same concrete change.", + "confidence": 0.95, + "canonical_issue_reason": null, + "canonical_pr_reason": "PR #47090 is the best canonical representative of the CohereASR duplicate cluster: it targets the same training-loss double-shift bug, explains the fix clearly, and uses the most complete loss-function call path.", + "best_issue_reason": null, + "best_pr_reason": "PR #47090 is the strongest overall representative among the duplicate candidates because it fixes the same concrete bug with the clearest rationale and preserves the intended loss-reduction behavior.", + "soft_edge_verdicts": [ + { + "left": "pull_request:44517", + "right": "pull_request:46836", + "accept": false, + "reason": "Different model additions (VibeVoice vs Unlimited OCR); same docs/auto-registration patterns, but not the same change." + }, + { + "left": "pull_request:46957", + "right": "pull_request:47161", + "accept": false, + "reason": "Both touch exporter-related code, but one is a finegrained FP8 kernel change and the other is exporter version gating; not the same bug or feature." + }, + { + "left": "pull_request:43451", + "right": "pull_request:46957", + "accept": false, + "reason": "Molmo2 model onboarding vs FP8/generation/export kernel work; no shared underlying change." + }, + { + "left": "pull_request:46957", + "right": "pull_request:47096", + "accept": false, + "reason": "OpenVINO exporter work and Kimi exportability are separate export paths with different code changes." + }, + { + "left": "pull_request:46658", + "right": "pull_request:46895", + "accept": false, + "reason": "Step-3.7 Flash model support and CohereASR loss fixing are unrelated." + }, + { + "left": "pull_request:46912", + "right": "pull_request:47077", + "accept": false, + "reason": "RoPE test coverage work vs Japanese DistilBERT documentation translation; unrelated." + }, + { + "left": "pull_request:43451", + "right": "pull_request:47142", + "accept": false, + "reason": "Molmo2 model addition vs pipeline generator cleanup; no duplicate change." + }, + { + "left": "pull_request:47096", + "right": "pull_request:47142", + "accept": false, + "reason": "Kimi export support vs pipeline generator handling; different code paths." + }, + { + "left": "pull_request:43451", + "right": "pull_request:47069", + "accept": false, + "reason": "Model onboarding vs a small configuration property typing fix; unrelated." + }, + { + "left": "pull_request:43451", + "right": "pull_request:46895", + "accept": false, + "reason": "Molmo2 addition vs CohereASR training-loss fix; unrelated." + }, + { + "left": "pull_request:47003", + "right": "pull_request:47142", + "accept": false, + "reason": "OpenVINO exporter addition and pipeline generator cleanup are distinct changes." + }, + { + "left": "pull_request:46895", + "right": "pull_request:47090", + "accept": true, + "reason": "Same CohereASR double-shift bug in the same files; both fixes adjust the training loss path for the same concrete issue." + }, + { + "left": "pull_request:46958", + "right": "pull_request:47090", + "accept": true, + "reason": "Same CohereASR loss-shift bug and same code path; different implementation style, but same underlying fix." + }, + { + "left": "pull_request:46895", + "right": "pull_request:46958", + "accept": true, + "reason": "Both are alternative fixes for the same CohereASR training-loss double-shift problem." + }, + { + "left": "pull_request:47077", + "right": "pull_request:47156", + "accept": false, + "reason": "Japanese DistilBERT docs translation vs a Furiosa typo/doc tweak in Optimum; unrelated." + }, + { + "left": "pull_request:47163", + "right": "pull_request:47166", + "accept": false, + "reason": "Paged-attention XPU test additions vs quantization test extensions; both are tests, but not the same bug/change." + }, + { + "left": "pull_request:46707", + "right": "pull_request:47034", + "accept": false, + "reason": "Native FSDP2 integration vs Neuron SFT experiment scripts; separate infrastructure work." + }, + { + "left": "pull_request:46658", + "right": "pull_request:46836", + "accept": false, + "reason": "Step-3.7 Flash model support vs Unlimited OCR model support; different models and code paths." + }, + { + "left": "pull_request:43451", + "right": "pull_request:46836", + "accept": false, + "reason": "Molmo2 and Unlimited OCR are separate model additions, despite similar registration boilerplate." + }, + { + "left": "pull_request:46658", + "right": "pull_request:47143", + "accept": false, + "reason": "Step-3.7 Flash model support vs NeuCodec model support; unrelated model additions." + }, + { + "left": "pull_request:46836", + "right": "pull_request:47143", + "accept": false, + "reason": "Unlimited OCR and NeuCodec are distinct model onboarding PRs." + }, + { + "left": "pull_request:46836", + "right": "pull_request:47163", + "accept": false, + "reason": "Unlimited OCR model addition vs paged-attention test changes; unrelated." + }, + { + "left": "pull_request:46196", + "right": "pull_request:47143", + "accept": false, + "reason": "SDPA dispatch support for wav2vec2-family models vs NeuCodec model addition; different changes." + }, + { + "left": "pull_request:47143", + "right": "pull_request:47163", + "accept": false, + "reason": "NeuCodec model addition vs paged-attention XPU tests; unrelated." + }, + { + "left": "pull_request:46196", + "right": "pull_request:47163", + "accept": false, + "reason": "Audio SDPA dispatch support vs paged-attention test additions; not the same underlying bug." + }, + { + "left": "pull_request:46836", + "right": "pull_request:47161", + "accept": false, + "reason": "Unlimited OCR model addition vs exporter minimum-version checks; unrelated." + }, + { + "left": "pull_request:43451", + "right": "pull_request:47077", + "accept": false, + "reason": "Molmo2 model onboarding vs Japanese translation docs; unrelated." + }, + { + "left": "pull_request:43451", + "right": "pull_request:47161", + "accept": false, + "reason": "Molmo2 model onboarding vs exporter version requirements; unrelated." + }, + { + "left": "pull_request:43451", + "right": "pull_request:46196", + "accept": false, + "reason": "Molmo2 model support vs SDPA dispatch support for audio models; unrelated." + }, + { + "left": "pull_request:40546", + "right": "pull_request:43451", + "accept": false, + "reason": "VibeVoice and Molmo2 are distinct model additions with different files and implementations." + }, + { + "left": "pull_request:47143", + "right": "pull_request:47161", + "accept": false, + "reason": "NeuCodec onboarding and exporter version enforcement are unrelated." + }, + { + "left": "pull_request:47161", + "right": "pull_request:47163", + "accept": false, + "reason": "Exporter dependency checks vs paged-attention test coverage; different concerns." + }, + { + "left": "pull_request:43451", + "right": "pull_request:46658", + "accept": false, + "reason": "Molmo2 and Step-3.7 Flash are separate model additions." + }, + { + "left": "pull_request:46658", + "right": "pull_request:47096", + "accept": false, + "reason": "Step-3.7 Flash model support vs Kimi exportability changes; unrelated." + }, + { + "left": "pull_request:43451", + "right": "pull_request:47143", + "accept": false, + "reason": "Molmo2 and NeuCodec are different model onboarding PRs." + }, + { + "left": "pull_request:43451", + "right": "pull_request:47096", + "accept": false, + "reason": "Molmo2 model addition vs exportable Kimi work; unrelated." + }, + { + "left": "pull_request:43451", + "right": "pull_request:47003", + "accept": false, + "reason": "Molmo2 model support vs OpenVINO exporter support; different changes." + }, + { + "left": "pull_request:47003", + "right": "pull_request:47096", + "accept": false, + "reason": "OpenVINO exporter introduction and Kimi export support are different exporter features." + }, + { + "left": "pull_request:43451", + "right": "pull_request:47034", + "accept": false, + "reason": "Molmo2 model addition vs Neuron training experiment; unrelated." + }, + { + "left": "pull_request:47034", + "right": "pull_request:47096", + "accept": false, + "reason": "Neuron training experiment vs Kimi export support; different code paths." + }, + { + "left": "pull_request:47003", + "right": "pull_request:47034", + "accept": false, + "reason": "OpenVINO exporter support and Neuron experiment scripts are unrelated." + }, + { + "left": "pull_request:47003", + "right": "pull_request:47166", + "accept": false, + "reason": "OpenVINO exporter work vs quantization test updates; not the same change." + }, + { + "left": "pull_request:47096", + "right": "pull_request:47143", + "accept": false, + "reason": "Kimi export support vs NeuCodec model onboarding; unrelated." + }, + { + "left": "pull_request:47003", + "right": "pull_request:47143", + "accept": false, + "reason": "OpenVINO exporter addition vs NeuCodec model addition; unrelated." + }, + { + "left": "pull_request:47096", + "right": "pull_request:47163", + "accept": false, + "reason": "Kimi export support vs paged-attention XPU tests; different concerns." + }, + { + "left": "pull_request:47003", + "right": "pull_request:47163", + "accept": false, + "reason": "OpenVINO exporter support vs paged-attention tests; unrelated." + }, + { + "left": "pull_request:47143", + "right": "pull_request:47166", + "accept": false, + "reason": "NeuCodec model addition vs quantization test changes; unrelated." + }, + { + "left": "pull_request:46836", + "right": "pull_request:47096", + "accept": false, + "reason": "Unlimited OCR model onboarding and exportable Kimi are unrelated." + }, + { + "left": "pull_request:46836", + "right": "pull_request:47003", + "accept": false, + "reason": "Unlimited OCR model onboarding vs OpenVINO exporter introduction; unrelated." + } + ] + }, + "evaluator_result": null, + "evaluator_used": false, + "retried": false, + "accepted_nontrivial_soft_edge": false, + "error_kind": null, + "error_message": null + }, + { + "label": "pull_request", + "component_index": 2, + "component_count": 33, + "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": "a284edac83525c4dcd397d77f0edfa36defa5d3b9e21cf13944f9945ae49386f", + "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": "These two PRs are unrelated feature additions in different model families, so they are not duplicates.", + "confidence": 0.99, + "canonical_issue_reason": null, + "canonical_pr_reason": "No canonical PR: 40755 adds TimesFM covariate forecasting, while 41224 adds DINOv3 image-classification support; they touch different code paths and solve different problems.", + "best_issue_reason": null, + "best_pr_reason": "Neither PR is a good global representative for a duplicate cluster because the changes are unrelated; if forced, 41224 is the more self-contained feature PR, but it does not subsume 40755.", + "soft_edge_verdicts": [ + { + "left": "pull_request:40755", + "right": "pull_request:41224", + "accept": false, + "reason": "Different underlying changes: TimesFM forecasting with covariates versus DINOv3 classification support. They modify different model implementations, docs, and tests, and could not plausibly be merged into one PR." + } + ] + }, + "evaluator_result": null, + "evaluator_used": false, + "retried": false, + "accepted_nontrivial_soft_edge": false, + "error_kind": null, + "error_message": null + }, + { + "label": "pull_request", + "component_index": 3, + "component_count": 33, + "review_unit_index": 1, + "review_unit_count": 1, + "status": "reviewed", + "reason": null, + "source": "cache", + "cache_hit": true, + "model": "gpt-5.4-mini?service_tier=flex", + "cluster_id": "cluster-41356-2", + "nodes": [ + "pull_request:41356", + "pull_request:44339" + ], + "soft_pairs": [ + "pull_request:41356|pull_request:44339" + ], + "prepared_review_unit_hash": "aead1308c167b0bae6ff11bf9358340baa5e76f816618383418b505a10a9cd01", + "component_budget": { + "node_count": 2, + "item_count": 2, + "soft_pair_count": 1, + "serialized_chars": 6050, + "estimated_input_tokens": 1513, + "estimated_eval_tokens": 3282 + }, + "budget": { + "node_count": 2, + "item_count": 2, + "soft_pair_count": 1, + "serialized_chars": 6050, + "estimated_input_tokens": 1513, + "estimated_eval_tokens": 3282 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": false, + "analyst_result": { + "summary": "Two DEIMv2 pull requests target the same model family, but they are not the same change: PR 41356 is an early/basic model + image processor add, while PR 44339 is a much broader follow-up that adds loss code, auto-mappings, conversion tooling, modular files, and more integration work.", + "confidence": 0.91, + "canonical_issue_reason": null, + "canonical_pr_reason": "PR 44339 is the more complete and representative DEIMv2 implementation; PR 41356 looks like an earlier partial version of the same model bring-up.", + "best_issue_reason": null, + "best_pr_reason": "PR 44339 is the strongest single artifact for the cluster because it subsumes the basic model work and includes the broader integration needed for a full Transformers addition.", + "soft_edge_verdicts": [ + { + "left": "pull_request:41356", + "right": "pull_request:44339", + "accept": false, + "reason": "Same model area, but not the same concrete change: 44339 expands the implementation substantially and adds new code paths beyond the basic PR 41356. They look like related iterations, not mergeable duplicates." + } + ] + }, + "evaluator_result": null, + "evaluator_used": false, + "retried": false, + "accepted_nontrivial_soft_edge": false, + "error_kind": null, + "error_message": null + }, + { + "label": "pull_request", + "component_index": 4, + "component_count": 33, + "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-41855-12", + "nodes": [ + "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" + ], + "soft_pairs": [ + "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:45655", + "pull_request:43823|pull_request:43995", + "pull_request:43995|pull_request:45213", + "pull_request:43995|pull_request:44827", + "pull_request:44827|pull_request:45170", + "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:43251|pull_request:43254" + ], + "prepared_review_unit_hash": "21bd118a02a57abbc59ec784caa415e04b6b3b0ed81f6bb5474bcce55dfc2b55", + "component_budget": { + "node_count": 12, + "item_count": 12, + "soft_pair_count": 17, + "serialized_chars": 31598, + "estimated_input_tokens": 7900, + "estimated_eval_tokens": 16056 + }, + "budget": { + "node_count": 12, + "item_count": 12, + "soft_pair_count": 17, + "serialized_chars": 31598, + "estimated_input_tokens": 7900, + "estimated_eval_tokens": 16056 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": false, + "analyst_result": { + "summary": "This cluster is mostly a false-positive mix of unrelated PRs spanning tokenizer methods, processor/model registration, export tests, new model additions, refactors, and kernel/loss plumbing. The only clear duplicate pair is the two `fixed_cross_entropy` PRs, which both extend `loss_utils.py` to pass supported `cross_entropy` kwargs for the same issue.", + "confidence": 0.93, + "canonical_issue_reason": null, + "canonical_pr_reason": "PR 43251 is the best canonical PR within the only true duplicate pair: it directly targets issue 43240, adds the missing `cross_entropy` kwargs in `fixed_cross_entropy`, and is the more complete/explicit fix compared with the narrower alternate implementation in 43254.", + "best_issue_reason": null, + "best_pr_reason": "PR 43251 is the strongest representative of the cluster\u2019s only real duplicate subcluster because it has the explicit issue link, broader fix semantics, and clear alignment with the bug report; the other PRs are unrelated or too broad to serve as a canonical duplicate.", + "soft_edge_verdicts": [ + { + "left": "pull_request:41855", + "right": "pull_request:43424", + "accept": false, + "reason": "Different problems: Mistral tokenizer method additions vs ExecuTorch dynamic-shape export testing. No shared code-path bug." + }, + { + "left": "pull_request:41855", + "right": "pull_request:43823", + "accept": false, + "reason": "Unrelated: tokenizer API gaps vs adding a new MobileLLM model package and weights. Not the same underlying change." + }, + { + "left": "pull_request:42668", + "right": "pull_request:43823", + "accept": false, + "reason": "Both touch processor/model auto-registration areas, but one is about robust `from_pretrained` handling while the other adds a new model. Different changes." + }, + { + "left": "pull_request:43251", + "right": "pull_request:43823", + "accept": false, + "reason": "`fixed_cross_entropy` kwargs handling is unrelated to adding MobileLLM model files and configs." + }, + { + "left": "pull_request:43251", + "right": "pull_request:44827", + "accept": false, + "reason": "Loss-function kwargs support is unrelated to Mistral4 test/model fixes." + }, + { + "left": "pull_request:43424", + "right": "pull_request:44827", + "accept": false, + "reason": "ExecuTorch dynamic-shape export tests are unrelated to Mistral4 model/test fixes." + }, + { + "left": "pull_request:43424", + "right": "pull_request:43823", + "accept": false, + "reason": "Different subsystems and different bugs: exportability tests vs new model addition." + }, + { + "left": "pull_request:43823", + "right": "pull_request:45655", + "accept": false, + "reason": "New MobileLLM model addition is unrelated to `cls.config` resolution order in `modeling_utils.py`." + }, + { + "left": "pull_request:43823", + "right": "pull_request:43995", + "accept": false, + "reason": "New model addition vs Falcon output-collection refactor; not the same code-path issue." + }, + { + "left": "pull_request:43995", + "right": "pull_request:45213", + "accept": false, + "reason": "Falcon model refactoring is unrelated to the model-creation skill/doc scaffold PR." + }, + { + "left": "pull_request:43995", + "right": "pull_request:44827", + "accept": false, + "reason": "Falcon standardized output interface refactor is unrelated to Mistral4 test fixes." + }, + { + "left": "pull_request:44827", + "right": "pull_request:45170", + "accept": false, + "reason": "Mistral4 tests/model changes are unrelated to the `layrnorm` spelling fix across CLIP-like models." + }, + { + "left": "pull_request:44827", + "right": "pull_request:46048", + "accept": false, + "reason": "Mistral4 test fixes are unrelated to adding the TDT loss kernel." + }, + { + "left": "pull_request:44827", + "right": "pull_request:45655", + "accept": false, + "reason": "Mistral4 fixes and `cls.config` resolution order are different bugs in different areas." + }, + { + "left": "pull_request:44827", + "right": "pull_request:45213", + "accept": false, + "reason": "Mistral4 test/model work is unrelated to the new model-creation skill PR." + }, + { + "left": "pull_request:43823", + "right": "pull_request:45213", + "accept": false, + "reason": "Adding MobileLLM is unrelated to a model-creation skill/documentation/tooling PR." + }, + { + "left": "pull_request:43251", + "right": "pull_request:43254", + "accept": true, + "reason": "Both PRs fix the same concrete bug in `src/transformers/loss/loss_utils.py`: forwarding supported kwargs (`weight`, `label_smoothing`) to `nn.functional.cross_entropy` for issue 43240." + } + ] + }, + "evaluator_result": { + "accept": true, + "feedback": "The summary is grounded in the packet, and the only accepted duplicate pair is well supported by the shared issue target and identical loss-utils bugfix. The false soft-edge verdicts are conservative and the reasons do not overstate equivalence." + }, + "evaluator_used": true, + "retried": false, + "accepted_nontrivial_soft_edge": true, + "error_kind": null, + "error_message": null + }, + { + "label": "pull_request", + "component_index": 5, + "component_count": 33, + "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": "1211cf2760ee585121d744d28bb3faa62d146c762f468f0eed2583c6b5376e35", + "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 touch GitHub Actions workflow dependencies, but they are different maintenance changes: one is a broad Node 24 compatibility upgrade for many actions, while the other is a separate bump of docker/build-push-action to newer versions. They should not be treated as duplicates.", + "confidence": 0.91, + "canonical_issue_reason": null, + "canonical_pr_reason": "PR 42975 is the better canonical representative because it has the clearest, broader stated purpose: upgrading workflow actions for Node 24 compatibility across many files.", + "best_issue_reason": null, + "best_pr_reason": "PR 42975 is the strongest standalone representative of the cluster since it addresses a concrete platform-compatibility goal rather than only refreshing a narrower action dependency.", + "soft_edge_verdicts": [ + { + "left": "pull_request:42975", + "right": "pull_request:42976", + "accept": false, + "reason": "Both are workflow dependency bumps, but they target different changes. 42975 upgrades several actions for Node 24 compatibility; 42976 specifically updates docker/build-push-action versions. They do not appear to fix the same underlying problem and would not naturally merge 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": 6, + "component_count": 33, + "review_unit_index": 1, + "review_unit_count": 1, + "status": "reviewed", + "reason": null, + "source": "llm", + "cache_hit": false, + "model": "gpt-5.4-mini?service_tier=flex", + "cluster_id": "cluster-42978-4", + "nodes": [ + "pull_request:42978", + "pull_request:43448", + "pull_request:45490", + "pull_request:46562" + ], + "soft_pairs": [ + "pull_request:42978|pull_request:43448", + "pull_request:42978|pull_request:45490", + "pull_request:43448|pull_request:45490", + "pull_request:45490|pull_request:46562" + ], + "prepared_review_unit_hash": "7ea4ada15fd5532a27102bd16f521e758e3bad537708b411ccbf37a9c1595651", + "component_budget": { + "node_count": 4, + "item_count": 4, + "soft_pair_count": 4, + "serialized_chars": 13044, + "estimated_input_tokens": 3261, + "estimated_eval_tokens": 6778 + }, + "budget": { + "node_count": 4, + "item_count": 4, + "soft_pair_count": 4, + "serialized_chars": 13044, + "estimated_input_tokens": 3261, + "estimated_eval_tokens": 6778 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": false, + "analyst_result": { + "summary": "These are four separate model-addition PRs that only overlap on shared framework registration/docs files; they do not fix the same bug or implement the same model change, so none should be clustered as duplicates.", + "confidence": 0.98, + "canonical_issue_reason": null, + "canonical_pr_reason": "No single canonical PR for the cluster: each PR adds a different model (ViT NEPA, Molmo, CTSM, Step3p7) with only generic infrastructure overlap.", + "best_issue_reason": null, + "best_pr_reason": "46562 is the strongest single representative if one must be picked: it is a substantial, self-contained model addition with broad integration across auto-mapping, processing, modeling, and tests.", + "soft_edge_verdicts": [ + { + "left": "pull_request:42978", + "right": "pull_request:43448", + "accept": false, + "reason": "Different model additions (ViT NEPA vs Molmo); shared auto/docs files are expected for registration, not evidence of the same underlying change." + }, + { + "left": "pull_request:42978", + "right": "pull_request:45490", + "accept": false, + "reason": "Different architectures and domains (vision model vs time-series CTSM); overlap is only in generic Transformers plumbing." + }, + { + "left": "pull_request:43448", + "right": "pull_request:45490", + "accept": false, + "reason": "Molmo and CTSM are unrelated model implementations; they share only common repo-wide integration files." + }, + { + "left": "pull_request:45490", + "right": "pull_request:46562", + "accept": false, + "reason": "CTSM and Step3p7 target different code paths and model families; common auto/mapping changes are not enough to treat them as duplicates." + } + ] + }, + "evaluator_result": null, + "evaluator_used": false, + "retried": false, + "accepted_nontrivial_soft_edge": false, + "error_kind": null, + "error_message": null + }, + { + "label": "pull_request", + "component_index": 7, + "component_count": 33, + "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-43291-7", + "nodes": [ + "pull_request:43291", + "pull_request:43665", + "pull_request:44815", + "pull_request:46380", + "pull_request:46539", + "pull_request:46886", + "pull_request:46922" + ], + "soft_pairs": [ + "pull_request:43291|pull_request:43665", + "pull_request:43291|pull_request:44815", + "pull_request:43291|pull_request:46380", + "pull_request:43291|pull_request:46922", + "pull_request:43665|pull_request:46539", + "pull_request:43665|pull_request:44815", + "pull_request:43665|pull_request:46380", + "pull_request:43665|pull_request:46922", + "pull_request:44815|pull_request:46539", + "pull_request:44815|pull_request:46380", + "pull_request:44815|pull_request:46922", + "pull_request:46380|pull_request:46539", + "pull_request:46380|pull_request:46922", + "pull_request:46539|pull_request:46922", + "pull_request:46886|pull_request:46922" + ], + "prepared_review_unit_hash": "2a7f30f19a63ee0c2cd04e5cca0fc50df2684d4a61e55ae4eb9ee88889c0b7ea", + "component_budget": { + "node_count": 7, + "item_count": 7, + "soft_pair_count": 15, + "serialized_chars": 19601, + "estimated_input_tokens": 4901, + "estimated_eval_tokens": 10058 + }, + "budget": { + "node_count": 7, + "item_count": 7, + "soft_pair_count": 15, + "serialized_chars": 19601, + "estimated_input_tokens": 4901, + "estimated_eval_tokens": 10058 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": false, + "analyst_result": { + "summary": "Only the two YOLOS PRs are true duplicates; the rest are generic fix/test PRs touching different models or subsystems and should stay separate.", + "confidence": 0.94, + "canonical_issue_reason": null, + "canonical_pr_reason": "PR 46922 is the most complete YOLOS fix: it addresses the device-mismatch path directly and also includes the no-split-module change, subsuming PR 46886.", + "best_issue_reason": null, + "best_pr_reason": "PR 46922 is the strongest representative for the cluster because it is the clearest and most complete concrete code fix among the candidates.", + "soft_edge_verdicts": [ + { + "left": "pull_request:43291", + "right": "pull_request:43665", + "accept": false, + "reason": "Both are vague test-fix PRs, but they touch different models and different test files; no shared underlying bug." + }, + { + "left": "pull_request:43291", + "right": "pull_request:44815", + "accept": false, + "reason": "Whisper tokenizer/test changes are unrelated to the FP8 dequant/loading fix." + }, + { + "left": "pull_request:43291", + "right": "pull_request:46380", + "accept": false, + "reason": "Whisper test cleanup and generation/watermarking logic are different code paths." + }, + { + "left": "pull_request:43291", + "right": "pull_request:46922", + "accept": false, + "reason": "Whisper tests versus YOLOS device-mismatch fix are unrelated." + }, + { + "left": "pull_request:43665", + "right": "pull_request:46539", + "accept": false, + "reason": "CLIP/ViT test changes are not the same issue as the Phimoe CI integration test fix." + }, + { + "left": "pull_request:43665", + "right": "pull_request:44815", + "accept": false, + "reason": "Generic test edits in vision models do not match the FP8 dequant/core loading change." + }, + { + "left": "pull_request:43665", + "right": "pull_request:46380", + "accept": false, + "reason": "Test-only changes for CLIP/ViT are unrelated to generation config and whisper generation fixes." + }, + { + "left": "pull_request:43665", + "right": "pull_request:46922", + "accept": false, + "reason": "Different models and different failure modes; only superficial similarity in being small fixes." + }, + { + "left": "pull_request:44815", + "right": "pull_request:46539", + "accept": false, + "reason": "FP8 dequant/loading logic and Phimoe integration CI setup are different problems." + }, + { + "left": "pull_request:44815", + "right": "pull_request:46380", + "accept": false, + "reason": "No shared concrete code path; FP8 quantization fix does not overlap with generation/watermarking and whisper changes." + }, + { + "left": "pull_request:44815", + "right": "pull_request:46922", + "accept": false, + "reason": "FP8 dequant fix and YOLOS device placement fix are unrelated." + }, + { + "left": "pull_request:46380", + "right": "pull_request:46539", + "accept": false, + "reason": "Generation config/watermarking and CI runner memory/device-map setup are different subsystems." + }, + { + "left": "pull_request:46380", + "right": "pull_request:46922", + "accept": false, + "reason": "Wholly different areas: generation config versus YOLOS model device handling." + }, + { + "left": "pull_request:46539", + "right": "pull_request:46922", + "accept": false, + "reason": "CI infrastructure for Phimoe is unrelated to the YOLOS device-mismatch bug." + }, + { + "left": "pull_request:46886", + "right": "pull_request:46922", + "accept": true, + "reason": "Same YOLOS device-mismatch/device_map='auto' bug; both target issue 46869 and the later PR adds the missing direct device fix alongside the same split-module guard." + } + ] + }, + "evaluator_result": { + "accept": true, + "feedback": "The summary is well-grounded in the packet: only the two YOLOS PRs share the same file and issue target, and the rest are clearly different models/subsystems. The soft-edge verdicts are conservative, with the one true duplicate pair correctly accepted and the others reasonably rejected as not mergeable into a single bugfix PR." + }, + "evaluator_used": true, + "retried": false, + "accepted_nontrivial_soft_edge": true, + "error_kind": null, + "error_message": null + }, + { + "label": "pull_request", + "component_index": 8, + "component_count": 33, + "review_unit_index": 1, + "review_unit_count": 1, + "status": "reviewed", + "reason": null, + "source": "llm", + "cache_hit": false, + "model": "gpt-5.4-mini?service_tier=flex", + "cluster_id": "cluster-43488-3", + "nodes": [ + "pull_request:43488", + "pull_request:43532", + "pull_request:46747" + ], + "soft_pairs": [ + "pull_request:43488|pull_request:46747", + "pull_request:43488|pull_request:43532", + "pull_request:43532|pull_request:46747" + ], + "prepared_review_unit_hash": "f11a602ea787e5434e84f5d13bebb9907a19dc87bc22a8cf6ca02e5d2eee3521", + "component_budget": { + "node_count": 3, + "item_count": 3, + "soft_pair_count": 3, + "serialized_chars": 6350, + "estimated_input_tokens": 1588, + "estimated_eval_tokens": 3432 + }, + "budget": { + "node_count": 3, + "item_count": 3, + "soft_pair_count": 3, + "serialized_chars": 6350, + "estimated_input_tokens": 1588, + "estimated_eval_tokens": 3432 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": false, + "analyst_result": { + "summary": "All three pull requests are unrelated '[don't merge]' test or demo changes. They touch different files and different workflows/code paths, so none are duplicates of each other.", + "confidence": 0.96, + "canonical_issue_reason": null, + "canonical_pr_reason": "No single PR represents the same underlying change as the others; they are separate test PRs with different purposes. If a representative is needed, PR 46747 is the cleanest single-file workflow tweak.", + "best_issue_reason": null, + "best_pr_reason": "PR 46747 is the most self-contained and concrete change: one workflow file with a narrow trigger update, unlike the broader or multi-file test PRs.", + "soft_edge_verdicts": [ + { + "left": "pull_request:43488", + "right": "pull_request:46747", + "accept": false, + "reason": "Different changes: 43488 is a setup/version/dependency test PR, while 46747 alters a security-gate workflow trigger. They do not fix the same bug or change." + }, + { + "left": "pull_request:43488", + "right": "pull_request:43532", + "accept": false, + "reason": "Only superficial overlap via the vit test file; 43488 is about repo-bot/version formatting and dependencies, while 43532 is a workflow/diff-display test PR." + }, + { + "left": "pull_request:43532", + "right": "pull_request:46747", + "accept": false, + "reason": "Both mention workflows, but they modify different workflow files for different behaviors. They are not the same underlying change and would not 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": 9, + "component_count": 33, + "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": "c19ae61ce20962412cc97de9c6cea0cbe99d1b92839240722c1ecbbef8a6893e", + "component_budget": { + "node_count": 2, + "item_count": 2, + "soft_pair_count": 1, + "serialized_chars": 6083, + "estimated_input_tokens": 1521, + "estimated_eval_tokens": 3298 + }, + "budget": { + "node_count": 2, + "item_count": 2, + "soft_pair_count": 1, + "serialized_chars": 6083, + "estimated_input_tokens": 1521, + "estimated_eval_tokens": 3298 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": false, + "analyst_result": { + "summary": "These are closely related pipeline-adding PRs in the same area, but they implement different features and should not be merged as duplicates: one adds Promptable Concept Segmentation, the other adds Promptable Visual Segmentation.", + "confidence": 0.95, + "canonical_issue_reason": null, + "canonical_pr_reason": "No true canonical duplicate PR here; if forced to pick one representative, PR 43613 is slightly broader and more central to the shared multimodal segmentation pipeline work.", + "best_issue_reason": null, + "best_pr_reason": "PR 43613 is the better standalone representative because it covers a larger surface area and adds more model/test integration, but it is still a distinct feature from PR 43612.", + "soft_edge_verdicts": [ + { + "left": "pull_request:43612", + "right": "pull_request:43613", + "accept": false, + "reason": "Reject: they target different concrete pipelines and APIs (concept segmentation vs visual segmentation). Shared docs, registry, and test scaffolding are not enough to treat them as the same underlying 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": 10, + "component_count": 33, + "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": "b7410709e84ca5c17e15c14e06dfcce691fe65753345b30a6cbe646e2efcad65", + "component_budget": { + "node_count": 2, + "item_count": 2, + "soft_pair_count": 1, + "serialized_chars": 4292, + "estimated_input_tokens": 1073, + "estimated_eval_tokens": 2402 + }, + "budget": { + "node_count": 2, + "item_count": 2, + "soft_pair_count": 1, + "serialized_chars": 4292, + "estimated_input_tokens": 1073, + "estimated_eval_tokens": 2402 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": false, + "analyst_result": { + "summary": "These two PRs both modify `trainer.py`, but they address different trainer behaviors: one changes gradient-accumulation loss scaling, the other refactors signature-column filtering for column removal. They are not the same underlying change.", + "confidence": 0.98, + "canonical_issue_reason": null, + "canonical_pr_reason": "PR #43651 is the clearer canonical item for the cluster because it makes a focused behavior change in training loss scaling and introduces a concrete hook for custom trainers.", + "best_issue_reason": null, + "best_pr_reason": "PR #43651 is the better representative because its purpose is narrowly defined and the diff directly changes the training-step loss-scaling path; PR #43663 is a separate refactor of input-column inspection.", + "soft_edge_verdicts": [ + { + "left": "pull_request:43651", + "right": "pull_request:43663", + "accept": false, + "reason": "Different code paths and different goals: loss scaling during training vs. signature column derivation for filtering. Shared filename alone is not enough to treat them as duplicates." + } + ] + }, + "evaluator_result": null, + "evaluator_used": false, + "retried": false, + "accepted_nontrivial_soft_edge": false, + "error_kind": null, + "error_message": null + }, + { + "label": "pull_request", + "component_index": 11, + "component_count": 33, + "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": "275dd27bff1549111c6a1485377a8853ae865ad9e5bfd1edee82e0c01b045b44", + "component_budget": { + "node_count": 2, + "item_count": 2, + "soft_pair_count": 1, + "serialized_chars": 3663, + "estimated_input_tokens": 916, + "estimated_eval_tokens": 2088 + }, + "budget": { + "node_count": 2, + "item_count": 2, + "soft_pair_count": 1, + "serialized_chars": 3663, + "estimated_input_tokens": 916, + "estimated_eval_tokens": 2088 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": false, + "analyst_result": { + "summary": "Both PRs touch `src/transformers/cli/serve.py` and mention `TypeAdapter`, but they appear to fix different annotation/import-time problems rather than the same concrete bug. One is a broader `from __future__ import annotations` typing cleanup, while the other specifically handles `TypeAdapter` when `pydantic` is absent.", + "confidence": 0.82, + "canonical_issue_reason": null, + "canonical_pr_reason": "PR 43842 is the more specific and better-scoped fix: it explicitly guards the `TypeAdapter` import under `TYPE_CHECKING` and quotes the annotation, directly addressing the optional-`pydantic` NameError path.", + "best_issue_reason": null, + "best_pr_reason": "PR 43842 is the strongest representative of this cluster because its title, issue target, and diff all align on the concrete `TypeAdapter`/missing-`pydantic` failure mode.", + "soft_edge_verdicts": [ + { + "left": "pull_request:43656", + "right": "pull_request:43842", + "accept": false, + "reason": "Related file and symbol, but the fixes target different runtime/type-annotation failure modes; they are not clearly the same bug and would not naturally 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": 12, + "component_count": 33, + "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": 33, + "review_unit_index": 1, + "review_unit_count": 2, + "status": "reviewed", + "reason": null, + "source": "llm", + "cache_hit": false, + "model": "gpt-5.4-mini?service_tier=flex", + "cluster_id": "cluster-43996-28", + "nodes": [ + "pull_request:43996", + "pull_request:43997", + "pull_request:44007", + "pull_request:44010", + "pull_request:44013", + "pull_request:44017", + "pull_request:44018", + "pull_request:44019", + "pull_request:44024", + "pull_request:44044", + "pull_request:44056", + "pull_request:44059", + "pull_request:44066", + "pull_request:44068", + "pull_request:44071", + "pull_request:44072", + "pull_request:44073", + "pull_request:44074", + "pull_request:44076", + "pull_request:44085", + "pull_request:44086", + "pull_request:44098", + "pull_request:44101", + "pull_request:44116", + "pull_request:44129", + "pull_request:44154", + "pull_request:44161", + "pull_request:44722" + ], + "soft_pairs": [ + "pull_request:43996|pull_request:44085", + "pull_request:43996|pull_request:44044", + "pull_request:44066|pull_request:44085", + "pull_request:44007|pull_request:44072", + "pull_request:44072|pull_request:44722", + "pull_request:44066|pull_request:44072", + "pull_request:44013|pull_request:44044", + "pull_request:44066|pull_request:44086", + "pull_request:44018|pull_request:44068", + "pull_request:44066|pull_request:44071", + "pull_request:44066|pull_request:44068", + "pull_request:44018|pull_request:44066", + "pull_request:44019|pull_request:44722", + "pull_request:44086|pull_request:44722", + "pull_request:44019|pull_request:44085", + "pull_request:44018|pull_request:44019", + "pull_request:44019|pull_request:44071", + "pull_request:44018|pull_request:44116", + "pull_request:44068|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:44098|pull_request:44101", + "pull_request:44085|pull_request:44086", + "pull_request:44059|pull_request:44086", + "pull_request:44059|pull_request:44722", + "pull_request:44018|pull_request:44071", + "pull_request:44017|pull_request:44161", + "pull_request:44017|pull_request:44073", + "pull_request:44017|pull_request:44722", + "pull_request:44068|pull_request:44071", + "pull_request:43997|pull_request:44073", + "pull_request:44017|pull_request:44071", + "pull_request:44071|pull_request:44086", + "pull_request:44073|pull_request:44161", + "pull_request:44073|pull_request:44101", + "pull_request:44059|pull_request:44071", + "pull_request:44056|pull_request:44101", + "pull_request:44018|pull_request:44059", + "pull_request:44059|pull_request:44066", + "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:44010|pull_request:44071", + "pull_request:44018|pull_request:44129", + "pull_request:44010|pull_request:44073", + "pull_request:43997|pull_request:44074", + "pull_request:44010|pull_request:44018", + "pull_request:43997|pull_request:44129", + "pull_request:44071|pull_request:44129", + "pull_request:43996|pull_request:44161", + "pull_request:43996|pull_request:44017", + "pull_request:44010|pull_request:44044", + "pull_request:44044|pull_request:44074", + "pull_request:44066|pull_request:44722", + "pull_request:44085|pull_request:44722", + "pull_request:44085|pull_request:44129", + "pull_request:43996|pull_request:44007", + "pull_request:43996|pull_request:44129", + "pull_request:44044|pull_request:44129", + "pull_request:44044|pull_request:44154", + "pull_request:44024|pull_request:44076", + "pull_request:44073|pull_request:44074" + ], + "prepared_review_unit_hash": "d9edcba9ca529763de101e1b6c708a577515a17856fa095ab9c9a44d26252512", + "component_budget": { + "node_count": 28, + "item_count": 28, + "soft_pair_count": 76, + "serialized_chars": 84111, + "estimated_input_tokens": 21028, + "estimated_eval_tokens": 42312 + }, + "budget": { + "node_count": 28, + "item_count": 28, + "soft_pair_count": 72, + "serialized_chars": 76933, + "estimated_input_tokens": 19234, + "estimated_eval_tokens": 38724 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": true, + "analyst_result": { + "summary": "This cluster is mostly a family of model-specific refactors that standardize output tracing / capture_outputs across Transformers models. Most PRs are not duplicates of each other because they touch different model code paths; the only clear overlaps are within the GPT-J/CodeGen subcluster and the GPT-Neo pair.", + "confidence": 0.91, + "canonical_issue_reason": null, + "canonical_pr_reason": "PR 44722 is the strongest representative of the overlapping GPT-J/CodeGen subcluster: it targets the same output-tracing refactor as the other GPT-J PRs, touches the same concrete files, and is the most complete/latest instance among the clearly overlapping changes.", + "best_issue_reason": null, + "best_pr_reason": "44722 is the best overall PR anchor for this cluster because it has the clearest concrete overlap with other accepted duplicates and the strongest explicit refactor framing; the rest of the cluster is too model-diverse to share one canonical PR cleanly.", + "soft_edge_verdicts": [ + { + "left": "pull_request:43996", + "right": "pull_request:44085", + "accept": false, + "reason": "Different model families (CVT/FNet vs GPT-J); same refactor theme is too broad to be a duplicate." + }, + { + "left": "pull_request:43996", + "right": "pull_request:44044", + "accept": false, + "reason": "Different models and code paths (CVT/FNet vs DeBERTa V2)." + }, + { + "left": "pull_request:44066", + "right": "pull_request:44085", + "accept": true, + "reason": "Both refactor GPT-J output tracing in the same file/code path; 44066 is a broader superset that also includes CodeGen." + }, + { + "left": "pull_request:44007", + "right": "pull_request:44072", + "accept": false, + "reason": "RegNet/ResNet vs EfficientNet are distinct implementations, not the same underlying change." + }, + { + "left": "pull_request:44072", + "right": "pull_request:44722", + "accept": false, + "reason": "EfficientNet and GPT-J/CodeGen are unrelated model paths." + }, + { + "left": "pull_request:44066", + "right": "pull_request:44072", + "accept": false, + "reason": "GPT-J/CodeGen output tracing vs EfficientNet output tracing are different concrete changes." + }, + { + "left": "pull_request:44013", + "right": "pull_request:44044", + "accept": false, + "reason": "MobileNetV2 vs DeBERTa V2; only the refactor pattern is shared." + }, + { + "left": "pull_request:44066", + "right": "pull_request:44086", + "accept": false, + "reason": "GPT-J/CodeGen refactor vs MGP-STR refactor; unrelated code paths." + }, + { + "left": "pull_request:44018", + "right": "pull_request:44068", + "accept": true, + "reason": "Both target GPT-Neo output tracing in the same model/file, and appear to be alternate versions of the same refactor." + }, + { + "left": "pull_request:44066", + "right": "pull_request:44071", + "accept": false, + "reason": "GPT-J/CodeGen vs MPT; different model families." + }, + { + "left": "pull_request:44066", + "right": "pull_request:44068", + "accept": false, + "reason": "GPT-J/CodeGen and GPT-Neo are different model code paths." + }, + { + "left": "pull_request:44018", + "right": "pull_request:44066", + "accept": false, + "reason": "GPT-Neo vs GPT-J/CodeGen; same issue theme but not the same change." + }, + { + "left": "pull_request:44019", + "right": "pull_request:44722", + "accept": false, + "reason": "ResNet output tracing is unrelated to GPT-J/CodeGen output tracing." + }, + { + "left": "pull_request:44086", + "right": "pull_request:44722", + "accept": false, + "reason": "MGP-STR and GPT-J/CodeGen are unrelated model implementations." + }, + { + "left": "pull_request:44019", + "right": "pull_request:44085", + "accept": false, + "reason": "ResNet and GPT-J are different model paths." + }, + { + "left": "pull_request:44018", + "right": "pull_request:44019", + "accept": false, + "reason": "GPT-Neo and ResNet are unrelated." + }, + { + "left": "pull_request:44019", + "right": "pull_request:44071", + "accept": false, + "reason": "ResNet vs MPT; same broad pattern but not the same concrete change." + }, + { + "left": "pull_request:44018", + "right": "pull_request:44116", + "accept": false, + "reason": "GPT-Neo vs Flaubert/XLM; different model families." + }, + { + "left": "pull_request:44068", + "right": "pull_request:44116", + "accept": false, + "reason": "GPT-Neo vs Flaubert/XLM; only the decorator refactor theme overlaps." + }, + { + "left": "pull_request:44068", + "right": "pull_request:44722", + "accept": false, + "reason": "GPT-Neo and GPT-J/CodeGen are different model code paths." + }, + { + "left": "pull_request:44071", + "right": "pull_request:44722", + "accept": false, + "reason": "MPT and GPT-J/CodeGen are unrelated implementations." + }, + { + "left": "pull_request:44018", + "right": "pull_request:44085", + "accept": false, + "reason": "GPT-Neo vs GPT-J; different model-specific refactors." + }, + { + "left": "pull_request:44018", + "right": "pull_request:44086", + "accept": false, + "reason": "GPT-Neo vs MGP-STR, unrelated." + }, + { + "left": "pull_request:44068", + "right": "pull_request:44086", + "accept": false, + "reason": "GPT-Neo vs MGP-STR, unrelated." + }, + { + "left": "pull_request:44071", + "right": "pull_request:44085", + "accept": false, + "reason": "MPT vs GPT-J; not the same code-path problem." + }, + { + "left": "pull_request:44086", + "right": "pull_request:44085", + "accept": false, + "reason": "MGP-STR and GPT-J are different model families." + }, + { + "left": "pull_request:44059", + "right": "pull_request:44086", + "accept": false, + "reason": "GPT2 vs MGP-STR; unrelated changes." + }, + { + "left": "pull_request:44018", + "right": "pull_request:44071", + "accept": false, + "reason": "GPT-Neo vs MPT, different models." + }, + { + "left": "pull_request:44017", + "right": "pull_request:44161", + "accept": false, + "reason": "SegFormer vs LongT5; same standardized-output theme but not the same underlying change." + }, + { + "left": "pull_request:44017", + "right": "pull_request:44073", + "accept": false, + "reason": "SegFormer vs VisualBERT are different model paths." + }, + { + "left": "pull_request:44017", + "right": "pull_request:44722", + "accept": false, + "reason": "SegFormer output tracing is unrelated to GPT-J/CodeGen." + }, + { + "left": "pull_request:44068", + "right": "pull_request:44071", + "accept": false, + "reason": "GPT-Neo vs MPT; different model-specific changes." + }, + { + "left": "pull_request:43997", + "right": "pull_request:44073", + "accept": false, + "reason": "RegNet vs VisualBERT are unrelated models." + }, + { + "left": "pull_request:44017", + "right": "pull_request:44071", + "accept": false, + "reason": "SegFormer vs MPT; same refactor pattern only." + }, + { + "left": "pull_request:44071", + "right": "pull_request:44086", + "accept": false, + "reason": "MPT vs MGP-STR; unrelated model code paths." + }, + { + "left": "pull_request:44073", + "right": "pull_request:44161", + "accept": false, + "reason": "VisualBERT vs LongT5; not the same concrete change." + }, + { + "left": "pull_request:44073", + "right": "pull_request:44101", + "accept": false, + "reason": "VisualBERT vs Flaubert/XLM; different model families." + }, + { + "left": "pull_request:44059", + "right": "pull_request:44071", + "accept": false, + "reason": "GPT2 vs MPT are distinct model implementations." + }, + { + "left": "pull_request:44056", + "right": "pull_request:44101", + "accept": false, + "reason": "MPNet vs Flaubert/XLM; unrelated." + }, + { + "left": "pull_request:44018", + "right": "pull_request:44059", + "accept": false, + "reason": "GPT-Neo vs GPT2; same general refactor theme, different models." + }, + { + "left": "pull_request:44059", + "right": "pull_request:44066", + "accept": false, + "reason": "GPT2 vs GPT-J/CodeGen; unrelated concrete changes." + }, + { + "left": "pull_request:44101", + "right": "pull_request:44161", + "accept": false, + "reason": "Flaubert/XLM vs LongT5; different model paths." + }, + { + "left": "pull_request:43997", + "right": "pull_request:44017", + "accept": false, + "reason": "RegNet vs SegFormer are unrelated implementations." + }, + { + "left": "pull_request:43997", + "right": "pull_request:44056", + "accept": false, + "reason": "RegNet vs MPNet; same theme, different code paths." + }, + { + "left": "pull_request:44071", + "right": "pull_request:44076", + "accept": false, + "reason": "MPT vs ImageGPT; different models." + }, + { + "left": "pull_request:44074", + "right": "pull_request:44101", + "accept": false, + "reason": "TextNet vs Flaubert/XLM are unrelated." + }, + { + "left": "pull_request:44056", + "right": "pull_request:44161", + "accept": false, + "reason": "MPNet vs LongT5; different model-specific refactors." + }, + { + "left": "pull_request:43997", + "right": "pull_request:44018", + "accept": false, + "reason": "RegNet vs GPT-Neo; unrelated changes." + }, + { + "left": "pull_request:43997", + "right": "pull_request:44161", + "accept": false, + "reason": "RegNet vs LongT5 are unrelated model families." + }, + { + "left": "pull_request:44074", + "right": "pull_request:44161", + "accept": false, + "reason": "TextNet vs LongT5; different code paths." + }, + { + "left": "pull_request:44010", + "right": "pull_request:44071", + "accept": false, + "reason": "SqueezeBERT vs MPT; unrelated models." + }, + { + "left": "pull_request:44018", + "right": "pull_request:44129", + "accept": false, + "reason": "GPT-Neo vs SpeechT5; different model families." + }, + { + "left": "pull_request:44010", + "right": "pull_request:44073", + "accept": false, + "reason": "SqueezeBERT vs VisualBERT; same refactor theme only." + }, + { + "left": "pull_request:43997", + "right": "pull_request:44074", + "accept": false, + "reason": "RegNet vs TextNet; unrelated model implementations." + }, + { + "left": "pull_request:44010", + "right": "pull_request:44018", + "accept": false, + "reason": "SqueezeBERT vs GPT-Neo are different code paths." + }, + { + "left": "pull_request:44044", + "right": "pull_request:44074", + "accept": false, + "reason": "DeBERTa V2 vs TextNet; unrelated." + }, + { + "left": "pull_request:44024", + "right": "pull_request:44076", + "accept": false, + "reason": "FocalNet vs ImageGPT; different model-specific changes." + }, + { + "left": "pull_request:44073", + "right": "pull_request:44074", + "accept": false, + "reason": "VisualBERT vs TextNet; not the same underlying change." + }, + { + "left": "pull_request:44044", + "right": "pull_request:44129", + "accept": false, + "reason": "DeBERTa V2 vs SpeechT5 are unrelated model paths." + }, + { + "left": "pull_request:44044", + "right": "pull_request:44154", + "accept": false, + "reason": "DeBERTa V2 vs VITS; different concrete refactors." + }, + { + "left": "pull_request:44066", + "right": "pull_request:44722", + "accept": true, + "reason": "Same GPT-J/CodeGen output-tracing refactor family, with matching filenames and overlapping implementation changes." + }, + { + "left": "pull_request:44085", + "right": "pull_request:44722", + "accept": true, + "reason": "Both are GPT-J output-tracing refactors on the same file and could plausibly be merged into one PR." + } + ] + }, + "evaluator_result": { + "accept": true, + "feedback": "The summary is grounded in the packet: it keeps the broad family-level refactor theme, but correctly narrows duplicate-like overlap to the same-model GPT-J/CodeGen and GPT-Neo subclusters. The accepted soft-edge pairs all look mergeable as single bugfix PRs, and the rejected pairs are conservatively separated by model family/code path. No obvious overstatement beyond the evidence." + }, + "evaluator_used": true, + "retried": false, + "accepted_nontrivial_soft_edge": true, + "error_kind": null, + "error_message": null + }, + { + "label": "pull_request", + "component_index": 13, + "component_count": 33, + "review_unit_index": 2, + "review_unit_count": 2, + "status": "reviewed", + "reason": null, + "source": "llm", + "cache_hit": false, + "model": "gpt-5.4-mini?service_tier=flex", + "cluster_id": "cluster-43996-28", + "nodes": [ + "pull_request:43997", + "pull_request:44010", + "pull_request:44056", + "pull_request:44073", + "pull_request:44074", + "pull_request:44129" + ], + "soft_pairs": [ + "pull_request: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": "f093ec7de7c9c150b2b9270c1ee4944e41cabfcb005e4a790d1db838f884651e", + "component_budget": { + "node_count": 28, + "item_count": 28, + "soft_pair_count": 76, + "serialized_chars": 84111, + "estimated_input_tokens": 21028, + "estimated_eval_tokens": 42312 + }, + "budget": { + "node_count": 6, + "item_count": 6, + "soft_pair_count": 4, + "serialized_chars": 12756, + "estimated_input_tokens": 3189, + "estimated_eval_tokens": 6634 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": true, + "analyst_result": { + "summary": "These PRs are all variants of the same broad refactor pattern\u2014migrating different models to standardized output tracing/capture decorators\u2014but they touch different model implementations and code paths, so they are not duplicates of one another.", + "confidence": 0.89, + "canonical_issue_reason": null, + "canonical_pr_reason": "No single PR is a true duplicate anchor, but PR #44010 is the most representative of the cluster theme: it explicitly refactors a model to standardized output collection decorators, has review activity, and is a clean example of the pattern seen across the others.", + "best_issue_reason": null, + "best_pr_reason": "PR #44129 is the strongest standalone candidate for triage because it includes an explicit issue target, a substantial output-tracing refactor in a concrete model path, and enough detail to understand the intended change without needing the rest of the cluster.", + "soft_edge_verdicts": [ + { + "left": "pull_request:44056", + "right": "pull_request:44073", + "accept": false, + "reason": "Both are standardized output-tracing refactors, but they target different models (MPNet vs VisualBert) and different internal layers. Same refactor theme, not the same underlying bug/change." + }, + { + "left": "pull_request:44056", + "right": "pull_request:44074", + "accept": false, + "reason": "These touch different model implementations (MPNet vs TextNet). They share a refactor pattern, but not a single concrete code-path fix." + }, + { + "left": "pull_request:43997", + "right": "pull_request:44010", + "accept": false, + "reason": "RegNet and SqueezeBert are unrelated model code paths. The similarity is only the migration pattern, so they should not be merged as duplicates." + }, + { + "left": "pull_request:44010", + "right": "pull_request:44129", + "accept": false, + "reason": "SqueezeBert and SpeechT5 are separate model stacks with different forward/output mechanics. Both are output-tracing refactors, but 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": "pull_request", + "component_index": 14, + "component_count": 33, + "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": "609d579b637123ea9204aaf283ee90fc2880469c617046a01dd4e01ee4373da3", + "component_budget": { + "node_count": 13, + "item_count": 13, + "soft_pair_count": 40, + "serialized_chars": 40172, + "estimated_input_tokens": 10043, + "estimated_eval_tokens": 20342 + }, + "budget": { + "node_count": 13, + "item_count": 13, + "soft_pair_count": 40, + "serialized_chars": 40172, + "estimated_input_tokens": 10043, + "estimated_eval_tokens": 20342 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": false, + "analyst_result": { + "summary": "This cluster is a set of parallel, model-specific output-tracing refactors, not duplicate fixes. The PRs share a pattern and tooling, but each one targets a different architecture and code path, so they should stay separate.", + "confidence": 0.93, + "canonical_issue_reason": null, + "canonical_pr_reason": "PR 44003 is the broadest representative in the set, spanning both `falcon_mamba` and `mamba` and showing the clearest instance of the shared output-tracing refactor pattern.", + "best_issue_reason": null, + "best_pr_reason": "PR 44003 is the strongest representative of this cluster because it covers multiple related Mamba-family files and the core tracing machinery, making it the best single exemplar of the refactor series.", + "soft_edge_verdicts": [ + { + "left": "pull_request:43998", + "right": "pull_request:44002", + "accept": false, + "reason": "Different model families (`timm_backbone` vs `upernet`); same refactor theme but not the same bug/change." + }, + { + "left": "pull_request:43998", + "right": "pull_request:44000", + "accept": false, + "reason": "`timm_backbone` and `vision_text_dual_encoder` are unrelated code paths; both only adopt output-tracing helpers." + }, + { + "left": "pull_request:43998", + "right": "pull_request:44001", + "accept": false, + "reason": "Separate models (`timm_backbone` vs `univnet`); no shared concrete defect." + }, + { + "left": "pull_request:43998", + "right": "pull_request:44003", + "accept": false, + "reason": "Different architectures and files; this is a pattern match, not a duplicate fix." + }, + { + "left": "pull_request:43998", + "right": "pull_request:44029", + "accept": false, + "reason": "`timm_backbone` and `rwkv` are distinct model implementations with different tracing behavior." + }, + { + "left": "pull_request:43999", + "right": "pull_request:44001", + "accept": false, + "reason": "`mobilenet_v1` and `univnet` touch unrelated forward paths; same tooling theme only." + }, + { + "left": "pull_request:43999", + "right": "pull_request:44002", + "accept": false, + "reason": "`mobilenet_v1` vs `upernet`; both refactor outputs, but not the same underlying change." + }, + { + "left": "pull_request:43999", + "right": "pull_request:44025", + "accept": false, + "reason": "Different vision models and different output contracts; not mergeable as one PR." + }, + { + "left": "pull_request:43999", + "right": "pull_request:44026", + "accept": false, + "reason": "`mobilenet_v1` and `vision_encoder_decoder` are separate code paths and refactor scopes." + }, + { + "left": "pull_request:43999", + "right": "pull_request:44027", + "accept": false, + "reason": "`mobilenet_v1` vs `speech_encoder_decoder`; parallel refactors, not duplicates." + }, + { + "left": "pull_request:44000", + "right": "pull_request:44001", + "accept": false, + "reason": "`vision_text_dual_encoder` and `univnet` are unrelated model families; only the refactor pattern matches." + }, + { + "left": "pull_request:44000", + "right": "pull_request:44002", + "accept": false, + "reason": "Different subsystems (`vision_text_dual_encoder` vs `upernet`); no shared concrete bug." + }, + { + "left": "pull_request:44000", + "right": "pull_request:44003", + "accept": false, + "reason": "`vision_text_dual_encoder` and `falcon_mamba/mamba` change different forward implementations." + }, + { + "left": "pull_request:44000", + "right": "pull_request:44025", + "accept": false, + "reason": "Distinct model families and filenames; same output-tracing cleanup, not the same change." + }, + { + "left": "pull_request:44001", + "right": "pull_request:44002", + "accept": false, + "reason": "`univnet` and `upernet` are unrelated; only generic tracing refactor overlap." + }, + { + "left": "pull_request:44001", + "right": "pull_request:44003", + "accept": false, + "reason": "Different architectures (`univnet` vs `mamba` family) and different code paths." + }, + { + "left": "pull_request:44001", + "right": "pull_request:44025", + "accept": false, + "reason": "Separate model implementations; no evidence they fix one shared defect." + }, + { + "left": "pull_request:44002", + "right": "pull_request:44029", + "accept": false, + "reason": "`upernet` and `rwkv` are different models with different output semantics." + }, + { + "left": "pull_request:44002", + "right": "pull_request:44003", + "accept": false, + "reason": "`upernet` vs `mamba` family; same refactor style, not same bug." + }, + { + "left": "pull_request:44003", + "right": "pull_request:44028", + "accept": false, + "reason": "`mamba` family and `superpoint` are unrelated implementations; no common code path." + }, + { + "left": "pull_request:44003", + "right": "pull_request:44004", + "accept": false, + "reason": "`mamba`/`falcon_mamba` and `codegen` are different architectures; not a single mergeable fix." + }, + { + "left": "pull_request:43998", + "right": "pull_request:44004", + "accept": false, + "reason": "`timm_backbone` and `codegen` touch unrelated model code; same theme only." + }, + { + "left": "pull_request:44001", + "right": "pull_request:44004", + "accept": false, + "reason": "`univnet` and `codegen` are different domains and output shapes; not duplicates." + }, + { + "left": "pull_request:44002", + "right": "pull_request:44004", + "accept": false, + "reason": "`upernet` and `codegen` are separate model paths with no shared defect." + }, + { + "left": "pull_request:44025", + "right": "pull_request:44026", + "accept": false, + "reason": "Different composite models (`depth_anything` vs `vision_encoder_decoder`); not the same change." + }, + { + "left": "pull_request:44025", + "right": "pull_request:44027", + "accept": false, + "reason": "`depth_anything` and `speech_encoder_decoder` are unrelated refactors." + }, + { + "left": "pull_request:44025", + "right": "pull_request:44028", + "accept": false, + "reason": "`depth_anything` and `superpoint` are different vision models with different output mechanics." + }, + { + "left": "pull_request:44025", + "right": "pull_request:44029", + "accept": false, + "reason": "`depth_anything` vs `rwkv`; distinct code paths and model families." + }, + { + "left": "pull_request:44025", + "right": "pull_request:44030", + "accept": false, + "reason": "`depth_anything` and `dpr` are unrelated model implementations." + }, + { + "left": "pull_request:44026", + "right": "pull_request:44029", + "accept": false, + "reason": "`vision_encoder_decoder` and `rwkv` are not the same bug/change." + }, + { + "left": "pull_request:44026", + "right": "pull_request:44027", + "accept": false, + "reason": "`vision_encoder_decoder` vs `speech_encoder_decoder`; similar wrappers, but separate model code." + }, + { + "left": "pull_request:44026", + "right": "pull_request:44028", + "accept": false, + "reason": "`vision_encoder_decoder` and `superpoint` are different model types and output paths." + }, + { + "left": "pull_request:44026", + "right": "pull_request:44030", + "accept": false, + "reason": "`vision_encoder_decoder` and `dpr` are distinct implementations; no shared concrete fix." + }, + { + "left": "pull_request:44027", + "right": "pull_request:44028", + "accept": false, + "reason": "`speech_encoder_decoder` and `superpoint` are unrelated; only the helper refactor overlaps." + }, + { + "left": "pull_request:44027", + "right": "pull_request:44029", + "accept": false, + "reason": "`speech_encoder_decoder` vs `rwkv`; different architecture and tracing behavior." + }, + { + "left": "pull_request:44027", + "right": "pull_request:44030", + "accept": false, + "reason": "`speech_encoder_decoder` and `dpr` are separate model paths, not duplicates." + }, + { + "left": "pull_request:44028", + "right": "pull_request:44029", + "accept": false, + "reason": "`superpoint` and `rwkv` are unrelated changes despite similar output-capturing work." + }, + { + "left": "pull_request:44028", + "right": "pull_request:44030", + "accept": false, + "reason": "`superpoint` and `dpr` target different models and output contracts." + }, + { + "left": "pull_request:44029", + "right": "pull_request:44030", + "accept": false, + "reason": "`rwkv` and `dpr` are different model families; not the same underlying 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": 15, + "component_count": 33, + "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": "371c7ea1367aeb62edd3c82403e7d5ddcf72a50f1f80a602fc95eafedcb8c3f3", + "component_budget": { + "node_count": 2, + "item_count": 2, + "soft_pair_count": 1, + "serialized_chars": 6551, + "estimated_input_tokens": 1638, + "estimated_eval_tokens": 3532 + }, + "budget": { + "node_count": 2, + "item_count": 2, + "soft_pair_count": 1, + "serialized_chars": 6551, + "estimated_input_tokens": 1638, + "estimated_eval_tokens": 3532 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": false, + "analyst_result": { + "summary": "Two separate model-addition PRs: one for PenguinVL and one for Voxtral TTS. They share Transformers boilerplate and auto-registration files, but they implement different models with different codepaths, so they are not duplicates.", + "confidence": 0.98, + "canonical_issue_reason": null, + "canonical_pr_reason": "PR #45401 is the more canonical choice for this cluster because it is narrowly scoped to Voxtral TTS, explicitly targets issue #45027, and has a focused implementation for a single model release.", + "best_issue_reason": null, + "best_pr_reason": "PR #45401 is the best representative PR: it has an explicit issue link, a tighter scope, and clearly documents the intended Voxtral TTS support. PR #44662 is a separate PenguinVL implementation and should not be grouped as the same change.", + "soft_edge_verdicts": [ + { + "left": "pull_request:44662", + "right": "pull_request:45401", + "accept": false, + "reason": "Different models and different underlying functionality. The overlap is only shared Transformers scaffolding (docs, auto-mappings, package registration), not the same 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": 16, + "component_count": 33, + "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": "40d8d2ccda757175996a7b5a715655e22f8acdf96fd72630b3f66cb75fb9857b", + "component_budget": { + "node_count": 2, + "item_count": 2, + "soft_pair_count": 1, + "serialized_chars": 4407, + "estimated_input_tokens": 1102, + "estimated_eval_tokens": 2460 + }, + "budget": { + "node_count": 2, + "item_count": 2, + "soft_pair_count": 1, + "serialized_chars": 4407, + "estimated_input_tokens": 1102, + "estimated_eval_tokens": 2460 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": false, + "analyst_result": { + "summary": "These are related by theme (module/kernel fusion) but they are not the same change: one introduces a generic module-fusion API, the other wires up n-to-1 kernel fusion through KernelConfig and integration code paths.", + "confidence": 0.91, + "canonical_issue_reason": null, + "canonical_pr_reason": "PR 45363 is the better representative of this cluster: it implements a concrete end-to-end fusion workflow across modeling/integration code, whereas PR 44979 is a more foundational API proposal in a separate code path.", + "best_issue_reason": null, + "best_pr_reason": "PR 45363 is the most user-facing and complete of the two, with explicit KernelConfig integration and example usage; PR 44979 is broader infrastructure rather than the specific fused behavior.", + "soft_edge_verdicts": [ + { + "left": "pull_request:44979", + "right": "pull_request:45363", + "accept": false, + "reason": "Different underlying changes: 44979 adds a generic module-fusion API and tests, while 45363 implements n-to-1 kernel fusion via KernelConfig in integrations/modeling utilities. They are conceptually adjacent but not the same fix/feature and would not be merged as one PR." + } + ] + }, + "evaluator_result": null, + "evaluator_used": false, + "retried": false, + "accepted_nontrivial_soft_edge": false, + "error_kind": null, + "error_message": null + }, + { + "label": "pull_request", + "component_index": 17, + "component_count": 33, + "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": "63d856062941e5b4b484b436b5b1167241b394c51f01f25ca638bfc18b700584", + "component_budget": { + "node_count": 2, + "item_count": 2, + "soft_pair_count": 1, + "serialized_chars": 4413, + "estimated_input_tokens": 1104, + "estimated_eval_tokens": 2464 + }, + "budget": { + "node_count": 2, + "item_count": 2, + "soft_pair_count": 1, + "serialized_chars": 4413, + "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 patch the same Mistral regex pre-tokenizer AttributeError in the same file and target the same issue; 45317 is the more complete version because it includes tests.", + "confidence": 0.98, + "canonical_issue_reason": null, + "canonical_pr_reason": "pull_request:45317 is the better canonical PR: it fixes the same AttributeError as 45086 and also adds test coverage, making it the more complete change set.", + "best_issue_reason": null, + "best_pr_reason": "pull_request:45317 is the strongest representative of the cluster because it covers the full fix plus tests, while 45086 is the narrower code-only variant.", + "soft_edge_verdicts": [ + { + "left": "pull_request:45086", + "right": "pull_request:45317", + "accept": true, + "reason": "Same concrete bug fix in _patch_mistral_regex on the same code path, same target issue, and same changed file; 45317 is essentially the tested version of 45086." + } + ] + }, + "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": 33, + "review_unit_index": 1, + "review_unit_count": 1, + "status": "reviewed", + "reason": null, + "source": "llm", + "cache_hit": false, + "model": "gpt-5.4-mini?service_tier=flex", + "cluster_id": "cluster-45115-2", + "nodes": [ + "pull_request:45115", + "pull_request:45569" + ], + "soft_pairs": [ + "pull_request:45115|pull_request:45569" + ], + "prepared_review_unit_hash": "2527672c54bb54a30aa71cc017871ea3c3fc6ec18a12cfceafbeafc7fb9a5b2d", + "component_budget": { + "node_count": 2, + "item_count": 2, + "soft_pair_count": 1, + "serialized_chars": 6066, + "estimated_input_tokens": 1517, + "estimated_eval_tokens": 3290 + }, + "budget": { + "node_count": 2, + "item_count": 2, + "soft_pair_count": 1, + "serialized_chars": 6066, + "estimated_input_tokens": 1517, + "estimated_eval_tokens": 3290 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": false, + "analyst_result": { + "summary": "Two distinct NemotronH PRs: one refactors NemotronH to inherit GraniteMoeHybrid and add conversion mapping, while the other splits NemotronH into Dense/Sparse variants with docs, auto-mapping, and tests. They overlap on shared model files but are not the same change.", + "confidence": 0.95, + "canonical_issue_reason": null, + "canonical_pr_reason": "PR #45569 is the better canonical PR because it represents the larger, more user-facing feature work: introducing NemotronHDense/NemotronHSparse, updating docs/auto mappings, and adding tests. PR #45115 is a narrower refactor/conversion-mapping adjustment.", + "best_issue_reason": null, + "best_pr_reason": "PR #45569 is the best representative of the cluster since it covers the main architectural split and packaging changes. PR #45115 is related but only addresses a smaller inheritance/conversion refactor.", + "soft_edge_verdicts": [ + { + "left": "pull_request:45115", + "right": "pull_request:45569", + "accept": false, + "reason": "Shared filenames indicate both touch NemotronH internals, but the concrete changes differ: #45115 refactors inheritance and conversion key renaming, while #45569 adds separate Dense/Sparse model variants plus docs and tests. They are not the same underlying fix/change 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": 19, + "component_count": 33, + "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": "d0a6182857cc95b1722edf81f70281e688fee164d42955df0c3673facee35649", + "component_budget": { + "node_count": 2, + "item_count": 2, + "soft_pair_count": 1, + "serialized_chars": 6080, + "estimated_input_tokens": 1520, + "estimated_eval_tokens": 3296 + }, + "budget": { + "node_count": 2, + "item_count": 2, + "soft_pair_count": 1, + "serialized_chars": 6080, + "estimated_input_tokens": 1520, + "estimated_eval_tokens": 3296 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": false, + "analyst_result": { + "summary": "These are related but not duplicates: PR #45332 is a broad heterogeneous model feature PR, while PR #45333 is a narrower configuration-only slice of the same overall theme.", + "confidence": 0.92, + "canonical_issue_reason": null, + "canonical_pr_reason": "PR #45332 is the better canonical candidate because it covers the full heterogeneous model feature set, including modeling, cache/masking/integration changes, not just per-layer configuration.", + "best_issue_reason": null, + "best_pr_reason": "PR #45332 is the strongest standalone PR because it contains the broader end-to-end implementation; PR #45333 is a narrower subset focused on configuration utilities.", + "soft_edge_verdicts": [ + { + "left": "pull_request:45332", + "right": "pull_request:45333", + "accept": false, + "reason": "They share the same feature area and some files, but #45332 is a much broader model-support PR while #45333 is configuration-only. They are not the same concrete change and would not plausibly collapse into 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": 33, + "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": "d177ba845645d9b4bad36a93d7351a2d2de2645f54528ed08ac9dc99c5169719", + "component_budget": { + "node_count": 2, + "item_count": 2, + "soft_pair_count": 1, + "serialized_chars": 5513, + "estimated_input_tokens": 1379, + "estimated_eval_tokens": 3014 + }, + "budget": { + "node_count": 2, + "item_count": 2, + "soft_pair_count": 1, + "serialized_chars": 5513, + "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 address different change scopes: one is a broad type-checking cleanup across many Transformer modules, while the other adds/adjusts typing protocols and targeted fixes around `modeling_utils`/PEFT integration.", + "confidence": 0.88, + "canonical_issue_reason": null, + "canonical_pr_reason": "PR 45425 is the better canonical representative because it is more focused and clearly centered on concrete typing API additions and related fixes, rather than a sweeping repo-wide type-checking pass.", + "best_issue_reason": null, + "best_pr_reason": "PR 45425 is the best standalone PR to keep: it has a narrower, more coherent purpose and is easier to evaluate/merge than the much broader type-checking cleanup in 45415.", + "soft_edge_verdicts": [ + { + "left": "pull_request:45415", + "right": "pull_request:45425", + "accept": false, + "reason": "Reject: these are not the same underlying change. 45415 is a broad typing cleanup across many files plus pyproject type-ignore config, while 45425 introduces typing helpers/protocols and targeted fixes in `_typing.py`, `modeling_utils`, PEFT, and flash-attention code. Shared typing files do not make them duplicate PRs." + } + ] + }, + "evaluator_result": null, + "evaluator_used": false, + "retried": false, + "accepted_nontrivial_soft_edge": false, + "error_kind": null, + "error_message": null + }, + { + "label": "pull_request", + "component_index": 21, + "component_count": 33, + "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": "d1b9a1f4a9d79bf968021b5960a0045ba48f17bb42a1bc61108f86dfd6e7738e", + "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": "Two open PRs both touch expert-parallel loading paths, but they address different bugs: one fixes EP + DeepSpeed ZeRO-3 loading via accelerate launch, the other fixes expert parameters being overwritten under EP + FSDP2/rank-0 broadcast. Shared modeling_utils edits are not enough to make them duplicates.", + "confidence": 0.91, + "canonical_issue_reason": null, + "canonical_pr_reason": "PR #45662 is the better canonical representative for the cluster because it targets the more concrete expert-parallel overwrite failure and spans multiple relevant loading/integration paths.", + "best_issue_reason": null, + "best_pr_reason": "PR #45662 is the strongest standalone fix: it addresses a specific distributed-loading bug with a clear failure mode and broader code impact than the narrower accelerate-launch loading patch.", + "soft_edge_verdicts": [ + { + "left": "pull_request:45548", + "right": "pull_request:45662", + "accept": false, + "reason": "Reject: they concern different underlying EP distributed-loading issues and would not plausibly be merged into one PR without conflating separate fixes." + } + ] + }, + "evaluator_result": null, + "evaluator_used": false, + "retried": false, + "accepted_nontrivial_soft_edge": false, + "error_kind": null, + "error_message": null + }, + { + "label": "pull_request", + "component_index": 22, + "component_count": 33, + "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": "6b5666e9a2bfe0bcfd0f9e191c3b55d053cba0c0d332cc510d4235abab565075", + "component_budget": { + "node_count": 2, + "item_count": 2, + "soft_pair_count": 1, + "serialized_chars": 4671, + "estimated_input_tokens": 1168, + "estimated_eval_tokens": 2592 + }, + "budget": { + "node_count": 2, + "item_count": 2, + "soft_pair_count": 1, + "serialized_chars": 4671, + "estimated_input_tokens": 1168, + "estimated_eval_tokens": 2592 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": false, + "analyst_result": { + "summary": "Both PRs target the same xdist-related captured_info logging collision around the testing debug artifact path; PR 45645 is the broader, more complete fix and subsumes the narrower file-write change in 45639.", + "confidence": 0.92, + "canonical_issue_reason": null, + "canonical_pr_reason": "PR 45645 is the best canonical representative because it addresses the core xdist collision in `testing_utils.py` and also updates CI log collection to read all `captured_info*.txt` files, making it the more complete end-to-end fix.", + "best_issue_reason": null, + "best_pr_reason": "PR 45645 is the strongest single artifact in the cluster: it fixes the underlying output-file collision and the downstream CI consumption path, whereas 45639 only covers the write-side change.", + "soft_edge_verdicts": [ + { + "left": "pull_request:45639", + "right": "pull_request:45645", + "accept": true, + "reason": "Same underlying bug: xdist-safe handling of `captured_info` debug logs. 45645 includes the same `testing_utils.py` fix plus the workflow change needed to consume per-worker files, so they are merge-compatible parts of one 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": 23, + "component_count": 33, + "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": 24, + "component_count": 33, + "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": "7d153538887ba8549ee47106bfa9a1123ed7ce9de7c749ac4c0e8f21c46b2cd3", + "component_budget": { + "node_count": 2, + "item_count": 2, + "soft_pair_count": 1, + "serialized_chars": 5526, + "estimated_input_tokens": 1382, + "estimated_eval_tokens": 3020 + }, + "budget": { + "node_count": 2, + "item_count": 2, + "soft_pair_count": 1, + "serialized_chars": 5526, + "estimated_input_tokens": 1382, + "estimated_eval_tokens": 3020 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": false, + "analyst_result": { + "summary": "Two PRs touch some of the same distributed/MoE files, but they are not the same change: #46126 is a focused EP + Trainer integration, while #46269 is a broad distributed branch base with many unrelated edits. They should not be deduplicated.", + "confidence": 0.93, + "canonical_issue_reason": null, + "canonical_pr_reason": "#46126 is the more specific, self-contained feature PR with a clear concrete goal (EP + Trainer integration). #46269 is a broad branch-base PR spanning many unrelated distributed changes, so it is a worse canonical representative.", + "best_issue_reason": null, + "best_pr_reason": "#46126 is the best PR candidate because it cleanly represents the underlying EP/Trainer feature work without the noise of unrelated branch-base churn. It is easier to triage and understand as the canonical change.", + "soft_edge_verdicts": [ + { + "left": "pull_request:46126", + "right": "pull_request:46269", + "accept": false, + "reason": "Reject: shared filenames and vague distributed-topic overlap are not enough. #46269 is a broad base branch with many unrelated modifications, not the same concrete bug fix or feature as #46126, so they 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": 25, + "component_count": 33, + "review_unit_index": 1, + "review_unit_count": 1, + "status": "reviewed", + "reason": null, + "source": "llm", + "cache_hit": false, + "model": "gpt-5.4-mini?service_tier=flex", + "cluster_id": "cluster-46138-10", + "nodes": [ + "pull_request:46138", + "pull_request:46432", + "pull_request:46501", + "pull_request:46502", + "pull_request:46503", + "pull_request:46504", + "pull_request:46554", + "pull_request:46558", + "pull_request:46563", + "pull_request:46716" + ], + "soft_pairs": [ + "pull_request:46138|pull_request:46554", + "pull_request:46138|pull_request:46558", + "pull_request:46138|pull_request:46563", + "pull_request:46138|pull_request:46716", + "pull_request:46138|pull_request:46504", + "pull_request:46432|pull_request:46503", + "pull_request:46432|pull_request:46501", + "pull_request:46432|pull_request:46502", + "pull_request:46432|pull_request:46554", + "pull_request:46432|pull_request:46558", + "pull_request:46501|pull_request:46503", + "pull_request:46501|pull_request:46502", + "pull_request:46501|pull_request:46504", + "pull_request:46501|pull_request:46554", + "pull_request:46502|pull_request:46503", + "pull_request:46502|pull_request:46504", + "pull_request:46502|pull_request:46554", + "pull_request:46503|pull_request:46504", + "pull_request:46503|pull_request:46554", + "pull_request:46504|pull_request:46554", + "pull_request:46504|pull_request:46558", + "pull_request:46554|pull_request:46558", + "pull_request:46554|pull_request:46563", + "pull_request:46554|pull_request:46716", + "pull_request:46558|pull_request:46563", + "pull_request:46558|pull_request:46716", + "pull_request:46563|pull_request:46716", + "pull_request:46504|pull_request:46563", + "pull_request:46432|pull_request:46563", + "pull_request:46504|pull_request:46716", + "pull_request:46432|pull_request:46716" + ], + "prepared_review_unit_hash": "57d2e4c49391c93d78699344d82a8682d498cb85564def9c7ed170c6550304e5", + "component_budget": { + "node_count": 10, + "item_count": 10, + "soft_pair_count": 31, + "serialized_chars": 26665, + "estimated_input_tokens": 6667, + "estimated_eval_tokens": 13590 + }, + "budget": { + "node_count": 10, + "item_count": 10, + "soft_pair_count": 31, + "serialized_chars": 26665, + "estimated_input_tokens": 6667, + "estimated_eval_tokens": 13590 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": false, + "analyst_result": { + "summary": "These are mostly unrelated workflow-maintenance PRs. Despite similar titles and some shared files, the diffs touch different jobs and different fixes, so none of the soft edges look like true duplicates.", + "confidence": 0.95, + "canonical_issue_reason": null, + "canonical_pr_reason": "No single PR is a true canonical duplicate anchor here; the set spans separate CI/workflow changes rather than one underlying fix.", + "best_issue_reason": null, + "best_pr_reason": "PR 46554 is the most substantive and security-focused change, but it is still not a duplicate of the others.", + "soft_edge_verdicts": [ + { + "left": "pull_request:46138", + "right": "pull_request:46554", + "accept": false, + "reason": "Different workflow files and different changes; not the same underlying bug or fix." + }, + { + "left": "pull_request:46138", + "right": "pull_request:46558", + "accept": false, + "reason": "Different workflows and unrelated updates; only broad CI similarity." + }, + { + "left": "pull_request:46138", + "right": "pull_request:46563", + "accept": false, + "reason": "Different workflows and unrelated updates; not mergeable as one fix." + }, + { + "left": "pull_request:46138", + "right": "pull_request:46716", + "accept": false, + "reason": "Different workflows and unrelated CI updates." + }, + { + "left": "pull_request:46138", + "right": "pull_request:46504", + "accept": false, + "reason": "Different workflow files; both are CI tweaks but fix different script behavior." + }, + { + "left": "pull_request:46432", + "right": "pull_request:46503", + "accept": false, + "reason": "Different workflow files and unrelated changes." + }, + { + "left": "pull_request:46432", + "right": "pull_request:46501", + "accept": false, + "reason": "Different workflows; one is SSH runner hardening, the other is anti-slop permission adjustment." + }, + { + "left": "pull_request:46432", + "right": "pull_request:46502", + "accept": false, + "reason": "Different workflows and different operational concerns; not the same fix." + }, + { + "left": "pull_request:46432", + "right": "pull_request:46554", + "accept": false, + "reason": "Different workflows and unrelated changes." + }, + { + "left": "pull_request:46432", + "right": "pull_request:46558", + "accept": false, + "reason": "Different workflows and unrelated changes." + }, + { + "left": "pull_request:46501", + "right": "pull_request:46503", + "accept": false, + "reason": "Different workflows and different purposes; only both are CI config changes." + }, + { + "left": "pull_request:46501", + "right": "pull_request:46502", + "accept": false, + "reason": "Different workflow files and unrelated fixes." + }, + { + "left": "pull_request:46501", + "right": "pull_request:46504", + "accept": false, + "reason": "Different workflow files; not the same underlying bug or change." + }, + { + "left": "pull_request:46501", + "right": "pull_request:46554", + "accept": false, + "reason": "Different workflow files and different concerns." + }, + { + "left": "pull_request:46502", + "right": "pull_request:46503", + "accept": false, + "reason": "Different workflow files and unrelated behavior changes." + }, + { + "left": "pull_request:46502", + "right": "pull_request:46504", + "accept": false, + "reason": "Different workflow files and different fixes." + }, + { + "left": "pull_request:46502", + "right": "pull_request:46554", + "accept": false, + "reason": "Different workflow files and unrelated changes." + }, + { + "left": "pull_request:46503", + "right": "pull_request:46504", + "accept": false, + "reason": "Different workflow files and different CI triggers/scripts; not one fix." + }, + { + "left": "pull_request:46503", + "right": "pull_request:46554", + "accept": false, + "reason": "Different workflow files and unrelated changes." + }, + { + "left": "pull_request:46504", + "right": "pull_request:46554", + "accept": false, + "reason": "Different workflow files and unrelated changes." + }, + { + "left": "pull_request:46504", + "right": "pull_request:46558", + "accept": false, + "reason": "Different workflow files and unrelated fixes." + }, + { + "left": "pull_request:46554", + "right": "pull_request:46558", + "accept": false, + "reason": "Same file, but different changes: one adds checkout-and-diff security checks, the other adjusts permissions/approval flow." + }, + { + "left": "pull_request:46554", + "right": "pull_request:46563", + "accept": false, + "reason": "Same file, but different changes: security-check logic versus reusable-workflow commentary." + }, + { + "left": "pull_request:46554", + "right": "pull_request:46716", + "accept": false, + "reason": "Same file, but different changes: workflow logic rewrite versus secrets handling." + }, + { + "left": "pull_request:46558", + "right": "pull_request:46563", + "accept": false, + "reason": "Same workflow file, but the patches address different aspects of the gate and are not one concrete fix." + }, + { + "left": "pull_request:46558", + "right": "pull_request:46716", + "accept": false, + "reason": "Same file, but different changes: permission/approval handling versus secrets inheritance." + }, + { + "left": "pull_request:46563", + "right": "pull_request:46716", + "accept": false, + "reason": "Same file, but different changes and different code paths." + }, + { + "left": "pull_request:46504", + "right": "pull_request:46563", + "accept": false, + "reason": "Different workflow files and different CI behavior changes." + }, + { + "left": "pull_request:46432", + "right": "pull_request:46563", + "accept": false, + "reason": "Different workflow files and unrelated changes." + }, + { + "left": "pull_request:46504", + "right": "pull_request:46716", + "accept": false, + "reason": "Different workflow files and unrelated changes." + }, + { + "left": "pull_request:46432", + "right": "pull_request:46716", + "accept": false, + "reason": "Different workflow files and unrelated changes." + } + ] + }, + "evaluator_result": null, + "evaluator_used": false, + "retried": false, + "accepted_nontrivial_soft_edge": false, + "error_kind": null, + "error_message": null + }, + { + "label": "pull_request", + "component_index": 26, + "component_count": 33, + "review_unit_index": 1, + "review_unit_count": 1, + "status": "reviewed", + "reason": null, + "source": "cache", + "cache_hit": true, + "model": "gpt-5.4-mini?service_tier=flex", + "cluster_id": "cluster-46174-2", + "nodes": [ + "pull_request:46174", + "pull_request:46780" + ], + "soft_pairs": [ + "pull_request:46174|pull_request:46780" + ], + "prepared_review_unit_hash": "a48ce4ca162f555682f79411d7229ac1ccc08d7822dfab654bb881639fccddbc", + "component_budget": { + "node_count": 2, + "item_count": 2, + "soft_pair_count": 1, + "serialized_chars": 6030, + "estimated_input_tokens": 1508, + "estimated_eval_tokens": 3272 + }, + "budget": { + "node_count": 2, + "item_count": 2, + "soft_pair_count": 1, + "serialized_chars": 6030, + "estimated_input_tokens": 1508, + "estimated_eval_tokens": 3272 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": false, + "analyst_result": { + "summary": "Two PRs target the same reported MPS crash in the sinusoidal position embedding helper, but they implement different fixes and are not safe to treat as duplicates.", + "confidence": 0.89, + "canonical_issue_reason": null, + "canonical_pr_reason": "PR 46174 is the broader, more complete fix for the float64-on-MPS crash: it updates the helper across all affected model files and keeps the internal math on CPU float64 before casting back.", + "best_issue_reason": null, + "best_pr_reason": "PR 46174 is the best representative of the cluster because it is the original, more comprehensive fix and has more discussion/review context than the narrower follow-up.", + "soft_edge_verdicts": [ + { + "left": "pull_request:46174", + "right": "pull_request:46780", + "accept": false, + "reason": "They share the same issue target and helper, but the patches are different solutions: 46174 avoids float64 on MPS by doing the arithmetic on CPU first, while 46780 switches the helper to use dtype for internal arithmetic. Those are alternative implementations, not one mergeable duplicate PR." + } + ] + }, + "evaluator_result": null, + "evaluator_used": false, + "retried": false, + "accepted_nontrivial_soft_edge": false, + "error_kind": null, + "error_message": null + }, + { + "label": "pull_request", + "component_index": 27, + "component_count": 33, + "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-46281-2", + "nodes": [ + "pull_request:46281", + "pull_request:46890" + ], + "soft_pairs": [ + "pull_request:46281|pull_request:46890" + ], + "prepared_review_unit_hash": "b5a044bef26c6d139cb7c86e609cceaa9a4d19efec2aa0f819288c607b0f4cfc", + "component_budget": { + "node_count": 2, + "item_count": 2, + "soft_pair_count": 1, + "serialized_chars": 4271, + "estimated_input_tokens": 1068, + "estimated_eval_tokens": 2392 + }, + "budget": { + "node_count": 2, + "item_count": 2, + "soft_pair_count": 1, + "serialized_chars": 4271, + "estimated_input_tokens": 1068, + "estimated_eval_tokens": 2392 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": false, + "analyst_result": { + "summary": "Both PRs harden local checkpoint/file resolution against path traversal, but they protect different inputs and code paths: sharded checkpoint `weight_map` handling in `hub.py` vs `transformers_weights` / explicit filename resolution in `modeling_utils.py`.", + "confidence": 0.91, + "canonical_issue_reason": null, + "canonical_pr_reason": "PR 46281 is the cleaner representative of the cluster theme: it directly fixes untrusted shard filenames in the sharded-checkpoint loader with a narrowly scoped change.", + "best_issue_reason": null, + "best_pr_reason": "PR 46890 is the broader security-hardening change and has more review activity, but it is still a separate code path from 46281.", + "soft_edge_verdicts": [ + { + "left": "pull_request:46281", + "right": "pull_request:46890", + "accept": false, + "reason": "Reject: both are path-traversal fixes, but they target different loaders and different untrusted fields. The diffs touch separate functions/files and would not naturally merge 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": 28, + "component_count": 33, + "review_unit_index": 1, + "review_unit_count": 1, + "status": "reviewed", + "reason": null, + "source": "llm", + "cache_hit": false, + "model": "gpt-5.4-mini?service_tier=flex", + "cluster_id": "cluster-46310-2", + "nodes": [ + "pull_request:46310", + "pull_request:47162" + ], + "soft_pairs": [ + "pull_request:46310|pull_request:47162" + ], + "prepared_review_unit_hash": "6fd7d2767fbdc41eaf510d2b0823e1de7cd3f6c8c4944044a16c52a9bc8ae603", + "component_budget": { + "node_count": 2, + "item_count": 2, + "soft_pair_count": 1, + "serialized_chars": 4883, + "estimated_input_tokens": 1221, + "estimated_eval_tokens": 2698 + }, + "budget": { + "node_count": 2, + "item_count": 2, + "soft_pair_count": 1, + "serialized_chars": 4883, + "estimated_input_tokens": 1221, + "estimated_eval_tokens": 2698 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": false, + "analyst_result": { + "summary": "Both PRs address the same off-by-one bug in MoonshineStreaming\u2019s right-side sliding window mask; PR 47162 is the broader, more complete fix and subsumes PR 46310.", + "confidence": 0.94, + "canonical_issue_reason": "Issue 46305 appears to be the underlying off-by-one right-window boundary bug in `sliding_window_mask_function` / `inner_mask`.", + "canonical_pr_reason": "PR 47162 is the better canonical PR because it fixes the same MoonshineStreaming bug, includes tests, and appears to supersede the narrower PR 46310 while also applying the same correction to Gemma4\u2019s analogous mask code.", + "best_issue_reason": "The shared issue target and identical boundary change strongly indicate a single underlying bug: the right window should include the boundary (`<=`) instead of excluding it (`<`).", + "best_pr_reason": "PR 47162 best represents the fix set because it contains the MoonshineStreaming correction plus validation coverage and broader related cleanup, whereas PR 46310 is only the narrower MoonshineStreaming-only version.", + "soft_edge_verdicts": [ + { + "left": "pull_request:46310", + "right": "pull_request:47162", + "accept": true, + "reason": "Both PRs fix the same concrete off-by-one in MoonshineStreaming\u2019s right-window mask; 47162 includes the same change and can plausibly subsume 46310." + } + ] + }, + "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": 33, + "review_unit_index": 1, + "review_unit_count": 1, + "status": "reviewed", + "reason": null, + "source": "llm", + "cache_hit": false, + "model": "gpt-5.4-mini?service_tier=flex", + "cluster_id": "cluster-46574-2", + "nodes": [ + "pull_request:46574", + "pull_request:46580" + ], + "soft_pairs": [ + "pull_request:46574|pull_request:46580" + ], + "prepared_review_unit_hash": "c30efa054957a9a0124b77df056586692208c785459d72940037b4db9c221faf", + "component_budget": { + "node_count": 2, + "item_count": 2, + "soft_pair_count": 1, + "serialized_chars": 4410, + "estimated_input_tokens": 1103, + "estimated_eval_tokens": 2462 + }, + "budget": { + "node_count": 2, + "item_count": 2, + "soft_pair_count": 1, + "serialized_chars": 4410, + "estimated_input_tokens": 1103, + "estimated_eval_tokens": 2462 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": false, + "analyst_result": { + "summary": "These two pull requests are unrelated fixes: one changes CodeLlama decode behavior to preserve real leading spaces, while the other adjusts tokenizer auto-selection for certain DeepSeek checkpoints. They should not be clustered as duplicates.", + "confidence": 0.98, + "canonical_issue_reason": null, + "canonical_pr_reason": "No single canonical PR: the two PRs address different bugs in different code paths (CodeLlama decoding vs. DeepSeek tokenizer-class selection).", + "best_issue_reason": null, + "best_pr_reason": "Neither PR is a better representative of the other; they solve separate issues and are not mergeable into one fix.", + "soft_edge_verdicts": [ + { + "left": "pull_request:46574", + "right": "pull_request:46580", + "accept": false, + "reason": "Different underlying problems and code paths: CodeLlama decoder stripping on decode versus DeepSeek auto-tokenizer misclassification. They do not appear to be 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": 30, + "component_count": 33, + "review_unit_index": 1, + "review_unit_count": 1, + "status": "reviewed", + "reason": null, + "source": "llm", + "cache_hit": false, + "model": "gpt-5.4-mini?service_tier=flex", + "cluster_id": "cluster-46773-2", + "nodes": [ + "pull_request:46773", + "pull_request:47087" + ], + "soft_pairs": [ + "pull_request:46773|pull_request:47087" + ], + "prepared_review_unit_hash": "68095d6ed89285c70a64f4ba626a139c6df548ebb72a6fecf54b4bd5c4d59561", + "component_budget": { + "node_count": 2, + "item_count": 2, + "soft_pair_count": 1, + "serialized_chars": 5432, + "estimated_input_tokens": 1358, + "estimated_eval_tokens": 2972 + }, + "budget": { + "node_count": 2, + "item_count": 2, + "soft_pair_count": 1, + "serialized_chars": 5432, + "estimated_input_tokens": 1358, + "estimated_eval_tokens": 2972 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": false, + "analyst_result": { + "summary": "These two PRs are related by theme (masking being skipped) but they fix different bugs in different code paths: 46773 targets batch-size-1 handling in linear-attention model forwards, while 47087 fixes recurrent-layer masking during continued/chunked forwards and cache continuation. They should not be deduplicated.", + "confidence": 0.97, + "canonical_issue_reason": null, + "canonical_pr_reason": "No single canonical PR: 46773 and 47087 address different masking regressions in different implementations, so neither can stand in for the other.", + "best_issue_reason": null, + "best_pr_reason": "47087 is the stronger standalone PR: it has an explicit issue link, a focused utility-level fix, and broader test coverage around the recurrent-mask continuation bug.", + "soft_edge_verdicts": [ + { + "left": "pull_request:46773", + "right": "pull_request:47087", + "accept": false, + "reason": "Same broad masking topic, but the underlying bugs and code paths differ: batch-size-1 linear-attention masking vs. recurrent continued-forward/cache-continuation masking." + } + ] + }, + "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": 33, + "review_unit_index": 1, + "review_unit_count": 1, + "status": "reviewed", + "reason": null, + "source": "llm", + "cache_hit": false, + "model": "gpt-5.4-mini?service_tier=flex", + "cluster_id": "cluster-47074-2", + "nodes": [ + "pull_request:47074", + "pull_request:47076" + ], + "soft_pairs": [ + "pull_request:47074|pull_request:47076" + ], + "prepared_review_unit_hash": "668ed5c6efe479c988678593021c5202e09f2218f8f2c6b0006d36060fead4e1", + "component_budget": { + "node_count": 2, + "item_count": 2, + "soft_pair_count": 1, + "serialized_chars": 4486, + "estimated_input_tokens": 1122, + "estimated_eval_tokens": 2500 + }, + "budget": { + "node_count": 2, + "item_count": 2, + "soft_pair_count": 1, + "serialized_chars": 4486, + "estimated_input_tokens": 1122, + "estimated_eval_tokens": 2500 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": false, + "analyst_result": { + "summary": "These are two unrelated pull requests in different multimodal code paths: one fixes a key mismatch in AnyToAnyPipeline, the other adds Gemma4 placeholder validation for audio/text inputs. They are not duplicates or mergeable into one change.", + "confidence": 0.97, + "canonical_issue_reason": null, + "canonical_pr_reason": "No canonical PR: 47074 and 47076 fix different bugs in different components, with no shared file targets or concrete code-path overlap.", + "best_issue_reason": null, + "best_pr_reason": "47076 is the stronger standalone fix in this cluster because it addresses a specific input-validation bug and includes companion test coverage; 47074 is a smaller pipeline key-normalization fix.", + "soft_edge_verdicts": [ + { + "left": "pull_request:47074", + "right": "pull_request:47076", + "accept": false, + "reason": "Different underlying problems and code paths: 47074 normalizes the AnyToAnyPipeline payload key from images to videos, while 47076 validates Gemma4 audio placeholder batching. They do not appear to be the same bug or a single combined 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": 32, + "component_count": 33, + "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-47105-2", + "nodes": [ + "pull_request:47105", + "pull_request:47129" + ], + "soft_pairs": [ + "pull_request:47105|pull_request:47129" + ], + "prepared_review_unit_hash": "27359829971ccb171755fa388e49d64d2b6c965d0a0a873579660fe8212f0b9d", + "component_budget": { + "node_count": 2, + "item_count": 2, + "soft_pair_count": 1, + "serialized_chars": 4182, + "estimated_input_tokens": 1046, + "estimated_eval_tokens": 2348 + }, + "budget": { + "node_count": 2, + "item_count": 2, + "soft_pair_count": 1, + "serialized_chars": 4182, + "estimated_input_tokens": 1046, + "estimated_eval_tokens": 2348 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": false, + "analyst_result": { + "summary": "These two PRs are unrelated: one re-enables a VITS forward test, the other fixes an RT-DETR v2 weight converter mapping issue. The soft similarity is superficial and should not be merged as a duplicate cluster.", + "confidence": 0.98, + "canonical_issue_reason": null, + "canonical_pr_reason": "No canonical PR: the changes target different models/code paths and do not appear to be the same underlying fix.", + "best_issue_reason": null, + "best_pr_reason": "Neither PR is a good global representative for the cluster because they solve unrelated problems in different subsystems.", + "soft_edge_verdicts": [ + { + "left": "pull_request:47105", + "right": "pull_request:47129", + "accept": false, + "reason": "Different artifacts and different fixes: VITS test expectations vs RT-DETR v2 converter key mapping. They do not look mergeable 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": 33, + "component_count": 33, + "review_unit_index": 1, + "review_unit_count": 1, + "status": "reviewed", + "reason": null, + "source": "llm", + "cache_hit": false, + "model": "gpt-5.4-mini?service_tier=flex", + "cluster_id": "cluster-47122-4", + "nodes": [ + "pull_request:47122", + "pull_request:47123", + "pull_request:47124", + "pull_request:47150" + ], + "soft_pairs": [ + "pull_request:47122|pull_request:47123", + "pull_request:47122|pull_request:47124", + "pull_request:47122|pull_request:47150", + "pull_request:47123|pull_request:47124", + "pull_request:47123|pull_request:47150", + "pull_request:47124|pull_request:47150" + ], + "prepared_review_unit_hash": "bbba2464a29c84a8a7f080bd010182fd91438f350ddf7586b2dbf2c3c2569d09", + "component_budget": { + "node_count": 4, + "item_count": 4, + "soft_pair_count": 6, + "serialized_chars": 9331, + "estimated_input_tokens": 2333, + "estimated_eval_tokens": 4922 + }, + "budget": { + "node_count": 4, + "item_count": 4, + "soft_pair_count": 6, + "serialized_chars": 9331, + "estimated_input_tokens": 2333, + "estimated_eval_tokens": 4922 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": false, + "analyst_result": { + "summary": "All four PRs target different models and different failure modes, so they are not duplicates of one another. The only code-change PR is the Whisper fix; the others are test/environment updates for unrelated models.", + "confidence": 0.97, + "canonical_issue_reason": null, + "canonical_pr_reason": "PR 47150 is the strongest representative of the set because it fixes an actual library code path in Whisper, whereas the others are model-specific test or environment adjustments. It is not a duplicate of the others, but it is the most substantive single change here.", + "best_issue_reason": null, + "best_pr_reason": "PR 47150 is the best standalone PR in this cluster: it changes core Whisper generation logic to address a concrete bug, while the other PRs only adjust tests or loading setup for different models.", + "soft_edge_verdicts": [ + { + "left": "pull_request:47122", + "right": "pull_request:47123", + "accept": false, + "reason": "Different models and different fixes: Edgetam drops an unexpected `original_sizes` kwarg in model forward, while MiniMax-M3-VL adjusts test loading with `offload_folder`. Not the same bug or change." + }, + { + "left": "pull_request:47122", + "right": "pull_request:47124", + "accept": false, + "reason": "Different subsystems and failure types: Edgetam fixes a backbone kwarg mismatch; Falcon-Mamba only updates expected generation outputs in tests." + }, + { + "left": "pull_request:47122", + "right": "pull_request:47150", + "accept": false, + "reason": "Unrelated model bugs: Edgetam input-kwarg cleanup versus Whisper max-length handling for segmented generation." + }, + { + "left": "pull_request:47123", + "right": "pull_request:47124", + "accept": false, + "reason": "Both are test-side changes, but for different models and different failure causes (`load_error` vs `output_mismatch`). No shared underlying bug." + }, + { + "left": "pull_request:47123", + "right": "pull_request:47150", + "accept": false, + "reason": "MiniMax-M3-VL test loading workaround is unrelated to Whisper generation-length logic." + }, + { + "left": "pull_request:47124", + "right": "pull_request:47150", + "accept": false, + "reason": "Falcon-Mamba expected-output refresh is a test expectation update, not the same as Whisper\u2019s source-code fix for max_length handling." + } + ] + }, + "evaluator_result": null, + "evaluator_used": false, + "retried": false, + "accepted_nontrivial_soft_edge": false, + "error_kind": null, + "error_message": null + } + ] +} diff --git a/snapshots/20260708T120111Z/analysis-runs/hybrid-model-20260708t120111z/manifest.json b/snapshots/20260708T120111Z/analysis-runs/hybrid-model-20260708t120111z/manifest.json new file mode 100644 index 0000000000000000000000000000000000000000..566f8d3e7c72af9385bbbc6503a31a30068ad415 --- /dev/null +++ b/snapshots/20260708T120111Z/analysis-runs/hybrid-model-20260708t120111z/manifest.json @@ -0,0 +1,14 @@ +{ + "analysis_id": "hybrid-model-20260708t120111z", + "artifacts": { + "hybrid": "snapshots/20260708T120111Z/analysis-runs/hybrid-model-20260708t120111z/analysis-report-hybrid.json", + "hybrid_reviews": "snapshots/20260708T120111Z/analysis-runs/hybrid-model-20260708t120111z/analysis-report-hybrid.llm-reviews.json" + }, + "channel": "canonical", + "model": null, + "published_at": "2026-07-08T12:14:09Z", + "repo": "huggingface/transformers", + "schema_version": 1, + "snapshot_id": "20260708T120111Z", + "variant": "hybrid" +} diff --git a/snapshots/20260708T120111Z/manifest.json b/snapshots/20260708T120111Z/manifest.json index d5fb1bba7b462b3dca46d591425436c0d06d12e5..3a2f161aa51aca4ba73f5c93f90b70b4c79540eb 100644 --- a/snapshots/20260708T120111Z/manifest.json +++ b/snapshots/20260708T120111Z/manifest.json @@ -32,6 +32,24 @@ "timeline_events": 985 }, "extracted_at": "2026-07-08T12:01:11Z", + "published_analysis": { + "canonical_analysis_id": "hybrid-model-20260708t120111z", + "runs": { + "hybrid-model-20260708t120111z": { + "analysis_id": "hybrid-model-20260708t120111z", + "artifacts": { + "hybrid": "snapshots/20260708T120111Z/analysis-runs/hybrid-model-20260708t120111z/analysis-report-hybrid.json", + "hybrid_reviews": "snapshots/20260708T120111Z/analysis-runs/hybrid-model-20260708t120111z/analysis-report-hybrid.llm-reviews.json" + }, + "channel": "canonical", + "manifest_path": "snapshots/20260708T120111Z/analysis-runs/hybrid-model-20260708t120111z/manifest.json", + "model": null, + "published_at": "2026-07-08T12:14:09Z", + "variant": "hybrid" + } + }, + "schema_version": 1 + }, "repo": "huggingface/transformers", "snapshot_id": "20260708T120111Z", "watermark": {